Contributors to V o l u m e 104

Article numbersare in parenthesesfollowingthe names of contributors.

Affiliationslisted are current.

RODIGER V. BATTERSBY (15), Abteilung SUSANNE FLYGARE (22), Department of

Klinische Biochemie, Medizinische Hoch- Pure and Applied Biochemistry, Chemical
schule Hannover, Hannover, Federal Re- Center, University of Lund, S-220 07
public of Germany Lund, Sweden
PERRY J. BLACKSHEAR (12), Diabetes Unit, ANNA J. FURTH (17), Department of Biol-
Massachusetts General Hospital, Boston, ogy, The Open University, Walton Hall,
Massachusetts 02114 Milton Keynes MK7 6AA, England
JOHN S. BLANCHARD (26), Department of OTHMAR GABRIEL (28), Department of Bio-
Biochemistry, Albert Einstein College of chemistry, Georgetown University
Medicine, Bronx, New York 10461 Schools of Medicine and Dentistry,
Washington, D.C. 20007
HILAR¥ BOLTON (17), Department of Biol-
ogy, The Open University, Walton Hall, JOHN E. GANDER (31), Department of Mi-
Milton Keynes MK7 6AA, England crobiology and Cell Science, University
WILLIAM M. BONNER (34), Laboratory of
of Florida, Gainesville, Florida 32611
Molecular Pharmacology, National Can- GARY C. GANZI (25), Millipore Corporation,
cer Institute, National Institutes of Ashby Road, Bedford, Massachusetts
Health, Bethesda, Maryland 20205 01730
ENRICO CAB1B (27), Laboratory of Biochem- GARY L. GILLILAND (23), Department of
istry and Metabolism, National Institute Physical Chemistry, Hoffman La-Roche,
of Arthritis, Diabetes, and Digestive and Inc., Nutley, New Jersey 07110
Kidney Diseases, National Institutes of DAVID GOLDMAN (30), Laboratory of Clini-
Health, Bethesda, Maryland 20205 cal Science, National Institute of Mental
GARY J. CALTON (24), Purification Engi- Health, National Institutes of Health, Be-
neering, Inc., 9505 Berger Road, Colum- thesda, Maryland 20205
bia, Maryland 21046 MILTON T. W. HEARN (9), St. Vincent's
ANDREAS CHRAMBACH (16), Section of School of Medical Research, Fitzroy, Vic-
Macromolecular Analysis, Laboratory of toria 3065, Australia
Theoretical and Physical Biology, Na- MARY J. HEEB (28), Department of Immu-
tional Institute of Child Health and Hu- nology, Research Institute of Scripps
man Development, National Institutes of Clinic, La Jolla, California 92037
Health, Bethesda, Maryland 20205 LEONARD M. HJELMELAND (16), Labora-
JoHN A. CUTTING (32), Houston, Texas tory of Vision Research, National Eye In-
77030 stitute, National Institutes of Health, Be-
DAVID R. DAVIES (23), Laboratory of Mo- thesda, Maryland 20205
lecular Biology, National Institute of Ar- CHRISTOPHER J. HOLLOWAY (15), Abteilung
thritis, Diabetes, and Digestive and Kid- Klinische Biochemie, Medizinische Hoch-
ney Diseases, National Institutes of schule Hannover, Hannover, Federal Re-
Health, Bethesda, Maryland 20205 public of Germany
ix
X CONTRIBUTORS TO VOLUME 104

VACLAV HO~EJ~f (14), Institute of Molecu- chemistry, University of Southampton,

lar Genetics, Czechoslovak Academy of Bassett Crescent East, Southampton S09
Sciences, 142 20 Praha 4, Czechoslovakia 3TU, England
KENNETH C. INGHAM (20), American Red ELBERT A. PETERSON (5), National Cancer
Cross Blood Services, Plasma Derivatives Institute, National Institutes of Health,
Laboratory, 9312 Old Georgetown Road, Bethesda, Maryland 20205
Bethesda, Maryland 20814 ITZHACK POLACHECK (27), Department of
GOTE JOHANSSON (21), Department of Bio- Clinical Microbiology, Hadassah Univer-
chemistry, Chemical Center, University sity Hospital, Ein Karem, Jerusalem,
of Lund, S-220 07 Lund, Sweden Israel
CHRISTOPH KEMPF (18), lnstitut far Hy- JENNIFER POTTER (17), Department of Biol-
giene und Medizinische Mikrobiologie, ogy, The Open University, Walton Hall,
Universitgit Bern, Friedbuehlstrasse 59, Milton Keynes MK7 6AA, England
Bern CH-3008, Switzerland JOHN D. PRIDDLE (17), Department of Gas-
RICHARD D. KLAUSNER (19), Laboratory of troenterology, Radcliffe Infirmary, Ox-
Biochemistry and Metabolism, National ford OX2 6HE, England
Institute of Arthritis, Diabetes, and Di- BERTOLD J. RADOLA (13), lnstitut fiir Le-
gestive and Kidney Diseases, National bensmitteltechnologie und Analytische
Institutes of Health, Bethesda, Maryland Chemie, Technische Universitiit Miin-
202O5 chen, D-8050 Freising-Weihenstephan,
JOACHIM KOHN (1), Department of Biophys- Federal Republic of Germany
ics, The Weizmann Institute of Science, FRED E. REGNIER (8), Department of Bio-
Rehovot 76100, Israel chemistry, Purdue University, West La-
PER-OLOF LARSSON (10, 22), Department of fayette, Indiana 47907
Pure and Applied Biochemistry, Chemical A. H. REISNER (29), CS1RO, Division of
Center, University of Lund, S-220 07 Molecular Biology, North Ryde N.S.W.
Lund, Sweden 2113, Australia
CHRISTOPHER R. LOWE (4), Department of JAIME RENART (33), lnstituto de Enzimolo-
Biochemistry, University of Southamp- gia y Patologia Molecular del C.S.I.C.,
ton, Bassett Crescent East, Southampton Facultad de Medicina, Universidad
S09 3TU, England Aut6noma de Madrid, Madrid 34, Spain
CARL R. MERRIL (30), Laboratory of Gen- JAMES S. RICHEY (11), Pharmacia Fine
eral and Comparative Biochemistry, Na- Chemicals, 800 Centennial Avenue, Pis-
tional Institute of Mental Health, Na- cataway, New Jersey 08854
tional Institutes of Health, Bethesda,
BENJAMIN RIVNAY (19), Department of
Maryland 20205
Membrane Research, The Weizmann In-
TALIA MIRON (1), Department of Biophys- stitute of Science, Rehovot 76100, Israel
ics, The Weizmann Institute of Science,
IGNACIO V. SANDOVAL (33), Section on En-
Rehovot 76100, Israel
zymes and Cellular Biochemistry, Labo-
KEAUS MOSBACH (2, 22), Department of ratory of Biochemistry and Metabolism,
Pure and Applied Biochemistry, Chemical National Institute of Arthritis, Diabetes,
Center, University of Lund, S-220 07 and Digestive and Kidney Diseases, Na-
Lund, Sweden tional Institutes of Health, Bethesda,
KURT NIESSON (2), Department of Pare and Maryland 20205
Applied Biochemistry, Chemical Center, SHMUEL SHALTIEL (3), Department of
University of Lurid, S-220 07 Lund, Chemical Immunology, The Weizmann
Sweden Institute of Science, Rehovot 76100,
JAMES C. PEARSON (4), Department of BiD- Israel
 CONTRIBUTORS TO VOLUME 104 xi

ANTHONY R. TORRES (5), Department of Jos VAN RENSWOUDE (18, 19), Laboratory
Laboratory Medicine, Yale University of Biochemistry and Metabolism, Na-
School of Medicine, New Haven, Con- tional Institute of Arthritis, Diabetes, and
necticut 06504 Digestive and Kidney Diseases, National
KLAUS UNGER (7), Institut far Anorgan- Institutes of Health, Bethesda, Maryland
ische Chemie und Analytisehe Chemie, 20205
Johannes-Gutenberg-Universitgit, 6500 C. TIMOTHY WEHR (6), Varian Associates,
Mainz, Federal Republic of Germany Walnut Creek Instrument Division, Wal-
MAR6ARET L. VAN KEUREN (30), The nut Creek, California 94598
Eleanor Roosevelt Institute for Cancer MEIR WILCHEK (1), Department of Biophys-
Research, 4200 East Ninth Avenue, Den- ics, The Weizmann Institute of Science,
ver, Colorado 80262 Rehovot 76100, Israel
 Preface

A dozen years after publication of "Enzyme Purification and Related

Techniques," Volume 22 of Methods in Enzymology, a number of novel
purification methods had been introduced which, along with subsequent
modifications of existing procedures, suggested the need for a supple-
ment. It is not that Volume 22 is outdated; almost all of the procedures
remain very much in use. Rather, it was felt that new, simpler, and possi-
bly more productive approaches that have been developed in the interim
should be made accessible in the Methods format.
Some of the subjects are entirely new. High-performance liquid chro-
matography, for example, is now being applied to proteins and, less fre-
quently, to enzymes. One major purification technique that has grown
steadily in importance since being introduced in Volume 22, affinity chro-
matography, merited a volume of its own (Volume 34); additional tech-
niques and a resum6 of existing ones are presented in detail here. The
solubilization of those proteins that are attached to membranes remains
difficult, but a variety of different approaches and detergents have been
introduced and are recorded in this volume. And a large number of meth-
ods of somewhat less protean specificity but general utility are also pre-
sented with the intent of rounding out our present knowledge of the state
of the art and science of protein purification.

WILLIAM B. JAKOBY

xiii
 METHODS IN ENZYMOLOGY

EDITED BY

Sidney P. Colowick and N a t h a n O. Kaplan

VANDERBILT UNIVERSITY DEPARTMENT OF CHEMISTRY
SCHOOL OF MEDICINE UNIVERSITY OF CALIFORNIA
NASHVILLE, TENNESSEE AT SAN DIEGO
LA JOLLA, CALIFORNIA

I. Preparation and Assay of Enzymes

II. Preparation and Assay of Enzymes
III. Preparation and Assay of Substrates
IV. Special Techniques for the Enzymologist
V. Preparation and Assay of Enzymes
VI. Preparation and Assay of Enzymes (Continued)
Preparation and Assay of Substrates
Special Techniques
VII. Cumulative Subject Index

XV
 METHODS IN ENZYMOLOGY
EDITORS-IN-CHIEF

Sidney P. Colowick Nathan O. Kaplan

VOLUME VIII. Complex Carbohydrates
Edited by ELIZABETH F. NEUEELD AND VICTOR GINSBURG

VOLUME IX. Carbohydrate Metabolism

Edited by WILLIS A. WOOD

VOLUME X. Oxidation and Phosphorylation

Edited by RONALD W. ESTABROOKAND MAYNARD E. PULLMAN

VOLUME XI. Enzyme Structure

Edited by C. H. W. HIRS

VOLUME XII. Nucleic Acids (Parts A and B)

Edited by LAWRENCE GROSSMAN AND KIVIE MOLDAVE

VOLUME XIII. Citric Acid Cycle

Edited by J. M. LOWENSTEIN

VOLUME XIV. Lipids

Edited by J. M. LOWENSTEIN

VOLUME XV. Steroids and Terpenoids

Edited by RAYMOND I . CLAYTON

VOLUME XVI. Fast Reactions

Edited by KENNETH KUSTIN

VOLUME XVII. Metabolism of Amino Acids and Amines (Parts A and B)

Edited by HERBERT TABOR AND CELIA WHITE TABOR

VOLUME XVIII. Vitamins and Coenzymes (Parts A, B, and C)

Edited by DONALD B. MCCORMICK AND LEMUEL D. WRIGHT

VOLUME XIX. Proteolytic Enzymes

Edited by GERTRUDE E. PERLMANN AND LASZLO LORAND
xvii
XVUl M E T H O D S IN E N Z Y M O L O G Y

VOLUME XX. Nucleic Acids and Protein Synthesis (Part C)

Edited by KIVIE MOLDAVE AND LAWRENCE GROSSMAN

VOLUME XXI. Nucleic Acids (Part D)

Edited by LAWRENCE GROSSMAN AND KIVIE MOLDAVE

VOLUME XXII. Enzyme Purification and Related Techniques

Edited by W1LLIAM B. JAKOBY

VOLUME XXIII. Photosynthesis (Part A)

Edited by ANTHONY SAN PIETRO

VOLUME XXIV. Photosynthesis and Nitrogen Fixation (Part B)

Edited by ANTHONY SAN PIETRO

VOLUME XXV. Enzyme Structure (Part B)

Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF

VOLUME XXVI. Enzyme Structure (Part C)

Edited by C. H. W. Hms AND SERGE N. T1MASHEFF

VOLUME XXVII. Enzyme Structure (Part D)

Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF

VOLUME XXVIII. Complex Carbohydrates (Part B)

Edited by VICTOR GINSBURG

VOLUME XXIX. Nucleic Acids and Protein Synthesis (Part E)

Edited by LAWRENCE GROSSMAN AND KIVIE MOLDAVE

VOLUME XXX. Nucleic Acids and Protein Synthesis (Part F)

Edited by K1VIE MOLDAVE AND LAWRENCE GROSSMAN

VOLUME XXXI. Biomembranes (Part A)

Edited by SIDNEY FLEISCHER AND LESTER PACKER

VOLUME XXXII. Biomembranes (Part B)

Edited by SIDNEY FLEISCHER AND LESTER PACKER

VOLUME XXXIII. Cumulative Subject Index Volumes I - X X X

Edited by MARTHA G. DENNIS AND EDWARD A. DENNIS
 METHODS IN ENZYMOLOGY xix

VOLUME XXXIV. Affinity Techniques (Enzyme Purification: Part B)

Edited by WILLIAM I . JAKOBY AND MEIR WILCHEK

VOLUME XXXV. Lipids (Part B)

Edited by JOHN M. LOWENSTEIN

VOLUME XXXVI. Hormone Action (Part A: Steroid Hormones)

Edited by BERT W. O'MALLEY AND JOEL G. HARDMAN

VOLUME XXXVII. Hormone Action (Part B: Peptide Hormones)

Edited by BERT W. O'MALLEY AND JOEL G. HARDMAN

VOLUME XXXVIII. Hormone Action (Part C: Cyclic Nucleotides)

Edited by JOEL G. HARDMAN AND BERT W. O'MALLEY

VOLUME XXXIX. Hormone Action (Part D: Isolated Cells, Tissues, and

Organ Systems)
Edited by JOEL G. HARDMAN AND BERT W. O'MALLEY

VOLUME XL. Hormone Action (Part E: Nuclear Structure and Function)

Edited by BERT W. O'MALLEY AND JOEL G. HARDMAN

VOLUME XLI. Carbohydrate Metabolism (Part B)

Edited by W. A. WOOD

VOLUME XLII. Carbohydrate Metabolism (Part C)

Edited by W. A. WOOD

VOLUME XLIII. Antibiotics

Edited by JOHN H. HASH

VOLUME XLIV. Immobilized Enzymes

Edited by KLAUS MOSBACH

VOLUME XLV. Proteolytic Enzymes (Part B)

Edited by LASZLO LORAND

VOLUME XLVI. Affinity Labeling

Edited by WILLIAM B. JAKOBY AND MEIR WILCHEK

VOLUME XLVII. Enzyme Structure (Part E)

Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF
XX METHODS IN ENZYMOLOGY

VOLUME XLVIII. Enzyme Structure (Part F)

Edited by C. H. W. HIRS AND SERGE N. T1MASHEFF

VOLUME XLIX. Enzyme Structure (Part G)

Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF

VOLUME L. Complex Carbohydrates (Part C)

Edited by VICTOR GINSBURG

VOLUME LI. Purine and Pyrimidine Nucleotide Metabolism

Edited by PATRICIA A. HOFFEE AND MARY ELLEN JONES

VOLUME LII. Biomembranes (Part C: Biological Oxidations)

Edited by SIDNEY FLEISCHER AND LESTER PACKER

VOLUME LIII. Biomembranes (Part D: Biological Oxidations)

Edited by SIDNEY FLEISCHER AND LESTER PACKER

VOLUME LIV. Biomembranes (Part E: Biological Oxidations)

Edited by SIDNEY FLEISCHER AND LESTER PACKER

VOLUME LV. Biomembranes (Part F: Bioenergetics)

Edited by SIDNEY FLEISCHER AND LESTER PACKER

VOLUME LVI. Biomembranes (Part G: Bioenergetics)

Edited by SIDNEY FLEISCHER AND LESTER PACKER

VOLUME LVII. Bioluminescence and Chemiluminescence

Edited by MARLENE A. DELUCA

VOLUME LVIII. Cell Culture

Edited by WILLIAM B. JAKOBYAND IRA PASTAN

VOLUME LIX. Nucleic Acids and Protein Synthesis (Part G)

Edited by KIVIE MOLDAVE AND LAWRENCE GROSSMAN

VOLUME LX. Nucleic Acids and Protein Synthesis (Part H)

Edited by KIVIE MOLDAVE AND LAWRENCE GROSSMAN

VOLUME 61. Enzyme Structure (Part H)

Edited by C. H. W. HIRS AND SERGE N. TIMASHEFF
 METHODS IN ENZYMOLOGY xxi

VOLUME 62. Vitamins and Coenzymes (Part D)

Edited by DONALD B. MCCORMICK AND LEMUEL D. WRIGHT

VOLUME 63. Enzyme Kinetics and Mechanism (Part A: Initial Rate and
Inhibitor Methods)
Edited by DANIEL L. PURICH

VOLUME 64. Enzyme Kinetics and Mechanism (Part B: Isotopic Probes

and Complex Enzyme Systems)
Edited by DANIEL L. PURICH

VOLUME 65. Nucleic Acids (Part I)

Edited by LAWRENCE GROSSMAN AND KIVIE MOLDAVE

VOLUME 66. Vitamins and Coenzymes (Part E)

Edited by DONALD B. MCCORMICK AND LEMUEL D. WRIGHI"

VOLUME 67. Vitamins and Coenzymes (Part F)

Edited by DONALD B. MCCORMICK AND LEMUEL D. WRIGHT

VOLUME 68. Recombinant DNA

Edited by RAY Wu

VOLUME 69. Photosynthesis and Nitrogen Fixation (Part C)

Edited by ANTHONY SAN PIETRO

VOLUME 70. Immunochemical Techniques (Part A)

Edited by HELEN VAN VUNAKIS AND JOHN J. LANGONE

VOLUME 71. Lipids (Part C)

Edited by JOHN M. LOWENSTEIN

VOLUME 72. Lipids (Part D)

Edited by JOHN M. LOWENSTEIN

VOLUME 73. Immunochemical Techniques (Part B)

Edited by JOHN J. LANGONE AND HELEN VAN VUNAKIS

VOLUME 74. Immunochemical Techniques (Part C)

Edited by JOHN J. LANGONE AND HELEN VAN VUNAKIS
xxii METHODS IN ENZYMOLOGY

VOLUME 75. Cumulative Subject Index Volumes XXXI, XXXII, and

XXXIV-LX
Edited by EDWARD A. DENNIS AND MARTHA G. DENNIS

VOLUME 76. Hemoglobins

Edited by ERALDO ANTONINI, LUIGI ROSSI-BERNARDI, AND EMILIA
CHIANCONE

VOLUME 77. Detoxication and Drug Metabolism

Edited by WILLIAM B. JAKOBY

VOLUME 78. Interferons (Part A)

Edited by SIDNEY PESTKA

VOLUME 79. Interferons (Part B)

Edited by SIDNEY PESTKA

VOLUME 80. Proteolytic Enzymes (Part C)

Edited by LASZLO LORAND

VOLUME 81. Biomembranes (Part H: Visual Pigments and Purple Mem-

branes, I)
Edited by LESTER PACKER

VOLUME 82. Structural and Contractile Proteins (Part A: Extracellular

Matrix)
Edited by LEON W. CUNNINGHAM AND DIXIE W. FREDERIKSEN

VOLUME 83. Complex Carbohydrates (Part D)

Edited by VICTOR GINSBURG

VOLUME 84. Immunochemicai Techniques (Part D: Selected Immunoas-

says)
Edited by JOHN J. LANGONE AND HELEN VAN VUNAKIS

VOLUME 85. Structural and Contractile Proteins (Part B: The Contractile

Apparatus and the Cytoskeleton)
Edited by DIXIE W. FREDERIKSEN AND LEON W. CUNNINGHAM

VOLUME 86. Prostaglandins and Arachidonate Metabolites

Edited by WILLIAM E. M. LANDS AND WILLIAM L. SMITH
 . ° .

METHODS IN ENZYMOLOGY XXIII

VOLUME 87. Enzyme Kinetics and Mechanism (Part C: Intermediates,

Stereochemistry, and Rate Studies)
Edited by DANIEL L. PURICH

VOLUME 88. Biomembranes (Part I: Visual Pigments and Purple Mem-

branes, II)
Edited by LESTER PACKER

VOLUME 89. Carbohydrate Metabolism (Part D)

Edited by WILLIS A. WOOD

VOLUME 90. Carbohydrate Metabolism (Part E)

Edited by WILLIS A. WOOD

VOLUME 91. Enzyme Structure (Part I)

Edited by C. H. W. HIRS AND SERGE N. T1MASHEFF

VOLUME 92. Immunochemical Techniques (Part E: Monoclonal Antibod-

ies and General Immunoassay Methods)
Edited by JOHN J. LANGONE AND HELEN VAN VUNAKIS

VOLUME 93. Immunochemical Techniques (Part F: Conventional Anti-

bodies, Fc Receptors, and Cytotoxicity)
Edited by JOHN J. LANGONE AND HELEN VAN VUNAKIS

VOLUME 94. Polyamines

Edited by HERBERT TABOR AND CELIA WHITE TABOR

VOLUME 95. Cumulative Subject Index Volumes 61-74 and 76-80

Edited by EDWARD A. DENNIS AND MARTHA G. DENNIS

VOLUME 96. Biomembranes [Part J: Membrane Biogenesis: Assembly

and Targeting (General Methods; Eukaryotes)]
Edited by SIDNEY FLEISCHER AND BECCA FLEISCHER

VOLUME 97. Biomembranes [Part K: Membrane Biogenesis: Assembly

and Targeting (Prokaryotes, Mitochondria, and Chloroplasts)]
Edited by SIDNEY FLEISCHER AND BECCA FLEISCHER

VOLUME 98. Biomembranes [Part L: Membrane Biogenesis (Processing

and Recycling)]
Edited by SIDNEY FLEISCHER AND BECCA FLEISCHER
xxiv METHODS IN ENZYMOLOGY

VOLUME 99. Hormone Action (Part F: Protein Kinases)

Edited by JACKIE n . CORBIN AND JOEL G. HARDMAN

VOLUME 100. Recombinant DNA (Part B)

Edited by RAY Wu, LAWRENCE GROSSMAN, AND KIVIE MOLDAVE

VOLUME 101. Recombinant DNA (Part C)

Edited by RAY Wu, LAWRENCE GROSSMAN, AND KIVIE MOLDAVE

VOLUME 102. Hormone Action (Part G: Calmodulin and Calcium-Binding

Proteins)
Edited by ANTHONY R. MEANS AND BERT W. O'MALLEY

VOLUME 103. Hormone Action (Part H: Neuroendocrine Peptides)

Edited by P. MICHAEL CONN

VOLUME 104. Enzyme Purification and Related Techniques (Part C)

Edited by WILLIAM B. JAKOBY

VOLUME 105. Oxygen Radicals in Biological Systems

Edited by LESTER PACKER

VOLUME 106. Posttranslational Modifications (Part A) (in preparation)

Edited by FINN WOLD AND KIVIE MOLDAVE

VOLUME 107. Posttranslational Modifications (Part B) (in preparation)

Edited by FINN WOLD AND KIVlE MOLDAVE
[1] AFFINITY CHROMATOGRAPHY 3

[1] A f f i n i t y C h r o m a t o g r a p h y
By M E I R W I L C H E K , 1 TALIA MIRON, and JOACHIM K O H N

Fifteen years after its introduction, affinity chromatography, la,2 a

method of purification based on biological recognition, has become a
major means for the purification of biologically active molecules. In 1974,
a volume in this series explored affinity chromatography, 3 and at that time
we were able to summarize the major literature in a table of fewer than
150 entries 4 and to present another 50 or so new examples. 3 The Medline
service now lists 1800 papers, published during a 3-year period, 1980-
1982, in which affinity chromatography is mentioned in the title.
Despite the enormous expansion of the field and the many materials
that have been purified, the directions for the investigation of methods
have not changed markedly. The same carriers, basically a n a g a r o s e , la
and the same activating reagent, cyanogen bromide, ~a,5 remain the pre-
dominant means for applying the technique. It is not surprising, therefore,
that the same concepts and difficulties remained. The problems reside
mainly in the lack of good protocols for efficient elution, particularly with
high-affinity interacting systems. Certain nonspecific interactions are in-
herent with cyanogen bromide systems due to ion-exchange and hydro-
phobic interactions, 6 both of which are obviously useful in their own right
and are covered elsewhere in this volume. In addition, there is the matter
of leakage of ligand from the carrier due to the unstable isourea linkage. 7
These two difficulties have been treated experimentally with some suc-
cess, as discussed subsequently.
It should be clear that means other than cyanogen bromide are being
made available, albeit at a slow pace. Indeed, other chapters in this vol-
ume treat such imaginative methods.
This chapter describes the chemistry of activation and coupling of
ligands and presents newer and effective methods for the preparation and
t Prepared during tenure as a Fogarty Scholar-in-Residence at the Fogarty International
Center of the National Institutes of Health, Bethesda, Maryland.
la p. Cuatrecasas, M. Wilchek, and C. B. Anfinsen, Proc. Natl. Acad. Sci. U.S.A. 61, 636
(1968).
2 p. Cuatrecasas and M. Wilchek, Biochem. Biophys. Res. Cornmun. 33, 735 (1968).
3 This series, Vol. 34, "Affinity Techniques."
4 M. Wilchek and W. B. Jakoby, this series, Vol. 34, p. 3.
5 R. Axrn, J. Porath, and S. Ernbiick, Nature (London) 214, 1302 (1967).
6 R. Jost, T. Miron, and M. Wilchek, Biochim. Biophys. Acta 362, 75 (1974).
7 M. Wilchek, T. Oka, and Y. J. Topper, Proc. Natl. Acad. Sci. U.S.A. 72, 1055 (1975).

Copyright© 1984by AcademicPress, Inc.

METHODSIN ENZYMOLOGY,VOL. 104 All rightsof reproductionin any formreserved.
ISBN 0-12-182004-1
 4 CHROMATOGRAPHY [1]

0
hydrolysis
O--C--NH~ carbamate (inert)

I!~ OH CNBr
~ 0 - - C_----N--
interehain
rearrange- !ii!i
ment
:::
ili|!i
NH
II
....
~ii:~ linear
imidocarbonate
(very slightly
reactive)
cyanate ester
(very reactive) cyclic
intrachain ~-- O\
*,- !~ ..C----
NH imidocarbonate
rearrange- iiiiO (slightlyreactive)
ment
J

~ O- ~IH
C-- NIl-Ligand
i}iii~
~}o~/C~--NH-Ligand •:,.~} II

isourea N-substituted N- substituted

derivative imidocarbonate earbamate
FIG. 1. Mechanism of activation of Sepharose by CNBr and subsequent coupling of
ligand. Heavy lines indicate major reaction pathways.

assay of affinity matrices. Specific ligands that have been coupled to

prepare affinity resins are so many and so varied that it has b e c o m e
unreasonable to present each one in detail. Nevertheless, knowledge of
the scope and variety of what has been done is obviously useful, even if
presented only in the form of representative examples. A selective list of
biomaterial that has been purified, along with comments on methods, is
appended.

C N B r Activation and Analysis

The preparation of an affinitychromatography column is a two-step
process. Reactive groups are firstintroduced into the chemically inert
polymeric resin, the process of activation. The second step is the covalent
attachment of a ligand to the activated resin. There is universal agreement
on the coupling condition: It must bc performed in a mild base. O n the
other hand, owing to limited knowledge of the mechanism of activation of
polysaccharides by CNBr, many different conditions for activation have
been described on the basis of "trial and error.''3
The study of the mechanism of activation of agarosc by C N B r (Fig. I)
became possible after the development of analytical methods for the de-
[1] AFFINITY CHROMATOGRAPHY 5

No, I-Is
C ~ N...CI'la

0
C
I
Et l
Etx [/Et

C
III
N
c
I

Ill
N
N
p-nitrophenyl N-cyanotriethyl- 1-cyano-4-dimethyl-
cyanate ammonium amino pvridinium
tetrafluoroborate tetrafluoroborate
FIG. 2. Molecular structure of p-nitrophenyl cyanatc (pNPC), N-cyanotriethylammo-
nium (CTEA), and 1-cyano-4-dimethylaminopyridiniumtetrafluoroborate (CDAP).

termination of each of the active and inactive species that are produced, a
Three major p/'oducts were found to be present: carbamates, imidocarbo-
nates, and cyanate esters. 9 The last two are chemically active. Cyanate
esters are stable below pH 4, whereas the imidocarbonates are most sta-
ble under basic conditions. By determining the amount of cyanate esters
and imidocarbonates present on the resin, based on these differences in
stability, it was possible to design and prepare carriers containing either
functional group, thereby enabling prediction of the coupling capacity of
activated resins toward ligands. Detailed analytical data of freshly acti-
vated Sepharose showed that 60-85% of the resin's total coupling capac-
ity is due to the formation of the cyanate esters (Fig. l) (cyanate ester
80%, imidocarbonates 20%).
Once this mechanism for activation by CNBr was established, it be-
came possible to develop new and more efficient activation procedures.~°
Aside from increased efficiency, activation reagents based on CNBr have
been developed that are nonvolatile solids and can bc stored and safely
handled without use of a fume hood, providing an attractive alternative to
the highly hazardous CNBr 11",b (Fig. 2).

Activation Procedures
Most C N B r activation procedures are modifications of the originally
described method, 5 in which inorganic bases, such as sodium hydroxide
s j. Kohn and M. Wilchek,Anal Biochem. 115, 375 (198i).
9 j. Kohn and M. Wilchek, Enzyme Microb. Technol. 4, 161 (1982).
l0 j. Kohn and M. Wilchek, Biochem. Biophys. Res. Commun. 107, 878 (1982).
ii, j. Kohn, R. Lengcr, and M. Wilchck, Appl. Biochem. Biotech. 8, 227 (1983).
nb j. Kohn and M. Wilchek, FEBS Left. 154, 209 (1983).
6 CHROMATOGRAPHY [I]

Activation
in
presence of ~ -OH + OH- O-
+ CNBr
~
~_
OCN + Br"

NaOH
reactive species

Activation CNBr +
in + ~ / + Br"
presence of y ~-OCN + TEA.HBr
~TEA /N.,,
Et l E t
TEA Et
reactive species

FIo. 3. Comparison of the mode of action of NaOH and triethylamine (TEA). NaOH
acts on the resin, causing an increase in the resin's nucleophilicity by formation of alkoxide
ions. TEA is proposed as operating on CNBr by increasing the electrophilicity of the cyano
moiety by complex formation.

or sodium carbonate, are used. 12 A new approach to the reaction of

agarose with cyanogen bromide involves the use of triethylamine (TEA)
as the "cyano-transfer" reagent at neutral pH ~° (Fig. 3). The procedure
requires less than 10% of the usual amount of cyanogen bromide, and the
resultant activated resins are free of imidocarbonates and carbamates,
containing only active cyanate esters. Extremely high coupling capacities
(75 gmol of ligand per gram of wet Sepharose 4B) are obtainable.

Method with Cyanogen Bromide and Triethylamine 1°

Wet Sepharose 4B (Pharmacia; 10 g) is washed and suspended in 10 ml
of 60% acetone. The mixture is cooled to -15 °. Depending on the desired
degree of activation (Fig. 4), the required volume of CNBr solution (I M
in acetone) is added to the agarose suspension. With cooling and vigorous
stirring, an equal volume of TEA solution (1.5 M in 60% acetone)
is added dropwise during 1-3 min. Best results are achieved with a
final molar ratio of CNBr to TEA of about 1 : 1.5. After 3 min, the entire
reaction mixture is quickly poured into 100 ml of ice-cold washing me-
dium (acetone : 0.1 N HC1, 1 : 1). The resin may be stored in washing
medium at about 0° for about I hr without significant loss of activation.
(During this period the coupling capacity of the activated resin can be
determined.) The reaction can also be performed at 5° without significant
loss in yield or purity.~°

12 S. C. March, I. Parikh, and P. Cuatrecasas, Anal. Biochem. 60, 149 (1974).

[1] AFFINITY CHROMATOGRAPHY 7

96 I"
en t~, S
je
o 80 previously
unavailable
"~ superactivated
64 resins

~ 48

32 . . . . . . . . . . . . . . . . . . . . . h--
tu S t r o n g activation
~>" __

-5 16 ed 'Moderate activation
/. . . . . . . . . .

Weak activation
0 I I I I I I I 1 1
0 10 20 30 40 50
mg CNBr/g drained Sepharose 4B

Fro. 4. Incorporation of active cyanate esters into Sepharose 4B as a function of the

concentration of CNBr.

Method with p-Nitrophenyl Cyanate ua'b

Sepharose 4B (10 g), washed successively with acetone : H20 (1 : 4,
v/v), acetone : n 2 0 (1 : 1), and, finally, acetone : n 2 0 (7 : 3), is resuspended
in 10 ml of acetone : H20 (7 : 3) and cooled to 0°. To this solution, the
desired volume of pNPC reagent [2.3 g of p-nitrophenyl cyanate (Sigma;
pNPC) dissolved in 10 ml of acetone to yield 1.4 M pNPC] is added with
vigorous stirring. This addition is followed by an equal volume of neat
TEA. In this manner, a fivefold molar excess of base over activating agent
is attained. After 10 min, the entire reaction mixture is poured into 200 ml
of ice-cold acid-treatment medium (acetone : 0.5 M HC1, 1 : 1), and al-
lowed to stand for 20 min at 0°. Thereafter, the resin is washed with ice-
cold acetone : H20 (1 : 1), followed by ice-cold water, and used for cou-
pling as described for CNBr-activated resins.l.3 For immediate coupling,
the resin is washed with cold 60% acetone, followed by quick washings
with cold 30% acetone, cold water, and the coupling medium.

Method with N-Cyanotriethylammonium (CTEA) and

1-Cyano-4-dimethylaminopyridinium Tetrafluoroborate (CDAP)
Drained Sepharose 4B (10 g) is washed sequentially with water, 30%
acetone, and 60% acetone and resuspended in 10 ml of 60% acetone. The
8 CHROMATOGRAPHY [ 1]

TABLE I
AMOUNT OF ACTIVATING AGENT REQUIRED FOR ACTIVATION OF 10 g OF
DRAINED SEPHAROSE4B

CTEA activa- CDAP activa-

Approximate tiona tiona
coupling
Degree of capacity CTEA b TEA b CDAP~ TEA b
activation (/~mol llgand/g resin) (mg) (ml) (mg) (ml)

Weak 5 60 0.6 25 0.2

Moderate 15 180 1.8 75 0.6
Strong 30 360 3.6 150 1.2

"The illustrative procedure was used as described.

b TEA, triethylammonium; TEA, triethylamine, 0.2 M aqueous solution.
c For convenient use, a stock solution of 1-cyano-4-dimethylaminopyridi-
nium tetrafluoroborate (CDAP) in acetonitrile (0.1 g/ml) was prepared.
This solution is stable at 4* for over a month.

suspension is cooled to 0°. Depending on the desired degree of activation,

the required amount of activating agent is added, followed by a corre-
sponding amount of TEA (see Table I). The TEA solution (0.2 M in water)
is added dropwise with vigorous stirring. After 2 min, the entire reaction
mixture is rapidly transferred into 100 ml of ice-cold washing medium
(acetone : 0.1 N HCL, 1 : 1). The resin may be stored in this manner for
over 1 hr without loss of activation. For coupling, the resin is washed with
a large volume of cold water followed by one quick wash with the cou-
pling medium. Coupling of ligand to CTEA- or CDAP-activated resins is
performed as described for CNBr-activated resins.
For prolonged storage of CNBr-activated resins, the resin is prepared
by extensive washing with storage medium (acetone:dioxane:H20,
60 : 35 : 5). When suspended in storage medium and stored airtight at -15
to - 2 0 °, the active cyanate ester content decreases at an approximate rate
of 10% per month. After prolonged periods of storage, the resin may be
reswollen for 5 min in cold washing medium and washed as described
above for immediate coupling.

Determination and Characterization of Activated Resins

Determination of the expected coupling capacity of the resin is some-
times crucial for successful application of affinity chromatography
columns. The determination of cyanate esters and imidocarbonates can
[1] AFFINITY CHROMATOGRAPHY 9

be performed in about an hour. The results are accurate and reproducible

and provide a quick method for resin characterization. 8

Determination of Total Nitrogen Content

The following method does not require the laborious Kjeldahl proce-
dure or Dumas microanalysis. A weighted amount of dry resin, about 15
mg, is placed in a 50-ml measuring bottle. H2SO4 (96%), 0.5 ml, is added,
and the mixture is heated in an oven to between 100° and 120° for 15 min.
Thereafter, 0.2 ml of 30% H202 is added slowly (vigorous reaction). The
mixture clears, and heating is continued for an additional 90 min. The
colorless solution is diluted to 50 ml with 0.2 M sodium acetate at pH 5.0.
Aliquots of known volume are used for determination of ammonia by the
ninhydrin reaction (see below); the amount measured is equal to the total
nitrogen content of the resin.

Determination o f Ammonia by Ninhydrin Reaction

Preparation of hydrindantin reagent: Hydrindantin (880 mg) and
ninhydrin (4 g) are dissolved in 100 ml of Methyl Cellusolve. Unless
stored airtight at - 2 0 °, the reagent will degrade during the course of a few
days. To I ml of sample solution, containing between 0.05 and 0.8/~mol of
NH4*, is added 1 ml of 2 M sodium acetate at pH 5.5 and 1.5 ml of
hydrindantin reagent. After mixing, the solutions should be a light red.
The stoppered test tubes are heated to 100° for 20 min. The contents of
each tube are diluted without prior cooling to 10 ml with cold formalde-
hyde diluent (0.5% solution of formaldehyde in isopropanol : water, I : 1).
The diluent reacts with excess hydrindantin so that a blue ninhydrin color
appears within a few seconds. Absorbance is measured at 570 nm, and the
amount of NH4 ÷ in the sample is determined from a calibration curve.

Determination of lmidocarbonates 8
Of all the nitrogen derivatives present on activated resins, only the
imidocarbonates will release ammonia on mild acid hydrolysis. Thus, the
amount of ammonia measured is equal to the amount of imidocarbonate
on the resin.
A sample of activated resin containing about 1-20/zmol of imidocar-
bonates is placed in a 25-ml measuring bottle. Acid (5 ml of 0.1 N HC1) is
added, and the mixture is heated at 40 ° for 30 min or allowed to remain at
room temperature for 60 min. The volume is brought to 25 ml with 0.2 M
sodium acetate at pH 5.5. After mixing, the suspension is allowed to settle
10 CHROMATOGRAPHY [ 1]

and aliquots of the clear supernatant liquid are used for determination of
NH4 ÷ by the ninhydrin reaction) 3

Determination of Cyanate Esters 8

Cyanate esters are determined by the use of pyridine and either di-
methylbarbituric acid or barbituric acid. With the possible exception of
traces of triazines, cyanate esters are the only species present on acti-
vated resin giving rise to a characteristic purple color with barbiturates in
pyridine. The reaction is highly sensitive (e588 = 137,000 liter mo1-1 cm -~)
so that as little as 5 nmol of cyanate ester can be determined. The method
is also applicable to cyanuric chloride-activated polysaccharides. The
color reaction can be used as a quick, qualitative test for the presence of
activation of agarose; for quantitative determinations of cyanate esters
directly on the activated resin; and as a quick test for the presence of
excess CNBr in washings after activation.
Qualitative Test Reagent. Pyridine (14 ml) is mixed with 6 ml of water
and 50 mg of barbituric acid or N,N'-dimethylbarbituric acid (Chemical
Dynamics Corporation, South Plainfield, NJ), resulting in a clear, color-
less solution that darkens slowly. For qualitative tests, coloration of the
test reagent is not important. To 10-20 mg of dry polysaccharide, or to
0.1-1.0 ml of swollen polysaccharide, 1-2 ml of qualitative reagent is
added. After shaking slightly, the presence of less than 5 nmol of cyanate
ester can be detected by the formation of a red-purple color, which
develops within 30 sec and reaches a maximum after l0 min.
Quantitative Test Reagent. N,N'-Dimethylbarbituric acid (500 mg;
recrystallized from water) is suspended in 5 ml of water. Cold, freshly
distilled pyridine (45 ml) is added and results in a clear, colorless solution.
Since some discoloration is evident even after storage at - 2 0 °, it seems
best to prepare the required amount of reagent mixture immediately prior
to use.
Quantitative Determination of Cyanate Esters. To 5-20 mg of dry,
activated resin, or an equivalent amount of wet resin, 5 ml of the quantita-
tive test reagent is added. The mixture is warmed to 40 ° for 25 rain in a
closed test tube with vigorous stirring. After 25 min, the mixture is diluted
to a convenient volume with water. For 10 mg of freshly activated resin,
dilution to 250-500 ml is usual in order to reduce the absorbance of the
purple solution to between 0.5 and 1.0. Filtered aliquots of the dilutions
are used to measure the absorbance at about 588 nm (e588 = 137,000).

13 S. Moore and W. H. Stein, J. Biol. Chem. 211, 907 (1954).

[1] AFFINITY CHROMATOGRAPHY 11

Coupling Capacity of an Activated Resin

Based on the assumption that coupling of ligand to cyanate esters and
to imidocarbonates occurs simultaneously and independently, the total
coupling capacity of a monovalent ligand to an activated resin is predicta-
ble. It is related to the amounts of cyanate esters and imidocarbonates
present on the resin according to the following equation.

Total coupling capacity = A(cyanate esters) + B(imidocarbonates)

Here, A and B represent the respective yields for the respective products
of the coupling procedure. Based on values A and B of 0.8 and 0.15,
respectively, the equation can be used to predict the coupling capacities
of specific samples of activated agarose.

N e w e r Methods of Activation

Despite its popularity, the CNBr method of activation has several

drawbacks. For example, coupling of amines to activated polysaccharides
introduces N-substituted isourea bonds that are not completely stable,
particularly in the presence of nucleophiles. 7 This small leakage of ligand
can lead to erroneous findings and, over a period of time, results in the
loss of capacity for adsorption of the desired protein. The isoureas are
positively charged at physiological pH (pK = 9.4), allowing such columns
to also exhibit ion-exchange properties that can interfere with the biospe-
cificity of the adsorbent. In order to eliminate the difficulties inherent in
the CNBr method, and provide other possibilities for binding ligands to
carrier, several methods have been developed. One of these uses CNBr to
activate polypyridine or other carriers containing the pyridine ring 14; this
is the reverse reaction of that used in the determination of the cyanic
esters. 15
Other new methods are based on well-established sugar chemistry in
which activation is performed in anhydrous media. Sulfonyl chloride is
used for activation in one such procedure, which is described in this
volume [2]. 16 For most of the other preparative procedures, the reactive
species are activated carbonates or carbamates. 17,18 Upon reaction with

14 F. Pittner, T. Miron, G. Pittner, and M. Wilchek, J. Solid Phase Biochem. 5, 167 (1980).
x5 F. Pittner, T. Miron, G. Pittner, and M. Wilchek, J. Am. Chem. Soc. 102, 2451 (1980).
t6 K. Nilsson and K. Mosbach, Biochem. Biophys. Res. Commun. 102, 449 (1981).
l: G. S. Bethell, J. S. Ayers, and W. S. Hancock, J. Biol. Chem. 254, 2572 (1979).
18 M. Wilchek and T. Miron, Biochem. Int. 4, 629 (1982).
12 CHROMATOGRAPHY [1]

OH + ~ 0 ~N
_ -- -- _ _ ~ O - - C - -

SCHEME 1.

amines, stable and uncharged carbamate (urethane) derivatives are ob-

tained. The reagents include 1,1'-carbonyldiimidazole, 17 p-nitrophenyl-
chloroformate, TMN-hydroxysuccinimidochloroformate, 18 and tri- or pen-
tachlorophenylchloroformate. TM With the last three, the extent of
activation and coupling can be measured spectrophotometrically.

Activation with 1,1'-Carbonyldiimidazole 17,19(Scheme 1)

Sepharose CL-6B (3 g of moist cake) is washed sequentially with 20 ml
each of water, dioxane-water (3 : 7), dioxane-water (7 : 3), and dioxane
and is suspended in 5 ml of dioxane. 1,1 '-Carbonyldiimidazole (120 mg) is
added, and the suspension is shaken at room temperature for 15 min. It is
washed with 100 ml dioxane and used immediately.
Cross-linked agarose was used, but normal agarose preparation is
equally effective. The amount of reagent can be varied depending on the
extent of activation required; up to 3 mmol of active groups can be intro-
duced per gram of dry matrix. The isolated matrix is stable under anhy-
drous conditions and may be stored in dioxane. These activated carriers
are also available commercially from Pierce Chemical Company.

Activation with Chloroformates TM (Scheme 2)

Sepharose (4B or CL-4B) is washed with acetone-water mixtures of
increasing concentration in acetone (1 : 3, 1 : 1, and 3 : 1) and, finally, with
dry acetone. The gel (10 g) is suspended in 10 ml of dry pyridine at 4°; the
chloroformate, dissolved in dry acetone, is added dropwise over a period
of 2 min. The reaction mixture is stirred for an additional 15-30 min at 4 °.
The activated polymer is washed sequentially with acetone, methanol,
and ether and then is dried in air. The activated carrier may be stored at 4 °
as a dry powder or in anhydrous solvents, such as isopropanol, acetone,
or dioxane. Fresh preparations can be used without drying with ether but
are washed with water before coupling.
Both Sepharose and cross-linked Sepharose can be used. However,
cross-linked Sepharose appears to be more stable in many organic sol-
vents. The extent of activation depends on the amount of chloroformate
~9M. T. W. Hearn, E. L. Harris, G. S. Bethell, H. S. Hancock, and J. A. Ayers,
J. Chromatogr. 218, 509 (1981).
[1] AFFINITY CHROMATOGRAPHY 13

o"

0 0

00 1 . ~ 0 ~ o

o o
+ + 4-

oc~~ o~ o~
+ + +
I:1:I
o o

l o~
Z

o~-~o I
e4
0 0 0
I I I
o=r~
I I I
0 0 0

c~
Z

o ~ : ~ o U
I
0 0 0
I i I
0--~ 0=~ 0 = r,.P
I I J

+ + +
14 CHROMATOGRAPHY [1]

added. Carrier containing up to 1.8 mmol of active groups per gram of

polymer can be obtained. Yields of up to 30% can be assumed, based on
the amount of chloroformate added. The swelling properties of the matrix
in water change only nominally with low levels of substitution, although
increasing substitution leads to decreased swelling of the gel. These acti-
vated carriers are commercially available from Sigma Chemical Com-
pany.
Activation of cellulose with the same reagents has also been de-
scribed; activation was performed in dioxane at 60° or for prolonged
periods at 25°. 20

D e t e r m i n a t i o n o f A c t i v e Groups on the Carriers

Active groups present on carriers prepared by reaction with p-ni-

trophenylchloroformate may be quantitated after hydrolysis of a weighed
sample with 0.2 N NaOH for 15 rain at room temperature. The free
p-nitrophenylate is measured spectrophotometrically at 400 nm (e =
17,000). The amount of hydroxysuccinimide on columns prepared with
hydroxysuccinimide chloroformate is determined by nitrogen analysis or,
after basic hydrolysis (0.2 N NH4OH), the absorbance of the supernatant
liquid may be measured spectrophotometrically at 261 nm (e = 9700). 2L22
Active groups on preparations with trichlorophenylchloroformate are es-
timated either by chlorine analysis or spectrophotometrically at 310 nm
(e = 5000) as a function of trichlorophenol released after hydrolysis with
0.2 N NaOH.
No direct method is available for the determination of the irnidazole-
activated resins. Nitrogen analysis and titration have been used to con-
firm the presence of imidazole on carriersJ 7

Polyacrylhydrazidoagarose
Polyacrylhydrazidoagarose (PAHOS), prepared by the CNBr method,
was described in this series. 23 Since then, this matrix has been applied to
purification, particularly in the fields of immunology24 and lectin recep-
tors .25 Although this material has greater stability when compared to nor-
20j. Drobnik, J. Labsky, H. Kudlvasrova,U. Sandek, and F. Svec, Biotechnol. Bioeng.
~¢4, 487 (1982).
2mT. Miron and M. Wilchek,Anal. Biochem. 126, 433 (1982).
P. M. Abdalla, P. K. Smith, and G. Royer, Biochem. Biophys. Res. Commun. 87, 736
(1979).
23M. Wilchekand T. Miron, this series, Vol. 34, p. 72.
24D. W. Sears, S. Young, P. H. Wilson, and J. E. Christiaansen,J. Immunol. 124, 2641
(1980).
25R. Lotan and R. L. Nicolson,Biochim. Biophys. Acta 559, 329 (1979).
[1] AFFINITY CHROMATOGRAPHY 15

mal CNBr-activated agarose, some leakage of ligands was observed.

Owing to the general usefulness of hydrazide matrices, it appeared worth-
while to develop a better procedure for their preparation. Periodate was
used to achieve an oxidized agarose from which hydrazides can be
formed. 26 This system is presented separately, since it is not a CNBr
preparation, nor are hydrazide columns, as such, a new development.

Preparation of PAHOS
Washed Sepharose 4B (10 g) is suspended in 30 ml of freshly prepared
0.25 M sodium periodate. The suspension is slowly stirred at room tem-
perature for 3 hr in the dark; the oxidized Sepharose is then washed with
cold water to remove excess periodate and its products. The gel is sus-
pended in three volumes of polyacrylhydrazide solution in water (0.1-
0.5%). Coupling is allowed to proceed overnight in the dark at room
temperature with slow stirring. The conjugate is washed with 0. l M so-
dium chloride until the washings are free of color after testing with trini-
trobenzenesulfonic acid (TNBS). 26 The conjugate is reduced with 0.3 M
sodium borohydride in 0.5 M Tris-chloride at about pH 8 for 3 hr at room
temperature. The reduced gel is washed with water on a sintered-glass
funnel. Although it is possible to prepare a carrier containing as much as
125/zmol of hydrazide per milliliter of Sepharose, it is suggested that
derivatization be limited to 15-25 /zmol of hydrazide per milliliter for
normal usage. The PAHOS prepared by this method has many advantages;
it is stable and colorless and is free of functional groups, such as carba-
mates and carbonates, that are produced by CNBr activation. These prep-
arations may be stored at 4° for at least 2 years.

Derivatization of PAHOS
Matrix with Spacers. PAHOS can be used directly for coupling of
compounds containing aldehyde groups such as glutaraldehyde or pe-
riodate-oxidized nucleic acids or can be modified further by reaction with
a variety of functional groups that have been described, z3 Carriers bearing
carboxyl groups, thiols, imidazoles, phenols, and amines can be pre-
pared. These reactive functional groups can be used to couple ligand and
protein via amide, thioester, and diazonium linkages among others.
Aldehyde-Containing PAHOS. PAHOS bearing a carbonyl group is
prepared by suspending the resin in three volumes of 7% glutaraldehyde
in water and slowly stirring for 2 hr at room temperature. Excess glu-
taraldehyde is removed by washing the gel with cold water on a sintered-
z6 T. Miron and M. Wilchek, J. Chromatogr. 215, 55 (1981).
16 CHROMATOGRAPHY [1]

glass funnel until no further reaction of glutaraldehyde with nitrophenyl-

hydrazide is detected. 26
Carboxyl-Containing PAHOS. Free carboxyl-containing resins are
prepared by acylation of the free hydrazide groups with an excess of
succinic anhydride. The resulting succinylhydrazides provide carboxyl
groups with long spacer arms. Succinic anhydride is usually added as 0.5
mmol per milliliter of packed hydrazido-Sepharose in an equal volume of
water. The pH of the reaction mixture is maintained at 6.0 by addition of 1
N NaOH and allowed to proceed at room temperature for 3 hr with slow
stirring. Completion of acylation is assayed with TNBS. The resin is
washed with 0.1 N HC1 and water.
Amino-Containing PAHOS. Compounds containing primary amino
groups may be coupled to the aldehyde resin by incubation at pH 7.0 for
16 hr at room temperature or, alternatively, by coupling to carboxyl deriv-
atives at pH 5.0 in the presence of a water-soluble carbodiimide. Coupling
to carboxyl derivatives is also performed under anhydrous conditions
using dioxane and dicyclohexylcarbodiimide. A large excess of diamine is
used to decrease cross-linking in cases in which free amino-terminal
groups are required. Such amino groups may be useful for further substi-
tution with different ligands. Methods for the use of carbodiimide re-
agents have been detailed. 27
Sulfhydryl-Containing PAHOS. Sulfhydryl groups are introduced by
treating the resin with N-acetylhomocysteine thiolactone. The thiolactone
(1 g) is added to a cold suspension of 10 ml of hydrazido-Sepharose in 20
ml of 1 M NaHCO3. The reaction mixture is stirred at 4° for 16 hr, and the
product is washed extensively with water and 0.1 N NaC1. The prepara-
tion can also be made by coupling of dithiodiglycolic acid in aqueous
solution using a water-soluble carbodiimide or in dioxane using dicyclo-
hexylcarbodiimide as described above for carboxyl coupling. After reac-
tion, the disulfide resin is reduced with 0.2 M mercaptoethanol for 30 min.

Characterization of Hydrazides and Derivatives

Free hydrazide or free amino groups on conjugates may be assayed by
reaction with an excess of TNBS. The quantity of unreacted TNBS is
measured and subtracted from its initial concentration. Samples of
washed gel (200 rag) are suspended in I ml of I% TNBS, to which is added
1 ml of saturated sodium tetraborate. The suspension is stirred for 30 rain
at 24° and washed with 23 ml of 0.2 M NaC1. Aliquots of the washings, 100
t~l, are incubated with 1.0 ml of 0.2 M adipodihydrazide and 0.9 ml of

27 S. Bauminger and M. Wilchek, this series, Vol. 70, p. 151.

[I] A F F I N I T Y CHROMATOGRAPHY 17

0
IF
,N~O--C--O~N02 --

0 O

Cl

c1

SCHEME3. Structure of active carriers.

saturated borate for 15-20 min. Absorbance is recorded at 500 nm (e =

16,500) and used for calculating the quantity of TNBS that had reacted
with adipodihydrazide.
Free sulfhydryl groups on the resin are determined with an excess of
5,5'-dithiobis(2-nitrobenzoic acid). Samples of washed gel (200 mg) are
suspended in 1 ml of 50 mM Ellman reagent in 0. I M potassium phos-
phate, pH 7.5, for 30 rain at room temperature and then filtered. The
absorbance of the filtrate is recorded (e4~2 = 13,600) and used for calculat-
ing the concentration of free sulfhydryl on the resin.

Coupling of Ligands to the Carriers

The major emphasis here has been on activation of the resin matrix.
The second step in the process is that of coupling of ligands to the matrix.
The structure formed on coupling of amino-containing ligands to CNBr-
activated agarose is presented in Fig. I; the products of coupling to car-
bonyldiimidazole and chloroformate-activated agarose are presented in
Scheme 3.
18 CHROMATOGRAPHY [ 1]

The coupling of ligands is usually performed in 0.1-0.2 M sodium

bicarbonate at pH 8-9 for 4 hr at room temperature or overnight at 4 °.
Precise directions for coupling can be found in Volume 34 of this series
but are briefly summarized below.

Coupling Amino-Containing Ligands to CNBr or

Carbonate-Activated Agarose 3
Alkylamines, amino acids, peptides, and proteins may be coupled in
0.1 M sodium bicarbonate (pH 8.0-9.0) for 20 hr at 4° with gentle agita-
tion. The residue is filtered and washed with 0.1 M NaHCO3 and with
distilled water. Excess active groups are removed by treatment for 5 min
with 0.01 N NaOH or by reaction for 1 hr at room temperature with either
0.1 N NH4OH or 0.1 M ethanolamine at pH 9.

Coupling of Nucleotides and RNA to PAHOS

Nucleosides, nucleotide phosphate coenzymes, and RNA possessing
vicinal free hydroxyl groups are oxidized with periodate and coupled to
PAHOS as described for other hydrazides. 25,26

Coupling of Proteins and Lectins to PAHOS

The aldehyde-containing resin, obtained after glutaraldehyde treat-
ment, is suspended in two volumes of 0.1 M potassium phosphate at pH
7.5, containing the appropriate protein (enzyme, lectins) at concentra-
tions of 5-10 mg/ml, and stirred slowly for 16 hr at 4 °. Conjugates are
washed with cold 10 mM potassium phosphate-0.14 M NaC1 at pH 7.4.
Excess aldehyde groups are reduced by addition of 0.5 mg of NaBH4 per
milliliter of solution for 2 hr at 4° or by treatment with 0.05 M hydrazine
hydrate at pH 8.0 for 2 hr at room temperature.

Determination of Coupled Ligand

The extent of ligand coupling can be estimated by several methods,
including total hydrolysis, radioactivity, spectroscopy, and the like. The
various different methods have been compiled in this series. 3

Adsorption and Elution

Conditions for the purification of specific protein must be tailored to
the properties of that protein; there is no all-purpose protocol that is
applicable. It is clear, then, that those conditions that allow complex
formation between the ligand and the protein must be manipulated with
[1] AFFINITY CHROMATOGRAPHY 19

respect to ionic strength, temperature, pH, concentration, presence of

metals, and flow rate. 3,28For high-affinity systems, any of the carriers and
even small amounts of ligand bound will be satisfactory. For low-affinity
systems, large amounts of ligand must be bound to carrier. 29'30

Elution
Elution of the adsorbed material from the column is the most impor-
tant step in successful purification by affinity chromatography. Again,
there is no protocol that serves as an example for elution, and individual
conditions must be improvised. Several general strategies for elution,
however, are commonly being used with success. These procedures in-
clude both specific and general elution conditions.
Specific elution refers to competition by hapten, ligands, or inhibitor
in solution. These are the best conditions for elution and, when success-
ful, strongly suggest that real affinity, bioaffinity, is involved. On the
other hand, if apparently specific ligands fail to elute the desired protein,
nonspecific means can be applied. Among the latter that have been used
successfully are solvent or buffer changes, pH or ionic strength altera-
tions, the application of chaotropic reagents, electrophoretic desorption,
temperature changes, reversible denaturation, and destruction of the
spacer arm. 31
A good practice in attempting elution of the adsorbed material is the
initial examination of several means of nonspecific elution. Nonspecific
elution can be considered as an extra step in washing the column if the
required compound is not eluted, i.e., it serves to remove contaminating
proteins. Therefore, elution washing should be tried immediately after the
sample has been applied and before a specific eluent is used. The washing
should be started with a high salt concentration in the buffer so that
species that bind because of their ion-exchange properties will tend to be
eluted. On the other hand, it may occasionally be advisable to start with a
very low salt concentration so as to elute species bound through hydro-
phobic interaction. If changes in ion strength are ineffective, different
buffers could be tried--e.g., a shift from a Tris to a borate buffer. Purifica-
tion is achieved by removing contaminating protein by these efforts or,
conversely, by selective removal of the desired protein. The next step in
such trials would be to change the pH, beginning with small changes

2a p. O'Carra, S. Barry, and T. Griffin, this series, Vol. 34, p. 108.

29 I. M. Chaiken, Anal. Biochem. 97, 1 (1979).
3o M. Wilchek and C. S. Hexter, Methods Biochem. Anal. 23, 347 (1976).
31 D. R. Absolom, Sep. Purif. Methods 10, 239 (1981).
20 CHROMATOGRAPHY [1]

followed by using acid or base, e.g., with 0.1 N acetic acid or 0.5 M
NHaOH, respectively; such experiments are predicated upon stability
requirements being met under these conditions. If elution of the desired
species does not occur, treatment with ethylene glycol or dioxane or with
an organic acid, such as propionic acid, may result in elution.
Specific elution is most desirable, so that competitive inhibitors, co-
factors, or haptens must be tried. If neither specific inhibitors nor the
nonspecific steps that have been outlined are effective, denaturing condi-
tions using urea or guanidine are in order; these reagents sometimes result
in reversible denaturation or in a time-dependent denaturation and, there-
fore, may be reversible if the eluent is collected in a dilution buffer imme-
diately after leaving the chromatography column.
Binding to affinity columns may be so tight as to preclude recovery of
an active enzyme or native protein. Under such circumstances ligands
with lower affinity must be tried. Alternatively, columns containing a
lower concentration of ligand, or ligand separated from the matrix by a
spacer arm, may be successful. One approach is to bind the ligand by
means of a hydrolyzable bond, either directly or to a destructible spacer
arm. Esters 32,33 are readily hydrolyzed with mild bases; vicinal hydroxyl
groups are cleared with periodate; and diazo bonds are reduced with
dithionates) 4 Columns of this sort are of only limited utility, since they
are destroyed by the "elution" procedure. All these strategies have been
described in Volume 34 of this series.
Electrophoresis has been used as a direct method of elution. 35 The
rationale behind this approach is that since proteins are charged at specific
pH values, they will migrate toward the appropriate electrode; affinity
forces should not be sufficiently strong to prevent this movement. This
method was applied mostly to antigen-antibody systems even though
success in other systems was also described. The procedure is mild and
the yields are very good. Since all carriers cannot be used and electro-
phoretic equipment is required, the procedure may not have broad appli-
cability.

Macromolecules Purified by Affinity Chromatography

Since the number of proteins that have been purified by affinity chro-
matography is very large, the methods that have been used previously
may serve to suggest approaches to those attempting novel purifications.
32R. J. Brown, N. E. Swaisgood,and H. R. Horton, Biochemistry 18, 4901 (1979).
33p. Singh, S. D. Lewis, and J. A. Shafer,Arch. Biochem. Biophys. 193, 284 (1979).
34p. Singh, S. D. Lewis, and J. A. Sharer, Arch. Biochem. Biophys. 203, 776 (1980).
35M. R. Morgan, P. J. Brown, M. J. Lieland,and P. D. Ocan, FEBS Lett. 87, 239 (1978).
[1] AFFINITY CHROMATOGRAPHY 21

A large number of selected affinity methods are recorded and summarized

in Tables I I - I V with the intent of providing some perspective as to avail-
able methods.

Enzymes
Among the many proteins purified by affinity chromatography, the
largest category consists of the enzymes. By reason of their discrete
specificity, a huge variety of ligands are potentially available; these in-
clude competitive enzyme inhibitors, coenzymes, and, to a lesser extent,
other substrates and cofactors. Purification of enzymes can be divided
into two categories. One is directed at enzymes of narrow specificity, i.e.,
enzymes purified with the aid of specific inhibitors or substrates. The
columns prepared with such ligands can be used only to purify the specific
enzyme for which the column was tailored. At times, specific columns
serve also for the purification of identical catalytic activities, but ones
derived from different organisms. The second category is designed for the
purification of a wide range of enzymes by using a general ligand, a ligand
common to a larger group of proteins. 36 Since about 30% of the known
enzymes require for their activity a coenzyme, i.e., a substrate common
to an entire class of reactions, immobilization of coenzymes on a carrier
will enable purification of a relatively large range of proteins on the same
column. The coenzymes in more common use are NAD, NADP, ATP,
and CoA, or derivatives of them which are themselves derivatives of
adenine nucleotides. Some idea of the variation possible, as well as indi-
cations for specific enzymes, are presented in Table II, which is arranged
alphabetically on the basis of the enzyme name. There is no attempt to be
all inclusive, and many meritorious preparations have not been included.

Immunoaffinity Chromatography
Affinity chromatography has been in use for many years for the purifi-
cation of antibodies by employing antigens bound to carriers. In recent
years, there has been a large increase in the use of immobilized antibodies
for purification of proteins, including enzymes. Under ideal conditions,
antibody columns allow separation of specific proteins from crude mix-
tures by a one-step procedure. Both conventional and monoclonal anti-
bodies have been used. A major difficulty in the use of immobilized anti-
bodies is the high affinity of antibody for antigens, thereby making the
recovery of active enzymes difficult owing to the drastic conditions re-
K. Mosbach, Adv. Enzymol. Relat. Areas Mol. Biol. 46, 205 (1978).
22 CHROMATOGRAPHY [1]

TABLE II
CONDITIONS FOR AFFINITYCHROMATOGRAPHYOF ENZYMES

Enzyme a Ligand Eluent

1. Acetylcholinesterase Acridinium, pyri- Decamethonium, methyl-

dinum, Con A o-mannoside
2. Acetyl-CoA carboxylase Biotin NaCI
3. N-Acetylgalactosamine UDP, apomucin NaCI
transferase
4. a-N-Acetylgalactos- CDP NaC1, CTP
aminide-(2-6)-sialyl-
transferase
5. N-Acetylglucosaminidase Thio-N-acetylglucos- NaC1
amide
6. N-Acetyl-fl-hexosamini- N-Acetyi-fl-D- Borate
dase A thioglucosamine
7. N-Acetyl-fl-hexosamini- Con A, 2-acetami- a-Methyl glucoside, N-
dase do-2-deoxy-fl- acetylglucosamine, pH,
glucosamine Triton X-100
8. Acid phosphatase Con A D-Mannose
9. Acid protease Pepstatin Urea gradient
10. Adenosine (phosph,.te) Inosine Adenosine, inosine
deaminase
11. Adenosine 5'-phospho- AMP AMP
sulfate sulfohydrolase
12. Adenine phosphoribo- AMP, GMP Mg, AMP
syltransferase
13. Adenosine kinase ATP, AMP Adenosine, pH
14. Adenosinetriphosphatase Calmodulin, ATP EDTA, ATP
15. Adenylate cylase Cholera toxin Tris forskolin
subunit GTP,
ATP forskolin
derivative
16. ADP-glucose pyrophos- 6-Phospho-l-pyro- AMP
phorylase (glucose-l- phosphate
phosphate adenylyl-
transferase)
17. ADP-ribosyltransferase Phenylagarose Propylene glycol
18. Alanine aminotransferase Cycloserine Pi
19. D-Alanine carboxypepti- Penicillins NaC1, hydroxylamine
dase
20. Alanine dehydrogenase NADP NaCl
21. Alcohol dehydrogenase AMP, NADP, ATP NAD, NADH, pyrazole,
KCI
22. Aldehyde dehydrogenase AMP, NAD NAD
23. Aldehyde reductase NADP KC1
[1] AFFINITY CHROMATOGRAPHY 23

TABLE II (continued)

Enzymea Ligand Eiuent

24. Aldose reductase 4-Carboxybenzalde- Pi

hyde
25. Alkaline phosphatase Phosphonic acid, Pi, a-methylmannoside
Con A
26. Amine oxidase Con A tx-Methyl-D-glucoside
27. Aminolevulinate dehydra- Aminophenyl, Dithiothreitol
tase (porphobilinogen mercuric acetate
synthase)
28. Aminopeptidase Leucylglycine NaC1, Zn
derivatives
29. a-Amylase Glycogen Glycogen
30. Anthranilate synthase Anthranilate pH
31. L-Arabinose kinase ATP NaC1
32. Arginase Antiarginase, L-Arg L-Arginine, NaC1
33. L-Arginine deaminase L-Arg L-Arg
34. Arylsulfatase A Con A subunit Sugar, Tris
35. Asparaginase L-Asp NaC1, Asp
36. Aspartase (aspartate L-Asp NaC1, Asp
ammonia-lyase)
37. Aspartate-fl-decarboxyl- L-Asp NaCI
ase
38. Aspartate transcarbamyl- L-Asp Asp, KC1
ase (aspartate carba-
moyltransferase)
39. Carbamoyl-phosphate Glu KC1
synthetase
40. Carbonic anhydrase Sulfonamides Perchlorate, KI, KCN,
(carbonate dehydratase) NaCI
41. Carboxypeptidase A Phenylpropionates, KCI
p-aminobenzyl
succinate
42. Carboxypeptidase B D-AFg, protease NaC1, pH
inhibitor
43. Carboxypeptidase N Aminobenzoyl- Guanidinoethyl-mercapto-
(arginine carboxypepti- arginine succinate
dase)
44. Catechol O-methyltrans- Catechols KC1
ferase
45. Cathepsin Pepstatin NaCI
46. Cellulase Cellulose Temperature
47. Choline acetyltransferase Coenzyme A Citrate-phosphate EDTA,
dithioerythritol
48. Choline dehydrogenase Choline Dithiothreitol

(continued)
24 CHROMATOGRAPHY [ 1]

TABLE II (continued)

Enzyme a Ligand Eluent

49. Cholinephosphate cytidyl- Glycerolphospho- NaCI

yltransferase choline
50. Chorismate synthase Pi Pi
51. Chorismate mutase- Phenylalanine Adamantane acetate
prephenate-dehydro-
genase
52. Chymotrypsin D-Tryptophan Acetic acid
methyl ester
53. Citrate synthase ATP ATP
54. Collagenase Collagen, Arg NaC1
55. Collagen galactosyltrans- Con A a-Methyl-l>mannoside,
ferase ethylene glycol
56. Collagen glucosyltrans- Collagen Collagen peptides, ethyl-
ferase ene glycol
57. Collagen glycosyltrans- UDPglucuronate Collagen peptides
ferase
58. Corrinoid enzyme Pteroylglutamate NaC1, pteroylglutamate
59. Creatine kinase p-Mercuribenzoate Mercaptoethanol, ATP, NaCl
60. Cyclic nucleotide phos- Phenylbutenolide KCI
phodiesterase troponin C
61. Cytochrome c oxidase Cytochrome c NaC1
62. Cytochrome reductase NADP, ADP NADP
63. Cytotoxic protease Soybean inhibitor pH 2
64. Deoxynucleotidyltrans- Oligo(dT) KC1
ferase (exo)
65. Deoxyribosylase , xdenines Deoxyinosine, guanidine
66. Dihydrofolate reductase Folate analogs Folate, pH, dihydrofolate
(tetrahydrofolate dehy-
drogenase)
67. Dihydropteridine reduc- 1,2-Naphtho- NADH, pH, Pi
tase quinone, metho-
trexate
68. Diol dehydrase Adenosylcobalamin Propane- 1,2-diol
69. Dipeptidyl peptidase 4-Phenylbutylgly- Ethylene glycol NaCI
cylproline
70. DNA-dependent DNA NaC1
ATPase II
71. DNA-dependent RNA tRNA (NH4)2SO4
polymerase
72. DNA ligase DNA KC1
73. DNA-nicking-closing DNA NaC1
enzyme
74. DNA polymerase DNA, pyrans, NaC1 or KCI
poly(N)
[1] AFFINITY CHROMATOGRAPHY 25

TABLE II (continued)

Enzymes Ligand Eluent

75. DNA polymerase/3 DNA NaCI

76. DNase DNA, d-DNA NaC1, KCI
77. DNA-unwinding DNA NaCi
enzyme II
78. Dopamine/3-hydroxylase Con A a-Methyl o-glucoside
(dopamine/3-monooxy-
genase)
79. Elastase Elastin, (Ala)3 Salt, Ala
80. Endonuclease DNA NaC1 or KC1
81. Endopeptidase Phenylbutylamine NaCI
82. Enterokinase (enteropep- p-Aminobenzami- NaC1, ethylene glycol,
tidase) dine benzamidine
83. Exonuclease I DNA NaC1
84. Fatty acid synthetase Pantetheine NADP pH, salt, NADP
85. Ferredoxin-NADP + re- Ferredoxin NaCI
ductase
86. Ferredoxin-nitrate reduc- Ferredoxin Dithionate
tase
87. Ferrochelatase Porphyrin Mesoporphyrin
88. Flavokinase (riboflavin Flavins Riboflavin
kinase)
89. FMN oxidoreductase AMP, NADP NAD, citrate
90. Formaldehyde dehydro- AMP NaCl
genase
91. Formate dehydrogenase AMP NaC1
92. Formiminotransferase Tetrahydrofolate Formiminoglutamate
93. Formylmethionine amino- N-Formylbestatin NaC1
peptidase
94. L-Fucose dehydrogenase AMP pH
95. a-L-Fucosidase Fucosamine thiofu- Fucose
copyranoside
96. Fumarase (fumarate Pyromelitate, 2-(5'- Malate, citrate
hydratase) pbenylpentyl)
fumarate
9% a-Galactosidase Galactopyranosyi- D-Galactose, methyl-D-
amine, Con A mannoside, NaC1
98. Acid a-galactosidase Thiolactopyranoside Galactonolactone
99. /3-Galactosidase Thiogalactoside Thiogalactoside, ethylene
glycol
100. fl-Galactosidase A2, A3 Con A a-Methylmannoside
101. Galactosaminidase (endo) Antifreeze-glyco- NaC1
protein
102. /3-Galactofuranosidase Galactomannan NaC1
(exo)

(continued)
26 CHROMATOGRAPHY [1]

TABLE II (continued)

Enzyme~ Ligand Eluent

103. /3-Galactosidase (neutral) Galactosylamine Citrate

104. D-Galactose kinase ATP NaCI
105. Galactosylhydroxylysyl- Collagen, UDPglu- Co!lagenpeptides, NaC1,
glucosetransferase cose, Con A a-methyl-D-glucoside,
ethylene glycol
106. Galactosyltransferase a-Lactalbumin UDPgalactose, NaC1, N-
UDP, mucin acetylglucosamine,
EDTA, Triton X-100,
sucrose
107. Glucokinase 2-Amino-2-deoxy- Glucose, KCl
glucose
108. Glucose isomerase Xylitol NaC1
109. Glucose-6-phosphate ADP, NADP NADP
dehydrogenase
110. Glucose-6-phosphatase Glucose 6-phos- KCI
phate
111. a-o-Glucosidase p-Aminophenyl- NaCI, sugars
maltoside,
phenylboronic
acid
112. Glucosyltransferase I, II Decosanates UDPglucose
113. Glutamate decarboxylase Ethyl glutamate NaC1
114. Glutamate dehydrogenase NADP, 2-AMP NaCI, NADP
115. Glutamate synthase 2',5'-ADP Glutamine, NADPH
116. y-Glutamylcysteine syn- ATP, cystamine ATP, dithiothreitol
thase
117. y-Glutamyl hydrolase Polyglutamate NaC1
118. Glutathione reductase ADP, glutathione NADPH, NADP, KC1
119. Glutathione transferase Bromosulfophthalein, KSCN, bromosulfo-
glutathione phthalein
120. Glyceraldehyde-3-phos- NAD, AMP, ATP NAD, NADH, glycerol 3-
phate dehydrogenase phosphate
121. Glycerol-3-phosphate ATP NADH
dehydrogenase
122. Glycineamide ribonu- Glycineamide, Pi
cleotide transformylase ribonucleotide
123. Glycogen phosphorylase AMP AMP
124. Glycosidases Sugar derivative Sugar derivative
125. Glyoxylase Glutathione pH
126. Glyoxylase I Glutathione Glutathione
127. Glyoxylase II Glutathione S-Octylglutathione
128. GMrganglioside-fl-galacto- Thiogalactoside, NaC1, a-methyl-D-manno-
sidase Con A side
129. Gramicidin-S synthetase Ornithine, D-phenyl- KC1
alanine
[1] AFFINITY CHROMATOGRAPHY 27

TABLE II (continued)

Enzyme a Ligand Ehient

130. GTP cyclohydrolase GTP, dihydrofolate GTP

131. Guanylate cyclase GTP EDTA, NaC1
132. Guanyloribonuclease Guanylyl-(2',5')- pH
(ribonuclease Tl) guanosine
133. Guanylyltransferase DNA NaC1
(mRNA)
134. Guanine aminohydrolase Guanine NaC1
(guanine deaminase)
135. N2-Guanine RNA-methyl- S-Adenosylhomo- NaC1
transferase cysteine
136. Hexokinase Glucosamines Glucose, MgATP
137. Hexosaminidase A,B /3-o-Glucosylamine pH
138. Hexosaminidase A Thiogalactoside 2- Gluconolactone, N-ace-
deoxy-l-thio-fl-o- tylglucosaminolactone
galactoside
139. Hexosaminidase P N-Acetylglucosyl- pH
amine
140. Histaminase (amine oxi- Cadaverine Heparin
dase)
141. L-Histidine 2-oxoghitarate Histidine NaC1
aminotransferase (histi-
dine aminotransferase)
142. Histidinol dehydrogenase His, histamine, His, imidazole, AMP
AMP
143. Hyaluronidase Con A, glycos- a-Methylglucoside, a-
aminoglycans methylmannoside NaC1
144. 3-Hydroxyacyl-CoA NAD Pi
dehydrogenase
145. to-Hydroxy-fatty-acid NADP KCi, NADP
NADP-oxidoreductase
146. 3-Hydroxy-3-methylghi- HHG-CoA HMG-CoA, glycerol
taryl-CoA reductase
147. 15-Hydroxyprostaglandin NAD NADH
dehydrogenase
148. 3ct-Hydroxysteroid dehy- NAD Pi
drogenase
149. 3/3-Hydroxysteroid dehy- NAD Pi
drogenase
150. 20a-Hydroxysteroid 1la-Hydroxyproges- pH
dehydrogenase terone
151. 20fl-Hydroxysteroid 1lc~-Hydroxypro- Mercaptoethanol
dehydrogenase gesterone
152. 3fl-Hydroxysteroid oxi- Cholesterol Triton X-100
dase

(contmued)
28 CHROMATOGRAPHY [1]

TABLE II (continued)

Enzyme a Ligand Eluent

153. Hypoxanthine-guanine GMP KC1, GMP

phosphoribosyltrans-
ferase
154. Indolyl-3-alkane c~-hy- Indole Acetate
droxylase
155. Inosinic acid dehydro- 8-Amino-GMP, IMP, XMP
genase AMP
156. Isocitrate dehydrogenase AMP NAD
157. ~x-Isopropylmalate Leu Glycerol
isomerase
158. Kallikrein Aprotinin, amino- Benzamidine
benzamidine
159. 2-Keto-3-deoxy-L-fuco- NAD NAD
nate NAD oxidoreduc-
tase
160. 2-Kynurenine 3-hydroxy- NADP NADP
lase (kynurenine 3-
monooxygenase)
161. Lactate dehydrogenase AMP, oxamate, NAD, NADH
ATP
162. Lipases (triacylglycerol Palmitate, Con A, Detergent, a-methyl-D-
lipase) cholate glucoside, cholate
163. Lipoamide dehydrogenase Lipoamide, NAD Salt
(dihydrolipoamide
reductase)
164. Lipoprotein lipase Heparin NaC1
165. Lipoxygenase Linolenate NaC1
166. Lysine tRNA-synthetase Lysine Lysine
167. L-Lysine 6-aminotrans- L-Lysylacetamido- NaCI
ferase dodecyl
168. Lysozyme Chitotetraose N- NaCI
acetyl-/3-glucos-
amide
169. Lysyl oxidase Collagen NaC1, urea
170. Luciferase FMN, benzylox- FMN, ethylene glycol, Pi
yanilin, benzyl-
amine
171. Malate dehydrogenase AMP, ADP, ATP, pH, NADH, NADP, NaCI
NAD
172. Malate thiokinase ADP CoA, ATP, malate
173. Malic enzyme NADP, ADP NADP
174. Malonyl-CoA NADP NaCI
decarboxylase
175. Maltodextrin Glycogen NaC1
phosphorylase
[1] AFFINITY CHROMATOGRAPHY 29

TABLE II (continued)

Enzymea Ligand Eluent

176. a-n-Mannosidase Mannosylamine, Mannose, NaC1

benzidine
177. Methionyl-tRNA synthe- AMP, tRNA L-Methionine
tase
178. 3-Methyladenine DNA- DNA NaC1
glycosylase
179. Methylmalonyl-CoA Vitamin BI2 Vitamin B12
mutase
180. Methyltransferase mRNA ADP NaCl
(nucleoside-2'-)
181. NS-Methyltetrahydrofo - Cobalamin NaCI, photolysis
late-homocysteine
methyltransferase
182. Monoamine oxidase Tyramine KC1
(amine oxidase)
183. Myosin light-chain kinase Light chain Pi
184. Myosin kinase (I, II) Calmodulin EGTA
185. NAD kinase NAD NAD
186. NADPH-adrenodoxin Adrenodoxin NaCI
reductase
187. NADPH-cytochrome c ADP, NADP AMP, 2-AMP, NADP
(P-450) reductase
188. NADPH-ferredoxin re- Adrenodoxin NaC1
ductase
189. NADPH-flavin oxidore- Flavin (FMN) FMN
ductase
190. NAD-protein ADP-ribo- AMP NH4CI
syltransferase
191. Neuraminidase Tyrosyl p-ni- pH, acetate, NaCI
trophenyloxamic
acid, colominic
acid
192. Neutral protease N-Phenylphos- pH
phenyl-Phe-Phe
193. Nitrite reductase Ferredoxin Pi
194. Nuclease (micrococcal) DNA KCI
195. Nuclease T TDP Guanidine
196. Nucleotidase Con A, 5'-AMP 5'-AMP
197. Nucleotide phosphodies- Polyhistidine pH
terase
198. Nucleotide phosphotrans- AMP H20
ferase
199. Nucleotide pyrophospha- AMP NaC1
tase

(continued)
30 CHROMATOGRAPHY [1]

TABLE II (continued)

Enzyme~ Ligand Eluent

200. Ornithine decarboxylase Pyridoxamine Pyridoxal phosphate

phosphate
201. Ornithine transcarbamyl- N-(Phosphonacetyl) Carbamoyl phosphate
ase L-Orn
202. Orotidine-5'-phosphate Azauridine 5'- NaCI
decarboxylase phosphate
203. Orotate phosphoribo- Orotidine 5-phos- Orotate-MgC12or 5-phos-
syltransferase phate phoribosyl-fl-D-ribose
diphosphate
204. Pepsin Poly(L-lysine) NaCl
205. Peroxidase Con A a-Methyl-D-mannoside
206. Phenylalanine ammonia- Phenylalanine Phenylalanine
lyase
207. Phenylalanine hydroxy- Pteridine derivative Phenylalanine, pH
lase
208. Phenylalanine-tRNA RNA KCl-glycerol
synthetase
209. Phosphatidylglycero- CDP-diglyceride CDP-diglyceride
phosphate synthetase
210. Phosphodiesterase Activator protein, Ca2÷, chelating agent,
regulatory pro- electrophoresis
tein
211. Phosphoenolypyruvate GTP, GMP KCI, IDP
carboxykinase
212. Phosphofructokinase AMP KCI, ATP
213. 6-Phosphogluconate NADP, ADP Citrate, NADP
dehydrogenase
214. Phosphoglucose ATP Glucose 6-phosphate
isomerase
215. Phosphoglycerate kinase ATP ATP (Mg2÷)
216. Phospholipase Con A, heparin t~-Methyi-D-pyranoside,
NaCI
217. Phospholipase A2 Glycero-3-phospho- EDTA
choline, phos-
phatidylcholine
218. Phospholipase C Lipoprotein NaC1, urea
219. Phosphoprotein Protamine, histone NaC1
phosphatase
220. Phosphorylase a AMP AMP
221. Phosphorylase b AMP AMP
222. Phosphorylase kinase Calmodulin Low Ca, EDTA
223. Poly(A) polymerase ATP, DNA, ATP, KCI
poly(A), RNA
[1] AFFINITY CHROMATOGRAPHY 31

TABLE II (continued)

Enzymea Ligand Eluent

224. Poly(adenosine diphos- DNA NaC1

phate rihose) poly-
merase
225. Polynucleotide phos- RNA NaC1
phorylase (polyribonu-
cleotide nucleoti-
dyltransferase)
226. Polynucleotide 5'-triphos- ADP, poly(U) NaC1
phatase
227. Post-proline cleaving Z-Prolyl-D-alanine Z-Prolylphenylalanine
enzyme
228. Prolyl hydroxylase Poly(L-proline), Poly(L-lysine), ethylene
collagen glycol
229. Propionyl-CoA Avidin Biotin
carboxylase
230. Prostaglandin cyclooxy- Flurbiprofen Flufenamic acid, ethylene
genase glycol
231. Proteases 4-Phenylbutyl- NaC1, urea mercuric
amine, soybean chloride Ca, etc.
trypsin inhibitor,
D-tryptophan-
methyl ester, etc.
232. Protein kinase (AMP- and cAMP, histone, cAMP, NaCI, c,GMP,
GMP-dependent) cGMP, catalytic guanidine NaCI, ethao
subunit, phos- nolamine
vitin, EGF,
casein
233. Protein phosphatase Histone NaCI
(phosphoprotein phos-
phatase)
234. Pteroyl-a-oligoglutamyl Oiigoglutamyl NaC1
endopeptidase peptides
235. Pullulanase a-Dextrin fl-Dextfin
236. Purine nucleoside phos- Formycin B Inosine
phorylase
237. Pyridine dinucleotide NAD, 2',5'-ADP NADPH, 2'-AMP
transhydrogenase
238. Pyridoxal kinase Pyridoxal Pyridoxine
239. Pyridoxamine-5-phosphate Phosphopyridoxyl Pyridoxal phosphate
oxidase
240. Pyruvate decarboxylase Thiamin pyrophos- EDTA
phate
241. Pyruvate dehydrogenase Thiochrome Salt

(continued)
32 CHROMATOGRAPHY [1]

TABLE II (continued)

Enzyme a Ligand Eluent

242. Pyruvate-UDP-N-acetyl- UDP NaC1

glucosaminetransferase
243. Renin Pepstatin, hemoglo- pH
bin octapeptide
244. Restriction endonuclease DNA P~, KC1
245. Reverse transcriptase Poly(U) Poly(G)
246. Ribonuclease CMP NaC1
247. Ribonuclease F~, F2 5'-GMP 2'(3)-GMP
248. Ribonuclease H Oligo(dT) EDTA
249. Ribonuclease T2 AMP 2'(3)-AMP, NaCI
250. Ribonuclease U2 AMP NaCI
251. Ribonucleotide reductase dGTP dGTP, urea
252. RNA ligase 2',5'-ADP Mg
253. tRNA nucleotidyltrans- tRNA NaCI-EDTA
ferase
254. RNA polymerase DNA, rifamycin, Salts
heparin
255. RNA polymerase III DNA (NH4)zSO4
256. RNA polymerase basic RNA (NH4)zSO4
protein
257. RNA polymerase (DNA DNA KC1
dependent)
258. Salicylate hydroxylase p-Aminosalicylate Salicylate
(salicylate 1-monooxy-
genase)
259. Sialidase (neuraminidase) a-Acid glycoprotein pH
260. Sialyl transferase (CMP- CDP CDP
N-acetyi-neuraminate-
galactosylglycoprotein
sialyltransferase)
261. Sorbitol dehydrogenase NAD NAD
(L-iditoldehydrogenase)
262. Spermine synthase Spermine Spermidine
263. Sphingomyelinase Con A c~-Methyl-o-mannoside
264. Succinyl-CoA acetoace- Acetoacetyl-CoA Acetoacetate
tyl-CoA transferase
265. Succinyl-CoA synthetase GDP GDP
266. Succinate-semialdehyde AMP AMP
dehydrogenase
267. Succinate thiokinase 3',5'-ADP CoA
(succinyl-CoA
synthetase)
268. Terminal deoxynucleoti- Oligo(dT) KC1
dyltransferase
[1] AFFINITY CHROMATOGRAPHY 33

TABLE II (continued)

Enzyme a Ligand Eluent

269. Thioredoxin reductase ADP NADPH

270. Threonine deaminase Val, Ile KC1, K3PO4
(threonine dehydratase)
271. Threonine dehydratase AMP AMP
272. Thrombin Aminobenzamidine Benzamidine
273. Thymidine kinase Uracil, glycoprotein NaCI, thymidine TMP
TMP
274. Thymidylate synthase Thymidylate aria- Pi, Tris
logs
275. Thyroid peroxidase Tyrosine pH
276. Transcarboxylase Avidin pH
(methylmalonyI-CoA
carboxyltransferase)
277. Transhydrogenase NAD NADH
[NAD(P) ÷ transhydro-
genase]
278. Triglyceride lipase (tri- Heparin NaCI
acylglycerol lipase)
279. Tryptophanase Pyridoxal phos- Pyridoxal phosphate
phate
280. Tryptophan 5-monooxy- Pteridine NaC1
genase
281. Tryptophan synthase Indolepropionate, Imidazole, indolepro-
tryptophanol-P panol-P
282. Tyrosinase Con A a-Methyl-D-mannoside
283. Tyrosine aminotransfemse Succinate pH
284. Tyrosine phenol-lyase Pyridoxal phos- Pyridoxal phosphate
phate
285. UDPgalactose 4-epi- UDP UMP, UDP
merase
286. UDPgalactose glycopro- a-Lactalbumin N-Acetylglucosamine
tein galactosyltrans-
ferase (glycoprotein/3-
D-galactosyltransferase)
287. UDPglucose dehydro- AMP AMP
genase
288. UDPglucuronyltransferase UDP UDPglucuronic acid
289. UDP-N-acetylenolpy- NADP NADPH
ruvylglucosamine re-
ductase
290. Urate oxidase 8-Aminoxanthine Urate
291. Uricase (urate oxidase) Urate, xanthine Xanthine
292. Urokinase /3-Naphthamidine pH
293. Valyl-tRNA synthetase tRNA KCI
34 CHROMATOGRAPHY [ 1]

References to TABLE II
a Key to references:
1. P. J. Morrod, A. G. Marshall, and D. G. Clark, Biochem. Biophys. Res. Commun.
63, 335 (1975); P. Ott, B. Jenny, and V. Brodbeck, Eur. J. Biochem. 57, 469 (1975);
T. L. Rosenberry and J. M. Richardson, Biochemistry 16, 3550 (1977); C. A. Reavil,
M. S. Wooster, and D. T. Plummer, Biochem. J. 173, 851 (1978); C. A. Reavil and
D. T. Plummer, J. Chromatogr. 157, 141 (1978); G. Webb and D. G. Clark, Arch.
Biochem. Biophys. 191, 278 (1978); A. S. Brooks, G. E. Tiller, and W. G. Struve,
Biochim. Biophys. Acta 615, 354 (1980); K. K. Parker, S. L. Chan, and A. J.
Trevor, Arch. Biochem. Biophys. 187, 322 (1978); T. H. Meedel, Biochim. Biophys.
Acta 615, 360 (1980).
2. J. S. Wolpert and M. L. Ernst-Fonberg, Biochemistry 14, 1095 (1975).
3. M. Schwyzer and R. L. Hill, J. Biol. Chem. 252, 2338 (1977); M. Sugiura, T.
Kawasaki, and I. Yamashina, J. Biol. Chem. 257, 9501 (1982).
4. J. E. Sadler, J. I. Rearick, and R. L. Hill, J. Biol. Chem. 254, 5934 (1977).
5. L. R. Glasgow, J. C. Paulson, and R. L. Hill, J. Biol. Chem. 252, 8615 (1977).
6. E. E. Grebner and I. Parikh, Biochim. Biophys. Acta 350, 437 (1974).
7. D. K. Banerjee and D. Basu, Biochem. J. 145, 113 (1975); R. G. Edwards, P.
Thomas, and J. M. Westwood, Biochem. J. 151, 145 (1975); M. Pokorny and
C. P. J. Glaudermans, FEBS Lett. 50, 66 (1975).
8. M. S. Saini and R. L. VanEtten, Arch. Biochem. Biophys. 191, 613 (1978).
9. E. Hackenthal, R. Hackenthal, and U. Hilgengeldt, Biochim. Biophys. Acta 522,
561 (1978); K. Yamamoto, N. Katsuda, and K. Kato, Eur. J. Biochem. 92, 499
(1978).
10. H. Rosemeyer and F. Seela, Anal. Biochem. 115, 339 (1981).
11. K. M. Rogers, G. F. White, and K. S. Dodgson, Biochim. Biophys. Acta 527, 70
(1978).
12. A. S. Olsen and G. Milman, Biochemistry 16, 2501 (1977); J. A. Holden, G. S.
Meredith, and W. H. Kelley, J. Biol. Chem. 254, 6951 (1979).
13. C. M. Andres and I. H. Fox, J. Biol. Chem. 254, 11388 (1979); R. L. Miller, D. L.
Adamczyk, and W. H. Miller, J. Biol, Chem. 254, 2339 0979); J. W. DeJong, E.
Keijzer, M. P. Uitendaal, and E. Harmsen, Anal. Biochem. 101, 407 (1980).
14. F. W. Hulla, M. Hodsel, S. Risi, and K. Dose, Eur. J. Biochem. 67, 469 (1976); K.
Gietzen, M. Tejcka, and H. U. Wolf, Biochem. J. 189, 81 (1980); V. Niggli, J. T.
Penniston, and E. Carafoli, J. Biol. Chem. 254, 9955 (1979).
15. V. Bennett, E. O'Keefe, and P. Cuatrecasas, Proc. Natl. Acad. Sci. U.S.A. 72, 33
(1975); N. I. Swislocki, T. Magnuson, and J. Tierney, Arch. Biochem. Biophys.
179, 157 (1977); C. Homcy, S. Wrenn, and E. Haber, Proc. Natl. Acad. Sci. U.S.A.
75, 59 (1978); T. Pfeuffer and H. Metzger, FEBS Lett. 146, 369 (1982).
16. T. Haugen, A. Ishaque, A. K. Chatterjee, and J. Preiss, FEBS Lett. 42, 205 (1974).
17. J. Moss, S. J. Stanley, and N. J. Oppenheimer, J. Biol. Chem. 254, 8891 (1979).
18. T. K. Korpela, J. Chromatogr. 143, 519 (1977); T. K. Korpela, A. E. Hinkkanen,
and R. P. Raunio, J. Solid Phase Biochem. 1, 215 (1977).
19. R. Yocum, P. M. Blumberg, and J. L. Strominger, J. Biol. Chem. 249,-4863 (1974);
M. Gorecki, A. Bar-eli, Y. Burstein, A. Patchornik, and E. B. Chain, Biochem. J.
147, 131 (1975); H. H. Martin, W. Schilf, and C. Maskos, Eur. J. Biochem. 71, 585
(1976); W. Schilf and H. H. Martin, Eur. J. Biochem. 105, 361 (1980); J. Coyette,
J. M. Ghuysen, and R. Fontana, Eur. J. Biochem. 88, 297 (1978).
[1] AFFINITY CHROMATOGRAPHY 35

References to TABLE II (continued)

20. D. Keradjopoulos and A. W. Holldorf, Biochim. Biophys. Acta 570, 1 (1979).
21. M. J. Comer, D. B. Craven, M. J. Harvey, A. Atkinson, and P. D. G. Dean, Eur. J.
Biochem. 55, 201 (1975); W. F. Bosron, T. K. Li, L. G. Lange, W. P. Dafeldecker,
and B. L. Vallee, Biochem. Biophys. Res. Commun. 74, 85 (1977); Boston et al.,
Proc. Natl. Acad. Sci. U.S.A. 74, 4378 (1977); C. N. Ryzewski and R. Pietruszko,
Arch. Biochem. Biophys. 183, 73 (1977); H. Von Bahr-Lindstrom, L. Andersson,
K. Mosbach, and H. Jornvall, FEBS Lett. 89, 293 (1978); A. J. L. Brown and C. Y.
Lee, Biochem. J. 179, 479 (1979); B. Tabakoff and J. P. yon Wartburg, Biochem.
Biophys. Res. Commun. 63, 957 (1975); C. Y. Lee, D. Charles, and D. Bronson, J.
Biol. Chem. 254, 6375 (1979).
22. R. J. S. Duncan, Biochem. J. 161, 123 (1977); N. J. Greenfield and R. Pietruszko,
Biochim. Biophys. Acta 483, 35 (1977); H. Nakayasu, K. Mihara, and R. Sato,
Biochem. Biophys. Res. Commun. 83, 697 (1978); K. Inoue, H. Nishimukai, and K.
Yamasawa, Biochim. Biophys. Acta 569, 117 (1979); K. A. Bostian and G. F. Betts,
Biochem. J. 173, 773 (1978); A. L. Bognar and E. A. Meighen, J. Biol. Chem. 253,
446 (1978).
23. B. Tabakoff and J. P. yon Wartburg, Biochem. Biophys. Res. Commun. 63, 957
(1975).
24. P. F. Kador, D. Carper, and J. H. Kinoshita, Anal. Biochem. 114, 53 (1981).
25. E. Mossner, M. Boll, and G. Pfleiderer, Hoppe-Seyler's Z. Physiol. Chem. 361, 543
(1980); F. G. Lehmann, Biochim. Biophys. Acta 616, 41 (1980); T. Komoda and Y.
Sakagishi, Biochim. Biophys. Acta 445, 645 (1976); M. Landt, S. C. Boltz, and
L. G. Butler, Biochemistry 17, 915 (1978); Y. Yokota, J. Biochem. (Tokyo) 83, 1293
(1978); B. O'Keefe and J. E. Kinsella, Int. J. Biochem. 10, 125 (1979).
26. J. J. Shiem, T. Tamaye, and K. J. Yasunobu, Biochim. Biophys. Acta 377, 229
(1975); P. Turini, S. Sabatini, O. Befani, F. Chimenti, C. Casanova, P. L. Riccio,
and B. Mondovi, Anal. Biochem. 125, 249 (1982).
27. S. R. Chandrika and G. Padmanaban, Biochem. J. 191, 29 (1980).
28. S. M. M. Basha, M. N. Horst, F. W. Bazer, and P. M. Roberts, Arch. Biochem.
Biophys. 185, 174 (1978); K. H. Rohm, Hoppe-Seyler' s Z. Physiol. Chem. 363, 641
(1982); C. Kettner, J. Rodriquez-Absi, G. I. Glover, and J. M. Prescott, Arch.
Biochem. Biophys. 162, 56 (1974).
29. R. Tkachuk, FEBS Lett. 52, 66 (1975).
30. T. H. Grove and H. R. Levy, Biochim. Biophys. Acta 397, 80 (1975).
31. P. H. Chan and W. Z. Hassid, Anal. Biochem. 64, 372 (1975).
32. R. Tarrab, B. Perez, and C. Lopez, Anal. Biochem. 65, 26 (1975); S. Traniello, R.
Barsacchi, E. Magri, and E. Grazi, Biochem. J. 145, 153 (1975).
33. T. Shibatani, T. Kakimoto, and I. Chibata, J. Biol. Chem. 250, 4580 (1975).
34. K. A. Balasubramanian and B. K. Bachhawat, Biochim. Biophys. Acta 4113, 113
(1975); A. A. Farooqui and P. N. Srivastava, Biochem. J. 181, 331 (1979); R. L. Van
Etten and A. Waheed, Arch. Biochem. Biophys. 202, 366 (1980).
35. T. Tosa, T. Sato, R. Sano, K. Yamamoto, Y. Matuo, and I. Chibata, Biochim.
Biophys Acta 334, 1 (1974).
36. Same as 35.
37. Same as 35.
38. G. S. J. Rao, M. S. Savithri, S. Seethalakshmi and N. A. Rao, Anal. Biochem. 95,
401 (1979).
(continued)
36 CHROMATOGRAPHY [1]

References to TABLE II (continued)

39. A. T. H. Abdelal and J. L. Ingraham, J. Biol. Chem. 250, 4410 (1975).
40. P. L. Whitney, Anal. Biochem. 57, 467 (1974); P. J. Wistrand, S. Lindahl, and T.
Wahlstrand, Eur. J. Biochem. 57, 189 (1975); P. J. Wistrand and T. Wahlstrand,
Biochim. Biophys. Acta 481, 712 (1977); J. T. Johansen, Carlsberg Res. Commun.
41, 73 (1976); R. S. Holmes, Eur. J. Biochem. 78, 511 (1977).
41. G. Oshima and K. Nagasawa, J. Biochem. (Tokyo) 81, 1285 (1977); L. B. Cueni,
T. J. Bazzone, J. F. Riordan, and B. L. Vallee, Anal. Biochem. 1117, 341 (1980).
42. Y. Narahashi, K. Yoda, and S. Honda, J. Biochem. (Tokyo) 82, 615 (1977); H. R.
Trayer, M. A. Winstanley, and I. P. Trayer, FEBS Lett. 83, 141 (1977).
43. T. H. Plummer and M. Y. Hurwitz, J. Biol. Chem. 253, 3907 (1978).
44. R. T. Borchardt, C. F. Cheng, and D. R. Thakker, Biochem. Biophys. Res. Com-
mun. 63, 69 (1975); R. T. Borchardt and C. F. Cheng, Biochim. Biophys. Acta 522,
49 (1978); P. A. Gulliver and K. F. Tipton, Eur. J. Biochem. 88, 439 (1978).
45. K. Yamamoto, N. Katsuda, and K. Kato, Eur. J. Biochem. 92, 499 (1978).
46. M. Nummi, M. L. Niku-Paavola, T. M. Enari, and V. Raunio, Anal. Biochem. 116,
137 (1981).
47. R. L. Ryan and W. O. McClure, Biochemistry 18, 5357 (1979).
48. H. Tsuge, Y. Nakano, H. Onishi, Y. Futamura, and K. Ohashi, Biochim. Biophys.
Acta 614, 274 (1980).
49. P. C. Choy and D. E. Vance, Biochem. Biophys. Res. Commun. 72, 719 (1976).
50. T. W. Cole and F. H. Gaertner, Biochem. Biophys. Res. Commun. 67, 170
(1975).
51. G. D. Smith, D. V. Roberts, and A. Daday, Biochem. J. 165, 121 (1977).
52. F. Pochon and J. G. Bieth, J. Biol. Chem. 257, 6683 (1982).
53. M. Lindberg and K. Mosbach, Eur. J. Biochem. 53, 481 (1975).
54. C. C. Huang and M. Abramson, Biochim. Biophys. Acta 384, 484 (1975); C. Gillet,
Y. Ecckhont, and G. Vaes, FEBS Lett. 74, 126 (1977); I. Emod and B. Keil, FEBS
Lett. 77, 51 (1977).
55. L. Risteli, Biochem. J. 169, 189 (1978).
56. H. Anttinen, R. Myllyla, and K. I. Kivinkko, Eur. J. Biochem. 78, 11 (1977); R.
Myllyla, H. Anttinen, L. Risteli, and K. I. Kivinkko, Biochim. Biophys. Acta 480,
113 (1977).
57. R. Myllyla, H. Anttinen, L. Risteli, and K. I. Kivinkko, Biochim. Biophys. Acta
480, 113 (1977).
58. F. K. Welty and H. G. Wood, J. Biol. Chem. 253, 5832 (1978).
59. R. J. Boegman, FEBS Lett. 53, 99 (1975); V. Madelian and W. A. Warren, Anal.
Biochem. 64, 517 (1975); N. Hall, P. Addis, and M. DeLuca, Biochem. Biophys.
Res. Commun. 76, 950 (1977); N. Hall, P. Addis, and M. DeLuca, Biochemistry 18,
1745 (1979); C.-Y. Lee, L. H. Lazarus, D. S. Kabakoff, P. J. Russell, M. Laver,
and N. O. Kaplan, Arch. Biochem. Biophys. 178, 8 (1977).
60. D. M. Watterson and T. C. Vanaman, Biochem. Biophys. Res. Commun. 73, 40
(1976); A. F. Prigent, G. Nemoz, and H. Pacheco, Biochem. Biophys. Res. Com-
mun. 95, 1080 (1980).
61. K. Bill, R. P. Casey, C. Brojer, and A. Azzi, FEBS Lett. 120, 248 (1980); R. B.
Gennis, R. P. Casey, A. Azzi, and B. Ludwig, Eur. J. Biochem. 125, 189 (1982); T.
Ozawa, M. Tanaka, and T. Wakabayshi, Proc. Natl. Acad. Sci. U.S.A. 79, 7175
(1982).
[1] AFFINITY CHROMATOGRAPHY 37

References to TABLE II (continued)

62. J. D. Dignam and H. W. Strobel, Biochemistry 16, 1116 (1977); D. A. Schafer and
D. E. Hulguist, Biochem. Biophys. Res. Commun. 95, 381 (1980).
63. V. B. Hatcher, M. S. Oberman, G. S. Lazarus, and A. I. Grayzel, J. Immunol. 120,
665 (1978).
64. S. Okamura, F. Crane, H. A. Messner, and T. W. Mak, J. Biol. Chem. 253, 3765
(1978).
65. J. Holguin and R. Cardinaud, Eur. J. Biochem. 54, 505 (1975).
66. D. L. Peterson, J. M. Gleisner, and R. L. Blakley, Biochemistry 14, 5261 (1975); D.
Baccanari, A. Phillips, S. Smith, D. Sinski, and J. BurchaU, Biochemistry 14, 5267
(1975); S. V. Gupta, N. J. Greenfield, M. Poe, D. R. Makulu, M. N. Williams, B. A.
Moroson, and J. R. Bertino, Biochemistry 16, 3073 (1977); R. S. Gupta, W. F.
Flintoff, and L. Siminovitch, Can. J. Biochem. 55, 445 (1977); T. M. Lincoln, W. L.
Dills, and J. D. Corbin, J. Biol. Chem. 252, 4269 (1977); R. Ferone and S. Roland,
Proc. Natl. Acad. Sci. U.S.A. 77, 5802 (1980); R. L. Then, Biochim. Biophys. Acta
614, 25 (1980).
67. C. G. H. Cotton and I. Jennings, Eur. J. Biochem. 83, 319 (1978); S. Webber, T. L.
Deits, W. R. Snyder, and J. M. Whiteley, Anal. Biochem. 84, 491 (1978).
68. T. Toraya and S. Fukui, J. Biol. Chem. 255, 3520 (1980).
69. R. M. Metrione, Biochim. Biophys. Acta 526, 531 (1978).
70. E. Richet and M. Kohiyama, J. Biol. Chem. 253, 7490 (1978).
71. B. Wittig and S. Wittig, Biochim. Biophys. Acta 520, 598 (1978).
72. M. Katouzian and J. C. David, Biochem. Biophys. Res. Commun. 82, 1168 (1978).
73. W. R. Bauer, E. C. Ressner, J. Kates, and J. V. Patzke, Proc. Natl. Acad. Sci.
U.S.A. 74, 1841 (1977).
74. D. J. Arndt-Jovin, T. M. Jovin, W. Bahr, A. H. Frischauf, and M. Marquardt, Eur.
J. Biochem. 54, 411 (1975); A. Matsukage, E. W. Bohn, and S. H. Wilson, Bio-
chemistry 14, 1006(1975); L. A. Salzman and L. McKerlie, J. Biol. Chem. 250, 5583
(1975); P. Chandra and L. K. Steel, Biochem. J. 167, 513 (1977); A. Baer and W.
Schiebel, Eur. J. Biochem. 86, 77 (1978); R. A. Hitzeman and A. R. Price, J. Biol.
Chem. 253, 8518 (1978); H. Joenje and R. M. Benbow, J. Biol. Chem. 253, 2640
(1978); W. Mastropaolo and C. A. Lang, J. Biol. Chem. 253, 1978 (1978); C. A.
Ross and W. J. Harris, Biochem. J. 171, 231 (1978); J. G. Chirikjian, L. Rye, and T.
S. Papas, Proc. Natl. Acad. Sci. U.S.A. 72, 1142 (1975); M. J. Modak and S. L.
Marcus, J. Biol. Chem. 252, 11 (1977); G. Bauer, G. Jilek, and P. H. Hofschneider,
Eur. J. Biochem. 79, 345 (1977); A. Hizi and W. K. Joklik, J. Biol. Chem. 252, 2281
(1977).
75. T. A. Kunkel, J. E. Tcheng, and R. R. Meyer, Biochim. Biophys. Acta 520, 302
(1978).
76. G. Sabeur, P. J. Sicard, and G. Aubel-Sadron, Biochemistry 13, 3203 (1974); H.
Tanaka, I. Sasaki, K. Yamashita, K. Miyazaki, T. Matuo, J. Yamashita, and T.
Hario, J. Biochem. (Tokyo) 88, 797 (1980); H. Tanaka, I. Sasaki, K. Yamashita, Y.
Matuo, J. Yamashita, and T. Horio, J. Biochem. (Tokyo) 91, 1411 (1982).
77. M. Abdel-Monem, M. C. Chanal, and H. Hoffmann-Berling, Eur. J. Biochem. 79,
33 (1977).
78. R. A. Rush, P. E. Thomas, S. H. Kindler, and S. Udenfriend, Biochem. Biophys.
Res. Commun. 57, 1301 (1974); R. P. Frigon and R. A. Stone, J. Biol. Chem. 253,
678O (1978).
(continued)
38 CHROMATOGRAPHY [ 1]

References to TABLE II (continued)

79. Y. Legrand, G. Pignaud, J. P. Caen, B. Robert, and L. Robert, Biochem. Biophys.
Res. Commun. 63, 224 (1975); K. Katagiri, T. Takeuchi, K. Taniguchi, and M.
Sasaki, Anal. Biochem. 86, 159 (1978); C. Gillet, Y. Ecckhont, and G. Vaes, FEBS
Left. 74, 126 (1977).
80. M. Mechali and A. M. DeRecondo, Eur. J. Biochem. 58, 461 (1975); S. Ljungquist,
J. Biol. Chem. 252, 2808 (1977); S. Riazuddin and L. Grossman, J. Biol. Chem. 252,
6280 (1977); V. Bibor and W. G. Verly, J. Biol. Chem. 253, 850 (1978).
81. K. D. Jany, H. Haug, G. Pfleiderer, and J. Ishay, Biochemistry 17, 4675 (1978).
82. D. W. A. Grant and J. Hermon-Taylor, Biochem. J. 147, 363 (1975); L. E. Ander-
son, K. A. Walsh, and H. Neurath, Biochemistry 16, 3354 (1977); D. W. A. Grant,
A. I. Magee, and J. Hermon-Taylor, Eur. J. Biochem. 88, 183 (1978).
83. R. K. Ray, R. Rueben, I. Molineux, and M. Gefter, J. Biol. Chem. 249, 5379
(1974).
84. F. A. Lornitzo, A. A. Qureshi, and J. W. Porter, J. Biol. Chem. 250, 4520 (1975);
A. A. Qureshi, M. Kim, F. A. Lornitzo, R. A. Jenik, and J. W. Porter, Biochem.
Biophys. Res. Commun. 64, 836 (1975); A. A. Qureshi, R. A. Jenik, M. Kim, F. A.
Lornitzo, and J. W. Porter, Biochem. Biophys. Res. Commun. 66, 344 (1975); W. I.
Wood, D. O. Peterson, and K. Bloch, J. Biol. Chem. 253, 2650 (1978).
85. M. Shin and R. Oshino, J. Biochem. (Tokyo) 83, 357 (1978).
86. C. Manzano, P. Candau, and M. G. Guerrero, Anal. Biochem. 90, 408 (1978).
87. K. Mailer, R. Poulson, D. Dolphin, and A. D. Hamilton, Biochem. Biophys. Res.
Commun. 96, 777 (1980).
88. A. H. Merrill and D. B. McCormick, J. Biol. Chem. 255, 1335 (1980).
89. E. Jablonski and M. DeLuca, Biochemistry 16, 2932 (1977).
90. L. Uotila and M. Koivusalo, Arch. Biochem. Biophys. 196, 33 (1979).
91. Same as 90.
92. K. Slavik, V. Zizkovsky, V. Siavikova, and P. Fort, Biochem. Biophys. Res. Com-
mun. 59, 1173 (1974).
93. H. Suda, K. Yamamoto, T. Aoyagi, and M. Umezawa, Biochim. Biophys. Acta
616, 60 (1980).
94. M. Endo and N. Hiyama, J. Biochem. (Tokyo) 86, 1959 (1979).
95. D. Robinson and R. Thorpe, FEBS Lett. 45, 191 (1974); J. A. Alhadeff, A. L.
Miller, H. Wenaas, T. Vedvick, and J. S. O'Brien, J. Biol. Chem. 250, 7106 (1975);
D. J. Opheim and O. Touster, J. Biol. Chem. 252, 739 (1977); J. A. Alhadeff and
G. L. Andrews-Smith, Biochem. J. 187, 45 (1980); S. F. Chien, and G. Dawson,
Biochim. Biophys. Acta 614, 475 (1980); R. S. Jain, A. Levy-Benshimol, C. A.
Buck, and L. Warren, J. Chromatogr. 139, 283 (1977).
96. S. Beeckmans and L. Kanarek, Eur. J. Biochem. 78, 437 (1977); S. Chaudhuri and
E. W. Thomas, Biochem. J. 177, 115 (1979).
97. N. Harpaz, H. M. Flowers, and N. Sharon, Eur. J. Biochem. 77, 419 (1977); J. W.
Kusiak, J. M. Quirk, R. O. Brady, and G. E. Mook, J. Biol. Chem. 253, 184 (1978).
98. A. L. Miller, R. G. Frost, and J. S. O'Brien, Biochem. J. 165, 591 (1977).
99. H. Hamazaki and K. Hotta, FEBS Lett. 76, 299 (1977); C. H. Kuo and W. W.
Wells, J. Biol. Chem. 253, 3550 (1978); P. V. Wagh, Biochim. Biophys. Acta 522,
515 (1978).
100. R. G. Frost, E. W. Holmes, A. G. W. Norden, and J. S. O'Brien, Biochem. J. 175,
181 (1978).
[1] AFFINITY CHROMATOGRAPHY 39

References to TABLE II (continued)

101. L. R. Glasgow, J. C. Paulson, and R. L. Hill, J. Biol. Chem. 252, 8615 (1977).
102. M. Rietschel-Berst, N. H. Jentoft, P. D. Rich, C. Pletcher, F. Fang, and J. E.
Gander, J. Biol. Chem. 252, 3219 (1977).
103. Y. Ben-Yoseph, E. Shapira, D. Edelman, B. K. Burton, and H. L. Nadler, Arch.
Biochem. Biophys. 184, 373 (1977).
104. P. H. Chan and W. Z. Hassid, Anal. Biochem. 64, 372 (1975).
105. H. Antinnen, R. Myllyla, and K. L Kivirikko, Biochem. J. 175, 737 (1978).
106. S. C. Magee, M. Ramswamp, and K. E. Ebner, Biochemistry 13, 99 (1974); I. H.
Fraser and S. Mookerjea, Biochem. J. 164, 541 (1977); A. Bella, J. S. Whitehead,
and Y. S. Kim, Biochem. J. 167, 621 (1977); Y. Legrand, G. Pignaud, and J. P.
Caen, FEBS Lett. 76, 294 (1977); C. A. Smith and K. Brew, J. Biol. Chem. 252,
7294 (1977); M. Pierce, R. D. Cummings, T. A. Cebula, and S. Roth, Biochim.
Biophys. Acta 571, 166 (1979); J. Mendicino, S. Sivakami, M. Davila, and V.
Chandrasekaran, J. Biol. Chem. 257, 3987 (1982).
107. M. J. Holroyde, J. M. E. Chesher, I. P. Trayer, and D. G. Walker, Biochem. J. 153,
351 (1976).
108. Y. H. Lee, P. C. Wankat, and H. A. Emery, Biotechnol. Bioeng. 13, 1639 (1976).
109. C.-Y. Lee, C. H. Langley, and J. Burkhart, Anal. Biochem. 86, 697 (1978); A.
Yoshida, J. Chromatogr. 114, 321 (1975); E. Burgissor and J. L. Fauchere, Heir.
Chim. Acta 59, 760 (1976); A. Morelli, U. Benatti, G. F. Gaetani, and A. DeFlora,
Proc. Natl. Acad. Sci. U.S.A. 75, 1979 (1978); N. Kato, H. Sahm, H. Schutte, and
F. Wagner, Biochim. Biophys. Acta 566, 1 (1979); M. L. Dan, J. L. Watson, R.
Delaney, and B. C. Johnson, J. Biol. Chem. 254, 9441 (1979).
110. P. R. Reczek and C. A. Villee, Biochem. Biophys. Res. Commun. 107, 1158
(1982).
111. J. Giudicelli, R. Emiliozzi, C. Vannier, G. DeBurlet, and P. Sudaka, Biochim.
Biophys. Acta 612, 85 (1980); T. A. Myohamen, V. Bouriotis, and P. D. G. Dean,
Biochem. J. 197, 683 (1981).
112. T. B. Breithaupt and R. J. Light, J. Biol. Chem. 257, 9622 (1982).
113. T. Yamaguchi and Y. Matsumura, Bioehim. Biophys. Acta 481, 706 (1977).
114. W. Leicht, M. M. Werber, and H. Eisenberg, Biochemistry 17, 4004 (1978); A. D.
McCarthy, J. M. Walker, and K. F. Tipton, Biochem. J. 191, 605 (1980); D. H.
Watson and J. C. Wootton, Bioehem. J. 167, 95 (1977).
115. C. N. G. Schmidt and L. Jervis, Anal. Bioehem. 104, 127 (1980).
116. R. Sekura and A. Meister, J. Biol. Chem. 252, 2599 (1977); G. Foure-Seelig and A.
Meister, J. Biol. Chem. 257, 5092 (1982).
117. M. Silink, R. Reddel, M. Bethel, and P. B. Rowe, J. Biol. Chem. 250, 5982
(1975).
118. V. P. Piglet and R. R. Conley, J. Biol. Chem. 252, 6367 (1977); I. Carlberg and B.
Mannervik, Biochim. Biophys. Acta 484, 268 (1977); G. Krohne-Ehrich, R. H.
Schirmer, and R. Untrucht-Crau, Fur. J. Biochem. 80, 65 (1977); V. Boggaram, T.
Brobjer, K. Larson, and B. Mannervik, Anal. Biochem. 98, 335 (1979); I. Carlberg
and B. Mannervik, Anal. Biochem. 116, 531 (1981); J. Danner, H. M. Lenhoff, and
W. Heagy, Anal. Biochem. 82, 586 (1977).
119. A. Grahnen and I. Sjoholm, Eur. J. Biochem. 80, 573 (1977); A. G. Clark and W. C.
Dauterman, Pept. Biochem. Physiol. 17, 307 (1982); P. C. Simons and D. L. V.
Jagt, Anal. Biochem. 82, 334 (1977); N. Pattinson, Anal. Biochem. 115, 424 (1981).
(continued)
40 c HROMATOGRAPHY [ 1]

References to TABLE II (continued)

120. C. R. Lowe, Eur. J. Biochem. 73, 265 (1977); J. D. Hillman, Biochem. J. 179, 99
(1979); J. I. Harris, J. D. Hocking, M. J. Runswick, K. Suzuki, and J. E. Walker,
Eur. J. Biochem. 108, 535 (1980).
121. D. W. Niesel, G. C, Bewley, S. G. Miller, F. B. Armstrong, and C. Y. Lee, J. Biol.
Chem. 255, 4073 (1980).
122. C. A. Capereli, C. Chettur, L. Y. Lin, and S. J. Benkovic, Biochem. Biophys. Res.
Commun. 82, 403 (1978).
123. N. B. Sorenson and P. Wang, Biochem. Biophys. Res. Commun. 67, 883 (1975).
124. L. R. Glasgow, Y. C. Paulson, and R. L. Hill, J. Biol. Chem. 252, 8615 (1977); T.
Mega and Y. Matsushima, J. Biochem. (Tokyo) 81, 571 (1977).
125. M. V. Kester and S. J. Norton, Biochim. Biophys. Acta 391, 212 (1975).
126. B. Oray and S. J. Norton, Biochim. Biophys. Acta 483, 203 (1977); N. Elango, S.
Janaki, and A. R. Rao, Biochem. Biophys. Res. Commun. 83, 1388 (1978); A. C.
Aronsson and B. Mannervik, Biochem. J. 165, 503 (1977); A. C. Aronsson, G.
Tibbelin, and B. Mannervik, Anal. Biochem. 92, 390 (1979); B. Oray and S. J.
Norton, Biochim. Biophys. Acta 611, 168 (1980).
127. B. Oray and S. J. Norton, Biochim. Biophys. Acta 611, 168 (1980).
128. J. K. Anderson, J. E. Mole, and H. J. Baker, Biochemistry 17, 467 (1978).
129. K. Hori, T. Kurotsu, M. Kanda, S. Miura, A. Nozoe, and Y. Saita, J. Biochem.
(Toyko) 84, 425 (1978).
130. J. E. Cone, J. Plowman, and G. Guroff, J. Biol. Chem. 249, 5551 (1974); R. L.
Then, Anal. Biochem. 100, 122 (1979).
131. D. L. Garbers, J. Biol. Chem. 251, 4071 (1976); M. Nakane and T. Deguchi, Bio-
chim. Biophys. Acta 52,5, 275 (1978); D. L. Garbers, J. Biol. Chem. 254, 240 (1979).
132. K. Ishiwata and H. Yoshida, J. Biochem. (Tokyo) 83, 783 (1978).
133. G. Monroy, E. Spencer, and J. Hurwitz, J. Biol. Chem. 253, 4481 (1978); S. Venka-
tesan, A. Gershowitz, and B. Moss, J. Biol. Chem. 255, 2829 (1980).
134. J. D. Bergstrom and A. L. Bieber, Prep. Biochem. 8, 275 (1978).
135. P. Izzo and R. Gantt, Biochemistry 16, 3576 (1977).
136. D. Moser, L. Johnson, and C.-Y. Lee, J. Biol. Chem. 255, 4673 (1980); C. L.
Wright, A. S. Warsy, M. J. Holroyde, and I. P. Trayer, Biochem. J. 175, 125 (1978);
E. Kopetzki and K. D. Entian, Anal. Biochem. 121, 181 (1982).
137. B. Geiger, Y. Ben-Yoseph, and R. Arnon, FEBS Lett. 45, 276 (1974).
138. K. Sandhoffl E. Conzelmann, and H. Nehrkorn, Hoppe Seyler's Z. Physiol. Chem.
358, 779 (1977); E. Conzelmann, K. Sandhoff, H. Nehrkorn, B. Geiger, and R.
Arnon, Eur. J. Biochem. 84, 27 (1978).
139. B. Geiger, E. Calet, and R. Arnon, Biochemistry 17, 1713 (1978).
140. S. B. Baylin and S. Margolis, Biochim. Biophys. Acta 397, 294 (1975).
141. A. J. Hacking and H. Hassall, Biochem. J. 147, 327 (1975).
142. J. K. Keesey, R. Bigelis, and G. R. Fink, J. Biol. Chem. 254, 7427 (1979).
143. C. H. Yang and P. N. Srivastava, Biochim. Biophys. Acta 391, 382 (1975); P. N.
Srivastava and A. A. Farooqui, Biochem. J. 183, 531 (1979); M. Lyon and C. F.
Phelps, Biochem. J. 199, 419 (1981).
144. T. Shimakata, Y. Fujita, and T. Kusaka, J. Biochem. (Tokyo) 86, 1191 (1979).
145. V. P. Agrawal and P. E. Kolattukudy, Arch. Biochem. Biophys. 191, 452 (1978).
146. G. C. Ness, C. D. Spindler, and M. M. Moffler, Arch. Biochem. Biophys. 197, 493
(1979); Z. H. Beg, J. A. Stonik, and H. B. Brewer, Biochem. Biophys. Res. Com-
mun. 107, 1013 (1982).
[1] AFFINITY CHROMATOGRAPHY 41

References to TABLE II (continued)

147. S. S. Braithwaite and J. Jarabak, J. Biol. Chem. 250, 2315 (1975).
148. M. Shikita and P. Talalay, Anal. Biochem. 95, 286 (1979).
149. Same as 148.
150. M. Moil and W. G. Wiest, J. Steroid Biochem. U , 1443 (1979).
151. F. Sweet and N. K. Adair, Biochem. Biophys. Res. Commun. 63, 99 (1975).
152. T. Kamei, Y. Takiguchi, H. Suzuki, M. Matsuzaki, and S. Nakamura, Chem.
Pharm. Bull. 26, 2799 (1978).
153. S. H. Hughes, G. M. Wahl, and M. R. Capecchi, J. Biol. Chem. 250, 120 (1975);
J. A. Holden and W. N. Kelley, J. Biol. Chem. 253, 4459 (1978); R. Schmidt,
H. Wiegand, and U. Reichert, Eur. J. Biochem. 93, 355 (1979).
154. G. Schmer and J. Roberts, Biochim. Biophys. Acta 527, 264 (1978).
155. Y. D. Clonis and C. R. Lowe, Eur. J. Biochem. 110, 279 (1980); P. E. Brodelius,
R. A. Lannom, and N. O. Kaplan, Arch. Biochem. Biophys. 18, 228 (1978);
H. J. Gilbert, C. R. Lowe, and W. T. Drabble, Biochern. J. 183, 481 (1979).
156. D. E. Nealon and R. A. Cook, Biochemistry 18, 3616 (1979).
157. R. Bigelis and H. E. Umbarger, J. Biol. Chem. 250, 4315 (1975).
158. R. Geiger, K. Mann, and T. Bettels, J. Clin. Chem. Clin. Biochem. 15, 479 (1977);
K. Uchida, M. Niimobe, M. Kato, and S. Fujii, Biochim. Biophys. Acta 614, 501
(1980).
159. N. A. Nwokoro and H. Schachter, J. Biol. Chem. 250, 6185 (1975).
160. Y. Nishimoto, F. Takeuchi, and Y. Shibata, J. Chromatogr. 169, 357 (1979).
161. N. O. Kaplan, J. Everse, J. E. Dixon, F. E. Stolzenbach, C. Y. Lee, C. T. Lee, S.
S. Taylor, and K. Mosbach, Proc. Natl. Acad. Sci. U.S.A. 71, 3450 (1974); C. R.
Lowe and K. Mosbach, Eur. J. Biochem. 52, 99 (1975); W. R. Ellington and G. L.
Long, Arch. Biochem. Biophys. 186, 265 (1978); S. Huang, C. Y. Lee, and S. S. L.
Li, Int. J. Biochem. 10, 279 (1979); A. R. Place, D. A. Powers, and C. Y. Lee, Anal.
Biochem. 83, 636 (1977); L. Jervis, Biochem. J. 197, 755 (1981).
162. Y. Horiuti and S. Imamura, J. Biochem. (Tokyo) 81, 1639 (1977); C.-S. Wang, Anal.
Biochem. 105, 398 (1980).
163. J. Visser and M. Strating, Biochim. Biophys. Acta 384, 69 (1975); J. Visser and M.
Strating, FEBS Lett. 57, 183 (1975).
164. C. Enholm, P. K. J. Kinnunen, J. K. Huttunen, E. A. Nikkila, and H. Ohta,
Biochem. J. 149, 649 (1975); D. Ganesan and H. B. Bass, FEBS Lett. 53, 1 (1975);
G. Bengtsson and T. Olivecrona, Biochem. J. 167, 109 (1977).
165. J. C. Allen, C. Eriksson, and J. R. Galpin, Eur. J. Biochem. 73, 171 (1977); T.
Sekiya, T. Kajiwara, and A. Hatanaka, Agric. Biol. Chem. 42, 677 (1978); I. Sha-
hin, S. Grossman, and B. Sredni, Biochim. Biophys. Acta 529, 300 (1978).
166. C. V. Dang and D. C. H. Yang, Biochem. Biophys. Res. Cornmun. 80, 709 (1978).
167. T. Yagi, T. Yamamoto, and K. Soda, Biochim. Biophys. Acta 614, 63 (1980).
168. J. Jolles and P. JoUes, Eur. J. Biochem. 54, 19 (1975); E. Junowicz and S. E.
Charm, FEBS Lett. 57, 219 (1975).
169. A. S. Narayanan, R. C. Siegel, and G. R. Martin, Arch. Biochem. Biophys. 162, 231
(1974).
17(I. C. A. Waters, J. R. Murphy, and J. W. Hastings, Biochem. Biophys. Res. Com-
mun. 57, 1152 (1974); J. C. Matthews, K. Hori, and M. J. Cormier, Biochemistry 16,
85 (1977); B. R. Branchini, T. M. Marschner, and A. M. Montemurro, Anal. Bio-
chem. 104, 386 (1980); T. F. Holzman and T. O. Baldwin, Biochemistry 21, 6194
(1982).
(continued)
42 CHROMATOGRAPHY [1]

References to TABLE II (continued)

171. D. Bout, H. Dupas, M. Capron, A. EI-Gazawi, Y. Carlier, A. Delacourte, and A.
Capron, lmmunochemistry 15, 633 (1978); I. P. Wright and T. K. Sundaram, Bio-
chem. J. 177, 441 (1979); C.-Y. Lee, C. H. Langley, and J. Burkhart, Anal. Bio-
chem. 86, 697 (1978); C.-Y. Lee, D. Charles, and D. Bronson, J. Biol. Chem. 254,
6375 (1979); M. Mevarech, H. Eisenberg, and E. Neumann, Biochemistry 16, 3781
(1977).
172. M. Elwell and L. B. Hersh, J. Biol. Chem. 254, 2434 (1979).
173. T. Caldes, H. R. Fatania, and K. Dalziel, Anal. Biochem. 100, 299 (1979); J. T.
Chang and G. G. Chang, Anal. Biochem. 121, 366 (1982).
174. Y. S. Kim and P. E. Kolattukudy, Arch. Biochem. Biophys. 190, 234 (1978); Y. S.
Kim, P. E. Kolattukudy, and A. Boss, Arch. Biochem. Biophys. 196, 543 (1979).
175. K. H. Schachtele, E. Schlitz, and D. Palm, Eur. J. Biochem. 92, 427 (1978).
176. N. C. Phillips, D. Robinson, and B. Winchester, Biochem. J. 153, 579 (1976); P. V.
Wagh, J. Chromatogr. 152, 565 (1978); N. C. Phillips, B. G. Winchester, and R. D.
Jolly, Biochem. J. 163, 269 (1977).
177. G. Fayat, M. Fromant, D. Kahn, and S. Blanquet, Eur. J. Biochem. 78, 333 (1977);
O. Kellermann, A. Brevet, H. Tonetti, and J. P. Waller, Eur. J. Biochem. 88, 205
(1978).
178. S. Riazuddin and T. Lindahl, Biochemistry 17, 2110 (1978).
179. V. V. Murthy, E. Jones, T. W. Cole, and J. Johnson, Biochim. Biophys. Acta 483,
487 (1977).
180. E. Barbosa and B. Moss, J. Biol. Chem. 253, 7692 (1978).
181. K. Sato, E. Hiei, S. Shimizu, and R. H. Abeles, FEBS Lett. 85, 73 (1978).
182. R. G. Dennick and R. J. Mayer, Biochem. J. 161, 167 (1977); C. M. Buess, J. K.
Price, B. D. Roberts, and W. R. Carper, Experientia 33, 163 (1977).
183. E. M. V. Pires and S. V. Perry, Biochem. J. 167, 137 (1977).
184. R. C. Bhalla, R. V. Sharma, and R. C. Gupta, Biochem. J. 203, 583 (1982).
185. M. Tseng, B. G. Harris, and M. K. Jacobson, Biochim. Biophys. Acta 568, 205
(1979).
186. T. Sugiyama and T. Yamano, FEBS Lett. 52, 145 (1975).
187. J. L. Vermillion and M. J. Coon, J. Biol. Chem. 253, 2694 (1978); Y. Yasukochi and
B. S. Siler Masters, J. Biol. Chem. 251, 5337 (1976); J. S. French and M. J. Coon,
Arch. Biochem. Biophys. 195, 565 (1979); K. M. Madyastha and C. J. Coscia, J.
Biol. Chem. 254, 2419 (1979); R. T. Mayer and J. T. Durrant, J. Biol. Chem. 2,54,
756 (1979).
188. J. I. Pedersen and H. K. Godager, Biochim. Biophys. Acta 525, 28 (1978).
189. G. A. Michaliszyn, S. S. Wing, and E. A. Meighen, J. Biol. Chem. 252, 7495 (1977).
190. R. Skorko, W. Zillig, H. Rohrer, H. Fujiki, and R. Mailhammer, Eur. J. Biochem.
799 55 (1977).
191. B. Venerando, G. Tettamanti, B. Cestaro, and V. Zambotti, Biochim. Biophys.
Acta 403, 461 (1975);Y. Uchida, Y. Tsukada, and T. SugJmori,J. Biochem. (Tokyo)
82, 1425 (1977).
192. B. Holmquist, Biochemistry 16, 4591 (1977).
193. S. Ida, J. Biochem. (Tokyo) 82, 915 (1977).
194. H. Tanaka, I. Sasaki, K. Yamashita, Y. Matuo, J. Yamashita, and T. Horio, J.
Biochem. (Tokyo) 91, 1411 (1982).
[1] AFFINITY CHROMATOGRAPHY 43

References to TABLE II (continued)

195. H. Taniuchi and J. L. Bohnert, J. Biol. Chem. 250, 2388 (1975).
196. J. Dornand, J. C. Bonnafons, and J. C. Mani, Eur. J. Biochem. 87, 459 (1978).
197. N. Miki, J. M. Baraban, J. J. Keirns, J. J. Boyce, and M. W. Bitensky, J. Biol.
Chem. 250, 6320 (1975).
198. E. F. Brunngraber and E. Chargaff, Proc. Natl. Acad. Sci. U.S.A. 74, 3226 (1977).
199. E. Bischoff, T. A. Tran-Thi, and K. F. A. Decker, Eur. J. Biochem. 51, 353 (1975).
200. R. J. Boucek and K. J. Lembach, Arch. Biochem. Biophys. 184, 408 (1977).
201. N. J. Hoogenraad, T. M. Sutherland, and G. J. Howlett, Anal. Biochem. 101, 97
(1980); M. Luisa de Martinis, P. Mclntyre, and N. Hoogenraad, Biochem. Int. 3,
371 (1981).
202. H. Rosemeyer and F. Seela, J. Med. Chem. 22, 1545 (1979).
203. G. Dodin, FEBS Lett. 134, 20 (1981).
204. P. F. Fox, J. R. Whitaker, and P. A. O'Leary, Biochem. J. 161, 389 (1977).
205. A. Signoret and J. Crouzet, Agric. Biol. Chem. 46, 459 (1982).
206. Y. Tanaka and I. Uritani, J. Biochem. (Tokyo) 81, 963 (1977).
207. R. G. H. Cotten and P. J. Grattan, Eur. J. Biochem. 60, 427 (1975); J. M. AI-Janabi,
Arch. Biochem. Biophys. 200, 603 (1980); S. Webber, G. Harzer, and J. M. White-
Icy, Anal. Biochem. 106, 63 (1980).
208. G. S. Swamy and D. T. N. Pillay, Z. Pflanzenphysiol. 93, 403 (1979).
209. T. Hirabayashi, T. J. Larson, and W. Dowhan, Biochemistry 15, 5205 (1976); T. J.
Larson, T. Hirabayashi, and W. Dowhan, Biochemistry 15, 974 (1976).
210. M. Miyake, J. W. Daly, and C. R. Creveling, Arch. Biochem. Biophys. 181, 39
(1977); R. L. Kincaid and M. Vaughan, Proc. Natl. Acad. Sci. U.S.A. 76, 4903
(1979).
211. G. Colombo, G. M. Carlson, and H. A. Lardy, Biochemistry 17, 5321 (1978);
B. S. T. Hung and R. Silverstein, Prep. Biochem. 8, 421 (1978).
212. H. Hengartner and J. I. Harris, FEBS Lett. 55, 282 (1975); J. Babul, J. Biol. Chem.
253, 4350 (1978).
213. N. Kato, H. Sahm, H. Schutte, and F. Wagner, Biochim. Biophys. Acta 566, 1
(1979).
214. C. Y. Lee, D. Charles, and D. Bronson, J. Biol. Chem. 254, 6375 (1979).
215. G. W. K. Kuntz, S. Eber, W. Kessler, H. Krietsch, and W. K. Krietsch, Eur. J.
Biochem. 85, 493 (1978); B. Pegoraro and C. Y. Lee, Biochim. Biophys. Acta 522,
423 (1978).
216. C. Ehnholm, W. Shaw, H. Greten, and W. V. Brown, J. Biol. Chem. 250, 6756
(1975).
217. C. O. Rock and F. Snyder, J. Biol. Chem. 250, 6564 (1975); R. M. Kramer, C.
Wuthrich, C. Bollier, P. R. Allegrini, and P. Zahler, Biochim. Biophys. Acta 507,
381 (1978).
218. T. Takahaski, T. Sugahara, and A. Ohsaka, Biochim. Biophys. Acta 351, 155
(1974); C. Little, B. Aurebekk, and A. B. Otnaess, FEBS Lett. 52, 175 (1975).
219. V. P. K. Titanji and S. Pahlman, Biochim. Biophys. Acta 525, 380 (1978).
220. P. Gergely, A. G. Castle, and N. Crawford, Int. J. Biochem. 10, 807 (1979).
221. S. Yonezawa and S. H. Hori, Arch. Biochem. Biophys. 181, 447 (1977).
222. R. K. Sharma, S. W. Tam, D. M. Waisman, and J. H. Wang, J. Biol. Chem. 255,
11102 (1980).

(contmued)
44 CHROMATOGRAPHY [ 1]

References to TABLE II (continued)

223. M. Grez and J. Niessing, FEBS Lett. 77, 57 (1977); J. R. Nevins and W. K. Joklik,
J. Biol. Chem. 252, 6939 (1977); D. Antoniades and O. Antonoglou, Biochim.
Biophys. Acta 519, 447 (1978).
224. K. Yoshsihara, T. Hashida, Y. Tanaka, H. Ohgushi, H. Yosihara, and T. Kamiya,
J. Biol. Chem. 253, 6459 (1978); T. Kristensen and J. Holtlund, Eur. J. Biochem. 88,
495 (1978).
225. H. Soreg and U. Z. Littauer, J. Biol. Chem. 252, 6885 (1977).
226. D. J. Tutas and E. Paoletti, J. Biol. Chem. 252, 3092 (1977).
227. T. Yoshimoto, M. Fischl, R. C. Orlowski, and R. Walter, J. Biol. Chem. 253, 3708
(1978).
228. L. Tuderman, E.-R. Kuutti, and K. I. Kivirikko, Eur. J. Biochem. 52, 9 (1975); K.
Tryggvason, K. Majamaa, J. Risteli, and K. I. Kivirikko, Biochem. J. 183, 303
(1979).
229. R. A. Cravel, K. F. Lam, D. Mahuran, and A. Kronis, Arch. Biochem. Biophys.
201, 669 (1980).
230. M. Hemler, W. E. M. Lands, and W. L. Smith, J. Biol. Chem. 251, 5575 (1976).
231. G. Feinstein and A. Janoff, Biochim. Biophys. Acta 403, 477 (1975); E. J. B. Fodor,
H. Ako, and K. A. Walsh, Biochemistry 14, 4923 (1975); S. Vavreinova and J.
Turkova, Biochim. Biophys. Acta 403, 506 (1975); D. Y. Twumasi and I. E. Liener,
J. Biol. Chem. 252, 1917 (1977).
232. W. L. Dills, J. A. Beavo, P. J. Bechtel, and E. G. Krebs, Biochem. Biophys. Res.
Commun. 62, 70 (1975); G. Shanker, H. Ahrens, and R. K. Sharma, Proc. Natl.
Acad. Sci. U.S.A. 76, 66 (1979); P. K. Tsung, T. Sakamoto, and G. Weissmann,
Biochem. J. 145, 437 (1975); G. Vauquelin, P. Geynet, J. Hanoune, and A. D.
Strosberg, Proc. Natl. Acad. Sci. U.S.A. 74, 3710 (1977); G. N. Gill, K. E. Holdy,
G. M. Walton, and C. B. Kanstein, Proc. Natl. Acad. Sci. U.S.A. 73, 3918 (1976);
Y. Suzuki, E. Itugaki, H. Mori, and T. Hosoya, J. Biochem. (Tokyo) 81, 1721
(1977); F. Farron-Furstenthal and J. R. Lightholder, FEBS Lett. 84, 313 (1977); W.
Thornburg, A. F. O'Malley, and T. J. Lindell, J. Biol. Chem. 253, 4638 (1978); Y.
Kitagawa and E. Racker, J. Biol. Chem. 257, 4547 (1982).
233. P. K. Tsung, T. Sakamoto, and G. Weissmann, Biochem. J. 145, 437 (1975).
234. P. K. Saini and I. H. Rosenberg, J. Biol. Chem. 249, 5131 (1974).
235. B. S. Enevoldsen, L. Reimann, and N. L. Hansen, FEBS Lett. 79, 121 (1977).
236. V. Zannis, D. Doyle, and D. W. Martin, J. Biol. Chem. 253, 504 (1978); W. R. A.
Osborne, J. Biol. Chem. 255, 7089 (1980); D. A. Wiginton, M. S. Coleman, and J. J.
Hutton, J. Biol. Chem. 255, 6663 (1980).
237. W. M. Anderson and R. R. Fisher, Arch. Biochem. Biophys. 187, 180 (1978); G.
Voordouw, S. M. Vandervies, J. K. Eweg, C. Veeger, J. F. L. Van Breeman, and
E. F. J. Van Bruggen, Fur. J. Biochem. 111, 347 (1980).
238. C. D. Cash, M. Maitre, J. F. Rumigny, and P. Mandel, Biochem. Biophys. Res.
Commun. 96, 1755 (1980).
239. Same as 238.
240. B. A. Klyashchitsky, V. F. Pozdnev, V. K. Mitina, A. I. Voskoboyev, and I. P.
Chernikevich, Bioorg. Khim. 6, 1572 (1980).
241. J. Visser, M. Strating, and W. Van Dongen, Biochim. Biophys. Acta 524, 37 (1978).
242. R. I. Zemmell and R. A. Anwar, J. Biol. Chem. 250, 3185 (1975).
[1] AFFINITY CHROMATOGRAPHY 45

References to TABLE II (continued)

243. K. Murakami and T. Inagami, Biochem. Biophys. Res. Commun. 62, 757 (1975); C.
Devaux, J. Manard, P. Sicard, and P. Corvol, Eur. J. Biochem. 64, 621 (1976); T.
Inagami and K. Murakami, J. Biol. Chem. 252, 2978 (1977); H. J. Chou, J. H.
Sharper, and R. I. Gregerman, Biochim. Biophys. Acta 524, 183 (1978); H. Yoko-
sawa, T. Inagami, and E. Haas, Biochem. Biophys. Res. Commun. 83, 306 (1978);
T. Matoba, K. Murakami, and T. Inagami, Biochim. Biophys. Acta 526, 560 (1978).
244. J. Reiser and R. Yuan, J. Biol. Chem. 252, 451 (1977).
245. M. G. Sarngadharan, V. S. Kalyanaraman, R. Rahman, and R. C. Gallo, J. Virol.
35, 555 (1980).
246. R. E. Scofield, R. P. Werner, and F. Wold, Anal. Biochem. 77, 152 (1977).
247. H. Yoshida, I. Fukuda, and M. Hushiguchi, J. Biochem. (Tokyo) 88, 1813 (1980).
248. J. G. Stavrianopoulos and E. Chargaff, Proc. Natl. Acad. Sci. U.S.A. 75, 4140
(1978).
249. S. Kanaya and T. Uchida, J. Biochem. (Tokyo) 90, 473 (1981).
250. T. Uchida and Y. Shibata, J. Biochem. (Tokyo) 90, 463 (1981).
251. P. J. Hoffman and R. L. Blakely, Biochemistry 14, 4804 (1975).
252. S. Sugiura, M. Suzuki, E. Ohtsuka, S. Nishikawa, H. Uemura, and M. Ikehara,
FEBS Lett. 97, 73 (1979).
253. P. Schofield and K. R. Williams, J. Biol. Chem. 252, 5584 (1977).
254. R. R. Burgess and J. J. Jendrisak, Biochemistry 14, 4634 (1975); K. Amemiya,
C. W. Wu, and L. Shapiro, J. Biol. Chem. 252, 4157 (1977); M. I. Goldberg,
J. C. Perriard, and W. J. Rutter, Biochemistry 16, 1655 (1977); K. H. Scheit and A.
Stutz, FEBS Lett. 50, 25 (1975); H. Sternbach, R. Engelhardt, and A. G. Lezius,
Eur. J. Biochem. 60, 51 (1975).
255. P. Hossenlopp, J. Sumegi, and P. Chambon, Eur. J. Biochem. 90, 615 (1978); E.
Gundelfinger, H. Saumweber, A. Dallendorfer, and H. Stein, Eur. J. Biochem. 111,
395 (1980); T. M. Wandzilak and R. W. Benson, Biochemistry 17, 426 (1978).
256. M. I. Goldberg, J. C. Perriard, and W. J. Rutter, Biochemistry 16, 1648 (1977).
257. K. K. Cheung and A. Newton, J. Biol. Chem. 253, 2254 (1978); P. A. Lowe, D. A.
Hager, and R. R. Burgess, Biochemistry 18, 1344(1979); T. Nakayama, V. William-
son, K. Burtis, and R. H. Doi, Eur. J. Biochem. 88, 155 (1978).
258. K. S. You and C. R. Roe, Anal. Biochem. 114, 177 (1981).
259. M. J. Geison, Biochem. J. 151, 181 (1975).
260. J. C. Paulson, W. E. Beranek, and R. L. Hill, J. Biol. Chem. 252, 2356 (1977).
261. N. Leissing and E. T. McGuinness, Biochim. Biophys. Acta 524, 254 (1978).
262. R. L. Pajula, A. Raina, and J. Kekoni, FEBS Lett. 90, 153 (1978).
263. J. W. Callahan, P. Shankaran, M. Khalil, and J. Gerrie, Can. J. Biochem. 56, 885
(1978).
264. A. Fenselau and K. Wallis, Biochem. Biophys. Res. Commun. 62, 350 (1975).
265. D. J. Ball and J. J. Nishimura, J. Biol. Chem. 255, 10805 (1980).
266. C. D. Cash, M. Maitre, L. Ossola, and P. Mandel, Biochim. Biophys. Acta 524, 26
(1978).
267. S. Barry, P. Brodelius, and K. Mosbach, FEBS Lett. 70, 261 (1976).
268. S. Okamura, F. Crane, H. A. Messner, and T. W. Mak, J. Biol. Chem. 253, 3765
(1978).
269. V. P. Pigiet and R. R. Conley, J. Biol. Chem. 252, 6367 (1977).

(continued)
46 CHROMATOGRAPHY [ 1]

References to TABLE II (continued)

270. K. Koerner, I. Rahimi-Laridjani, and H. Grimminger, Biochim. Biophys. Acta 397,
220 (1975).
271. R. Bhadra and P. Datta, Biochemistry 17, 1691 (1978).
272. W. H. Holleman and L. J. Weiss, J. Biol. Chem. 251, 1663 (1976).
273. R. A. Madhav, M. L. Coetzee, and P. Ove, Arch. Biochem. Biophys. 200, 99 (1980).
274. J. M. Whiteley, I. Jerkunica, and T. Deits, Biochemistry 13, 2045 (1974); B. J.
Dolnick and Y. C. Cheng, J. Biol. Chem. 252, 7697 (1977); W. Rode, K. J. Scanlon,
J. Hynes, and J. R. Bertino, J. Biol. Chem. 2,54, 11538 (1979); C. K. Banerjee, L. L.
Bennett, R. W. Brockman, B. P. Sani, and C. Temple, Anal. Biochem. 121, 275
(1982).
275. K. Yamamoto and L. J. DeGroot, 7. Biochem. (Tokyo) 91, 775 (1982).
276. M. Berger and H. G. Wood, 7. Biol. Chem. 250, 927 (1975).
277. L. N. Y. Wu, R. M. Pennington, T. D. Everett, and R. R. Fisher, J. Biol. Chem.
257, 4052 (1982).
278. C. Enholm, P. K. J. Kinnunen, J. K. Huttunen, E. A. Ni~ila, and M. Ohta,
Biochem. J. 149, 649 (1975).
279. S. Ikeda, H. Hara, S. Sugimoto, and S. Fukui, FEBS Lett. 56, 307 (1975).
280. H. Nakata and H. Fujisawa, Eur. J. Biochem. 122, 41 (1982).
281. D. H. Wolf and M. Hoffmann, Eur. J. Biochem. 45, 269 (1974); M. P. Gschwind, U.
Gschwind, C. H. Paul, and K. Kirschner, Eur. J. Biochem. 96, 403 (1979).
282. K. Nishioka, Eur. J. Biochem. 85, 137 (1978).
283. P. Donner, H. Wagner, and H. Kroger, Biochem. Biophys. Res. Commun. 80, 766
(1978).
284. Same as 281.
285. C. R. Geren and K. E. Ebner, J. Biol. Chem. 252, 2082 (1977).
286. C-. A. Smith and K. Brew, J. Biol. Chem. 252, 7294 (1977).
287. C. R. Geren, C. M. Oloman, D. C. Primrose, and K. E. Ebner, Prep. Biochem. 7,
19 (1977).
288. B. Burchell. FEBS Lett. 78, 101 (1977); B. Burchell, Biochem. J. 173, 749 (1978).
289. R. A. Anwar and M. Vlaovic, Can. J. Biochem. 57, 188 (1979).
290. T. Watanabe and T. Suga, Anal. Biochem. 89, 343 (1978); T. Watanabe and T.
Suga, Anal. Biochem. 861, 357 (1978).
291. F. Batista-Viera, R. Ax6n, and J. Carlsson, Prep. Biochem. 7, 103 (1977).
292. B. Astedt, L. Holmberg, G. Wagner, P. Richter, and J. Ploug, Thromb. Haemosta-
sis 42, 924 (1979).
293. E. M. Clarke and J. R. Knowles, Biochem. J. 167, 419 (1977).
[1] AFFINITY CHROMATOGRAPHY 47

TABLE Ill
PROTEINSPURIFIEDBY IMMUNOAFFINITYCHROMATOGRAPHYON
ANTIBODYCOLUMNS

Proteina Eluent

I. Acetyl-fl-glucosaminidase NaC1
2. Adenosine deaminase Urea
3. /3-Adrenergic receptor Propranolol
4. Alkaline phosphatase High pH
5. t~-Fetoprotein Urea
6. t~rMacroglobulin Low pH
7. Aminopeptidase Tris-HCl
8. Angiotensin I converting enzyme MgC12
9. Astroglial protein Acetic acid-urea
10. Clg, Clr, Cls subcomponents EDTA, ethylene glycol
11. Complement (C5) KBr
12. Carcinoembryonic antigen Urea
13. Cardiotoxin Acetic acid-NaCl
14. Catalase Glycerol
15. Choriogonadotropin MgCI2
16. Cellulase Low pH
17. Cyclic nucleotide phosphodiesterase EGTA
18. Estrogen receptor NaSCN
19. Factor VIII NH4SCN
20. Factor V NaC1
21. Factor IX NaC1, KSCN
22. Glucocorticoid receptor Acetic acid
23. /3-Glucuronidase Urea
24. Hepatitis A virus NaI
25. H-2Kk antigen TNP-40, NaC1
26. HLA-B and B Diethylamine pH 11.5
27. HCG-receptor complex HCG
28. Insulin receptor MgC12
29. Interferon Ethylene glycol + citrate
30. Kallikreins Guanidine
31. Legumin KSCN
32. Lymphocyte
33. Monoamine oxidase KSCN
34. Membrane glycoproteins Diethylamine
35. Myoglobin KCN-pH
36. Myosin Guanidine
37. Neutrophil migration inhibition factor (NIF) Urea
38. Phosphotyrosine proteins Phenyl phosphate
39. Poliovirus KSCN
40. Properdin NaCI
41. Ribonuclease H MgCI2
42. Serine acetyltransferase O-Acetylserine

(continued)
 TABLE Ill (continued)

Protein Eluent

43. Sm and RNP antigens Urea, guanidine

44. Shiga toxin MgC1
45. Somatostatin Acetic acid
46. Terminal deoxynucleotidyltransferase NH4OH
47. Tetrodotoxin pH
48. Thyrotropin Guanidine
49. TSH Glycine-HC1
50. Transplantation antigen
51. Trypsin inhibitor Propionic acid
52. Urokinase Glycine-HCl

a Key to references:
1. R. P. Erickson and R. Sandman, Experientia 33, 14 (1977).
2. W. P. Schrader and A. R. Stacy, J. Biol. Chem. 252, 6409 (1977).
3. C. M. Frazer and J. C. Venter, Proc. Natl. Acad. Sci. U.S.A. 77, 7034
(1980).
4. J. J. Hughes and S. E. Charm, Biotechnol. Bioeng. 21, 1439 (1979).
5. U. H. Stenman, M. L. Sutinen, R. K. Selander, K. Tontii, and J. Schro-
der, J. Immunol. Methods 46, 337 (1981).
6. J. E. McEntire, J. lmmunol. Methods 24, 39 (1978).
7. H. Sj6str6m, O. Noren, L. Jeppesen, M. Stuan, B. Svensson, and L.
Christiansen, Eur. J. Biochem. 88, 503 (1978).
8. J. J. Lanzillo, R. P. Cynkin, and B. L. Fanburg, Anal. Biochem. 103, 400
(1980).
9. D. C. Rueger, D. Dahl, and A. Bignani, Anal. Biochem. 89, 360 (1978); D.
Dahl, Biochim. Biophys. Acta 622, 9 (1980).
10. Y. Moil, J. Chromatogr. 189, 428 (1980).
11. R. A. Wetsel, M. A. Jones, and W. P. Kolb, J. Immunol. Methods 35, 319
(1980).
12. E. Engvall, J. E. Shiveli, and M. Wronn, Proc. Natl. Acad. Sci. U.S.A.
75, 1670 (1978).
13. E.-H. Wong, C.-L. Ho, and K.-T. Wang, J. Chromatogr. 154, 25' (1978).
14. S. Webber and J. M. Whiteley, Biochem. Biophys. Res. Commun. 97, 17
(1980).
15. C. Y. Lee, S. Wong, A. S. K. Lee, and L. Ma, Hoppe Seyler's Z. Physiol.
Chem. 358, 909 (1977).
16. U. Hakansson, L. G. Fagerstam, L. G. Pettersson, and L. Anderson,
Biochem. J. 179, 141 (1979).
17. R. Scott, R. S. Hansen, and J. A. Beavo, Proc. Natl. Acad. Sci. U.S.A.
79, 2788 (1982).
18. B. Moncharmont, J.-L. Su, and I. Parikh, Biochemistry 21, 6916 (1982).
19. J. L. Lane, H. Ekert, and A. Vafiadis, Thromb. Haemostasis 42, 1306
(1979).
20. J. A. Katzmann, M. E. Neshiem, L. S. Hibbard, and K. G. Mann, Proc.
Natl. Acad. Sci. U.S.A. 78, 162 (1981).
21. A. H. Goodall, G. Kemble, D. P. O'Brien, E. Rawlings, F. Rotblat, G. C.
Russell, G. Janossy, and G. D. Tuddenham, Blood 59.. 664 (1982).
22. H. J. Eisen, Proc. Natl. Acad. Sci. U.S.A. 77, 3893 (1980).

48
References to TABLE III (continued)
23. F. E. Brot, C. E. Bell, and W. S. Sly, Biochemistry 17, 385 (1978).
24. Y. Elkana, A. Thornton, and A. J. Zuckerman, J. Immunol. Methods 25,
185 (1979).
25. J. E. Mole, F. Hunter, J. W. Paslay, A. S. Bhown, and J. C. Bennett, Mol.
lmmunol. 19, 1 (1982).
26. P. Parham, J. Biol. Chem. 254, 8709 (1979).
27. K. Metsikko and H. Rajaniemi, FEBS Lett. 1116, 193 (1979).
28. L. C. Harrison and A. Itin, J. Biol. Chem. 255, 12066 (1980).
29. D. S. Secher and D. C. Burke, Nature (London) 285, 446 (1980).
30. K. M. Gautvik, L. Johansen, K. Svindah|, K. Nustad, and T. B. Or-
stavik, Biochem. J. 189, 153 (1980); R. Geiger and B. Clausnitzer, Hoppe
Seyler's Z. Physiol. Chem. 362, 1279 (1981).
31. R. Casey, Biochem. J. 177, 509 (1979).
32. D. M. Raulet, P. D. Gottlieb, and M. J. Bevan, J. Immunol. 12,5, 1136
(1980).
33. R. M. Denney, R. R. Fritz, N. T. Patel, and C. W. Abell, Science 215,
1400 (1982).
34. E. N. Hughes and J. T. August, J. Biol. Chem. 257, 3970 (1982).
35. H. K. B. Simmerman, C.-C. Wang, E. M. Horwitz, J. A. Berzofsky, and
F. R. N. Good, Proc. Natl. Acad. Sci. U.S.A. 79, 7739 (1982).
36. S. Sartore, L. D. Libera, and S. Schiaffino, FEBS Lett. 106, 197 (1979).
37. R. H. Weisbart, A. J. Lusis, A. Kacena, L. Spolter, P. Eggena, and D. W.
Golde, Clin. lmmunol. Immunopathol. 22, 408 (1982).
38. A. H. Ross, D. Baltimore, and H. N. Eisen, Nature (London) 294, 654
(1981).
39. F. Brown, B. O. Underwood, and K. H. Fantes, J. Med. Virol. 4, 315
(1979).
40. C. K. Ogle, J. D. Ogle, and J. W. Alexander, Immunochemistry 14, 341
(1977).
41. W. Biusen, J. Biol. Chem. 257, 7106 (1982).
42. P. A. Baecker and R. T. Wedding, Anal. Biochem. 102, 16 (1980).
43. P. J. White, W. D. Gardner, and S. O. Hoch, Proc. Natl. Acad. Sci.
U.S.A. 78, 626 (1981); R. R. C. Buchanan, P. J. W. Venables, A. Morgan,
N. A. Staines, P. R. Smith, and R. N. Maini, Clin. Exp. lmmunol. 51, 8
(1983).
44. A. D. O'Brien, G. D. LaVeck, D. E. Griffin, and M. R. Thompson, Infect.
Immun. 30, 170 (1980).
45. M. Lauber, M. Camier, and P. Cohen, Proc. Natl. Acad. Sci. U.S.A. 76,
6004 (1979).
46. S. Srivastava, J. Y. H. Chan, and F. A. Siddiqui, J. Biochem. Biophys.
Methods 2, 1 (1980).
47. H. Nakayama, R. M. Withy, and M. A. Raftery, Proc. Natl. Acad. Sci.
U.S.A. 79, 7575 (1982).
48. F. Pekonen, D. M. Williams, and B. D. Weintraub, Endocrinology (Balti-
more) 1116, 1327 (1980).
49. G. Ponsin and R. Mornex, Acta Endocrinol. 93, 430 (1980).
50. G. C. DuBois, L. W. Law, and E. Apella, Proc. Natl. Acad. Sci. U.S.A.
79, 7669 (1982).
51. R. S. Corfman and G. R. Reeck, Biochim. Biophys. Acta 715, 170 (1982).
52. T. C. Wun, L. Ossowski, and E. Reich, J. Biol. Chem. 257, 7262 (1982).

49
50 CHROMATOGRAPHY [ 1]

TABLE IV
RECEPTORAND BINDINGPROTEINSPURIFIEDBY AFFINITYCHROMATOGRAPHY

Receptor ° Ligand Fluent

1. Acetylcholine Toxin Decamethonium

2. a2-Adrenergic 3-Benzazepine Agonists, antagonists
3. fl-Adrenergic AIprenolol, acebutolol Isoproterenol, alprenolol
4. Androgen Heparin, DNA Heparin, NaC1
5. Asialoglycoprotein Asialo-orosomucoid Galactose
6. Cyclic nucleotide cAMP cAMP, cGMP
7. 1,25-Dihydroxy- DNA, heparin, calciferol KC!
vitamin D
8. Epidermal growth factor Wheat germ agglutinin N-Acetylglucosamine
9. Estradiol Estradiol derivative Estradiol, KCI
oligo(dT)
10. Fc IgG Urea
11. Glucocorticoid DNA, ATP NaCI, pyridoxal phosphate
12. Glycine Aminostrychnine Glycine
13. Growth hormone Growth hormone Urea, MgCl
14. IgE IgE, phenylarsonate IgE Acetic acid, KSCN phenyl-
azotyrosine
15. Insulin Insulin Guanidine, KSCN
16. Intrinsic factor Intrinsic factor-B12 EDTA
17. Macrophage (C3) (C3) Acetic acid
18. Opiate Amidomorphine Levorphanol
19. Progesterone ATP KC1
20. Prolactin Prolactin Urea, MgCl
21. Transcobalamin II Transcohalamin II EDTA
22. Transferrin Transferrin Transferrin, electrophoresis
23. Triiodothyronine (T3) T3 T3
24. TSH TSH KSCN
Binding protein
25. Actin Deoxyribonuclease I, myo- Guanidine ATP
sin
26. Androgen Dihydrotestosterone Dihydrotestosterone
27. C-reactive Phosphorylcholine Ca2+
28. C3b component Factor H NaCl
29. Calmodulin Phenothiazine EDTA
30. Corticosteroid-binding Androstene, cortisol Cortisol
globulin
31. cAMP cAMP Urea, cAMP
32. Cytokinin Benzyladenine isopen- NaOH, KC1
tenyladenosine
33. Dextran Sephadex Guanidine
34. DNA DNA NaCI
35. Elongation factors GDP GDP
Tu, Ts
[1] AFFINITY CHROMATOGRAPHY 51

TABLE IV (continued)

Receptor ~ Ligand Eluent

36. Fibronectin (fragments) Gelatin Urea, arginine

37. Guanosine GDP GDP
38. Hemopexin Heme, hematin Gly-HC1, urea
39. Initiation factor elF-3 rRNA KCI
40. Low- and high-density Cholic acid Triton X-100, hydroxyl-
lipoproteins amine
41. Meromyosin Actin ADP, ATP
42. Messenger ribonucleo- Oligo(dT) Temperature change
proteins
43. Migration inhibitory Glycolipids KSCN
factor
44. Myosin ADP, actin ADP, ATP
45. Neurophysin Met-Tyr-Phe Acetic acid
46. Nonhistone chromatin DNA NaC1
47. Penicillin Ampicillin, 6-APA Hydroxylamine
48. Phosphatidylcholine Phosphocholine Deoxycholate
49. Retinal Prealbumin NaCI
50. Riboflavin Flavin NaCI
51. r-Ribosomal proteins rRNA KCI-EDTA
52. t-Ribosomal proteins tRNA Urea-LiC1
53. mRNA cap mTGDP KC1
54. Serotonin Serotonin Ca2+
55. Sex steroid Dihydrotestosterone Dihydrotestosterone
56. Tubulin Lactoperoxidase NaCI
57. Vimentin DNA NaC1
58. Z-DNA Poly(dG-dC) NaC1

a Key to references:
1. J. P. Merlie, J.-P. Changeux, and F. Gros, J. Biol. Chem. 253, 2882 (1978); D.
Barffeld and S. Fucs, Biochem. Biophys. Res. Commun. 89, 512 (1979); N. Kalderon
and I. Silman, Biochim. Biophys. Acta 465, 331 (1977).
2. J. W. Regan, N. Barden, R. J. Lefkowitz, M. G. Carson, R. M. DeMafinis, A. J.
Krog, K. G. Holden, W. O. Matthews, and J. P. Hieble, Proc. Natl. Acad. Sci.
U.S.A. 79, 7223 (1982).
3. M. G. Caron, Y. Srinivasan, J. Pitha, K. Kociolek, and R. J. Lefkowitz, J. Biol.
Chem. 254, 2923 (1979); G. Vauguelin, P. Geynet, J. Hanoune, and A. D. Strosberg,
Eur. J. Biochem. 98, 543 (1979); C. J. Homcy, S. G. Rockson, J. Countaway, and
D. A. Egan, Biochemistry 22, 660 (1983).
4. E. Mulder, J. A. Foekens, M. J. Peters, and H. J. Van Der Molen, FEBS Lett. 97,
260 (1979); S. T. Hiremath, R. M. Loor, and T. Y. Wang, Biochem. Biophys. Res.
Commun. 97, 981 (1980).
5. W. E. Pricer and G. AshweU, J~ Biol. Chem. 251, 7539 (1976); A. L. Schwartz, A.
Marshak-Rothstein, D. Rup, and H. F. Lodish, Proc. Natl. Acad. Sci. U.S.A. 78,
3348 (1981).

(continued)
52 CHROMATOGRAPHY [1]

References to TABLE IV (continued)

6. J. Ramseyer, C. B. Kanstein, G. M. Walton, and G. N. Gill, Biochim. Biophys. Acta
446, 358 (1976).
7. J. W. Pike and M. R. Haussler, Proc. Natl. Acad. Sci. U.S.A. 76, 5485 (1979); S.
Sharpe, C. J. Hillyard, M. Szelke, and I. MacIntyre, FEBS Lett. 75, 265 (1977).
8. R. A. Hock, E. Nexo, and M. D. Hollenberg, Nature (London) 277, 403 (1979); S.
Cohen, G. Carpenter, and L. King, J. Biol. Chem. 255, 4834 (1980).
9. G. Redeuilh, R. Richard-Foy, C. Secco, V. Torelli, R. Bucourt, E. E. Baulieu, and
H. Richard-Foy, Eur. J. Bioehem. 106, 481 (1980); S. Thrower, C. Hall, L. Lira, and
A. N. Davison, Bioehem. J. 160, 271 (1976); P. Hubert, J. Mester, B. Dellacherie, J.
Neel, and E. E. Baulieu, Proc. Natl. Acad. Sci. U.S.A. 75, 3143 (1978).
10. T. Suzuki, R. Sadasivan, G. Wood, and W. Bayer, Mol. lmmunol. 17, 491 (1980).
11. H. ,I. Eisen and W. H. Alinsmann, Biochem. J. 171, 177 (1978); O. Wrange, J. C.
Duke, and J. A. Gustafsson, J. Biol. Chem. 254, 9284 (1979); V. K. Moudgil and
J. K. John, Biochem. J. 190, 809 (1980).
12. F. Pfeiffer, D. Graham, and H. Betz, J. Biol. Chem. 257, 9389 (1982).
13. M. J. Waters and H. G. Friesen, J. Biol. Chem. 254, 6815 (1979).
14. A. Kulczycki and C. W. Parker, J. Biol. Chem. 254, 3187 (1979); D. H. Conrad and
A. Froese, J. Immunol. 120, 429 (1978); J. Kanellopoulos, G. Rossi, and H. Metzger,
J. Biol. Chem. 254, 7691 (1979).
15. S. Jakobs, E. Hazum, Y. Schechter, and P. Cuatrecasas, Proc. Natl. Acad. Sci.
U.S.A. 76, 4918 (1979); S. T. Hiremath, P. F. Pilch, and P. Czech, J. Biol. Chem.
255, 1732 (1980).
16. S. Yamada, H. Itaya, O. Nakazawa, and M. Fukuda, Biochim. Biophys. Acta 496,
571 (1977); G. Marcoullis and R. Grasbeck, Biochim. Biophys. A cta 499, 309 (1977);
I. Kouvonen and R. Grasbeck, Biochem. Biophys. Res. Commun. 86, 358 (1979).
17. R. J. Schneider, A. Kulczycki, S. K. Law, and J. P. Atkinson, Nature (London) 290,
789 (1981).
18. J. M. Bidlack, L. G. Abood, P. O. Gyimah, and S. Archer, Proc. Natl. Acad. Sci.
U.S.A. 78, 636 (1981).
19. V. K. Moudgil and D. O. Tort, Proc. Natl. Acad. Sci. U.S.A. 73, 3443 (1976); J. B.
Miller and D. O. Tort, Biochemistry 17, 173 (1978).
20. D. S. Liscia and B. K. Vanderhaar, Proc. Natl. Acad. Sci. U.S.A. 79, 5930 (1982).
21. P. A. Seligman and R. H. Allen, J. Biol. Chem. 253, 1766 (1978).
22. J. A. Feraandez-Pol and D. J. KIos, Biochemistry 19, 3904 (1980); B. S. Stein and
H. H. Sussman, J. Biol. Chem. 258, 2668 (1983).
23. J. W. Apriletti, N. L. Eberhard, K. R. Latham, and J. D. Baxter, J. Biol. Chem. 256,
12094 (1981).
24. G. F. Fenzi, E. Macchia, L. Bartalena, F. Manzani, and A. Pincheria, FEBS Lett.
88, 292 (1978).
25. C. D. Strader, E. Lazarides, and M. A. Raftery, Biochem. Biophys. Res. Commun.
92, 365 (1980); C. L. Leonardi and R. W. Rubin, Anal. Biochem. 118, 58 (1981).
26. N. A. Musto, G. L. Gunsalus, and C. W. Bardin, Biochemistry 19, 2853 (1980).
27. J. E. Volanakis, W. L. Clements, and R. E. Schrohenloher, J. lmmunol. Methods
23, 285 (1978).
28. J. D. Scott and J. E. Fothergill, Biochem. J. 205, 575 (1982).
29. G. A. Jamieson and T. C. Vanaman, Biochem. Biophys. Res. Commun. 90, 1048
(1979); C. R. Caldwell and A. Hang, Anal. Biochem. 116, 325 (1981).
[1] AFFINITY CHROMATOGRAPHY 53

References to TABLE IV (continued)

30. K. E. Mickelson and U. Westphal, Biochemistry 18, 2685 (1979); IA. K. Mahajan,
R. B. Billiar, and A. B. Little, J. Steroid Biochem. 13, 67 (1980).
31. J. M. Trevillyan and M. L. Pall, J. Biol. Chem. 257, 3978 (1982).
32. K. Yoshida and T. Takegami, J. Biochem. (Tokyo) 81, 791 (1977); C. Chen, O. K.
Melitz, B. Petschow, and R. L. Eckert, Eur. J. Biochem. 108, 379 (1980).
33. M. M. McCabe, R. H. Hamelik, and E. E. Smith, Biochem. Biophys. Res. Commun.
78, 273 (1977).
34. S. O. Hoch and E. McVey, J. Biol. Chem. 252, 1881 (1977).
35. G. R. Jacobson and J. P. Rosenbusch, FEBS Lett. 79, 8 (1977).
36. L.-H. E. Hahn and K. M. Yamada, Proc. Natl. Acad. Sci, U.S.A. 76, 1160 (1979);
M. Vuento and A. Vaheri, Biochem., J. 183, 331 (1979).
37. D. Ricquier, C. Gervais, J. C. Kader, and P. Hemon, FEBS Lett. 101, 35 (1979).
38. J. Suttner, F. Hrkal, and F. Vodrazka, J. Chromatogr. 131, 453 (1977); K. W. Olsen,
Anal. Biochem. 109, 250 (1980).
39. O. Nygard and P. Westermann, Biochim. Biophys. Acta 697, 263 (1982).
40. A. Wichmann, Biochem. J. 181, 691 (1979).
41. H. R. Trayer, M. A. Winstanley, and I. P. Trayer, FEBS Lett. 83, 141 (1977).
42. S. K. Jain, M. G. Pluskal, and S. Sarkar, FEBS Lett. 97, 84 (1979).
43. D. Y. Lin, J. R. David, and H. G. Remold, Nature (London) 296, 78 (1982).
44. M. A. Winstanley, D. A. P. Small, and I. P. Trayer, Eur. J. Biochem. 98, 441 (1979).
45. I. M. Chaiken, Anal. Biochem. 97, 302 (1979).
46. T. L. Thomas and G. L. Patel, Proc. Natl. Acad. Sci. U.S.A. 73, 4369 (1976).
47. T. Tamura, H. Suzuki, Y. Nishimura, J. Mizoguchi, and Y. Hirota, Proc. Natl.
Acad. Sci. U.S.A. 77, 4499 (1980); H. Amanuma and J. L. Strominger, J. Biol.
Chem. 255, 11173 (1980).
48. L. I. Barsukov, C. W. Dam, L. D. Bergelson, G. I. Muzja, and K. W. A. Wirtz,
Biochim. Biophys. Acta 513, 198 (1978).
49. G. Fex, P. A. Albertsson, and B. Hansson, Fur. J. Biochem. 99, 353 (1979).
50. J. A. Froehlich, A. H. Merril, C. O. Clogett, and D. B. McCormick, Comp. Bio-
chem. Physiol. 66, 397 (1980).
51. N. Ulbrich, A. Lin, and I. G. Wool, J. Biol. Chem. 254, 8641 (1979); H. R. Burrell
and J. Horowitz, Eur. J. Biochem. 75, 533 (1977).
52. M. Yukioka and K. Omori, FEBS Lett. 75, 217 (1977).
53. N. Sonenberg, K. M. Rupprecht, S. M. Hecht, and A. J. Shatkin, Proc. Natl. Acad.
Sci. U.S.A. 66, 4345 (1979).
54. A. Rotman, Brain Res. 146, 141 (1978).
55. P. H. Petra and J. Lewis, Anal. Biochem. 105, 165 (1980).
56. B. Rousset and J. Wolff, J. Biol. Chem. 266, 11677 (1980).
57. W. J. Nelson, C. E. Vorgias, and P. Traub, Biochem. Biophys. Res. Commun. 106,
1141 (1982).
58. A. Nordheim, P. Tesser, F. Azorin, Y. H. Kwon, A. Moiler, and A. Rich, Proc.
Natl. Acad. Sci. U.S.A. 79, 7729 (1982).
54 CHROMATOGRAPHY [ 1]
quired for elution of such avid pairs. It is not unusual to find eluents that
consist of chaotrophs (5 M KSCN), low pH (2.2) or high pH (11.5)
buffers, urea (3.5-8 M) or guanidine (6 M). Despite such harsh condi-
tions, it is of interest that antigenicity is retained in most cases. Because
of the harsh conditions, methods have been developed in which antibod-
ies to the contaminating proteins are prepared and the protein is purified
by "reverse immunoadsorption"; i.e., contaminating proteins are ad-
sorbed while the required protein is unretarded and comes through in the
eluent. 37 Another approach utilizes immobilized protein A 38'39 o r antihap-
ten antibody 4°,41 to bind antibody-antigen or hapten antibody-antigen
complexes; it is a purified complex that is eluted, not the antigen.
The introduction of monoclonal antibodies, i.e., homogeneous im-
munoglobulins directed against a specific antigenic determinant, 42,43 de-
creased the possibility of contamination with antibodies directed against
proteins other than that desired (see this volume [24]). Since the number
of monoclonal antibodies produced against a specific protein are many,
selection of low-affinity antibodies for immobilization is desirable. The
lower affinity allows milder conditions for elution and less possibility of
irreversible denaturation. Monoclonal antibodies are being adopted rap-
idly for purification in all fields of biochemistry. Indeed, if the trend
continues and mild conditions for elution become the norm, a shift from
affinity chromatography to immunoaffinity chromotography will become
apparent. Some of the proteins purified by antibodies, more recently with
monoclonal antibodies, are summarized in Table III.

Lectins
Lectins are sugar-binding, cell-agglutinating proteins of nonimmune
origin and of wide distribution. They are being utilized extensively as
macromolecular carbohydrate-specific reagents. The ability to combine
with carbohydrates specifically and reversibly enables their purification
on immobilized sugars**; conversely, immobilization of the lectin itself

37 j. A. Weare, J. T. Gafford, N. S. Ur, and E. G. Erdos, Anal. Biochem. 123, 310 (1982).
38 D. M. Gorsten and J. J. Marchalans, J. Immunol. Methods 24, 305 (1978).
39 C. Schneider, R. A. Newman, D. R. Sutherland, U. Asser, and M. F. Greaves, J. Biol.
Chem. 257, 10766 (1982).
4o M. Wilchek and M. Gorecki, FEBS Lett. 31, 149 (1973).
41 j. Kanellopoulos, G. Rossi, and N. Metzger, J. Biol. Chem. 254, 7691 (1979).
42 G. Kohler and C. Milstein, Nature (London) 256, 495 (1975).
43 j. W. Goding, J. Immunol. Methods 39, 285 (1980).
44 H. Lis and N. Sharon, J. Chromatogr. 215, 361 (1981).
[1] AFFINITY CHROMATOGRAPHY 55

allows isolation of a variety of carbohydrate contaminating compounds. 45

The use of lectins for affinity chromatography of glycoconjugates is ac-
complished readily, since glycoconjugates do not interact very strongly
with lectins (K = 102 to 104) and, therefore, can be displaced readily from
affinity columns by specific sugars at neutral pH. Of the many lectins
available, two are used frequently in affinity purification of carbohy-
drates: concanavalin A and Ricinus communis agglutinin. Since lectins
are generally used at the early stages of purification of glycoproteins, and
require additional procedures for complete purification, details of specific
methods are not included. The general method for their use and applica-
tion has been reviewed. 45

Receptor-Binding Proteins
A receptor is a molecule that recognizes a specific chemical entity,
binds to it, and initiates a series of biochemical events resulting in a
characteristic physiological response. 46 The interaction between hor-
mone, neurotransmitter, or drug and the respective receptor is selective
and specific. Due to the selectivity, interaction between the ligand and its
receptor is one of high affinity, among the highest known among biologi-
cal interactions. Therefore, harsh elution conditions are being used to
elute receptors from columns containing a selective affinity ligand.
In overcoming some of these difficulties, the use of monoclonal anti-
bodies is being introduced. The rationale behind this approach depends on
the ability to prepare monoclonal antibodies against cell surface antigens.
The monoclonal antibody will inhibit binding of the ligand, e.g., a hor-
mone, to the receptor; will interact specifically with the receptor; and can
be used, therefore, to isolate the receptor. In Table IV examples are
presented of the use of specific ligands and monoclonal antibodies for the
isolation of receptors and other binding proteins.

45 R. Lotan and G. L. Nicolson, Biochim. Biophys. Acta 559, 329 (1979).

46 M. D. Hollenberg and P. Cuatrecasas, Methods Cancer Res. 12, 317 (1976).
56 CHROMATOGRAPHY [2]

[2] I m m o b i l i z a t i o n o f L i g a n d s w i t h O r g a n i c
Sulfonyl Chlorides
By KURT NILSSON and KLAUS MOSBACH

A number of methods for binding biomolecules to solid phases are in

use, many of which are summarized in Volumes 34 and 44 of this series,
as well as in this volume [1]. Few, however, can satisfy all of the following
requirements: high yields of coupled product at neutral pH, mildness (i.e.,
avoidance of denaturation or changes in the properties of the biomole-
cule), and stability of linkages between support and ligand. Organic sul-
fonyl chlorides, such as p-toluenesulfonyl chloride (tosyl chloride) and
2,2,2-trifluoroethanesulfonyl chloride (tresyl chloride), can be used to
convert hydroxyl groups into good leaving groups (sulfonates) that, on
reaction with nucleophiles, allow stable linkages to be formed between
the nucleophile and the initial hydroxyl group-carrying carbon. These
reagents appear to be suitable for the immobilization of enzymes and
affinity ligands to supports bearing a hydroxyl group, such as agarose,
cellulose, glycophase glass, and others. 1-3 The activation and coupling of
the ligand to the support are thought to involve the following steps.
Activation:
Support-CH2OH + R-SO2CI ---> Support-CH2OSO2-R

Coupling:
Support-CH2OSO2-R + H:N-ligand ---> Support-CH2-NH-ligand + HOSO2-R
Support-CH2OSO2-R + HS-ligand ~ Support-CH2-S-ligand + HOSO2-R
(R = CH2CF3 or CrH4CH3)

The influence of the R group on the reactivity of the sulfonate has been
studied with soluble organic molecules containing hydroxyl groups? It
was found that the relative reactivities for CH3C6H4---, CF3CHz---, and
CF3 sulfonate esters are 1 : 100:4000. Tosylates and tresylates are most
convenient reagents for enzyme and ligand immobilization.
Unlike tresylation, the introduction of the tosyl ester can be conve-
niently followed with UV spectroscopy (e26~= 480 M -~ cm-~). In addition,
tosyl chloride is inexpensive, a distinct advantage over tresyl chloride.
I K. Nilsson and K. M o s b a c h , Eur. J. Biochem. 112, 397 (1980).
2 K. Nilsson, O. N o r r l r w , and K. M o s b a c h , Acta Chem. Scand. Ser. B B35, 19 (1981).
3 K. Nilsson a n d K. M o s b a c h , Biochem. Biophys. Res. Commun. 102, 449 (1981).
4 R. K. Crossland, W. E. Wells, and V. J. Shiner, Jr., J. Am. Chem. Soc. 93, 4217 (1971).

Copyright © 1984by AcademicPress, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
[2] IMMOBILIZATION WITH SULFONYL CHLORIDES 57

The drawback to the method lies in its lower reactivity, being only slightly
more reactive than that of epoxy groups. Tosyl chloride constitutes an
alternative to tresyl chloride when the protein or affinity ligand to be
coupled is stable at pH 9-10.5, i.e., the optimal range for coupling of
primary amino groups to tosyl-Sepharose. 2
Tresyl chloride forms sulfonates much more readily than tosyl chlo-
ride. It is possible to obtain about 0.6 mmol of sulfonates per gram of dry
Sepharose 6B in 10 min using 2.5 mmol of tresyl chloride, whereas 12
mmol of tosyl chloride is needed for the same degree of substitution. The
reactivity of the resulting Sepharose tresyl groups is very high, allowing a
75-100% yield in protein coupling after 1 hr at pH 7.5 in the cold. This
efficiency has been found to be equal to or even greater than that of
CNBr-activated Sepharose. 3
The relationship between the degree of support activation (sulfona-
tion) obtained and the amount of sulfonyl chloride used during the activa-
tion process seems to be almost linear over a broad substitution range
(100-1000/xmol of ester per gram of dry Sepharose), thereby allowing
ready prediction and reproducibility of the degree of activation in contrast
to CNBr. Highly substituted gels do not diminish in coupling capacity
over a period of several weeks when stored under acidic conditions (1 mM
HC1) in the cold. Only a few percent of the sulfonates are lost owing to
hydrolysis when a tresylated gel is stored at pH 7.5 in the cold overnight,
in sharp contrast to CNBr-activated Sepharose, which must be used di-
rectly after activation. Freeze-dried tresylated preparations (such as sil-
ica) have been stored for a year in a desiccator without losing coupling
capacity.
Thiols and primary amino groups are the most reactive nucleophiles
with sulfonate esters on gels, thiols showing the highest reactivity. Amino
acid analyses of bound proteins indicate that imidazole and tyrosine hy-
droxyl groups can also displace the esters. This broad range of reactive
nucleophiles can be a great advantage when a very tight binding between
protein and support is desired, i.e., for stabilization or "freezing" of the
protein conformation. Furthermore, one is not restricted to amino group-
carrying spacers when constructing affinity ligands. With the use of thiol
spacers, uncharged S - - C bonds are obtained. The sulfonyl chloride
method mentioned above can also be utilized for activation of water-
soluble polymers, which are very useful in biochemical separations. Poly-
ethylene glycol (PEG) is frequently used in phase partitioning systems for
the separation of biomolecules and cells 5'6 ; tresyl chloride can be used for
5 p../~. Albertsson, "Partition of Cell Particles and Macromolecules." Wiley, New York,
1960.
6 F. Biickmann, G. Johansson, and M. Mort, Macromol. Chem. 182, 1379 (1981).
58 CHROMATOGRAPHY [2]

activation of PEG and for subsequent coupling of PEG to proteins, affin-

ity ligands, and cells. 7

Experimental Procedures

Preparation of Tresylated Supports (Activation)

Activation is performed in a water-free solvent in order to avoid hy-
drolysis of sulfonyl chloride. Pyridine is added to neutralize liberated
protons. Higher yields are obtained when using dried solvents rather than
commercially available solvents. It is especially important to use dried
solvents when low levels of activation are desired. The activation proce-
dure should be performed in a well-ventilated hood.
In a typical procedure, 10 g (wet) Sepharose 4B is washed succes-
sively with 100 ml of each of the following: 30 : 70 and 70 : 30 of acetone :
water (v/v), twice with acetone, and three times with dry acetone (dried
with a molecular sieve overnight, 25 g 4A per liter of acetone). The gel is
then transferred to a dried beaker containing 3 ml of dry acetone and 150
/.d of dry pyridine (dried with a molecular sieve). During magnetic stirring,
100/xl of tresyl chloride (Fluka AG, Buchs, Switzerland; also available
from Fluorochem Ltd., UK) is added dropwise. After 10 min at room
temperature the gel is washed twice with 100 ml of each of the following:
acetone, 30 : 70 and 50 : 50; 70: 30 of 5 mM HC1 : acetone (v/v); and 1 mM
HCI. The product is stored at 4° until used. The amount of introduced
tresyl groups is determined on freeze-dried gel by elemental analysis for
sulfur. The procedure described above will yield about 450/xmol of tresyl
groups per gram of dry gel.
Cellulose is treated in a similar manner. Glass and silica, however,
must first undergo a "coating reaction" before they can be activated. In
this procedure a hydrophilic layer consisting of glycerylpropyl groups is
introduced to the surface of these particles. 8 Glass treated in this manner
is commercially available from Pierce, U.S.A. (Glycophase glass) and
silica from Merck, West Germany (LiChrosorb Diol). In a typical activa-
tion, 2 g of the dry support is washed three times with 50 ml each of dried
acetone. The moist gel, about 5 g, is added to a dry beaker containing 2.5
ml of dry acetone and 130/.d of pyridine. With magnetic stirring, tresyl
chloride (90/zl) is added to the suspension. After 15 min at 0 ° the gel is
washed as described above for Sepharose. This procedure will introduce

7 K. Nilsson and K. Mosbach, to be published.

8 F. E. Regnier and R. Noel, J. Chromatogr. Sci. 14, 316 (1976).
[2] IMMOBILIZATION WITH SULFONYL CHLORIDES 59

about 130/xmol of tresyl groups per gram of dry silica (300-/~ pores). If the
activated preparations are not used directly for coupling, they should be
washed with water, 50 : 50 (v/v) water: acetone, and acetone and dried for
storage.

Preparation of Tosylated Supports

Sepharose is transferred to dry acetone as has been described above.
Tosyl chloride (0.6 g) is dissolved in 3 ml of dry acetone, and 8 g of moist
Sepharose CL-6B (in dry acetone) is added. After addition of 1 ml of
pyridine, the reaction is continued for 1 hr at room temperature with
continuous magnetic stirring. The activated gel is transferred to 1 mM
HC1 after the washing steps described for preparation of tresylated sup-
ports. The gel is now ready for use. This procedure will introduce about
0.6 mmol of tosyl groups per gram of dry Sepharose. The degree of
substitution by tosyl groups is determined by adding 40 mg of tosyl-
Sepharose to one cuvette and 40 mg of unreacted Sepharose CL-6B to a
reference cuvette, both in a solution (1 ml) of glycerol-water, 87: 13,
(v/v). The difference spectrum is recorded between 300 and 250 nm. The
extinction coefficient for ethyl tosylate at 261 nm (480 M -1 cm -1) is used
to calculate the tosyl group content. Alternatively, the degree of activa-
tion may be determined by elemental analysis for sulfur. Tosylated gels
can be stored for several weeks in 1 mM HC1 at 4° without losing coupling
capacity.

Preparation of Tresylated Polyethylene Glycol (PEG)

PEG-4000 (4 g) is dissolved in 10 ml of dichloromethane. 7 Pyridine
(250 /~l) and tresyl chloride (220 /~l) are added at 0°. The reaction is
allowed to continue at 20 ° for 1.5 hr before the solvent is evaporated to
dryness under reduced pressure. The product is dissolved in 60 ml of
ethanol containing 250/~l of concentrated HC1 and is precipitated over-
night at - 18°. The supernatant liquid is discarded after centrifugation, and
another 60 ml of ethanol containing 50/~l of concentrated HC1 is added.
The precipitation procedure is repeated six times, or until no pyridine is
detected in the UV (255 nm). After precipitation in ethanol without HC1,
and drying, about 2.5 g of white crystalline product is generally obtained.
Elemental analysis for sulfur indicates that about 1.6 mmol of tresyl
groups is formed per millimole of PEG; i.e., 80% of the hydroxyl groups
in PEG had been transformed into tresyl esters. The same principle can be
used for tresylation of PEG-6000; with PEG-6000, methanol is preferable
to ethanol.
60 CHROMATOGRAPHY [2]

Coupling of Affinity Ligands and Proteins

The procedures for coupling proteins or ligands to sulfonated supports
are the same as for CNBr-activated supports. Any buffer that does not
contain strong nucleophiles can be used in the coupling step, i.e., phos-
phate, HEPES, carbonate, and so forth. After coupling, the gel is prefer-
entially treated with Tris-HCl buffer to remove unreacted tresyl groups
that might interfere in the subsequent use of the product. The time neces-
sary for this treatment varies with the number of tresyl groups that had
been added. Treatment with 0.2 M Tris-HCl, pH 7.5-8.5, for 5 hr at 4 ° is
usually sufficient. For higher activated gels, longer incubation periods are
recommended. Tosyl-activated supports have to be treated at higher pH
and temperature for longer time periods (pH 9-10, 20-40 °, 15 hr) in order
to remove unreacted sulfonates. Mercaptoethanol is found to be more
efficient than Tris buffer for removing tosyl groups.
Similar procedures can be applied to glass and silica. If, however, the
activated glass or silica has been freeze dried, it is first washed with cold
coupling buffer, transferred to the coupling vessel, and deaerated under
reduced pressure before addition of the protein or ligand solution.

Examples
Coupling of N6-(6-Aminohexyl)-5'-AMP to Tresyl-Sepharose. The
AMP analog (Sigma, 16 mg) was dissolved in I ml of 0.2 M sodium
phosphate containing 0.5 M sodium chloride, pH 8.2 (coupling buffer) and
added to 1 g of wet tresyl-Sepharose that had been briefly washed with
cold coupling buffer) Coupling proceeded with gentle agitation by tum-
bling for 16 hr at 4°. The gel was treated with 0.2 M Tris, pH 8.5, for 5 hr
(20°) and washed with 0.2 M sodium acetate-0.5 M NaC1, pH 3.5, 0.5 M
NaC1, distilled water, and 0.2 M phosphate, pH 7.5. The amount of bound
AMP analog was determined from the UV spectrum obtained with 40 mg
of AMP-gel and 40 mg of unreacted Sepharose added to 1 ml of glycerol-
water (87: 13) in the sample and reference cuvettes, respectively (e268 =
17,600). The content of the bound analog could also be determined from
elemental analysis of nitrogen on a sample of gel that had not been treated
with Tris-HCl.
Coupling of Soybean Trypsin Inhibitor (STI) to Tresyl-Sepharose.
Soybean trypsin inhibitor (type I-S, Sigma, 10 mg) was dissolved in 0.2 M
sodium phosphate at pH 7.5. 3 Tresyl-Sepharose was briefly washed with
the cold coupling buffer and added to the protein solution. Coupling pro-
ceeded with gentle agitation overnight at 4°. The gel was treated with
[2] IMMOBILIZATION
WITH SULFONYL CHLORIDES 61

Tris-HC1 and washed as described above. The amount of bound STI

could be determined either by UV measurements, elemental analysis of
nitrogen, or amino acid analysis (see Table II).
Coupling of Hexokinase and Trypsin to Tresyl-Sepharose. Hexo-
kinase (yeast, type III, Sigma, 10 mg) was immobilized to l g of tresyl-
Sepharose, by the method used for trypsin inhibitor, in 0.2 M HEPES at
pH 7.0, containing 15 mM MgCl2.3 After coupling, the gel was washed as
above for trypsin inhibitor, omitting the Tris and acetate buffers. The
amount of bound enzyme was determined by amino acid analysis. Assay
of enzyme activity was done in a cuvette with a magnetic stirring device
(20 mg of gel in 3 ml of assay medium) as outlined in the Worthington
manual. 9 Trypsin (bovine pancreas, type III, Sigma, 10 mg), was coupled
and washed, using the same procedure as for trypsin inhibitor. 3
Coupling of N6-(6-Aminohexyl)-5'-AMP to Tosyl-Sepharose. Tosyl-
Sepharose (0.7 g) was washed briefly with 0.25 M NaHCO3 at pH 10.5,
and added to 0.3 ml of coupling buffer containing 20 mg of the AMP-
analog.l Coupling was allowed to proceed with gentle agitation overnight
at 40°. The gel was then treated with 0.8 M mercaptoethanol, pH 10, for 15
hr at 40 ° and washed as recommended for the tresyl-coupled analog. The
AMP content (5/zmol/g wet gel) was determined from the UV spectrum as
described for tresyl-Sepharose. To minimize interference from possible
remaining tosyl groups, the absorption at 290 nm was used (e290 = 3300
M-1 cm-1).
Coupling of Soybean Trypsin Inhibitor (STI) to Tosyl Sepharose. The
protein (115 mg), was dissolved in 3 ml of 0.25 M NaHCO3 at pH 9.5, and
10 g of the activated gel, briefly washed with coupling buffer, was added)
Coupling was allowed to proceed for 20 hr with gentle agitation at 20°.
Remaining interfering tosyl groups were removed by adding 20 ml of 0.8
M Tris-HCl at pH 9.5. After 30 hr at 20 °, the gel was washed as described
above for coupling of STI to tresyl-Sepharose. In this way about 3.5 mg of
the inhibitor was bound per gram of wet gel.
Coupling of Horse Liver Alcohol Dehydrogenase to Tosyl-Sepharose.
Alcohol dehydrogenase from horse liver (Sigma) was immobilized, by the
same method as was trypsin inhibitor, in 0.2 M sodium phosphate at pH
7.5.1 After coupling, the gel was washed three times with 10 gel volumes
of 0.1 M NaHCO3 at pH 8.5, 0.5 M NaCI, and 0.1 M sodium phosphate at
pH 7.5. The amount of enzyme bound was determined by amino acid
analysis.

9 "Worthington Enzyme Manual" (L. A. Decker, ed.), p. 66. Worthington, Freehold, New
Jersey, 1977.
62 CHROMATOGRAPHY [2]

TABLE I
ACTIVATION OF DIFFERENT SUPPORTS WITH TRESYL CHLORIDE a

Tresyl Tresyl
chloride used groups introduced b Yield
Support (mmol/g dry support) (mmol/g dry support) (%)

Sepharose 4B 0.70 0.11 16

Sepharose 4B 2.70 0.45 17
Sepharose 4B 4.50 0.80 18
Sepharose 4B 9.00 1.30 15
Sepharose CL-6B 2.00 0.45 23
Sepharose CL-6B 4.00 0.98 25
Sepharose CL-6B 6.00 1.10 18
Diol-silica 300~ 0.16 0.05 31
Diol-silica 300 0.41 0.13 31
Diol-silica 300 2.45 0.40 16
Separon HEMAd 0.44 0.14 32
Separon HEMA 2.50 0.78 31

a Data are reproduced, with permission, from Nilsson and Mosbach 3 and Nils-
son and tarsson."
b The activation time was 10 min for Sepharose and 15 min for silica and
Separon HEMA.
c Glycerlypropyl-silica with 300-A pores; 10-/~m particles; about 400/zmol of
diol groups per gram of dry weight, i1
d A special type of hydroxyethyl methacrylate for HPLC (Chemapol, Prague)
with 1000-/~ pores; 7-10/~m particles; 2.2 mmol of hydroxyethyl groups per
gram of polymer.

Coupling and Activity Yield

Table I s h o w s the d e g r e e s o f activation o b t a i n e d w h e n S e p h a r o s e ,
glycerylpropyl-silica, and h y d r o x y e t h y l m e t h a c r y l a t e w e r e activated
with v a r y i n g a m o u n t s o f tresyl chloride. 3,1°,11 As c a n be seen, the activa-
tion level can be adjusted o v e r a v e r y b r o a d range simply b y c h a n g i n g the
a m o u n t o f sulfonyl chloride.
B e c a u s e o f the high reactivity o f tresylated supports, a low activation
level [i.e., 1 5 0 - 3 0 0 / z m o l o f tresyl g r o u p s per g r a m o f d r y S e p h a r o s e or
5 0 - 1 0 0 / x m o l p e r g r a m o f d r y silica (100-300/~)] is effective in obtaining
a d e q u a t e yields o f c o u p l e d p r o d u c t s (Table II). As indicated in Table II,
90% coupling o f p r o t e i n A w a s a c h i e v e d with 150/zmol o f tresyl g r o u p s
p e r g r a m o f d r y S e p h a r o s e ~2 and a l m o s t 100% yield w a s o b t a i n e d with
~0K. Nilsson and K. Mosbach, unpublished results, 1981.
ii K. Nilsson and P.-O. Larsson, Anal. Biochem., in press (1983).
Jz M. Ramstorp, K. Nilsson, R. Mosbach, and K. Mosbach, in preparation.
[2] IMMOBILIZATION WITH SULFONYL CHLORIDES 63

TABLE II
COUPLING OF PROTEINS TO DIFFERENT SUPPORTS ACTIVATED WITH TRESYL CHLORIDEa

Tresyl
groups on
activated Ligand
support Time for bound
(mmol/g couplingb (mg/g dry Yield
Support dry support) Ligand (hr) pH support) (%)

Sepharose 4B 0.15 Protein A 15 8.5 70 90

Sepharose 4B 1.30 Concanavalin A 15 7.5 360 97
Sepharose 4B 1.30 STF 15 7.5 195 68
Sepharose 4B 1.30 STI 15 8.5 211 74
Sepharose 4B 1.30 STI 1.5 8.5 200 70
Sepharose CL-6B 1.30 Nucleotided 15 8.2 87 32
Diol-silica 100e 0.45 STI 20 8.0 250 83
Diol-silica 300 0.13 STI 20 8.0 115 33
Diol-silica 300 0.40 STI 20 8.0 280 80
Diol-silica 300 0.40 STI 1 8.0 220 63
Diol-silica 1000 0.06 STI 20 8.0 46 92
Diol-silica 100 0.45 Nucleotidea 20 7.5 13 80
Separon HEMA 0.14 Protein A 12 8.0 10 100

a Data are reproduced, with permission, from Nilsson and Mosbach, 3 Nilsson and Lars-
son, tl and Ramstorp e t al. ~2
b All couplings were performed at 4° except for protein A and for the AMP analog to diol-
silica, which were at room temperature. The estimated amount of diol in the silanized
silicas was 450, 400, and 50/zmol per gram of support for 100-, 300-, and 1000-/~ pore
silicas, respectively. The maximum theoretical loadings of STI c on diol-silicas 100, 300,
and 1000 were estimated to be 250, 270, and 50 mg, respectively.
c STI, soybean trypsin inhibitor.
d N6_(6_Aminohexyl).5,.AMP.
e Glyceryl-propyl-silica (10 gm).

hydroxyethyl methacrylate. The method is rapid, even at neutral pH. If,

however, very high protein loadings are desired, the activation level has
to be increased. Thus, almost all of the available surface of glyceryl-
propyl-silica 100 and 1000 could be covered with soybean trypsin inhibitor
by using the tresyl chloride activation method (250 and 50 mg of protein
per gram of dry gel, respectively). The amount of tresyl groups on these
highly activated silicas corresponded to the original amount of diol groups
(Table II). 11
Trypsin was bound with a 70% yield at pH 8.2 to tresyl-Sepharose,
whereas hexokinase was bound with a 53% yield at pH 7.0. 3 The specific
activities of bound trypsin and hexokinase relative to the free enzyme
were 33 and 26%, respectively. These values are in agreement with results
64 CHROMATOGRAPHY 12]

TABLE III
COUPLING OF HORSE LIVER ALCOHOL DEHYDROGENASE
TO DIOL-SILICA 1000~

Tresyl
chloride Relative
for Enzyme Enzyme activity
activation added coupled b bound to
(/zmol/g (mg/g dry (mg/g dry free enzyme c
dry silica) silica) silica) (%)

100 21 11 100
300 21 21 95

Data are reproduced, with permission, from Nilsson

and Larsson. 11
b Immobilization was in 0.25 M phosphate at pH 8.0
for 20 hr at room temperature together with 2 mM
NADH and 100 mM isobutyramide.
c Activity was measured with 9 mM NAD and 9 mM
ethanol in 0.25 M sodium phosphate at pH 7.8.

obtained when the enzymes were immobilized to CNBr-activated

Sepharose. 1° Apart from the enzymes listed above under Examples, a
number of other enzymes have been immobilized in high yields to tresyl-
Sepharose, such as T4 DNA ligase,13 horse liver alcohol dehydrogenase, 14
and chymotrypsin.l°
Direct quantitation of remaining intrinsic activity after immobilization
to Sepharose is difficult because of diffusional hindrances caused by the
relatively large Sepharose particles (40-120/.~m). A better assessment
would be possible by immobilization to smaller particles. As shown in
Table III, virtually complete retention of added horse liver alcohol dehy-
drogenase activity was obtained after covalent attachment to tresyl chlo-
ride-activated glycerylpropyl-silica (1000-A pores, 10/zm), despite quite a
high loading of enzyme. H Furthermore, the dissociation constants of
NADH for free and immobilized alcohol dehydrogenase were very similar
(the dissociation constants were determined from Scatchard plots of
bound radioactive nucleotide to the enzyme gelsll). This appears to be a
unique demonstration of complete retention of catalytic activity and bind-
ing characteristics of an enzyme after immobilization.

13 L. Biilow and K. Mosbach, Biochem. Biophys. Res. Cornmun. 107, 458 (1982).
t4 M.-O. Mfinsson, N. Siegbahn, and K. Mosbach, Proc. Natl. Acad. Sci. U.S.A. 811, 1487
(1983).
[2] IMMOBILIZATION WITH SULFONYL CHLORIDES 65

Tosyl-Sepharose was found to bind horse liver alcohol dehydrogenase

in high yield (60%) at pH 7.5, with the same retention of activity as tresyl-
Sepharose. l The high efficiency with tosyl-Sepharose was, in this case,
attributed to the many free thiol groups on the surface of this enzyme.
Thiol compounds have been shown to couple about twice as efficiently as
primary amino group-containing compounds to both tosyl- and tresyl-
Sepharose. 2,1° Very large amounts of ligand can be bound when coupling
takes place in DMF; about 40 /zmol of hexylamine per gram of wet
Sepharose was introduced when the ligand was allowed to react at 60 ° in
DMF with tosyl-Sepharose. 2 Details of activation and coupling yields by
the tosyl chloride method have been presented.l,2

Affinity Chromatography
To facilitate the removal of interfering sulfonates after coupling and to
minimize the risk of altering the properties of the immobilized protein, it
is recommended that the level of activation be kept quite low, between
150 and 400/zmol of sulfonate groups per gram of dry Sepharose and 50-
150/zmol of sulfonate groups per gram of dry glycerylpropyl-silica (100-
or 300-/~ pores). Some examples of affinity chromatographic behavior
after immobilization of ligands to sulfonyl chloride-activated supports are
given below.
Nr-(6-Aminohexyl) analogs of 5'-AMP and 2' ,5'-ADP, bound to tosyl-
or tresyl-Sepharose, 1,1° showed similar affinity characteristics as when
they were immobilized to CNBr-activated Sepharose.15 For example, 5'-
AMP and 2',5'-ADP analogs coupled to tosyl-Sepharose, specifically
bound NAD+-dependent lactate dehydrogenase, and NADH-dependent
6-phosphogluconate dehydrogenase, respectively, from a mixture of the
two enzymes, l Furthermore, the immobilized AMP analog was applied
for the single-step purification of lactate dehydrogenase from crude beef
heart extract) The AMP analog binding capacity for lactate dehydro-
genase, when bound to tresyl-Sepharose (3/xmol of ligand per gram of wet
support), was 6.5 mg per gram of wet support. 1°
Soybean trypsin inhibitor, immobilized to tosyl- or tresyl-Sepharose,
showed intact affinity properties, similar to CNBr-coupled gels) ,3 The
capacity of a preparation containing 11 mg of inhibitor bound to tresyl-
Sepharose (1 ml) was 9 mg of trypsin. 3
The lectin concanavalin A (Con A) efficiently bound (97% yield) to
tresyl-Sepharose (Table II) with retained affinity properties) On applica-
tion of commercially available horseradish peroxidase (grade II,
Boehringer) to a column containing Con A-Sepharose, impurities were
15 p. Brodelius, P.-O. Larsson, and K. Mosbach, Eur. J. Biochem. 47, 81 (1974).
66 CHROMATOGRAPHY [2]

unretarded and bound peroxidase was completely eluted with buffer con-
taining 0. I I M mannose. The A403:A280 ratio was 3.15, which corresponds
well with literature data for the pure homogeneous peroxidase. ~6
Protein A was also found to bind very efficiently to tresyl-Sepharose
with retained binding characteristics for immunoglobulin G (IgG).lZ Care
was taken not to immobilize protein A with too many linkages, since
difficulties have been encountered in eluting all of the bound IgG under
the latter circumstances. Protein A was, therefore, coupled to Sepharose
CL-4B with 150 ~mol of tresyl groups per gram of dry support at pH 8.5
overnight at room temperature. The yield for coupling was 90% (Table II).
The gel, 1 ml containing 2.5 mg of protein A, was packed in a column, and
12.5 mg of human serum IgG was applied at pH 7. Immunoglobulin lack-
ing specificity for protein A passed through unretarded (0.5 mg). The
bound IgG was completely eluted on lowering the pH to 2.2 with 0.2 M
glycine-HC1. The capacity of the column was very high; i.e., 36 mg of
IgG. Under these circumstances leakage was observed of 0.05% of pro-
tein A immobilized to tresyl-Sepharose when maintained at pH 7.0 and 37°
for 8 days. In general, affinity ligands bound by single-point attachment to
tresyl-Sepharose show slightly more leakage under these conditions. It
may well be that this slight leakage is due to instability of the support
chains per se. These questions are presently under investigation. 3

High-Performance Liquid Affinity Chromatography

N6-(6-Aminohexyl)-5'-AMP (I 6 mg per gram of dry silica) was coupled
to tresyl-activated glycerylpropyl-silica (LiChrosorb Diol, 5 /zm, 100-A
pores, 450/.~mol of tresyl groups per gram of dry support). As can be seen
from Table II, very high yields of bound ligand (80%) were obtained when
coupling proceeded overnight at room temperature and pH 7.5. 3 The
AMP-silica was used in an HPLC system allowing highly rapid separa-
tions to be achieved (see Fig. 1).
Concanavalin A also has been bound to tresyl chloride-activated silica
and used in an HPLC system for the successful separation of glucosides. ~7
The possibility of utilizing an enzyme, horse liver alcohol dehydro-
genase, immobilized to tresylated silica for affinity separations on an
analytical scale, together with its use in the screening of Kd for retained
material, was demonstrated.l~ Details of the utilization of these materials
for affinity HPLC are presented in this volume [10].
16 H. Theorell and A. C. Maehly, Acta Chem. Scand. 4, 422 (1950).
17 A. Borchert, P.-O. Larsson, and K. Mosbach, J. Chromatogr. 244, 49 (1982).
[2] IMMOBILIZATION WITH SULFONYL CHLORIDES 67

100"

• E
E
• •
A
< < <
Z Z Z

E E E
g
L-

0
n

. . . . . . •" "..-.t A
\
1
o ~ ;o is
Time {min)
FIG. l. Separation of bovine serum albumin (25/~g), beef heart lactate dehydrogenase (5
/~g), and horse liver alcohol dehydrogenase (25/~g) on Nr-(6-aminohexyl)-AMP-silica by the
high-performance liquid affinity chromatography technique. The sample (450/~l) was applied
at room temperature to a column (0.5 x 10 cm) containing 29/~mol of AMP analog per gram
of dry support. The flow rate was 1 ml/min, which gave a pressure drop over the column of 7
MPa. The column effluent (0.05 M sodium phosphate, pH 7.5) was monitored at 280 nm and
then mixed, on-line, with assay reagents for dehydrogenases and measured at 340 nm.
Reproduced, with permission, from Nilsson and Mosbach)

Stabilization of Enzyme Configuration

Horse liver alcohol dehydrogenase (Boehringer), 5 mg per gram of dry
gel, was immobilized in its ternary complex conformation with NADH
and isobutyramide to tresyl chloride-activated glycerylpropyl-silica with
100-/~ pores (LiChrosorb Diol, Merck) at pH 8.0, 20° overnight) 8 After
is K. Nilsson and K. Mosbach, in preparation.
68 CHROMATOGRAPHY [2]

rigorous washing, an enzyme preparation was obtained having lower dis-

sociation constants (50%) for NADH and NAD when compared with a
preparation immobilized together with ADP-ribose, which is known not
to induce a conformational change. Immobilization of the enzyme to silica
with a low amount of tresyl groups does not yield such an effect. The
binding of AMP, however, was the same for all preparations; this is
expected, since the part of the dehydrogenase responsible for AMP bind-
ing does not undergo gross conformational change.19 The activity of the
"frozen" preparation is also lower than that for the other preparations,
possibly owing to a decrease in the rate of dissociation of the coenzyme.
The stability of enzymes in organic solvents can be increased by their
immobilization to tresyl-Sepharose. The higher the activation of the sup-
port by treatment with increasing tresyl chloride concentration, the better
the stability. This was shown for the synthesis of N-acetyl-L-tyrosine
ethyl ester from N-acetyl-L-tyrosine and ethanol 2° and for peptide synthe-
sis. 18Thus, after 3 days in 50% DMF-0.2 M NaHCO3 at pH 9, the remain-
ing activity for synthesis of N-acetyl-L-phenylalanine glycinamide from
N-acetyl-L-phenylalanine methyl ester and glycinamide was 40% for
chymotrypsin immobilized to highly activated tresyl-Sepharose, com-
pared to 10% when immobilized to Sepharose with a low activation level.18
A stabilizing effect on the activity of T4 DNA ligase after immobiliza-
tion to tresyl-Sepharose was also reported.13 The stabilization was greater
with tresyl- than with CNBr-bound enzyme. The sulfonate method has
been used as a facile and effective means for immobilization of proteins
(IgG, Con A, lactate dehydrogenase) to glycerylpropyl-coated silicon
plates. El The interactions of the immobilized proteins with added biomole-
cules have been studied by ellipsometry. 21

Comments
Tresyl chloride can be used with advantage for coupling of pH-sensi-
tive ligands or proteins with high yields, even at neutral pH. In the pH
range 9 to 10.5, tosylated supports, similar to epoxy gels but easier to
prepare, are effective for coupling ligands that are stable in this range. We
consider the sulfonyl chloride method a facile one for immobilization,
with advantages over the widely used CNBr method. The degree of acti-
vation is easy to predict and to reproduce, and the activated gel does not
lose coupling capacity when stored for several weeks in aqueous media.
19 H. E. Eklund and C.-I. Brfind6n, J. Biol. Chem. 254, 3458 (1979).
2o T. Mori, K. Nilsson, P.-O. Larsson, and K. Mosbach, in preparation.
2~ C. F. Mandenius, S. Welin, B. Danielsson, I. Lundstrfm, and K. Mosbach, Anal. Bio-
chem., in press (1983).
[3] HYDROPHOBIC CHROMATOGRAPHY 69

The tresyl chloride method is ideally suited for the activation of sorbents
applicable to high-performance liquid affinity chromatography and for
activation of polyethylene gycol. In addition to amino groups, thiol groups
react efficiently with sulfonate esters, enabling coupling of such a group
carrying affinity ligands and proteins.

[3] H y d r o p h o b i c C h r o m a t o g r a p h y
By SHMUEL SHALTIEL1

Until the late 1960s, commonly used methods for the separation of
proteins from each other were based on differences in their size, their net
charge, their solubility, and their stability to high temperature or to ex-
treme pH values. An important step forward was made with the develop-
ment of affinity chromatography as a general method for protein purifica-
tion. 2-5 This method makes use of the biorecognition of enzymes,
antibodies, and receptors for their biospecific ligands, covalently linked to
an inert matrix backbone. In some instances, the purification was signifi-
cantly improved upon interposing a hydrocarbon chain (an "arm") be-
tween the ligand and the matrix. 2 It was assumed that these arms relieve
the steric restrictions imposed by the matrix backbone and allow an in-
creased flexibility and availability of the ligand. 6 It was also assumed that
such hydrocarbon chain arms do not damage the inert nature of the ma-
trix, a condition that must be fulfilled to preserve a biospecific adsorption.
With water-soluble proteins, the latter assumption seemed reasonable at
the time, as it had been found in the case of lysozyme that "most of the
markedly nonpolar and hydrophobic side c h a i n s . . , are shielded from
the surrounding liquid by more polar parts of the molecule" and that, as
predicted by Sir Eric Rideal and Irving Langmuir, "lysozyme is quite well
described as an oil drop with a polar coat. ''7

i Incumbent of the HeUa and Derrick Kleemkan Chair in Biochemistry at the Weizmann
Institute of Science, Rehovot, Israel. This chapter was written while serving as a Scholar-
in-Residence at the Fogarty International Center of the National Institutes of Health,
Bethesda, Maryland.
2 p. Cuatrecasas, M. Wilchek, and C. B. Anfinsen, Proc. Natl. Acad. Sci. U.S.A. 61, 636
(1968).
3 p. Cuatrecasas and C. B. Anfinsen, Annu. Reo. Biochem. 40, 259 (1971).
4 W. B. Jakoby and M. Wilchek, eds., this series, Vol. 34.
5 M. Wilchek, T. Miron, and J. Kohn, this volume [1].
6 p. Cuatrecasas, J. Biol. Chem. 245, 3059 (1970).
7 D. C. Phillips, Sci. Am. November, p. 78 (1966).

Copyright © 1984 by Academic Press) Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All fights of reproduction in any form reserved.
1SBN 0-12-182004-1
[4] IMMOBILIZEDDYES 97

[4] Affinity C h r o m a t o g r a p h y on I m m o b i l i z e d D y e s
By CHRISTOPHERR. LOWE and JAMES C. PEARSON

The development of group-specific or general-ligand affinity adsorb-

ents in which the immobilized ligand interacts specifically and reversibly
with a wide range of complementary proteins has proved to be a valuable
addition to the techniques available for enzyme purification. 1-3 In recent
years, some of the more widely used ligands for application in this type of
affinity chromatography have been a variety of reactive triazine-based
textile dyes immobilized to agarose and other matrices. 3-8 In particular,
immobilized Cibacron Blue F3G-A appears to be an especially effective
adsorbent for the purification of pyridine nucleotide-dependent oxidore-
ductases, phosphokinases, coenzyme A-dependent enzymes, hydrolases,
acetyl-, phosphoribosyl-, and aminotransferases, RNA and DNA nucle-
ases and polymerases, restriction endonucleases, synthetases, hydroxy-
lases, glycolytic enzymes, phosphodiesterases, decarboxylases, sulfohy-
drolases, a number of blood proteins including serum albumin, clotting
factors, serum lipoproteins, transferrin, and complement factors, and a
host of other seemingly unrelated proteins including interferon and phy-
tochrome. 3,6,9 Other triazine dyes such as Procion Red H-E3B, Procion
Red H-8BN, Procion Yellow MX-8G, Procion Scarlet MX-G, Procion
Green H-4G, and Procion Brown MX-5BR have also been found to be
suitable ligands for the selective purification of individual pyridine nucleo-
tide-dependent dehydrogenases, phosphokinases, plasminogen, carboxy-
peptidase G2, alkaline phosphatase, and L-aminoacyl-tRNA synthe-
tases.8,~0-18

i C. R. Lowe and P. D. G. Dean, FEBS Lett. 14, 313 (1971).

2 K. Mosbach, H. Guilford, R. Ohlsson, and M. Scott, Biochem. J. 127, 625 (1972).
3 C. R. Lowe, in " A n Introduction to Affinity Chromatography: Laboratory Techniques in
Biochemistry and Molecular Biology" (T. S. Work and E. Work, eds.). North-Holland
Publ., Amsterdam, 1979. See also Vol. 34 of this series.
4 R. L. Easterday and I. M. Easterday, Adv. Exp. Med. Biol. 42, 123 (1974).
5 S. T. Thompson, K. H. Cass, and E. Stellwagen, Proc. Natl. Acad. Sci. U.S.A. 72, 669
(1975).
6 p. D. G. Dean and D. H. Watson, J. Chromatogr. 165, 301 (1979).
7 C. R. Lowe, D. A. P. Small, and A. Atkinson, Int. J. Biochem. 13, 33 (1981).
s j. Baird, R. F. Sherwood, R. J. G. Cart, and A. Atkinson, FEBS Lett. 70, 61 (1976).
9 E. Gianazza and P. Arnaud, Biochem. J. 201, 129 (1982).
i0 y . D. Clonis and C. R. Lowe, Biochim. Biophys. Acta 659, 86 (1981).
i1 A. J. Turner and J. Hryszko, Biochim. Biophys. Acta 613, 256 (1980).
12 D. H. Watson, M. J. Harvey, and P. D. G. Dean, Biochem. J. 173, 591 (1978).

Copyright © 1984 by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
98 CHROMATOGRAPHY [4]
The use of synthetic dyes such as Cibacron Blue F3G-A and dyes of
the Procion series as ligands for affinity chromatography offers several
advantages over the more conventional immobilized coenzyme and other
biological group-specific media. For example, the protein-binding capaci-
ties of immobilized dye adsorbents exceed those of natural biospecific
media by factors of 10 to 100 and the low capital cost, general availability,
and ease of coupling to matrix materials represent major advantages of
dyes for large-scale affinity chromatography.7,~9 Furthermore, synthetic
dyes are largely resistant to chemical and enzymic degradation and the
triazine bond, an essential feature of many reactive dyes, is less prone to
solvolysis than the isouronium linkage introduced during CNBr activation
of polysaccharides.3,2° In addition, the characteristic spectral properties
of the dyes permits ready monitoring of ligand concentrations and identifi-
cation of column materials. TM Finally, the general applicability and re-
usability of the adsorbents together with the ease of elution of bound
proteins, often with good recoveries, ensures excellent prospects for the
further exploitation of affinity chromatography on immobilized dyes. This
chapter describes the chemical nature of reactive dyes, procedures for
their immobilization to matrix materials and the subsequent use of the
dyed adsorbents in affinity chromatography and related techniques. In
addition, advice is offered on the selection of the appropriate dye and
matrix, and recommendations are made as to operational parameters for
the adsorption and selective desorption of proteins from immobilized dye-
stuffs in order to achieve optimal purification or resolution of the desired
proteins.

Structure and Chemistry of Reactive Dyes

The reactive triazine dyes exploited in the purification of a host of

proteins by affinity chromatography were originally developed at Impe-
rial Chemical Industries (ICI) in the 1950s and were destined primarily for
the textile industry. The Procion range of reactive dyes is composed of

13 C. R. Lowe, M. Hans, N. Spibey, and W. T. Drabble, Anal. Biochem. 104, 23 (1980).

14 V. Bouriotis and P. D. G. Dean, J. Chromatogr. 206, 521 (1981).
15 y . D. Clonis, M. J. Goldfinch, and C. R. Lowe, Biochem. J. 197, 203 (1981).
16 C. J. Bruton and A. Atkinson, Nucleic Acids Res. 7, 1579 (1979).
i7 p. Hughes, C. R. Lowe, and R. F. Sherwood, Biochim. Biophys. Acta 700, 90 (1982).
18 E. E. Farmer and J. S. Easterby, Anal. Biochem. 123, 373 (1982).
19 M. D. Scawen, J. Darbyshire, M. J. Harvey, and A. Atkinson, Biochem. J. 203, 699
(1982).
20 C. R. Lowe, Int. J. Biochem. 8, 177 (1977). See also Wilchek et al., this volume [1].
21 y . Hey and P. D. G. Dean, Chem. Ind. (London) 20, 726 (1981).
[4] IMMOBILIZED DYES 99

various polysulfonated chromophores in a comprehensive range of shades

linked either to reactive dichlorotriazinyl functional groups by an ami-
noether bridge (Procion MX dyes) or to the less reactive monochloro-
triazinyl group (Procion H, HE, or P dyes). Dichlorotriazinyl dyes are
readily prepared from chromophores containing a primary or secondary
amino group by reaction with an equimolar proportion of cyanuric chlo-
ride (1,3,5-sym-trichlorotriazine) for 1-2 hr at pH 6-7 and 0-5 °. Mono-
chlorotriazinyl dyes are conveniently prepared by replacing the second
chlorine of the triazinyl ring of dichlorotriazinyl dyes by a nonlabile sub-
stituent such as an alkoxy, aryloxy, amino, arylamino, or o,m,p-sulfoani-
lino group at pH 6-7 and 30-40 °. An alternative, and often preferred,
procedure is to introduce the nonlabile substituent into the triazine ring
first, whence the resulting substituted cyanuric chloride is condensed with
an appropriate chromophoric base. 22
Chlorotriazine dyes encompass a complete range of shades in both the
dichlorotriazine (Procion MX) and monochlorotriazine (Procion H, HE,
or P) series. The reactive dyes available commercially from IC123 are
derived from three principal types of chromophore: azo, anthraquinone,
and phthalocyanine. The anthraquinone chromophore is responsible for
bright blue shades with absorption maxima in the range of 600-630 nm,
the phthalocyanines producing bright turquoise shades. Green dyes are
derived from mixed chromophores of the anthraquinone-stilbene, anthra-
quinone-azo, or phthalocyanine-azo classes, whereas the majority of the
yellow, orange, and red with absorption maxima in the range 380-450 nm
are derived from the azo class. Rubine, violet, blue, navy, deep brown, or
black dyes are invariably composed of 1 : 1 or 1 : 2 metal complexes of
o,o'-dihydroxyazo or o-carboxy-o'-hydroxyazo chromophores. Figure
1 illustrates the structures of representative dyes. Structures of the major-
ity of the commercial dyes are undisclosed, and therefore this figure in-
cludes only examples of dyes to be found in the public domain and patent
literature.
22 The chemistry and detailed preparation procedures for a number of reactive dyes are
described in "The Chemistry of Synthetic Dyes" (K. Venkataraman, ed.), Vol. 6. Aca-
demic Press, New York, 1972.
23 The chlorotriazine dyes mentioned in this chapter are available from Imperial Chemical
Industries p.l.c. (Organics Division, P.O. Box 42, Hexagon House, Blackley, Manches-
ter, M9 3DA, U.K.) under the trade name Procion. Other reactive dyes based on 2,4,5-
trichloropyrimidinyl groups are available from Ciba-Geigy and Sandoz, respectively, un-
der the commercial names Reactone and Drimarene; Bayer markets dyes based on the
2,3-dichloroquinoxaline reactive group (Levafix E dyes) and on alkylsulfonyl-substituted
pyrimidines (Levafix P range); Hoechst produces a range of Remazol dyes containing
vinyl sulfone, sulfatoethyl sulfone, or/3-chloroethylsulfone groups. The relative reactivi-
ties of different commercial ranges of reactive dyestuffs are given in J. A. Fowler and W.
J. Marshall, J. Soc. Dyers Colourists 80, 858 (1964).
100 CHROMATOGRAPHY [4]

~
0 NH2
S03H
.. ~ N----'~CI
0 N H ~ / x~,--NH.-~-- N

503H NH-'~.~SO3H {o/m/p)

0 MH2

N~', ~*
NH~NH-~ N
I~.~ o~" N~,<,

~.SO:H OH III-I-.,,i//~-'-~kNCi

,-700,',-=<,,._0
H03S S03H

0 Cu 0

S03H .°~" .L ,.
S03n .=x.
--Ci

3 \CI

SO-~H ~ N--,,/ct

. k.J..J ,>=-' '-~,.,

- T NHCOCH3
S03H
FiG. 1. Structures of several representative tfiazine dyes: (a) Cibacron Blue F3G-A; o-
isomer: C.]. 61211, Procion Blue H-B; m, p-isomers: C.I. 61211. (b) Procion Blue MX-R:
C.]. 61205. (c) Procion Red H-3B: C.I. 18159. (d) Procion Rubine MX-B: C.I. 17965. (e)
Procion Scarlet MX-G: C.I. 17908. (f) Procion Yellow H-A: C.I. 13245.
[4] IMMOBILIZED DYES 101

Reactive dye nomenclature is generally based on the commercial

name and Colour Index (C.I.) generic name and number, 24 although some
confusion may arise when a dye is synthesized by both of the major
manufacturers, ICI and Ciba. F o r example, the yellow azoic dye, reactive
yellow 3 (C.I. 13245), may also be referred to as Procion Yellow H-A
(ICI) or Cibacron Yellow RA (Ciba). The use of different isomeric sub-
stituents in dye manufacture can create further confusion: although the
anthraquinone dye Cibacron Blue F3G-A (Ciba) contains an o-aminoben-
zenesulfonate substituted on the triazine ring (Fig. 1) and the equivalent
product from ICI, Procion Blue H-B, a meta and para mixture o f isomers,
both products share the same Colour Index generic name (reactive blue 2)
and n u m b e r (C.I. 61211). Some caution should be exercised, therefore, in
the interpretation o f data obtained with dyes from different commercial
sources.

Purification and Properties of Reactive Dyes

Commercially supplied dye powders usually contain a number of addi-
tives to ensure adequate operational storage and handling properties.
Thus, dye powders incorporate a buffer such as phosphate, since the
stability of reactive chlorotriazine dyes is maximal at pH 6.4. Diluents,
such as sodium chloride, and traces of surface-active agents are also
generally added to standardize the shade in different dye batches and to
produce a nondusty product, respectively. The crude dye powder may be
washed on a sintered funnel with diethyl ether prior to dissolution in
distilled water and precipitation o f the dye with ethanol, acetone, or, as
the K ÷ salt, with potassium acetate to r e m o v e some of these contami-
nants. 7,25 Virtually all of the commercially available dyes are chro-
mophorically heterogeneous when examined by thin-layer chromatogra-
phy 17'26'27 o r high-performance liquid chromatography. 7,28 The principal

24"Colour Index," 3rd Ed. The Society of Dyes and Coloufists, Bradford, U.K., 1971.
z5A. Atkinson, J. E. McArdell, M. D. Scawen, R. F. Sherwood, D. A. P. Small, C. R.
Lowe, and C. J. Bruton, in "Affinity Chromatographyand Related Techniques" (T. C. J.
Gribnau, J. Visser, and R. J. F. Nivard, eds.), p. 399. Elsevier, Amsterdam, 1982.
26p. Hughes, R. F. Sherwood, and C. R. Lowe, Biochem. J. 205, 453 (1982).
27j. E. C. McArdell, A. Atkinson, and C. J. Bruton, Eur. J. Biochem. 125, 361 (1982).
A typical high-performance liquid chromatography system would comprise chromatogra-
phy on a C~8Bondapak (5-/zm)column at 1000psi (70 MPa) and 1 ml rain-~ in an isocratic
system, such as water: acetonitrile : acetic acid : triethylamine (65 : 35 : 1.0 : 0.5, by vol-
ume) (unpublished observations).
102 CHROMATOGRAPHY [4]

chromophores may also be purified by conventional column chromatogra-

phy on silica gel, 29 cellulose, 3° alumina, z7 or Sephadex LH-20. 7'z8,31
A typical procedure currently used in our laboratory for the routine
purification of the triazine dyes Procion Red H-8BN, Procion Yellow
H-A, Procion Green H-4G, and Cibacron Blue F3G-A by preparative
thin-layer chromatography ~7,z8is performed on kieselge160 plates (20 x 20
cm; 2-mm thickness; E. Merck, Darmstadt, FRG). The plates are dried at
80 °, and crude dye (20 mg) in deionized water (0.5 ml) is loaded at 20-22 °
as a narrow band across the base of the plate. Ascending chromatography
in 2-butanol-l-propanol-ethyl acetate-water (20 : 40 : 10 : 30 by volume)
for Cibacron Blue F3G-A and Procion Yellow H-A or 2-butanol-l-pro-
panol-ethyl acetate-water (20 : 35 : I0 : 35 by volume) for Procion Green
H-4G and Procion Red H-8BN is performed overnight. The principal dye
band is carefully scraped from the plate and suspended in deionized wa-
ter; silica gel is removed by filtration through Whatman No. 1 filter paper,
and the filtrate is lyophilized. Purified dyes should be stored dry after
lyophilization in sealed vials protected from light.
Applications in affinity chromatography do not normally require
highly purified dyes~ ns most of the minor chromophoric impurities do not
covalently attach to support matrices. For analytical applications, how-
ever, or where precise quantitative information on dye-protein interac-
tions is required, the exclusive use of purified preparations is recom-
mended. One unfortunate consequence of the impurity of commercial dye
powders and batch-to-batch variation is that either definitive molar ab-
sorption coefficients are not available or the literature contains widely
different values for the same supposed dye. Table I lists the spectral
properties of some of the more commonly exploited triazine dyes.

Immobilization of Reactive Dyes to Matrices

Chlorotriazine dyes have been covalently attached to a variety of
support matrices including agarose, 3,8-18 Sephadex, 4,32,33 beaded cellu-
lose, 34 metal oxides, 35 polyacrylamide, 36 Sephacryl S-200, 6 Spheron, 6
29B. H. Weber, K. Willeford, J. G. Moe, and W. Piszkiewicz, Biochem. Biophys. Res.
Cornmun. 86, 252 (1979).
3oR. A. Edwards and R. W. Woody, Biochem. Biophys. Res. Cornmun. 79, 470 (1977).
31L. A. Haft and R. L. Easterday, in "Theory and Practice in AffinityTechniques" (P. V.
Sundaram and F. Eckstein, eds.), p. 23. Academic Press, New York, 1978.
32G. Sand, H. Brocas, and C. Erneux, Acta Endocrinol. 82, $204 (1976).
33F. Qadri and P. D. G. Dean, Biochem. J. 191, 53 (1980).
D. Mislovicova, P. Gemeiner, L. Kuniak, and J. Zemek, J. Chromatogr. 194, 95 (1980).
35C. R. Lowe, FEBS Lett. 1116,405 (1979).
36H.-J. B6hme, G. Kopperschl~iger,J. Schulz, and E. Hofmann, J. Chromatogr. 69, 209
(1972).
[4] IMMOBILIZED DYES 103

TABLE I
PROPERTIES OF SOME TRIAZINE DYES

Molar absorption
Mr hmax coefficient
Reactive dye (Na salt) (nm) (M -~ cm 1)

Procion Yellow H-A 578.5 385 8,900

Procion Red H-E3B -- 522 30,000 a
Procion Brown MX-5BR 588.2 530 15,000
Procion Red H-8BN 801.2 546 21,300
Cibacron Blue F3G-A 773.5 610 13,600
Procion Blue MX-R 635.9 600 4,100
Procion Green H-E4BD -- 630 20,800
Procion Green H-4G 1760.1 675 57,400

a A molar absorption coefficient of 11,300 M -1 cm -1 is quoted in the

product literature of the Amicon Corporation, Lexington, MA.

glass, 37 microparticulate silica, 38-4° and agarose-polyacrylamide (Ultro-

gel) copolymers. 6 Of these, agarose would appear to be the best general-
purpose matrix at present, although the cross-linked dextran Sephadex
G-200 has been reported to give better capacities for proteins of relatively
low molecular weight, s,33 Triazine dyes may be covalently attached to
support matrices either directly by reaction with hydroxyl groups on the
matrix backbone or indirectly with spacer molecules or dextran-dye con-
jugates. However, since comparisons of the relative properties of directly
and indirectly coupled triazine dye affinity adsorbents have given contra-
dictory results, the ease and simplicity with which triazine dyes may be
coupled directly to polysaccharides makes this approach the universally
accepted procedure. The following protocol has been found to work con-
sistently well in our hands. 13,15,17
Sepharose 4B (5 g) is thoroughly washed on a sintered-glass funnel
with 1-5 liters of distilled water under reduced pressure. Surplus water is
removed by gentle suction, and the moist gel cake is transferred to a
stoppered glass vial containing dye (50 mg) in distilled water (5 ml). The
gel suspension is tumbled for 30 min, NaC1 solution (22% w/v, 1 ml) is
added to give a final concentration of 2% (w/v) NaC1, 41 and the gel suspen-

37 p. A. Anderson and L. Jervis, Biochem. Soc. Trans. 6, 263 (1978).

38 C. R. Lowe, M. Glad, P. O. Larsson, S. Ohlson, D. A. P. Small, A. Atkinson, and K.
Mosbach, J. Chromatogr. 215, 303 (1981). See also this volume [2].
39 D. A. P. Small, C. R. Lowe, and A. Atkinson, J. Chromatogr. 216, 175 (1981).
4o S. Rajgopal and M. Vijayalakshmi, J. Chromatogr. 243, 164 (1982).
41 This procedure "salts out" the dye onto the matrix backbone and thereby facilitates
preferential reaction between the chlorotriazine ring of the dye and matrix-borne nu-
 104 CHROMATOGRAPHY [4]
sion is gently tumbled for an additional 30 min prior to adding solid
Na2CO3 to a final concentration of 1% (w/v). The gel is tumbled for a
period of time that depends on the type of dye and the incubation temper-
ature. Typically, dichlorotriazinyl dyes require 2-4 hr at room tempera-
ture at 20 °, whereas satisfactory coupling of monochlorotriazinyl dyes
may be achieved by incubation for 3-5 days at 20°, for 2 days at 37°, or for
2 hr at 60°.
Other alkalies, such as NaOH, may be substituted for Na2CO3 in the
above protocol. A final concentration of 0.01 M NaOH in the coupling
step is optimal for dichlorotriazinyl dyes and 0.05-0.2 M for monochloro-
triazinyl d y e s Y ,42 Under the above conditions the immobilization of di-
chlorotriazinyl dyes, such as Procion Red MX-2B, is rapid and virtually
complete in 2 hr at 25-35 °. The reaction with agarose displays a distinct
optimum for alkali concentration at 10 mM, and at higher concentrations
rapid hydrolysis of the second chlorine on the triazine ring ensues and
gives a sharp decline in the yield. Residual unreacted chlorines in the
coupled dye may be converted to hydroxyl groups by incubating the
matrix at pH 8.5 and room temperature for 2-3 days, or to amino groups
by reaction with 2 M NH4C1 at pH 8.5 for 4 hr at 200. 42 Immobilization of
monochlorotriazinyl dyes, such as Procion Red H-3B, is slow at room
temperature and plateaus after about 3 days. At 60° the reaction rate is 3-
4 times that at 20°. The amount of dye immobilized to agarose is relatively
insensitive to alkali concentrations in the range of 0.05-0.2 M although
there is a slight fall at higher concentrations (0.4 M). The amount of dye
incorporated into the matrix is a linear function of dye concentration up to
2 mg of dye per milliliter; above this value the relationship becomes
progressively nonlinear, with concentrations above 4 mg of dye per milli-
liter resulting in barely detectable increases in incorporation.
The reactivity of the relatively inert monochlorotriazinyl dyes may be
enhanced by a factor of four- to eightfold in the presence of certain ter-
tiary amines, hydrazines, tetrazenes, or hydrazones when used in the
dyeing solution in quantities of 0.1-10% (w/w) of the dye employed. 43
Pyridine, trimethylamine, N,N'-dimethylhydrazine, tetramethyltetra-
zene, and 1,4-diazabicyclo[2.2.2]octane (DABCO) are especially effective
auxiliary compounds in this respect.

cleophiles rather than solvolysis. Although significant coupling can be achieved in the
absence of NaC1, optimal dye coupling requires a salt concentration greater than 0.25 M.
42 A. Atkinson, P. M. Hammond, R. D. Hartwell, P. Hughes, M. D. Scawen, R. F. Sher-
wood, D. A. P. Small, C. J. Bruton, M. J. Harvey, and C. R. Lowe, Biochem. Soc. Trans.
9, 290 (1981).
43 D. Hildebrand, in "The Chemistry of Synthetic Dyes" (K. Venkataraman, ed.), Vol. 6,
p. 361. Academic Press, New York, 1972.
[4] IMMOBILIZED DYES 105

After covalent attachment of the reactive dye, the gel is copiously

washed on a sintered-glass funnel with distilled water, 1 M KCI, and
distilled water again until the washings are free of color. Gels are stored as
a moist cake or as a suspension in water at or below 4 ° in the presence of a
small amount of microbial growth inhibitor, such as 0.02% NAN3.

Determination of Immobilized Dye Concentration

The immobilized dye concentration is best determined by acid hydro-
lysis of the agarose gel. 13,15,~7Gel sucked moist on a sintered funnel (30
rag) is incubated at 37° with 0.6 ml of 5 M HCI for 5 rain, whence 2.4 ml of
2.5 M sodium phosphate at pH 7.5 is added, and the absorbance of the
resulting solution is determined at the hma~ of the dye. 44 The dye concen-
tration on the gel is calculated from the known molar absorbance coeffi-
cients (Table I) and should typically lie in the range of 1-10/zmol of dye
per milliliter (or gram wet weight) of gel. Where the structures of the dyes
are undisclosed, it is common practice to quote immobilized dye concen-
trations as milligrams of dye coupled per milliliter (or gram wet weight) of
gel.

Considerations Relating to the Selection of Dyes as Ligands for

Affinity Chromatography
There are few guidelines on which to base a rational choice of a dye for
a putative protein purification. A number of reactive dyes, including Ciba-
cron Blue F3G-A, Procion Red H-E3B, Procion Blue MX-4GD, Procion
Green H-E4BD, and Procion Red H-8BN, have been shown to possess
the general ability to bind to a wide range of NAD- and NADP-dependent
d e h y d r o g e n a s e s 3,6,7,13,17,18,25,42 and aminoacyl-tRNA synthetases. 45 Apart
from these examples, however, the task of selecting a suitable immobi-
lized dye remains very much an empirical process involving the testing of
a number of dyes for the ability to bind the protein in question. Often it is
adequate simply to test one or two of the more commonly used dyes,
since these are found to give satisfactory results in a surprisingly large
number of cases.
The suitability of a range of dyes may be conveniently tested by estab-
lishing a series of small columns, each containing 500 mg moist weight, of

44 These conditions refer to conventional agarose adsorbents. Cross-linked matrices

(Sepharose 4B-CL, 6B-CL) require higher temperatures or longer hydrolysis times to
effect dissolution of the gel. In all cases the final solution is buffered at pH 7.5 in order to
avoid problems of color change in some dyes at extremes of pH.
45 C. J. Bruton and A. Atkinson, Nucleic Acids Res. 7, 1579 (1979).
106 CHROMATOGRAPHY [4l
agarose-bound dye equilibrated with an appropriate buffer, in Pasteur
pipettes plugged with glass WOO1. 46 m small sample containing the protein
of interest is applied to each column, and the flow is interrupted for 10 min
to allow equilibration of the sample with the dye matrix. Nonadsorbed
protein is purged with 5-10 ml of equilibration buffer. Elution of specifi-
cally bound protein is then effected with 1 M KC1 in the same buffer and
collected in a single fraction. Both the void and the eluate fractions are
assayed for protein and activity-enabling recoveries, purification factors,
and capacities to be speedily assessed. 7 The results from such a screening
program generally suggest several dyes as being worthy of further study.
For example, the dye screen may indicate that a tandem system involving
both "negative" and "positive" chromatography steps might be worth
considering. In this case, the sample is applied first to a "negative"
column that does not bind the protein of interest but adsorbs other pro-
teins. Subsequently, the void volume eluate is applied directly to a second
"positive" column. The latter column binds the desired protein and on
elution yields purified protein. The preliminary screening operation may
also suggest when sequential immobilized dye adsorbents may best be
employed 1° to enhance purification.
A number of examples of such dye-screening procedures have been
published 1°,13,45although in many cases their value is diminished by insuffi-
cient characterization of the dye matrix in terms of the concentration of
immobilized d y e . 47,48 The concentration of dye immobilized on agarose
matrices is quite variable 7,1°,13 and must be taken into account when com-
paring different immobilized dye adsorbents. An alternative and preferred
procedure for screening dyes involves the characterization of dye-protein
interactions in terms of kinetic inhibition constants or dissociation con-
stants for a range of dyes in free solution. 49,5°As a rule of thumb, there is a
general trend of protein-dye dissociation constants (Kd) across the color
spectrum. Thus, blue and green dyes tend to bind more tightly to pig heart
lactate dehydrogenase (Kd < 8/zM) than yellow dyes (Kd > 36/xM), and
orange and red dyes display intermediate affinities (10 ~M < Kd < 33
/~M). 5° These observations account for the popularity of dyes such as
Cibacron Blue F3G-A, Procion Blue MX-4GD, Procion Green H-E4BD,
46 Disposable polypropylene columns (Econo-columns) comprising a 0.8 x 4 cm column
holding up to 2 ml of chromatographic medium and including an integral 10-ml reservoir,
porous polyethylene bed support disk, and Luer column tip are available from Bio-Rad
Laboratories, Richmond, CA. These products offer an effective replacement for Pasteur
pipettes and other such improvised columns.
47 V. Bouriotis and P. D. G. Dean, J. Chromatogr. 206, 521 (1981).
48 K. G. McFarthing, S. Angal, and P. D. G. Dean, Anal. Biochem. 122, 186 (1982).
49 A. Ashton and G. Polya, Biochern. J. 175, 501 (1978).
50 y . D. Clonis and C. R. Lowe, Biochem. J. 191, 247 (1980).
[4] IMMOBILIZEDDYES 107

TABLE II
REPORTED OPERATIONAL PARAMETERS FOR ADSORPTION OF
PROTEINS TO IMMOBILIZED DYES

Condition Data

Equilibration buffers
Molarity 5-100 mM
pH 5.5-9.0
Composition Tris, phosphate, acetate, triethanolamine,tricine,
HEPES, MOPS, bicarbonate
Temperature 4-50°
Additives EDTA, KC1,NaCI, NH4CI,(NH,)2SO4, 2-mercap-
toethanol, dithiothreitol,cysteine,thioglycerol,
phenylmethylsulfonylfluoride,urea, sucrose,
glycerol, ethyleneglycol,detergents, chloram-
phenicol, amphotericinB
Metal ions Na, K, Mg, Ca, Sr, Ba, Zn, Cu, Co, Ni, Mn, Fe, AI,
Fe, Cr
Sample applied per bed volume 0.03-950 mg/ml

Procion Red H-E3B, and Procion Red H-8BN for dye-ligand chromatog-
raphy.

Operational Parameters in Immobilized Dye Chromatography

The successful operation of a chromatographic separation on immobi-
lized dyes requires that careful consideration be given to a number of
variables, such as sample size, flow rate, buffer composition, pH, ionic
strength, and temperature. In this respect, chromatography on dye ad-
sorbents closely parallels the approach used in conventional biospecific
affinity chromatography) However, the mixed electrostatic-hydrophobic
characteristics of most textile dyes suggest that manipulation of the pH,
ionic strength, and temperature of the column irrigants is a worthwhile
exercise in improving adsorption and subsequent elution of complemen-
tary proteins. Table II lists a selection of conditions that have been used
to adsorb proteins to immobilized-dye adsorbents. Generally speaking,
irrigant buffers would have molarities in the range of 5-100 mM, pH
values of 5.5-9.0, 51 and a variety of additives depending on the system in
question. The effect of temperature on the adsorption to immobilized-dye
adsorbents is equally unpredictable; sometimes raising the temperature
may increase affinity in one instance and lower it in another. Neverthe-

51The affinityof proteins for immobilizeddyes is generallyreducedon raisingthe pH of the

equilibrating buffer,t7 althoughthere are exceptions.21
108 CHROMATOGRAPHY [4]
TABLE III
REPORTED ELUENTSOF PROTEINSFROMIMMOBILIZEDTRIAZINEDYE
AFFINITYADSORBENTS

Method Data

Nonspecific
Molarity 0.025-6 M
Salts NaCI, KCI, CaC12, NH4C1, (NH4)2504
Polyols Glycerol, ethylene glycol
Chaotropes Urea, NaSCN, KSCN
Chelating agents EDTA, 2,2'-bipyridyl, pyridine dicarboxylic acid
Detergents 0.1% (w/v) SDS, Triton X-100
Specific elutants
Molarity 0.001-25 mM
Composition Coenzymes, nucleotides, polynucleotides, ternary com-
plexes, substrates, dyes
Electrophoretic desorption

less, operating temperatures in the range of 0-50 ° have been exploited;

room temperature often gives more satisfactory results than 4° provided
that the stability of the protein is not compromised.
Table III lists a number of methods that may be used to desorb bound
proteins from immobilized dye adsorbents. Two particularly favored
techniques that have been universally exploited in dye chromatography
are pH and salt elution. Biomolecules adsorbed to immobilized dye matri-
ces may be eluted by increasing the ionic strength either in steps, pulses,
or gradients 9,13,17,5z or by raising the concentration of buffer ions. 45 In
addition, many substrates, coenzymes, and other biospecific ligands have
been used to elute bound proteins, although the precise conditions se-
lected remain empirical. In some instances, a suitable combination of
substrates and coenzymes may elute the desired protein in good yield,
whereas the individual components alone may be singularly ineffective.
For example, the elution of Escherichia coli adenylosuccinate synthetase
from immobilized Procion Blue H-B is efficiently executed by quaternary
complex formation with IMP, GTP, and L-aspartate, but with much lower
yields with individual substrates or with combinations of pairs of sub-
strates, lO
The excellent durability of both dyes and agarose supports toward
chemical and biological degradation permits the immobilized dye columns
to be reused many times. However, repeated use of the columns with
crude protein extracts can lead to fouling with denatured proteins, lipids,
lipoproteins, and other lipophilic materials. Regular rejuvenation of the
52p. G. H. Byfield, S. Copping, and W. A. Bartlett, Biochem. Soc. Trans. 10, 104 (1982).
[4] IMMOBILIZEDDYES 109

columns with 8 M urea in 0.5 M NaOH, detergents (1%, w/v, sodium

dodecylsulfate or Triton X-100), ethylene glycol (<50%, v/v), or potas-
sium thiocyanate (1-3 M KSCN) is recommended. In extreme cases of
lipid fouling, washing the gel with a chloroform-methanol solution will
generally restore the adsorbent to its original utility.

Examples of the Application of Immobilized Dyes in

Affinity Chromatography
The literature abounds with examples in which immobilized dyes have
been exploited to purify individual proteins. 3,6,7,53The following examples
have been selected to illustrate specific facets of affinity chromatography
on immobilized dyes.

Purification of lnosine 5'-Monophosphate Dehydrogenase from E. coli

on Immobilized Procion Yellow MX-8G 13
A glass microcolumn (5 × 50 mm) packed with 0.5 g (moist weight) of
Sepharose 4B-bound Procion Yellow MX-8G (2.1 mg of dye per gram of
wet gel; 2.8/xmol of dye per gram of wet gel) is equilibrated at 4° with 20
mM potassium phosphate at pH 7.4 containing 15% (v/v) ethylene gly-
col. 54 A cell-free extract of E. coli (0.4 ml) that had been previously
dialyzed against 100 volumes of 20 mM potassium phosphate at pH 7.0 is
applied to the top of the column. The flow is interrupted for 10 min to
allow adsorption of the protein; nonadsorbed protein is removed by wash-
ing with the equilibration buffer (7.5 ml). Elution is effected with a linear
gradient of KC1 (0 to 1 M; 20-ml total volume) in 20 mM potassium
phosphate, pH 7.4, containing 15% (v/v) ethylene glycol. Fractions (1.5
ml) are collected at a flow rate of 7.5 ml/hr. Under these conditions, IMP
dehydrogenase is quantitatively eluted with a 14-fold overall purification
factor to yield a product of approximately 90% purity.

The Large-Scale Purification o f E. coli Adenylosuccinate Synthetase

Using Consecutive Dye-Agarose Affinity Adsorbents l°
A cell-free extract from a 40-liter culture of E. coli (Mr 1068) is par-
tially purified by adsorption and elution from DEAE-cellulose using a
NaCI gradient, 1° and the pooled fractions, containing adenylosuccinate
synthetase activity, are immediately precipitated with (NH4)2504. Dialy-
53 I. G. I. Pesylakas, O. F. Sudzhyuvene, and A. A. Glemzha, Appl. Biochem. Microbiol.
17, 337 (1981).
54 Ethylene glycol (15% v/v) is included throughout the column irrigants to reduce nonspe-
cific hydrophobic interactions between the protein and the dyed agarose.
110 CHROMATOGRAPHY [4]
sis against 1000 volumes of 50 mM Tris-HCl buffer, pH 7.5, containing
1.5 mM MgC12 and 2.0 mM dithiothreitol, yields a preparation showing a
7.6-fold increase in specific activity over the cell-free extract and an over-
all 90% yield.
The dialyzed preparation (7.1 ml, 4800 units of adenylosuccinate syn-
thetase activity; 284 mg of protein) is applied to a column (1. I × 16.5 cm;
16 g, moist weight, of gel) of agarose-bound Procion Red H-3B (1.4 ~mol
of dye per gram, moist weight, of gel), equilibrated with the same Tris buf-
fer, and is used for dialysis. The sample is applied to the column, flow is
stopped for 5 min to allow equilibration, and nonadsorbed protein is re-
moved by washing with the Tris buffer (90 ml). Adenylosuccinate synthe-
tase activity is eluted with a linear gradient of KC1 (0 to 1 M; 120 ml total
volume) and collected at a flow rate of 30 ml/hr. Pooled, active fractions
are concentrated to a 2-ml volume through an ULVAC ultrafiltration
membrane and dialyzed overnight against 1 liter of the Tris buffer. The
dialysate (2.8 ml; 2128 units of enzyme activity; 16.8 mg of protein) is
applied to a column (1.1 × 13.5 cm; 12.8 g of gel, moist weight) of agarose-
bound Procion Blue H-B (0.8 tzmol of dye per gram of gel, moist weight)
equilibrated with 50 mM Tris-HCl at pH 7.5 containing 1.5 mM MgCI2 and
2.0 mM dithiothreitol. Flow is interrupted for 5 min, and nonadsorbed
protein is removed with 60 ml of the equilibration buffer. Adenylosucci-
nate activity is eluted with 50 ml of equilibration buffer containing IMP (4
mM), GTP (2 mM), and L-aspartate (10 mM) at a flow rate of 40 ml/hr.
Adenylosuccinate synthetase is pooled and concentrated (1.8 ml) to give
the final preparation composed of 3.6 mg of protein and 1829 units of
activity.

Purification of Yeast Hexokinase by Mg2+-Promoted Binding to

Procion Green H-4G 15
Preliminary studies on the inactivation of yeast hexokinase by a range
of triazine dyes showed that, of those tested (Procion Green H-4G, Blue
H-B, Turquoise H-A, Turquoise H-7G, Brown H-2G, Green H-E4BD,
Red H-E3B, Yellow H-5G, and Yellow H-A), the enzyme was inactivated
most rapidly by Procion Green H-4G. Almost complete protection from
inactivation was afforded by 20 mM ATP, suggesting that the inactivation
by Procion Green H-4G was active-site directed. A most interesting prop-
erty of this dye was that in the presence of 10 mM MgC12 the rate of
inactivation by Procion Green H-4G dropped to 20% of that recorded in
the absence of Mg 2+. The effect was unique for this dye and associated
with a fivefold decrease in the apparent dissociation constant of the d y e -
enzyme complex. 15These unique properties of Procion Green H-4G were
[4] IMMOBILIZED DYES 111

exploited in the purification of yeast hexokinase by affinity chromatogra-

phy on Procion Green H-4G-Sepharose 4B.
A concentrated crude yeast extract (50 mg in 1 ml of 30 mM Tris-HC1
at pH 7.5 containing I0 mM MgCl2) is dialyzed overnight against 1 liter of
the same buffer at 4 °. A sample (400/.d; 24.3 units of hexokinase activity;
10.5 mg of protein) of the extract is applied to a column (10 × 30 mm;
2.5 g, of gel moist weight) of Sepharose 4B-immobilized Procion Green
H-4G (2.6/¢mol of dye per gram of gel, moist weight) equilibrated with the
Tris buffer. The column is washed with about 12 ml of the same buffer.
Elimination of MgC12 from the equilibration buffer elutes approximately
7.5% of the applied hexokinase activity, whereas a pulse (5 ml) of 20 mM
ATP-10 mM MgCl2 is required to elute the remaining enzyme activity.
The peak fraction eluted after the ATP pulse displayed a specific activity
sevenfold greater than that of the crude yeast extract.
Other examples of metal ion-promoted protein adsorption to immobi-
lized triazine dyes have been examined for potential exploitation in affin-
ity chromatography 17,26and high-performance liquid affinity chromatogra-
phy. 38'39In particular, low concentrations of first row transition metal ions
such as Zn 2+, Co 2+, Mn 2+, Ni 2+, Cu 2+ and, to a lesser extent, the group
IIA ions, Ca z+ and Mg 2÷, promote the binding of proteins to a number of
immobilized triazine dye affinity adsorbents. 17 For example, Zn 2+ pro-
motes binding of carboxypeptidase G2, alkaline phosphatase, and yeast
hexokinase to immobilized Procion Red H-8BN, Procion Yellow H-A, and
Cibacron Blue F3G-A, respectively. The binding of ovalbumin to immobi-
lized Cibacron Blue F3G-A and Procion Orange MX-G is selectively en-
hanced in the presence of A13+. In the case of ovalbumin and alkaline
phosphatase, the effect is almost totally specific for the protein, metal ion,
and dye, whereas with carboxypeptidase G2 and hexokinase, metal ions
such as Co 2÷, Ni 2÷, Mn 2÷, Cu 2+, Ca 2÷, and Mg 2÷ also promote binding to
varying degrees; almost all other monovalent and trivalent metal ions
appear to be ineffectual. Proteins bound to immobilized triazine dyes by
metal ion-promoted interactions can subsequently be eluted with appro-
priate chelating agents of the amine, aminocarboxylate, or substituted
pyridine classes. 17

The Purification of 3-Hydroxybutyrate Dehydrogenase and Malate

Dehydrogenase from Rhodopseudomonas spheroides7A9
A cell-free extract from R. spheroides is applied to a 1.5-liter column
of agarose-bound Procion Red H-3B; 3-hydroxybutyrate dehydrogenase
is eluted with 1 M KC1, and malate dehydrogenase is subsequently recov-
ered by including 2 mM NADH in the elution buffer. Pooled and dialyzed
112 CHROMATOGRAPHY [4]
fractions containing malate dehydrogenase activity are applied to a
column (1 liter) of immobilized Procion Blue MX-4GD, and the enzyme is
eluted with a linear gradient of KCI (0 to 700 mM; 8 liters) to yield 1 g of
homogeneous enzyme in 70% overall yield. The fractions containing 3-
hydroxybutyrate dehydrogenase activity eluted from Procion Red H-3B
are applied to a column (1 liter) of immobilized Procion Blue MX-4GD
and subsequently eluted with 2 mM NADH to yield 300 mg of homoge-
neous enzyme in 80% overall yield. 55

Other Applications of Triazine Dyes in Biotechnology

Dye columns have been exploited for the removal of contaminants
from almost homogeneous protein preparations, 56 especially the removal
of serum albumin from samples of IgG, 57the production of apoenzymes,S8
the resolution of functionally active from inactive proteins 59 and isoen-
zymes, 38 and the fractionate multienzyme complexes. 6° Reactive dyes
bound to microparticulate silica have been extensively used to resolve
enzymes and other proteins by analytical 38,39 and preparative 6~ high-per-
formance liquid affinity chromatography. Triazine dyes have been used as
chromogenic substrates for enzymes62 and lectins, 63 as ligands for novel
electrochemical sensors, 35,64 as histochemical stains and colored protein
markers, 7 and as active-site probes and irreversible affinity labels.15,5°,65

Comments
The three principal characteristics of reactive dyes, namely their abil-
ity to bind a plethora of proteins, their spectral properties, and their
reactivity with nucleophilic groups, ensure a wide range of applications
55 This protocol for the purification of 3-hydroxybutyrate dehydrogenase compares favor-
ably with the conventional scheme involving nine separate steps with an overall yield of
9%.
56 M. Dao, J. Watson, R. Delaney, and B. Johnson, J. Biol. Chem. 254, 9441 (1979).
57 j. Travis, J. Bowen, D. Tewksbury, D. Johnson, and R. Pannell, Biochem. J. 157, 301
(1976).
5s S. T. Thompson, R. Cass, and E. Stellwagen, Anal. Biochem. 72, 293 (1976).
59 p. Menter and W. Burke, Fed. Proc. Fed. Am. Soc. Exp. Biol. 38, 673 (1979).
R. DeAbreu, A. Dekok, and C. Veeger, FEBS Left. 82, 89 (1977).
61 D. A. P. Small, A. Atkinson, and C. R. Lowe, J. Chromatogr. 266, 151 (1983).
62 B. Klein and J. A. Foreman, Clin. Chem. 26, 250 (1980).
63 B. S. Rathaur, G. S. Khatri, K. C. Gupta, G. K. Narang, and N. K. Mathur, Anal.
Biochem. 112, 55 (1981).
64 C. R. Lowe and M. J. Goldfinch, Biochem. Soc. Trans. 11, 448 (19,83).
65 D. A. P. Small, C. R. Lowe, A. Atkinson, and C. J. Bruton, Eur. J. Biochem. 128, 119
(1982).
[5] DISPLACEMENT CHROMATOGRAPHY 113

for these dyes. The purification of proteins by biospecific affinity chroma-

tography is possibly the broadest and best documented application to
date, although their potential in other areas of biotechnology should not
be underestimated. 7 This growing interest in the application of triazine
dyes in preparative and analytical biotechnology has prompted investiga-
tions into the basis of the protein-dye interactions. Thus, it has been
suggested that the polyanionic aromatic dye chromophores mimic the
overall shape, size, and charge distribution of the naturally occurring
biological heterocycles, such as the nucleotides and coenzymes. Indeed,
studies by induced circular dichroism, 3°,66 X-ray crystallography,67 and
irreversible enzyme inactivation 15,5°,65 all confirm this suggestion. How-
ever, the available evidence indicates that Cibacron Blue F3G-A and
other triazine dyes are not highly specific analogs of nucleotide coen-
zymes and that there is some degree of latitude in binding reactive dyes to
active-site regions of proteins. Nevertheless, reactive dyes clearly have a
bright future in preparative and analytical biochemistry.
R. A. Edwards and R. W. Woody,Biochemistry 18, 5197 (1979).
67j. F. Beillmann,J.-P. Samama,C.-I. Brhndrn,and H. Eklund,Eur. J. Biochem. 102, 107
(1979).

[5] D i s p l a c e m e n t C h r o m a t o g r a p h y o f P r o t e i n s
By ELBERTA. PETERSON and ANTHONY R. TOaREs

Whereas the separation of substances in elution chromatography, the

original and most widely used form of chromatography, depends upon
differences among the sample components with respect to their ability to
compete with the eluent for sites on the adsorbent, displacement chroma-
tography is based on direct competition among the adsorbed components,
themselves, including substances of intermediate affinity that may be in-
troduced as spacers. This competition gives rise to a train of contiguous
bands having successively higher affinities for the adsorbent, all being
driven through the column at the same velocity by the continuous addi-
tion of a solution containing a substance that has the highest affinity in the
system. In the classical displacement train, first described by Tiselius I for
sugars chromatographed on a column of carbon with phenol as the "de-

mA. Tiselius, Ark. Kemi Mineral. Geol. 16A, n18 (1943).

Copyright © 1984by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All fights of reproduction in any form reserved.
ISBN 0-12-182004-1
[5] DISPLACEMENT CHROMATOGRAPHY 1 13

for these dyes. The purification of proteins by biospecific affinity chroma-

tography is possibly the broadest and best documented application to
date, although their potential in other areas of biotechnology should not
be underestimated. 7 This growing interest in the application of triazine
dyes in preparative and analytical biotechnology has prompted investiga-
tions into the basis of the protein-dye interactions. Thus, it has been
suggested that the polyanionic aromatic dye chromophores mimic the
overall shape, size, and charge distribution of the naturally occurring
biological heterocycles, such as the nucleotides and coenzymes. Indeed,
studies by induced circular dichroism, 3°'66 X-ray crystallography,67 and
irreversible enzyme inactivation 15'5°,65all confirm this suggestion. How-
ever, the available evidence indicates that Cibacron Blue F3G-A and
other triazine dyes are not highly specific analogs of nucleotide coen-
zymes and that there is some degree of latitude in binding reactive dyes to
active-site regions of proteins. Nevertheless, reactive dyes clearly have a
bright future in preparative and analytical biochemistry.
R. A. Edwards and R. W. Woody, Biochemistry 18, 5197 (1979).
67 j. F. Beillmann, J.-P. Samama, C.-I. Br~indrn, and H. Eklund, Eur. J. Biochern. 102, 107
(1979).

[5] D i s p l a c e m e n t C h r o m a t o g r a p h y of P r o t e i n s
By ELBERTA. PETERSON and ANTHONY R. TORRES

Whereas the separation of substances in elution chromatography, the

original and most widely used form of chromatography, depends upon
differences among the sample components with respect to their ability to
compete with the eluent for sites on the adsorbent, displacement chroma-
tography is based on direct competition among the adsorbed components,
themselves, including substances of intermediate affinity that may be in-
troduced as spacers. This competition gives rise to a train of contiguous
bands having successively higher affinities for the adsorbent, all being
driven through the column at the same velocity by the continuous addi-
tion of a solution containing a substance that has the highest affinity in the
system. In the classical displacement train, first described by Tiselius ~for
sugars chromatographed on a column of carbon with phenol as the "de-

i A. Tiselius, Ark. Kemi Mineral. Geol. 16A, hi8 (1943).

Copyright © 1984by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All fights of reproduction in any form reserved.
ISBN 0-12-182004-1
114 CHROMATOGRAPHY [5]

veloper," and observed by others in the displacement chromatography of

hydrocarbons, z fatty acids, 3 steroids: and amino acids and peptides, 5,6
the separated components emerge as contiguous plateaus, each higher in
concentration than the preceding one when plotted on a mass basis. The
concentration of each is a reflection of the affinity of that component for
the adsorbent under the prevailing conditions and serves to identify the
substance. The area under the plateau is proportional to the amount,
reducing the determination of quantity to a measure of height and length
(concentration and volume). So long as the concentration of the devel-
oper remains constant, increasing the amount of any component in the
sample increases the length of its plateau but does not affect the concen-
tration at which it emerges.
Displacement chromatography has been applied to separations of such
widely different classes of substances as amino acids, 7,8 rare earths, 9 and
oligonucleotides 1° on ion exchangers, and it has been used on reversed-
phase columns designed for high-performance liquid chromatography to
fractionate a variety of organic substances on a preparative scale. H
From the beginning, it was recognized that displacement chromatogra-
phy offered several advantages over elution chromatography: (1) the elim-
ination of tailing caused by convex isotherms, (2) a high capacity resulting
from high concentrations attainable in the separated bands and main-
tained by the band-sharpening that prevented widening of bands once the
train had been fully developed, (3) the potential for spacing the compo-
nents, and (4) the ease of measuring the quantity of each component in
analytical applications. The mixing that occurred at the boundaries of the
contiguous plateaus was later overcome by including other substances
with intermediate affinities to serve as spacers that could readily be re-
moved from the sample components after chromatographic separation) ,8
Application of these principles has been limited by the problem of finding
suitable spacers and developers, although the search is aided in the case
of small molecules by the fact that the positions of the components in the
train can be determined from their isotherms, 1as measured on the adsorb-

2 S. Claesson, Ann. N.Y. Acad. Sci. 49, 183 (1948).

3 R. T. Holman and L. Hagdahl, J. Biol. Chem. 182, 421 (1950).
4 j. G. Hamilton, Jr., and R. T. Holman, Arch. Biochem. Biophys. 36, 451 (1952).
A. Tiselius and L. Hagdahl, Acta Chem. Scand. 4, 394 (1950).
6 j. Porath, Acta Chem. Scand. 6, 1237 (1952).
7 S. M. Partridge and R. C. Brimley, Biochem. J. 51, 628 (1952).
8 D. L. Buchanan, J. Biol. Chem. 229, 211 (1957).
9 F. H. Spedding and J. E. Powell, in "Ion Exchange Technology" (F. C. Nachod and J.
Schubert, eds.), Chap. 15. Academic Press, New York, 1962.
l0 G. G. Brownlee, F. Sanger, and B. G. Barrell, J. Mol. Biol. 34, 379 (1968).
i1 C. Horvath, A. Nahum, and J. H. Frenz, J. Chromatogr. 218, 365 (1981).
[5] DISPLACEMENT CHROMATOGRAPHY 115

ent to be employed in chromatography. In ion exchange, the dissociation

constants of the organic acids and bases serve as a guide, s
The situation is quite different in the ion-exchange chromatography of
macromolecular polyelectrolytes such as proteins. Under pH and salt
conditions that permit adsorption of most of the proteins in a biological
mixture, only a few, if any, will be in finite adsorption equilibrium with
the adsorbent; the others will be either not adsorbed or so tightly bound
by multiple bonds that they will not migrate appreciably in the initial
buffer in the absence of displacers. In the latter case adsorption isotherms
will not be helpful. Also, the number of different molecular species in a
natural sample is likely to be very large, with few, if any, present in
quantities sufficient to form plateaus. Nevertheless, all the other advan-
tages of displacement over elution cited above for small molecules apply
to the displacement chromatography of proteins, provided suitable sys-
tems can be developed to achieve it. Two approaches have been taken to
accomplish this: one is the use of the complex mixtures of ampholytes
that have been developed for isoelectric focusing, and the other involves
the use of polyanions prepared specifically for the displacement chroma-
tography of proteins. In the discussion of these systems and their applica-
tions, care will be taken to distinguish elution from displacement. 1 Al-
though the word "elute," according to its dictionary meaning, "to wash
away," is a term broad enough to cover any chromatographic process, its
use will be confined to the process in which the sample components
compete with eluent molecules, which have lower affinity for the adsorb-
ent than their own, a process that has been historically associated with the
word "elution." When, however, substances are caused to migrate along
the column because of competition with substances having affinities for
the adsorbent that are greater than their own, the word "displace" will be
used. At the same time, it must be kept in mind that both mechanisms can
operate in different parts of the chromatogram. If the concentration of the
developer is so low that the developer front moves more slowly than the
rates at which low-affinity components are eluted by the background salt,
these will move ahead as eluted peaks.

Ampholyte Displacement Chromatography

Leaback and Robinson 12introduced a procedure in which the ampho-
lytes developed for isoelectric focusing are used as displacers of proteins
on ion exchangers. By this means they were able to separate two forms of
the B variant of pig epididymal/3-N-acetyl-D-hexosaminidase that could
12 D. H. Leaback and H. K. Robinson, Biochem. Biophys. Res. Commun. 67, 248 (1975).
116 CHROMATOGRAPHY [5]
not be resolved by isoelectric focusing. An excellent separation was ob-
tained when 1 ml of the partially purified enzyme preparation was applied
to a 45 × 11 mm column of CM-cellulose (Whatman CM-52) that had been
equilibrated at 5° with 0.02 M Tris-HC1 at pH 7.6, and 4% (w/v) of pH 8-
10 Ampholine (LKB) was passed through. Efforts to resolve the two
forms by elution from a similar column with a shallow gradient of NaC1
were unsuccessful. A gradient of increasing pH would have been more
appropriate for this comparison in view of the fact that the first isoenzyme
emerged in the steep pH gradient imposed by the Ampholine solution.
These authors concluded that the two proteins were displaced selectively
by amphoteric ions having closely similar isoelectric points, and this ex-
planation has largely been accepted by others who have used similar
procedures in a variety of applications) 2-19
Figure 1 shows one such application reported by Young and Webb, 13
the separation of human serum albumin from ot-fetoprotein on DEAE-
cellulose (Whatman DE-52) with 4% (w/v) pH 4-6 Ampholine in l0 mM
sodium phosphate at pH 7.8. The emergence of the o~-fetoprotein in a
number of peaks and shoulders led the authors to suggest that this protein
occurs in multiple forms, as do many serum proteins, frequently based on
differences in the content of sialic acid. 2° Although this is likely to be the
case, the possibility that this multiplicity might be the result of the forma-
tion of complexes between certain ampholytes and the protein should be
considered. Since the peaks shown in the figure are based on immunologi-
cal assays, no determination of the presence or the absence of other
protein in the background can be made. Page and Belles-Isles ~4employed
a similar procedure as a final step in the purification of mouse ~-fetopro-
tein from hepatoma tissue and obtained a product that gave a single band
on electrophoresis after treatment with sodium dodecyl sulfate.
Although the ampholyte solution absorbs too strongly at 280 nm to
permit monitoring at that wavelength, 18absorbance at 415 nm was used by
Chapuis-Cellier et al. ~5 to delineate the profiles of hemoglobin variants
(stabilized with KCN) that were separated with 1.2% pH 7-9 Ampholine
on DE-52 columns. Relatively smooth pH gradients were generated by
passing the Ampholine, adjusted to pH 7.5, 7.8, or 8.0, through columns
13 j. L. Young and B. A. Webb, Anal. Biochem. 88, 619 (1978).
t4 M. Page and M. Belles-Isles, Can. J. Biochem. 56, 853 (1978).
t5 C. Chapuis-Cellier, A. Francina, and P. Arnand, Protides Biol. Fluids 27, 743 (1980).
1~A. Francina, E. Dorleac, and H. Cloppet, J. Chromatogr. 222, 116 (1981).
17j. L. Young and B. A. Webb, Protides Biol. Fluids 27, 739 (1980).
is j. L. Young and B. A. Webb, Sci. Tools 25, 54 (1978).
19j. L. Young, B. A. Webb, D. G. Coutie, and B. Reid, Biochem. Soc. Trans. 6, 1051 (1978).
2o L. Anderson and N. G. Anderson, Proc. Natl. Acad. Sci. U.S.A. 74, 5421 (1977).
[5] DISPLACEMENT CHROMATOGRAPHY 117

30g ISO

./!
c 200 I00 r,
a,

I00 SO

* z~ I ' I I I I
// 30 SO 70 90

Fraction number
FIG. 1. Ampholyte displacement chromatography of amniotic fluid on DEAE-cellulose.
A 1 : 4 dilution of amniotic fluid was dialyzed against 10 mM sodium phosphate, pH 7.8, and
3 ml was applied to a 10 x 0.9 cm column of DE-52 equilibrated with the same buffer. Then a
4% (w/v) solution ofpH 4-6 Ampholine in the pH 7.8 buffer was pumped into the column at
13.5 ml/hr. Fractions (0.625 mi) were assayed for a-fetoprotein (AFP) (A) and albumin (A)
by immunological methods. From Young and Webb. 13

equilibrated with 0.05 M Tris-HCl at pH 8.9, and the separated hemoglo-

bins migrated within these gradients. Under such conditions the separa-
tion process appeared to be essentially chromatofocusing. Francina et
al.n6 found the relative mobilities of hemoglobin variants that were sepa-
rated on an anion-exchange paper (Whatman DE-8 l) to be linearly related
to the isoelectric points when either 1.2 or 4% concentrations of pH 6-8
Ampholine (not adjusted in pH) were used, a result compatible with either
displacement or elution (including chromatofocusing in its most general
sense). However, Young and Webb 17 have shown that when 4 or 6%
solutions of Ampholine are used, a region of relatively constant pH is
formed in which protein peaks can migrate, with narrower peaks of other
proteins appearing where the pH changes sharply at either end of this
region. Moreover, they found that the pH did not have to fall to a value at
or near the isoelectric point of a protein in order for that protein to be
released from the adsorbent. Also, in other work TM in which the same
authors determined the positions of a number of proteins in a chromato-
118 CHROMATOGRAPHY [5]

gram of maternal serum by immunological means, they observed that not

all of these emerged in the order of their isoelectric points. Again, the
result is compatible with both displacement and elution mechanisms, al-
though not generally expected in chromatofocusing in its special sense,
where salt is absent or very low in concentration. However, all these
mechanisms are affected by the presence of background salt, which will
cause the release of the protein from an anion exchanger above its isoelec-
tric point, where selection is made on the basis of the number of remain-
ing effective charges. The isoelectric point is merely a condition where
positive and negative charges are balanced; it does not reflect differences
in density, distribution, or total numbers of charges, all of which affect the
number of effective bonds that can be made with the adsorbent at pH
values away from the isoelectric point. Since adsorption involves interac-
tions with regions of the protein rather than the entire surface of the
molecule, taken as a whole, as is the case in the electrophoretic methods
that are used to determine the isoelectric point, one can expect some
proteins that have unusual concentrations of like charge in certain regions
of their molecules to remain adsorbed at their isoelectric points (in low
salt) or even beyond it.
Decreasing the concentration of a displacement developer, increasing
the background salt concentration, or altering the pH can convert a dis-
placement process into an elution process, at least in part. Therefore, to
identify a displacement mechanism it is necessary to determine, with
relatively pure proteins so that plateaus can be formed under suitable
circumstances, the shapes of the bands and how they are affected by
changes in the sample load and in the concentration of the displacers, as
has been done in the case of the carboxymethyldextrans that are de-
scribed in the next section.

Carboxymethyldextran Displacers

Principle
Any soluble polyanion having a sufficient number of negative charges
per molecule to match those offered to a cluster of charges on an anion
exchanger by a given protein can be expected to compete with that pro-
tein in a displacement train, though differences in the spacing of the
charges may give one or the other an advantage. If the polyanion can offer
an additional negative charge, it is likely that it will move the protein
ahead of it on a column that has sufficient resolving power to discriminate
[5] DISPLACEMENT CHROMATOGRAPHY 119

between them. Any series of soluble polyanions that do not interact by

nonionic mechanisms with the adsorbent or with each other and have
sufficient charge to be adsorbed should form a displacement train based
on differences in their ion-exchange affinities for the adsorbent, i.e., on
differences in the number of effective negative charges per molecule. It is
also reasonable to assume that proteins added to the column will find
positions in that train according to their own abilities to compete, as all
the components of the train migrate through the column in finite adsorp-
tion equilibrium with the adsorbent. Such a system of polyanions can be
made up of molecules of uniform molecular weight with varying densities
of charge, or of varying molecular weight with uniform density of charge,
or molecules that are heterogeneous in both respects. To facilitate the
characterization of the polyanions and the control of their affinities, the
first of these alternatives was chosen, in principle. Specifically, carboxy-
methyldextrans (CM-Ds) 21 were selected to constitute the polyanion se-
ries because dextran is hydrophilic and is commercially available in rela-
tively narrow ranges of molecular weight, and carboxymethyl groups can
be readily attached to it by reaction with monochloroacetate under alka-
line conditions. The number of carboxymethyl groups incorporated per
average dextran chain (or per unit of mass) can be controlled by limiting
the amount of monochloroacetate used, the time of reaction, or the tem-
perature of reaction. The amount of water and alkali in the reaction mix-
ture also affects the incorporation.
Actually, a CM-D that has been prepared under fixed conditions of
reaction is, nevertheless, heterogeneous with respect to the affinity of its
constituent molecules, and this heterogeneity is very much greater in the
low-affinity range than at higher levels of substitution. In addition to the
variation in affinity due to heterogeneity in molecular weight, statistical
variation in the degree of substitution is to be expected in an incompletely
substituted product, and differences in the reactivity of highly branched
and sparsely branched dextran chains affect the distribution of the charge-
able groups among the chains. Although the chosen starting material is
nominally 10,000 in molecular weight, the lowest in the series offered by
Pharmacia Fine Chemicals, it contains substantial quantities of dextran
smaller than 5000 or larger than 15,000, according to the manufacturer's
analysis. The unfractionated product might therefore appear to be in the
category of the third alternative: heterogeneous with respect to both mo-
lecular weight and density of charge. However, the very high-molecular-
weight components that might lead to excessive bonding, high viscosity,

21 E. A. Peterson, Anal. Biochem. 90, 767 (1978).

120 CHROMATOGRAPHY [5]

and difficulty of migration in gel electrophoresis are excluded. Unfrac-

tionated preparations may include so wide a range of affinities that only a
small fraction accurately represents the value that is used to characterize
the preparation, whatever index of affinity may be employed. Conse-
quently, they are inefficient in spacing proteins that differ only slightly in
affinity. However, these widely overlapping products can be used in com-
bination to provide a partial or full-range continuum of affinity, permitting
a scan of a mixture of proteins or a preliminary determination of the
affinity index of the CM-D that follows or precedes a protein of interest in
a displacement train.
Since the ability to space the proteins in the chromatogram with ap-
propriate displacers is a major advantage of displacement chromatogra-
phy, it is desirable to have at hand a series of preparations each of which
represents a narrow segment of the whole spectrum of potentially re-
quired affinities, so that proteins with closely similar affinities for the
adsorbent can be caused to emerge from the column well separated. Such
preparations also make possible the focusing of the resolving power of the
system on the narrow range of affinities represented by a single protein of
interest and its neighboring impurities.

Preparation of CM-Ds
In early work, 2~ a number of preparations of CM-D were produced by
mixing monochloroacetic acid, water, dextran, and lithium hydroxide in
various proportions and allowing the resulting solution to stand, sealed at
room temperature, for periods ranging up to 72 hr. A purification proce-
dure involving precipitation of the CM-D by ethanol provided the solid
product, initially in the free-acid form, separated from lithium chloride,
glycolic acid, and residual unreacted monochloroacetic acid. Very highly
substituted products had to be precipitated with acetone instead of etha-
nol because of their solubility in the latter. The CM-Ds were finally con-
verted to their potassium salts (which gave better precipitates than the
sodium salts) for characterization and storage because long-term storage
of the solid in its free-acid form at 4° was found to produce changes in its
solubility, apparently as the result of cross-linking by ester bonds.
This method of synthesis, which has the advantage of requiring only
the simplest of laboratory equipment and is relatively efficient in the use
of reactants, is still employed to prepare the relatively high-affinity CM-
Ds that serve as final displacers. At their high level of substitution, the
original product is usually sufficiently homogeneous for this purpose.
However, CM-Ds that are to be used as spacers are now made by a
reaction procedure that first produces a series of preparations that repre-
[5] DISPLACEMENT CHROMATOGRAPHY 121

sent overlapping segments of a continuum of affinities, populated with

molecules at every level of substitution below a chosen limit. These are
later fractionated by displacement chromatography in water on DEAE-
cellulose to obtain a series of narrow-range CM-Ds that are suitable even
for spacing proteins that have closely similar affinities.
The initial synthesis is carded out under conditions that produce an
approximately linear time vs reaction curve, 2~ the reaction mixture being
withdrawn continuously 2z or at short intervals 23 to receivers that contain
acid to stop the reaction. Thus, several separate preparations are pro-
duced that together compose a distribution of molecules that have been
exposed to progressively longer reaction periods. After purification by
alcohol precipitation, these are fractionated by displacement chromatog-
raphy on a DEAE-cellulose column. In order to avoid the disturbing
effects of the rather high concentrations of displaced counterion that are
developed when high concentrations of CM-D are used, 2~the adsorbent in
its free-base form is titrated in water with the free-acid form of the CM-D
preparation that has the lowest level of substitution. This results in the
binding of an amount of CM-D that is comparable to the dry weight of the
adsorbent itself. This complex of adsorbent and low-affinity CM-D is
packed into a column, and solutions of the potassium salts of CM-Ds
having progressively higher affinity are subsequently pumped in. The
CM-D is the only anion, provided the preparations employed are free of
chloride and organic acids.
Figure 2 illustrates the shape of the refractive index profile of the CM-
D after such a fractionation on DEAE-cellulose (Whatman DE-23). It is
essentially the mass profile of a CM-D displacement train in water, dem-
onstrating that, in the absence of competing salt, the capacity of DEAE-
cellulose for CM-D increases as the affinity of the CM-D decreases, i.e.,
as the number of carboxyl groups on the dextran molecule decreases.
This is attributable to the fact that a larger mass of dextran must accom-
pany the carboxyl groups that are necessary to saturate the adsorbent
with low-affinity CM-D. 21 There is a high degree of conformational ac-
commodation of both the charged adsorbent chain and the CM-D to the
spacing of charges on the other, leading to a very high efficiency in the
matching of charges on the CM-D with those on the adsorbent. Since
some of the components of the train were applied after earlier, low-affin-
ity ones had emerged, each period during which a different preparation
was being pumped in represented a segment of the train in which a differ-
ent CM-D with a different affinity was serving as a "final displacer."
22 E. A. Peterson and A. R. Tortes, Anal. Biochem. 130, 271 (1983).
A. R. Torres and E. A. Peterson, J. Biochem. Biophys. Methods 1, 349 (1979).
122 CHROMATOGRAPHY [5]

_ 13

1
1"340 1
pH

o o 20

II O u O 0.11N KOH
X_ 10
RPV

,,,, ,,v ,
1.330 r ° ~ ' ' ~ ' ' 0
0 10 20 30 40 50 60

FRACTION

FIG. 2. Fractionation of carboxymethyldextran (CM-D) preparations into narrow-range

CM-D by displacement chromatography. A water slurry of DE-23 in the free-base form was
adjusted to pH 5.6 with a salt-free 6% solution of CM-D (RPV = 9) in the free-acid form, and
this suspension was packed into a column to form a 115-ml bed. Water solutions of the
potassium salts of CM-D preparations were then pumped successively into the column at 50
ml/hr: about 150 ml each of 6% II (RPV = 14), 3.6% III (23), and 3% IV (27). CM-D
remaining on the column was displaced by 0.11 N KOH. The volume of each fraction was
12.5 ml, but pairs were pooled before determination of the reciprocal of the pellet volume
(RPV). The heavy line represents the refractive index (n); the thin line, the pH; and the
circles, RPV. From Peterson and Torres. 22

Therefore, the successive preparations of CM-D were applied at concen-

trations that would drive the various segments at approximately the same
rate in order to minimize changes in the shape of the train. The raised
plateau at the end of the train resulted from the use of a concentration of
KOH that was about 20% too high in displacing the final CM-D from the
column.
Recovery of the fractionated CM-D, all of which emerged in the form
of its potassium salt, was quantitative: 34 g of CM-D was applied and 34 g
was recovered, including 2.6 g of unadsorbed material and 1.2 g of high-
affinity CM-D that was recovered in the alkaline end fractions after re-
moval of alkali by treatment with the acid form of Bio-Rex 70 (Bio-Rad).
The solutions were stored in the frozen state, without preservatives.

Characterization of the CM-D

The open circles (RPV) in Fig. 2 represent values obtained in an assay
that provides an indirect measure of the carboxyl group content of the
CM-D, and they therefore serve as an affinity index. In early work 21,23the
[5] DISPLACEMENT CHROMATOGRAPHY 123

CM-D preparations were characterized by a specific absorption index,

AEEo/An(where An is the difference between the refractive index of the
CM-D solution and that of water), based on the strong absorbance of the
carboxyl group at that wavelength. Values of A220]An were found to be
essentially proportional to the NaC1 concentration at which preparations
of CM-D were eluted from DEAE-cellulose, and they served to character-
ize pure CM-D with respect to affinity. However, this index could not be
used to evaluate the affinity of CM-D emerging in effluent fractions con-
taining proteins separated by displacement because of the overwhelming
absorption, at this wavelength, by proteins and other substances intro-
duced with the sample. Fortunately, the observation that the extent of
shrinkage of DEAE-Sephadex A-50 when saturated with CM-D was in-
versely related to the A220/An of the CM-D provided the basis for an assay
that could be applied to both pure CM-D and that in protein-containing
fractions. 2z
To 2.5-ml portions of 1% solutions of CM-D in Bauer-Schenck tubes
(3-ml glass centrifuge tubes with long, narrow tips graduated from 0 to 0.4
ml in 0.004-ml divisions) are added 0.50-ml portions of a 1% suspension of
DEAE-Sephadex A-50 containing sufficient acetate buffer to bring the
mixture to pH 5. After equilibration on a rotator for 1 hr, the suspension is
centrifuged for 20 min at 2000 rpm in a Model PR-2 International centri-
fuge, and the volumes of the pellets are read to the nearest microliter. The
reciprocal relationship of these volumes to the A22o/An values suggests
that the differences arise from a limitation of the extent of shrinkage by
the effective volume of the adsorbed CM-D molecules as the electrostatic
forces that caused the initial extensive swelling of the DEAE-Sephadex
A-50 beads are strongly diminished or eliminated by saturation of the
adsorbent with CM-D. Since, at saturation, approximately the same num-
ber of charges on the adsorbent are utilized by CM-D of different levels of
substitution, 21 the extent of this limitation is a direct function of the equiv-
alent weight of the CM-D employed. The reciprocal of the pellet volume
(RPV) is therefore a direct function of the carboxyl content of the CM-D
and, accordingly, serves as an index of affinity.
The assay is carried out at pH 5, where the affinity of CM-D for
DEAE-Sephadex A-50 is very strong, since the adsorbent is maximally
charged and the affinity of most proteins is very weak or nil because of an
adverse balance of positive and negative charges. This evaluation of the
CM-D can therefore be carried out on protein-containing fractions from a
displacement chromatogram as well as on pure CM-D. It is essential that
an excess of CM-D be provided, since the DEAE-Sephadex beads must
be saturated. A larger amount of low-affinity CM-D is required than high-
affinity for this purpose, but the concentrations of CM-Ds in the effluent
124 CHROMATOGRAPHY [5l
fraction of a fully developed displacement train on DEAE-cellulose at
moderate pH and at low salt are automatically adjusted by the displace-
ment process to provide the necessary levels so long as the concentration
of the final displacer is, itself, high enough (e.g., 0.5%) to saturate the
beads in the assay.

Protein Separations
Principle. When human serum or other biological mixtures containing
colored anionic proteins are applied to a column of fibrous DEAE-cellu-
lose under conditions permitting adsorption, the proteins arrange them-
selves in a series of colored bands. As additional sample enters the
column, each band widens and shifts position to accommodate the ex-
panding widths of the bands behind it, providing evidence that protein-
protein displacement is occurring. This phenomenon has even been used
to separate the various components of bovine plasma albumin, the frac-
tions being recovered by sectioning the column and eluting the protein
from each section. 24 The CM-Ds provide a means of continuing the dis-
placement process that starts with the application of the protein sample,
causing the otherwise immobilized protein bands to move through the
column in a displacement train of CM-D that also spaces them to achieve
useful separations. The electrophoretic evaluation of every fraction in the
chromatogram is facilitated by the fact that the conductivity of the CM-D
is low enough not to interfere with electrophoresis, though its mobility is
so high that it migrates with the discontinuity in disc gel electrophoresis,
as revealed by the highly refractile band observed at that position. There-
fore, all of the chromatographic fractions can be applied directly to the gel
without dialysis. Since displacement chromatograms provide fractions at
suitable concentrations, concentrating procedures have not been neces-
sary.
The serum proteins composed too complex a mixture to allow confir-
mation that the operating mechanism was displacement, apart from the
observation that the array of colored bands moved down the column at
constant velocity and constant width. 2~Attention was therefore turned to
the study of " p u r e " proteins such as purified ovalbumin and fl-lactoglob-
ulin with the purpose of determining whether they would behave chro-
matographicaUy in the manner observed by Tiselius ] in the displacement
development of mixtures of small, reversibly adsorbed molecules. Exper-
iments with 50 mg of ovalbumin on a 1.4-ml column of DE-52 showed that
the protein emerged as a plateau when followed by 0.1% CM-D (an un-
fractionated preparation having an RPV of 18) and as a peak when a 0.2%
24 R. W. Hartley, Jr., E. A. Peterson, and H. A. Sober, Biochemistry 1, 60 (1962).
[5] DISPLACEMENT CHROMATOGRAPHY 125

solution of the same CM-D was used. Twice as much of the 0.1% CM-D
was required as 0.2%, in agreement with the displacement requirement
for saturating the column. When the load was increased to I00 mg, and
followed by the 0.1% CM-D, the emerging plateau was twice as wide as
when 50 mg was used and its height was the same. The position of the rear
margin of both plateaus was the same; the extra load was added to the
front of the plateau. These observations 23 clearly establish displacement
as the mechanism operating in these experiments. However, attempts to
demonstrate protein-protein displacement with a mixture of 25 mg each
of ovalbumin and fl-lactoglobulin A, displaced by 0.1% CM-D having an
RPV of 20, failed. The proteins emerged with little resolution. A good
separation was achieved when the experiment was repeated with an un-
fractionated CM-D having an intermediate affinity (RPV = 15), intro-
duced before the final displacer to serve as a spacer between the proteins.
The two proteins emerged as separated peaks, not plateaus. The individ-
ual loads (25 mg) were too small to maintain a plateau against the margin-
mixing effects of passage through a bed of particles. However, the CM-D
used as a spacer was undoubtedly a more efficient displacing agent than
the larger and more rigid protein molecules.
Spacing. The potential for spacing chromatographed proteins is one of
the great attractions of displacement. This offers the possibility of insert-
ing sufficient nonprotein substance between two difficultly resolvable pro-
teins without excessively widening the peaks, as would be the case if a
longer column or a flatter gradient were used in elution chromatography.
This requires that the column be capable of distinguishing each protein
from the spacer between them, a requirement that is greater in terms of
resolving power than simply distinguishing one protein from the other.
However, any deficiency in such resolution will leave one or more of the
proteins contaminated with some of the spacer CM-D instead of with the
other protein. So long as enough spacer is used, the volume of effluent
separating the peaks will prevent their being mixed by the dispersive
mechanisms associated with passage through the adsorbent bed, and the
column can be made as long as is necessary to achieve the separation
without widening the bands.
Figure 3 illustrates the use of a narrow-range CM-D as a spacer in
separating an artificial mixture of the genetic variants of fl-lactoglobulin
on DE-52. 22 These variants, A and B, differ in their number of carboxyl
groups: A has 56 and B has 54 per molecule of 36,000 daltons, 25 and their
isoelectric points are about 0.1 pH unit apart. 26In panel A of Fig. 3,120 mg

25 K. A. Piez, E. W. Davis, J. E. Folk, and J. A. Gladner, J. Biol. Chem. 236, 2912 (1961).
26 j. Bours, Sci. Tools 20, 29 (1973).
126 CHROMATOGRAPHY [5]

1.5 A IB

ee
0 c~ o
o

< 0.5

! ,
0 10 20 30 40 0 10 20 30 40 50
FRACTION
FIG. 3. Spacing of fl-lactoglobulins A and B with different amounts of carboxymethyldex-
tran (CM-D) on DEAE-cellulose. In each experiment a 7-ml column of DE-52, equilibrated
with 20 mM sodium phosphate, pH 6, was used to chromatograph 20 mg of an equal mixture
of the A and B variants applied in 2 ml of the starting buffer. In panel A, 10 ml of a 1.2%
solution (120 mg) of narrow-range CM-D having an RPV of 11 was used as spacer; in panel
B, 20 ml of the same solution (240 mg) was used. In both cases, the final displacer was a 0.5%
solution of unfractionated CM-D with an RPV of 29. The flow rate was 10 ml/hr, and the
fraction volume was 2.5 ml. All solutions contained the pH 6 phosphate buffer. The heavy
line delineates absorbance at 280 nm, the thin line absorbance at 625 nm after reaction of 10-
/~1 samples with anthrone as a measure of carbohydrate, and the dashed line absorbance at
750 nm after reaction of 50-/xl samples in the Lowry assay. The circles represent RPV and
the arrows indicate the points at which the final displacer was started. From Peterson and
Torres. 22

of narrow-range CM-D having an RPV of 11 was used to space A and B

(10 mg of each) in displacement chromatography on a 7-ml column of DE-
52 initially equilibrated with 20 mM sodium phosphate, pH 6.0. The
spacer CM-D emerged between and within the protein peaks, as shown by
the thin solid line representing the anthrone reaction 27 and the open cir-
cles indicating RPV values of CM-D in the fractions. In panel B of Fig. 3
twice as much spacer was employed and the peak centers are almost
twice as far apart. Lowry determinations 2s of the protein content (dashed
line) indicated that the recovery was 98%. Gel electrophoresis identified
the first peak as the B form and the second as A. In both experiments, the
RPV value of the CM-D that emerged between the proteins was 11.

27 T. A. Scott, Jr. and E. H. Melvin, Anal. Chem. 25, 1656 (1953).

O. H. Lowry, N. J. Rosebrough, A. L. Farr, and R. J. Randall, J. Biol. Chem. 193, 265
(1951).
[5] DISPLACEMENT CHROMATOGRAPHY 127

The amount of spacer CM-D used in these experiments was substan-

tial, and a means to reduce the requirement would be desirable. One
approach involves increasing the molarity of the background buffer con-
centration. In one set of experiments, 22 raising the molarity of the buffer
from 20 to 40 mM sodium phosphate resulted in a 60% increase in the
separation of the peaks of A and B because the capacity of the adsorbent
for the more lightly adsorbed CM-D and protein was much more strongly
reduced than its capacity for the more tightly bound CM-D and protein,
which was hardly affected at all. The increase in buffer concentration is
advantageous in that it affords better control of the pH and provides a flux
of small ions that can be expected to speed equilibration, but since loss of
capacity diminishes the resolving power of the column, an optimum
buffer concentration must exist for each separation, depending on the
affinity of the protein to be purified. Careful adjustment of the buffer
concentration is likely to be a useful option.
An increase in the pH of the buffer affects the protein and the CM-Ds
quite differently. Although the charge on the CM-Ds is not affected by
such a change, the positive charge on most anion exchangers used for
protein chromatography is diminished and their capacity for adsorbing
any given polyanion of fixed charge is reduced. On the other hand, the
increase in net negative charge on the proteins is generally more effective
than the decrease in the positive charge on the adsorbent, so their affini-
ties usually increase. This has two effects: a rise in pH can significantly
lower the capacity of a column for binding CM-D without diminishing its
capacity for binding protein, and a specific CM-D will match a specific
protein in affinity only at a certain pH. As the pH rises, the affinity of that
protein will be matched by a CM-D with a somewhat higher RPV value.
The lowering of the capacity for CM-D also lowers the resolving power of
the column for CM-D, but if the spacers used are sufficiently homoge-
neous in affinity, this should not be a handicap.
Experiments with DEAE-Trisacryl M (LKB) provide a special exam-
ple of this in separations of the fl-lactoglobulin A and B in 20 mM sodium
phosphate at pH 6.0. It was found that only 25 mg of spacer CM-D (RPV
= 1l) was required for this column to provide one-third more spacing of A
and B than was obtained on DE-52 with 240 mg under the same condi-
tions, despite the fact that the former contained twice as many milliequiv-
alents of ion-exchange capacity as the latter. The explanation appears to
depend on two effects: there was an unusually high rise in pH (0.7 unit) as
a result of the collapse of electrostatic shielding of the chargeable groups
on the adsorbent as the CM-D entered the beads, and the effect of this pH
rise on the capacity for CM-D was unusually great because approximately
half of the chargeable basic groups in this adsorbent are very weak (pK' =
128 CHROMATOGRAPHY [5]

6.2). Thus, the diminished capacity for CM-D rendered the spacer much
more effective in moving the first peak ahead of the second. There was
too little spacer in the fractions to permit determination of RPV values by
our usual procedure.
On DE-52, CM-D is apparently held on many sites that are not accessi-
ble to proteins or not effective for binding them. On the other hand,
DEAE-BioGel A (DEAE-agarose made by Bio-Rad), with an ion-ex-
change capacity about one-tenth that of DE-52 for a given volume of
packed bed, has about 40% of the capacity of DE-52 for adsorbing/3-
lactoglobulin, but its capacity for adsorbing a given CM-D, relative to that
of DE-52, is much smaller. The ratio of the number of extra CM-D-
binding sites to the number of sites that can bind proteins as well as CM-D
must be much smaller in this adsorbent. It therefore offers a means of
reducing the amount of CM-D required for a given spacing of the sepa-
rated proteins. Experimental trial has confirmed that only about one-
fourth as much spacer is needed to separate A and B on this adsorbent as
is needed on DE-52, and the peaks are very sharp. 22 The same spacer
CM-D (RPV = 11) serves to separate these proteins at pH 6 on such
widely different adsorbents as DE-52, DEAE-Trisacryl M, DEAE-BioGel
A, and DEAE-Sephacel, although each of these adsorbents requires a
different salt concentration in gradient elution. However, only the /3-
lactoglobulin variants have been studied on all of these adsorbents, so a
similar conclusion regarding proteins in general cannot be drawn as yet.
Because of its low capacity for adsorbing CM-D, one cannot expect
DEAE-BioGel A to contribute adequately to the further resolution of the
spacer CM-Ds, so the spacers employed with this adsorbent must be
sufficiently homogeneous to do the job. For more difficult separations, a
small precolumn packed with a more effective adsorbent (about one-
fourth the size normally used) could presumably serve to fractionate fur-
ther the reduced amount of spacer that is needed. The DEAE-BioGel A
column would then function as a spacing enhancer as well as contribute to
the separation of the proteins from the CM-D.
Resolution. Experience has shown that columns packed with DE-52,
DEAE-Trisacryl M, or DEAE-Sephacel are capable of resolving CM-Ds
that differ very slightly in RPV, thus establishing positions representing
affinity values that will be recognized by proteins in the sample. The latter
two adsorbents are particularly effective. The resolution of the CM-Ds by
displacement on a given column may well be better than the column's
resolution of proteins by elution, for the CM-Ds are relatively small,
flexible chains that have access to or at least can utilize more adsorption
sites than the proteins. Moreover, in contrast to the situation in elution,
where migration depends on a progressive decrease in the capacity of the
[5] DISPLACEMENT CHROMATOGRAPHY 129

column for the adsorbed substances as the eluent increasingly dominates,

the capacity of the column in displacement chromatography remains the
same throughout the experiment. For any specific component, the same
number of sites are available per unit of volume to be competed for with
members of the train that are rapidly selected by the process to resemble
that component more and more with respect to affinity for those sites. The
CM-Ds should be able to compete with their equals and inferiors among
the proteins no matter what sites on the ion exchanger may be involved,
so long as the binding forces are ionic. The tailing that arises from convex
isotherms, and is only partially compensated for by the relatively flat
gradients required for high resolution in elution chromatography, is not a
problem in displacement.
In the isolation of minor protein components such as enzymes and
protein markers in disease, the column is sometimes deliberately over-
loaded with respect to proteins having lower affinities than the desired
protein in order to be able to apply a larger amount of the latter. Even
though the separation may be completed by elution chromatography, it is
displacement that is relied upon to achieve the fractionation that permits
the overloading to occur without loss of the desired protein. However,
CM-Ds are more efficient displacers, better results being obtained when
they are used. In unpublished work, the authors undertook to isolate a
minor protein from the serum of a person afflicted with psoriasis. This
protein was known only as a small spot appearing in two-dimensional
electrophoresis gels with an incidence correlated with the disease. The
first step in its purification was the displacement chromatography of 6 ml
of the dialyzed serum on a 7-ml column of DEAE-Sephacel at pH 7, using
narrow-range spacer CM-Ds with RPV values of 5.5 to 7.0. Fractions
were examined for the presence of the psoriasis-associated protein by
two-dimensional electrophoresis, since its position in such a gel was the
only criterion by which it could be recognized. The CM-D did not inter-
fere with electrophoresis in either dimension. 29 The protein emerged with
CM-D having an RPV of 6.5 at the very end of a highly compressed group
of peaks and shoulders representing most of the protein applied. The
fractions that contained the desired product were pooled (less than 12 ml),
and chromatographed on a similar column at pH 8, with spacer CM-Ds
ranging in RPV from 6.7 to 9. The psoriasis-associated protein emerged
early in a series of crowded peaks, the CM-D having an RPV of 7.0.
Almost pure, it was among those components least affected by the rise in
pH; most of the others required CM-D with an RPV of 8 or more. The
CM-Ds in the displacement train provided a continuous array of positions

29 A. R. Torres, G. G. Krueger, and E. A. Peterson, Clin. Chem. 28, 998 (1982).

130 CHROMATOGRAPHY [5]

1.2

0.8

LIJ
Z
30
Q0
a~
o ~° o
0.4 20

RPV

10

0 10 2O 30 40 50
FRACTION
FIG. 4. Fractionation of ovalbumin by displacement chromatography on DEAE-cellu-
lose. A solution of 50 mg of ovalbumin in 2 ml of 10 mM sodium phosphate, pH 7.5, was
applied to a 7-ml column of DE-52 equilibrated with the same buffer. Four narrow-range
carboxymethyldextrans (CM-Ds) were used as spacers: 4 ml of 1.9% CM-D (RPV = 7.1), 9
ml of 1.8% (7.7), 8 ml of 1.6% (8.3), and 5 ml of 1.4% (9.2), totaling 434 mg of CM-D. The
final displacer was a 0.5% solution of unfractionated CM-D with an RPV of 27. The heavy
line represents absorbance at 280 nm; the thin line, absorbance at 625 nm after reaction of
10-/xl samples with anthrone; and the circles, RPV. 27 From Peterson and Torres. 2z

along the abscissa of the chromatogram that represented affinities that

could be matched by the proteins in the sample. The larger the amount of
CM-D in a given affinity range, the wider was the spacing of these posi-
tions and the greater the amplification of the differences in affinity in-
duced in the proteins by the change in pH.
Another model protein mixture has provided an opportunity to dem-
onstrate the resolving power of the CM-D displacement system. Figure 4
illustrates the separation obtained when 50 mg of a highly purified com-
mercial ovalbumin preparation, claimed by the supplier (Sigma) to be 99%
ovalbumin on the basis of electrophoretic analysis, was chromatographed
on a 7-ml column of DE-52 in 10 mM sodium phosphate at pH 7.5, using
four narrow-range CM-Ds (RPV = 7-9, total 434 mg). 22 Three forms of
[5] DISPLACEMENT CHROMATOGRAPHY 131

ovalbumin were expected: A~, A2, and A3, containing two, one, and no
phosphate groups, respectively,3° with isoelectric points of neighboring
forms differing by about 0.1 pH unit) ~However, six definite peaks appear
in the A280profile, and each is represented by protein that migrates within
the range attributable to ovalbumin in the patterns obtained by electro-
phoresis in polyacrylamide gels (not shown). The appearance of dimers at
the end of the chromatogram, in fraction 40, is not unexpected, but the
presence of so many components in purified ovalbumin was not antici-
pated. However, Iwase et al. 32 have reported that differences in carbohy-
drate composition as well as phosphate content provide a basis for hetero-
geneity in gel electrophoresis of ovalbumin, though not in isoelectric
focusing. They concluded from the latter observation that the electropho-
retic heterogeneity contributed by the carbohydrate was not the result of
differences in charge. This suggests that a partial masking of the charges
on the protein by carbohydrate chains, leading to intermediate affinities,
may be the basis of some of the chromatographic heterogeneity seen in
Fig. 4. Vanecek and Regnier33 have obtained gradient elution profiles of
ovalbumin of similar complexity on an analytical scale, employing a com-
mercial HPLC ion-exchanger (see this volume [8]). However, relative to
the size of the column, the protein load was one-fiftieth of that used in
Fig. 4.
Removal o f CM-Ds from Isolated Proteins. Separation of the protein
from the CM-D that accompanies it in the fractions of a displacement
chromatogram is likely to occur whenever those fractions are subjected to
further purification by other procedures, e.g., electrophoresis, gel filtra-
tion (depending on the molecular weight of the protein), or chromatogra-
phy on other types of adsorbents. If the pH can be safely lowered to
render the protein cationic, passage of the fraction through a small
column of CM-cellulose will result in the adsorption of the protein while
the CM-D passes through unadsorbed. Sharp elution with a suitable salt
solution will then yield the protein in highly concentrated form. On the
other hand, one can use a column of DEAE-cellulose or similar anion
exchanger, at a pH that makes the protein cationic, to bind the CM-D and
allow the protein to pass through unadsorbed. This procedure promises to
be the basis of a convenient recycling of the spacer CM-D, a matter that is
likely to be of interest in large-scale separations.
The high-affinity CM-D that saturates the column after all the proteins

30 G. E. Perlmann, J. Gen. Physiol. 35, 711 (1952).

3~ M. B. Rhodes, P. R. Azari, and R. E. Feeney, J. Biol. Chem. 230, 399 (1958).
32 H. Iwase, Y. Kato, and K. Hotta, J. Biol. Chem. 256, 5638 (1981).
33 G. Vanecek and F. E. Regnier, Anal. Biochem. 109, 345 (1980).
132 CHROMATOGRAPHY [5]

have been driven from it is readily recovered by displacing it with a

dilute solution of alkali.
Scale of Operation. Displacement chromatography of proteins can be
carried out at any scale. Its classical advantages of high capacity and
concentrated bands are particularly attractive for preparative work, and
its inherent band-sharpening may be of special value in the multiple
columns used in very large-scale separations by utilizing entry into each
successive column to correct any skewing that might have occurred in
passage through the preceding one. Large-scale separations of proteins by
the procedures described in this chapter have not yet been attempted, but
there is reason to believe that fewer problems will be encountered in
scaling up displacement separations than in scaling up those based on
elution chromatography, since effective systems will be smaller. The use-
fulness of displacement on a repetitive, large-scale basis will, however,
depend on the refinement of procedures for recycling the spacers as well
as the final displacers. Proper recycling of the spacers can be expected to
improve their performance inasmuch as they should become more homo-
geneous with respect to affinity than the initial supply.
At the other end of the scale, in analytical applications such as high-
performance liquid chromatography (HPLC), displacement chromatogra-
phy offers the advantage of requiring very much smaller columns for a
given load and relatively simple apparatus. The potential for spacing the
peaks for maximum efficiency is also attractive. Although experience in
this area is meager, preliminary trials in collaboration with Dr. Vernon
Alvarez have been promising. Thus, a very good separation of the A and
B forms of fl-lactoglobulin was achieved on a 0.1-ml silica-based ion-
exchange column (Aquapore AX-300, sold as a microbore guard column
by Brownlee) in 35 min, applying solutions containing 0.4 mg of the pro-
tein, 3 mg of the CM-D spacer (RPV = 12), and 8 mg of the final displacer
(RPV = 30) at 5.5 ml/hr with a Pharmacia P-3 peristaltic pump. The entire
chromatogram encompassed only 3.2 ml, and the concentrations of the
two bands were high. After the separation the column was rapidly
stripped of CM-D with a salt solution before reequilibration with the
buffer for reuse.
Given the objectives of analytical HPLC, one is unlikely to use protein
samples or conditions that will produce plateaus instead of peaks. How-
ever, it must be kept in mind that in displacement the position of a peak is
not fixed, as in elution, by standardizing the conditions under which the
samples are run. Any substance in the sample with an affinity higher than
that of a given protein will add to the displacement of that protein, and a
very marked change in the distribution of low- and high-affinity compo-
nents of a sample might produce a significant shift in its position. How-
[6] HPLC: CAREOF COLUMNS 133

ever, by developing a standard CM-D train that contains many times as

much CM-D as the quantity of protein in the samples, one should be able
to make such shifts insignificant.
In the work of Francina et al. 16 mentioned above, microgram quanti-
ties of hemoglobin were applied to DE-81 paper strips for separation by
ampholytes. Presumably, thin-layer chromatography of other proteins on
other thin-layer media is feasible, since spacers can be easily applied.
However, procedures for detecting enzyme activity must take into ac-
count the fact that much of the enzyme in a displaced band is not immobi-
lized on the adsorbent and can readily be washed away by the reagent
solution. Also, the dyes that are used to stain proteins are bound by
positively charged adsorbents, and produced intensely colored back-
grounds.
Where the objective is the separation of isoenzymes to provide infor-
mation of clinical interest, the use of microcolumns from which individual
forms of the enzyme can be displaced by carefully selected narrow-range
spacers is more promising. The suppression of tailing and the control of
band width characteristic of displacement are of importance in such appli-
cations.

[6] H i g h - P e r f o r m a n c e L i q u i d C h r o m a t o g r a p h y :
Care of Columns
By C. TIMOTHY WEHR

The high-performance liquid chromatography (HPLC) columns now

used in protein isolation and characterization represent a considerable
investment by the user compared to the low-pressure supports they are
replacing. Careful use should allow these columns to give satisfactory
performance from several months to over a year. This review is meant to
provide information on the maintenance and troubleshooting of commer-
cial HPLC columns to enable the user to achieve maximum performance
and column lifetime. In addition to the specific recommendations on rou-
tine operation and column repair described below, a troubleshooting
guide for defective column performance is outlined in Table I. Since the
column is an integral part of a larger system, symptoms caused by pump,
injector, or detector conditions may be misdiagnosed as column failure
and are included in the table. Only the chromatographic modes commonly
used in protein or peptide separations (size exclusion, ion exchange, re-

Copyright © 1984by Academic Press, Inc.

METHODS IN ENZYMOLOOY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
[6] HPLC: CAREOF COLUMNS 133

ever, by developing a standard CM-D train that contains many times as

much CM-D as the quantity of protein in the samples, one should be able
to make such shifts insignificant.
In the work of Francina et al. 16 mentioned above, microgram quanti-
ties of hemoglobin were applied to DE-81 paper strips for separation by
ampholytes. Presumably, thin-layer chromatography of other proteins on
other thin-layer media is feasible, since spacers can be easily applied.
However, procedures for detecting enzyme activity must take into ac-
count the fact that much of the enzyme in a displaced band is not immobi-
lized on the adsorbent and can readily be washed away by the reagent
solution. Also, the dyes that are used to stain proteins are bound by
positively charged adsorbents, and produced intensely colored back-
grounds.
Where the objective is the separation of isoenzymes to provide infor-
mation of clinical interest, the use of microcolumns from which individual
forms of the enzyme can be displaced by carefully selected narrow-range
spacers is more promising. The suppression of tailing and the control of
band width characteristic of displacement are of importance in such appli-
cations.

[6] H i g h - P e r f o r m a n c e L i q u i d C h r o m a t o g r a p h y :
Care of Columns
By C. TIMOTHY WEHR

The high-performance liquid chromatography (HPLC) columns now

used in protein isolation and characterization represent a considerable
investment by the user compared to the low-pressure supports they are
replacing. Careful use should allow these columns to give satisfactory
performance from several months to over a year. This review is meant to
provide information on the maintenance and troubleshooting of commer-
cial HPLC columns to enable the user to achieve maximum performance
and column lifetime. In addition to the specific recommendations on rou-
tine operation and column repair described below, a troubleshooting
guide for defective column performance is outlined in Table I. Since the
column is an integral part of a larger system, symptoms caused by pump,
injector, or detector conditions may be misdiagnosed as column failure
and are included in the table. Only the chromatographic modes commonly
used in protein or peptide separations (size exclusion, ion exchange, re-

Copyright © 1984by Academic Press, Inc.

METHODS IN ENZYMOLOOY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
 134 CHROMATOGRAPHY [6l

versed phase) are considered; information on the use of adsorption or

normal-phase HPLC columns has been presented. 1-3

Column Installation
High-performance liquid chromatography columns consist of micro-
particulate (3- to 50-/xm particle diameter) materials based on silica or or-
ganic polymers (resins) most commonly packed at high pressure into
stainless steel tubing. Although most HPLC packings have the high me-
chanical strength needed to withstand the pressure drops of 150-5000 psi
encountered at typical flow rates, column beds will not tolerate undue
mechanical shock. For this reason, columns should be handled carefully
during installation and removal and should be stored so as to avoid shocks
that might cause disruption of the packed bed. The majority of commer-
cially available HPLC columns are "universal," that is, they may be
installed on any HPLC pump or injector either directly or with a simple
adaptor. However, when fitting a new column to an existing system, care
must be taken to ensure that the connections between injector and column
do not introduce dead volume. Inlet nut and ferrule must be mated with
the column terminator, unions and connectors should be zero- or low-
dead volume type, and connecting tubing should be of low internal diame-
ter (-0.3 mm). Connections should be sufficiently tight to prevent leaks
(approximately finger-tight plus a quarter-turn). If overtightening is re-
quired to prevent leakage, it suggests a mismatch between terminator and
inlet nut or ferrule that can be remedied only by replacement. Fitting
mismatch can result (in addition to leakage) in band broadening or car-
ryover due to excessive dead volume or unswept voids. Since there is to
date no standardization of terminators and fittings among manufacturers,
it is recommended that, in cases where columns from different sources
are used on the same HPLC system, separate transfer lines be fabricated
for each column type using fittings supplied by the column manufacturer.
Most commercial HPLC columns are shipped containing an organic
solvent designated in the column installation instructions. If this solvent is
not compatible with the mobile phase solvent, the column must be washed
with an intermediate solvent prior to use. For example, reversed-phase
columns shipped in methanol or acetonitrile should be washed with water
before introduction of buffers or salts. Columns shipped in hexane (e.g.,

L. R. Snyder and J. J. Kirkland, "Introduction to Modem Liquid Chromatography," 2nd

ed., pp. 782-823. Wiley, New York, 1979.
2F. M. Rabel, J. Chromatogr. Sci. 18, 394 (1980).
3D. J. Runser, "Maintaining and TroubleshootingHPLC Systems," pp. 69-90. Wiley, New
York, 1981.
[6] HPLC: CaRE Or COl.~r~Ns 135

nitrile or amino columns) must be washed with propanol or tetrahydro-

furan before introduction of methanol, acetonitrile, or aqueous solvents.
When a new column is installed, it should not be connected to the detec-
tor until several milliliters of wash solvent or mobile-phase solvent have
passed through the column to prevent residual packing solvent or silica
fines from entering the detector.
The trend in HPLC continues to be toward the use of high efficiency
columns packed with supports of 5-/zm particle diameter or less. These
columns may be of standard 25- or 30-cm lengths for difficult separations
requiring very high efficiency. Short columns 4-15 cm in length and
packed with 5-/zm or sub-5-/zm materials may be operated at normal flow
rates (-1.0 ml/min) to achieve moderate efficiency and low solvent con-
sumption or at high flow rates (2-5 ml/min) to obtain very short analysis
times. The user should be aware that as column efficiency increases and
column length decreases, the extra-column contributions to band broad-
ening will become more significant. Use of these columns demands that
extra-column dead volume be kept to a minimum by reducing the length
of transfer lines, using small inner diameter (i.d.) tubing, and installing
appropriate guard columns. With microbore HPLC columns of 2-mm i.d.
or less, extra-column effects are sufficiently severe that a conventional
liquid chromatograph would require extensive modifications, including
use of low-volume injectors, microbore transfer lines, and detectors with
micro flow cells (<5/xl) and rapid time constants (<100 msec).

Column Testing
A newly purchased HPLC column should always be tested upon re-
ceipt using the manufacturer's suggested test compounds and separation
conditions. Manufacturers generally test columns to minimize batch-to-
batch variations in stationary-phase characteristics and to verify that indi-
vidual columns meet performance specifications; test compounds are se-
lected that effectively probe such stationary-phase characteristics as bed
efficiency, selectivity, free silanol content, and contamination by trace
metals. Test compounds are usually small molecules selected for their
stability, low toxicity, and availability in the user's laboratory. In some
cases, column packings designed expressly for chromatography of pro-
teins may be batch-tested by the manufacturer with a protein test mixture
to check selectivity or recovery. The user should test a column upon
receipt to verify that the column meets the published specifications and
has not been damaged in shipment; in some cases, failure to test a column
within a given time period may void the manufacturer's warranty. Vari-
ances of up to 10% from the manufacturer's test results should not be
136 CHROMATOGRAPHY [61

considered indicative of a defective column since they may reflect extra-

column effects, different flow accuracy, or small differences in solvent
composition. In addition to confirming column integrity, testing a new
column provides a baseline for monitoring column performance during
use. In some cases, it may be advisable also to test a new column with
compounds more chemically related to the user's samples. Since interac-
tions of polypeptides with HPLC stationary phases are in some respects
different from those of small molecules,4 it is unlikely that a manufactur-
er's test results will adequately probe such column characteristics as
protein selectivity and recovery. A test sample made up of standard pro-
teins or peptides may be more relevant in establishing a performance
baseline and monitoring column aging.

Column Operation

Sample and Mobile-Phase Preparation

The most common cause of early column failure is lack of care in
solvent selection and preparation and inadequate cleanup of biological
samples. Solvents may contain dissolved impurities that can be retained
on the stationary phase and gradually alter the phase characteristics or
elute as spurious or "ghost" peaks. Undissolved particulates in the mo-
bile phase can plug system hydraulic components or the column head,
resulting in excessive operating pressure. The first step in avoiding these
problems is to use solvents of high purity (HPLC grade or spectroscopic
grade) and filter them prior to use through an appropriate micropore (0.22
or 0.45/zm) filter. Where pumps with ball-and-seat inlet check valves are
used, solvents must be thoroughly degassed prior to use to prevent loss of
pump prime by cavitation. Degassing is also advisable when using station-
ary phases susceptible to oxidation, e.g., amino phases, or when using
low UV detection (<--200 nm), where absorbance by dissolved oxygen can
give rise to noisy or drifting baselines. Degassing mobile-phase solvents
will minimize bubble formation due to outgassing in the detector cell,
although this can be prevented by creating 20-100 psi resistance on the
outlet side of the detector with a restrictor or with about 3 meters of 0.3
mm i.d. tubing.
Since peptides and protein fragments may not always contain chromo-
phores absorbing at the commonly used detection wavelengths (254 and
280 nm), low UV detection in the 205- to 220-nm range is often used. Many
of the mobile-phase modifiers (alcohols, organic acids) employed in pep-

4 G. Vanecek and F. E. Regnier, Anal. Biochem. 121, 156 (1982).

[6] HPLC: CAREOF COLUMNS 137

tide chromatography have UV cutoffs in this spectral range, and some

baseline offset in gradient elution must be expected. Solvent impurities,
however, become a severe problem at these wavelengths, giving rise to
noisy baselines or spurious peaks that may be incorrectly diagnosed as
column or pump disorders. Water is the most notorious offender, often
containing trace organic impurities that become concentrated on re-
versed-phase columns and appear as spurious peaks during gradient elu-
tion. This can even be a problem with commercial HPLC-grade water
whose optical purity is specified at longer wavelengths, e.g., 254 nm.
Trace impurities are best removed by passing the water or aqueous mobile
phase through an activated charcoal or reversed-phase column prior to
use. The same procedure can be used to remove UV-absorbing impurities
from mobile-phase additives such as buffer salts and ion pair agents.
Solvent impurities are also a problem in ion-exchange chromatogra-
phy employing phosphate gradient elution. Phosphate salts contain UV-
absorbing impurities that cause severe baseline offset at high detector
sensitivity. This problem is compounded by the tendency of these impuri-
ties to collect on the column and elute at high ionic strength; thus, the
baseline offset increases with column use. Solvent purification techniques
have been devised using recrystallization and ion-exchange cleanup 5 or
passage through a chelating resin, 6 but none of them completely elimi-
nates the difficulty. The best remedy is periodic stripping of adsorbed
impurities from the column (see Column Repair, below).
Where impurities in the weak (A) solvent in gradient elution can be
removed by passage through an appropriate adsorbent bed, a stripper
column packed with the adsorbent material can be placed in-line on the
outlet side of the " A " pump in a multipump gradient HPLC system.
However, the breakthrough volume of the stripper column must be accu-
rately determined to avoid elution of impurities during an analysis.
Where mixtures of an aqueous buffer and an organic solvent are used,
as in reversed-phase chromatography, the user must be careful to avoid
precipitation of buffer salts in the system. Premixed solvents should be
filtered through a micropore filter. If aqueous-organic mixtures are pro-
portioned from separate reservoirs by the HPLC pumping system, precip-
itation may occur in hydraulic components leading to reduced seal life or
component failure. If precipitation occurs in the column bed, column
failure is a certainty. The surest means for avoiding buffer-organic sol-
vent incompatibility is to use as the strong eluent a premixed combination
of buffer and organic modifier filtered prior to use. Neither the column nor

5 H. W. Shmukler, J. Chromatogr. Sci. 8, 581 (1970).

6 G. Karkas and G. Germerhausen, J. Chromatogr. 214, 267 (1981).
 138 CHROMATOGRAPHY [6]
the HPLC system should be stored while containing buffers, and when a
system is converted from one application to another, e.g., from ion-ex-
change to reversed-phase chromatography, the entire system should be
flushed with water before solvent and column are changed.
Column failure may arise from the gradual accumulation of particulate
material originating from wear of the system seals or microbial growth in
the mobile-phase solvents. This can be avoided by installation of an in-
line filter either before the injector or between injector and column; the
latter arrangement will protect against particulates originating from the
sample or from wear of the injector seal. In-line filters, available from
several manufacturers, have easily replaceable frits in the 0.5- to 2-tzm
pore size range and add only a minimal amount of dead volume to the
system. Microparticulate precolumns and guard columns also serve as
effective solvent filters.
A number of mobile-phase additives may cause column failure by
irreversibly changing the stationary-phase characteristics or by accelerat-
ing column degradation. Ion-pairing agents or detergents that have bulky
hydrophobic groups (e.g., camphorsulfonic acid, sodium dodecyl sulfate)
may irreversibly partition into the stationary phase of steric exclusion and
reversed-phase columns, changing the phase chemistry and reducing ap-
parent pore volume. In applications requiring the uses of these compo-
nents, it is recommended that the column be dedicated for this use. An-
ionic detergents cannot be used as mobile-phase components with
anion-exchange columns. Cationic alkylamines used as ion-pairing agents
or competing bases for silanol complexation (e.g., triethylamine, tetra-
methylammonium salts) tend to accelerate dissolution of the silica sup-
port in silica-based columns when used at neutral pH. It is recommended
that these agents not be used at a mobile-phase pH of 6 or greater and that
they be flushed from the column immediately afterward. The use of halide
salts with stainless steel HPLC pumping systems and columns has long
been a point of concern in protein purification, particularly since many
ion-exchange procedures developed on carbohydrate-based gels employ
chloride eluents as high as 1-2 M. Actually, the 316 stainless steel used in
most HPLC components is reasonably resistant to chloride at neutral pH,
and component lifetime should not be significantly reduced if such solu-
tions are flushed from the system after use. Chloride at acidic pH should
n o t be used in HPLC systems. If the investigator is uncertain about the
solvent compatibility of a particular instrument or component, the manu-
facturer should be consulted.
It is generally recommended that samples injected onto HPLC
columns be as free as possible of contaminating material to minimize
interferences in detection and to prevent unnecessary adsorption of sam-
16] HPLC: CAREOF COLUMNS 139
pie components on the column. This often is not possible when HPLC is
used for purification of polypeptides from biological fluids or extracts.
However, samples containing significant amounts of lipids should be ex-
tracted with a nonpolar solvent (ether, hexane) prior to injection. Size-
exclusion HPLC provides a rapid means of removing high-molecular-
weight material before injection onto ion-exchange or reversed-phase
columns. Samples should always be filtered to remove particulates, and
guard columns should always be used to prevent strongly retained compo-
nents from reaching the analytical column.
Chromatographic resolution will be affected by sample volume and
sample concentration. In steric exclusion chromatography, sample vol-
ume should be less than 1% of the column permeation volume (e.g., a
sample volume of about 100 ~1 for a standard 8-mm x 25-cm analytical
column), and sample loads of between 1 and 5 mg can typically be ap-
plied. In ion-exchange and reversed-phase chromatography using gradi-
ent elution, the sample must be applied in a weak solvent (typically the
initial or A solvent); here sample volume is not critical and several millili-
ters or more can be applied to an analytical column if necessary. Because
of the high capacity of ion-exchange and reversed-phase HPLC supports,
relatively large sample loads, 5-15 mg total protein or greater, can be
injected, and often analytical columns can be used for semipreparative
applications if gradient elution is employed. When such columns are oper-
ated isocratically, resolution will be more sensitive to sample volume,
sample concentration, and sample matrix effects, particularly for low k'
compounds.

Guard Columns
Guard columns are used to protect the analytical column from particu-
late material or strongly retained contaminants in the sample matrix; they
are by far the most effective means of preserving and extending the life of
the analytical column. Guard columns are placed between the injector and
analytical column and are discarded or repacked often to ensure that
contaminants are not eluted. The frequency of replacement will vary
depending on column capacity, solvent strength, sample type, and sample
load; typically, the life of a guard column is between 10 and 50 injections.
Selection of the appropriate guard column packing and column configura-
tion must take into account two considerations. First, the packing should
be similar in chemical structure to the analytical stationary phase so that
the selectivity of the system is not altered. This is not always possible,
since not all manufacturers supply guard column analogs of analytical
packings and, when they are available, phase characteristics of the guard
 140 CHROMATOGRAPHY [6]
material may not be identical to the analytical packing. If a low-capacity
guard material is used, selectivity mismatch is not as critical as with a
high-capacity material. In steric exclusion chromatography, mismatches
in guard column pore diameter and pore volume can affect protein elution
profiles. The second consideration in selecting a guard column is loss of
efficiency; ideally, this should not exceed 10% in order to preserve overall
resolution. This is an important point when short, high-efficiency analyti-
cal columns are used, since extra-column contributions to band broaden-
ing are more significant. Here, particular care should be taken if a high-
capacity microparticulate guard column is used. A high-capacity guard
column should have a height equivalent of a theoretical plate (HETP)
value similar to that of the analytical column. For example, coupling of a
3-cm guard column packed with 10-/~m material to a 15-cm, 5-/~m analyti-
cal column can result in a 30 to 40% loss in efficiency.
Two types of guard column packings are currently available: pellicular
materials and fully porous microparticulate packings. Pellicular packings
consist of solid-core beads (usually 37-40/zm in diameter) with the sta-
tionary phase bonded to or polymerized on the surface as a "pellicle."
Because of their large particle size, pellicular packings are easily packed
in dry form by the user without any special equipment; replacement cost
for material in a 4 × 50 mm guard column is about $4. Since the surface
area of a peUicular packing is small compared to a porous microparticle,
capacity is quite low. Thus, such columns must be repacked frequently,
but their effect on selectivity is not large. Two manufacturers (Scientific
Systems, Inc., State College, PA, and Upchurch Scientific, Bremerton,
WA) have introduced low-volume guard columns that couple directly to
the analytical column; these can be packed with pellicular materials and
used with 5-/xm and sub-5-~m analytical columns to achieve low effi-
ciency losses of 5% or less. The need for frequent replacement should be
offset by the ease of repacking.
Microparticulate guard columns are packed with the same material
used in analytical columns and are designed for situations in which high
efficiency is required. Because they must be slurry-packed, they are usu-
ally obtained from the manufacturer as prepacked columns or cartridges.
Brownlee Laboratories (Santa Clara, CA) offers prepacked disposable
guard cartridges that fit into a reusable holder, and they are available with
a variety of 5- and 10-/zm reversed-phase, ion-exchange, and steric-
exclusion packings. Since microparticulate guard columns are of high
capacity, a close selectivity match with the analytical column is impor-
tant, and dead volume in connectors and transfer lines must be minimized
to reduce extra-column band broadening. When an exact selectivity
match is not possible, a guard column support that is less retentive than
[6] HPLC: CAREOF COLUMNS 141

the analytical packing should be used. In some cases, saturated micropar-

ticulate guard columns can be regenerated off-line by washing with a
series of strong solvents.
Guard columns can also be used for the concentration or cleanup of
samples. In a typical application, the guard column replaces the sample
loop in a six-port rotary injection valve. The sample is introduced through
the injection port and concentrated on the guard column; poorly retained
contaminants may be eluted with a weak solvent, followed by valve rota-
tion and introduction of the sample onto the analytical column with a
stronger solvent.

Saturator Columns
A major limitation of silica-based HPLC supports is the instability of
silica under alkaline conditions. Ion-exchange and size-exclusion chroma-
tography of native proteins usually requires the use of mobile phases in a
pH range of 7 to 8, conditions that reduce the lifetime of silica columns.
Strong anion-exchange (SAX) columns with bonded quaternary ammo-
nium phases are particularly vulnerable to high pH. Column life can be
extended by saturating the mobile phase with dissolved silica to suppress
degradation of the analytical support. This can be done by adding silica to
the mobile phase during preparation or, more commonly, by installing an
in-line saturator or solvent-conditioning precolumn before the injector,
which is packed with porous silica. Large-particle (30-50/xm) silica is
generally used for ease of packing and to reduce system pressure. Al-
though it has been reported that use of a saturator may extend column life
more than 10-fold, 7 lifetime extensions of 2- to 5-fold are more typical
using a 4 x 300 mm precolumn. The precolumn should be maintained at
the same temperature as the analytical column and should be topped off
periodically to replace dissolved packing. Use of a saturator precolumn
carries several disadvantages. First, a silica-saturated mobile phase may
precipitate if allowed to stand in the system; the saturator column should
be removed and the HPLC flushed prior to shutdown. Second, dissolution
of the silica packing can generat~ fines that, if passed through the exit frit,
can plug the injector, transfer lines, or analytical column. It is recom-
mended that a small-pore (0.5-/.~m) frit be installed in the precolumn outlet
terminator. Third, voids created by dissolution of the precolumn packing
will increase the dead volume of the system, leading to poor reproducibil-
ity in gradient elution. The precolumn should be inspected and repacked
often to prevent this. Fourth, the precolumn can act as a trap, concentrat-
ing solvent impurities that later elute when a stronger solvent is intro-
7 j. G. Atwood, G. J. Schmidt, and W. Slavin, J. Chrornatogr. 171, 109 (1979).
142 CHROMATOGRAPHY [6]

duced. The analytical column should be removed initially during solvent

changeover to prevent its contamination. Finally, the presence of silicates
in isolated sample components may alter biological activity or interfere
with subsequent chemical characterization.

Gradient Elution
Gradient elution is employed in the vast majority of HPLC applica-
tions in protein isolation and characterization. Most commercial gradient
HPLC systems have sufficiently high solvent proportioning precision to
provide retention time reproducibility of a few percent or better in gradi-
ent elution. However, to achieve this level of performance, the user must
make sure that the column is reequilibrated at initial conditions at the
beginning of each analysis and that the gradient elution protocol including
column regeneration and equilibration is exactly repeated for each trial.
When solvents of widely different strengths are used, the column should
be regenerated with a reverse gradient to initial conditions, followed by
equilibration with 5-10 column volumes of the initial solvent. Regenera-
tion and equilibration can be done at elevated flow rates to reduce turn-
around time. In ion-exchange chromatography, column equilibration can
require up to 50 column volumes or more if initial and final solvents differ
in pH. It is not uncommon that ion-exchange columns, even when stored
in the initial solvent, will exhibit retention time variations in the first
analysis of the day and stabilize over successive runs. Operation of re-
versed-phase columns with ion-pairing agents or buffers will also require
extended equilibration periods. For maximum reproducibility, the con-
centration of the ion-pairing agent should be kept constant across the
solvent gradient.

Operation Limits
Column supports and stationary phases differ in their limitations to
pressure, flow rate, temperature, and pH. This information should be
supplied in installation instructions delivered with the column and should
be reviewed by the user to ensure that the column warranty is not voided.
Some column packings are incompatible with certain mobile phases--for
example, amine phases react with carbonyl groups; therefore, aldehydes
or ketones cannot be used as mobile-phase modifiers. In cases where the
user is unsure of incompatibility with an unusual solvent, the manufac-
turer should be contacted.
[6] HPLC: CAREOF COLUMNS 143

Column Storage
For short periods, e.g., overnight, it is best to store silica-based
columns with an organic solvent or mixtures of an organic solvent and
water. Columns should not be stored with buffer or salt solutions; if it is
necessary to leave buffer in the column for equilibration overnight, a low
flow of about O. 1 ml per minute should be maintained to prevent precipita-
tion within the column or within hydraulic components of the system. For
long-term storage, silica-based columns should be flushed with water if
they have been used with buffers or salts and then filled with an organic
solvent (methanol or acetonitrile). Since stationary phases can act as
substrates for microbial growth, aqueous solvents are not recommended
for storage. Column terminators should be tightly capped as drying may
cause changes in bed geometry, and the storage solvent should be indi-
cated on a label attached to the column. Columns based on polystyrene
resins or hydrophilic organic polymers should be stored with solvents
recommended by the manufacturers. Columns should be kept in a place in
which they will not be exposed to potential shock or extremes of temper-
ature.
It is good practice to maintain a log in which column use and perfor-
mance are detailed. This should include the type of column and its serial
number, along with initial test data; the history of its analytical use,
including sample type, mobile-phase conditions, and operator; test data
from periodic performance checks; details on column repair.

Troubleshooting the Column

The most common problems encountered in the operation of HPLC
columns are outlined in Table I, along with suggested diagnoses and treat-
ments. Several of these problems occur frequently in chromatography of
proteins and peptides, particularly in the reversed-phase mode.
Spurious or "ghost" peaks can arise from mobile-phase impurities,
from sample carryover in the injector or transfer lines, or from elution of
adsorbed sample components in subsequent runs. Mobile-phase impuri-
ties are a common source of ghost peaks in gradient elution when detec-
tion in the low UV is used, and they typically originate from the water.
This can be diagnosed by pumping water through the column for varying
time periods, followed by blank gradients. An increase in ghost peak
height as a function of the volume of water used is indicative of impure
water. Mobile-phase additives (ion-pair agents, buffers, salts) can be
checked individually in the same manner. Spurious peaks arising from
 144 CHROMATOGRAPHY [6]

TABLE I
TROUBLESHOOTINGHPLC COLUMNS

Circumstances Possible cause Remedy

A. Abnormal operating pressure

Rapid increase in pressure Plug in detector, column, Working back from detec-
injector, or chromato- tor to pump, break each
graph fitting to localize source
of resistance. If local-
ized to column, flush or
replace frit. Otherwise,
backflush or replace
plugged component
Gradual increase in pres- 1. Plugged inlet frit or 1. Replace frit; remove
sure column head top of column bed and
repack
2. Plugged outlet frit 2. Replace frit
Reduced pressure Leak or defective check Tighten fittings; with
valve volatile solvents, slow
leaks can be detected by
cold fitting; check pump
seals and check valves
Fluctuating pressure Defective hydraulic system Check inlet and outlet
check valves, pulse
damper, etc.
B. Decreasing retention time
Pressure increasing; all Flow rate too high Check flow volumetric
peaks including solvent rate; repair or reset
front affected flow
Efficiency also decreasing, Losing stationary phase 1. Guard column is con-
solvent front unaffected sumed
2. Replace inlet frit and
top off
3. Replace column
Column pressure decreas- Column temperature in- Control column tempera-
ing, solvent front unaf- creasing ture
fected
Retention time varies with Injection solvent decreas- Use starting mobile phase
injection volume ing column activity as injection solvent
Large sample Column overload Use smaller sample
Gradient elution Column recondition in- Use longer reconditioning
complete
Regeneration not repro- Use more care in recondi-
ducible tioning program
Gradient elution or pro- Increase in percentage Control solvent composi-
portioned solvents organic modifier tion, prepare new mix-
ture, check performance
of solvent proportioning
system
[6] H P L C : CARE OF COLUMNS 145

TABLE I (continued)

Circumstances Possible cause Remedy

Prolonged use Column contaminated 1. Flush with strong

solvent
2. Replace precolumn
3. Replace analytical
column
Spiking sample with au- Components in sample or I. Prepare standard in
thentic knowns gives injection solvent are same controlled matrix
different Tr than stan- altering chromatography 2. Remove matrix inter-
dards alone. (matrix effect) ference by sample
cleanup
3. Use internal standard
technique
Ion pair or ion exchange Concentration of ion pair Increase concentration of
chromatography reagent too low; buffer ion pair agent; check
pH or ionic strength buffer composition
incorrect
Retention time of some Sample changing with Dilute or concentrate
(not all) peaks changes concentration sample to minimize
with sample amount effects; change sample
solvent

C. Increasing retention time

Pressure lower than nor- Flow rate too low Check volumetric flow
mal, solvent front also rate; reset flow or repair
affected
Backpressure increasing, Column contaminated 1. Replace precolumn
flow rate OK, poor peak 2. Replace inlet filter
shape, solvent front
unaffected
Liquid found or cold Leak nr defective check Tighten fittings; slow leaks
fitting; flow rate lower valve of volatile solvents can
than expected be detected by cold
fitting. Inspect check
valve
Reversed-phase chroma- Decrease in concentration Control solvent composi-
tography of nonpolar solvent by tion
evaporation or incorrect
preparation
Gradient elution or iso- One solvent not flowing
cratic with proportioned owing to:
solvents 1. Empty reservoir 1. Fill reservoir
2. Plugged inlet frit 2. Clean frit
3. Air bubble in line 3. Purge line
4. Inlet valve plugged or 4. Consult manual
damaged
5. Programming error 5. Check program

(continued)
146 CHROMATOGRAPHY [6]

TABLE I (continued)

Circumstances Possible cause Remedy

Widely varying environ- Column temperature Control column tempera-

mental temperature change ture
Gradient elution Variations in column Use more care in repro-
regeneration program ducing reconditioning
program
Spiking sample with au- Components in sample or 1. Prepare standards in
thentic knowns gives injection solvent are same matrix
different retention time altering the chromatog- 2. Remove matrix inter-
than standards alone raphy (matrix effects) ference by sample
cleanup
3. Use internal standard
technique
Column pressure rising Column temperature de- Control column tempera-
creasing ture
Tubing starting to plug Working from detector to
pump, break each fitting
and determine if there is
excessive resistance to
flow anywhere but in
the column; if found,
investigate and repair
Ion-pair or ion-exchange Concentration of ion pair Reduce concentration of
chromatography reagent too high; buffer ion pair agent; change
pH or ionic strength buffer composition
incorrect

D. Decreasing resolution
Trend observed over Slow buildup of contami- See Part B above
several runs nants on column
Retention times un- Column efficiency de- Check column head for
changed creasing void; refill if necessary;
replace column or pre-
column
Gradient elution or pro- Efficiency unchanged, Pump malfunction; con-
portioned solvents retention or selectivity firm that each pump is
changed delivering correct
amount of solvent
Retention time shorter Flow too high Check volumetric flow
rate; reset flow or repair
Gradient elution Regeneration program Lengthen regeneration
inadequate
Changed or rechanged Change in solvent compo- Use more care in prepar-
solvent reservoir re- sition ing eluent
cently
Large sample load Column overload Dilute sample
[6] HPLC: CARE OF COLUMNS 147

TABLE I (continued)

Circumstances Possible cause Remedy

E. Peak tailing
All peaks affected Void at top of column 1. Fill in void with suit-
able packing or filler
2. Avoid high flow rates,
excessive pressures,
high mobile phase pH
3. Replace column
All peaks affected Poorly packed column Confirm by comparison
with good column.
Repair column or re-
place column
All peaks affected Frits contaminated or Replace fittings or frits
partially plugged
Most or all peaks affected Adsorbed contaminants on 1. Flush with strong
column giving mixed solvent
mechanism separations 2. Use precolumn; replace
as required
Reversed-phase chroma- Acidic or basic samples Add competing base, e.g,,
tography; only a few interacting with silanol 0.1% (CH3)4NC1
peaks affected groups
One peak affected Unresolved minor compo- Improve resolution: longer
nent column, different mobile
phase
Large sample load Column overloaded May be acceptable; de-
crease sample size
One peak affected Sample decomposing on 1. Decrease column tem-
column perature
2. Decompose sample to
stable product
3. Change mobile phase
4. Change column packing
Reversed-phase ion-pair Slow reaction kinetics 1. Use column with
chromatography monolayer coverage
2. Increase concentration
of ion pairing agent
Ion-exchange chromatog- Secondary adsorption 1. Increase column tem-
raphy effects perature
2. Add low concentration
of organic modifier such
as isopropyl alcohol to
solvent
Chromatography of surfac- Micelle formation Use more polar mobile
tants phase
All peaks affected, partic- Dead volume introduced in Remove dead volume,
ularly early eluting instrument inspect fittings, use
peaks 0.010-in. tubing, keep
lines short

(contmued)
148 CHROMATOGRAPHY [6]

TABLE I (continued)

Circumstances Possible cause Remedy

F. Fronting peaks
All peaks affected Poorly packed column Replace or repack column

G. Double or split peaks

All peaks affected Void or channel in column 1. Check for void at top
bed of column, fill if found
2. Replace column
All peaks affected Inlet frit plugged or con- Replace or clean frit
taminated

H. Ghost peaks
Reversed-phase gradient Impurities in water 1. Confirm by pumping
elution varying volumes of
aqueous solvent prior to
gradient
2. Use HPLC grade water
3. Clean up water by
passage over reversed-
phase support
Gradient or isocratic Sample carryover Flush injector with strong
elution solvent
Gradient or isocratic Elution of sample compo- 1. Repeat elution program
elution nents on column with blank injections
2. Flush column with
strong solvent

I. Truncated peak
Concentrated or large Detector overloaded I. Decrease sample size
sample 2. Use shorter path length
cell
3. Use different wave-
length

J. Negative peaks
Near Vp (size exclusion 1. Refractive index of 1. Use same solvent for
chromatography) or injection solvent differ- sample and mobile
solvent front (other ent than in mobile phase phase
modes) 2. Temperature fluctuation 2. Water jacket detector
associated with injection flow cell and column

K. Drifting baseline
1. Late eluting compo- I. Flush column with
nents strong solvent
2. Temperature fluctuation 2. Control ambient tem-
perature or detector
flow cell temperature
[6] H P L C : CARE OF COLUMNS 149

TABLE I (continued)

Circumstances Possible cause Remedy

3. Dissolved oxygen in 3. Degas mobile phase

mobile phase (low UV
detection)
4. Washout of previous 4. Purge entire system
solvent in pump or with fresh solvent
column
5. Residual immiscible 5. Wash system with
solvent in pump or commonly miscible
column solvent, e.g., propanol
6. Selective evaporation of 6. Cap solvent reservoirs
volatile mobile-phase
component
L. Baseline noise
1. Bubbles in detector cell 1. Degas solvents, install
restrictor on detector
outlet
2. Leakage of packing 2. Replace outlet termina-
material from column tor frit
3. Dirty or misaligned 3. Clean or realign cell
detector cell according to manufac-
turer's instructions
4. Failing detector lamp 4. Replace lamp
5. Defective pulse damper 5. Replace damper
6. Poor instrument 6. Provide adequate
grounding, faulty re- grounding; clean or
corder connections replace connections

unswept volume between injector and column can be diagnosed by thor-

oughly flushing (in the worst cases, backflushing) transfer lines and injec-
tor; improperly seated tubing and ferrules on injector loops or transfer
lines are often the cause. Occasionally, proteins that fail to elute quantita-
tively by the end of a gradient will appear as ghost peaks at the same
elution position in subsequent blank gradients, with peak height decreas-
ing progressively. In such instances the column can be washed between
runs by injecting a volume (up to several milliliters) of a strong solvent
between analyses.
Tailing peaks and split peaks occur frequently with biological materi-
als in HPLC. Peak tailing can arise from sample overload, a poorly
packed or degraded column, buildup of adsorbed contaminants on the
stationary phase, or extra-column effects. In reversed-phase chromatog-
raphy of polypeptides, tailing is often a sign of interaction between resid-
150 CHROMATOGRAPHY [61

ual silanol groups on the stationary phase and basic amino acid side
chains. The effect can be minimized by operating at acidic pH to suppress
silanol ionization, by adding a competing base such as triethylamine or a
tetramethylammonium salt to the mobile phase, and by using a column
with high surface coverage that has been end-capped (undergone second-
ary silanization to reduce the number of residual free silanols.) Split peaks
or doublets arise by the formation of multiple flow paths through the
column, usually by settling or degradation of the column bed to form
voids or channels. Peak doublets can also occur as sample contaminants
partially plug the inlet flit or column head, creating partial flow resistance.
The effect may be accompanied by a gradual rise in operating pressure
and can be cured by frit replacement or, as a last resort, removing and
repacking the top millimeter of the column bed.

Column Repair
Careful column operation will prolong column lifetime to a year or
longer. However, column performance will degrade as strongly adsorbed
sample components accumulate on the stationary phase and as a void is
formed by the gradual dissolution of the support. Often column life can be
extended for old or abused columns by simple operations such as washing
and backflushing, frit replacement, and bed repair.

Frit Cleaning and Replacement

To remove a column flit, tightly secure the column tube and carefully
remove the top terminator fitting, being careful not to disturb the column
bed. Replaceable frits may sometimes remain in the terminator body; they
can usually be dislodged with the tip of a spatula blade or by passing a 20-
or 24-gauge wire through the terminator inlet. Although replaceable stain-
less steel frits may sometimes be cleaned by immersing them in a sonic
bath containing 3-6 N nitric acid, it is easier simply to install a new flit.
When refitting the terminator, be sure the new flit is aligned and that
surfaces are free of packing material to permit proper seating. Where the
terminator ferrule has been distorted, it may be difficult to achieve a leak-
free connection with the new flit; installation of double flits will some-
times allow a tight seal to be formed.
Terminators with pressed-in flits can be cleaned by pumping a nitric
acid solution through the terminator. If this procedure is ineffective (as
indicated by continued high backpressure during flushing), the terminator
can be installed on an empty guard column and pumped in the reverse
direction to backflush the frit. The analytical column should be capped
[6] HPLC: CAREOF COLUMNS 151

with a spare terminator during frit cleaning to prevent drying or disruption

of the bed.

Column Backflushing
When a column exhibits excessive operating pressure due to partial
plugging of the inlet frit or column head, it is often possible to dislodge the
material by reversing the direction of flow and backflushing with the
mobile phase or a wash solvent; this is a simple remedy when a column
begins to show a gradual rise in pressure, and it avoids the hazards of
opening a terminator. However, several precautions should be observed.
First, the pressure rise must not be accompanied by a loss in efficiency.
Efficiency loss is indicative of voids in the bed, and backflushing may
disturb the bed so as to prevent repair by topping off. Second, not all
manufacturers recommend backflushing of columns; the user should refer
to the column installation instructions or contact the manufacturer. Third,
the column should be disconnected from the detector during backflushing
to prevent contaminants or particulates from entering the flow cell. Occa-
sionally, increasing pressure can arise from blockage of the outlet termi-
nator, particularly when fines are generated by degradation of silica sup-
ports during operation at high pH, high temperature, or with cationic
ion-pair agents. Under such conditions, backflushing will produce only a
partial or transient reduction in pressure, and the outlet terminator frit
should be replaced.

Repair of Voids and Bed Irregularities

Symptoms such as peak broadening, peak tailing, or split peaks sug-
gest that a void or channel may have been formed in the column bed. If
these are observed with a new column upon receipt or during the first few
injections of normal operation, it indicates a damaged or poorly packed
column, which should be returned. Voids that develop with extended use
can be repaired by topping off; original performance will probably not be
recovered, but the improvement in resolution is often acceptable, particu-
larly in gradient elution.
After removal of the inlet terminator, the column bed should be level
and flush with the face of the column tube. A light gray or brown discolor-
ation on the bed surface is normal after extended use, but dark discolor-
ation indicates a contaminated bed, which should be removed and re-
packed. Even small depressions in the bed can result in significant band
broadening and should be filled; large voids of a centimeter or more
probably cannot be repaired successfully. Before filling the void, the bed
152 CHROMATOGRAPHY [6]
should be leveled using a square blade; a small flat-tip laboratory spatula
trimmed to the column internal diameter works well. The void can be
filled with glass beads, a microparticulate packing, or a pellicular guard
column packing. Glass beads are not recommended because, unless they
have been completely silanized, they will adsorb proteins. If a microparti-
culate material is used, it should be prepared as a slurry in an organic
solvent and added dropwise to the column, allowing excess solvent to
permeate the bed. Since microparticulate packings are high capacity, the
stationary phase should be the same as that of the column to prevent
selectivity changes. In most cases, a pellicular packing will serve as the
best repair material since loss in efficiency due to a poorly repacked void
is not as significant with a low-capacity packing. Once the void is filled,
the bed surface should be leveled flush with the face of the tubing, excess
packing material removed from the tubing surface, and the terminator
replaced. After the repaired column has been tested, the terminator
should be removed and the bed checked for settling. If settling has oc-
curred, the top-off procedure should be repeated.

Column Washing
With extended use, the buildup of strongly retained material such as
lipids or basic and hydrophobic proteins will cause increased operating
pressure and altered chromatographic behavior of an HPLC column. Of-
ten column performance can be recovered by stripping adsorbed material
with one or a series of strong solvents. The key in rejuvenating a contami-
nated column lies in knowing the nature of the contaminants and an
appropriate strong solvent. Lipids can be removed by washing with non-
polar solvents such as methanol, acetonitrile, or tetrahydrofuran. There is
no solvent or series of solvents that will universally strip all adsorbed
proteins from HPLC stationary phases, but Table II lists several strong
eluents or solubilizing agents that have been used in specific instances for
stripping proteins from HPLC columns. Some columns, particularly those
based on organic polymer supports rather than silica, are not compatible
with these solvents, and the user should contact the manufacturer regard-
ing solvent compatibility. In most cases, neat organic solvents such as
acetonitrile or methanol are not strong eluents in protein chromatography
and, therefore, are not effective stripping solvents; 50:50 mixtures of
organic solvents with a buffer, organic acid, or ion-pairing agent serve as
better stripping agents for reversed-phase and size-exclusion columns. It
has been observed s that repeated gradients followed by retrogradients

8 F. E. Regnier, this series, Vol. 91, p. 165. See also this volume [8].
[6] HPLC: CARE OF COLUMNS 153

TABLE II
WASH SOLVENTS FOR HPLC COLUMNSa

Solvent Composition

Reversed-phase and size-exclusion columns

Acetic acid 1% in water
Trifluoroacetic acid 1% in water
0.1% Aqueous trifluoroacetic acid/propanol b 40:60
TEAP~/propanoP 40:60
Aqueous urea or guanidine 5-8 M
Aqueous sodium chloride, sodium phos- 0.5-1 M
phate, sodium sulfate
DMSO-water' 50 : 50
Ion-exchange columns
Acetic acid 1% in water
Phosphoric acid 1% in water
Aqueous sodium chloride, sodium phos- 1-2 M
phate, or sodium sulfate

Consult manufacturer for column compatibility.

b High viscosity solvent; pump at reduced flow rate to pre-
vent overpressure.
c Triethylamine-phosphoric acid; adjust 0.25 N phosphoric
acid to pH 2.5 with triethylamine.

between aqueous trifluoroacetic acid (TFA) and TFA-propanol can be

effective in regenerating contaminated HPLC columns. Although deter-
gents such as sodium dodecyl sulfate and Triton are good protein sol-
vents, they tend to be strongly retained on HPLC stationary phases and
may irreversibly change column characteristics; detergent cleanup should
be used as a last resort and be followed by extensive washing with water
and methanol.
Several precautions should be observed during column cleanup. If the
initial wash solvent is not compatible or miscible with the mobile phase,
the column should be flushed with an intermediate solvent. Similarly, sets
of wash solvents used in series must be compatible or, if not, interspersed
with a mutually compatible flushing solvent. For example, after washing
with high salt, urea, or guanidine, the column m u s t be purged with 5-10
volumes of water before the introduction of any organic solvent. Simi-
larly, if nonpolar solvents such as hexane or methylene chloride are used
to strip lipids from a reversed-phase column, a propanol purge is neces-
sary before the introduction of aqueous solvents. To avoid shocking the
column bed, it is advisable to introduce wash solvents with gradients over
5-10 column volumes. Viscous solvents such as dimethyl sulfoxide
154 CHROMATOGRAPHY [7]
(DMSO)-water, methanol-water, and propanol-water mixtures should
be pumped at reduced flow rates to prevent high pressures.
Successful regeneration of a contaminated column can be a time-con-
suming process. With microprocessor-controlled HPLC systems, multi-
step washing sequences can be programmed and run automatically over-
night. Alternatively, column regeneration can be carried out off-line with
an inexpensive low-pressure pump at reduced flow rates.

[7] H i g h - P e r f o r m a n c e Size-Exclusion C h r o m a t o g r a p h y
By KLAUS UNDER

Since 1980 the classical gel filtration technique employing soft and
semirigid organic gels for protein characterization and purification has
received progressively greater competition from high-performance size-
exclusion chromatography (HPSEC). The breakthrough of HPSEC is as-
sociated with the development of highly efficient buffer-compatible
columns operating at elevated back pressures. The columns are packed
with microparticulate organic-based or silica-based particles of tailor-
made graduated pore size and hydrophilic surface composition. High-
resolution separation of proteins on these columns is attained by adjusting
appropriate operating conditions. The proteins elute in the sequence of
decreasing molecular weight and size. In its simplest version, the reten-
tion of the proteins in HPSEC is considered to be a selective permeation
of biopolymeric solutes through the pores of the particles of the column
bed; smaller proteins should penetrate a larger portion of the pore volume
of the packing and hence will be retarded longer than larger proteins. Two
limiting cases arise: (1) proteins so large that they are excluded from the
pores will be eluted with a volume equal to the interparticle volume of the
column, V0; (2) small proteins that totally permeate the specific pore
volume of the packing, Vi, will elute with Ve = V0 + Vi. The total accessi-
ble volume is then equal to Vt = V0 + Vi. The intraparticle volume of the
column is termed Vi. Between these two limits, a linear relationship is
established between the logarithm of molecular weight (M) of the protein
and its elution volume, Ve. This retention mechanism is distinctly differ-
ent from other modes of column liquid chromatography, such as adsorp-
tion and ion exchange, and it entails several advantages. Proteins are
eluted quickly with relatively narrow bands. A predictable volume inter-

Copyright © 1984 by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All fights of reproduction in any form reserved.
ISBN 0-12-182004-1
154 CHROMATOGRAPHY [7]
(DMSO)-water, methanol-water, and propanol-water mixtures should
be pumped at reduced flow rates to prevent high pressures.
Successful regeneration of a contaminated column can be a time-con-
suming process. With microprocessor-controlled HPLC systems, multi-
step washing sequences can be programmed and run automatically over-
night. Alternatively, column regeneration can be carried out off-line with
an inexpensive low-pressure pump at reduced flow rates.

[7] H i g h - P e r f o r m a n c e Size-Exclusion C h r o m a t o g r a p h y
By KLAUS UNDER

Since 1980 the classical gel filtration technique employing soft and
semirigid organic gels for protein characterization and purification has
received progressively greater competition from high-performance size-
exclusion chromatography (HPSEC). The breakthrough of HPSEC is as-
sociated with the development of highly efficient buffer-compatible
columns operating at elevated back pressures. The columns are packed
with microparticulate organic-based or silica-based particles of tailor-
made graduated pore size and hydrophilic surface composition. High-
resolution separation of proteins on these columns is attained by adjusting
appropriate operating conditions. The proteins elute in the sequence of
decreasing molecular weight and size. In its simplest version, the reten-
tion of the proteins in HPSEC is considered to be a selective permeation
of biopolymeric solutes through the pores of the particles of the column
bed; smaller proteins should penetrate a larger portion of the pore volume
of the packing and hence will be retarded longer than larger proteins. Two
limiting cases arise: (1) proteins so large that they are excluded from the
pores will be eluted with a volume equal to the interparticle volume of the
column, V0; (2) small proteins that totally permeate the specific pore
volume of the packing, Vi, will elute with Ve = V0 + Vi. The total accessi-
ble volume is then equal to Vt = V0 + Vi. The intraparticle volume of the
column is termed Vi. Between these two limits, a linear relationship is
established between the logarithm of molecular weight (M) of the protein
and its elution volume, Ve. This retention mechanism is distinctly differ-
ent from other modes of column liquid chromatography, such as adsorp-
tion and ion exchange, and it entails several advantages. Proteins are
eluted quickly with relatively narrow bands. A predictable volume inter-

Copyright © 1984 by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All fights of reproduction in any form reserved.
ISBN 0-12-182004-1
[7] HPLC: SIZEEXCLUSION 155

val for the elution operates when V0 and Vt of the column are established.
Calibration of the column using the plot of log M against V~ for standard
proteins enables the molecular weight of an unknown solute to be esti-
mated by its elution volume. Apart from the selection of a proper column,
no intensive method development is required. The eluents used are
buffers of appropriate pH and ionic strength, and the necessary instru-
mentation is also simple. These features more than compensate for the
disadvantages involved, such as the limited peak capacity of HPSEC
(intrinsic to the method) and restrictions in the pH range of the eluent
when applying silica-based packings.

Fundamental Aspects
Although the area of application of HPSEC is very broad, two general
aims may be distinguished. In analytical separation and identification,
small sample sizes are injected and the primary objective is to gain opti-
mum resolution of components during a reasonable period of time. In
preparative separation, the purpose is to isolate large amounts per unit
time at a given purity and at reasonable cost.
Under analytical conditions, the elution volume of the protein on an
HPSEC column obeys Eq. (1), provided size exclusion is the only separa-
tion mechanism.
Ve = V0 + KsEcVi (1)
where KSEC denotes the distribution coefficient of the protein and ex-
presses its degree of permeation. KsEc is defined by Eq. (2), which equals
zero for a totally excluded protein, and unity for a totally permeating
solute. Appropriate markers for V0 of HPSEC columns are horse apoferri-
tin (467,000 daltons), j pig thyroglobulin ~ (670,000), bovine glutamic de-
hydrogenase I (998,000), and calf thymus DNA. 2 Dextran Blue of 2 × 106
daltons was found to be less suitable owing to its tailing peak. L2 As Vt
markers, sodium azide and DNP-alanine I are usually preferred. Other
solutes have also been recommended as Vt markers to check the extent of
non-size-exclusion effects on retention. 2,3
KsEc = (Ve - Vo)/(Vt - Vo) (2)
To construct a calibration curve, size-exclusion data are arranged by
plotting log M vs V~ or KSEC (Fig. I).
M. E. Himmel and P. G. Squire, Int. J. Pept. Protein Res. 17, 365 (1981).
2 E. Pfankoch, K. C. Lu, F. E. Regnier, and H. G. Barth, J. Chromatogr. Sci. 18, 430
0980).
3p. Roumeliotisand K. K. Unger, J. Ckromatogr. 218, 535 (1981).
156 CHROMATOGRAPHY [7]

105

104
o

103 • • __..tO.._

10 15 20 m! Ve
I ~ ~ , ' 015 ~ , , I
0 Io KSE c

FIG. 1. Calibration curve of native proteins on a TSK 3000 SW column (7.5 × 600 mm).
Eluent consists of 10 mM potassium phosphate at pH 7.0, 100 mM KCI, and 0.02% NAN3.
Data are calculated from Table I of Himmel and Squire ~ with permission of the publisher.

The interesting part of the calibration curve is the linear intermediate

portion, i.e., the working range of the column. Above the exclusion limit
all excluded proteins coelute with the same Ve. The same is valid at the
lower end for the totally permeating solutes. Typically, an HPSEC
column with a microparticulate organic-based or silica-based packing
elutes proteins of about 1.5-2 decades of molecular weight across the
linear range. This means that a total of two or three columns, each with an
appropriate calibration curve, is required to cover the whole molecular
weight range of about 4 decades from 1000 to 10 million.
A large number of theoretical models have been developed to relate
the molecular size and shape of polymeric solutes to the pore size of the
packing and to predict the experimental calibration curve with sufficient
accuracy. 4 As a rule of thumb, two silica-based HPSEC packings having a
mean pore diameter of 10 and 80 nm will span the whole molecular weight
range for protein separation.
Since the elution volume of protein is affected by both size and shape,
the calibration curve log M vs Ve is valid only for proteins of the same
shape, e.g., globular. This aspect has been thoroughly examined for na-
tive ~ and denatured 5 proteins on TSK-SW columns. Himmel and Squire 1
stated: "When using a calibration curve Fv vs M 1/3 and a well-calibrated

4 W. W. Yau, J. J. Kirkland, and D. D. Bly, "Modern Size Exclusion Chromatography,"

p. 27. Wiley, New York, 1979.
N. Ui, Anal. Biochem. 97, 65 (1979).
[7] HPLC: SIZE EXCLUSION 157

column, the estimated error in calculating molecular weight expressed as

standard deviation is 14%." Fv is given by Eq. (3).
Fv = (Veu3 - VoU3)/(Vt v3 - Vov3) (3)
Deviations from linearity (i.e., much larger errors) are often caused by
proteins being involved in non-size-exclusion interactions on HPSEC
columns. Ionic interactions between proteins and charged surface sites of
the packing result either in a decrease or increase of V~, depending on
whether repulsion or attraction forces operate. With eluents of high ionic
strength, hydrophobic interactions between the protein and hydrophobic
surface groups of the packing increase the elution volume. These second-
ary effects can be suppressed or minimized by appropriate manipulation
of the eluent composition. 6 On the other hand, they may be utilized in
some cases to generate a superior selectivity than is expected from pure
size exclusion alone. 3
The selectivity of an HPSEC column is expressed by the slope of the
linear part of the calibration curve D2; i.e., the smaller the slope, the
better is the discrimination between proteins of different molecular
weight. The primary interest in HPSEC is aimed at gaining resolution, Rs.
For two solutes of similar peak width and height, Rs is defined by Eq. (4)
Ve(2)- Ve0)
Rs = 2(oq + 002) Ve(2) > Ve0) (4)

where 001 and 0°2 are the standard deviation of peak 1 and peak 2, respec-
tively. Furthermore,
M = D i e -D2ve (5)
where D1 is the intercept of the calibration curve. Combining Eq. (4) and
Eq. (5) yields Eq. (6).
In (M2/M1) (6)
Rs = 2D2(001 + 0"2)
High resolution is achieved on a column with a small value ofD2(00~ + 002).
D2 depends on a proper choice of column packing material and is propor-
tional to the reciprocal of the column length, L. ty is proportional to the
square root of the column length. 7 Thus increase in column length in-
creases the resolution.
A special situation, compared to other modes of HPLC, is met in
HPSEC with regard to peak dispersion. Peak dispersion is expressed by
the standard deviation of the peak (00viin volume units), given by Eq. (7)
6 H. G. Barth, J. Chromatogr. Sci. 18, 409 (1980).
7 W. W. Yau, J. J. Kirkland, and D. D. Bly, "Modern Size Exclusion Chromatography,"
p. 102. Wiley, New York, 1979.
158 CHROMATOGRAPHY [7]
[7] HPLC: SIZEEXCLUSION 159

o'vi = G~i>/N~ ~/2 (7)

where N is the number of theoretical plates per column length. The aim is
to keep trv~ to a minimum or to generate a large plate number.
Totally permeating solutes exhibit the highest plate number per
column length, whereas N usually diminishes with decreasing elution
volume of the proteins at a constant flow rate of the eluent. The drop in N
with increasing molecular weight of the solute is due to the inverse rela-
tionship between the size of the molecule and the diffusion coefficient.
Low diffusivity, however, is associated with slow mass transfer and en-
hanced peak broadening. In order to compensate for this undesirable
effect, the flow rate must be reduced by a factor of about I00, compared to
normal HPLC of monomeric solutes. Since the resolution R~ is propor-
tional to N 1/2, [combining Eqs. (6) and (7) and assuming trvt = O'vz and
N~ = Nz], R~ is improved with reduced flow rate. This is shown in Fig. 2.
The relation between the elution volume (Ve), the elution time (tel and
the flow rate (fv) is given by Eq. (8).
v~ = ~ t e (8)
Since the elution volume remains constant, a reduction in flow rate causes
an increase in t~. This is exemplified by the data in Fig. 2. The required
resolution must be weighed against an acceptable analysis time.
Isolation of proteins by means of HPSEC involves an optimization
governed by the following quantities: resolution, sample throughput per
unit time, and purity of protein collected. When the required sample mass
and its purity are fixed, the problem reduces to how much sample size can
be tolerated in column load while maintaining the resolution attained
under analytical conditions. Assuming that the elution volume remains
unaffected by increasing sample size, a loss in resolution is expected to
occur only through a decrease in the plate number of the HPSEC column.
Defining a 20% decrease of the plate number as a loadability limit, the
mass load at constant injection volume was found to be 0.1 mg of protein
per gram of column packing material s for LiChrosorb Diol.
A TSK 3000 SW column of 600-mm length and 7.5-ram i.d. contains
about I0 g (for 21.5-mm i.d., about 80 g) of silica support. Mass loadability
then calculates to 1 mg (or 8 rag) under these conditions. Increasing the
s p. Roumeliotis and K. K. Unger, J. Chromatogr. 185, 445 (1979).

FIG. 2. Effect of eluent flow rate, fv, on the resolution of proteins. Elution with 0.1 M
sodium phosphate-0.1 M NaCl at pH 6.8 with a TSK 3000 SW column (7.5 × 600 mm) and
monitored at 206 nm. The chromatogram was kindly supplied by Dr. F61di, LKB Instrument
GmbH, Karlsruhe, FRG.
160 CHROMATOGRAPHY [7]

sample concentration also increases the viscosity. As an approximate

guide, the relative viscosity of the sample (compared to the eluent) should
be less than 2, corresponding to about 70 mg/ml for proteins in dilute
buffers. 9 Higher concentrations may also change the hydrodynamic vol-
ume of biopolymeric solutes and thus affect the elution volume. Another
means for circumventing mass overload is to inject a large volume of a
dilute sample. Again, however, a limit exists that is given by the volume
loadability of the column. This is the sample volume, at a given mass of
solute, that causes a 20% decrease of the column plate number. This value
was found to be about 1-2% of the column bed volume 8 ; for a column of
600 × 7.5 mm, the volume amounts to 270-540/xl. On the other hand, a
high dilution of the solute is unfavorable when isolation of the protein is
considered.
In practice, one can overload the HPSEC column to some extent and
accept a certain loss in resolution. Then the central portion of the peaks is
collected and the fractions are recycled for repetitive separation. This is
the preferred technique for the isolation of 10- to 100-mg quantities of
proteins from complex mixtures. Ultimately, however, this approach is
costly, because expensive columns are employed.
When analytical conditions no longer apply and real column overload
is apparent, the resolution equation defined above loses its original mean-
ing. The elution volume VE is then defined as the volume at the half-height
of the ascending part of the plug of solute emerging from the column. The
separation volume, Vs~p, is then given by VE(2) -- VE~)(VE~2)> VE~j)) (see
Pharmacia~°). Typical cases for this approach are group separation and
desalting processes, where a high resolution exists under analytical condi-
tions. Sample volumes that correspond to up to 30% of the column bed
volume are applied in such processes, l° Under these circumstances mi-
croparticulate packings no longer offer any advantage and large-bore
columns packed with coarse particles of about a 50-/zm mean diameter are
preferred. As a consequence of increasing particle diameter the column
back pressure is drastically reduced and much simpler instrumentation is
required. Typical packings for this purpose are the Fractogel TSK types
marketed by E. Merck, Darmstadt, FRG. 11

Equipment
HPSEC is carried out with conventional HPLC equipment composed
of the following basic components: solvent reservoir, high-pressure

9 H. G. Barth, J. Chromatogr. Sci. 18, 409 (1980).

~0Pharmacia Fine Chemicals, "Gel Filtration, Theory and Practice." Uppsala, Sweden.
l~ E. Merck, Fractogel TSK, Reagenzien Merck, Darmstadt, FRG.
[7] HPLC: SIZEEXCLUSION 161

pump, injection system, column, detector, recorder, and appropriate

data-handling systems) 2 A fraction collector connected to the detector
cell outlet is employed when isolation is considered. To check the purity
of fractions separated, gel electrophoresis is employed or, alternatively,
HPLC using a reversed-phase or ion-exchange column. The equipment
can be either modular (i.e., assembled of single components) or a com-
pletely integrated apparatus. Since the eluents are buffers, often of high
salt concentration, corrosion of stainless steel parts becomes noticeable
on long-term use. For that reason pumps have been developed with ce-
ramic-lined cylinders and sapphire pistons and valve seals, e.g., the LKB
2150 HPLC pump. Similarly, columns made of high-pressure-resistance
glass have been developed, e.g., those manufactured by Laboratorni Pris-
toge, Prague, Czechoslovakia. However, a complete apparatus of inert
equipment is not yet commercially available. Among other decisive de-
sign criteria controlling analytical accuracy, speed of analysis, and sepa-
ration versatility are the following: a precise and stable flow rate, precise
sampling, efficient columns with graduated exclusion limits and sufficient
lifetime, low dead volume connections, and good temperature control.
Commercially available pumps are of three types: reciprocating, posi-
tive displacement, and constant pressure. A detailed comparison of the
different types under the aspects of resettability, drift, short-term preci-
sion, accuracy, and cost has been presented.~2 There are two important
aspects to consider when employing pumps in HPSEC. As the working
volume of HPSEC columns is rather small compared to other HPLC
columns, a high precision of the flow rate produced is required. For
instance, a TSK 3000 SW column of 600 × 7.5 mm possesses a working
volume equal to the intraparticle volume of 13 ml. Within that volume,
proteins are eluted spanning a molecular weight of 40,000 to 400,000. This
means that the elution volume, measured with a precision of about --+1%,
generates a precision of about 10% in molecular weight.~ Furthermore, for
full utilization of the performance of HPSEC columns for high-resolution
separation of proteins, flow rates of the order of 10-100 ktl/min are neces-
sary for 6-mm bore columns. Most of the currently available HPLC
pumps produce flow rates in the range of 0.1-10 ml/min; i.e., columns are
operated under far from ideal conditions.
To maintain the efficiency of the column (see this volume [6]), samples
are injected by means of an injector as a sharp plug with a reproducibility
of better than +2%. Sample volume varies between 10 and a few hundred
microliters. Of the two injection systems, the syringe-septumless device

12 W. W. Yau, J. J. Kirkland, and D. D. Bly, "Modem Size Exclusion Chromatography,"

p. 123. Wiley, New York, 1979.
162 CHROMATOGRAPHY [7]
and the microsampling detector valve, the latter offers higher reproduc-
ibility and flexibility. By interchanging the loops, up to a few millimeters
can be introduced onto the column; automatic versions of the sampling
device are available.
Columns made of stainless steel and closed by appropriate low dead-
volume end fittings are usually 4-25 mm in bore and 300-600 mm in length.
These are packed with particles of a mean diameter of 5-15/zm, mostly 10
/zm, and narrow size distribution of about 10-20% standard deviation
from the mean. Packings are either organic- or silica-based; in both cases
the pore size and size distribution are the most significant packing proper-
ties. These are controlled in the case of organic gels by the degree of
cross-linking and in the case of porous silica, by the method of synthesis
and the nature of the aftertreatment. To become suitable as HPSEC pack-
ings for proteins, the surface of these particles must be hydrophilic, free
of charged surface sites, and chemically stable toward the buffered
eluents. Hydroxyl, ether, and amide functional groups anchored by a
hydrocarbon spacer to the surface of the matrix are preferred. The exact
bulk and surface composition is not indicated in most of the commercial
packings. Detailed examinations have been carried out on diol and amide-
modified silica packings. ~3,~4The problems arising in the modification of
silicas lie in attaining a chemically stable and dense surface layer that
essentially protects the residual silanols.
Silanol groups cause adsorption and denaturation of proteins and act
as weak cation-exchange sites at pH > 7, giving rise to ionic interactions.
Owing to the chemical stability of the S i - - O - - S i - - C link and to the
solubility of silica itself, the packings are employed in a pH range of 2 to
8.5. The major aspect in the synthesis of organic-based packings is to
overcome the swelling property by appropriate cross-linking. The organic
packings also may carry functional groups that undergo protonation or
deprotonation, e.g., amino and carboxyl groups. Thus organic-based
types of HPSEC packings are applicable in the wider pH range of 2
through 13, although stability problems remain. For instance, the TSK-
type PW packings, being a semirigid hydrophilic cross-linked gel contain-
ing - - C H 2 C H O H C H 2 0 - - groups as main constituent, do not tolerate salt
concentrations higher than 0.5 M. 15
As most proteins absorb in the UV, photometers, preferably of vari-
able wavelength, are usually used as detectors. The latest developments

z3 D. P. Herman and L. R. Field, J. Chromatogr. Sci. 19, 470 (1981).

t4 H. Engelhardt and D. Mathes, J. Chromatogr. 185, 305 (1979).
~5C. T. Wehr, T. V. Alfredson, and L. Tallman, "Varian, Liquid Chromatography," No.
LC-134. Varian Associates, Walnut Creek, CA, 1982.
[7] HPLC: SIZE EXCLUSION 163

in this field are the diode array spectrophotometers that scan the whole
spectrum (190 to 700 nm) of a solute emerging from the column in 10 nsec.
Flow fluorometers are also applied in HPSEC, whereby UV-activated
solutes emit a fraction of absorbed light as fluorescence. In this way a
noise-level concentration of l 0 - l l g/ml can be achieved under appropriate
conditions. Solutes that do not fluoresce can be converted into fluorescent
components by means of an appropriate postcolumn reaction although
difficulties arise in ensuring a sufficient conversion (adjustment of the
residence time of reactants) and in avoiding large dead v o l u m e s . 16A7
The principle of one-line postcolumn reaction has also been used in
developing a direct or coupled enzyme assay using photometric detec-
tors.~8

Column Selection and Maintenance

The choice of an appropriate HPSEC column and its operation under
optimum conditions is the key to a successful separation of proteins (also
see this volume [6]). The table lists the commercially available organic-
based and silica-based packings and columns. The majority of HPSEC
separations are carried out on the TSK columns of Toyo Soda. Thus most
of the work on column characterization and use will refer to this type.
The type of column chosen should exhibit a fractionation range that
covers the molecular weight range of the proteins to be resolved. This
column property is independent of the particle size of the packing and the
column length and is solely a function of the pore size distribution for
silica-based packings or the degree of cross-linking for organic-based
packings.
The most frequent range in which size-exclusion separations of pro-
teins are carried out covers 10,000 to 500,000 daltons. This is almost
achieved by a single column: the TSK 3000 SW has a fractionation range
of 40,000 to 400,000. The mean pore diameter of this packing is estimated
to be about 30 nm. With materials of large pore size, e.g., 100-nm pore
diameter, the exclusion limit is extended to above 1 million daltons. Small
proteins of <40,000 daltons require packings of about 5-rim mean pore
diameter. Special effects arise in the application of small and large pore
size packings, associated with their structure. Since the mean pore diame-
ter of a rigid porous packing such as silica is inversely proportional to the
specific surface area, 5-nm pore size packings offer high surface areas of

t6 R. W. Frei and A. H. M. T. Scholten, J. Chromatogr. Sci. 17, 152 (1979).

1~j. F. K. Huber, K. M. Jonker, and H. Poppe, Anal. Chem. 52, 2 (1980).
is T. D. Schlabach and F. E. Regnier, J. Chromatogr. 158, 349 (1978).
 164 CHROMATOGRAPHY [7]

(,-q ¢,q

I I I I I I I I I I I I I

X X X X X X X X X X X X X X X X X'~ X

,d
0

Z
<
+1 +1 +l +1 +l +l +i +1 +1 +l +l
~ Z Z Z Z Z
Z

<

~.~
~2R~ ~ i i I i ~ Z Z Z Z Z Z ~zzzzz¢
M
<

<
<> "0 '~

<

0
(D
~.~ ~ o= ~ o

I1)

o~
[7] HPLC: SIZE EXCLUSION 165

l l l l l l l l I
, o .>,
S~ S~
ZZ ZZ

r.d =.
°~

+1 +1 +1 +1 +1 +1

~ z ~ ~ ~

•8 8 '="~
o=~ ,-, ,,,

£
o N

_~..n 5 1 0

[,0
Z ~ ~o

o,o
0='~, z~
166 CHROMATOGRAPHY [7]
the order of 400 to 500 m2/g, possibly slightly less after modification. This
gives rise to adsorption effects in protein separation by HPSEC that can
be decreased to some extent by manipulation of the composition of the
eluent. For large-pore silicas of 100-nm pore size, problems of mechanical
stability of the particles bring about limitation in the pressure and flow
rate used in the slurry packing method.
Commercial columns differ in their working volume at a given frac-
tionation range, i.e., differences in the intraparticle volume of the
columns. For comparison, Vi is related to the Vo of the column. The ratio
Vi/Vo, the "phase ratio," varies between 0.6 and 1.70, the latter being the
highest value attainable for silica packings. 1.2 As the intraparticle volume
determines both separation capacity and column selectivity, these can be
improved by coupling two separate columns of the same type or by apply-
ing a column of 600-mm length, instead of 300 mm. It will be evident that
by doubling the column length the slope of the calibration curve D2 (in
milliliters of Ve per decade of molecular weight) becomes smaller by a
factor of 2 while the range of separation remains the same; resolution is
improved as expected from Eq. (6).
To characterize tb.e proteins by molecular weight, a calibration curve
of log M vs Ve (Fig. 2) is needed. It is advisable to gauge the calibration
plot of the column employed by measuring the elution volume of standard
proteins at the given eluent composition, rather than relying on literature
data. This procedure should be repeated from time to time to monitor
possible changes in the column or packing structure. Although the com-
mercial HPSEC columns show high recoveries for native proteins, 2 the
data are measured for a given eluent composition and may change under
new conditions.
When required, the fractionation range can be extended by coupling
columns of different and descending pore sizes. The whole separation
range is spanned with three columns of 5-, 30-, and 100-nm pore size. Yau
et al. 19 have shown that the individual calibration curves should be adja-
cent but not overlapping, in order to obtain the broadest range and
maximum linearity of the calibration curve. The intraparticle volume of
each column should be equal, since sigmoidal calibration curves would be
obtained otherwise.
Since the pressure Ap across the column is proportional to column
length, Ap increases when columns are assembled. Longer columns also
enhance the analysis compared to short columns. The relationship be-

t9 W. W. Yau, J. J. Kirkland, and D. D. Bly, " M o d e m Size Exclusion Chromatography," p.

267. Wiley, New York, 1979.
[7] HPLC: SIZE EXCLUSION 167

tween these parameters is given by Eq. (9), where ~ is the viscosity of the
eluent, ~b' is the flow resistance factor of the column (about 500-1000), dp
is the mean particle diameter of the packing, and tm is the elution time of a
totally permeating solute.
Ap = [(rldp')/tm](L/dp) 2 (9)
To keep an HPSEC column functioning properly for a long time, a
series of precautions have to be considered (see this volume [6]). Primary
attention has to be paid to the eluent composition. Prepacked HPSEC
columns of TSK types are filled with a 0.05% sodium azide in aqueous
solution on delivery. Solvent replacement with an appropriate buffer is
carried out at a low flow rate, e.g., 0.5 ml/min. Column lifetime is limited
by both the chemical resistance of the packing and the stability of stain-
less steel column toward corrosive reagents. The pH range available for
the TSK PW type of packings is much wider than that of the TSK SW
type owing to the enhanced solubility of silica above pH 8.0 and the
sensitivity of the S i - - O - - S i - - C link in alkaline media. High salt concen-
trations are often applied to suppress ionic interactions between the sup-
port and the proteins. To avoid a rise in viscosity and precipitation of salt
due to temperature changes, the concentration should be kept below
0.5 M. Column life is drastically reduced when proteins are separated in
their denatured form by applying sodium dodecyl sulfate (SDS), highly
concentrated guanidinium hydrochloride, or urea. When changing to such
denaturing conditions, it should be noted that the elution volume of the
proteins on the same column becomes shorter owing to the unfolded state
of the protein. In principle, HPSEC columns can be operated above room
temperature to reduce the viscosity of the eluent, to increase the plate
count by enhanced diffusivity of proteins, and to prevent possible adsorp-
tion.
For TSK PW and TSK SW columns, a maximum flow rate of 1.2 ml/
min (analytical column) and of 8.0 ml/min (preparative column) is tolera-
ble. In routine work it is advisable to run the column overnight at a low
flow rate of about 0.5 ml/min. On longer storage the buffer must be re-
placed by rinsing the column with dilute buffer and finally conditioning
with an aqueous solution of 0.05% sodium azide.
Column life is also dependent on the purity of the samples that are
introduced. Solutions to be injected should be free of fines and colloidal
constituents. Samples must be dissolved in the eluent. Degrading of the
column is prevented to some extent by inserting a short guard column
between the injector and the column itself. Guard columns are packed
with the same material; since their capacity is limited, they must be re-
placed from time to time.
168 CHROMATOGRAPHY [7l

Operating Conditions

Having decided on the type and dimension of column, an eluent is

chosen that is composed of a buffer, a salt, and, if needed, a stabilizer.
Typical buffers of pH 5-8 are phosphate, Tris acetate, citrate, and ace-
tate. The ionic strength is adjusted by adding sodium chloride, sodium
sulfate, ammonium acetate, and ammonium formate. The addition of salts
suppresses ionic interactions between the charged surface sites of the
HPSEC packing and the protein. Depending on the pH and the isoelectric
point of the protein, the elution volume increases or decreases with in-
creasing ionic strength. 2°,2~ The charge and type of the ions added also
affect the hydration of the protein according to the chaotropic series,
which may lead to changes in the elution volume. 3
For separation of denatured proteins, 0.1% aqueous SDS, 6 M guani-
dine hydrochloride, or 6 M urea solutions are applied as eluent. SDS is
preferred for purification of membrane proteins, which do not dissolve in
common buffers. Kato e t al. 22-24 established the calibration plots of TSK-
gel SW columns for a series of proteins under denaturing conditions. In
eluents containing SDS, the sodium phosphate concentration was shown
to have a significant effect on the slope of the calibration curve.
Optimum performance of HPLC column, i.e., minimum peak disper-
sion and highest plate count N, is known to be attained at values of the
reduced linear velocity (v) between 1 and 10 for monomeric solutesY
Equation (10) defines v, where
v = (ttdp)/Dim (10)
u is the linear velocity of the eluent, d o is the mean particle diameter of
packing, and D~m is the diffusion coefficient of solute in the eluent. With
dp = l0/zm and Dim --- I X 10 -11 m2/sec for a protein 26the linear velocity of
the eluent calculates to 0.0001 cm/sec (v = l) and 0.001 cm/sec (v -- 10).
This corresponds to a volume flow rate (fv) of 2-20 p,l/min for a 7.5-mm
i.d. column. As most HPSEC separations are carried out at flow rates of
about 0.5-1.0 ml/min for a 7.5-mm i.d. column, these columns are oper-
ated far from their optimum. It must be said, however, that such low flow

20 D. E. Schmidt, R. W. Giese, D. Conron, and B. L. Karger, Anal. Chem. 52, 177 (1980).
21 p. Roumeliotis, K. K. Unger, J. Kinkel, G. Brunner, R. Wieser, and G. Tschank, in
"High Pressure Liquid Chromatography in Protein and Peptide Chemistry" (F. Lott-
speich, A. Henschen, and K. Hupe, eds.), p. 71. de Gruyter, Berlin, 1981.
22 y . Kato, K. Komiya, H. Sasaki, and T. Hashimoto, J. Chromatogr. 193, 29 (1980).
23 y. Kato, K. Komiya, H. Sasaki, and T. Hashimoto, J. Chromatogr. 193, 458 (1980).
~4 y. Kato, K. Komiya, H. Sasaki, and T. Hashimoto, J. Chromatogr. 190, 297 (1980).
25 p. A. Bristow and J. H. Knox, Chromatographia 10, 279 (1977).
26 j. M. Schurr, CRC Crit. Reo. Biochem. 4, 371 (1977).
[7] HPLC: SIZE EXCLUSION 169

rates would produce long analysis times; assuming an elution volume of

15 ml for a protein on a 600 × 7.5 mm HPSEC column atfv = 20/A/rain,
the analysis time calculates [Eq. (8)] to 12.5 hr. In order to complete an
HPSEC chromatogram in a more reasonable time (a few hours), the flow
rate is increased to about 100 gl/min. This has to be paid for with a loss in
plate count and loss in resolution. On a TSK-gel SW 3000, 600 × 7.5 mm
column, the following plate counts were achieved for human albumin in a
phosphate buffer of pH 6.8:3300 (1 ml/min), 4100 (0.8 ml/min), 4900 (0.4
ml/min), 8500 (0.2 ml/min), 11,100 (0.1 ml/min) (data kindly supplied by
LKB Instruments GmbH, Munich, FRG). Roumeliotis and Unger 8 re-
ported values for N of 4000 and 9000 for chymotrypsinogen A on LiChro-
sorb Diol columns of 250 × 6 mm and 250 × 23.5 ram, dp 5/~m, at flow
rates of 2 and 21 ml/min. When plotting the plate number of proteins on a
given column against the flow rate, N is seen to decrease linearly with
increasing fv.z The decrease is more pronounced for proteins of high
molecular weight.
Since a gain in resolution is the primary objective, the dependence of
Rs on the molecular weight of proteins at a given column is of interest.
Data are available on the TSK-gel SW type of columns. 24 The resolution
in this case was calculated by
R~p = (Ve~2) - V~z))/[2(o'1 + o'~)(log M1/M2)] (11)
Eq. (11), which yields values up to 7 for a pair of proteins on G 2000 SW
and G 3000 SW columns but drops to 3 on a G 4000 SW column. For each
column, the function Rsp of proteins passes through a characteristic maxi-
mum on plotting against M. The value of M at Rsp(max) corresponds to the
value in the middle of the linear part of the calibration plot. In conclusion,
proteins that differ in molecular weight by a factor of 2 may be resolved.

Applications
HPSEC columns are applied in four areas of protein characterization
and purification: prefractionation, analytical separation, molecular weight
determination, and preparative isolation. For prefractionation, collected
fractions are further separated on ion-exchange or reversed-phase
columns. Analytical separations are carried out, for example, in the moni-
toring of the time course of enzymic reactions. Molecular weight determi-
nation of proteins by means of a calibration curve requires carefully puri-
fied standards and precise measurement of Ve or te. The most frequent use
of HPSEC is the isolation of milligram quantities of proteins for further
examination. Examples of such use are available. 27
27 C. T. Wehr, R. L. Cunico, G. S. Ott, and V. G. Shore, Anal. Biochem. 125, 386 (1982).
170 CHROMATOGRAPHY [8]

[8] H i g h - P e r f o r m a n c e I o n - E x c h a n g e C h r o m a t o g r a p h y
By FRED E. REGNIER

Liquid chromatography of proteins is a separation process based on

differential rates of molecular migration through a bed of particles. When
purifying a single substance by this process, the objective is to choose
conditions and materials that maximize the difference between the migra-
tion of this substance and all others in the sample. The difference in
migration between the substance being purified and that of any other
component in the sample depends on a number of variables, including the
resolving power of the system. Resolution (R0 of a particular component
from any other is expressed by Eq. (1)
R~ = 2(Ve2 - Vel)/(AVe I + A V e 2) (1)
in which Vel and Ve2 refer to the elution volumes of the first and second
components, respectively, to elute from the column. Peak width for each
of these components is designated, with the appropriate subscript, as
AVe. The discussion here will deal with the rationale used in column
selection, operation of ion-exchange columns, and maximization of reso-
lution through manipulation of Ve and AVe values.
Obtaining successful ion-exchange separations of proteins from a se-
ries of different samples has two requirements: a good column and an
understanding of the ion-exchange process. The first requirement is gen-
erally satisfied by the purchase of a commercial column, but the second
requires an intellectual commitment from the investigator. Actually,
many individuals fail to see the need for the second requirement since
they view chromatography as a procedure that can be accomplished by a
recipe. Although much can be done by recipe, that approach is as restric-
tive as an attempt to paint a landscape in monochrome. Each protein is
unique and, therefore, offers unique opportunities for its purification. The
chromatographer must be aware of the means for finding and exploiting
these opportunities.

Retention
Protein retention on an ion-exchange column is obviously the result of
electrostatic interactions in which retention increases in proportion to the
charge density on both the ion-exchange matrix and the protein. Since
proteins are amphoteric, their net charge and charge density will vary

Copyright © 1984 by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
[8] HPLC: ION EXCHANGE 171

with the pH of the solution. Two additional variables that may influence
retention are the nature of the ion-exchange support and the distribution
of charge in the protein. Some ion-exchange materials have broad titra-
tion curves that will cause ion-exchange ligand density to vary with the
solution pH. The ionic properties of the ion-exchange material and of
proteins may vary independently with pH. That all the charges on the
surface of a protein may not be able to interact with the ion-exchange
matrix simultaneously is an important consideration. That is, two proteins
of identical charge can interact differently with an ion-exchange support
because of differences in the distribution of charges rather than the net
charge alone.
In order to control the retention process in ion-exchange chromatogra-
phy and to manipulate elution volumes of the various protein compo-
nents, it is necessary to examine the ionic properties of proteins. Proteins
are amphoteric species that bear a net positive charge under acidic condi-
tions and a net negative charge under basic conditions and are isoelectric
at a specific intermediate pH designated as the isoelectric point (pI).
Titration curves indicate that the amount of charge on proteins increases
steadily as the pH of the enveloping solution deviates from the pI and that
each protein has a unique titration curve. This implies that a pH could
probably be found at which the difference in charge between the protein
being purified and all other proteins will be at a maximum. That pH would
obviously be the one at which the protein should be chromatographed on
an ion-exchange column. Two factors complicate this rationale: all the
charged groups within a protein will not be able to interact simultaneously
with the ion-exchange matrix, and charge distribution on the surface of all
proteins may not be uniform. Ion-exchange characteristics are based only
on those groups of the protein that are capable of interacting with the ion-
exchange matrix. Ion-exchange surfaces can recognize the difference be-
tween two proteins that are identical in every respect except that one has
uniform charge distribution and the other is asymmetric. (Significantly,
electrophoretic systems will not distinguish between these two species
because both have the same net charge and pI.) In contrast, two mole-
cules having a different pI and net charge could possibly cochromato-
graph if they have the same surface characteristics. From these examples,
it can be seen that any comparison between electrophoresis and ion-
exchange chromatography has the inherent flaw that the two techniques
are effective by different mechanisms using charge differences as the
basis for separation.
The difference in separation mechanism between ion-exchange and
electrophoretic techniques is most prominent at the pI where the net
charge of a protein is zero and where it does not migrate in an electric
172 CHROMATOGRAPHY [8]
- Lactoglobulin
35 I i I i I ~ I
l
I
I
I
30 l ao

25
/
T
I
//
I 20
i--
o

1'5 s,x
~z
IN
10 I0 0 Z
I-
bJ
I-
UJ

I--
U
0 0 .J
2 4 6 8 I0 W
pH
FIG. I. Chromatographic retention as compared to titration and electrophoretic mobility
curves for fl-lactoglobulin, pl = 5.1. SAX refers to the Pharmacia Mono Q column; SCX, to
the Pharmacia Mono S column. A 40-min linear gradient to 1.0 M NaC1 was used at eluent
pH 3.0 on the SCX column.

field. The absence of electrophoretic migration does not mean that the
protein is devoid of charge. There will still be cationic and anionic groups
within the molecule at its pl, and if their distribution is sufficiently asym-
metric, i.e., if some of the charges are clustered, they will cause ionic
interactions with an ionic surface. Thus, it is not surprising that 75% of
the proteins examined by Kopaciewicz ~were retained on both anion- and
cation-exchange columns at their pI.
The term "ion-exchange retention," as used here, refers to either the
retention time or the elution volume of a protein from a column that has
been eluted with a gradient. Two major modes of gradient elution are used
in protein separations: ionic strength gradients, and pH gradients. Ionic
strength gradients are by far the most common and the easiest to gener-
ate. Chromatofocusing provides a form of pH gradient elution and will be
discussed separately. The reason for separating most proteins by gradient
elution will also be discussed. To examine the difference between electro-
phoretic and ion-exchange separations, Kopaciewicz ~ compared the

] W. Kopaciewicz, M. A. Rounds, J. Fausnaugh, and F. E. Regnier, J. Chrornatogr. 2,66, 3

(1983).
[8] HPLC: ION EXCHANGE 173

charge, electrophoretic mobility, and chromatographic retention of/3-

lactoglobulin at a number of pH values. The results, shown in Fig. 1,
indicate that retention on both strong cation- and anion-exchange
columns occurs at the pI at which net charge and electrophoretic mobility
are zero. This would indicate that/3-1actoglobulin represents a protein in
which the charge distribution is asymmetric. The figure also shows that
chromatographic retention begins to plateau several pH units above the p!
of/3-1actoglobulin whereas its charge and electrophoretic mobility con-
tinue to increase. Further analysis of the titration and retention data indi-
cates that the protein had a total of 17 negatively charged residues at pH
10 but that only 6 were involved in the ion-exchange process.
Plots of chromatographic retention versus pH are referred to as r e t e n -
tion m a p s ~ (see also this volume [11]). Comparison of the retention maps
(Fig. 2) of several proteins indicates that many proteins are retained on
both strong anion- and cation-exchange columns near their pI; that reten-
tion on both types of columns increases as mobile-phase pH is moved
away from the pl; that each protein has a unique retention map, just as it
has a unique titration curve; and that there is an optimum pH that pro-
duces the largest difference in retention of a specific component from
others in a mixture.
Questions relating to properties of the ion-exchange support and oper-
ating parameters to resolution are discussed separately.
Macromolecules have been separated chromatographically on the ba-
sis of charge, hydrophobicity, size, and bioaffinity. Obviously, charge
characteristics are the most important in ion-exchange separations, but
hydrophobicity and molecular size also play a role. For example, mole-
cules that are of limited water solubility may interact with an ionic matrix
by a hydrophobic mechanism in addition to coulombic forces. In extreme
cases, a protein may be so hydrophobic that it lacks water solubility.
Separation of proteins of limited water solubility, such as membrane pro-
teins, will also be discussed (see also this volume [16] and [18]). The
contributions of molecular size to ion-exchange chromatography, in con-
trast to hydrophobic effects, are usually evident in the capacity of an ion-
exchange medium. The pore diameter of a support must be matched to the
molecular size to obtain maximum ion-exchange capacity. The role of
these variables should become clearer from the discussion that follows.

Column Selection
The most important question confronting the neophyte chromatog-
rapher is that of the nature of the column to be used. For a rational
choice, the chromatographer should have some knowledge of the protein
 174 CHROMATOGRAPHY [8]

• I I l ! I I ! I [ I I I I

20 ANION EXCHANGE ANION EXCHANGE

pI<7 pI:>7
16

12

._~
E ~L

20 CATION EXCHANGE -- GE
pI<:7
16

12

| I I I I I I l

4 5 6 7 8 9 1 0 3 4 5 6 7 8 9 1 0
pH pH

Flo. 2. Retention maps of some common proteins. Five acidic and five basic proteins
were chromatographed on both a strong anion (Mono Q) and strong cation (Mono S) ex-
change column. The symbols represent proteins within the vertical column. A suitable
buffering ion (10 mM) was chosen for each pH. Proteins were eluted with a 20-rain linear
gradient from 0 to I = 0.5 M NaCI at a flow rate of I rnl/min. Anion exchange: ©, lysozyme;
A, cytochrome c; [3, ribonuclease; O, chymotrypsin; A, carbonic anhydrase; cation ex-
change: (D, /3-1actoglobulin; A, soybean trypsin inhibitor; G, ovalbumin; A, a-amylase;
0 , conalbumin.

that is to be separated in terms of its pI, size, and even hydrophobicity.

The retention maps of proteins in Fig. 2 indicate that greatest retention on
cation-exchange columns occurs below the isoelectric point, where the
protein has a net positive charge, whereas the largest retention on anion-
exchange columns occurs above the isoelectric point, where proteins
are negatively charged. For example, a protein with a pl of 7 would be
retained on a cation-exchange column below pH 7 and on an anion-
[8] HPLC: ION EXCHANGE 175

exchange column above pH 7. Less than 5% of the proteins examined are

an exception to this statement. 1 In the event that the chromatographer
does not know the pl of the protein he is trying to separate, the informa-
tion may be obtained either chromatographically or electrophoretically. A
simple technique for approximating pI is to chromatograph the protein on
both anion- and cation-exchange columns over a range of 3 pH units at
near-physiological conditions. Retention properties of the protein may be
mapped and the most suitable column selected by this simple process. A
superior method for determining pI is the isoelectric focusing titration
curve of Rhigetti. 2 In addition to giving the isoelectric points of all pro-
teins in a sample, information on the charge characteristics of all sample
proteins at any pH from 3 to 10 is also provided. Although there is not a
direct correlation between the charge characteristics as observed in elec-
trophoresis and chromatographic retention, electrophoresis data suggest
the most useful pH at which to begin ion-exchange separations. A number
of instances have been observed in which electrophoresis was used to
predict optimum pH for ion-exchange chromatography.3 Once it has been
established that either an anion- or cation-exchange column is the more
likely to provide the desired separation, the experimenter may proceed
with column selection.
Four basic types of high-performance ion-exchange chromatography
(HPIEC) columns are available: (I) weak anion-exchange (WAX), (2)
strong anion-exchange (SAX), (3) weak cation-exchange (WCX), and (4)
strong cation-exchange (SCX) materials. Unfortunately, the terms
"strong" and "weak" are often misunderstood in reference to ion-ex-
change chromatography. They are not intended to suggest the strength
with which a substance is retained on the ion-exchange matrix, as the
name would imply. Rather, they indicate the degree of stationary phase
ionization at various extremes of pH. Strong cation- and anion-exchange
materials are usually sulfonates and quaternary amines, respectively;
both remain fully ionized within the normal operating range of high-per-
formance ion-exchange chromatography. In contrast, weak cation- and
anion-exchange supports usually contain carboxyl groups or primary and
secondary amines, respectively, as the ionizable species. The pKa of the
average weak cation-exchange support is roughly 4, and that of a weak
anion-exchange material is in the range of 8 to 10. Some ion-exchange
materials even have very broad titration curves that span 5 pH units. 4

~"P. G. Righetti and J. W. Drysdale, in "Isoelectric Focusing," p. 341. North-Holland

Publ., Amsterdam, 1976.
Technical Bulletin, "The Pharmacia HPLC System." Pharmacia Fine Chemicals AB,
Uppsala, Sweden.
4 A. J. AIpert and F. E. Regnier, J. Chromatogr. 185, 375 (1978).
176 CHROMATOGRAPHY [8]

Obviously, as the operating pH of a column approaches within a few pH

units of the pK~, charge on the ion exchange matrix decreases along with
ligand density. Since the charge density of both the support and the pro-
tein vary with the pH of the mobile phase in an unrelated manner, reten-
tion characteristics of proteins on weak ion-exchange columns are less
predictable than with strong ion exchangers. There are instances in which
the retention of a protein will be greatest at an intermediate pH and will
diminish toward the pH extremes as ionization of the WAX or WCX
support collapses. As a general rule, it is best to carry out the initial
separations of a protein on a strong ion-exchange material, although this
is not meant to suggest that the strong ion-exchange support will neces-
sarily be superior to a weak ion-exchange column in resolution. The sub-
ject of resolution is discussed below.
A number of ion-exchange columns are now available (Table I). There
are two broad classes of ion-exchange supports in terms of support matri-
ces: those totally organic and those that are inorganic with an organic
surface coating. The inorganic supports are available in a series of pore
diameters ranging from 100 to 4000 ~ and particle sizes from 3 to 30/zm.
The only microparticulate organic-based support currently available for
proteins is the Pharmacia MonoBead material; it is available in a strong
cation- and anion-exchange form in addition to a weak anion-exchange
material for chromatofocusing (see also this volume [11]). Support pore
diameter in the MonoBeads is reported 3 to be approximately 800/~, with
the matrix chemically stable between the limits of pH 2 and 12. This is
considerably broader than the pH 2 to 8 limit of silica-based materials.
Present silica supports have an upper pH limit of 8 or 9 because of the
increased erosion of the silica matrix at basic pH.
An important question centers about the pH to be used in protein
separations. Chromatographic band broadening, the tendency of enzymes
to denature, and the amount of retained material at the inlet of ion-ex-
change columns increase substantially when columns are operated more
than 2 pH units from physiological pH. This does not mean that ion-
exchange separations are limited to the limited pH 5 to 9 range, but rather
that most separations can be achieved within this range.
Pressure limits and expected column life are generally not available
from manufacturers (see also this volume [6]). Since most separations
may be achieved at less than 30 atmospheres with properly designed
columns, mechanical stability is probably not a problem with any of the
current semirigid and rigid packing materials. Expected column life, in
contrast, is more difficult to quantitate since it depends on operating
conditions, type of sample applied, and column care. There is now evi-
dence that many columns will survive the analysis of 500 or more serum
samples at pH 7 if routinely cleaned when resolution begins to fail.
[8] HPLC: ION EXCHANGE 177

× ××x~
<<<<
r/l

<<
oc5- ~ z z ~ A A A A o
o
e.
0
<<< <<<<
g ZZZ ZZZZ

ZZ ~
u~

la.
la0
O
r~ < "~
O o

r.~

ZZZZ
e~ ~~~~

m
Z
< ©
0

u~
0

0
d o

z .~
178 CHROMATOGRAPHY [8]

Loading Capacity
An additional parameter in column selection is that of the amount of
protein that must be separated. In current practice, a 4.6 x 250 mm
column is considered to be one for analytical use. Most ion-exchange
columns of this size will have a loading capacity ranging up to 10 mg of
protein without loss of resolution. As sample loading is increased, resolu-
tion decreases steadily up to loads of 40 or 50 mg, at which point protein
often saturates most of the length of the column and normally retained
proteins begin to break through with other nonretained material. The
relative degree of separation between the proteins of interest determines
to a large extent whether such a large sample may be accommodated.
Loads of 50 mg may give very acceptable separation of proteins that
separated widely in an analytical sample, whereas a 15-mg load may
produce unacceptable resolution between peaks that are immediately ad-
jacent. The loading capacity of any support is due to a combination of
variables that includes the ratio of support pore diameter to solute diame-
ter, support surface area, and ligand density. The loading capacities pre-
sented here are in terms of what might be expected with a protein (Mr =
50,000) on a 300-/~ pore diameter support. Proteins of higher molecular
weight would be expected to exhibit lower loading because they cannot
penetrate the support matrix. Smaller proteins might load more heavily
because they have greater access to the support surface inside pores.
At submicrogram loadings, sample recovery may decline because the
surface area-to-solute mass ratio in the column is very large; the column is
undefloaded. The small number of imperfections in a support that irre-
versibly adsorb or denature protein can dominate the separation in an
underloaded column. It is better to use a smaller column for micropre-
parative work. Submicrogram samples also begin to exceed the detection
limits of the system unless such sensitive indicators as radioactivity, en-
zyme activity, or fluorescence are being monitored.
Columns of 50-ram length or less are becoming common in analytical
separations because column length contributes little in the resolution
of proteins. The loading capacity of a 4.2 × 50 mm column is in the
range of 1-2 mg of protein. The principal advantages of short columns are
that they are less expensive, mechanical stress on the packing bed is
reduced, they may be eluted more quickly, they are easier to pack, and
solutes may be eluted in smaller volumes of mobile phase thereby de-
creasing sample dilution and increasing detector sensitivity. When the
intent is to separate very large quantities of protein, the chromatographer
must go to the true preparative-scale column. It is to be expected that the
loading capacity of ion-exchange columns will increase with the square of
[8] HPLC: ion EXCrtANGE 179

the column radius. Thus, by increasing column dimensions from 4 mm

(analytical dimensions) to 10 mm (preparative dimensions), loading capac-
ity is expected to increase sixfold. The design and application of prepara-
tive ion-exchange columns to protein separations remains in an early
stage, but there is sufficient knowledge available to begin a discussion of
the subject.
Two types of ion-exchange users are anticipated for the future: those
scaling up an analytioal procedure to produce a few grams or less of
protein, and the industrial user who intends to produce tens of grams to
kilograms of protein per hour. Their needs and their approaches will be
completely different. In the first case, simply scaling up the analytical
separation by using the same packing material and a two- to fourfold
larger column would probably provide the best solution. The need for
redesigning a separation system and elution protocol is thereby avoided
for occasional use. The cost per unit mass of protein isolated may be
much more expensive by this approach than with the large, preparative
system required in the commercial production of proteins, but it requires
much less development.
With the advent of genetic engineering, the ability to prepare kilogram
quantities of protein is in demand. The preparative systems used in the
purification of these proteins will be designed for specific separations.
Even the support materials may be optimized for loading capacity and the
resolution of a single protein species.

Elution
A major initial concern of a chromatographer attempting a separation
on a new type of column is "how to get a peak." Since ion-exchange
chromatography of proteins has been in use for the past 25 years and is
quite similar to HPIEC, a large body of information is available on ion-
exchange elution protocols. As a first suggestion, it is recommended that
the chromatographer draw upon his own experience or search the litera-
ture for separations similar to the one he is undertaking. Conditions simi-
lar to those reported with conventional gel-type ion-exchange columns
usually work well on HPIEC columns. The most widely used technique
for eluting ion-exchange columns over the last 25 years has been with
ionic-strength gradients at a fixed pH. Although pH gradients have been
successful, they are more difficult to produce and generally give poorer
resolution than ionic-strength gradients. The possible exception would be
in chromatofocusing columns.
The initial step in gradient elution of increasing the ionic strength of an
ion-exchange column is to load the sample on the column in a buffer at a
180 CHROMATOGRAPHY 18]

low ionic strength and at a pH appropriate for retention to occur. Buffer

concentrations of I0-20 m M are generally used, although up to 50 mM
buffer is appropriate when proteins are strongly retained. Desorption and
elution of proteins is achieved by gradually increasing the ionic strength
of the mobile phase. Proteins are generally displaced from the column in
the order of their increasing charge, although the actual number of
charges in contact with the surface are more important in determining
retention than is net charge.
The next consideration is the selection of a mobile phase consisting of
a buffer and displacing ions. Since the ionic strength of the buffer is low,
the pKa of the buffer chosen should be within 1 pH unit of the operating
pH of the system. Manufacturer literature supplied with the column
should also be consulted, since specific buffers are occasionally recom-
mended.
If there is no information on which to base a selection of the displacing
ion or salt, it is best to start with simple salts such as sodium chloride,
sodium acetate, or, possibly, sodium bromide. Sodium chloride has been
so widely successful in conventional gel-type columns that its use is rec-
ommended in HPIEC as well. However, it should be recognized that
halide ions erode stainless steel surfaces of pumping systems unless the
metal is passivated. The problems of metal erosion and passivation tech-
niques have been discussed in this series. 5 Selection and use of a variety
of other salts will be discussed below in the section dealing with optimiza-
tion. The strong or displacing solvent (solvent B) usually consists of 0.5-
1.0 M displacing ion added to the initial buffer (solvent A). Seldom will a
protein be so strongly adsorbed to a column that more than 1.0 M salt is
required for elution; when a case of strong retention is encountered,
bringing the operating pH of the column closer to the pI of the protein will
diminish retention. In the event that all protein is eluted from the column
by the time the gradient has reached 50% solvent B, it would be desirable
to dilute solvent B until the most strongly retained proteins elute near 90%
B in gradient elution. Although step gradients are commonly used in
conventional gel-type columns and, on occasion, in HPIEC columns,
complex mixtures are usually best resolved by a continuous gradient.
Analytical columns may be eluted at 1 ml/min with a linear gradient
ranging from 0 to 100% B in 20-30 min. A rationale for using different
mobile-phase velocities and gradient slopes will be presented in the dis-
cussion of optimization.

5 This series, Vol. 91, p. 137.

[8] HPLC: ION EXCHANGE 181

A note of caution should be added with regard to particulate matter in

buffers, since both the column frits and the column bed have the proper-
ties of a filtration medium. Any particulate matter in the mobile phases
will accumulate in the column and eventually plug it. Proper care and
preparation is discussed in detail in this volume [6].

Chromatofocusing
As noted, charged proteins may be adsorbed to an ion-exchange sup-
port and eluted with either a pH or salt gradient. Gradients are normally
generated externally and fed into the ion-exchange column to effect elu-
tion. Sluyterman and his co-workers 6-8 have developed a new technique
for generating a pH gradient: chromatofocusing. A pH gradient is formed
by pumping a buffer with a large number of different charged species into
a column that has a natural buffering capacity. For example, if the pH of
a weak anion-exchange column is adjusted to 8 and polybuffcr--a term
introduced by Pharmacia Fine Chemicals to indicate a buffer of many
charged components--of pH 5 is pumped into the inlet of the weak anion-
exchange column, the most acidic components of the buffer will be ad-
sorbed and other components will migrate farther down the column be-
fore being adsorbed. The net effect of this process on pH is that the
ion-exchange groups at the head of the column will be titrated and the pH
of the medium will gradually be raised to that of the inlet buffer. This
titration process will be repeated in all segments of the column until the
whole length of the column is, in the example, at pH 5. Because the
titrating buffer enters the head of the column, titration will occur sequen-
tially in column segments and a pH gradient will be established across the
length of the column. If a protein is introduced into the column during this
titration process, it will migrate along the pH gradient until it reaches a
point in the column where the pH is sufficiently high that it will become
negatively charged. This pH will be near, but not necessarily at, the pI of
the protein because some proteins have intensely charged areas on their
surface that bind to an ionic surface even at their pI. As the pH gradient
moves down the column, a protein is alternately adsorbed and desorbed
in such a manner as to have a focusing or concentrating effect on sample
components. Proteins will move along the column at or near their pl.
As in other types of ion-exchange chromatography, resolution may be
influenced by mobile-phase velocity and gradient slope. Figure 3 shows
L. A. E. Sluyterman and J. Wijdenes, J. Chromatogr. 150, 31 (1978).
7 L. A. E. Sluyterman and J. Wijdenes, J. Chromatogr. 206, 429 (1981).
8 L. A. E. Sluyterman and J. Wijdenes,J. Chromatogr. 206, 441 (1981).
182 CHROMATOGRAPHY [8]

, i t ,BI i t i c

E
cO.J. 5

"T
N 0. 4 4 t
t,- I
3~

L
0.3 IlI

~0.2 2

mO 0.1
I
J I
0o s ~ m5 20 2S 30
TIME (rain}

FIG. 3. Chromatofocusing profiles showing the effect of flow rate on resolution of peaks
ofconalbumin. A 4% suspension of conaibumin (20/,l) was chromatographed at (A) 2, (B) l,
and (C) 0.5 ml/min. The pH gradient (---) was formed using a 1 : 10 dilution of PB 96
polybuffer, pH 6.0, after the column was equilibrated in 0.25 M imidazole-acetic acid at pH
7.4. I, II, and III represent the major components of conalbumin.

the relationship between mobile-phase velocity and resolution of compo-

nents in a conalbumin sample at the same gradient slope (percentage of
change per unit volume of mobile phase). Flow rates even lower than 0.5
ml/min are useful if the pumping system retains its accuracy at these low
velocities.
Gradient slope is controlled in chromatofocusing by the concentration
of polybuffer being pumped into the column. As the buffer becomes more
dilute, a greater volume will be required to titrate the support and the pH
change of the eluent per unit volume of buffer added to the column will be
smaller. It is apparent that the amount of buffer required to titrate the
column will be a function of the charge density of the support. Since
charge density varies considerably among commercial supports, the
column manufacturer's literature should be consulted for guidance, or
several dilutions of the polybuffer examined to achieve the desired gradi-
ent slope.
Only two types of columns have currently been reported as useful:
Pharmacia Mono P and the Synchrom AX series. Both are weak anion-
exchange supports. Mono P is a proprietary material of undisclosed pore
diameter; presumably, the pore diameter is 800 ~ as in the Mono Q
support. The Synchropak AX materials have a polyethylenimine phase
bonded to silica and are available on 100-, 300-, 500-, or 1000-~ diameter
porosity. It has been reported s that polyethylenimine is useful in chroma-
[8] HPLC: ION EXCHANGE 183

tofocusing because it has a very broad titration curve and that the pore
diameter contributes to resolution in chromatofocusing. Substantially bet-
ter resolution of estrogen receptor proteins were obtained on 500-.~ pore
diameter supports than 300 A.9
Although it is possible to prepare polybuffers in the laboratory, they
are generally inferior to the products of Pharmacia. l° Pharmacia recom-
mends that only a polybuffer be used over an interval of 3 pH units. To
cover the full range of pH within which one must work, they offer three
polybuffers: polybuffer 74 for the pH 7 to 4 range; polybuffer 96 for the
pH 9 to 6 range; and Pharmalyte for the pH 10.5 to 8 range. Pharmacia
polybuffers are supplied as sterile filtered solutions at 0.075 mmol per pH
unit per milliliter.
Before reusing a chromatofocusing column, all residual components
from the previous sample should be removed by adjusting to pH 3 or less,
using 1 M salt, or by a combination of low pH and high ionic strength.
Two to five column volumes should be sufficient for recycling, depending
on the nature of the sample. After this step, the column may be brought
back to the initial pH required for chromatofocusing.
The advantages of chromatofocusing are in provision of another
method of ion-exchange chromatography that has a selectivity different
from that of ionic strength gradient elution and requires only a single
pump to form a gradient. The principal disadvantage is that fractionation
is achieved near the isoelectric point of a protein at low ionic strength,
both of which favor denaturation. As yet, there are relatively few reports
of high-performance chromatofocusing. The true utility of the technique
compared to ionic strength gradient elution remains largely undeter-
mined.

Optimization

One of the questions regarding any separation is whether it is the best

that can be achieved. This is particularly pertinent when trying to purify a
protein with the minimum number of steps or when a coeluting substance
prevents analytical quantitation. The answer to the question requires a
more detailed discussion of resolution. Unfortunately, there is no single
means by which it may be determined that a column is operating at maxi-
mum resolution or to optimize resolution. Two sets of variables control
resolution in a chromatographic system: those that are inherent to the
column and those that are subject to experimental control. The emphasis

9 T. W. Hutchens, R. D. Wiehle, N. A. Shahabi, J. V. Evan, and J. L. Wittliff, J. Chroma-

togr. 2,66, 115 (1983).
~0G. Wagner and F. E. Regnier, Anal. Biochem. 126, 37 (1982).
184 CHROMATOGRAPHY [81
here is with the latter set of variables, among which the most important
are mobile-phase velocity, ionic strength, pH of the mobile phase, and the
displacing ions.
From the resolution equation [Eq. (1)] it may be seen that resolution
can be maximized by decreasing the peak width (aVe) or by maximizing
the difference in elution volume (Ve2 - Vel) between peaks. These two
variables of a separation are often referred to as band spreading and
selectivity, respectively. The tendency of chromatographic columns to
spread solute bands (increase AVe) is well known and easily explained.
The operational variables that contribute strongly to band spreading are
mobile-phase velocity (flow rate), diffusion coefficient of the protein, and
viscosity of the mobile phase. Mobile-phase velocities of 7.2 ml/min per
square centimeter of cross-sectional area in an ion-exchange column are
average. Columns operated at this velocity with a 3 to 6% per minute
gradient slope will give good separations in 15-30 min. Decreasing the
mobile-phase velocity to 3.6 ml/min per square centimeter (0.5 ml/min in a
4.2 x 50 mm column) has been shown to increase resolution 1.5-fold when
a 1.66%/min gradient slope was used. Total separation time in this case
would be 60 min. Decreasing mobile phase velocity still more to 1.8 ml/
min per square centimeter (0.25 ml/min in the 4.2 × 50 mm column), and
using a 0.75%/min gradient slope with a separation time of 3 hr produced a
separation only 12% better than that at 1 hr. Available information sug-
gests that the small additional increase in resolution that may be gained
beyond separation times of 60 min does not warrant that investment of
time.
Obviously, the enhancement of resolution gained by decreasing mo-
bile-phase velocity was accomplished by allowing solute molecules longer
for equilibration between the stationary and mobile phase. Anything that
hinders this equilibration process will diminish resolution. Subambient
column operation is one such means. As column temperature is lowered,
mobile-phase viscosity increases, with a concomitant decrease in solute
diffusion coefficient. Unless there is compelling evidence that enzyme
activity is lost in transit through the HPLC column at ambient tempera-
ture, subambient column operation is not recommended.
Reference has been made to a relationship between gradient slope and
flow rate in resolution, i.e., the percentage of solvent change per minute
relative to the flow rate. It is important that these two variables be bal-
anced in a manner that allows elution of one substance from the column
before another is desorbed. Gradient slope relative to mobile-phase veloc-
ity determines the number of column volumes of mobile phase that will be
pumped through the column to effect elution of adsorbed material. The
examples given in this section used a gradient slope relative to velocity
[8] H P L C : ION EXCHANGE 185

that produced a 30 column-volume gradient. This means that there was a

3.3% salt gradient across the column. In a 25-cm column, the gradient
would be 0.13%/cm; and in a 5-cm column, a 0.66%/cm gradient. Elution
volumes equal to 20-40 column-volumes are routine with 5-cm columns.
The number of elution volumes used to develop a column is easily calcu-
lated by dividing the total elution volume by the liquid volume of the
column. In general, the liquid volume of an anion-exchange column is 65-
70% of the volume of the empty column.

Selectivity
Selectivity is the word usually used to describe the differential affinity
of a column for two solutes. As the difference in retention between two
solutes increases, the term Ve2 - Ve1 in Eq. (1) will increase, with a
concomitant increase in resolution. To vary the differential affinity of a
column for two solutes, the investigator must find some property in their
interaction with the support that may be made different. The two vari-
ables in HPIEC of proteins that have been found to be the most useful in
altering selectivity are the pH of the mobile phase and the nature of
displacing ions. As shown in the retention maps (Fig. 2), the difference in
retention between solute species varies with the pH of the mobile phase,
which can be traced to unique differences in the pH titration curves of
proteins. Although ion-exchange retention is not directly proportional to
net charge, the titration curve indicates that differential ionization of ionic
groups is taking place. By changing the pH of the mobile phase, the
possibility of differentially altering the number of sites at which two pro-
teins are adsorbed to a surface may be explored. It will be recalled from
the section on retention that retention on strong anion-exchange columns
is increased as the pH of the mobile phase is elevated, whereas retention
on strong cation-exchange columns increases with a decrease in mobile-
phase pH. The opportunity exists for achieving multiple purification steps
on a single column by operating it with several mobile phases of different
pH. Impurities coeluting from an ion-exchange column operated at one
pH have a reasonable possibility of being separated from the desired
component at another pH.
The displacing salt present in the strong buffer is equal in importance
to pH in controlling retention and selectivity. H For a variable that is so
important, selection of a displacing ion is, unfortunately, complex. This is
illustrated in Tables II and III, in which the retention time of a number of
proteins is examined on anion- and cation-exchange columns. It is seen

H W. Kopaciewicz and F. E. Regnier, Anal. Biochem. 133, 251 (1983).

186 CHROMATOGRAPHY [8]

TABLE II
INFLUENCE OF VARIOUS ANIONS ON THE RETENTION AND RESOLUTION OF
SELECTED PROTEINSa

Relative displacing power c

Anions b Rs Rs
(sodium salt) Ova a STI a Chy' Cyt ce Lys e Ova/STl Cyt c/Lys

Bromide 0.70 1.00 0.93 0.77 1.0 4.1 2.2

Phosphate 0.59 0.74 1.00 1.00 0.90 ND y 3.8
Citrate 1.00 0.80 0.64 0.44 0.51 5.3 3.6
Chloride 0.60 0.88 0.87 0.66 0.80 5.4 3.6
Perchlorate 0.81 0.88 ND ND ND 4.3 ND
Fluoride 0.45 ND 0.74 0.56 0.58 ND 6.7
Bicarbonate 0.61 0.79 ND ND ND 4.1 ND
Tartrate 0.70 0.69 0.64 0.56 0.51 4.5 4.4
Sulfate 0.73 0.70 0.70 0.60 0.61 6.1 3.0
Formate 0.54 0.74 ND ND ND 3.5 ND
Acetate 0.54 0.59 0.75 0.73 0.70 5.7 4.1
Propionate 0.54 0.56 ND ND ND 4.2 ND

The following abbreviations are used for proteins: Ova, ovalbumin; STI, soybean
trypsin inhibitor; Chy, chymotrypsin; Cyt c, cytochrome c; and Lys, lysozyme.
b Chromatography was performed at pH 8. The ionic strength of buffer B was 0.5.
c Unity refers to the strongest displacing salt for that protein; 0 the least.
d Determined on the Q-300 silica-based strong anion-exchange column.
e Determined on a Pharmacia Mono S strong cation-exchange column.
s ND, Not determined.

that different ions affect the chromatographic behavior uniquely. For soy-
bean trypsin inhibitor, bromide was the most powerful displacing anion in
contrast to citrate for ovalbumin. The relative order of displacing power
was also seen to vary between anions from a low of 3.5 with formate to a
high of 6.1 with sulfate, using an anion-exchange column and the same
pair of proteins. In contrast, sulfate produced the lowest selectivity and
fluoride the highest with the cytochrome c/lysozyme pair on a strong
cation-exchange column. The ability to alter selectivity by changing dis-
placing ions is of obvious utility in purification. For cations, similar phe-
nomena were observed. It may be generally stated that magnesium and
bromide ions are among the most powerful displacing agents as con-
trasted to lithium and fluoride ions, which are among the weakest. When
salts are used at the same concentration, strong displacing agents de-
crease retention and compress the chromatogram, whereas weak displac-
ing agents expand the chromatogram. Compression or expansion of elu-
tion profiles has minimal influence on resolution unless there is a change
in the retention of one protein relative to another.
[8] HPLC: IoN EXCHArq6E 187

TABLE III
INFLUENCE OF VARIOUS CATIONS ON THE RETENTION A N D RESOLUTION OF FIVE
SELECTED PROTEINSa

Relative displacing power b

gs Rs
Cations c Ova d STI a Chy e Cyt c" Lys ~ Ova/STI Cyt c/Lys

Chloride salt STI Lys

Lithium 0.68 0.57 0,60 0.56 0.54 5.2 3.1
Sodium 0.82 0.72 0,93 0.97 0.98 5.4 3.6
Potassium 0.80 0.69 0.93 0.98 0.98 5.0 4.0
Ammonium 0.83 0.74 0.86 0.94 0.93 5.5 3.7
Magnesium 1.0 0.94 1.00 1.00 1.00 3.1 2.8
Calcium 1.0 1.0 ND f ND ND 2.8 ND
Acetate salt
Sodium 0.67 0.47 0.81 0.88 0.85 5.7 4.1
Magnesium 0.83 0.57 ND ND ND 4.2 ND

See footnote a of Table II.

b Unity refers to the strongest displacing salt for that protein; 0, the least.
c Chromatography was performed at pH 8. The ionic strength of buffer B was 0.5.
d Determined on the Q-300 strong anion-exchange support.
e Determined on the Pharmacia Mono S strong cation-exchange column.
/ ND, not determined.

Hydrophobic Proteins
The principal problems with hydrophobic proteins are their limited
solubility in water and their tendency to aggregate. In size-exclusion chro-
matography [2 (see also this volume [7]), a number of denaturing agents
and various organic acid-alcohol combinations have been used success-
fully. The choice of a mobile phase is more dependent on the unique
requirements of the protein for dissociation and solubilization than on the
chromatography system. In contrast, ion-exchange systems are much
more restrictive. Mobile-phase additives that modify the solubility of a
protein cannot be ionic; large concentrations of ionic solubilizing agents
would swamp the requisite ionic character of ion-exchange columns and
prevent electrostatic interaction of the protein with the column. This
problem has been circumvented in part by the use of organic solvents or
nonionic detergents as solubilizing agents (see this volume [16]).
Despite the hydrophobicity of the Mr 8000 subunit of the chloroplast

~2j. D. Pearson, E. Pfankoch, and F. E. Regnier, "Food Constituents and Food Residues:
Their Chromatographic Determination." Dekker, New York, 1983.
 188 CHROMATOGRAPHY [8]
coupling complex, Tandy et al.13 have reported that ion-exchange separa-
tions could be obtained in high concentrations of chloroform and metha-
nol. Proteins were applied to a weak anion-exchange column (0.41 ×
10 cm SynChropak AX 300) in CHCI3-CH3OH, and the column was
eluted according to the following protocol: 15 min of isocratic elution with
CHCI3-CH3OH (2 : 1); 10 min of isocratic elution with CHC13-CHaOH
(1 : 1); and, finally, gradient elution in 25 min from CHCI3-CH3OH-H20
(3 : 3 : 1) to the same eluent containing 20 mM ammonium acetate. Elution
from the column during the gradient was on the basis of charge.
In the presence of 0. I% Triton X-100, envelope proteins from Sendai
virus have been fractionated on a strong anion-exchange column (0.5 ×
5 cm, Mono Q) and eluted with a gradient ranging from 0.02 M sodium
phosphate at pH 7.2, containing 0.15 M sodium chloride and 0. I% Triton
X-100, to the same solution except for the addition of 1.5 M sodium
chloride.14 A similar type of separation of bovine viral diarrhea proteins
has been reported with Berol as the detergent) 5
Six-molar urea has been used to dissociate alkaline phosphate com-
plexes. 16 Although this reagent works well and has little influence on the
ion-exchange process, the high concentration of urea results in high vis-
cosity and, therefore, large column back pressure. Urea is recommended
only when other techniques are ineffective.

Cleaning Columns
During the course of many analyses, residual protein will accumulate
on the column. This may be the result of denaturation, low solubility in
the mobile phase, or very strong interaction with the support surface. If
residual protein is not periodically removed, the column loses ion-ex-
change capacity and resolution and can eventually plug. When perfor-
mance tests indicate deterioration of resolution, residual materials should
be eluted from ion-exchange columns by extremes in pH, high ionic
strength, or organic solvents (see also this volume [6]). A combination of
these techniques using gradients of 0.1 to 0.5% trifluoroacetic acid
(TFA) to 60% propanol in 0.1 to 0.5% TFA may be preferred. This solvent
system is widely used in reversed-phase chromatography and appears to
be able to solubilize both the most ionic and hydrophobic proteins. Five to

13 N. Tandy, F. E. Regnier, and R. Dilly, J. Chromatogr. (in press).

14 G. W. Welling, G. Kroen, and S. Welling-Webster, J. Chromatogr. (in press).
t5 p. Karsncs, J. Moreno-Lopez, and T. Kristiansen, J. Chromatogr. (in press).
16T. D. Schlabach, "High Speed Isoenzyme Profiling by HPLC." Ph.D. Thesis, Purdue
University, Lafayette, Indiana, 1978.
[8] HPLC: ION EXCHANGE 189

10 repetitive gradient cycles with the TFA-propanol mixture is usually

sufficient for cleaning. In the event that some proteins remain adsorbed,
substitution of 60% formic acid for TFA will probably remove them.
However, it is advisable to consult the manufacturer before using such
severe solvents because some support matrices are destroyed under very
acidic conditions.

Sample Preparation

The amount of sample preparation required in HPIEC depends on the

origin and history of the sample. Obviously, all particulate matter must be
eliminated; thus, all cell extracts require preliminary treatment.
When the protein has been previously subjected to column, salt, or
solvent fractionation, the ionic strength and pH of the sample must be
adjusted to allow electrostatic association with the column. Although
solvent adjustment of this sort may be accomplished in many ways, dialy-
sis seems to be the most common. If the sample is very dilute and ionic
strength has been reduced to less than 20 mM of buffer by dialysis, large
volumes of sample may be applied to a column. For example, 300-500 ml
of a solution containing 0.2 mg of protein per milliliter can be pumped
onto a 1 × 25 cm polyanion silica column and gradient eluted to produce a
normal chromatogram. 17

Column Switching
Purification of proteins is usually a labor-intensive, multiple-column
process. In addition to the time required for chromatography, consider-
able additional time is consumed in collecting and preparing samples for
the next chromatographic step. Part of this expenditure of labor may be
circumvented if the sample can be directly transferred from one column to
another. The coupling of two different types of columns with the aid of
switching valves has been shown 18 to be applicable in the transfer of a
protein directly from a size-exclusion to an ion-exchange column without
an intermediate collection step. Direct transfer from ion-exchange to re-
versed-phase columns was equally successful. Although not widely used
currently, apparatus for column switching is being introduced commer-
cially and should become more common in the near future.

t7 W. Kopaciewicz and F. E. Regnier, Anal. Biochem., submitted for publication.

is W. Kopaciewicz and F. E. Regnier, Anal. Biochem. 129, 472 (1983).
190 CHROMATOGRAPHY [9]

[9] R e v e r s e d - P h a s e H i g h - P e r f o r m a n c e
Liquid Chromatography
B y MILTON T. W. HEARN

In recent years, reversed-phase high-performance liquid chromatogra-

phy with porous, microparticulate, chemically bonded alkylsilicas has
emerged as a very rapid and selective method for peptide and protein
purification. 1,2Although the role of a variety of mobile-phase parameters,
such as the concentration of the organic solvent modifier, the pH, and the
buffer ion composition on peptide resolution in general is now well docu-
mented, the purification of proteins by similar reversed-phase techniques
still requires careful optimization on a case-by-case basis. Investigations
with a variety of silica-based hydrocarbonaceous supports with nominal
pore diameters encompassing the range of 6 to 50 nm have allowed sev-
eral crucial stationary phase characteristics to be identifiedr-5 for a num-
ber of specific structural features of proteins based on studies using se-
lected protein standards. For peptides and proteins separated on
adequately bonded alkylsilicas, retention correlates well with the topo-
graphic surface polarity of these molecules. As a consequence, chromato-
graphic behavior for a group of peptides under a particular set of elution
conditions can largely be described in terms of the concepts of linear
elution development with selectivity following regular free-energy rela-
tionships, which can often be directly estimated from consideration of the
amino acid composition and sequence differences. Although these empiri-
cal relationships have enabled, in a variety of selected peptidic cases, the
prediction of elution orders from topological indices, ~ the current data
base does not yet adequately take into account amino acid residue posi-
tional effects or changes in the unique secondary or tertiary structural
features of a peptide or protein that may arise with different elution condi-
tions.
Recognition of the synergistic features of mobile-phase and stationary-
phase parameters and their adequate control such that the chromato-

I M. T. W. Hearn, Adv. Chromatogr. 20, 1 (1982).

2 M. T. W. Hearn, F. E. Regnier, and C. T. Wehr, Am. Lab. (Boston) 14, 18 (1982).
J. D. Pearson, N. T. Lin, and F. E. Regnier, in "High Performance Liquid Chromatogra-
phy of Proteins and Peptides" (M. T. W. Hearn, F. E. Regnier, and C. T. Wehr, eds.),
p. 81. Academic Press, New York, 1983.
4 M. T. W. Hearn and B. Grego, J. Chromatogr. 288 (in press).
5 M. J. O'Hare, M. W. Capp, E. C. Nice, N. H. C. Cooke, and B. G. Archer, in "High
Performance Liquid Chromatography of Proteins and Peptides" (M. T. W. Hearn, F. E.
Regnier, and C. T. Wehr, eds.), p. 161. Academic Press, New York, 1983.

Copyright © 1984 by Academic Press, Inc.

METHODS 1N ENZYMOLOGY, VOL. 104 All fights of reproduction in any form reserved.
ISBN 0-12-182004-1
[9] HPLC: REVERSEDPHASE 191
graphic distribution ideally involves a single (conformationally stable)
molecular species is of the utmost importance for the separation of large
polypeptides and proteins by the reversed-phase techniques. The great
versatility of peptide separation by reversed-phase HPLC arises from the
ease with which mobile-phase elutropicity can be manipulated. However,
the present popularity of several elution conditions that were developed
initially with peptide separations in mind, and their application largely in
an empirical manner, has tended to obscure the requirements for optimal
resolution and recovery of larger polypeptides and proteins with alkylsil-
ica supports. In this chapter, focus is on the selection of mobile-phase
and stationary-phase parameters that lead to improved strategies for the
purification of both native and denatured proteins by reversed-phase
methods. General reviews on the use of HPLC in protein chemistry have
been published6,7 and one is extensively recorded in this volume [6]-[8].
Shorter discussions of procedures for resolution optimization, including
pairing-ion modulation of retention behavior of polypeptides and proteins
on alkylsilicas, have also appeared in the scientific literature, g-~°

Background
There are basically two types of interactive retention mechanisms that
can be proposed for protein separation by reversed-phase HPLC: parti-
tion and adsorption. The concept of liquid-liquid partition implies that a
homogeneous solution of a protein in the nonpolar phase occurs, and
further, the concept requires that the intercept (C) of the linear plot (with
slope equal to unity) of the logarithmic capacity factor versus the logarith-
mic partition coefficient is less than the phase ratio, ~, of the chromato-
graphic system. Experimentally, the observed C values for polypcptides
and even for amino acids and small peptides separated on alkylsilicas are
much larger than the • values. As a consequence, the retention process
should be more appropriately characterized as involving adsorption pro-
cesses and alkylsllicas, based on bonded monolayers of a nonpolar phase,
considered from a formal point of view as diffuse solid adsorbents. In
contrast to classical liquid-solid chromatography of proteins on neat sil-
ica or glass surfaces, where the interactions between solute and adsorb-

S. Stein, in "The Peptides" (E. Gross and J. Meienhofer, eds.), Vol. 4, p. 73. Academic
Press, New York, 1981.
M. T. W. Hearn, in "High-Performance Liquid Chromatography: Advances and Perspec-
tives" (C. Horvath, ed.), p. 87. Academic Press, New York, 1983.
s M. T. W. Hearn and B. Grego, J. Chromatogr. 1,55, 125 (1983).
9 R. A. Barford, B. L. Sliwinski, and H. L. Rothbart, J. Chromatogr..235, 281 (1982).
10 M. T. W. Hearn and B. Grego, J. Chromatogr. 218, 497 (1981).
192 CHROMATOGRAPHY [9]
ent involve strong polar forces, protein-stationary-phase interactions in
reversed-phase HPLC are predominantly due to weak, nonspecific van
der Waals forces. However, with most bonded alkylsilicas currently
available, mixed retention mechanisms involving nonpolar and polar con-
tributions operate. The intensity of the polar contributions in the retention
process has been found to be directly related to the extent of alkyl ligand
coverage and the chemical nature of the mobile-phase components that
can adsorb to the surface of the nonpolar support.

Protein Separation by Reversed-Phase Chromatography

To date, most protein separations on hydrocarbonaceous silica-based
surfaces have been phenomenological in nature with the methodology
applied largely in an empirical manner. Recent experience has shown that
the silica-based supports with pore diameters in the range 6-10 nm are far
from optimal for the chromatography of proteins with regard to resolution
and high recovery. Further, the separation of proteins on alkylsilicas with
water-organic solvent combinations often involves much more complex
phenomena than is evident with peptides, e.g., adsorption-desorption ki-
netics which may be slow compared to the chromatographic time scale.
Interpretation of the structure-retention relationships of proteins sepa-
rated on alkylsilicas, in contrast to the reversed-phase HPLC of peptides,
is thus complicated by multiple chromatographic and nonchromato-
graphic variables that may affect either protein structure in solution or
protein structure at a polar liquid-hydrocarbonaceous surface interface.
Proteins exhibit great structural diversity, with the native three-dimen-
sional structure very sensitive to changes in the microenvironment of
solvation. Contact with organic solvents, adsorption at air-liquid or liq-
uid-solid interfaces, high ionic strengths of dissociating salt species, or
extremes of pH may cause denaturation and loss of biological activity. As
a consequence, several practical requirements must be met for successful
reversed-phase HPLC separations of proteins.
The first, and most important, requirement when preservation of bio-
logical function is a necessary goal of the chromatographic separation is
the issue of irreversible conformational changes induced by mobile-phase
composition or the characteristics of the stationary phase. Analysis of the
effects of mobile-phase composition on biological activity in preliminary
incubation experiments aids the selection of suitable eluents but does not
guarantee optimal resolution or recovery.
A second requirement that must be satisfied is associated with the
phenomenon of multisite attachment of protein solutes to alkylsilica sta-
tionary phases. The concept of cooperative multisite interactions between
[9] HPLC: REVERSEDPHASE 193

proteins, such as glycolytic enzymes and n-alkylated agaroses, has been

well developed by Jennissen, H,~2 and similar behavior is clearly evident
with large-pore diameter alkylsilicas. This behavior invariably results in a
pronounced sensitivity of protein elution on the composition of the mobile
phase and is particularly noticeable when isocratic behavior and gradient
retention behavior for a group of proteins are compared. With water-
organic solvent eluents, this behavior is manifested by severe dependen-
cies of the logarithmic capacity factor on the volume percentage of the
organic solvent modifier. Furthermore, owing to heterogeneous protein-
stationary-phase interactions in the multisite attachment mode, broad-
ened or multiple peak shapes can occur.
A third requirement arises with complex mixtures where, from practi-
cal necessities, it is important to ensure that chromatographic retention of
a specific protein is essentially mediated by a single molecular species.
Variations in protein shape, extent of ionization, and conformation may
be induced by either the eluent composition or stationary-phase charac-
teristics. Size-exclusion phenomena as well as distorted peak shapes may
result. Because proteins under these conditions may exhibit unique popu-
lations of reversible or irreversible states involving native, partially un-
folded, and fully unfolded conformers in solution, different manifestations
of these complex surface-directed interactive phenomena can occur under
a given set of chromatographic conditions. Differences in interactive af-
finities, attendant on deformational changes induced by eluent or station-
ary-phase effects, which may reveal additional sequestered hydrophobic
or polar domains of a polypeptide or protein, will affect chromatographic
performance. As a consequence, optimization of chromatographic perfor-
mance with porous, microparticulate alkylsilicas requires the same level
of tactical consideration as, for example, is used to define the best strat-
egy for approaching the amino acid sequence determination of a given
protein. The availability of the starting material, the end use of the puri-
fied protein, the number of discrete high-resolution separation steps that
may be required to achieve the best purification factors, will all determine
the most appropriate choice of chromatographic conditions.
Specific clues guiding the rational selection of chromatographic condi-
tions for optimal resolution and recovery come from physical measure-
ments of solubility, exploratory analytical elution experiments, evalua-
tion of selectivity matrices (and, in particular, the evaluation of ~-plots4.7),
and attention to the susceptibility of some proteins to respond to pairing-
ion stabilization of their secondary and tertiary structure, which may thus
11 H. P. Jennissen, Biochemistry 15, 5683 (1976).
12 H. P. Jennissen, J. Chromatogr. 159, 71 (1978).
194 CHROMATOGRAPHY 19]
facilitate their separation. Attention must also be given to methods for
eluent preparation as well as procedures for handling samples at the pre-
and postcolumn stages. In a study with several different protein hor-
mones, 13,~4improved methods of handling the chromatographic fractions
greatly facilitated the recovery of the proteins in biologically active form,
particularly in circumstances where the abundance of the desired compo-
nents was low and their initial dilution high, i.e., < 10/~g/ml. Practically, it
is obviously expedient to optimize selectivity with volatile buffers where
possible. Although such choice of eluents should be contemplated at an
exploratory stage of an investigation, their use may not necessarily be
desirable with respect to the highest resolution or recovery.
The most efficient strategy for successful reversed-phase HPLC purifi-
cations of proteins is one that continuously redefines the practical limita-
tions of the elution conditions as the molecular features of the protein of
interest become more evident. Recovery for many proteins from alkylsili-
cas can be improved by the addition of stabilizing cations or cofactors to
the eluent. For example, the addition of 2 mM CaC12 to water-organic
solvent mixtures enhanced the recovery of both trypsin and parvalbumin
from reversed-phase silicas.15-~7 Reappraisal of strategy may necessitate
the use of alternative columns of different dimensions, chosen according
to rational resolution optimization requirements and a battery of different
alkylsilica phases, characterized in terms of ligand coverage and porosity
at different stages of the purification scheme. Rigorous adherence to a
single elution condition or stationary phase cannot be expected to gener-
ate efficient high-resolution separations except for the most straightfor-
ward problems, typified by the separation of denatured polypeptides
where it may not be essential to recover all components.

Consideration of Retention Behavior

Protein selectivity in this method is dependent on the different distri-
bution coefficients established by the several protein species present be-
tween a polar, water-rich mobile phase and a nonpolar stationary phase
solvated by extractable mobile-phase components. On the assumption
that protein retention involves the distribution of the solute between an
~3M. Dobos, H. K. Burger, M. T. W. Hearn, and F. J. Morgan, J. Mol. Cell. Endocrinol. 31,
187 (1983).
~4G. S. Baldwin, B. Grego, M. T. W. Hearn, J. A. Knesel, F. J. Morgan, and R. J. Simpson,
Proc. Natl. Acad. Sci. U.S.A. 80, 5276 0983).
1~ K. Titani, J. Sasagawa, K. Resing, and K. A. Walsh, Anal. Biochem. 123, 408 (1982).
~6W. S. Hancock, C. A. Bishop, R. L. Prestidge, and M. T. W. Hearn, Anal. Biochem. 89,
203 (1978).
IT M. W. Berchtold, K. J. Wilson, and C. W. Heizmann, Biochemistry 21, 6552 (1982).
[9] HPLC: REVERSEDPHASE 195

interfacial monolayer (m/s) and a mobile eluent (m), a capacity factor, k~,
for a sparingly soluble protein solute, i, can be described by Eq. (1).
In ki' = ln( Cm/s,i/Cm,'~ = ln(ui,~ra/Cm/s) -- ln(u~m/s/Cm) (1)
Cm/s,i and Cm,i are the number of moles of the solute in the stagnant
interfacial phase and the mobile phase; U~,mand Ui~m/sare the solute activity
coefficients in the mobile phase and the bulk solvated stationary phase;
and nm,s and nm are the number of moles of adsorbed and mobile eluent
components. If AAi is the area occupied by a protein molecule on the
nonpolar surface, ui~,m/~is the corresponding solute activity coefficient at
the interface of a solute monolayer, and Ym/s and Yi,s are the surface
tensions of the adsorbent-eluent and adsorbent-solute interfaces, then
retention can be given by Eq. (2).

In kf = In(nays~rim) + (In u .~
l,ln - In Ui~m/~)+ AAiN(ym/s - Yi,s)/RT (2)

Several corollaries arise on the basis of this displacement-adsorption

model. First, the contact area, AAi, for a protein will usually be considera-
bly larger than that occupied by a desorbing agent, such as an organic
solvent molecule. As a consequence, the binding of a polypeptide or
protein to the nonpolar surface will displace multiple molecules of the
eluent, e.g., an alcohol or acetonitrile. As multisite binding becomes
more heterogeneous owing to unfolding transitions in the bulk mobile
phase or at the interface, the assumption that all the molecules of a partic-
ular protein interact in an identical manner with the same number of sites
on the surface--a basic tenet of linear elution development of low-molec-
ular-weight solutes--will clearly no longer be valid. Furthermore, desorp-
tion of these heterogeneously bound species will require a mobile phase of
higher eluotropic strength even when a single molecular conformer is the
dominant binding species. This "unzippering" requirement underlies the
well-documented observation that proteins can usually be successfully
chromatographed on alkylsilicas only under gradient elution conditions.
For example, Lewis et al. is noted that bovine serum albumin eluted from
a Cs column in about 13 min with an isocratic mobile phase of 0.5 M
formic acid-0.4 M pyridine at pH 4.0, containing 34% 1-propanol, but did
not elute when the mobile phase contained 32% 1-propanol. In fact, such
behavior is typical of globular proteins eluted from alkylsilicas. Typical
examples of the pronounced biomodal dependencies of retention on or-
ganic solvent content have been reported for growth hormone, collagen
is R. V. Lewis, A. Fallon, S. Stein, K. D. Gibson, and S. Udenfriend, Anal. Biochem. 104,
153 (1980).
196 CHROMATOGRAPHY [9]

chains, ribonuclease, and phosphorylase a. 19-21Where bimodal dependen-

cies exist, inappropriate changes in solvent composition may account for
the resolution of native and unfolded forms of the same protein under
certain gradient elution conditions.
Second, protein retention is strongly dependent on the extent to which
mobile-phase components can be extracted onto the stationary phase [the
first term of Eq. (2)]. Retention modulation will occur as a consequence of
the extraction of buffer ion species, which may engage in ion-pair or
dynamic ion-exchange phenomena, as well as owing to the extraction of
organic solvent molecules. The influence of these secondary adsorption
isotherms on polypeptide retention has been extensively documented ex-
perimentally. 1 Changes in retention of fully permeating polypeptide or
protein solutes can thus be equated with changes in the extracted solvent
density and differences in solvent eluotropic strengths. The pronounced
effect on stationary phase porosity and matrix field effects due to eluent
component isotherms has been accommodated4,19 into the so-called soft-
ball model of protein retention with alkylsilicas.
Finally, retention of proteins on alkylsilicas is anticipated from a theo-
retical analysis of activity coefficient dependencies, as well as from more
empirical treatments of experimental data, to be the sum of the retention
components due to size-exclusion and interactive processes that involve
not only the nonpolar topographic features of the protein molecule itself,
but also composite polar electrostatic and hydrogen-bonding characteris-
tics. The general empirical expression for protein retention to alkylsilicas
has the form of Eq. (3).
k' = prk r + pek'e + phk~ + psk's (3)
The subscripts r, e, h, and s refer to the hydrophobic, electrostatic, hy-
drogen-bonding, and size-exclusion components of retention, and pr, Pc,
Oh, and Ps are the weighted mole fractions of the solute in each retention
mode. Despite the inherent difficulties in evaluating the k~ and k~ terms in
Eq. (3) due to the complexity of protein structure in solution and of the
multiple participating variables, including ionic strength, pH, dielectric
constant, etc., which influence the electrostatic and hydrogen bonding
terms, a beginning has been made by several investigators 19,2°,22in includ-
ing these effects in the displacement model for protein separation on
alkylsilicas. For a specified column, temperature, and buffer composition,
t9 M. T. W. Hearn and B. Grego, J. Chromatogr. (in press).
2o M. T. W. Hearn and B. Grego, J. Chromatogr. 266, 75 (1983).
2t K. Cohen, B. L. Karger, K. ScheUenberg, B. Grego, and M. T. W. Hearn, submitted for
publication.
z2 X. Geng and F. E. Regnier, Proc. Am. Pept. Assoc. (in press).
[9] HPLC: REVERSEDPHASE 197

the linearized form of the dependency of retention on solvent composition

has been shown ~° to follow Eq. (4).
k~,i = kw,ie-siq' + (ko,rBid,/)-1 + ks,if(q/) (4)
where k ' , ko, and k~ are the capacity factors in neat water (related to
the solubility parameter 8w), in the final organic solvent modifier (an ex-
trapolated value related to k~ and k~) and due to size exclusion, respec-
tively. The variables, S, B, and f are solute and condition dependent.
Since the volume fraction of organic modifier, t0s, is directly related to the
bulk surface tension of the eluent, y, through the modified Fowkes' equa-
tion, then protein retention on alkylsilicas may also be evaluated in terms
of changes in intrinsic surface tensions associated with interaction at the
liquid-solid interface.

Mobile Phase Dependencies

Although literature values of the microscopic surface tensions for pro-
teins under reversed-phase conditions are generally not readily available,
they can be derived from plots of the logarithmic selectivity parameter
(In a = z) versus the eluent surface tension at fixed temperature with a
defined stationary phase. Alternatively, they may be estimated from con-
tact angle measurements. A major interpretative concern that such ap-
proaches raise has been the desire to express solute activity coefficients
predominantly in terms of mobile-phase parameters. In this treatment,
protein retention is equated with the composite free-energy changes asso-
ciated with the formation of a cavity in the eluent of dimensions sufficient
to accommodate the protein molecule, with the van der Waals interac-
tions between the solute and eluent and with the electrostatic interactions
between the eluent and solute. Retention in this model may then be
viewed as a consequence of the propensity of a protein solute to be
expelled from the mobile phase with the stationary phase playing a pas-
sive role as a protein acceptor-displacer surface. Composite strong (elec-
trostatic, hydrogen bonding) and weak (hydrophobic) forces that occur in
the mobile phase are manifested by intrinsic differences in interfacial
surface tension and dielectric constant. Although it was originally as-
sumed that only weak dispersive forces were associated with the station-
ary phase, recent detailed studies 3,~9,2°with well-characterized stationary
phases have demonstrated that secondary effects, due to the stationary
phase ligand coverage and matrix preparation and pretreatment, can sig-
nificantly influence protein selectivity. Undoubtedly, the most successful
theoretical approach to date that accounts for solvent-mediated interac-
tions and the effect of salts on the macroscopic properties of the eluent
 198 CHROMATOGRAPHY [9]

has been the solvophobic model proposed by Horvath and co-workers, z3

According to this model, selectivity for pairs of solutes eluted under
isocratic conditions may be given by Eq. (5).
~'i,j = [A(AGvdw)i,j + yN(AAj - z~u4_i)]/RT (5)
Relative retention of a protein with isocratic eluents y and z on a given
column can be given by Eq. (6).
ri,y/z = C + Aai(KySy - K ~ z ) (6)

The parameter K*corrects for the effect of curvature of the cavity surface
on the macroscopic surface tension of the eluent and is both solute and
eluent dependent. Evaluation of protein selectivity on alkylsilicas in
terms of the cavity factor, K*, and interfacial contact area, AA, from plots
of r versus y with slope given by the to factor 19,2°permits the evaluation of
difference in van der Waals energies of interaction, A(AGvdw), as well as
the free energies of association (AGas,o¢) for different polypeptide or pro-
tein solutes.
When isocratic retention for polypeptides and proteins over an appro-
priate 0-value range obeys Eq. (7)
lnk~ = I n k ' l,W - S t 0 (7)
(as is often observed with polypeptides and globular proteins for the range
1 < k' < 10 under optimized low pH conditions), then the value of capac-
ity factor at the column inlet for linear solvent strength gradient elution
starting from neat aqueous conditions is given by Eq. 8.
In k~ = In k;,w - b(t/to) (8)
and the gradient retention time is given by Eq. (9) 20

/g,i = (to~b) log[2.3bki,w(tsec/to)] + 1 + t,ec + tL (9)

where b = SVm(dO/dt)F -1, F = volumetric flow rate, S = solvent desorp-
tion parameter of a specific protein, to = VoF - j , tsec = retention time for
size exclusion alone, tL = lag time for the gradient system, Vi = mobile-
phase pore volume, Vo = mobile-phase interstitial volume, and Vm = total
mobile-phase volume (equal to Vo + Vi).
From Eqs. (7)-(9), it can be seen that for polypeptides and proteins
with large kw and S values, very little elution development is expected.
The observed sensitivity of protein elution to mobile-phase composition,
particularly the very narrow operational range of solvent strength condi-
tions that can be exploited under appropriate circumstances of isocratic
23c. Horvath, W. Melander, and I. Molnar, J. Chromatogr. 125, 129 (1976).
[9] HPLC: REVERSEDPHASE 199

and batch elution, is in accord with these predictions. Linear time-based

gradient elution is not anticipated to yield optimal resolution for complex
mixtures of proteins. Rather, nonlinear gradients that accommodate s-
value differences are required. Since k" values are usually very large
[e.g., kw (lysozyme) > 1000] the solute band will not migrate significantly
during passage of the lag volume of mobile phase through the band center.
This means that the protein zone is subjected to microscopic solvent
strength increases during gradient elution, which results in band compres-
sion factors substantially different from those observed for small mole-
cules (e.g., Mr = 1000). The consequence of this behavior is a pronounced
dependence of peak height (and recovery) on flow rate. Other corollaries
that emerge directly from Eqs. (7)-(9) relate to optimal column dimen-
sions from the S values versus Vmanalysis, optimal flow rate for different
gradient d~b/dt rates, and solvent strength limits for particular tg values.
Experimental data obtained with large polypeptides and globular proteins
separated on alkylsilicas under denaturing, low pH elution conditions
have been in accord with these theoretical predictions. 4,2°,2~,24

Selection of Chromatographic Parameters

The success of polypeptide and protein separation on alkylsilica sup-
ports depends to a very large extent on the choice of mobile-phase condi-
tions. For gradient elution with organic solvent modifiers, appropriate
selection of the initial and final compositions is important if optimal
resolution and recovery are to be achieved. Relative retention is depen-
dent on the surface area of the nonpolar adsorbent within the column and
the chemical nature of the bonded surface. For most investigators without
access to "in-house" silica-bonding technology, control over the different,
and frequently complex, distribution coefficients can be achieved only
through manipulation of the eluent composition. This limitation immedi-
ately preempts the question of the purpose of the separation. Many hy-
drocarbonaceous silica-based supports exhibit chemical instability at high
pH values that effectively limits their use to below pH 7.5. With mixed
solvent combinations under acidic conditions, many proteins are dena-
tured. It is often feasible, however, to reconstitute proteins isolated by
reversed-phase HPLC into functional proteins either by the addition of
stabilizing cofactors, cations, or other reagents to the eluent or by careful
preferential removal of the organic solvent modifier under nitrogen, or by
pH and buffer ion changes immediately after chromatographic separation.
Examples of the use of one or other of these manipulations can be found
N. H. C. Cooke, B. G. Archer, M. J. O'Hare, E. C. Nice, and M. Capp, J. Chromatogr.
2,55, 115 (1983).
200 CHROMATOGRAPHY [9]
in studies with purification of enzymes such as trypsin 15or papain, 25 ribo-
somal proteins, 26 calcium-binding proteins, 17 and protein hormonesJ 4
Not infrequently, impaired recovery of biological activity can be at-
tributed to handling procedures subsequent to chromatographic fractiona-
tion, such as adsorption to glass or plastic ware following lyophilization.
Difficulties in the purification of human prostatic acid phosphatase 27 and
follicular inhibins ~3 on reversed-phase silicas were circumvented by ap-
propriate handling procedures. A characteristic of protein chromatogra-
phy on alkylsilicas for which poor mass and activity recovery are found is
the appearance of ghost peaks in a blank gradient run immediately after
the initial protein separation has been completed. Often this carryover
effect will give rise to multiple peaks, the later eluting peaks correspond-
ing to irreversibly denatured protein. The presence of ghost peaks can be
usually traced back to the use of stationary phases with bonded nonpolar
surface areas inappropriate for the specific separation. 2°,z~ Stationary
phases with either lower surface areas or, alternatively, decreased ligand
densities frequently circumvent this difficulty without a major adjust-
ment in mobile-phase composition being necessary.

Influence of Organic Solvent Composition

All the water-miscible organic solvents have found use for protein
separation in reversed-phase HPLC. Their efficacy in terms of eluotropic-
ity can be related directly to the nature of the relationship between the
volume fraction (usually expressed as the percentage of solvent) and the
bulk eluent surface tension. Comparative studies with methanol, acetoni-
trile, and 1- or 2-propanol have confirmed that the slope of the retention
dependency, log k' versus qJ, follows the eluotropic strength, e°, of the
organic solvent; i.e., the slope is larger for 1-propanol than, for example,
methanol, the relative retention decreasing as e° decreases. Although
relative retention for a particular polypeptide decreases in the order meth-
anol < ethanol < acetonitrile < 1-propanol or 2-propanol, it does not
follow that resolution will improve with more effective desorbing solvents
such as the propanols. In fact, ternary mixtures containing water and two
organic solvents, e.g., acetonitrile-1-propanol or ethanol-l-propanol, at
compositions chosen to maximize the In k' versus 3' slope of Eq. (7), will
often result in improved selectivity; proteins are thereby eluted with mo-

2~ S. A. Cohen, S. Dong, K. Benedek, and B. L. Karger, in "Affinity Chromatography and

Biological Recognition" (I. Chaiken, M. Wilchek, and I. Parikh, eds.). Academic Press,
New York (in press).
26 A. R. Kerlavage, C. J. Weitzmann, T. Hasan, and B. S. Cooperman, J. Chromatogr. 266,
225 (1983).
27 M. P. Strickler, J. Kintzios, and M. J. Gemski, J. Liq. Chromatogr. 5, 1921 (1982).
[9] HPLC: REVERSEDPHASE 201

bile phases of higher water content than can be achieved with either
individual solvent. Such approaches have been successfully applied in a
number of studies on the isolation of membrane proteins. 28
Because of the magnitude of the S and k~, values for large polypeptides
and proteins, gradient elution provides the most powerful routine method
for their separation on alkylsilicas. Based on exploratory gradient experi-
ments, however, optimized batch elution conditions can be employed for
larger scale preparative separations. In agreement with theory, resolu-
tion is dependent on the gradient steepness parameter, b. Peak capacity
follows an inverse asymptotic dependence on b [Eq. (8)]. With b values
up to about 1.5, peak capacity gradually decreases; above about 2.0,
significant loss in resolution occurs. Since the b term is related to the flow
rate as well as to the rate of change in percentage of the organic modifier,
comparable resolution can be achieved when (d~b/dt)F -1 is held constant.
This can be achieved under conditions of long separation times, i.e., a
shallow gradient and low flow rate, or short separation times with a high
flow rate and steep gradient. Practical constraints will apply in individual
cases, but flow rates in excess of 50 ml/min can be utilized over suitable
d~b/dt ranges for preparative separations, particularly where inappropriate
dwell times for a protein mixture on the column may lead to impaired
recoveries associated with either dissociation or denaturation. A further
corollary arises insofar as solute detectability is also asymptotically de-
pendent on the gradient steepness parameter. Irrespective of whether
spectrophotometric or biological (immunological or biological assay pro-
cedures) detection methods are used, advantage should be taken of the
optimal b value compatible with resolution so that the band width of the
eluting protein solute experiences the maximum band compression. Un-
der such conditions, recovery is favored provided shear degradation or
irreversible changes in conformation do not occur. To a large extent, the
commonly used gradient conditions (l%/min and l ml/min for a 25 ×
0.4 cm column) should be viewed as representing only a starting point for
exploratory experiments designed to establish the correct limits of mo-
bile-phase gradient composition. Once established, significantly reduced
separation times can then be achieved ~9,24by rigorous optimization proce-
dures.
When applied in strictly aqueous solutions, most polypeptides and
proteins are strongly adsorpted to most modern microparticulate re-
versed-phase packing materials. The magnitude of the binding effect is
dependent on the number and accessibility of the hydrophobic domains in
these biopolymers. Most water-soluble proteins contain a fairly large
2s S. Welling-Wester, T. Boer, G. W. Welling, and J. B. Wilterdink, in "Proceedings of the
Second International Symposium on HPLC of Proteins, Peptides and Polynucleotides"
(M. T. W. Hearn, F. E. Regnier, C. J. Wehr, eds.). Elsevier, Amsterdam, 1983. In press.
202 CHROMATOGRAPHY [9]
number of sites of weak affinity for hydrocarbons whereas some proteins,
e.g.,/3-1actoglobulin and the serum lipoproteins, have hydrophobic bind-
ing sites of high affinity. When high-adsorption isotherms are observed, as
with very hydrophobic polypeptides and proteins, special care must be
given to the parameters that influence the sorption-desorption events.
Included would be the initial volume fraction or the rate of change of
volume fraction of the organic solvent modifier in isocratic or gradient
elution, respectively, and pH and buffer ion effects. If these parameters
are not controlled with precision, poor resolution, associated with large
and erratic k' values, will occur as a symptom of "irreversible" binding.
Although the magnitude of the solute-ligand interactions is expected to
decrease with increasing organic solvent modifier in the mobile phase, the
low solubility parameters of polypeptides and proteins (and of some
buffer components) in mobile phases containing high levels of the com-
mon organic solvents, i.e., where tO < 0.75, limits the secondary solvent
composition. The problems associated with limited solubility of polypep-
tides and proteins in hydroorganic eluents can be partially circumvented
by choosing a different modifier combination of lower eluotropic proper-
ties. In this regard, the use of mixed 1- or 2-propanol-based eluents with
detergents has proved to be efficacious for some macroglobulin separa-
tions. 29With labile proteins, optimal elution conditions for analytical sep-
arations may be incompatible with larger scale preparative isolations or
with subsequent assay methods; a compromise between resolution and
recovery must be found. Illustrative of this approach is the purification of
ovine thyrotrophin from a pituitary extract on octyl- and octadecylsilica
supports. 3° On a 50-nm pore diameter, end-capped octylsilica, this glyco-
protein hormone could be readily fractionated using a triethylammonium
formate-acetonitrile gradient eluent at pH 4.0 to pH 6.5, whereas sub-
unit dissociation appears to occur in triethylammonium phosphate at pH
3.0. Similar results have been obtained with luteotropin and human chori-
onic gonadotropin. 31

Influence of Buffer and Ionic Strength

The application of solvophobic theory and pairing-ion concepts32 has
enabled much of the retention behavior, on reversed-phase columns, of
polypeptides and proteins in the presence of different ionic buffers and

29 R. A. Barford, B. J. Sliwinski, A. C. Breyer, and H. L. Rothbart, J. Chromatogr. (in

press).
30 M. T. W. Hearn, P. G. Stanton, and B. Grego, J. Chromatogr. (in press).
3t G. J. Putterman, M. B. Spear, K. S. Meade-Cobun, M. Widra, and C. V. Hixson, J. Liq.
Chromatogr. 5, 715 (1982).
32 M. T. W. Hearn, Adv. Chromatogr. 18, 1 (1980).
[9] HPLC: REVERSEDPHASE 203

salts to be rationalized and judicious choice made for analytical and mi-
cropreparative elution conditions. The addition of miUimolar concentra-
tions of many inorganic salts to the mobile phase tends to decrease rela-
tive retention and improve peak shape, presumably owing to suppression
of polar interactive equilibria. At high salt concentrations, e.g., >100
mM, salting-out effects become evident. This has been exploited with
butyl- and phenyl-bonded TSK-G3000SW, a silica-based packing material
with a 250-/~ mean pore diameter. 33 With this support, it was possible to
separate proteins with high efficiency under mild eluting conditions simi-
lar to those conventionally used in hydrophobic interaction chromatogra-
phy. Linear gradient elution has been possible with ammonium sulfate
concentration decreasing from 2 to 0 M in I00 mM phosphate buffer at pH
6.0. With organic acids and bases (and their salts) relative retention of
proteins on alkylsilicas can be increased or decreased depending on the
chemical nature and polarity of the ionic substance. Considerable flexibil-
ity is thereby introduced into the chromatographic separation by the use
of suitable buffer co- and counterions. With appropriate choice of the
polarity and concentration of the added buffer co- and counterions, spe-
cific electrostatic and solution field interactions can be optimized that lead
to apparent increases or decreases in protein polarity and hence changes
in relative retention. Since they are directly lyophilizable, the volatile
organic acids (formic, acetic, trifluoroacetic, and heptafluorobutyric) and
their alkylammonium salts have been widely utilized. The most appropri-
ate concentration for optimal resolution can be evaluated from pairing ion
considerations and ~" p l o t s . 4A9 The nonvolatile inorganic salts, such as
phosphate and perchlorate, as well as the nonvolatile organics, such as
dodecyl sulfate and heptane sulfonate, also exhibit similar relationships
(asymptotic retention versus concentration) with proteins.
Because of the ease of desalting samples by reversed-phase systems,
removal of inorganic salts from recovered fractions is usually straightfor-
ward. A large variety of ionic and neutral species, which at pH ~< 7.0
modify the retention characteristics of polypeptides and proteins, have
been described.l.7 Table I lists a selection of co- and counterions that have
been used successfully in reversed-phase separations of proteins (see
footnotes 7 and 32 for more extensive reviews of applications and condi-
tions). Because of the surfactant properties of many of the more hydro-
phobic anionic and cationic pairing ions, disaggregations of protein-pro-
tein noncovalent complexes can be achieved, although this may be at the
expense of protein denaturation. However, a dramatic improvement in
peak shape and recovery for polypeptides and proteins is often seen2°,34
33 y. Kato, T. Kitamura, and T. Hashimoto, J. Chromatogr. 266, 49 (1983).
34 j. Rivier, J. Liq. Chromatogr. 1, 343 0978).
204 CHROMATOGRAPHY [9]

TABLE I
ANIONIC AND CATIONICa Co- AND COUNTERIONS
USEFUL IN RP-HPLC SEPARATIONSOF PROTEINS

Anion Cation

Formate R4N +
Acetate R3N*H
Propionate RzN+H~
Trifluoroacetate RN+H3
Heptafluoroacetate Inorganic M +, M 2+
C3-Cn sulfonate N+H4, pyridine
C3-CI2 sulfate
Phosphate
Perchlorate
Bicarbonate

a The chain length for the cationic reagents usually

ranges from CI to C7 (see footnotes 1, 4, and 32
for a compendium of applications).

with these reagents. Above critical concentrations with nonionic and

ionic surfactants, micellar chromatography of proteins occurs on alkylsili-
cas. z°'z9 Of the organic acids, trifluoroacetic acid in particular has been
found to induce desirably reversed-phase selectivity for protein separa-
tion owing to its effectiveness as a solubilizing ion-pairing reagent. Simi-
larly, alkylammonium phosphates such as triethylammonium phosphate,
typically at 0.1 M, have provided alternative isolative capabilities, partic-
ularly in the range pH 3-5.

Influence of pH
With proteins, as with peptides, pH values in the range of 2-3 tend to
give better resolution on alkylsilicas. The possibility that a specific protein
will be insoluble at or near its pI value must be taken into account when a
pH is chosen. Similarly, dissociation of multisubunit proteins or the un-
folding of the tertiary structure may occur at low pH. The adverse effects
of reduced solubility, dissociation, or unfolding arising from low pH mo-
bile phases can be readily assessed in static experiments. A suitable
choice of pH, permitting local resolution and optimum recovery, is rarely
a practical problem. Although low pH values favor suppression of ioniza-
tion of carboxyl groups (as well as silanol groups on the stationary phase),
careful handling is required when proteins are exposed to aqueous organic
solvent mixtures at low pH. For example, both human growth hormone
[9] HPLC: REVERSEDPHASE 205

and insulin deamidate 35,36at a moderate rate at pH 2.5 and residue cleav-
age at acid-labile sites may also occur, as appears to be the case with
apolipoproteins. 37 Significant broadening of the eluted peak can occur
with only modest increases in pH, e.g., with ACTH1_39,38 whereas other
proteins require very low pH values for elution, e.g.,/3-microglobulin. 39
Pyridine-acetic acid or formic acid systems, as used for human fibroblast
interferon4° in the range of pH 3.5-5.0, provide comparable results to
alkylammonium phosphate of lower pH value, with the added benefit of
volatility. For proteins that are unstable at a low pH, these systems offer
considerable flexibility, particularly when on-line fluorescent detection is
available. When combinations of pH and pairing-ion modulation under
gradient elution conditions are used, it is possible to separate very closely
related molecules. For example, Terabe et al. 4~ have used butane sulfo-
nate to resolve Thr30-bovine insulin, bovine insulin, human insulin, and
porcine insulin, the last two differing by only I out of 51 amino acid
residues. Related techniques have been used for the isolation of insulins
from animal pancreas, 42 corticotropin analogs, 43 relaxins, 44 the neurophy-
sins, 45 the bovine pancreatic trypsin inhibitor aprotinin, 46 ferritin, 47 colla-'
gen c~-chains, TMthyroglobulins, 48 cytochrome c, 34,47albumins, 38:7 and acyl
carrier proteins and analogs. 49

Column Material
Several important practical benefits arise as a result of the high affinity
of polypeptides and proteins for alkylsilicas. First, the solute mixture can
be loaded onto a column in a large volume of a very dilute solution and

35 B. Grego, F. Lambrou, and M. T. W. Hearn, J. Chromatogr. 266, 89 (1983).

M. E. F. Beimond, W. A. Sipman, and J. Olivie, J. Liq. Chromatogr. 2, 1407 (1979).
37 W. S. Hancock, C. A. Bishop, A. M. Gotto, D. R. K. Harding, S. M. Lamplugh, and J. T.
Sparrow, J. Lipid Res. 16, 250 (1981).
3s M. J. O'Hare and E. C. Nice, J. Chromatogr. 171, 209 (1979).
39 V. L. Alvarez, C. A. Roitsch, and O. Henriksen, Anal. Biochem. 115, 353 (1981).
40 S. Stein, C. Kenny, H. J. Friesen, J. Shively, U. Del Valle, and S. Pestka, Proc. Natl.
Acad. Sci. U.S.A. 77, 5716 (1980).
41 S. Terabe, R. Konaka, and K. Inouye, J. Chromatogr. 172, 163 (1979).
42 M. T. W. Hearn, B. Grego, J. Cuttield, and M. Mclnnes, in preparation.
43 H. P. J. Bennett, J. Chromatogr. 266 (in press).
44 j. R. Walsh and H. D. Niall, Endocrinology 107, 1258 (1980).
45 j. A. Glasel, J. Chromatogr. 145, 469 (1978).
46 K. Krummen, J. Liq. Chromatogr. 3, 1243 (1980).
47 W. Monch and W. Dehnen, J. Chromatogr. 147, 415 (1978).
40 A. J. Paterson and M. T. W. Hearn, J. Protein Chem., submitted for publication.
49 W. S. Hancock, C. A. Bishop, R. L. Prestidge, D. R. K. Harding, and M. T. W. Hearn,
Science 200, 1168 (1978).
206 CHROMATOGRAPHY [9]
concentrated in situ. Second, high loadings for some solutes can be
achieved; i.e., a loading up to 50 mg of ribonuclease did not noticeably
affect the resolution on an analytical column (25 x 0.4 m, Si 100 A,
octadecyl, dp 5 /~m). 5° Third, the sample capacity of large proteins is
greater on macroporous than on mesoporous reversed phases, although at
comparable loadings similar elution times are observed with otherwise
identical chromatographc conditions.
Studies by Lewis et al., TM Regnier et al., 3 and Karger and Hearn 2°'21
have delineated some of the options for high-capacity supports for the
reversed-phase HPLC of large proteins. High coverage, 25- to 50-nm nomi-
nal pore diameter, silica bonded with butyl or octyl phases and end-
capped, in particular, appear to be very useful supports for the chroma-
tography of semipreparative quantities of proteins exhibiting good
recoveries, high sample capacities, and good resolution. For stationary
phases bonded to the same alkyl chain density, the chain length has been
found to have only a small influence on relative retention but does affect
recoveries. Several studies have implied that the resolution and recover-
ies are higher with short-chain alkyl ligands. For example, Rubinstein 5°
has favored LiChrosorb RP-8 over RP-18 or RP-2 for the fractionation of
proteins of Mr > 30,000.
The changes in elutlon behavior noted in these and other early studies
can now be more readily accommodated in terms of the extent of carbon
coverage of the stationary phase. What is apparent is that selectivity is
influenced by the chemical characteristics of the parent silica matrix and
the bound ligand: cyanopropyl-, octyl-, and octadecylsilicas, for example,
show noticeable selectivity differences to the diphenyl- or cyclohexyl-
silicas. Sample loading also affects relative retention of proteins with
alkylsilicas, but, interestingly, overloading does not result in total loss of
resolution as may occur in conventional size-exclusion or ion-exchange
chromatography. Similarly, column length does not appear to be impor-
tant for protein separation. A 10-fold increase in column length increased
the gradient elution resolution of human growth hormone (Mr = 22,000)
and the human growth hormone (Mr = 2 0 , 0 0 0 ) 19 by only 11%, and similar
results have been observed 3 for the separation of bovine serum albumin
and ovalbumin. In contrast, relative retention of proteins on alkylsilicas is
directly dependent on pore size, and, in particular, on the accessible
bonded surface area of the stationary phase. As a consequence, those
alkylsilicas that currently afford the greatest utility for protein separation
all exhibit surface areas in the range 60-100 m2/g with ligand densities of
2.7-2.5/xmol of alkyl chain/m2. For many commercial reversed phases

50 M. Rubinstein, Anal. Biochem. 98, 1 (1979).

[9] HPLC: REVERSEDPHASE 207

with nominal pore diameters of 80-100 A, it is probable that many protein

molecules are entropically excluded from the pore interior, with conse-
quent impairment of sample capacity.4 Protein solutes that do interact
strongly with small-pore silicas may not be recovered in good yield. It is
also apparent from two studies3,19 that wide-pore (200-500/~) silicas are
not necessarily suitable for every case of protein separation.
Data obtained3,4,19 with polypeptide hormones and protein standards
on reversed phases of different pore diameters, but identical alkyl chain
length, have indicated that the resolving power of the system decreases as
hydrophobicity of the ligand increases, but that the retention dependence
on solvent content is essentially independent of pore size at low sample
loadings, i.e., that the solute S values are constant. These results stress
the importance of choosing elution conditions that generate adequate se-
lectivity factors [Aa/a; cf. Eq. (5)] if resolution between two proteins of
similar hydrophobicity is to be achieved. For these reasons, attention
must be given to the potential available for optimization of resolution
arising from the manipulation of secondary chemical equilibria, the tem-
perature, the flow rate, the eluotropic properties of the organic solvent
modifier, and the extent of nonpolar ligand coverage on the stationary
phase.

Influence of Flow Rate and Temperature

Because of the slower rates of diffusion of macroglobulins, lowering
the mobile-phase velocity generally improves the resolution. The flow
rate and eluent composition had a significant influence on the column
efficiency for bovine serum albumin, for example, on an Ultrasphere-octyl
(dp 5 tzm, 10 nm) column) ~Three solvent systems (15 rnM HaPO4, 0.1%
HCOOH, and 500 mM HCOOH-400 mM pyridine) were compared in this
study, using a 1-propanol gradient. Under all flow rate conditions the
formic acid-pyridine buffer showed the highest efficiency, the phosphoric
acid system the lowest. These results are in accord with the expected
differences in the molarity of the buffer. More interesting, however, was
the finding that the differences in plate number (N) were accentuated as
the flow rate was decreased; e.g., the plate numbers for phosphoric acid
system showed little affect when the flow rate was decreased from 75 to
15 ml/hr, whereas the formic acid-pyridine system showed a threefold
increase.
Although column efficiency for polypeptides and proteins appears to
increase at higher temperature,21'z5the lability of many biologically func-
51 B. N. Jones, R. J. Lewis, S. Paabo, K. Kojima, S. Kimura, and S. Stein, J. Liq. Chroma-
togr. 3, 1373 (1980).
208 CHROMATOGRAPHY [9]

t.l
e~

•~.~ . ~
. ~ ~ ~
~ . ~
m

m
>
. ~'~
u2

N _ ~'- ~ ~
m ~zz ~
2
- ~ ~ ~ ~

~i ~ ~
Nu

< . r...)
.~.~
<

r~ ~

M
g
0

ua
oo
ua r~ oo ® _ ~r,.)

m
M
u,l

< z~ ~
[9] HPLC: REVERSEDPHASE 209

*:t

¢,q

o_., ~z~ ~i~

•~ ~
z

~ ~..~ a ~z~-d~ 4,sz

d~ 0 0 I0~" I I

N ~N ~ N

2~ ~ 0
o ~
F,

: zo ~.~
210 CHROMATOGRAPHY [9]

¢;

O~go~

~ Z F

•~.~ e
~ - ~- Z ~
"~ .fi ~.~

~r,.)

.r, ~-~, ~ - ~ ' ~ ~

~5

•~ .~
z ~s ~o~4g
M
.~,
O

r~a~

--~-- o. q ,-2. q --. ~ "

. ~ ;:I:I
e~
r~

_=
0
,.a ~. <>-~:~ ~-Z< :s.< :s.
~-~
E

e~ a~ ~ "~ "~,

~.-~ ~ ~ .¢ ~ o ~.~.~
[9] HPLC: REVERSEDPHASE 211

tional proteins may limit the utility of this approach for improving resolu-
tion to temperatures below 50 ° . In some cases, low temperatures, e.g., 5 ° ,
may be essential for preservation of biological activity. Although elevated
temperatures are normally thought to reduce ionic interactions between
the solutes and the stationary phase, temperature-induced changes in
secondary equilibria and protein dynamics in the mobile phase will also
have a significant effect on retention and resolution.

Comments
Some of the potential of reversed-phase HPLC in the purification of
polypeptides and proteins can be demonstrated from the selected examples
summarized in Table I I : 1-79 It should be emphasized that most of the

52 j. Meienhofer, T. F. Gabriel, J. Michalewsky, and C. H. Li, in "Peptides 78," p. 243.

Wroclaw Univ. Press, Wroclaw, Poland, 1979.
53 M. Rubinstein, S. Stein, L. D. Gerber, and S. Udenfriend, Proc. Natl. Acad. Sci. U.S.A.
74, 3052 (1977).
M. T. W. Hearn and B. Grego, J. Liq. Chromatogr. (in press).
55 M. T. W. Hearn and B. Grego, submitted.
56 H. P. J. Bennett, C. A. Browne, and S. Solomon, Proc. Natl. Acad. Sci. U.S.A. 78, 4713
(1981).
57 W. Richter and P. Schwandt, J. Neurochem. 36, 1279 (1981).
58 E. C. Nice, M. Capp, and M. J. O'Hare, J. Chromatogr. 147, 413 (1979).
~9 W. S. Hancock, H. J. Pownall, A. M. Gotto, and J. T. Sparrow, J. Chromatogr. 216, 285
(1981).
60 S. J. Skinner, B. Grego, M. T. W. Hearn, and C. G. Liggins, J. Chromatogr. (in press).
61 A. FaUon, R. A. Lewis, and K. D. Gibson, Anal. Biochem. 110, 318 (1981).
62 W. A. Schroeder, J. B. Shelton, J. R. Shelton, and D. Powers, J. Chromatogr. 174, 385
(1979).
63 S. Terabe, H. Nishi, and T. Ando, J. Chromatogr. 212, 293 (1981).
64 M. Green and K. H. Brackmann, Anal. Biochem. 124, 209 (1982).
63 M. A. Anzano, A. B. Roberts, J. M. Smith, L. C. Lamb, and M. B. Sporn, Anal.
Biochem. 125, 217 (1982).
66 j. j. L'Italien and J. L. Strickler, Anal. Biochem. 127, 198 (1982).
67 j. Spiess, J. Rivier, C. Rivier, and M. Vale, Proc. Natl. Acad. Sci. U.S.A. 78, 6517 (1981).
68 S. D. Power, M. A. Lochrie, and R. O. Poyton, J. Chromatogr. 266, 585 (1983).
69 S. W. Herring and R. K. Enns, J. Chromatogr. 266, 249 (1983).
7o D. Voskamp, C. Olieman, and H. C. Bayerman, Recl. Tray. Chim. Pays-Bas 99, 105
(1980).
71 U. Certa and G. J. Ehrenstein, Anal. Biochem. 118, 147 (1981).
72 G. J. Xu, E. Hannappel, J. Morgan, J. Hempstead, and B. L. Horecker, Proc. Natl.
Acad. Sci. U.S.A. 79, 4006 (1982).
73 K. Krupen, B. A. Araneo, L. Brink, J. A. Kapp, S. Stein, K. J. Weider, and D. R. Webb,
Proc. Natl. Acad. Sci. U.S.A. 79, 1254 (1982).
74 j. B. McManon, J. G. Farrelly, and P. T. Iype, Proc. Natl. Acad. Sci. U.S.A. 79, 456
(1982).
212 CHROMATOGRAPHY 110]

polypeptides or proteins successfully purified by this technique have rela-

tively low molecular weights and conformation and aggregation states
that are not strongly dependent on intramolecular interactions induced by
the elution conditions. H o w e v e r , the usefulness of reversed-phase H P L C
as an integral part of the purification of proteins in their native form with
unaltered physiological properties is increasingly being established. Be-
cause of the versatility of the technique, membrane and transmembrane
proteins can be resolved in the presence and in the absence of detergents
(see also this volume [18]). At this stage, the greatest strength that these
procedures currently offer to the protein chemist is in the microprepara-
tive isolation of proteins or protein fragments required for structural eluci-
dation. As a consequence, reversed-phase methods are increasingly re-
placing conventional means as an efficient prelude to structural studies.

Acknowledgment
The support of the National Health and Medical Research Council of Australia is grate-
fully acknowledged.

75K. J. Wilson, M. W. Berchtold, P. Zumskin, S. Klause, and G. J. Hughes, in "Methods in

Protein Sequence Analysis" (M. Elizinga, ed.).
76j. C. Chan, N. G. Seidah, C. Gianoulakis, A. Belanger, and M. Chretien, J. Clin. Endo-
crin. Metab. 51, 364 (1980).
77A. Marquardt and G. J. Todaro, J. Biol. Chem. 2,57, 5220 (1982).
78M. T. W. Heam, P. A. Smith, and A. K. Mallia, Biosci. Rep. 2, 247 (1982).
79G. E. Gerber, R. J. Anderegg, W. C. Herlihy, C. P. Gray, K. Biemann, and H. G.
Khorana, Proc. Natl. Acad. Sci. U.S.A. 76, 227 (1979).

[10] H i g h - P e r f o r m a n c e Liquid Affinity Chromatography

B y PER-OLOF LARSSON

High-performance liquid affinity chromatography combines with the

remarkable specificity of (bio)affinity techniques with the efficiency, sen-
sitivity, and speed o f operation of H P L C techniques.

Preparation of Adsorbents
The affinity adsorbents described herein are based on porous silica,
which has excellent mechanical properties. Unfortunately, silica is unsta-
ble toward alkaline conditions and therefore should be used only for brief
212 CHROMATOGRAPHY 110]

polypeptides or proteins successfully purified by this technique have rela-

tively low molecular weights and conformation and aggregation states
that are not strongly dependent on intramolecular interactions induced by
the elution conditions. H o w e v e r , the usefulness of reversed-phase H P L C
as an integral part of the purification of proteins in their native form with
unaltered physiological properties is increasingly being established. Be-
cause of the versatility of the technique, membrane and transmembrane
proteins can be resolved in the presence and in the absence of detergents
(see also this volume [18]). At this stage, the greatest strength that these
procedures currently offer to the protein chemist is in the microprepara-
tive isolation of proteins or protein fragments required for structural eluci-
dation. As a consequence, reversed-phase methods are increasingly re-
placing conventional means as an efficient prelude to structural studies.

Acknowledgment
The support of the National Health and Medical Research Council of Australia is grate-
fully acknowledged.

75K. J. Wilson, M. W. Berchtold, P. Zumskin, S. Klause, and G. J. Hughes, in "Methods in

Protein Sequence Analysis" (M. Elizinga, ed.).
76j. C. Chan, N. G. Seidah, C. Gianoulakis, A. Belanger, and M. Chretien, J. Clin. Endo-
crin. Metab. 51, 364 (1980).
77A. Marquardt and G. J. Todaro, J. Biol. Chem. 2,57, 5220 (1982).
78M. T. W. Heam, P. A. Smith, and A. K. Mallia, Biosci. Rep. 2, 247 (1982).
79G. E. Gerber, R. J. Anderegg, W. C. Herlihy, C. P. Gray, K. Biemann, and H. G.
Khorana, Proc. Natl. Acad. Sci. U.S.A. 76, 227 (1979).

[10] H i g h - P e r f o r m a n c e Liquid Affinity Chromatography

B y PER-OLOF LARSSON

High-performance liquid affinity chromatography combines with the

remarkable specificity of (bio)affinity techniques with the efficiency, sen-
sitivity, and speed o f operation of H P L C techniques.

Preparation of Adsorbents
The affinity adsorbents described herein are based on porous silica,
which has excellent mechanical properties. Unfortunately, silica is unsta-
ble toward alkaline conditions and therefore should be used only for brief
[10] HPLC: AFFINITY 213

periods with buffers of pH 8 o r a b o v e . 1 An alternative packing material

with lower pressure resistance but better stability toward hydrolysis is
exemplified by cross-linked hydroxymethyl methacrylate. 2
Porous silica is commercially available in several pore sizes (60-4000
/~). Ordinarily, HPLC material is based on silica with 60- or 100-A pores,
a size sufficient for separating small molecules. In affinity chromatogra-
phy larger pores are needed since the separated molecules or the silica-
bound molecules are of high molecular weight. In order to achieve unhin-
dered diffusion in the pores, the pore size must be considerably larger
than the chromatographed molecules3,4; a pore size of 300-1000 A should
be adequate for most applications. When separating very large entities
such as immune complexes, the 4000-~ pore size would be advantageous.
Native silica contains acidic silanol groups, among other surface
groups, that cause strong, often irreversible, adsorption of proteins. 5
Therefore an important task when preparing affinity adsorbents from sil-
ica is to mask such binding sites that would otherwise interfere with the
reversible binding process. In the present case this is accomplished by
coating the silica with glycerylpropyl groups (diol-silica; Fig. 1), thereby
producing a hydrophilic and nonionic surfaceY
Coupling to epoxy-silica is the simplest of the three routes outlined in
Fig. 1. The poor reactivity makes it best suited for strong nucleophiles or
for stable molecules that will withstand rather drastic conditions. The
tresyl chloride method is efficient and suitable for coupling of all types of
molecules containing primary amino groups. The aldehyde method in-
volves several steps but generally gives good results, although reduction
with NaBH4 may be deleterious. Both the tresyl chloride method and the
aldehyde method may use commercially available diol-silica as a starting
material. 6

Preparation of Epoxy-Silica (Fig. I)

Silica (10 g) with 1000-A pores 7 is placed in a 500-ml three-necked
flask. The flask is connected to a vacuum line and heated to 150° for 4 hr to
i K. K. Unger, "Porous Silica: Its Properties and Use as Support in Column Liquid Chro-
matography." Elsevier, Amsterdam, 1979.
2 j. Turkovfi, K. Blahfi, and K. Adamamov~i, J. Chromatogr. 236, 375 (1982).
3 P.-O. Larsson, M. Glad, L. Hansson, M.-O. M~nsson, S. Ohlson, and K. Mosbach, in
"Advances in Chromatography" (J. C. Giddings, E. Grushka, J. Cazes, and P. R. Brown,
eds.), p. 41. Dekker, New York, 1983.
4 R. R. Walters, J. Chromatogr. 249, 19 (1982).
5 S. H. Chang, K. M. Gooding, and F. E. Regnier, J. Chromatogr. 120, 321 (1976).
6 For example, LiChrosorb Diol (100-A pore size), manufactured by E. Merck, Darmstadt,
FRG.
7 LiChrospher Si 1000 (10-/.~m particles; surface area 20 m2/g) from E. Merck, Darmstadt,
FRG.
214 CHROMATOGRAPHY [ 10]

Z
0 II
z ..~ z
0 0
z z '~
0 0 z~ ~..

)_. i
_,~_ -05-,
o ~ - o

, El
-b3-

1- ~ o ~_~
0-0 ~: , z
(~ "r "1-
i" -r O-O -
o-o~ 7- :~
n,- (.D ,o
o ! _, o ~
--r u.l -r
~d

'
-~- ~ ~ -~-
-i- -r ~
ILl 0-0
I
0
/I
OT
"° ~ o~
r
z o

<- 0

, 2
0
+ [...
g 0
[10] HPLC: AFFINITY 215

remove adsorbed moisture. After cooling to 50-100 °, the vacuum is dis-

connected and the flask is immediately fitted with a reflux condenser and
is provided with a CaCI2 drying tube and stirrer. Sodium-dried or molecu-
lar sieve-dried toluene (150 ml) is added, followed by 2 ml of y-glycidoxy-
propyltrimethoxysilane 8 and 0.05 ml of triethylamine. 9 The mixture is
refluxed with stirring for 16 hr. The epoxy-silica obtained is washed on a
glass filter with toluene, acetone, and ether and dried under reduced
pressure.
Well-coated silica should contain about 2.5/~mol of epoxy groups per
square meter. The epoxy group content is determined by titration of
hydroxyl ions released by the reaction of epoxy groups with thiosulfate3:
Epoxy-silica (25-100 mg) is carefully slurried in 2 ml of water, the pH is
adjusted to 7.0, and 1 ml of 3 M sodium thiosulfate at pH 7 is added to
start the reaction; 100 mM HCI is provided, preferably with an automatic
titrator, to maintain the pH at 7.0 for 1 hr. The consumption of HC1 is a
direct measure of the epoxy group content.

Preparation of Glycerylpropyl-Silica, Diol-Silica (Fig. 1)

Epoxy-silica (5 g) is suspended in 500 ml of 10 mM H2SO4 and heated
to 90° for 1 hr. The glycerylpropyl-silica (diol-silica) obtained is filtered;
washed sequentially with water, ethanol, and ether, and dried under re-
duced pressure.

Preparation of Aldehyde-Silica (Fig. 1)

Diol-silica (5 g) is suspended in 100 ml of acetic acid-water (90 : 10),
5 g of sodium periodate is added, and the suspension is stirred for 2 hr at
room temperature. The resultant aldehyde-silica is washed sequentially
with water, ethanol, and ether and dried under reduced pressure.

Preparation of Tresyl-Silica (Fig. 1)

Tresyl chloride (2,2,2-trifluoroethanesulfonyl chloride), which is used
in the activation process, is readily hydrolyzed by water; dry solvents
(molecular sieve 4/~) and a minimum of exposure to humid conditions are
imperative. The activated silica, on the other hand, is stable in aqueous
solutions at low pH (pH 3). 1°
8 Dow Coming Z-6040, Midland, MI.
9 When activating silica with larger surface area, e.g., LiChrospher Si 300 and LiChrospher
Si 100 (area --~ 250 m2/g), five times the amount of silane and amine was used.
lo The tresyl chloride activation method is described by Nilsson and Mosbach in this volume
[2].
216 CHROMATOGRAPHY [ 1O]

Diol-silica (2 g; 1000-A pores) is washed three times with 50 ml of dry

acetone on a glass filter. The wet silica (5 g) is immediately transferred to
a vessel containing 2.5 ml of dry acetone and 200/xl of dry pyridine. The
suspension is cooled to 0°, and 65 /xl of tresyl chloride H is added with
vigorous stirring. As soon as the addition is made, the stirring speed is
slowed to avoid generation of "fines." After 20 min the silica is washed
on a glass filter with acetone, acetone-5 mM HCI (I : I), 5 mM HC1, and,
finally, acetone followed by drying under reduced pressure. If the acti-
vated silica is to be used within a day, the last acetone wash and the
drying process are unnecessary. The degree of activation may be deter-
mined by elemental analysis of sulfur.

Coupling to Epoxy-Silica (Fig. 1)

Coupling to epoxy-silica is a comparatively slow process. As high a
concentration as possible of the ligand should therefore be attempted. The
method is exemplified here by coupling of a spacer-provided NAD, Nr-[N -
(6-aminohexyl)carbamoylmethyl]-NAD.12 Epoxy-silica (100-A pores; 1 g)
is suspended in 3 ml of 0.1 M sodium pyrophosphate at pH 8, containing
20 mg of the NAD analog. The suspension is gently shaken for 5 days at
room temperature. The NAD-silica obtained is filtered, washed with wa-
ter, and suspended in 100 ml of dilute H2SO4 (pH 2) and heated to 50° for 4
hr to hydrolyze excess epoxy groups to the diols. The NAD-silica is
finally washed with water, pyrophosphate buffer, water, ethanol, and
ether and dried under reduced pressure. The NAD-analog content may be
determined from UV spectra to be 1/zmol per gram of silica.

Coupling of Alcohol Dehydrogenase to Tresyl-Silica (Fig. 1)

Tresyl-silica (1000-/~ pores, 10/zm, 0.7 g dry weight) is suspended in 2
ml of 0.4 M sodium phosphate at pH 7.0, containing 0.2 M isobutyramide
and 2 mM NADH. The suspension is deaerated for 5 min. Horse liver
alcohol dehydrogenase, B 15 mg in 2 ml, is added, and the coupling is
allowed to proceed for 20 hr at room temperature. Remaining tresyl
groups are subsequently removed by treatment with 0.2 M Tris-HC1 con-
taining 1 mM dithioerythritol at pH 8 for 1 hr. The gel is then washed
extensively with 0.1 M sodium phosphate at pH 7.5, containing 0.5 M
,1 Obtained from Fluka, Buchs, Switzerland.
12 K. Mosbach, P.-O. Larsson, and C. R. Lowe, this series, Vol. 44, p. 859. The compound
is available from Sigma, St Louis, MO.
13 Boehringer, Mannheim, FRG. The commercial preparation was dialyzed against 0.075 M
sodium phosphate at pH 7.9, and cleared by centrifugation.
[10] HPLC: AFFINITY 217

NaC1 and 1 mM dithioerythritol, followed by washing with the same

solution but without NaCI. The alcohol dehydrogenase-silica may be
stored at 4 ° until used. Enzyme content can be determined by amino acid
analysis or from UV spectra and is generally of the order of 20 mg per
gram of dry silica, i.e., 100% coupling yield.

Coupling of Concanavalin A to Aldehyde-Silica (Fig. 1)

Aldehyde-silica (1000-,~ pores; 10/.tm; 2 g dry weight) is suspended in
10 ml containing 1 mM CaCI2, 1 mM MnCI2, and 0.1 M sodium phosphate
at pH 6.0. The suspension is deaerated under reduced pressure for 5 min.
Concanavalin A (150 mg) in 20 ml of the above solution is added, followed
by 75 mg of NaCNBH3 (to reduce the Schiff base formed to a secondary
amine). The suspension is stirred very gently at 4° overnight.14 The solu-
tion is adjusted to pH 8.0, and 50 mg of NaBH4 is added in portions during
30 min (to reduce remaining aldehyde groups to their alcohols). After 1 hr
the concanavalin A-silica is washed on a glass filter with the above buf-
fered solution supplemented with 0.5 M NaCI. The concanavalin A-silica
(about 60 mg per gram of silica; determined spectrophotometrically) may
be stored at 4 ° until used.

Determination of Silica-Bound Ligands

Ligand density often may be determined spectrophotometrically by
one or both of the following methods:
Method 1. Ligand-substituted silica (10-200 mg) is carefully sus-
pended in 2.5 ml of saturated aqueous sucrose solution in a spectropho-
tometer cell. The cell is repeatedly evacuated to ensure complete filling of
the pores. Spectra are recorded using a proper reference, and the ligand
concentration is calculated. The sucrose solution has approximately the
same refractive index as the silica. Light scattering from silica particles
suspended in saturated sucrose is therefore negligible, allowing meaning-
ful spectra to be obtained. 3
Method 2. Ligand-substituted silica (1-100 mg) is heated to 60° for 30
min in 1 ml of 1 M NaOH. To the solubilized silica gel are added 4 ml of
0.1 M sodium phosphate at pH 7, 1 ml of 1 M HCI, and water to a final
volume of 10 ml. A reference cuvette is prepared by parallel treatment of
a blank silica. The spectrum is recorded, and the ligand density is calcu-
lated. The neutralization step provides a defined pH and avoids alkaline
attack on the spectrophotometer cells.
14Coupling at room temperature for 2 hr will give approximately the same result.
218 CHROMATOGRAPHY [ 1O]

Packing of Columns
Stainless steel columns are used that are 50 or 100 mm in length, ]-in.
o.d., 5-mm i.d., and provided with compression fittings. 15 A 50-mm
column has an inner volume of 1 ml; when packed, it contains 0.44 g of
silica (1000-]~ pores) and has a liquid-phase volume of 0.80 ml (pore
volume + interstitial volume).
Packing of 50-mm Column. Derivatized silica (0.7-I g) is suspended in
5 ml of buffer containing 50% sucrose. The suspension is treated for 5 min
in an ultrasonic bath and then immediately poured into the packing de-
vice. The packing device is prepared from a 250-mm, i-in. stainless steel
tube, connected at one end to the column to be packed via a bored-
through union. After filling, the other end of the packing device is con-
nected to a standard HPLC pump, and the pumping is started immedi-
ately. The pump should deliver buffer at a pressure of about 4000 psi.
After 30 min of pumping, packing is considered to be complete and the
column may be disconnected and fitted with the remaining end piece.

Chromatographic Procedures
Normal HPLC equipment may be used, generally at room tempera-
ture.
The flow rate is usually maintained within 0.1-3 ml/min for a standard
5-mm i.d. column. A typical value is 1 ml/min, which gives a pressure
drop over a 50-mm column loaded with 5-/zm particles of about 500 psi.

Applications of High-Performance Liquid Affinity Chromatography

Table I gives a summary of applications with high-performance liquid
affinity chromatography.

Separation of Dehydrogenases on Silica-Bound AMP

AMP is an inhibitor, competitive with NAD for many dehydro-
genases, and silica-bound AMP might, therefore, be expected to resolve
this class of enzymes. AMP-silica has been prepared by coupling a
spacer-provided AMP analog, N6-(6-aminohexyl)-AMP, to tresyl-acti-
vated silica. 16 When a mixture of serum albumin, liver alcohol dehydro-
genase, and lactate dehydrogenase was injected into a column prepared
with this analog, albumin eluted unretarded, whereas both dehydro-

15 hetp, Macclesfield, Cheshire, U.K.

16 P.-O. Larsson and K. Mosbach, Biochem. Soc. Trans. 9, 285 (1981).
[10] H P L C : AFFINITY 219

TABLE I
HIGH-PERFORMANCE LIQUID AFFINITY CHROMATOGRAPHY WITH
SILICA-BOUND LIGANDS

Silica-bound ligands Interacting molecules

Nucleotides
AMP ~ Alcohol dehydrogenase
NAI~ Lactate dehydrogenase
Dyes
Cibachron Blue F3G-A c Dehydrogenases, kinases, others
Procion Blue MX-R d Lactate dehydrogenase
Procion Green MX-5BR a Hexokinase
Procion Brown MX-5BR d Tryptophanyl-tRNA synthetase
Procion Yellow H-A d Carboxypeptidase
Procion Red H-8BN d Alkaline phosphatase
Antibodies
Antihuman serum albumin a Albumin
Antibovine insulin" (monoclonal) Insulin
Proteins
Soybean trypsin inhibitor s Chymotrypsin
Bovine serum albuming oL-Amino acids
Alcohol dehydrogenase h Nucleosides, nucleotide dyes
Concanavalin A i Carbohydrates, glycoenzymes
Protein Ai Immunoglobulins
Miscellaneous
Boronic aci& Nucleotides, carbohydrates
Glucoseamine t Concanavalin A
L-Phe-D-Phe-OCH3 m (methacrylate support) Pepsin
Thymine n (methacrylate support) Nucleic acid derivatives

" S. Ohlson, L. Hansson, P.-O. Larsson, and K. Mosbach, FEBS Lett. 93, 5 (1978).
b Larsson et al)
c C. R. Lowe, M. Glad, P.-O. Larsson, S. Ohlson, D. A. P. Small, T. Atkinson,
and K. Mosbach, J. Chromatogr. 215, 303 (1981).
d D. A. P. Small, T. Atkinson, and C. R. Lowe, J. Chromatogr. 216, 175 (1981).
• j. R. Sportsman and G. Wilson, Anal. Chem. 52, 2013 (1980).
f V. Kasche, K. Buchholz, and B. Galunsky, J. Chromatogr. 216, 169 (1981).
g S. Allenmark, Chem. Scr. 20, 5 (1982).
h K. Nilsson and P.-O. Larsson. 19
i A. Borchert et alfl 7
J S. Ohlson and U. Niss, Swedish Patent application, 8104876-1 (1981).
k M. Glad, S. Ohlson, L. Hansson, M.-O. M~nsson, and K. Mosbach, J. Chroma-
togr. 200, 254 (1980).
t R. R. Walters. 4
m j. Turkovfi et al. 2
n y . Kato, T. Seita, T. Hashimoto, and A. Shimizu, J. Chromatogr. 134, 204
(1977).
220 CHROMATOGRAPHY [ 10]

genases were strongly adsorbed. A specific elution method, ternary com-

plex formation, released the dehydrogenases. Addition of NAD and oxa-
late eluted lactate dehydrogenase, whereas addition of NAD and pyrazole
eluted alcohol dehydrogenase. The NAD concentration used, 0.1 mM,
was in itself too weak to effect desorption.

Separations with Antibody-Silica

The specific interaction between an antibody and its antigen allows
very effective separations. In illustration, a mixture of two serum al-
bumins, human and bovine, was injected in a 5 x 50 mm antihuman serum
albumin-silica column (particle size, I0/xm; pore size, 60 A). The bovine
albumin was eluted unretarded, whereas human albumin was strongly
adsorbed. To elute the human albumin, drastic conditions in the form of a
pulse of 0.2 M glycine-HC1 at pH 2.2 were required. By the pulse elution
technique, the separation process could be managed within 5 min. The
acid eluent had no apparent adverse effects, since the procedure could be
repeated more than 20 times without any observable decrease in perfor-
mance of the system.

Separations on Concanavalin A-Silica

The lectin concanavalin A was coupled to silica (60 mg of lectin per
gram of silica), and its specificity for a-D-mannose and a-D-glucose was
used for the separation and purification of carbohydrates and glycopro-
teins. 17A sample of commercial horseradish peroxidase (4.1 mg in 4.1 ml
of buffer) was injected in a concanavalin A column (5 x 50 mm;
particle size, 10/zm; pore size, 1000/~) operated at room temperature at a
flow rate of 1 ml/min. Approximately 50% of the total protein, but less
than 2% of the peroxidase, passed the column unretarded.
A pulse of the competitively acting a-methylglucoside was subse-
quently used to elute pure peroxidase. The separation illustrates that even
a small analytical column can be used for preparative purposes, at least in
the final stages of a purification process. Aside from a twofold purification
in 20 min, the enzyme became concentrated in the process, 90% of the
enzyme being collected at four times its original concentration.

Reversed Affinity Chromatography with Silica-Bound

Alcohol Dehydrogenase
Enzymes are seldom used as adsorbents for compounds of low molec-
ular weight in conventional affinity chromatography.18 Although several
17 A. Borchert, P.-O. Larsson, and K. Mosbach, J. Chromatogr. 244, 49 (1982).
is K. Das, P. Dunnill, and M. D. Lilly, Biochim. Biophys. Acta 397, 277 (1975).
[10] HPLC: AFFINITY 221

ADP
AMP

0.02

E
0
¢D
cJ 0.01
ADPR

0.00
i i I I I
0 1 2 :3 4 5
MIN
FIG. 2. Separation of adenine nucleotides on alcohol dehydrogenase-sflica. Column:
50 × 5 mm. Content: Silica (1000/~) with 12 mg of horse liver alcohol dehydrogenase per
gram of silica coupled after tresyl chloride activation. Binding site concentration: 120/~M.
Sample: 0.5 nmol of AMP, 0.2 nmol of ADP and 1 nmol of ADP-ribose in 15/zl of sodium
phosphate at pH 7.5. Mobile phase: 0.25 M sodium phosphate at pH 7.5, containing 1 ~ M
ZnSO4. Flow rate: 1.0 ml min-L Reproduced from Nilsson and Larsson, ~9with permission.

difficulties apply, the basic problem is that the capacity, at least on a

weight basis, is very limited in such reversed affinity systems. This obvi-
ously has consequences in preparative applications although problems
may also arise in an analytical context. For example, poor resolution due
to overloading may occur. However, in combination with the HPLC tech-
nique, with its emphasis on resolution and sensitivity, the reversed affin-
ity chromatography concept could be useful.
An example is given in Fig. 2 of a mixture of adenine nucleotides
chromatographed on a horse liver alcohol dehydrogenase silica column.19
The separation pattern obviously reflects the strength of the interaction
between the silica-bound enzyme and the chromatographed nucleotide. A
quantitative relationship between the retention (k') of a substance, the
dissociation constant, Kd, for the binary complex involved, and the ligand
density, [HLADH], has been verified for alcohol dehydrogenase-silica~9:
k' = [HLADH]/Kd. The capacity factor, k', is commonly used in HPLC
contexts and refers to the retention of a substance in column volumes. It
may be experimentally calculated from k ' = ( V e - V o ) / V o , where Ve is the
elution volume for the compound in question and Vo is the elution volume
for a nonadsorbed, nonexcluded compound. If the ligand density and the
dissociation constant are known, the chromatographic behavior could be
predicted from the equation. Conversely, knowledge of ligand density and
chromatographic behavior allow calculation of the Kd constant. Interest-
19 K. Nilsson and P.-O. Larsson, Anal. Biochem. 133 (1983).
222 CHROMATOGRAPHY [ 1O]

TABLE II
CHROMATOGRAPHIC DATA FOR ADENINE NUCLEOTIDES ON ALCOHOL
DEHYDROGENASE-SILICAa

KO (/~M)

Compound Spacer composition k'-value Calculated Literature

AMP 2.15 134 70

N6-(6-Aminohexyl)- HzN(CH2)6-- 5.6 51 32
AMP
N6-(2-Aminoethyl)- H2N(CH2)2-- 0.28 1030
AMP
NS-Carboxymethyl - HOOCCH2-- 2.7 107
AMP
ADP 0.19 1500 390
ATP 0.06 5000 --
ADP-ribose 7.5 38 18
NAD 2.4 120 96
N6-[N-(6-Amino- H2N(CH2)6NHCOCH2-- 1.56 185 --
hexyl)carbamoyl-
methyl]-NAD
N6-[N-(2-Amino- H2N(CHE)2NHCOCH2-- 0.40 720
ethyl)carbamoyl-
methyl]-NAD

a The adsorbent contained 21 mg of horse liver alcohol dehydrogenase per gram of silica
(1000/~; 10/zm).

ingly, it has been shown that alcohol dehydrogenase immobilized to silica

by the tresyl chloride method preserves many of its native properties? °
Thus, dissociation constants referring to silica-bound enzyme may be
valid also for the free enzyme. This obviously opens up the possibility of
rapid screening of a large number of compounds in order to determine
their Kd. Table II exemplifies this idea: the k' values, the calculated
dissociation constants, and the corresponding literature data are pre-
sented for several adenine nucleotides of potential interest as affinity
ligands. The chromatographically derived dissociation constants are gen-
erally higher than those presented previously. These observations may be
due to chromatography being carded out with high ionic strength buffers,
whereas data from the literature refer mainly to weak buffers. Table II
also indicates that Kd values for very weakly retained substances are less
reliable, probably because of the greater imprecision in the determination
of k' for these compounds. Nevertheless, Table II reveals interesting
affinity differences for the several analogs--differences that may be ex-
[11] OPTIMALpH FOR ION EXCHANGERS 223

plained by the attraction or the repulsion between the N 6 substituents and

certain amino acid residues on the enzyme surface.

Comments
High-performance liquid affinity chromatography as described here
allows very rapid separation, making the technique suitable for analytical
as well as micropreparative purposes. Moderate scaling up should give
preparative systems with high capacity. The technique provides a con-
venient means of obtaining quantitative information about biological com-
plexes.

Acknowledgment
Discussions with S. Birnbaum, A. Borchert, M. Glad, and L. Hansson are gratefully
acknowledged.

[11] O p t i m a l p H Conditions for Ion E x c h a n g e r s on

M a c r o p o r o u s Supports
By J A M E S S. RICHEY

Two relatively recent techniques are available for determining the

most suitable pH conditions for performing ion-exchange chromatogra-
phy with macroporous media: the electrophoretic titration curve ap-
proach and the retention mapping method. The ready determination of
this parameter should provide optimal conditions for truly high-perfor-
mance ion-exchange chromatography of biologically active macromole-
cules. However, both methods are known to be effective only with mac-
roporous media; whether these approaches are reliable or have predictive
value for other HPLC or conventional ion-exchange materials is not
tested (see also [8]).

Macroporous Ion-Exchange Matrix

Recently, a major change in the polymer chemistry industry has al-
lowed the production of a rigid chromatographic support media with very
large pore structure. The macroporous structure of these beads allows
rapid movement of both fluid and large molecules through packed beds
with very low resistance. Another factor, the monodispersity or very

Copyright © 1984by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
[11] OPTIMALpH FOR ION EXCHANGERS 223

plained by the attraction or the repulsion between the N 6 substituents and

certain amino acid residues on the enzyme surface.

Comments
High-performance liquid affinity chromatography as described here
allows very rapid separation, making the technique suitable for analytical
as well as micropreparative purposes. Moderate scaling up should give
preparative systems with high capacity. The technique provides a con-
venient means of obtaining quantitative information about biological com-
plexes.

Acknowledgment
Discussions with S. Birnbaum, A. Borchert, M. Glad, and L. Hansson are gratefully
acknowledged.

[11] O p t i m a l p H Conditions for Ion E x c h a n g e r s on

M a c r o p o r o u s Supports
By J A M E S S. RICHEY

Two relatively recent techniques are available for determining the

most suitable pH conditions for performing ion-exchange chromatogra-
phy with macroporous media: the electrophoretic titration curve ap-
proach and the retention mapping method. The ready determination of
this parameter should provide optimal conditions for truly high-perfor-
mance ion-exchange chromatography of biologically active macromole-
cules. However, both methods are known to be effective only with mac-
roporous media; whether these approaches are reliable or have predictive
value for other HPLC or conventional ion-exchange materials is not
tested (see also [8]).

Macroporous Ion-Exchange Matrix

Recently, a major change in the polymer chemistry industry has al-
lowed the production of a rigid chromatographic support media with very
large pore structure. The macroporous structure of these beads allows
rapid movement of both fluid and large molecules through packed beds
with very low resistance. Another factor, the monodispersity or very

Copyright © 1984by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
224 CHROMATOGRAPHY [1 1]

dns negative
electrode

electrophoresis
dimension
sampleapplication

3
zone after formation
of pH gradient
pH gradient positive
10 electrode
l
protein
charge +
+

5
O

pH
protein 4-
charge
5
5

70

4-

pH

protein
4-

c h a r g e + t ~ d

~+

pH
[11] OPTIMALpH FOR ION EXCHANGERS 225

narrow particle size distribution of this rigid material, is responsible for

drastically reduced back pressures and allows elution equipment more
compatible with recovery of enzymic activity than the stainless steel nor-
mally used with very high-pressure pumping systems. A review of the
properties of these macroporous exchange materials is available. 1 The
large pore structure and monodisperse particle size distribution are
thought to be the prime factors responsible for the ability of the two
techniques described here to function in a predictive and reliable manner
to aid in the determination of optimal separation conditions on macropo-
rous ion exchangers for macromolecules. 2

Electrophoretic Titration Curves

This method, first described in 1977,3 provides a relatively simple
determination of the pH-dependent mobility of components within a pro-
tein mixture. The objective, of course, is the determination of the most
suitable pH at which to perform a chromatographic separation. The prin-
cipal steps of this two-dimensional electrophoretic technique are outlined
in Fig. 1. Briefly, the first dimension involves the creation of a pH gradi-
ent by isoelectric focusing with carrier ampholytes in a polyacrylamide or
agarose gel matrix. Once the pH gradient is established, sample is applied
in a trough across the gradient and electrophoresis is performed at right
angles. Righetti and Gianazza 4 recognized the potential of the technique
in predicting conditions for charge-dependent separations and showed in
1979 that the sigmoidal curves generated by this method were propor-
I j. Richey, Am. Lab. 14, 104 (1982).
2 Commercial availability of these supports is presently limited to the MonoBead column
series from Pharmacia Fine Chemicals, Piscataway, New Jersey. Mono Q is a strong anion
exchanger, Mono S is a strong cation exchanger, and Mono P is a chromatofocusing
medium.
3 A. Rosengren, B. Bjellqvist, and V. Gasparic, "Electrofocusing and Isotachophoresis."
de Gruyter, Berlin, 1977.
4 p. G. Righetti and E. Gianazza, in "Electrophoresis '79," p. 23. de Gruyter, Berlin, 1980.

FIG. 1. Generation of titration curves by two-dimensional electrophoresis is useful for

prediction of chromatographic pH-dependent charge mobility. (a) A pH gradient is formed
with carrier ampholytes in an electric field in the horizontal dimension, followed by sample
application in a central trough. Electrophoresis is then performed in the vertical dimension.
(b) A titration curve of this mixture, indicating that optimal separation should take place at
lower than neutral pH (pH 4-5) where proteins are positively charged, i.e., a cation ex-
changer should be used. (c) In this case, optimal separation will be at a pH greater than 7.0
and an anion exchanger should be used to resolve the net negatively charged proteins. (d)
Where titration curves are essentially parallel, separation near neutral charge, i.e., near the
pl, is optimal.
226 CHROMATOGRAPHY [ 11 ]

tional to the theoretical titration curve of the protein. Proportionality,

however, is not sufficient for prediction. In order to exploit the method
fully, two factors must be resolved: sieving effects caused by the electro-
phoresis gel matrix should be eliminated, and the chromatographic media
must separate on the basis of net protein charge (not localized charge) and
without extraneous sieving or hydrophobic interaction.
The polyacrylamide electrophoretic gel matrix described here has a
pore structure that allows free mobility of proteins up to 1.5 x 105
daltons. 5 The agarose matrix described here is recommended for mixtures
containing larger proteins up to about 6.6 x 105 daltons. By selecting and
using the two matrices, sieving effects within the electrophoresis matrix
are minimized.

Eleetrophoretic Titration Curve Method

Materials
All focusing and electrophoreses are performed on a Pharmacia Flat-
bed Apparatus FBE-3000 with an ECPS 3000/150 power supply and VH-1
volt-hour integrator. A circulating bath is required for temperature con-
trol and cooling. Gels have been cast in a custom-made casting frame,
which forms a trough for sample application.

Gel Preparation
Polyacrylamide Gels. Prepared from a stock solution (10% acryla-
mide, 3% bisacrylamide); the stock solution is deionized by mixing 1 g of
Ambedite MB-3 per 100 ml of solution for 1 hr. Solution may be stored
refrigerated, over MB-3, for up to 1 week. A solution sufficient for two
gels contained 22.5 ml acrylamide stock, 12 ml glycerol (50%), and 3 ml of
Pharmalyte 3-10. The solution is diluted to 45 ml, filtered through What-
man #1 paper, and degassed. One hundred microliters of a 60 mg/ml
ammonium peroxydisulfate solution is added to the mixture and the solu-
tion is quickly applied to the casting frame after insertion of Silane 174
(Pharmacia)-treated glass plates. Polymerization is complete and stable
after 90 min at room temperature. Gels may be stored in a humidity
chamber at 4 °.
Agarose Gels. Agarose-Sephadex gels are prepared from a solution
containing 0.45 g Agarose-IEF (Pharmacia), 0.75 g Sephadex G-200 Su-
perfine, and 4.5 g D-sorbitol. The solution is diluted to 45 ml with hot

5 A. B. Bosisio, C. Loeherlein, R. S. Snyder, and P. G. Righetti, J. Chromatogr. 189, 317

(1980).
[11] OPTIMALpH FOR ION EXCHANGERS 227

water, boiled, and cooled to 70 ° before 3 ml of Pharmalyte 3-10 is added.

Gel Bond (Marine Colloids), cut to the size of the casting mold, is allowed
to adhere to the frame with the hydrophobic side to the frame. The
agarose-Sephadex solution is then injected into the casting frame. After
hardening at room temperature for 45 min, the gels may be stored in a
moist environment at 4 °.

Running and General Staining Conditions

Both types of gels are electrofocused on the flatbed apparatus at 12°,
with the sample trough perpendicular to the electrodes. The anode elec-
trode strip is soaked in I M phosphoric acid for agarose gels or 40 mM
aspartic acid for polyacrylamide gels. Cathode strips are always prepared
by immersing in 1 M sodium hydroxide solution.
Focusing is complete for agarose gels (size: 80 x 80 mm) at 7 W
constant power for 750 V-hr whereas polyacrylamide gels are run at 15 W
for 750 V-hr. Second-dimension conditions start with sample application
and rotation of the gel a quarter of a turn, such that the sample trough is
parallel to the electrode wicks; the wicks are freshly soaked again in their
respective solutions. Agarose gels are electrophoresed at I000 V for I00
V-hr and polyacrylamide gels require 1000 V for 150 V-hr.
Both types of gels may then be fixed for 30 min in 10% trichloroacetic
acid and 5% sulfosalicylic acid. Agarose gels are dried by blotting for 30
min, followed by drying in a hot air stream until dry to the touch. Both
types of gels may be stained in either 0.2% Coomassie Brilliant Blue
R-250 or 0.2% Page Blue 83 in 35% methanol and 10% acetic acid. Protein
material in agarose gels binds stain rapidly, in about 10 min. Poly-
acrylamide gels require much longer and are normally left overnight (6-18
hr). Both types of gels may then be destained in 35% methanol and 10%
acetic acid.
In practice, specific detection methods such as zymogram staining or
fluorescent-labeled substrate-binding techniques are used to locate the
patterns of specific enzymes.6 Comparison of a nonspecifically stained gel
to one stained for enzyme activity presents the full picture in terms o f p H -
dependent mobility in relation to the other proteins in the mixture.

Retention Mapping
Most easily described as "running the sample at a number of different
pH values," the special properties of the strong anion and cation ex-
6 M. J. Heeb and O. Gabriel, this volume [28]. See also this series, Vol. 22 [40].
228 CHROMATOGRAPHY [ 11]

pH 7.3 pH 7.8

/
5

.5

2,3
6

J 7

ly J
pH 8 . 5 d
2,3
pH
9.0

8.5-

8.0-

7.5-

L0-

6.5-

6.0-

5.5-

5.0-
i

10 15 20
Retention Vol (ml)

FIG. 2. Retention mapping method for chromatographic optimization using a Pharmacia

FPLC System with the fast anion exchanger, Mono Q, at a flow rate of 1 ml per minute. Pro-
tein: (1) carbonic anhydrase (carbonate dehydratase), 1 mg/ml; (2) conalbumin type 2, 2
mg/ml; (3) transferrin, I mg/ml; (4) ovalbumin, 3 mg/ml; (5) a-lactalbumin, 1 mg/ml; (6) bovine
serum albumin, 5 mg/ml; (7) trypsin inhibitor, 4 mg/ml. (a) Buffer A: 20 mM triethanolamine
[11] OPTIMALpH FOR ION EXCHANGERS 229
changers' rapid separation times make this approach practical. All buffers
for the various pH values have been tested by the manufacturer and their
recommendations are supplied with the product. Preparation of four or
five of these buffers for one or both ion exchangers is sufficient for obtain-
ing practical retention map information. Macroporous ion exchangers
provide consistent, interpretable results within a time frame that makes
the method very attractive (10-20 min per trial). Figure 2a, b, and c
provides examples of chromatograms run at different pH values, and Fig.
2d represents the retention volume as a function of pH. It is a simple
matter to examine the curves and select the pH most appropriate for
optimal resolution of the components of interest.

Retention Mapping and Electrophoretic Titration Curves: An Example

H a f t et al. 7 have reported a study of the behavior of a number of

samples with the macroporous ion exchangers and the value of both of
these techniques in predicting conditions for optimal separation. One sys-
tem, the lactate dehydrogenase isoenzymes from beef heart and muscle
prepared by a quick-freeze, slow-thaw method, 8 was chosen because of
the very high degree of physical similarity among the isomeric subunits
and the possible resultant isoenzymes. Freezing and thawing produces
hybrids containing subunits from both native sources [beef heart (H) and
muscle (M)] in all the tetrameric conformations, H3M, HEME, HM3, H 4 ,
and M4. The investigators reasoned that chromatographic behavior would
not be complicated by size or shape differences, but would be influenced
primarily by surface charge. The lactate dehydrogenase, then, should
serve as a model system for comparing retention mapping and electropho-
retic titration curves.
The agarose-Sephadex electrophoretic titration curve (Fig. 3) reveals
the five major bands to be fairly well resolved in the anionic side of the
gel. Although polyacrylamide gels were reported to show slightly sharper
bands, splitting of the M3H and MEHEbands could just barely be discerned
7 L. A. Haft, L. G. Fagerstam, and A. R. Barry, J. Chromatogr., 266, 409 (1983).
s A. Stolzenbach, in "Methods in Enzymology: Carbohydrate Metabolism" (Willis A.
Wood, ed.), p. 287. Academic Press, New York, 1966.

chloride, pH 7.3; buffer B: buffer A + 0.35 M NaC1; gradient elution: 0-20 ml (0-100% B).
(b) pH 7.8; all other conditions as in (a). (c) Buffer A: 20 mM diethanolamine chloride, pH
8.5; buffer B: buffer A + 0.35 M NaC1; all other conditions as in (a). (d) Elution volume
plotted as a function of pH. Each curve is representative of the charge-dependent mobility of
the protein it represents.
230 CHROMATOGRAPHY [11]

/
/
/
/

/
/
,/
/'
/

:c

H-'~'l~ 10 ~ o lo EL
p Volume, ml

ETC of LDH Isoenzymea in Anion Exchange of LDH

Agarose Sephadex Isoenzyme$ on Mono Q,
pH 8.5
FIG. 3. Electrophoresis titration curve of lactate dehydrogenase isoenzymes in an
agarose-Sephadex gel matrix compared to anion exchange chromatographyat pH 8.5 on
Mono Q, a fast anion-exchangematerial.

whereas they are plainly seen here. Notice the smearing of protein mate-
rial in the cationic portion. This is valuable information to the chromato-
grapher because it immediately suggests that the mixture is unstable or
insoluble in this pH region. Chromatographic manipulations of this en-
zyme in the acidic regions must therefore be avoided because little or no
sample recovery could be expected. Using this information, a logical
choice for performing ion-exchange chromatography would be an anion
matrix, i.e., one charged positively, developed under basic conditions.
The results of four trials at pH 8.5 through 10.0 are presented in Fig. 4.
Overall resolution is clearly superior in the separation at pH 8.5, as pre-
dicted from the electrophoretic curve.
Retention mapping was performed in the range of pH 5.5 to 10 with
retention volume, in milliliters, plotted against the pH of the separation
for M3H, MH3, and H4 (Fig. 5a) and the two bands of M2H2 (Fig. 5b). That
the M2H2 band splits in this manner is not as curious as it might seem at
first. Although the subunit structures are precisely identical, two different
conformational states may exist that result in a different net surface
charge, hence a different retention volume.
[11] OPTIMALpH FOR ION EXCHANGERS 231

=: :£

Y
Y
v /

c D

< =-
=' =" =" =E

C) I'0
Volume,ml Volume, ml
Fio. 4. High-performance anion exchange chromatography on Mono Q of lactate dehy-
drogenase isoenzymes at several pH values. Buffers: (A) 20 mM 1,3-diaminopropane, pH
10.0; (B) 20 mM piperazine, pH 9.5; (C) 20 mM ethanolamine, pH 9.0; (D) 20 mM diethanol-
amine, pH 8.5. Flow rate, 1 mi/min. All were developed in their respective buffers with a
linear salt gradient to 0.35 M in NaCI.

Conclusions
Obtaining electrophoretic titration curves provides "advanced" infor-
mation regarding stability and resolution of proteins in a complex mix-
ture. Specific detection methods applicable to polyacrylamide or agarose
232 CHROMATOGRAPHY [1 1]

.o it')

~E~ "

~o

./ I~.

~..= E~

7 ~ eo

o~

i i
6 ® ~ 4 &
lUJ ~ e m n l o ^ UOh~UelaN

{3 If)
J J

(0

/ / '~.,~
f,, ,~-=

CO

./..7- .7
O)

2
I I I I I I l I I
•t o,I o ~o ~o ~ o~ o

I m ' euJnlO^ u o ! l u a ~ , e ~
[11] OPTIMALpH FOR ION EXCHANGERS 233

may be used for electrophoretic titration curves. However, if such tech-

niques are not available, retention mapping provides another, also empiri-
cal, alternative with actual fractions of recovered material available for
specific assay.
Although published uses of these techniques are as yet infrequent, the
producer of the macroporous exchangers reports numerous communica-
tions of helpful applications and has published on the subject.l,7
[12] POLYACRYLAMIDE GEL ELECTROPHORESIS 237

[12] S y s t e m s for P o l y a c r y l a m i d e G e l E l e c t r o p h o r e s i s
By PERRY J. BLACKSHEAR

Separation of proteins in complex mixtures by polyacrylamide gel

electrophoresis remains a powerful and versatile analytical technique.
Advances in the use of sensitive protein staining procedures (see this
volume [28-32]) and autoradiography (see this volume [34]) have greatly
increased the sensitivity of the primary technique. Perhaps equally impor-
tant are extensions of gel electrophoresis for specialized purposes, includ-
ing two-dimensional electrophoresis, partial proteolytic digestion within
the stacking gel, electrophoretic transfer to paper for immunoblotting
procedures, and others.
Presented here are some of the polyacrylamide gel electrophoretic
separation methods that we have found to be useful for a variety of
purposes. No attempt is made to review the many electrophoretic proce-
dures in use today, although some of them are recorded elsewhere in this
volume. Most of the procedures presented here deal with electrophoresis
in slab gels, but virtually all the comments are applicable to electrophore-
sis with rod gels.

General Principles
Most gel systems widely used today make use of the strategy for zonal
or discontinuous electrophoresis in polyacrylamide gels developed by
Ornstein I and Davis 2 and Raymond and Weintraub. 3 The advantages in-
herent in the use of a stacking gel with multiphasic buffer systems have
been described in detail by Chrambach and Rodbard 4 and include the
ability to study very dilute samples; the ability to concentrate proteins
into very thin starting zones of high local protein concentration, resulting
in improved resolution; and the provision of a dye front, which allows
calculation of relative mobilities of proteins, as well as a means of follow-
ing the progress of a given electrophoretic run.
The principle behind the use of a stacking gel is derived from the so-
called Kohlrausch regulating function. In brief, at the pH of the stacking
gel, the sample ion (in most cases, a mixture of proteins or their detergent
I L. Ornstein,Ann. N.Y. Acad. Sci. 121, 321 (1964).
2 B. J. Davis, Ann. N.Y. Acad. Sci. 121, 404 (1964).
3 S. Raymondand L. Weintraub, Science 130, 711 (1959).
4A. Chrambachand D. Rodbard, Science 172, 440 (1971).

Copyright © 1984 by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
238 ELECTROPHORESIS [12]
derivatives) is introduced near the boundary of two ions of the same sign
as the sample ion at that pH. One ion migrates faster than, and the other
slower than, the sample ion when a current is passed through the mixture,
and the sample ion, being of intermediate mobility, is sandwiched be-
tween them, forming a narrow zone or band of high concentration. When
the sample ion migrates into the running or separating gel, under condi-
tions of different pH and/or pore size, the trailing ion continuously over-
takes and passes the sample ion, "establishing a comparatively uniform
voltage gradient in which electrophoretic separation of samples occurs."z
A further refinement that has attained widespread use is the introduc-
tion of electrophoresis of sodium dodecyl sulfate (SDS) derivatives of
proteins by Shapiro, Vifiuela, and Maizel in 19675 and subsequent descrip-
tions of the utility of this system as a means of separating proteins by their
subunit molecular weight, largely independent of tertiary conformation,
amino acid composition, or isoelectric point. 6 Denaturation and binding of
SDS to proteins and peptides generally result in a relatively uniform
negative charge, since most peptides bind about 1.4 mg of SDS per milli-
gram of protein; furthermore, changes in electrophoretic behavior result-
ing from different tertiary protein structures are largely obviated because
of the uncoiling introduced by SDS binding. Thus, when combined with
discontinuous or zonal electrophoresis, SDS electrophoresis in polyacryl-
amide gels combines the advantages of a stacking system with those of
the combined gel filtration and electrophoretic properties of the separat-
ing gel, resulting in sharp separation of proteins based largely on their
subunit molecular weights. The gel systems described by Laemmli,7 who
added SDS to the Tris-glycine buffer system of Ornstein and Davis, and
by Neville, 8 using an SDS-Tris-borate buffer, are probably the most
versatile and widely used types of denaturing, discontinuous electropho-
resis.
Other systems have since been developed for more specialized pur-
poses. For example, linear or curvilinear gradient gels, in which the acryl-
amide concentration increases linearly in the separating gel (often from 3
to 30%) have become useful for maximizing separation of protein bands in
a complex mixture of proteins of widely varying molecular weights. Other
buffer systems employing SDS-urea, SDS-agarose, or other combina-
tions have been useful in the study of small proteins and peptides. Acidic
detergents such as tetradecyltrimethylammonium bromide (TDAB) or

5 A. L. Shapiro, E. Vifiuela, and J. V. Maizel, Biochem. Biophys. Res. Commun. 28, 815
(1967).
6 K. Weber and M. Osborn, J. Biol. Chem. 244, 4406 (1969).
7 U. K. Laemmli, Nature (London) 277, 680 (1970).
8 D. M. Neville, J. Biol. Chem. 246, 6328 (1971).
[12] POLYACRYLAMIDE GEL ELECTROPHORESIS 239

hexadecylpyridinium chloride (HDPC) have found acceptance for the

separation of membrane acylphosphate intermediates in a discontinuous
system. Careful use of disulfide reducing agents with and without SDS has
been successful in determination of the subunit structure of a number of
complex proteins. Finally, two-dimensional electrophoresis in which pro-
teins are separated by isoelectric focusing and then by SDS-polyacryl-
amide gel electrophoresis has revolutionized the separation of proteins in
crude mixtures.
The emphasis here is on discontinuous or zonal electrophoresis in
which a stacking gel is used. These systems have so many advantages
over continuous electrophoresis that there are few situations in which the
latter is necessary. However, Hames 9 has noted that the concentration of
a protein that occurs during stacking can occasionally lead to protein
aggregation or precipitation in nondenaturing solutions. This leads to fail-
ure of the protein to enter the gel, as well as to artifacts such as streaking
in the path of the sample lane. Where a continuous electrophoresis system
seems advantageous, almost any buffer can be used, provided the proteins
of interest retain either a positive or a negative charge; this can be accom-
plished using one of the three buffers described below for nondenaturing
gels. Such systems require that the smallest possible volume of sample be
applied and that the sample buffers be of low ionic strength to ensure
maximal band sharpness despite the absence of a stacking gel.

Nondenaturing Gels
When it is necessary to separate intact proteins, especially oligomeric
proteins, by a nondestructive means for later assessment of biological
activity, gels must be prepared under nondenaturing conditions. Since
denaturation of protein is not involved, protein separation depends on a
combination of differences in molecular size and shape as well as charge.
Separation by size is accomplished by varying the pore size of the acryl-
amide polymer as a function of both the concentration of the acrylamide
(range about 3-30%, w/v) and the amount of cross-linker used. In general,
the higher the acrylamide concentration, the smaller the proteins that
remain in the gel; this can be counteracted by decreasing the amount of
cross-linker used, which in turn increases the degree of gel swelling dur-
ing standard staining and washing procedures. Separation by charge in
nondenaturing gel systems is permitted because the protein separation
can be performed at any pH between 3 and 11, to allow for maximal
charge differences between neighboring protein species. For these rea-
9 B. D. Hames, in "Gel Electrophoresisof Proteins: A PracticalApproach" (B. D. Haines
and D. Rickwood,eds.), p. 1. IRL Press, Oxfordand WashingtonD.C., 1981.
240 ELECTROPHORESIS [ 12]
sons, Gabriel 1° has recommended that several properties of an enzyme be
established before it is subjected to electrophoresis to determine, for ex-
ample, its purity in a heterogeneous solution. These include the molecular
weight of the protein, its isoelectric point, and, if enzymic activity is to be
retained, its stability in the range of pH 4 to 9. Once these parameters are
known, optimal resolution can be obtained by varying certain components
within a single gel system. For example, the initial choice of a gel system
for separation of an acidic protein of molecular weight 20,000 might be a
system operating at an alkaline pH (i.e., 9) with an acrylamide concentra-
tion of 12-15%.
In this series, Gabriel 1°has described in detail the gel formulations and
buffer systems to be used in the construction of three gel systems for
discontinuous, nondenaturing electrophoresis at both acid and alkaline
pH. Detailed formulations for these gel systems, summarized from previ-
ously published methods, have been compiled in a more convenient for-
mat by Hames, 9 and these details are described below. Many hundreds of
possible buffer combinations can be generated by computer analysis of
discontinuous gel systems, 4 and suitable combinations of pH, buffer com-
ponents, and gel constituents are obtainable for almost every require-
ment. It should be noted that biologically active proteins such as enzymes
often require special electrophoretic conditions in order to retain activity
after separation. For example, the heat lability of some enzymes requires
that electrophoresis be conducted most conveniently in a refrigerated
room or, if necessary, using circulating water of 0-4 ° in the cooling
jacket of the slab gel apparatus. In addition, ammonium persulfate, an
agent commonly used in the gel polymerization reaction, often interferes
with enzyme activity after elution from the gel. For this reason, the tables
include the use of the photoactivated polymerizing agent riboflavin, as
well as ammonium persulfate, for use in the preparation of the stacking
gel for the various systems; if ammonium persulfate affects enzyme activ-
ity, then preelectrophoresis can be used to remove it from the gel if a
continuous electrophoresis system is used. Dithiothreitol (40-100 mM) or
2-mercaptoethanol (up to 1 M) may be included in the sample buffer to
reduce some disulfide linkages, although detergent denaturation is often
necessary for complete disulfide reduction.
Using the compilation of Hames,9 the stock solutions in Table I should
be prepared. These can be combined with the appropriate buffers to form
the solutions for the three types of discontinuous, nondenaturing electro-
phoresis described in Table II. Depending on the desired acrylamide con-
centration, the solutions can be combined to form the gels and buffers in
Table III. Details of solution preparation are included in the tables in most

1o O. Gabriel, this series, Vol. 22, p. 565.

[12] POLYACRYLAMIDE GEL ELECTROPHORESIS 241

TABLE I
STOCK SOLUTIONS FOR NONDENATURING DISCONTINUOUS ELECTROPHORESISa

Solution Formula

1. Acrylamide-bisacrylamide, 30 : 0.8 30 g of acrylamide, 0.8 g of bisacrylamide

per 100 ml; store at room temperature
2. TEMED (N,N,N',N'-tetramethyl- Use as supplied; store in the cold in dark
ethylenediamine) bottle
3. Ammonium persulfate, 1.5%, w/v 1.5 g in 100 ml; make up fresh on day of
electrophoresis
4. Riboflavin, 0.004%, w/v 4 mg in 100 ml
5. SDS, 10%, w/v 10 g in 100 ml; filter through Whatman No.
1 paper before use; store at room tem-
perature

a From Hames.9

cases; details of solution mixing, gel pouring, sample loading, etc., are
described in greater detail in the following section on electrophoresis
under denaturing conditions.

Denaturing Gel Systems

Protein separation for analytical purposes by denaturing, discontinu-
ous electrophoresis has achieved widespread popularity. It is a relatively
rapid, inexpensive, and reproducible means of separating hundreds of
proteins in large numbers of heterogeneous samples, largely on the basis
of their molecular weights. The gels can then be analyzed for protein
staining by several techniques of differing sensitivity and specificity for
determination, for example, of protein purity during a purification proce-
dure. Selected protein bands can be cut out of the gel and used for the
determination of radioactivity from a variety of in vivo and in vitro label-
ing techniques, amino acid compositions, and sometimes sequences, for
preparation of antibody, and, occasionally, for assay of retained enzyme
activity. Dried gels can be used for autoradiography of proteins labeled by
similar techniques.
Among the advantages of denaturing gel electrophoresis using sodium
dodecyl sulfate (SDS) is that most proteins in crude mixtures are soluble
in SDS and bind it avidly; even the most basic proteins are converted to
their acidic SDS derivatives and thereby migrate toward the anode at pH
7-9. Since SDS derivatives are largely unfolded, they migrate closer to
their true subunit molecular weight and are more susceptible to reduction
by disulfide-reducing agents, so that protein subunit structure can be more
accurately assessed. The addition of hot SDS to an enzymic reaction or to
mixtures of proteins is usually sufficient to stop enzyme activity, thereby
242 ELECTROPHORESIS [12]

TABLE II
BUFFERS FOR NONDISSOCIATINGDISCONTINUOUSSYSTEMSa

Buffer Preparation

High pH discontinuousb: stacks at pH 8.3,

separates at pH 9.5
Stacking gel buffer (Tris-HCl at pH 6.8) 6.0 g of Tris in 40 ml titrated to pH 6.8
6.8) with 1 M HC1; bring to 100 ml

Resolving gel buffer (Tris-HCl at pH 8.8) 36.3 g of Tris and 48.0 ml of 1 M HCI
brought to 100 ml; titrate to pH 8.8 with
HCi, if necessary
Reservoir buffer (Tris-glycine at pH 8.3) At the correct concentration for use; 3 g of
Tris and 14.4 g of glycine in 1 liter
Neutral pH discontinuousc: stacks at pH
7.0, separates at pH 8.0
Stacking gel buffer (Tris-phosphate at 4.95 g of Tris in 40 ml titrated to pH 5.5
pH 5.5) with 1 M phosphoric acid; bring to
100 ml

Resolving gel buffer (Tris-HC1 at pH 7.5) 6.85 g of Tris in 40 ml of water titrated to

pH 7.5 with 1 M HCI; bring to 100 ml
Reservoir buffer (Tris-diethylbarbiturate 5.52 g of diethylbarbituric acid and 1.0 g of
at pH 7.0) Tris to 1 liter

Low pH discontinuousd: Stacks at pH 5.0,

separates at pH 3.8
Stacking gel buffer (acetic acid-KOH at 48.0 ml of 1 M KOH and 2.9 ml of glacial
pH 6.8) acetic acid; bring to 100 mi

Resolving gel buffer (acetic acid-KOH 48.0 ml of 1 M KOH and 17.2 ml of glacial
at pH 4.3) acetic acid are mixed and diluted to 100
ml
Reservoir buffer (acetic acid-fl-alanine 31.2 g of fl-alanine and 8.0 mi of glacial
at pH 4.5) acetic acid to 1 liter

Reprinted from B. D. Hames, in "Gel Electrophoresis of Proteins: A Practical Ap-

proach" (B. D. Hames and D. Rickwood, eds.), p. 30. IRL Press, Oxford and Washing-
ton D.C., 1981, with permission.
b Davis.2
c D. E. Williams and R. A. Reisfeld, Ann. N.Y. Acad. Sci. 121, 373 (1964).
d R. A. Reisfeld, V. J. Lewis, and D. E. Williams, Nature (London) 195, 281 (1962).

halting the effects of proteases, protein kinases, and protein phospha-

tases.
The main drawback of the technique is that denaturation with SDS and
similar detergents is usually irreversible, and most proteins cannot be
[12] POLYACRYLAMIDE GEL ELECTROPHORESIS 243

,.A "~ I

II
o ..Q

al

r~

r~
r,.
• °
o

z
e4 ~
Z

z
°

.1 i

[-..
z • .
~ . m ~° ~
Z

.o~,~ o=o,.
.<

0EL
.°
ul

G
oo
. ~

• ~
244 ELECTROPHORESIS [12]
recovered with intact biological activities, although several exceptions
have been recorded. Variable binding of SDS to some proteins, particu-
larly low-molecular-weight acidic and basic proteins, has been described;
it often results in inaccurate molecular weight determinations. A number
of posttranslational modifications, including glycosylation and phospho-
rylation, can affect electrophoretic mobility; the latter modification af-
fects mobility of certain proteins, but not all, probably by affecting the net
charge of the SDS derivative. Some of these factors can be used to advan-
tage; others can be obviated by using another type of denaturing electro-
phoresis.
One SDS-polyacrylamide gel system that is in use in our laboratory,
and some of its variations, is presented here in detail. Subsequently,
several other types of denaturing electrophoretic systems, those useful in
specialized applications, are also described.

SDS-Discontinuous Electrophoresis
As noted, the anionic detergent SDS denatures proteins by binding
avidly to them at about 1.4 mg of SDS per milligram of protein, resulting
in strongly negatively charged proteins at neutral pH that are of approxi-
mately uniform charge per unit of protein. Proteins are exposed to SDS,
usually 1-2% (w/v), for 3 min at 100° to effect complete denaturation. This
is performed with or without reducing agents, generally dithiothreitol (40-
100 mM) or 2-mercaptoethanol (up to 1 M), to reduce disulfide bonds as
needed. A substance of high specific gravity [ 10% (w/v) glycerol or 0.25 M
sucrose] is added to the protein sample to allow it to sink to the bottom of
the sample well in the stacking gel, and a tracking dye (pyronine Y or
bromophenol blue) is added to allow visualization of the fastest migrating
components in the mixture. The negatively charged SDS-protein deriva-
tives are then subjected to an electric current at pH 7-9, causing them to
migrate toward the anode. The pore size of the polyacrylamide matrix
allows components of lower molecular weight to migrate faster, thus sep-
arating the protein components in the mixture according to the apparent
molecular weights of their SDS derivatives. Apparent molecular weights
of sample proteins can be determined by comparison with protein stan-
dards of known molecular weight.
At the end of the run, when the dye front has reached a predetermined
point near the end of the gel, the current is turned off, the separating gel is
separated from the stacking gel, and the former is subjected to protein
fixation and staining or to whatever further analytical techniques are de-
sired. Fixation and staining of proteins in polyacrylamide gels are dis-
cussed in this volume [29-32].
[12] POLYACRYLAMIDE
GEL ELECTROPHORESIS 245

TABLE IV
STOCKSOLUTIONSFOR TRIs-GLYCINE(LAEMMLI)DENATUPaNG(SDS)
DISCONTINUOUSELECTROPHORESISa

Solution Preparation

1. Acrylamide-bisacrylamide, 37.5 : 1.0 37.5 g of acrylamide and 1 g or 0.5 g of

or 37.5 : 0.5 bisacrylamide to 100 ml; store at 20°
2. Lower gel buffer, 1.5 M Tris-HCl, Tris, 181.5 g. Bring to pH 8.8 and to 1
pH 8.8 liter; store at 4°
3. Upper gel buffer, 0.5 M Tris-HC1, "Iris, 60.5 g. Bring to pH 6.8 with concen-
pH 6.8 trated HC1 and to 1 liter; store at 4°
4. SDS, 20% (w/v) SDS, 100 g, in 500 ml. Filter through
Whatman No. 1 before use; store at 20°
5. TEMED, 0.5% (v/v) TEMED, 0.5 ml in 100 ml; store in dark
bottle at 4°
6. Ammonium persulfate, 1.5% Ammonium persulfate, 1.5 g in 100 ml;
make up fresh
7. Running buffer Tris, 12 g, and glycine, 57.6 g, in water;
add 10 ml of 20% SDS and bring to 2
liters
8. Gel overlay, 0.1% SDS SDS, 100 mg in 100 ml; dispense from
spray bottle
9. Sample "quench" 20% (w/v) SDS, 6.0 ml; 1 M dithiothreitol,
4.8 ml; 0.1% (w/v) pyronine Y, 1.2 ml;
and 50% (w/v) sucrose, 8.0 ml. Bring to
20 ml
10. Stain Glacial acetic acid, 100 ml; methanol, 500
ml; H20, 400 ml; Coomassie Brilliant
Blue, 1-2 g. Use about 500 ml per gel;
stain for 45-60 min at 20°
11. Destain 10% (v/v) acetic acid; use about 500 ml per
gel, change as needed

a Modified slightly from U. K. Laemmli, Nature (London) 277, 680 (1970).

Specific Details of Procedure

Preparation of the Gel. T h e e l e c t r o p h o r e s i s s y s t e m d e s c r i b e d in detail
h e r e is a m i n o r m o d i f i c a t i o n o f the S D S - T r i s - g l y c i n e s y s t e m o f L a e m m l i 7
( T a b l e s I V a n d V). S i m i l a r p r o c e d u r e s c a n b e u s e d with the S D S - T r i s -
b o r a t e systemS; s t o c k s o l u t i o n s a n d gel f o r m u l a s for this s y s t e m are de-
s c r i b e d in T a b l e s V I a n d VII. C o m m e r c i a l l y a v a i l a b l e p r e p a r a t i o n s o f
a c r y l a m i d e a n d N , N ' - m e t h y l e n e b i s a c r y l a m i d e ( b i s a c r y l a m i d e ) , as well as
o t h e r gel c o m p o n e n t s , are o f a n a d e q u a t e level o f p u r i t y w i t h o u t r e c r y s t a l -
lization. S t a n d a r d slab gel a p p a r a t u s , a v a i l a b l e f r o m B i o - R a d , H o e f f e r , or
o t h e r s u p p l i e r s , i n w h i c h the s e p a r a t i n g a n d s t a c k i n g gels are s u c c e s s i v e l y
246 ELECTROPHORESIS [ 12]

TABLE V
FORMULASFOR TRIS-GLYCINE(LAEMMLI)DENATURING(SDS)
DISCONTINUOUSELECTROPHORESIS

Final percentage of
acrylamide desired

7.2% 9.0% 12.0% 15.0% 20.0%

A. Lower (separating) gel, for one slab of average size

Mix first
I. Acrylamide-bisacrylamide a (ml) 5.75 7.2 9.6 12.0 16.0
2. Lower gel buffer (ml) 7.5 7.5 7.5 7.5 7.5
3. SDS, 20% (ml) 0.15 0.15 0.15 0.15 0.15
4. Degassed water (ml) 13.45 12.0 9.6 7.2 3.2
Add to polymerize
5. Ammonium persulfate, 1.5% (ml) 0.75 0.75 0.75 0.75 0.75
6. TEMED, 0.5% (ml) 2.4 2.4 2.4 2.4 2.4

B. Upper (stacking) gel, for one slab of average size

Mix first
1. Acrylamide-bisacrylamide (ml) 0.8
2. Upper gel buffer (ml) 2.5
3. SDS, 20% (ml) 0.05
4. Degassed water (ml) 5.15
Add to polymerize
5. Ammonium persulfate, 1.5% (ml) 0.7
6. TEMED, 0.5% (ml) 0.8

a For 15 and 20% gels, 37.5 : 0.5 or 37.5 : 0.25 may be needed.

poured b e t w e e n a mold formed b y two glass plates, separated by spacers

of 0.75- to 3-ram thickness, are all adequate. The thin gels (0.75 mm) are
n e c e s s a r y if the sensitive silver stain of Merril e t ai. H is to be used; 3-mm
gels are valuable for large sample volumes or for preparative purposes.
Once the gel apparatus has b e e n arranged with extremely clean and
grease-free glass plates, the separating gel is poured. Table IV lists the
stock solutions, and Table V lists the c o m p o n e n t s of these gels as pre-
p a r e d in our laboratory; it should be stressed that the higher ratios o f
a c r y l a m i d e - b i s a c r y l a m i d e are used when the final acrylamide concentra-
tion is greater than about 12%. At high acrylamide concentrations, the
gels b e c o m e stiffer and m o r e brittle and require more cooling during the
run; in addition, at high ratios of acrylamide to bisacrylamide, gels tend to
destain m o r e quickly and swell considerably during destaining in 7 - 1 0 %

11C. R. Merril, D. Goldman, S. A. Sedman, and M. H. Ebert, Science 211, 1437 (1981). See
also this volume [30].
[12] POLYACRYLAMIDE GEL ELECTROPHORESIS 247

acetic acid. The deionized water used should be degassed prior to use; the
ammonium persulfate solution should be prepared on the day of use.
To prepare the gel, mix the appropriate components together in a
small flask, swirl rapidly without causing bubble formation or aeration,
and pipette rapidly a suitable volume of the final solution into the mold
formed by the glass plates. No bubbles or leaks should be present. To
ensure a flat junction between separating and stacking gels, we routinely
spray the inside of one plate with a gel overlay solution consisting of 0.1%
(w/v) SDS, which will maintain a flat meniscus on top of the separating gel
during polymerization. For most gels, polymerization usually is complete
within 30-60 min at room temperature; this can he checked by noting a
discrete line of separation between the gel and the overlay or by tilting the
mold to make sure that the gel is polymerized. If it is covered by a
reasonable volume of overlay to prevent drying out, the separating gel can
be stored at 4° for up to a week without evident ill effects.
After polymerization of the separating gel is complete, the gel overlay
is poured off and the stacking gel mixture is pipetted into the mold. If
several samples are to be evaluated "combs" providing a template for I0-

TABLE VI
STOCK SOLUTIONS FOR TIns-BORATE (NEVILLE) DENATURING (SDS)
DISCONTINUOUS ELECTROPHORESISa

Solution Preparation

1. Acrylamide-bisacrylamide, 40 : 1.5 40 g of acrylamide and 1.5 g of bisacrylamide

in 100 ml of water; store at room tempera-
ture
2. Lower (separating) gel buffer 318.5 g of Tris in water. Adjust to pH 9.81
with 1 N HCI and bring to 1 liter; store in
cold
3. Upper (stacking) gel buffer 270.6 g of Tris brought to pH 6.1 with 1 N
H2SO4 and diluted to 1 liter; store in cold
4. Upper tray (cathode) stock (10×) 99.3 g of Tris and 49.5 g of boric acid brought
to pH 8.64 with 1 N NaOH. Add 100 ml of
20% SDS and dilute to 2 liters; store at
room temperatuie. Dilute 1 : 10 for use
5. Lower tray (anode) buffer Dilute lower gel buffer 1 : 10 for use
6. SDS, 20% (w/v) See Table IV
7. TEMED, 0.5% (v/v) See Table IV
8. Gel overlay See Table IV
9. Sample " q u e n c h " See Table IV
10. Stain See Table IV
11. Destain See Table IV

a Modified slightly from D. M. Neville, J. Biol. Chem. 246~. 6328 (1971).

248 ELECTROPHORESIS [ 12]

TABLE VII
FORMULAS FOR TRIS-BORATE (NEVILLE) DENATURING(SDS)
DISCONTINUOUSELECTROPHORESIS

Final
percentage
of acrylamide
desired

7.2 9.0 12.0

A. Lower (separating) gel, for one slab of average size

Mix first
1. Acrylamide-bisacrylamide (ml) 5.4 6.75 9.0
2. Lower gel buffer (ml) 15.0 15.0 15.0
3. Degassed water (ml) 6.6 5.25 3.0
Add to polymerize
4. Ammonium persulfate, 1.5% (ml) 1.5 1.5 1.5
5. TEMED, 0.5% (ml) 1.5 1.5 1.5

B. Upper (stacking) gel, for one slab of average size

Mix first
1. Acrylamide-bisacrylamide (ml) 1.15
2. UBper gel buffer (ml) 1.5
3. Degassed water (ml) 10.1
Add to polymerize
4. Ammonium persulfate, 1.5% (ml) 1.2
5. TEMED, 0.5% (ml) 1.05

20 sample wells can be inserted into the liquid stacking gel solution before
polymerization begins. If a flat surface is desired for, for example, a first-
dimension isoelectric focusing gel, then gel overlay solution is again
sprayed on top of the running gel before polymerization to ensure a
smooth, fiat meniscus. A similar procedure can be followed if the gel is to
be used for preparative electrophoresis of larger volumes. For this pur-
pose, the stacking gel is poured and covered with overlay, leaving suffi-
cient space between the plates for application of the sample in a volume of
up to 3 ml. Once polymerization of the upper or stacking gel is complete,
usually taking longer than the lower gel but still less than an hour at room
temperature, the template or comb is removed, the upper and lower as-
pects of the gel are placed in running buffer, and sample application can
proceed.
Preparation of the Sample and Sample Loading. When preparing
crude protein mixtures or solutions for electrophoresis, the various sam-
ples are routinely diluted until they are of the same protein concentration
or trichloroacetic acid-precipitable radioactivity, and then one-fifth vol-
[12] POLYACRYLAMIDE GEL ELECTROPHORESIS 249

ume of a "quench" solution (Table IV) is added to each. Generally, the

ratio of SDS to protein in the final sample should be at least 3 : I (by
weight). The final dithiothreitol concentration should be at least 40 mM to
ensure complete reduction of both interchain and intrachain disulfide
bonds. The samples are normally placed in a boiling water bath for 3-5
min, allowed to cool, and then loaded onto the gel lanes after centrifuga-
tion of the samples to remove any insoluble components. In general, not
more than 250/.Lg of a complex mixture of protein in a volume of 50/zl or
less should be loaded into 4-mm wells in a 1.5-mm-thick gel; 1 mg of
protein, or more, in less than 200/xl can usually be loaded into a l-cm well
in a 1.5-mm gel provided the stacking gel is sufficiently long. The high
specific gravity of the sucrose makes the samples sink to the bottom of the
well, so that sample gels are not necessary. We have found that capillary
pipettes with a hand-held applicator are the best means for loading sam-
ples, since they fit nicely between the plates separated by 1.5-mm spac-
ers, and they can be inserted directly into the sample wells. Again, air
bubbles should be avoided during sample loading.
Running the Gel. After samples have been loaded, the entire gel appa-
ratus is placed into a box so that both the upper and lower aspects of the
gel are exposed to running buffer (Table IV). The lower buffer chamber is
connected to the anode, so that negatively charged SDS derivatives mi-
grate downward when the current is turned on. Air bubbles should be
removed from the lower aspect of the gel, using a piece of rubber tubing or
a bent Pasteur pipette. The current is then turned on; the procedure is
adequate at 25 mA per side during the stacking phase and 30-50 mA per
side during the separating phase of the run, providing the voltage remains
below 200 V. Higher voltages are associated with excessive heat produc-
tion and sometimes blurring of protein bands. At high acrylamide concen-
trations (greater than 20%, w/v), it is often necessary to run the gel at 4 °
with a cooling jacket to prevent excessive heating of the gel. In most
cases, a single gel run takes 3-4 hr; if necessary, gels can be run overnight
at low levels of current without apparent loss of resolution.
After the completion of the electrophoretic run, the gels are removed
from the plates, stained, and destained. For routine use, gels are im-
mersed for about 1 hr at room temperature in a staining solution consist-
ing of 50% (v/v) methanol, 10% (v/v) acetic acid, 40% (v/v) water, and
0.1% Coomassie Brilliant Blue, with moderate shaking. Destaining is car-
ried out overnight in 10% acetic acid in water. Other treatment of the
destained gels can also be carried out, including removal of phosphory-
lated nucleic acids by treatment of the gel with 5% trichloroacetic acid at
90 ° for 30 min, ~2hydrolysis of phosphoserine by immersion in 1 N KOH at
12 S. Auerbach and T. Pederson, Biochem. Biophys. Res. Commun. 63, 149 (1975).
250 ELECTROPHORESIS [ 12]

55° for 2 hr, 13 or even radioiodination of proteins within the gel. 14 The
destained gels may then be analyzed by densitometry, or the stained
bands may be cut out and analyzed or subjected to liquid scintillation
counting. The gels may be dried, subjected to direct or indirect autora-
diography or fluorography, or analyzed in a number of other ways as
described in Section IV of this volume.
Molecular Weight Determinations of Unknown Proteins. SDS-poly-
acrylamide gel electrophoresis is useful for determining molecular
weights of proteins and their subunits, although some proteins behave
anomalously. In general, proteins are separated on a slab gel in which one
or more lanes have been devoted to the electrophoresis of protein stan-
dards, i.e., proteins of relative purity of known subunit molecular
weights. A number of suitable proteins and their subunit molecular
weights (after disulfide reduction) are listed in Table VIII; for routine use,
both Sigma and Bio-Rad provide kits containing five or six proteins of
well-defined molecular weights in the region of interest.
After completion of electrophoresis, staining, and destaining, relative
mobilities are calculated for each of the standard proteins and the un-
known proteins of interest. Relative mobility (Rf) is defined as the dis-
tance migrated by protein divided by the distance migrated by tracking
dye, where distance refers to the distance from the junction of the stack-
ing and separating gels. If the lOgl0 of the molecular weights of the stan-
dard proteins is plotted as a function of their Rf, a straight line is usually
formed that encompasses about the middle 80% of the area of the separat-
ing gel. The apparent molecular weights of the sample proteins can then
be determined by matching their Rf value with the appropriate point on
the standard curves. Hames 9 has estimated that linear standard curves
can be obtained within the following useful ranges at three different acryl-
amide concentrations: 5% for molecular weights 60,000 to 212,000; 10%
for 18,000 to 75,000; and 15% for 15,000 to 45,000. Increasing the acrylam-
ide-to-bisacrylamide ratio from the standard ratio results in an increase in
the size of proteins for which a linear relationship between molecular
weight and Rf obtains, for any given acrylamide concentration.
This general method, using uniform separating gels (rather than gradi-
ent gels) is not very useful for proteins of less than about 10,000 molecular
weight. This problem has been approached by Swank and Munkries, 15
who used SDS-polyacrylamide gel electrophoresis in the presence of 8 M
urea in gels composed of 12.5% acrylamide and 1.25% bisacrylamide. An

i~ j. A. Cooper and T. Hunter, Mol. Cell Biol. 1, 165 (1981).

14 j. H. Elder, R. A. Pickett, J. Hampton, and R. A. Lerner, J. Biol. Chem. 252, 6510 (1977).
15 R. T. Swank and K. D. Munkries, Anal. Biochem. 39, 462 (1971).
[12] POLYACRYLAMIDE
GEL ELECTROPHORESIS 251

TABLE VIII
MOLECULARWEIGHT(mr) OF POLYPEPTIDESTANDARDSa

Polypeptide Mr

Myosin (rabbit muscle) heavy chain 212,000

RNA polymerase (E. coli)/3'-subunit 165,000
fl-subunit 155,000
fl-Galactosidase (E. coli) 130,000
Phosphorylase a (rabbit muscle) 92,000
Bovine serum albumin 68,000
Catalase (bovine liver) 57,500
Pyruvate kinase (rabbit muscle) 57,200
Glutamate dehydrogenase (bovine liver) 53,000
Fumarase (pig liver) 48,500
Ovalbumin 43,000
Enolase (rabbit muscle) 42,000
Alcohol dehydrogenase (horse liver) 41,000
Aldolase (rabbit muscle) 40,000
RNA polymerase (E. coli) a-subunit 39,000
Glyceraldehyde-3-phosphate dehydrogenase (rabbit muscle) 36,000
Lactate dehydrogenase (pig heart) 36,000
Carbonic anhydrase 29,000
Chymotrypsinogen A 25,700
Trypsin inhibitor (soybean) 20,100
Myoglobin (horse heart) 16,950
a-Lactalbumin (bovine milk) 14,400
Lysozyme (egg white) 14,300
Cytochrome c 11,700

a The data for Mr are in the presence of excess~thiol reagent. Re-

printed from B. D. Hames, in "Gel Electrophoresis of Proteins: A
Practical Approach" (B. D. Hames and D. Rickwood, eds.), p. 39.
IRL Press, Oxford and Washington D.C., 1981, with permission.

essentially linear relationship was observed between the log of the molec-
ular weight and Rf of oligopeptides of molecular weights between 1500
and 15,000 using this system.
In addition to low-molecular-weight peptides, anomalous migration
behavior can be exhibited by very basic proteins, very acidic proteins,
glycoproteins, membrane proteins, phosphoproteins, and proteolipids.
For glycoproteins, one approach to the true molecular weight has been
the use of many different acrylamide concentrations as described by Se-
grest and Jackson. 16 Some of the methods in use for molecular weight
determination in instances of anomalous migration behavior will be de-

16j. p. Segrest and R. L. Jackson, this series, Vol. 28, p. 54.

252 ELECTROPHORESIS [12]
scribed further below. Use of gradient gels for this application is dis-
cussed in the following section.
Gradient Gels. Gradient gels, in which the concentration of acrylam-
ide increases regularly from top to bottom of the separating gel, have
gained popularity for several reasons. First, protein molecular weights
can be estimated for a much broader range of proteins in a single gel than
can be measured in a uniform gel. Second, the separation of protein bands
from neighboring bands is better, and each band is sharper than in nonop-
timal zones of uniform gels. Finally, some anomalous migration behavior
may be decreased on gradient gels. However, band separation in the
optimum area of a uniform gel remains superior to that achieved in gradi-
ent gels. Thus, gradient gels are most useful when performing initial stud-
ies of protein mixtures encompassing a large range of molecular weights.
In practice, linear gradients are formed by adding equal volumes of a
final separating gel mixture (i.e., after addition of both ammonium per-
sulfate and TEMED) to each side of a common gradient-forming device.
A common recipe for a single slab gel of 20 ml would involve 10 ml of a 3%
acrylamide solution and 10 ml of a 30% solution. As the mixed solution is
withdrawn by gravity or through the use of a peristaltic pump, it is infused
into a slab gel apparatus in such a way as to prevent aeration and bubble
formation. When the entire gel is poured, the gel overlay and stacking gel
are poured in the usual manner. Obviously, the gradient mixer should be
arranged so that the 30% acrylamide solution reaches the bottom of the
gel first, rather than the other way around.
Any of the gel recipes described in the foregoing sections can be used
for gradient gels, although the discontinuous SDS systems have probably
been used most widely. It is wise to decrease the amount of TEMED used
in the gel recipes so that more time is available for pouring the gel before it
polymerizes. Some authors recommend dissolving sucrose in the gel solu-
tion with the higher concentration of acrylamide, to form a density gradi-
ent with the acrylamide gradient that will prevent mixing with added
solution during pouring of the gel. We have not found this to be nec-
essary.
For determination of molecular weights in a gradient gel, the log of the
molecular weights of standard proteins run in a neighboring track on a
slab gel is plotted vs the log of acrylamide concentration of the gel, assum-
ing a linear gradient. The molecular weights of unknown proteins are
determined by measuring the Rf for each protein, assigning it to the appro-
priate value of acrylamide concentration, and reading the apparent molec-
ular weight off the standard curve. Hames 9 has listed the approximate
ranges of molecular weights that can be determined for different acrylam-
ide gradient gels. These include: 7 to 25% acrylamide for molecular
[12] POLYACRYLAMIDE
GEL ELECTROPHORESIS 253

TABLE IX
STOCK SOLUTIONSFOR HDPC (CATIONICDETERGENT)DENATURING
DISCONTINUOUSELECTROPHORESISa

Solution Preparation

1. Acrylamide-bisacrylamide, 37.5 : 0.5 37.5 g of acrylamide, 0.5 g of bisacryl-

amide in 100 ml; store at 20°
2. Phosphate, 1.5 M, pH 2 Adjust pH of phosphoric acid (85% = 14.7
M) to 2.0 with KOH; bring to 1.5 M
3. Ascorbic acid, 1 (w/v) 1.0 g of ascorbic acid in 100 ml
4. Ferrous sulfate (0.003, w/v, 3 mg FeSO4 • 7 H20 in 100 mi
FeSO4 • 7 H20)
5. Hexadecylpyridinium chloride, 595 mg in 10 ml
0.175 M
6. Hydrogen peroxide, 0.03 (w/v), H20: Dilute 1 ml of 30% H202to 1 liter before
use
7. Acrylamide-bisacrylamide, 30 : 2.5 30 g of acrylamide and 2.5 g of bisacryl-
amide in 100 ml; store at 20°
8. Phosphate buffer, 0.5 M, pH 4 Adjust pH of phosphoric acid (85% = 14.7
M) to 4.0 with KOH; dilute to 0.5 M
9. Running buffer Hexadecylpyridinium chloride, 2.5 g, and
glycine, 11.25 g, to pH 3.0 with HDPO4;
bring to 2 liters
10. Sample buffer 0.78 ml of 2-mercaptoethanol, 4.28 g of
sucrose, 10 ml of hexadecylpyridinium
chloride (0.175 M), 0.5 g of pyronine Y,
10 ml of phosphate buffer at pH 4 (No.
8). Bring to 50 ml and adjust pH to 4
11. Separating gel overlay 1.26 ml of phosphate buffer at pH 4 (No.
8) and 0.1 ml of hexadecylpyridinium
chloride (0.175 M); bring to 5 ml

a Modified from a similar system described by A. Amory, F. Foury, and A. Goffeau, J.

Biol. Chem. 255, 9353 (1980), largely according to the modification of M. D. Resh, J.
Biol. Chem. 257, 6978 (1982).

weight 14,000 to 330,000; 5 to 20% for 14,000 to 210,000; 3 to 30% for

13,000 to 950,000.

Discontinuous Denaturing Electrophoresis with Cationic Detergents

O c c a s i o n a l l y , a d e n a t u r i n g e l e c t r o p h o r e s i s s y s t e m is r e q u i r e d for the
s e p a r a t i o n o r d e t e r m i n a t i o n o f the m o l e c u l a r weight o f a p r o t e i n o r its
s u b u n i t s t h a t e x h i b i t a n o m a l o u s b e h a v i o r o n SDS gels. This c a n o c c u r
b e c a u s e o f i n c o m p l e t e S D S b i n d i n g d u e to e x t r e m e s of isoelectric p o i n t o f
the p r o t e i n , e i t h e r acidic o r b a s i c , p r o t e i n g l y c o s y l a t i o n , m e m b r a n e asso-
254 ELECTROPHORESIS [12]

TABLE X
FORMULAS FOR HDPC (CATIONIC DETERGENT) DENATURING
DISCONTINUOUSELECTROPHORESIS

Final percentage of acrylamide

desired

5% 7.5% 10% 15% 20%

A. Lower (separating) gel, for one slab of average size

Mix first
1. Acrylamide-bisacrylamide, 37.5 : 0.5 (ml) 4 6 8 12 16
2. Degassed water (ml) 15.2 13.2 11.2 7.2 3.2
3. Phosphate buffer, pH 2 (ml) 1.5 1.5 1.5 1.5 1.5
4. Ascorbic acid, 1% (ml) 2.3 2.3 2.3 2.3 2.3
5. Ferrous sulfate, 0.003% (ml) 2.3 2.3 2.3 2.3 2.3
Degas x 2 min
6. Hexadecylpyridinium chloride, 0.175 M (ml) 0.4 0.4 0.4 0.4 0.4
Degas x 2 min; then add to polymerize
7. Hydrogen peroxide, 0.03% (ml) 2.3 2.3 2.3 2.3 2.3
Degas x 30 sec, then pour separating gel

B. Upper (stacking) gel, for one slab of average size (4% acrylamide)
Mix first
1. Acrylamide, 30 : 2.5 (ml) 0.67
2. Degassed water (ml) 1.83
3. Phosphate buffer, pH 4 (ml) 1.26
4. Ascorbic acid, 1% (ml) 0.38
5. Ferrous sulfate, 0.003% (ml) 0.38
Degas for 2 min
6. Hexadecylpyridinium chloride, 0.175 M (ml) 0.10
Degas for 2 miD; then add to polymerize
7. Hydrogen peroxide, 0.03% (ml) 0.38
Degas for 30 sec; then pour stacking gel

ciation, or other factors. Because of such difficulties, denaturing systems

have been devised for discontinuous electrophoresis using the cationic
detergents tetradecyltrimethylammonium bromide 17 or cetylthiomethyl-
ammonium bromide is for the separation of the phosphorylated form of
yeast plasma membrane ATPase and histories, respectively. We have
used a modification of the former system using the cationic detergent
hexadecylpyridinium chloride (HDPC); because of errors in the printing
of the method of gel preparation of Amory et al., ~7details of our modifica-
tion of this method are shown in Tables IX and X. Amory et al. 17 have

t7 A. Amory, F. Foury, and A. Goffeau, J. Biol. Chem. 255, 9353 (1980).

ts V. V. Schmatchenko and A. J. Varshavskey, Anal. Biochem. 8.~, 42 (1978).
[12] . POLYACRYLAMIDEGEL ELECTROPHORESIS 255

provided good evidence that in this system, as with SDS-polyacrylamide

gel electrophoresis, proteins are denatured by the detergent and migrate
according to the molecular weights of their HDPC derivatives; the Rf of
proteins subjected to electrophoresis using this system appear to be re-
lated in a linear way to the log~0 of the true molecular weights of the
proteins. Obviously, when a cationic detergent in used, the polarity of the
electrophoresis chamber is reversed compared to SDS electrophoresis;
i.e., the HDPC-proteins migrate toward the cathode instead of the anode.
Otherwise, gel preparation, sample loading, and the like are as described
above for SDS electrophoresis.

Other Discontinuous Denaturing Systems

Several other discontinuous systems have been described for special
applications of detergent electrophoresis. As mentioned above, Swank
and Munkries 15have used a combination of SDS with 8 M urea to separate
and determine the molecular weights of peptides of molecular weights less
than 10,000. For some proteins, a urea-glycerol system has apparently
been useful in maximizing the separation of native and phosphorylated
forms, e.g., with myosin light chains. 19 For separation of histones, a
number of techniques have been used successfully, as reviewed by Hardi-
son and Chalkley2°; these include acetic acid-urea and Triton-acetic
acid-urea systems. For some membrane proteins, nonionic detergents
such as Triton X-100 and sodium deoxycholate can be used to both solubi-
lize the proteins and substitute for SDS in the electrophoresis buffers. 21,2z
Details of these and related techniques can be found in the original refer-
ences and several excellent general r e v i e w s . 9,1°,z3,24

19W. T. Perrie and S. V. Perry, Biochem. J. 19, 31 (1970).

z0R. Hardisonand R. Chalkley,Methods Cell Biol. 17, 235 (1978).
zl B. Dewald, J. T. Dulaney,and O. Touster, this series, Vol. 32, p. 82.
A. C. Newbyand A. Chrambach,Biochem. J. 177, 623 (1979).
A. H. Gordon, "Electrophoresisof Proteins in Polyacrylamideand Starch Gels," Vol. 1,
Pt. 1. North-HollandPubl., Amsterdam, 1975.
z4 R. C. Allen and H. R. Maurer (eds.). "Electrophoresis and Isoelectric Focusing in
PolyacrylamideGel." de Gruyter, Berlin, 1974.
256 ELECTROPHORESIS [ 13]

[13] H i g h - R e s o l u t i o n P r e p a r a t i v e I s o e l e c t r i c F o c u s i n g
By BERTOLD J. RADOLA

Isoelectric focusing presents an ingenious addition to modern separa-

tion methods for the fractionation and isolation of enzymes. Among
charge fractionation methods preparative isoelectric focusing is particu-
larly attractive owing to high resolution and load capacity. The method
concentrates and separates amphoteric substances in a stable pH gradient
according to differences in isoelectric points (pI).l Isoelectric focusing is a
steady-state method that offers two distinct advantages over the more
usual kinetic methods of electrophoresis. (1) After the steady state has
been attained, the process is time-independent so that zone definition
does not deteriorate with further passage of time. In practice some minor
deterioration becomes apparent after extended focusing periods, particu-
larly in gel-stabilized systems, owing to the operation of certain
nonideal effects. 2 In most experiments this deterioration is almost negligi-
ble. (2) Since solutes migrate from all positions of the separation system
toward the final steady-state position, no definite starting zone is re-
quired, and the initial solute mixture may indeed occupy the entire vol-
ume of the separation system. Minor components, sometimes accompa-
nied by major contaminants, are concentrated at their pI and may be
isolated. Isoelectric focusing is the predominant charge fractionation
method, mainly because of high resolution. Proteins differing in their pI
values by 0.001 to 0.01 in pH are resolved by variations of this method. 3-6
The number of protein zones found in one dimension can be of the order
of 100-120. High-resolution preparative isoelectric focusing reaches the
indicated resolution in terms of resolvable pI differences and number of
zones under conditions of high total load. 7,8

i H. Rilbe, in "Isoelectric Focusing" (N. Catsimpoolas, ed.), p. 13. Academic Press, New
York, 1976.
2 H. Rilbe, in "Electrofocusing and Isotachophoresis" (B. J. Radola and D. Graesslin,
eds.), p. 35. de Gruyter, Berlin, 1977.
30. Vesterberg and H. Svensson, Acta Chem. Scand. 20, 820 (1966).
4 R. C. Allen, R. A. Hasley, and R. C. Talamo, Am. J. Clin. Pathol. 62, 732 (1974).
5 R. Charlionet, J. P. Martin, R. Sesbour, P. J. Madec, and F. Lefebvre, J. Chromatogr.
176, 89 (1979).
6 B. Bjellqvist, K. Ek, P. G. Righetti, E. Gianazza, A. G6rg, R. Westermeier, and W.
Postel, J. Biochem. Biophys. Methods 6, 317 (1982).
7 M. D. Frey and B. J. Radola, Electrophoresis 3, 216 (1982).
8 M. Fiieger, M. D. Frey, and B. J. Radola, Electrophoresis 5 (in press).

Copyright © 1984 by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
[13] PREPARATIVE ISOELECTRIC FOCUSING 257

The potential of isoelectric focusing for preparative separations was

recognized early, 9,1° but applications have remained modest. Depending
on the scale of fractionation, all preparative isoelectric focusing tech-
niques can be classified into two categories: (1) techniques for laboratory-
scale fractionation of milligram quantities with 1 g as an upper limit; in
only a few applications has 0.5-1 g of protein actually been separated H-~7;
(2) techniques for large-scale fractionation of gram amounts (and more)
with the potential of industrial applications. 18-2° Both groups of tech-
niques comprise a plethora of diverse systems, but most do not seem to
have been applied owing to unsolved practical problems. The two most
widely used methods of preparative isoelectric focusing use density-gradi-
ent columns ~° and layers of granulated gels.14,2~ The former technique has
remained unchanged over the past years and so have its shortcom-
ings.of. 14,22
Although granulated gels of the Sephadex and BioGel series are also
imperfect, they afford the most challenging potential for preparative iso-
electric focusing. High-resolution isoelectric focusing in layers of granu-
lated gels has evolved from the previously described technique by the
following modifications. Resolution is improved by using high field
strengths (100-300 V/cm) in thin (0.2- to 1-mm) instead of the thick (2- to
12-mm) gel layers employed thus far. Loading is limited to 1 g because of
inherent limitations in heat dissipation in the thicker gel layers that are
required for larger amounts. Increased flexibility is achieved by using dry,
9 H. Rilbe, Protides Biol. Fluids 17, 369 (1970).
10 O. Vesterberg, this series, Vol. 22, p. 559.
11 j. S. Fawcett, in "Isoelectric Focusing" (J. P. Arbuthnott and J. A. Beeley, eds.), p. 23.
Butterworth, London, 1975.
12 H. Rilbe and S. Pettersson, in "Isoelectric Focusing" (J. P. Arbuthnott and J. A. Beeley,
eds.), p. 44. Butterworth, London, 1975.
13 j. p. Arbuthnott, A. C. McNiven, and C. J. Smyth, in "Isoelectric Focusing" (J. P.
Arbuthnott and J. A. Beeley, eds.), p. 212. BuUerworth, London, 1975.
14 B. J. Radola, in "Isoelectric Focusing" (N. Catsimpoolas, ed.), p. 119. Academic Press,
New York, 1976.
15 j. Bours, M. Grabers, and O. Hockwin, in "Electrophoresis '79" (B. J. Radola, ed.),
p. 529. de Gruyter, Berlin, 1980.
16 R. van Driel, J. Mol. Biol. 138, 27 (1980).
17 j. p. Liberty and M. S. Miller, J. Biol. Chem. 255, 1023 (1980).
18 j. S. Fawcett, in "Isoelectric Focusing" (N. Catsimpoolas, ed.), p. 173. Academic Press,
New York, 1976.
19 M. Bier, N. B. Egen, T. T. Allgyer, G. E. Twitty, and R. A. Mosher, in "Peptides:
Structure and Biological Function" (E. Gross and J. Meinenhofer, eds.), p. 79. Pierce
Chemical, Rockford, Illinois, 1979.
2o M. Jonsson and H. Rilbe, Electrophoresis 1, 3 (1980).
21 B. J. Radola, Biochim. Biophys. Acta 386, 181 (1975).
22 M. Jonsson, J. St~hlberg, and S. Fredriksson, Electrophoresis 1, 113 (1980).
258 ELECTROPHORESIS [13]
rehydratable gels 8 on polyester films instead of wet gel layers on glass
plates or in troughs. Rapid focusing with resultant short residence times
of the separated samples in the gel layer is possible using shorter separa-
tion distances, extended prefocusing and cascades, or a combination of
these approaches. Detection of proteins and enzymes is improved by new
print techniques utilizing cellulose acetate membranes, 7 trichloroacetic
acid-impregnated paper strips, 8 or high-resolution enzyme visualization
techniques. 23The time for detection by these techniques is shortened to a
few minutes.

Apparatus
Most horizontal systems for analytical flatbed isoelectric focusing can
be used without adaptation or with only minor adaptation for high-resolu-
tion preparative isoelectric focusing. The main requirements for prepara-
tive work are high efficiency of the cooling system; adequate loading
capacity provided by the surface of the cooling plate; movable electrodes
for increased flexibility in the choice of separation distance; and a power
supply capable of yielding 3000-6000 V. The horizontal flatbed apparatus
excels over systems with a cylindrical geometry for either density gradi-
ents ~° or polyacrylamide gels24-26 in more efficient heat dissipation and
versatility. Cooling is most favorable with thin layers owing to a high ratio
of cooling surface to total volume. Some features of the most widely used
advanced systems for isoelectric focusing are summarized in Table I. The
maximum voltage of 3000 V provided by most commercial systems may
prove to be too low for high-resolution preparative isoelectric focusing.
An option is to combine the flatbed apparatus with power supplies manu-
factured locally or available for other electrophoretic techniques, e.g.,
high-voltage paper electrophoresis or nucleic acid sequencing. In working
with voltages of 6000-8000 V, the safety features of the commercially
available flatbed apparatus should be carefully reevaluated. Troughs and
fractionation grids are obsolete in high-resolution preparative isoelectric
focusing. The 0.2- to 1-mm layers are prepared more conveniently on plas-
tic supports 8 or thin glass plates 7 than in troughs. The grid is incompatible
with high-resolution preparative isoelectric focusing because it divides
the layer arbitrarily into a number of segments with inevitable remixing of
components separated in situ.
23 A. Kinzkofer and B. J. Radola, Electrophoresis 4 (in press).
A. Chrambach, T. M. Jovin, P. J. Svendsen, and D. Rodbard, in "Methods of Protein
Separation" (N. Catsimpoolas, ed.), Vol. 2, p. 27. Plenum, New York, 1976.
25 B. An der Lan and A. Chramba~h, in "Gel Electrophoresis of Proteins: A Practical
Approach" (B. D. Hames and D. Rickwood, eds.), p. 157. IRL Press, London, 1981.
26 N. Y. Nguyen and A. Chrambach, J. Biochem. Biophys. Methods 1, 171 (1979).
[13] PREPARATIVE ISOELECTRIC FOCUSING 259

TABLE I
FLATBED APPARATUSFOR ISOELECTRIC FOCUSING

Power
supplya and Maximum
maximum Focusing Cooling loadingb
Company voltage unit plate (mg protein)

Bio-Rad Laboratories, 3000/300, 1415 12.5 x 22 cm 270

Richmond, CA 3000 V 1415 12.5 x 43 cm 540
(Glass and
plastic)
Desaga, Heidelberg, Desatronic, Mediphor 12.5 x 26.5 cm 330
West Germany 6000 V Desaphor 12.5 x 26.5 cm --
20 x 26.5 cm 530
16.5 x 30 cm 800
26.5 x 40 cm 1060
(Aluminum
coated with
epoxide)
LKB, Bromma, 2197c, Ultrophor 12.6 x 26.1 cm 330
Sweden 2500 V (Aluminum
insulated
with glass)
Pharmacia, Uppsala, ECPS d, FBE 3000 25 x 25 cm 625
Sweden 3000/150, (Aluminum
3000 V with
replaceable
Teflon
coating)

a All power supplies have adequate power outputs (100-300 W) and operate in three
modes: constant power, constant voltage, or constant current.
b Calculated for high-resolution isoelectric focusing at a loading capacity of 10 mg of
protein per milliliter of focusing gel volume in a 1-mm gel layer. All systems can be
operated at 5-10 times higher total loading in 5-10-mm gel layers at lower final field
strengths and inferior resolution. High resolution and increased total loading can be
achieved by cascade focusing combining in two steps, isoelectric focusing in thick and
thin gel layers (for details see text).
c Macrodrive power supply for 5000 V is available for nucleic acid sequencing.
d Optional extra: the VH-I integrates volts with time and records volt-hours. This en-
ables running conditions to be controlled accurately, particularly when the voltage
changes with time.

Generation of pH Gradients

T h e g e n e r a t i o n o f s t a b l e p H g r a d i e n t s is t h e k e y p r o b l e m i n i s o e l e c t r i c
f o c u s i n g . O n l y t h r e e a p p r o a c h e s ( 1 - 3 b e l o w ) a r e o f p r a c t i c a l v a l u e at
present.
260 ELECTROPHORESIS [13]

1. pH gradients are usually formed with the aid of mixtures of syn-

thetic carrier ampholytes. 3 Several commercial products [Ampholine
(LKB), Servalyt (Serva), and Pharmalyte (Pharmacia)] are available, dif-
fering with respect to synthesis 27-29 and physicochemical properties. 3° In
addition to wide-range carrier ampholytes for the generation of steep pH
gradients covering 5-6 pH units (e.g., the range pH 3-10), restricted pH
ranges are offered covering 2-2.5 pH units, 1 pH unit, and even only 0.5
pH unit (Serva). Servalyt T carrier ampholytes, which are much cheaper
than other products, appear to be particularly well suited for preparative
isoelectric focusing. Small amounts of colored material present in Serva-
lyt T, as well in most laboratory-made preparations, 3~-34do not interfere
with preparative isoelectric focusing, with the rare exception of protein
detection at 280 nm in situ. 21 Carrier ampholytes of any restricted pH
range can be easily prepared from the commercial products, or even a
blend of different products, for an increased number of carrier ampholyte
species, 35 by preparative isoelectric focusing. 21,36
Most of the desirable properties of carrier ampholytes defined for
analytical work 37 apply also to preparative isoelectric focusing. A critical
property is the molecular weight distribution of carrier ampholytes in
those applications in which they have to be removed from the focused
proteins. In contrast to many manufacturers' assertions, such removal is
not as simple a procedure as is suggested. Nonstandard batches, uncon-
trolled changes on storage, and the presence of species of high molecular
weight may seriously hamper the removal of carrier ampholytes from
focused proteins. 38-41 If high-molecular-weight species are detected, e.g.,

27 O. Vesterberg, Acta Chem. Scand. 23, 2653 (1969).

28 N. Grubhofer and C. Borja, in "Electrofocusing and Isotachophoresis" (B. J. Radola and
D. Graesslin, eds.), p. 1 ll. de Gruyter, Berlin, 1977.
29 K. W. Williams and L. S6derberg, Int. Lab. 1, 45 (1979).
30 W. J. Gelsema, C. L. de Ligny, and N. G. van der Veen, J. Chromatogr. 173, 33 (1979).
31 S. N. Vinogrador, S. Lowenkron, H. R. Andonian, H. R. Baghshaw, J. Felgenhauer, and
J. Pak, Biochem. Biophys. Res. Commun. 54, 501 (1973).
32 R. Charlionet, J. P. Martin, R. Sesbou6, P. J. Madec, and F. Lefebvre, J. Chromatogr.
176, 89 (1979).
33 W. W. Just, Anal. Biochem. 102, 134 (1980).
34 S. B. Binion and L. S. Rodkey, Anal. Biochem. 112, 362 (1981).
35 B. J. Thompson, M. J. Dunn, A. H. M. Burghes, and V. Dubowitz, Electrophoresis 3, 307
(1982).
36 A. Kinzkofer and B. J. Radola, Electrophoresis 2, 174 (1981).
37 O. Vesterberg, in "Isoelectric Focusing" (N. Catsimpoolas, ed.), p. 53. Academic Press,
New York, 1976.
38 G. Baumann and A. Chrambach, Anal. Biochem. 64, 530 (1975).
39 W. Otavsky and J. W. Drysdale, Anal. Biochem. 65, 533 (1975).
4o K. Goerth and B. J. Radola, in "Electrophoresis '79" (B. J. Radola, ed.), p. 955.
de Gruyter, Berlin, 1980.
[13] PREPARATIVE ISOELECTRIC FOCUSING 261

with the aid of thin-layer gel chromatography,4°,4~ their removal is recom-

mended; gel chromatography or, preferably, ultrafiltration through mem-
branes with appropriate retention characteristics4° are suggested.
2. Buffer electrofocusing uses mixtures of amphoteric or nonampho-
teric buffers as carrier constituents. 25,4z With mixtures containing 2-14
constituents, narrow pH ranges are generated, and with a 47-component
buffer mixture, gradients between pH 3 and 10 can be obtained. 43 Most of
the work with buffer electrofocusing is confined to analytical separations,
but their utility for preparative work has also been demonstrated. 26,44
There appears to be no limit to the choice of the desired pH range in
buffer electrofocusing. Addition of acidic constituents to the buffer mix-
ture causes a shift in the acidic direction, whereas basic constituents shift
toward the basic direction. A single constituent, added in large amounts
compared to the concentration of other constituents, flattens the pH gra-
dient in the vicinity of its steady-state position. 45
Buffer electrofocusing offers a number of advantages over isoelectric
focusing utilizing synthetic carrier ampholytes. 43 The average molecular
weight of the buffer constituents is 150, so that these components can be
easily removed. There is no evidence that the buffer constituents bind to
the proteins under electrofocusing conditions, although interactions be-
tween buffer constituents cannot be excluded. Buffer electrofocusing pro-
vides a higher degree of reproducibility due to the defined composition of
the buffer mixtures.
3. Immobilized pH gradients are prepared with the aid of acidic and
basic acryloyl derivatives with defined pK values. 6 These derivatives
are used to generate two buffer solutions, which are linearly mixed
and copolymerized with acrylamide and N,N'-methylene bisacrylamide.
When an electric field is applied to the immobilized pH gradient,
amphoteric substances are focused within their pl ranges while other
charged solutes collect at the electrodes. Immobilized pH gradients
are claimed to overcome problems associated with pH drift and irregulari-
ties in zone formation. Their major advantage is increased resolution in
extremely flat pH gradients. Their disadvantages include prolonged focus-
ing time, high electroendoosmosis due to H ÷ and OH- imparting a net
charge to the medium at below pH 5 and above pH 9, and longer time for
sample entry into the gel with the risk of protein precipitation and re-

41 B. J. Radola, Electrophoresis 1, 43 (1980).

42 A. Chrambach, L. Hjelmeland, and N. Y. Nguyen, in "Electrophoresis '79" (B. J. Ra-
dola, ed.), p. 3. de Gruyter, Berlin, 1980.
43 C. B. Cuono and G. A. Chapo, Electrophoresis 3, 65 (1982).
44 R. L. Prestidge and M. T. W. Hearn, Anal. Biochem. 97, 95 (1979).
45 M. L. Caspers, Y. Posey, and R. K. Brown, Anal. Biochem. 79, 166 (1977).
262 ELECTROPHORESIS [ 13]
stricted flexibility with respect to sample application and gel matrix selec-
tion. These disadvantages limit the utility of immobilized pH gradients for
preparative work. An attractive feature is that neither carrier ampholytes
nor buffer constituents need be removed from the focused proteins. How-
ever, elution of proteins may be expected to be difficult, and soluble
nonproteinaceous impurities derived from the polyacrylamide gel are
likely to contaminate the eluate. 46

Gel Matrices
Isoelectric focusing requires a nonrestrictive anticonvective gel. Non-
restrictive gels are imperative because molecular sieving will retard mi-
gration of proteins, resulting in prolonged focusing periods. The migration
velocities decrease as proteins approach their p/s, and any molecular
sieving effect not only will retard the migration of the proteins, but may
appear to halt it entirely, giving rise to fallacious pI values. 25 However, a
steady state need not be attained because many proteins are sufficiently
separated under nonequilibrium conditions. The following gels, arranged
according to decreasing restrictiveness, are suitable for anticonvective
stabilization: polyacrylamide gels for proteins with molecular weights up
to 500,000; agarose for molecules up to several millions and particles 30-
80 nm in radius, 47 and granulated gels, which may be used for all mole-
cules but not for cells.

Polyacrylamide Gel
Continuously polymerized polyacrylamide gels dominate in analytical
applications of isoelectric focusing, but they have failed to attract much
interest in preparative separations. Cylindrical gels have a poor geometry
for heat dissipation and are also afflicted with problems of wall adherence
that can disturb the operation of preparative columns. 24,25 Flatbed poly-
acrylamide gels were only occasionally used for preparative separa-
tions. 48,49Gels of rather low total monomer and cross-linking concentra-
tions are usually used in analytical isoelectric focusing, and a gel
composed of 5% T and 3% CBis (cross-linked with N,N'-methylene
bisacrylamide 5°) has become particularly popular) 1 A drawback of poly-
acrylamide gels is the difficult recovery of separated proteins. 46 Proteins

46 N. Y. Nguyen, J. DiFonzo, and A. Chrambach, Anal. Biochem. 1116, 78 (1980).

47 p. Serwer and S. J. Hayes, Electrophoresis 3, 80 (1982).
'~ D. H. Leaback and A. C. Rutter, Biochem. Biophys. Res. Commun. 32, 447 (1968).
49 D. Graesslin, H. C. Weise, and M. Rick, Anal. Biochem. 71, 492 (1976).
50 S. Hjertrn, Arch. Biochem. Biophys. Suppl. 1, 147 (1962).
51 O. Vesterberg, Biochim. Biophys. Acta 257, 11 (1972).
[13] PREPARATIVE ISOELECTRIC FOCUSING 263

that are electrophoretically extracted from polyacrylamide gels may be

contaminated with nonproteinaceous impurities.

Agarose
Agarose with low electroendoosmosis has been proposed as an alter-
native to polyacrylamide gels in analytical isoelectric focusing of high-
molecular-weight proteins) 2,53 In a few reports, agarose has also been
used for preparative isoelectric focusing) 4-56 High resolution, easy han-
dling, nontoxicity, and absence of molecular sieving for high-molecular-
weight proteins are some of its advantages. Although the purified or
charge-balanced agaroses are claimed to fulfill many requirements of a
good anticonvective support, there is evidence for certain disadvantages.
A severe degree of surface flooding at the cathode, water transport to
both electrodes resulting in distorted pH gradients, protein loss into the
water accumulated on the gel surface, and protein trailing at the edges of
the gel are some of the shortcomings) 5,54,57 To overcome these draw-
backs, a composite agarose-Sephadex matrix was developed)7 Higher
field strength than in agarose could be used for improved resolution. In
analytical experiments, photopolymerized composite agarose (0.5%)-
polyacrylamide gels (2.5%) proved to be superior to each of the single gels
for the analysis of crude tissue extracts containing a wide molecular
range) 8 Preparative isoelectric focusing of immunoglobulins was im-
proved by adding 0.5% non-cross-linked polyacrylamide to 1% agarose) 9
A drawback of all agarose-containing gels is the unsatisfactory protein
recovery from macerated gels, which may have to be digested for ex-
tended periods with a mixture of agarase-hemicellulase) 6

Granulated Gels
Horizontal layers of granulated gels of the Sephadex or BioGel type
were introduced for anticonvective stabilization of the pH gradient with
the intent of overcoming some of the limitations of both the density gradi-

52 A. Ros6n, K. Ek, and P./~man, J. Immunol. Methods 28, 1 (1979).

s3 C. A. Saravis, M. O'Brien, and N. Zamcheck, J. Immunol. Methods 29, 91, 97 (1979).
54 G. C. Ebers, G. P. Rice, and H. Armstrong, J. Immunol. Methods 37, 315 (1980).
55 C. Chapuis-Cellier and P. Arnaud, Anal. Biochem. 113, 325 (1981).
56 W. D. Cantarow, C. A. Saravis, D. V. Ives, and N. Zamcheck, Electrophoresis 3, 84
(1982).
57 A. Manrique and M. Lasky, Electrophoresis 2, 315 (1981).
s8 E. A. Quindlen, P. E. McKeever, and P. L. Kornblith, in "Electrophoresis '81" (R. C.
Allen and P. Arnaud, eds.), p. 539. de Gruyter, Berlin, 1981.
59 R. McLachlan and F. N. Cornell, in "Electrophoresis '82" (D. Stathakos. ed.), 13. 697.
de Gruyter, Berlin, 1983.
264 ELECTROPHORESIS [13]
ent technique and continuously polymerized polyacrylamide gels. 2L6°,61
Granulated gels have also been used for preparative isoelectric focusing in
columns 62 and in continuous-flow configurations, ~8'63 but the horizontal
systems offer distinct advantages over the vertical, closed systems. Gran-
ulated gels excel over other gel matrices in a number of properties: high
load capacity, quantitative elution of the focused proteins from the gel,
simple handling, absence of molecular sieving for high-molecular-weight
proteins (making these gel particularly suitable for isoelectric focusing of
molecules >500,000), and availability of granulated gels ready for use,
partly in a prewashed form, with well-defined chemical and physical prop-
erties. Their inertness toward biopolymers under a wide range of condi-
tions is well established owing to their widespread use in gel chromatogra-
phy. The focused proteins and enzymes can be conveniently and rapidly
located in the gel with the print technique. Drawbacks to granulated gels
have been reported. Preparation of a gel bed with optimum consistency
has been considered difficult55:,64 or laborious. 65 Mixtures of Sephadex
and Pevikon 64 [a copolymer of poly(vinyl chloride) and poly(vinyl ace-
tate) 66] and Pevikon alone 65 were suggested as possible supports. Inferior
resolution and bad printing properties are shortcomings of Pevikon-con-
taining layers. Loss of resolution due to diffusion on protein detection
with the paper print technique was a drawback of granulated gels in a
comparative study of different gel matrices .57 With the new printing tech-
niques, there is essentially no loss in resolution. 7,8
Sephadex G-200 is the gel of choice for most applications. It exhibits
the best load capacity and handling properties and the highest water re-
gain and is most economical. 7,8 Sephadex G-200 was not practical owing
to unsatisfactory printing properties, 21 but the new printing techniques 7,8
have overcome this limitation. Sephacryl S-200, prepared by covalently
cross-linking allyl dextran with N,N'-methylene bisacrylamide, can be
handled as conveniently as Sephadex G-200 and also has good printing
properties but inferior resolution. The higher G-numbered Sephadex gels
may contain as much as 10% free dextran, which would contaminate the
eluates. The enzymically resistant polyacrylamide gel BioGel P-60 is po-
tentially useful, superior to Sephadex in work with crude preparations of

60 B. J. Radola, Biochim. Biophys. Acta 295, 412 (1973).

61 B. J. Radola, Ann. N.Y. Acad. Sci. 209, 127 (1973).
62 T. J. O'Brien, H. H. Liebke, H. S. Cheung, and L. K. Johnson, Anal. Biochem. 72, 38
(1976).
63 G. Hedenskog, J. Chromatogr. 107, 91 (1975).
64 W. I. Otavsky, T. Bell, C. Saravis, and J. W. Drysdale, Anal. Biochem. 78, 301 (1977).
65 B. M. Harpel and F. Kueppers, Anal. Biochem. 104, 173 (1980).
66 H. J. Miiller-Eberhard, Scand. J. Clin. Lab. Invest. 12, 33 (1960).
[13] PREPARATIVE ISOELECTRIC FOCUSING 265

cellulases and hemicellulases. 7 All granulated gels must be extensively

washed with distilled water before use to remove charged solutes interfer-
ing with the formation of the pH gradient. 67 Optimum results are obtained
with gels with a dry bead diameter of 10-40/xm ("Superfine" or minus
400 mesh).

Rehydratable Gels
Until recently, granulated gels were prepared as wet layers on a glass
plate or in a trough 21; they could not be stored. Preparation of rehydrat-
able layers is simple and allows storage. After spreading the gel suspen-
sion of the correct consistency over a support, the gel is dried in air. The
dry gel firmly adheres to the support, is mechanically stable, and can be
preserved indefinitely.
Instead of glass plates or troughs, the rehydratable gel layers are prefer-
ably prepared on a plastic film. Best results are obtained with 100-/~m
polyester films (Mylar D, Du Pont) treated with alkali to impart hydro-
philic properties to the film. 41 Two commercially available supports
(GelBond for agarose, from Marine Colloids; and Gel-Fix, from Serva)
are also suitable.
Rehydratable gels can be prepared with carrier ampholytes, which,
owing to their hygroscopic properties, ensure the residual moisture neces-
sary for storage.
Even greater versatility is provided by preparing "empty" gels, gels
without added carrier ampholytes, but supplemented with 1-2% glycerol.
Before use, the rehydratable gels, containing carrier ampholytes, are
sprayed with an amount of water, calculated from the surface and thick-
ness of the gel layer; empty gels are sprayed with a 2-3% solution of
carrier ampholytes. Any formulation of carrier ampholytes, supple-
mented if necessary with such additives as urea, can be used for rehydra-
tion.

Load Capacity
In order to compare isoelectric focusing in systems employing a differ-
ent geometry, pH gradient, or other forms of anticonvective stabilization,
load capacity is defined as the amount of protein (in milligrams) per millili-
ter of focusing volume. 21 Load capacity is calculated by dividing the total
load by total volume of the gel layer. The appearance of straight zones is
used as the criterion for determining the highest permissible protein load
capacity. Overloading results in irregular zones, which, with additional
67 A. Winter, in "Electrofocusing and Isotachophoresis" (B. J. Radola and D. Graesslin,
eds.), p. 433. de Gruyter, Berlin, 1977.
266 ELECTROPHORESIS [ 13]

protein, cause the major zones to decay into droplets. The irregularities of
the major zones in some parts of the gel layer usually have no detrimental
effect on zone definition of minor components in other parts of the gel.
Thus, load capacity for total protein is much higher when minor compo-
nents are to be separated from an excess of major components rather than
when all protein zones have to be well defined. 21
Load capacity has been determined for natural and artificial mixtures
of proteins as well as for single protein and carrier ampholytes of different
pH ranges using 0.3- to 1-mm layers of Sephadex G-200 and BioGel P-60.
The highest loads are attained for protein mixtures with a uniform distri-
bution of protein zones over a wide pH range, e.g., crude tissue extracts.
The decisive parameter for preparative systems is the amount of material
to be fractionated, which, depending on the load capacity of the system,
requires a specific focusing volume. In early work with granulated gels,
thick (2-mm) layers w e r e u s e d , 68 and subsequent separations also em-
ployed thick layers of up to 12 m m . 14'21'61 The notion that preparative
isoelectric focusing requires thick layers became established. More recent
work appears to have made the thick layer obsolete. 7,8 In high-resolution
preparative isoelectric focusing, thin layers afford the following advan-
tages: higher field strengths due to more efficient heat dissipation; better
resolution as the result of higher field strength and absence of skew zones;
shorter focusing time; easier and more rapid gel preparation; better mois-
ture control; reduced cost due to lower consumption of carrier ampho-
lytes, buffer, and gels; improved recovery and higher concentrations of
the recovered proteins; and reduced risk of inactivation of labile sub-
stances, e.g., as a result of chelating activity of carrier ampholytes. High-
resolution preparative isoelectric focusing should be carried out at as high
a load level as possible. The optimum capacity is 3-5 mg protein per
milliliter of gel bed volume, 10-20 mg/ml remaining well tolerated al-
though this is dependent on the specific properties of the separated mate-
rial. An increase of total load to more than 1 g can be achieved by select-
ing an apparatus with larger surface of the cooling plate (see Table II) or,
preferably, by applying a two-step cascade with prefractionation of up to
several grams in the first step, followed by high-resolution isoelectric
focusing of selected parts of the gel in the second step.

Resolution

The excellent resolution of analytical isoelectric focusing is a chal-

lenge for any preparative focusing method. Resolution in isoelectric fo-
68 B. J. Radola, Biochim. Biophys. Acta 194, 335 (1969).
[13] PREPARATIVE ISOELECTRIC FOCUSING 267

cusing depends on several factors: design of apparatus, anticonvective

stabilization of the pH gradient, parameters of the separation process
(field strength and shallowness of the pH gradient), and properties of the
separated material (diffusion coefficient and mobility slope in the vicinity
of pI).~ Apparatus in which a continuous pH gradient is built up, e.g., all
gel or density gradient systems, afford superior resolution to instruments
with a segmented design such as zone convection isoelectric focusing69-71
and multicompartment electrolysis apparatus. 2°,72The popularity of ana-
lytical gel-stabilized systems stems from an operational advantage of gel
matrices, namely, rapid fixation of the focusing pattern by conversion of
the diffusible species into insoluble precipitates. 73 The main drawback of
the segmented design is mixing of the zones within a single compartment
and retarded equilibration; the latter may result from the use of mem-
branes or other devices.
Resolution can be influenced by the field strength and the shallowness
of the pH gradient. 1 Whereas high field strengths are being increasingly
applied in analytical work, 36,74,75preparative isoelectric focusing has been
carried out at more moderate field strengths, mainly because of difficult
heat dissipation. The added advantage of high field strength is a shorter
focusing time; long focusing times repeatedly have been held as a short-
coming of preparative isoelectric focusing. 18'22'56 Extended residence
times of labile substances at extreme pH values or close to their pI incurs
the risk of inactivation. By reducing the thickness of the gel layer, field
strengths of I00-500 V/cm can be applied in preparative isoelectric focus-
ing with resultant improved resolution over a drastically shortened period
of time. 7,8 In prefocused gels over a 10-cm separation distance, the resi-
dence time of the sample is decreased to only 30-40 min under steady-
state conditions for granulated gels. The most conspicuous effect of high
field strength is improved resolution. Proteins differing by only 0.01 to
0.15 pH are resolved on 40-cm gels using a pH 4 to 6 gradient. 7 This
resolution had been achieved previously only with the analytical system.
In preparative experiments, two components should not only be visibly
resolved but also be amenable to elution from the gel layer by a simple

69 E. Valmet, Sci. Tools 15, 8 (1969).

7o j. Bours, in "Isoelectric Focusing" (N. Catsimpoolas, ed.), p. 209. Academic Press, New
York, 1976.
71 R. Quast, in "Electrokinetic Separation Methods" (P, G. Righetti, C. J. van Oss, and J.
W. Vanderhoff, eds.), p. 221. Elsevier/North-Holland, Amsterdam, 1979.
72 M. Jonsson and S. Fredriksson, Electrophoresis 2, 193 (1981).
73 M. D. Frey and B. J. Radola, Electrophoresis 3, 27 (1982).
74 R. C. Allen, Electrophoresis 1, 32 (1980).
75 T. L~ts, I. Olsson, and L. S6derberg, Anal. Biochem. 101, 449 (1980).
268 ELECTROPHORESIS [ 13]
slicing technique. The main argument for using longer separation dis-
tances is that zones can be handled more easily on elution.
There are several approaches toward improved resolution by flatten-
ing the pH gradient.
1. Selection of narrow-range carrier ampholytes. These are either
commercially available or can be prepared by fractionation of the com-
mercial products by preparative isoelectric focusing. 21,36
2. Increased separation distance. For longer separation distances the
gradient is flattened linearly and resolution is improved if focusing is
carried out at the same field strength. For a 40-cm separation distance, pH
gradients are flattened from 0.15 pH/cm to 0.025 pH/cm for wide range
and 0.5 to 1 pH range carrier ampholytes, respectively.
3. Cascade focusing combines in a two-step or multistep procedure,
prefractionation of the sample with a fractionation of the carrier ampho-
lytes. In the first step the sample is focused at a high load and with lower
resolution in a steep pH gradient. In subsequent steps, parts of the gel
layer, enriched with the components of interest, are transferred to a pre-
focused narrow-range pH gradient. The carrier ampholytes, transferred
with the sample, flatten the pH gradient and greatly improve resolution.
In an experiment with a crude fungal enzyme resolution was improved to
as little as 0.0013-0.005 pH. 8
4. Addition of separators. Single or multiple amphoteric substances
added in large amounts (5-50 mg/ml) to carrier ampholytes flatten the pH
gradient in the vicinity of their steady-state positions. At present, manipu-
lation of the pH gradient has to be conducted in a systematic but empirical
manner because our understanding of the mechanism of gradient forma-
tion remains inadequate for predicting the gradient from the pK values of
the separatorsY
5. Buffer isoelectric focusing. With some buffer mixtures in cylindri-
cal polyacrylamide gels, using a 14-cm separation distance, the pH gradi-
ent was flattened to 0.02-0.04 p H / c m . 26,76 By addition or deletion of buffer
constituents, the course of the pH gradient may be manipulated.
6. Local increase of gel v o l u m e 77 o r concentrationTM of carrier ampho-
lytes. Both approaches have been described for analytical isoelectric fo-
cusing, but with 0.2- to 0.3-mm layers they could be useful also in prepar-
ative separations.
7. Continuous displacement and pH of the anolyte. The pH gradient

76 N. Y. Nguyen and A. Chrambach, Electrophoresis 1, 14 (1980).

77 K. Altland and M. Kaempfer, Eiectrophoresis 1, 57 (1980).
78 T. L ~ s and I. Olsson, Anal. Biochem. 114, 167 (1981).
[13] PREPARATIVE ISOELECTRIC FOCUSING 269

can be flattened by suitable choice of anolyte and catholyte, both chosen

so that they fall within the pH range of the gradient. 79-81
8. Immobilized pH gradients. With the aid of Immobilines (LKB) the
most shallow pH gradients can be created, thereby improving resolution
and increasing the distance between separated zones. This will facilitate
isolation of zones without contamination by adjacent zones. Although
potentially interesting, limitations to the use of immobilized pH gradients
exist (see section on granulated gels, above).
The advantages of increased field strength and flat pH gradients are
documented for analytical isoelectric focusing. 36,74,75The few data avail-
able indicate that the approaches outlined should be equally successful for
preparative separation. The degree to which gradient shallowness is desir-
able remains unclear. 25 Retarded migration of proteins with flat titration
curves near their pI, impeded entrance of sample into the gel, increased
band width, and prolonged focusing time are all negative factors of iso-
electric focusing in very flat pH gradients. While annoying in the most
restrictive gel matrix, i.e., polyacrylamide gels of standard composition,
some of these limitations may be alleviated by optimizing the separation
with respect to selection of gel matrix, field strength, and mode of genera-
tion of the pH gradient. Even more important is to disregard the notion
that preparative isoelectric focusing has to attain the steady state. By
adopting the strategy outlined in this section, good separation may be
expected for most samples also under nonequilibrium conditions.

Detection of Proteins
Prior to recovery, the focused proteins and enzymes must be located
in the gel layer by detection techniques that should be rapid, simple, and
preferably nondestructive. Speed is important because keeping the gels
without voltage or at reduced voltage will broaden the zones as a result of
diffusion, an effect less conspicuous with long separation distances. The
gel layer may be rapidly frozen if this is compatible with the separated
material. There are several approaches to locating proteins and enzymes
in horizontal gel layers (see also this volume [28]-[32]).
1. Transparent zones. At high protein loading the major components
are visible in the gel layer after focusing as transparent zones owing to

79 A. G. McCornick, L. E. M. Miles, and A. Chrambach, Anal. Biochem. 75, 314 (1976).

80 A. G. McComick, H. Wachslicht, and A. Chrambach, Anal. Biochem. 85, 209 (1978).
81 B. An der Lan and A. Chrambach, Electrophoresis 1, 23 (1980).
270 ELECTROPHORESIS [13]
changes in refraction relative to the surrounding gel. 7"8'21 This permits
rapid visual identification, pI determination, and direct isolation by gel
slicing.
2. Membrane and paper printing is the most versatile technique for
protein location. Originally, paper prints were obtained with chromato-
graphic papers. The drawbacks of chromatographic paper for printing are
diffuse zones resulting from the coarse structure of paper relative to the
gel matrix, limited applicability to some gel matrices, e.g., Sephadex G-
200, and long visualization time. These shortcomings are overcome by
cellulose acetate membranes 7 or trichloroacetic acid-impregnated paper:
Ponceau S, rather than the more sensitive triphenylmethane d y e s , 41,73 is
preferred for staining at high protein load. With membrane printing, the
total time required for fixation, staining, and destaining is 2-3 min. De-
staining depends on the chemical properties of the carrier ampholytes 4~
and is most rapid for Servalyt. Narrow (1-2 cm) strips are sufficient for
printing, only small amounts of proteins being removed. Membrane print-
ing is nondestructive but the trichloroacetic acid from the paper exerts a
fixative effect on the proteins in the gel layer. The stained prints are a
convenient document that can be preserved easily and evaluated densito-
metrically. Location of radioactivity in the print by a strip scanner has
been reported. 82
3. Ultraviolet densitometry is more sensitive than zone transparency,
but less sensitive than staining of a print, and requires an expensive
instrument.
4. A topographic method is based on the fluorescence of Servalyt
carrier ampholytes in a paper print. 83-85
5. Fluorescence. Without printing, proteins can be visualized in the
gel layer with 8-anilino-l-naphthalenesulfonic acid by spraying a water
solution of the reagent on the gel surface. 86
6. Enzyme visualization. Substrate-impregnated p a p e r s , 61,87 dimen-
sionally stable polyamide membranes, 23 or 100- to 200-/xm ultrathin
agarose layers 23 containing a high concentration of the substrate and cou-
pling dyes can be used for enzyme location.

A. J. MacGillivray and D. Rickwood, in "Isoelectric Focusing" (J. P. Arbuthnott and J.

A. Beeley, eds.), p. 254. Butterworth, London, 1975.
s3 j. Bonitati, J. Biochem. Biophys. Methods 2, 341 (1980).
J. Bonitati, J. Biochem. Biophys. Methods 4, 49 (1981).
J. Bonitati, B. Sabatino, and J. B. Van Liew, Electrophoresis 3, 326 (1982).
86 W. E. Merz, U. Hilgenfeldt, M. DOrner, and R. Brossmer, Hoppe-Seyler's Z. Physiol.
Chem. 355, 1035 (1975).
s7 H. Delincre and B. J. Radola, Anal. Biochem. 48, 536 (1972).
[13] PREPARATIVE ISOELECTRIC FOCUSING 271

7. Activity determination in eluates. In those cases in Which visualiza-

tion reactions are not available, the enzyme activity has to be determined
in eluates of gel segments, with some unavoidable zone remixing within a
single segment.

Recovery
Recovery in preparative isoelectric focusing will depend on a number
of factors that are related either to the proper separation, including elution
from the gel, or to additional steps that may be necessary for removal of
the carrier ampholytes or concentration of the isolated fractions. Elution
from granulated gels is simple, rapid, and quantitative. A loss of recovery
at this step is negligible in comparison with elution from compact poly-
acrylamide gels. 46 The recovery of isoelectrically homogeneous proteins
was studied in preparative refocusing experiments for which protein re-
covery of 85-92% was found. 61 For crude protein mixtures recoveries of
80-90% were determined by eluting all proteins simultaneously from gel
strips removed lengthwise from the layer21; this approach gives a more
reliable estimate than procedures in which protein recovery is calculated
by summation of the protein content of individual isolated fractions, as The
total activity in a portion of the gel layer is therefore a means of checking
inactivation inherent to the separation process. Recovery may depend
strongly on load capacity s9 ; for Pronase E at loads from 0.5 to 10 mg per
milliliter of gel suspension, recovery of activity increases with increasing
load from 14 to 80%.
The chelating properties 9° of the carrier ampholytes have been impli-
cated as resulting in the dependence of enzyme recovery on the ratio of
enzyme to the carrier ampholytes. 89
Since extreme pH values during focusing may cause denaturation, the
sample is applied at a sufficient distance from the electrodes. Components
that are focused at extreme pH values should be protected by efficient
temperature control and a short focusing period. The risk of denaturation
can be lowered somewhat by establishing a pH gradient, prefocusing in
the absence of the sample. The residence time of the sample is thereby
reduced, in some instances at the expense of not reaching steady-state
conditions.

H. Delinc6e and B. J. Radola, Fur. J. Biochem. $2, 321 (1975).

B. J. Radola, in "Isoelectric Focusing" (J. P. Arbuthnott and J. A. Beeley, eds.), p. 182.
Butterworth, London, 1975.
H. Davies, Protides Biol. Fluids 17, 389 (1970).
272 ELECTROPHORESIS [ 13]

Dialysis, H,39.91electrodialysis,92 ultrafiltration, H salting out,93 gel chro-

matography, 94 ion-exchange chromatography, 38 hydrophobic interaction
chromatography, 95 and two-phase extraction with n-pentano196 have been
suggested for the removal of carrier ampholytes. A recently described
technique is based on electrophoresis of carrier ampholytes through a
dialysis membrane into a filter paper sheet soaked with buffer. 7 The pro-
teins are retained by the membrane and can be recovered with nearly
100% yield. The technique is simple, flexible with respect to gel volume
and processing of multiple samples. The technique can be easily used with
most commercially available equipment for flatbed isoelectric focusing
with buffer vessels of sufficient capacity.

Technique 7,8
Preparation o f Gel Suspension. Sephadex G-200 (Superfine) or BioGel
P-60 (minus 400 mesh) is suspended in distilled water, and the swollen gel
is washed on a sintered-glass funnel with 20 volumes of deionized water.
To avoid mechanical damage of the gel beads, water is added without
stirring the gel. After washing, the thick gel suspension is dehydrated with
several changes of methanol and dried in a vacuum oven at 35°. From
the dry, washed gel suspensions for coating are prepared according to
Table II.
Preparation o f Rehydratable Gels. Mylar D polyester films (100/zm)
are treated with 6 N NaOH for 15 min to render the surface hydro-
philic. 8,4~ The films are washed with tap water, followed by deionized
water, and are dried at 80°. Gel-Bond films for agarose (Marine Colloids)
and Gel-Fix films (Serva) may be also used. The required amount of a
deaerated (water pump) gel suspension is spread with the aid of a glass
rod over the plastic film. The gels are dried with the aid of a hot fan for 1
hr, or without a fan at room temperature overnight. Gels without carrier
ampholytes should be supplemented with 1-2% glycerol. The dry gel can
be preserved without noticeable changes for at least several months when
stored at room temperature. Prior to use, the gels containing carrier am-
pholytes are sprayed in a zigzag course with a water spray. Gels without
ampholytes are sprayed with a 2% (w/v) solution of carrier ampholytes. A
small excess of the solvent will produce a shiny surface, After drying for a
9~j. F. Gierthy, K. A. O. Ellem, and J. R. Kongsvik,Anal. Biochem. 98, 27 (1979).
92T. G. Bloomsterand D. W. Watson,Anal. Biochem. 113, 79 (1981).
93p. Nilsson, T. Wadstr6m, and O. Vesterberg, Biochim. Biophys. Acta 221, 146 (1970).
94O. Vesterberg, Sci. Tools 16, 24 (1969).
95W. J. Gelsema, C. L. De Ligny,and W. M. Blanken,J. Chrornatogr. 196, 51 (1980).
96H. P. K6st and E. K6st-Reyes, in "Electrophoresis '79" (B. J. Radola, ed.), p. 565.
de Gruyter, Berlin, 1980.
[13] PREPARATIVE ISOELECTRIC FOCUSING 273

TABLE II
COMPOSITION OF THE GEL SUSPENSION FOR PREPARATION OF GEL LAYERS OF
GRANULATED GELS a

Amount of dry gel

40% Solution
Thickness BioGel Sephadex Distilled of carrier
of layer P-60 G-200 Wet gel water b ampholytes
(/~m) (mg) (mg) (ml) (ml) (/zl)

300 207 120 3 5.0 150

500 345 200 5 7.5 250
1000 690 400 10 12.5 500

° The amounts are for 10 × 10 cm gels.

b Different amounts of water are recommended to facilitate spreading of the
gel suspension on the plastic support or glass plate.

few minutes, the gel is ready for use. Alternatively, wet gel layers may be
used by coating the films or glass plates and drying the layer with a fan
until irregular I- to 3-mm fissures appear at the edges of the gel. 6° Gel layers
dried to a consistency defined by this easily recognizable criterion do not
flow when inclined to an angle of ->45°.
Electrode Strips and Solutions. Strips (1 cm in width) of MN 866 paper
(Macherey & Nagel) are soaked with the following solutions: 25 mM
aspartic acid and 25 mM glutamic acid (anolyte) and 2 M ethylenediamine
containing 25 mM arginine and 25 mM lysine (catholyte). Care should be
taken to establish good contact of the strips to the gel layer. Adjustable
platinum ribbon or wire electrodes are placed on the electrode strips and
weighed to ensure good contact.
Sample Application. The sample is applied as a streak on the gel
surface with a microscope slide or a rectangular glass plate of a width 2
cm less than that of the gel layer (as compensation for possible edge
effects). Approximately 30/~1 of solution per centimeter of applicator is
easily deposited by this technique. Greater sample volumes can be ap-
plied in intervals of 0.5-1 cm by this means or by using commercially
available sample applicators (Bio-Rad, LKB, and Pharmacia). The sam-
ple should be applied over a 20-25% distance of the total separation
length from the electrodes, preferably in a part of the gel layer in which
components of interest are not expected after focusing. Although proteins
from dilute samples can be concentrated by preparative isoelectric focus-
ing, the more practical way is to concentrate the dilute samples prior to
sample applications, e.g., by ultrafiltration, and to apply 2-10% protein
solutions. All samples should be desalted to -<0.05 M salt content.
274 EL~CTROPHORESIS [13]

Isoelectric focusing. A cooling solution at 4° is circulated through the

apparatus from a constant-temperature circulating bath. Typical running
conditions for different separation distances are (a) 10 cm: prefocusing at
50 V/cm, 20-30 min, final field strength 300 V/cm, 2000 to 3000 V-hr; (b)
20 cm: prefocusing at 20-50 V/cm, final field strength 300 V/cm, 10,000 to
13,000 V-hr; (c) 40 cm: prefocusing at 20-40 V/cm, 16 hr, final field
strength 100-200 V/cm, 30,000 to 40,000 V-hr. With adequate cooling,
0.05-0. I W/cm 2 is tolerated.
Printing. Of the several cellulose acetate membranes tested, Sartorius
membranes (catalog No. 11106) have optimal printing properties. Strips, 1-
2 cm wide and of appropriate length, are rolled from one end onto the gel
layer and gently pressed to ensure uniform wetting of the membrane. Care
should be taken not to entrap air between the membrane and the gel. After
contact for 1 min, which suffices for uniform wetting, the membrane is
removed and placed for 1-2 min in 10% trichloroacetic acid (w/v).
Adhering gel particles are removed by washing. The strips are stained
for 1 min in 1% Ponceau S in 10% trichloroacetic acid and destained for a
few seconds in 5% acetic acid. Prints are dried in air, yielding a white
background. Staining intensity decreases on drying, which is of advantage
in work at high protein loads.
Removal of Carrier Ampholytes. Carrier ampholytes are removed
from the gel electrophoretically with a dialysis membrane. A sheet of filter
paper, Whatman No. 3 or Macherey & Nagel MN 827, is wetted with 0.05
M Tris-HC1 at pH 8.5 and blotted with dry filter paper to remove excess
liquid. The wet paper is mounted on a 1-mm thin glass plate on the cooling
plate of the focusing chamber and connected with pads of MN 866 paper
with the electrode vessels containing the same buffer as the paper. Visk-
ing dialysis membrane is placed on the buffered paper and pressed to
establish good contact. A rectangular 2-mm silicone rubber strip with an
appropriate slit in the middle, about 0.5 to 1 cm × 5 cm, is laid on the
dialysis membrane. The gel, liquefied with a small amount of water, is
transferred into the slit, and electrophoresis is conducted at 800-1000 V
for 40-60 min. The dialysis membrane retains the protein while the carrier
ampholytes migrate into the paper. The efficiency of the removal is
checked by drying the paper at 120° for 5-10 min and staining with 1%
Amido Black 10B in methanol-acetic acid-water (45:5:25, v/v/v) fol-
lowed by destaining in the same solvent. With Amido Black 10B approxi-
mately 0.5-1/zg of carrier ampholytes are detected per square centimeter.
Thus, failure to stain the paper beneath the slit indicates a reduction of
carrier ampholytes to <0.01%.
Elution of Proteins. The gel with the focused proteins, directly from
the gel layer or after electrophoretic removal of carrier ampholytes, is
[14] AFFINITY ELECTROPHORESIS 275

transferred into a centrifuge tube and suspended in a 1 : 1.5 ratio in dis-

tilled water. After centrifugation for 10 min at 35,000 g at 2°, the superna-
tant fluid is collected and the sediment suspended in a fresh amount of
water. In the two to three combined supernatant fractions, protein is
determined against a blank of carder ampholytes of the same pH range as
that used for the focusing procedure. Alternatively, proteins are eluted in
a minicolumn assembly consisting of cotton-plugged Eppendorf pipette
tips supported in a test tube. A 1.5 : 2 ratio of eluent to gel is sufficient for
protein elution.
Determination ofpH Gradient. The pH gradient is determined directly
in the layer in a part of the gel containing the focused proteins. Measure-
ment of pH at the edge of the layer or in protein-free parts of the gel may
introduce great errors owing to a different distribution of carder ampho-
lytes.

[14] A f f i n i t y E l e c t r o p h o r e s i s
By V,~CLAV HOi~FAgf

The principle of affinity electrophoresis is simple: a macromolecule

migrates electrophoretically in a gel medium containing effectively immo-
bilized ligands ("affinity gel") capable of interaction with the migrating
macromolecule. As a result of this interaction, the macromolecule is more
or less retarded in the affinity gel as compared to a control gel. Such
control gels include media devoid of the immobilized ligand or media
containing an immobilized mock ligand incapable of complex formation
with the migrating macromolecule; the mobility I of other macromolecules
contained in the sample is not affected in affinity gels. Thus, affinity
electrophoresis is essentially analogous to affinity chromatography, since
both methods are based on separation of macromolecules due to their
affinity toward a ligand immobilized on a solid-phase carrier. Since re-
views on affinity electrophoresis are available, 2-4 this chapter concen-
t As used here, mobility denotes the distance migrated by the protein during the duration o f
the experiment under a specific set of conditions; the term is not intended to imply the
physicochemical quantity of electrophoretic mobility as estimated by free-flow electro-
phoresis.
2 V. Hoi~ejgf, Anal. Biochem. 112, 1 (1981).
3 V. Ho~ejgf and J. Kocourek, this series, Vol. 34, p. 178.
4 T. C. BOg-Hansen, in "Proceedings o f the Third International Symposium on Affinity
Chromatography and Molecular Interactions" (J. M. Egly, ed.), p. 399. INSERM Sympo-
sium Series, Pads, 1979.

Copyright © 1984by AcademicPress, Inc.

METHODS IN ENZYMOLOGY,VOL. 104 All rightsof reproductionin any form reserved.
ISBN 0-12-182004-1
[14] AFFINITY ELECTROPHORESIS 275

transferred into a centrifuge tube and suspended in a 1 : 1.5 ratio in dis-

tilled water. After centrifugation for 10 min at 35,000 g at 2°, the superna-
tant fluid is collected and the sediment suspended in a fresh amount of
water. In the two to three combined supernatant fractions, protein is
determined against a blank of carder ampholytes of the same pH range as
that used for the focusing procedure. Alternatively, proteins are eluted in
a minicolumn assembly consisting of cotton-plugged Eppendorf pipette
tips supported in a test tube. A 1.5 : 2 ratio of eluent to gel is sufficient for
protein elution.
Determination ofpH Gradient. The pH gradient is determined directly
in the layer in a part of the gel containing the focused proteins. Measure-
ment of pH at the edge of the layer or in protein-free parts of the gel may
introduce great errors owing to a different distribution of carder ampho-
lytes.

[14] A f f i n i t y E l e c t r o p h o r e s i s
By V,~CLAV HOi~FAgf

The principle of affinity electrophoresis is simple: a macromolecule

migrates electrophoretically in a gel medium containing effectively immo-
bilized ligands ("affinity gel") capable of interaction with the migrating
macromolecule. As a result of this interaction, the macromolecule is more
or less retarded in the affinity gel as compared to a control gel. Such
control gels include media devoid of the immobilized ligand or media
containing an immobilized mock ligand incapable of complex formation
with the migrating macromolecule; the mobility I of other macromolecules
contained in the sample is not affected in affinity gels. Thus, affinity
electrophoresis is essentially analogous to affinity chromatography, since
both methods are based on separation of macromolecules due to their
affinity toward a ligand immobilized on a solid-phase carrier. Since re-
views on affinity electrophoresis are available, 2-4 this chapter concen-
t As used here, mobility denotes the distance migrated by the protein during the duration o f
the experiment under a specific set of conditions; the term is not intended to imply the
physicochemical quantity of electrophoretic mobility as estimated by free-flow electro-
phoresis.
2 V. Hoi~ejgf, Anal. Biochem. 112, 1 (1981).
3 V. Ho~ejgf and J. Kocourek, this series, Vol. 34, p. 178.
4 T. C. BOg-Hansen, in "Proceedings o f the Third International Symposium on Affinity
Chromatography and Molecular Interactions" (J. M. Egly, ed.), p. 399. INSERM Sympo-
sium Series, Pads, 1979.

Copyright © 1984by AcademicPress, Inc.

METHODS IN ENZYMOLOGY,VOL. 104 All rightsof reproductionin any form reserved.
ISBN 0-12-182004-1
276 ELECTROPHORESIS [ 14]

trates mainly on one of the current experimental modifications of the

method, i.e., affinity-chromatography-like technique.

Experimental Modifications of Affinity Electrophoresis

Several variants of affinity electrophoresis exist that differ in the man-
ner of immobilization of the ligand in the gel, in the medium, and in the
buffer system used.

Immunoelectrophoresis-Like Modification
This is sometimes called crossed immuno-affino-electrophoresis4-6 and
is closely related to some immunoelectrophoretic techniques, namely,
those in which antibody is incorporated into the agarose gel and interacts
with the migrating antigen. Usually, lectins are incorporated into the gel
instead of antibodies although other interacting proteins would serve. The
conditions (pH, electroendoosmosis) are chosen so that the lectin has a
very low mobility. Such an affinity gel causes retardation of those glyco-
proteins that form complexes with the incorporated lectin; control gel is
either devoid of the lectin or contains a carbohydrate inhibitor of the
lectin. The position of the glycoprotein zones after electrophoresis is most
conveniently detected by crossed immunoelectrophoresis, i.e., electro-
phoresis of the separated proteins in a second perpendicular dimension
into the agarose slab gel with incorporated, precipitating polyvalent anti-
serum against the sample. The lectin concentration may be chosen so that
lectin-glycoprotein precipitates are formed during electrophoresis in a
manner analogous to antigen-antibody precipitates. 7 Instead of merely
incorporating the lectin in agarose gel, a slurry of ligand-substituted beads
in melted agarose can be used for preparation of affinity gels. 8

Affinity Electrophoresis in Polyacrylamide Gel

In contrast to the above-mentioned technique, applicable to the study
of interactions between two macromolecules, e.g., lectin and glycopro-
tein, polyacrylamide gel is usually used as a carrier of various ligands,
e.g., enzyme substrates and inhibitors, or haptens interacting with lectins
or antibodies. Such affinity gels are employed in the studies of the respec-
5 T. C. B0g-Hansen, Anal. Biochem. 56, 480 (1973).
6 T. C. BCg-Hansen, in "Electrophoresis, A Survey of Techniques and Applications," Part
B (Z. Deyl, ed.), p. 219. Elsevier, Amsterdam, 1983.
v T. C. BCg-Hansen, O. J. Bjerrum, and C. H. Brogren, Anal. Biochem. 81, 78 (1977).
s M. Raftell, lmmunochemistry 14, 787 (1977).
[14] AFFINITY ELECTROPHORESIS 277

tive protein-ligand interactions. Immobilization of the ligand in the poly-

acrylamide gel matrix can be achieved by several means.
1. A polymerizable derivative of the ligand can be prepared, e.g., an
allyl- or acryloyl derivative, and copolymerized with the monomers nor-
mally used during preparation of polyacrylamide gel (acrylamide and bis-
acrylamide). This procedure has been discussed in detail in this series. 3
2. A soluble macromolecular derivative of the ligand can be prepared,
which is then added to the polymerization mixture normally used for
preparation of polyacrylamide gels. After completion of polymerization,
the macromolecular carrier, if sufficiently large, remains entrapped within
the polyacrylamide gel network and is effectively immobilized. Several
types of macromolecular carriers have been used. Polysaccharides can be
used directly without modification when proteins bind to them. 9-~2 Alter-
natively, a polysaccharide or synthetic polymer can be substituted with
the ligand, 13-~5 or a soluble copolymer can be prepared by copolymeriza-
tion of acrylamide with a suitable derivative such as allyl or acryloyl. 16-19
3. Instead of soluble macromolecular derivatives of the ligand, beads
substituted with the ligand can be entrapped in the polyacrylamide gel
matrix. A simple procedure ensures homogeneously tight packing of the
beads within the gel and eliminates potential problems due to irregular
sedimentation of the beads before polymerization occurs. 2°

Affinity E l e c t r o p h o r e s i s in A g a r o s e or A g a r o s e - P o l y a c r y l a m i d e Gels
with Covalently B o u n d L i g a n d 2°
For the purposes of affinity chromatography, the ligands are usually
coupled to suitable beaded gels, agarose gel beads generally being the
material of choice. However, most of the reactions used for "activation"
of agarose result in partially cross-linked gels that do not melt upon heat-
ing. If periodate oxidation is used for activation with subsequent reduc-

9 H. Stegemann, Hoppe-Seyler's Z. Physiol. Chem. 348, 951 (1967).

t0 K. Takeo and S. Nakamura, Arch. Biochem. Biophys. 153, 1 (t972).
~l K. Takeo and E. A. Kabat, J. Immunol. 121, 2305 (1978).
12C. Borrebaeck and M. E. Etzler, FEBS Lett. 117, 237 (1980).
13K. t~ei~ovsk~,M. Tichfi, V. Hoi~ej~i,and J. Kocourek, J. Biochem. Biophys. Methods 3,
163 (1980).
14V. ~efovsk~, M. Tichli, J. Turltovfi,and J. Labsk~,,J. Chromatogr. 194, 175 (1980).
t5 M. Tich~i,V. Hoi~ej~f,and J. Barthovfi,Biochim. Biophys. Acta 534, 58 (1978).
16V. Hoi~ej~f,P. Smolek, and J. Kocourek, Biochim. Biophys. Acta 538, 293 (1978).
17K. Nakamura, A. Kuwahara, H. Ogata, and K. Takeo, J. Chromatogr. 192, 351 (1980).
is p. Masson, A. Privat de Garilhe, and P. Burnat, Biochim. Biophys. Acta 701, 269 (1982).
19 Jang-Lin Chen and H. Morawetz, J. Biol. Chem. 256, 9221 (1981).
2oV. Hoi'ej~t,M. Tichfi, P. Tich~, and A. Hol~, Anal. Biochem. 125, 358 (1982).
278 ELECTROPHORESIS [14]

tive amination by cyanoborohydride for binding of an amino derivative of

the ligand to the "activated" aldehydic derivative of agarose beads, non-
cross-linked, normally melting ligand-substituted beads are obtained that
can be used conveniently in affinity electrophoresis. They can be simply
melted on a water bath, mixed in a suitable ratio with unsubstituted
agarose, and, after cooling, they result in homogeneous affinity gels. Al-
ternatively, these agarose gel derivatives can be used for preparing mixed
agarose-polyacrylamide gels. The last method of ligand immobilization
and affinity gel preparation, described in greater detail below, seems to be
easy and generally applicable.

Affinity Isoelectric Focusing

Immobilization of a ligand in a gel medium containing carrier ampho-
lytes yields affinity gels applicable to affinity isoelectric focusing, z~ In this
method, a combination of gel isoelectric focusing and affinity interaction,
the ligand-binding protein present in a complex mixture is again "cap-
tured" in the gel near the start. The noninteracting proteins are sharply
resolved and focus normally at the positions corresponding to their iso-
electric points. The ligand may be immobilized by some of the methods
mentioned above, although care must be taken to avoid gel media with
increased electroendoosmosis, the latter being incompatible with the iso-
electric focusing procedure. 2°

Applicability and Limitations of Affinity Electrophoresis

The phenomenon of retardation of a specific ligand-binding protein
under the conditions of affinity electrophoresis can be exploited either
qualitatively or quantitatively. 22

Qualitative Applications
Affinity electrophoresis, or affinity isoelectric focusing, can be used
for detection and identification of a ligand-binding protein in a complex
protein mixture; for detection of the presence of inactive admixtures in
purified preparations of the ligand-binding proteins, and for estimation of
their ligand-binding heterogeneity; and for checking the results of chemi-
cal modification reactions affecting, presumably, the ligand-binding site.
The method is also useful in testing materials to be used for affinity
chromatography. In all these applications, the patterns observed on con-
2~v. Hofej~fand M. Tich~,Anal. Biochem. 116, 22 (1981).
22For a review with referencesto more particularexamples, see Hoi~ej~f.2
[14] AFFINITY ELECTROPHORESIS 279

trol and affinity gels are compared and the zones retarded on the affinity
gels are readily identified.

Quantitative Applications
The degree of retardation of a ligand-binding protein in an affinity gel
depends primarily on the concentration of the immobilized ligand and on
the value of the dissociation constant of the protein-ligand complex.
Thus, the dependence of mobility on the concentration of the immobilized
ligand can be used for estimation of apparent dissociation constant of the
protein-immobilized ligand complexes, as shown first by Gerbrandy and
Doorgeest 23 and Takeo and Nakamura) ° This factor may also be useful
for estimation of dissociation constants of protein-free (mobile) ligand
complexes, if the free ligand, in addition to the immobilized one, is
present in the affinity gel. 24 The basic theoretical background necessary
for extraction of quantitative information from affinity electrophoresis
experiments performed under various conditions has been developed. 25-27
These quantitative applications of affinity electrophoresis fall beyond the
scope of this chapter, and the relevant references can be found else-
where .2

Preparation of Some Macromolecular Carriers of the Ligands

Water-Soluble Copolymers of Acrylamide with an Unsaturated
Derivative of the Ligand 15-18
Acrylamide (0.5 g; an example of a copolymerizable ligand deriva-
tive), 16 allylglycoside (1 g), and ammonium persulfate (10 mg) are dis-
solved in I0 ml of water and heated for 5 rain on a boiling water bath. The
viscous product is diluted, extensively dialyzed against water, and lyophi-
lized. The copolymer contains approximately one carbohydrate residue
per six acrylamide units. The ligand content can be regulated by the
amount of the ligand monomer used in the reaction. The molecular weight
of the copolymer may be varied by altering the concentration of the
polymerization catalyst (persulfate) and probably also by temperature and
duration of the polymerization reaction. Studies on the kinetics of acryl-
amide polymerization may be relevant for optimization of the preparation
of copolymers. 2s
z3 S. J. Gerbrandy and A. Doorgeest, Phytochemistry 11, 2403 (1972).
V. Hoi~ej~I, M. Tich~, and J. Kocourek, Biochim. Biophys. Acta 499, 290 (1977).
25 V. Hoi'ej~f, J. Chromatogr. 178, 1 (1979).
26 V. Hot~ej~f and M. Tich~l, J. Chromatogr. 216, 43 (1981).
27 V. Matou~ek and V. Ho~ej~l, J. Chromatogr. 245, 271 (1982).
C. Gelfi and P. G. Righetti, Electrophoresis 2, 213 (1981).
280 ELECTROPHORESIS [14]

Ligand-Substituted Dextransl3
A solution of 0.1 M NalO4 and 1% dextran T-500 (Pharmacia) is main-
tained at 25° for 1 hr and then exhaustively dialyzed against water. A 1%
solution of this oxidized dextran, 20 ml, is mixed with 20 ml of a 10-50
mM solution of the ligand bearing a primary amino group in 0.2 M carbon-
ate-bicarbonate at pH 9.2 and stirred at room temperature for 2 hr. Then
10 mg NaBH4 is added in small portions during 10 min; after another 10
min, 0.1 ml of acetone is added. The reaction mixture is dialyzed exhaus-
tively against distilled water and lyophilized. The ligand content of the
product is approximately 10% (w/w). Commercially available Blue Dex-
tran, i.e., Cibracron Blue F3G-substituted dextran (Pharmacia), may be
used for affinity electrophoresis of proteins interacting with this dye? 5

Ligand-Substituted Meltable Derivatives of Agarose G el 2°,29

Washed Sepharose 4B (5 ml) is added to 5 ml of 0.5 M NaIO4. The

suspension is agitated for 2 hr at room temperature, after which the oxi-
dized gel is thoroughly washed with water and stored at 4°. The gel (2 ml)
is washed with 0.5 M potassium or sodium phosphate at pH 6 and added
to 2 ml of the same buffer containing 20 mM ligand bearing a primary
amino group and 20 mM NaCNBH3. The mixture is agitated for 24 hr at
room temperature. After washing, the gel is stored in 0.1 M Tris-HC1 at
pH 8.2 containing 0.1% NAN3. Typically, the concentration of immobi-
lized ligand in the gel is 1-2 mM. 3°
Triazine dyes may be covalently bound to a nonactivated agarose gel
to yield meltable derivatives applicable to affinity electrophoresis. 31

Procedure for Affinity Electrophoresis

Affinity electrophoresis is performed exactly as is the corresponding
"parent" electrophoretic method--i.e., a tube or slab gel arrangement
can be used. Glass tubes 2-2.5 mm in diameter are convenient, espe-
cially owing to the low consumption of the immobilized ligand deriva-
tives. Some ligands, or soluble high-molecular-weight ligand derivatives,
may inhibit acrylamide polymerization. Should this occur, it becomes
necessary to increase the concentration of the polymerization catalyst.

29 I. Parikh, S. March, and P. Cuatrecasas, this series, Vol. 34, p. 77.

3o If the ligand possesses suitable spectral properties, its concentration in the substituted
beads can be estimated most conveniently by spectrophotometry of solubilized gel solu-
tions or directly by spectrophotometry of the agarose gel bead suspension in 50°% glycerol
solution, using the free ligand solution as a standard.
31 S. J. Johnson, E. C. Metcalf, and P. D. G. Dean, Anal. Biochem. 109, 63 (1980).
[15] PREPARATIVEISOTACHOPHORESIS 281

When meltable agarose derivatives are used for immobilization of the

ligand in polyacrylamide affinity gels, polymerization is performed at
higher temperatures (60-70 °) to prevent premature gelling of agarose2°;
under such conditions, 5- to 10-fold lower concentrations of polymeriza-
tion catalysts must be used. Generally, the catalyst concentration that is
optimal for polymerization of polyacrylamide-based affinity gels must be
found empirically.
The affinity and control gels to be compared should be run under
identical conditions, i.e., simultaneously in the same electrophoretic
tank. The properties of the control gel should be as similar as possible to
those of the affinity gel. The control gel should contain an immobilized
mock ligand very similar to the affinity ligand. The mobility of the inter-
acting protein zone should be related to the mobility of a noninteracting
reference protein that is present in all the samples.

[15] P r e p a r a t i v e I s o t a c h o p h o r e s i s
By CHRISTOPHER J. HOLLOWAY and RODIGER V. BATTERSBY

Isotachophoresis is the least well known of modern electrophoretic

techniques. Although the theoretical groundwork for the method was
published in 1897, isotachophoresis has been applied practically to any
significant extent only since the early 1960s. It is the exception rather than
the rule to find an article devoted to this technique on a preparative scale,
particularly with regard to protein separations, since isotachophoresis is
most commonly associated with the analytical capillary equipment, I in
which small ionic species, such as nucleotides, 2 amino acids, or small
peptides, 3 rather than macromolecules, present the majority of applica-
tions. Nevertheless, there are very good reasons for use of isotachophore-
sis in protein separation and purification. An attempt has been made here
to present the possible advantages and useful areas of application insofar
as this is necessary for satisfactory exploitation of the method.
That isotachophoresis is only one of several suggested names for this
technique causes some difficulty in literature searches. Other synony-
mous terms are ionic migration method, displacement electrophoresis,

i F. M. E v e r a e r t s , J. L. Beckers, T. P. E. M. Verheggen, J. Chromatogr. Libr. 6 (1976).

2 C. J. Holloway a n d J. L[istorff, Electrophoresis 1, 129 (1980).
3 C. J. Holloway a n d V. Pingoud, Electrophoresis 2, 127 (1981).

Copyright © 1984by AcademicPress, Inc.

METHODS IN ENZYMOLOGY,VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
[15] PREPARATIVEISOTACHOPHORESIS 281

When meltable agarose derivatives are used for immobilization of the

ligand in polyacrylamide affinity gels, polymerization is performed at
higher temperatures (60-70 °) to prevent premature gelling of agarose2°;
under such conditions, 5- to 10-fold lower concentrations of polymeriza-
tion catalysts must be used. Generally, the catalyst concentration that is
optimal for polymerization of polyacrylamide-based affinity gels must be
found empirically.
The affinity and control gels to be compared should be run under
identical conditions, i.e., simultaneously in the same electrophoretic
tank. The properties of the control gel should be as similar as possible to
those of the affinity gel. The control gel should contain an immobilized
mock ligand very similar to the affinity ligand. The mobility of the inter-
acting protein zone should be related to the mobility of a noninteracting
reference protein that is present in all the samples.

[15] P r e p a r a t i v e I s o t a c h o p h o r e s i s
By CHRISTOPHER J. HOLLOWAY and RODIGER V. BATTERSBY

Isotachophoresis is the least well known of modern electrophoretic

techniques. Although the theoretical groundwork for the method was
published in 1897, isotachophoresis has been applied practically to any
significant extent only since the early 1960s. It is the exception rather than
the rule to find an article devoted to this technique on a preparative scale,
particularly with regard to protein separations, since isotachophoresis is
most commonly associated with the analytical capillary equipment, I in
which small ionic species, such as nucleotides, 2 amino acids, or small
peptides, 3 rather than macromolecules, present the majority of applica-
tions. Nevertheless, there are very good reasons for use of isotachophore-
sis in protein separation and purification. An attempt has been made here
to present the possible advantages and useful areas of application insofar
as this is necessary for satisfactory exploitation of the method.
That isotachophoresis is only one of several suggested names for this
technique causes some difficulty in literature searches. Other synony-
mous terms are ionic migration method, displacement electrophoresis,

i F. M. E v e r a e r t s , J. L. Beckers, T. P. E. M. Verheggen, J. Chromatogr. Libr. 6 (1976).

2 C. J. Holloway a n d J. L[istorff, Electrophoresis 1, 129 (1980).
3 C. J. Holloway a n d V. Pingoud, Electrophoresis 2, 127 (1981).

Copyright © 1984by AcademicPress, Inc.

METHODS IN ENZYMOLOGY,VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
282 ELECTROPHORESIS [15]
cons electrophoresis, ionophoresis, omegaphoresis, and steady-state
stacking. From the last, it will become apparent that the concentrating
step of the well-known Ornstein 4 and Davis 5 disc electrophoresis, i.e.,
steady-state stacking, is actually an isotachophoretic configuration.
Two reviews on preparative-scale isotachophoresis that contain prac-
tical details have been published, but these represent mainly the approach
of the individual g r o u p s . 6'7 We present here the methodology for the
better known preparative procedure in a specially designed apparatus and
we also describe a method on a flat gel slab that can easily be performed
with standard equipment of most laboratories involved in protein chem-
istry.

Basic Principles of Isotachophoretic Separations

We briefly outline here the essentials of the principles of isota-
chophoresis. A more comprehensive theoretical treatment ~and extensive
descriptions of the principles have been presented. 8,9

Definition of the Isotachophoretic System

In contrast to conventional zonal electrophoresis, in which sample
ions migrate through a background electrolyte of uniform composition,
isotachophoresis uses a discontinuous electrolyte system. The sample
ions are preceded in the migration stack by a leading ion (L) of the same
charge quality, but with higher effective mobility than all the sample ions
of interest. The displacing electrolyte terminating the stack (T) is a spe-
cies of lower effective mobility than all the sample ions in the stack. The
system is buffered at the required pH by a common counterion. Owing to
this arrangement, the samples are forced to migrate in the steady state at
the same velocity; hence the name isotachophoresis.

Some Properties of the Isotachophoretic System

Isotachophoresis is normally performed under conditions of constant
current. In the fundamental equation governing all electrophoretic separa-
tions [Eq. (1)],
4 L. Ornstein, Ann. N . Y . Acad. Sci. 121, 321 (1964).
5 B. J. Davis, Ann. N . Y . Acad. Sci. 121, 404 (1964).
6 N. Y. Nguyen and A. Chrambach, in "Gel Electrophoresis of Proteins" (B. D. Haines
and D. Rickwood, eds.), p. 145. IRL Press, London, 1981.
7 p. j. Svendsen, in "Electrophoresis, A Survey of Techniques and Applications" (Z. Deyl,
ed.), p. 345. Elsevier, Amsterdam, 1979.
8 C. J. Holloway and I. Trautschold, Z. Anal. Chem. 311, 81 (1982).
9 S. G. Hjalmarsson and A. Baldesten, CRC Crit. Rev. Anal. Chem. 261 (1981).
[15] PREPARATIVEISOTACHOPHORESIS 283

v = thE (1)
v is the migration velocity, th the effective mobility, and E the electric
field strength. Since v is constant, the electric field strength is inversely
proportional to the effective mobility. Thus, in the stack, the field strength
increases stepwise from L to T at each zone boundary.
Since electrophoretic systems are subject to Joule heating effects, in
isotachophoresis the temperature pertaining in the zones will increase
stepwise with decreasing mobility, i.e., with increasing field strength. For
this reason, a vertical apparatus is always arranged such that the hotter
terminating electrolyte is uppermost. In this way, convectional distur-
bances are minimized. An additional property of the isotachophoretic
system can be deduced from Ohm's law. Under conditions of constant
current, with the stepwise increase in field strength, the conductivity will
decrease at each zone boundary from high to low mobility. These proper-
ties are exploited in analytical capillary equipment as detectors, but have
not yet been incorporated in preparative apparatus.

The Steady State

Conventional zone electrophoresis never reaches a steady state. In
isoelectric focusing, the steady state is reached when proteins are focused
in a narrow band at a position in the pH gradient equal to their isoelectric
points. In isotachophoresis, the steady state is reached when no partially
mixed zones are present in the stack. If two sample ions have differing
effective mobilities in the mixed state under the conditions of the separa-
tion, then they are completely separable. If they have identical effective
mobilities in the mixed state, then they are inseparable. "Partial" separa-
tion means that the steady state has not been reached. A lower sample
load or greater separation volume is then required.
An important consideration in connection with the steady state is the
regulation of zone concentrations in isotachophoresis. The so-called "be-
harrliche Function" laid down by Kohlrausch in 1897 defines the condi-
tions at a boundary between two electrolyte solutions with a common
counterion migrating in an electric field. As described by Eq. (2),
CL mL (reX + mR)
- + (2)
CX (mL + mR) mx
c represents the concentration of the two ions L and X, m represents their
effective mobilities, and R refers to the common counterion, This form of
the equation is a simplification, not taking into account the possible differ-
ing electric charges of sample and leading ions.
284 ELECTROPHORESIS [ 15]

Having set the concentration of leading ion, the concentration of the

following sample zone in the steady state is regulated to a set concentra-
tion, which is a function of the several mobilities of the components. In
turn, therefore, the concentration of each component of the stack is regu-
lated. In the steady state, the stack migrates in zones with unchanging
concentrations and the stack itself has a constant length. This property
has been used to determine whether the steady state has been reached. 6
Practically, the steady state, i.e., the best possible separation, must be
reached before sample zones are eluted from the apparatus. Each system
has a defined load capacity, which, if exceeded, will not allow the steady
state to be reached. The load capacity increases with increasing separa-
tion volume. In theory, it does not vary with the applied separation cur-
rent, since the total coulombs applied up to elution will remain the same.
However, increasing the leading electrolyte concentration is a way of
increasing the load capacity, since more coulombs are required for com-
plete migration through the apparatus. However, increasing the concen-
tration of leading electrolyte also leads to higher concentrations in the
sample zones (from the Kohlrausch function), and solubility problems
may be encountered with proteins. The load capacity also depends on the
actual differences in mobilities of the sample ions. A large number of
species with only slightly differing effective mobilities will require a much
greater separation volume than do a smaller number of separands of
widely differing mobilities.

The Spacer Principle

In zone electrophoresis, the sample (protein) zones migrate at differ-
ent velocities and move apart during the separation process. In isota-
chophoresis, this is not the case. If the sample stack contains only pro-
tein, a protein continuum between leading and terminating electrolytes is
formed. The zone boundaries cannot be distinguished by staining proce-
dures, and elution of the pure zones becomes impossible. Thus, the
"spacer technique" is often applied. This involves adding nonprotein
species to the sample with effective mobilities between those of the se-
parands of interest. Amino acids are an example of such spacers. How-
ever, finding suitable spacers in preparative systems is a laborious matter
of trial and error, although some catalogs of spacers have been pub-
lished, l0 In principle, carder ampholytes can be applied to form a more or
less continuous spacer mobility gradient. Since these mixtures contain a
large number of compounds of closely related mobilities, difficulties in
load capacity and separation quality can result.
l0 S. Husmann-Holloway and E. Borriss, Z. Anal. Chem. 311, 465 (1982).
[15] PREPARATIVEISOTACHOPHORESIS 285

The spacer principle can be employed much more usefully for selec-
tive stacking or unstacking. If a separand of primary interest has high
mobility and migrates at the front of the stack, the terminating electrolyte
can be replaced by another of high mobility, ideally one slightly lower
than that of the separand. This separand will migrate in the stack as
previously, although many of the other components of the sample of
lower mobility will be excluded from the stack; the separand in question is
said to be selectively stacked. The separand can be selectively unstacked
if it has lower mobility than the other components of the sample mixture.

The Zone "Sharpening" Effect

The main reason for the low resolving power of zone electrophoresis is
that no property of the system can counteract diffusion. In isoelectric
focusing, diffusion is countered by the continuous migration into a point
of zero charge. In isotachophoresis, diffusion is also continuously coun-
tered, providing high resolution; the zone "sharpening" effect is respon-
sible. Suppose that a sample ion, X, diffuses into the following zone, Y,
where it will find itself in an environment of higher electric field strength
than its own zone in the stack. According to Eq. 1, the fundamental
migration equation, X will be accelerated over and above the "isota-
chophoretic" migration velocity and will overtake the X/Y zone boundary
before being forced back into its own zone.

Procedures in a Column Apparatus with Polyacrylamide Gel

By far the greatest number of investigations involving preparative
isotachophoresis report the use of commercially available instruments for
isotachophoresis in columns. Such an apparatus is shown schematically
in Fig. 1. The column is filled with the stabilizing medium, e.g., polyacryl-
amide. The prerequisite here is low electroendoosmosis. In principle, the
newer types of agarose with electroendoosmotic values approaching zero
could be employed. Some investigators have used granulated dextran-
type gels, or density gradients with, for example, sucrose. However,
polyacrylamide is the most common gel, whereby rather low concentra-
tions may be selected to minimize molecular sieving.
The gel is prepared in the leading electrolyte system, and the electrode
region above the gel is filled with terminating electrolyte. In any vertical
apparatus, as explained previously, it is important that the terminator be
placed above the leading system, i.e., migration is in a downward direc-
tion, since the temperature of the terminating zone is highest. With the
terminator above the leading system, therefore, convectional effects are
286 ELECTROPHORESIS [ 15]

q- q-
T

$
L $
L_

A A

I 11"
FIG. 1. Schematic representation of preparative isotachophoresis in a column apparatus.
Column I shows the initial situation, in which the sample mixture (shaded region) has been
introduced into the upper part of the gel cylinder. The upper electrode compartment (in this
case, the cathode) is filled with terminating electrolyte (T); the lower gel region is filled with
leading electrolyte (L). The lower electrode compartment is filled with anolyte (A), which
also serves as elution buffer (E). In column II, the sample species have separated into the
stack (S) and migrate down the gel cylinder. The upper gel region is displaced by T. When
the separands migrate out of the gel (column III), they are eluted. For simplicity, the cooling
mantle of the column has not been shown, although this is an essential component of the
equipment.

l o w e r than t h e y w o u l d be with the r e v e r s e configuration and an u p w a r d

migration. T h e s a m p l e is applied to the u p p e r gel surface before adding
terminating electrolyte. It is p r e f e r a b l e to dissolve the s a m p l e or to dia-
lyze it against the t e r m i n a t i n g electrolyte.
Since p r o t e i n s e p a r a t i o n s are u n d e r discussion, s o m e sort o f s p a c e r
m i x t u r e will h a v e to be a d d e d to the sample. A l t h o u g h c a r d e r a m p h o l y t e
m i x t u r e s are m o s t f r e q u e n t l y e m p l o y e d , p r o b l e m s o f load c a p a c i t y c a n
arise if e x c e s s i v e a m o u n t s are applied. As far as the c o l u m n a p p a r a t u s is
c o n c e r n e d , a m p h o l y t e s m a y lead to swelling of p o l y a c r y l a m i d e gels. Such
effects h a v e led to a m o v e m e n t a w a y f r o m glass c o l u m n s , since poly-
[15] PREPARATIVEISOTACHOPHORESIS 287

acrylamide adheres strongly to glass. Swelling of the gel will give rise to
distortion, whereby the expansion takes place through the center of the
gel cylinder. The upper surface of the gel becomes convex, and the zones
curve. Plastic columns have proved to be more satisfactory, since little if
any adherence of the gel is observed; polyacrylamide is free to expand
and contract uniformly.
When current is applied, the sample ions migrate down the column
and begin to separate. Eventually, they emerge from the gel into an elu-
tion chamber of small volume, through which elution buffer is passed at a
constant flow rate. The elution tube can be passed through a detector to a
fraction collector.

Procedures
Electrolyte Systems. One of the major difficulties associated with iso-
tachophoresis is the selection of suitable leading and terminating electro-
lytes. Frequently, this is a matter of trial and error, although guidelines
are given in Table I. All these systems are designed for anionic separa-
tions and may be unsuitable for highly basic proteins. For this latter group
a cationic system, such as that presented in a later section on flatbed
procedures, can be used.
In principle, the leading electrolyte is simply chosen from the list in
Table I, according to the required pH. A system at lower pH will be
suitable for selectively collecting acidic proteins; a higher pH yields a
broader spectrum of separands. For separation of immunoglobulins, for
example, a system at a higher pH is essential. The isoelectric points of the
proteins of interest are a useful parameter in selecting the electrolyte. The
terminating electrolytes are chosen for selectively stacking or unstacking
sample proteins. In order to obtain stacking of as many separands as
possible, the EACA system, which has the lowest mobility in the list, is
chosen. Clearly, the higher mobility glycine terminator is suitable for
selective stacking of the more acidic proteins.
Gel Preparation. The following solutions are required for the poly-
acrylamide gel: acrylamide, 33% w/v in distilled water; Bis, 1% (w/v) in
distilled water (these solutions are stable for up to 10 days at 4°); riboflavin
5'-phosphate, 0.008% (w/v) in distilled water; ammonium persulfate 0.1%
(w/v) in distilled water (these solutions must be made up immediately be-
fore preparing the gel). Note that photopolymerization is preferable.
The following recipe for the gel is given for a final total volume of 100
ml. Using these proportions, the total quantity can be varied according to
the dimensions of the column apparatus.
288 ELECTROPHORESIS [ 15]

TABLE I
ELECTROLYTE SYSTEMS FOR ANIONIC SEPARATIONS IN
COLUMN ISOTACHOPHORESISa

Approximate pH Leading ion Counterion

A. Leading electrolytes b
4 HAc, 30 ml Tris, 10 g
6 MES, 73 g Tris, 15 g
7 1 M H3PO4, 300 ml Tris, 60 g
8 1 M H3PO4, 300 ml Tris, 120 g
8.5 1 M HCI, 600 ml Tris, 220 g
9 1 M HCI, 600 ml Ammediol, 120 g
B. Terminating electrolytes C
8.5 Glycine, 15 g Tris, 3.0 g
8.7 Alanine, 17.5 g Tris, 2.5 g
8.9 EACA, 30 g Tris, 1.5 g

a Abbreviations: HAc, glacial acetic acid; MES, 2-(N-

morpholino)ethanesulfonic acid; Tris, tris(hydroxymethyl)-
aminomethane; ammediol, 2-amino-2-methylpropane-l,3-
diol; EACA, 6-aminohexanoic acid.
b The quantities given are for a final stock volume of 1000 ml.
Leading and counter species are mixed and diluted to 1000
ml by the addition of distilled water. The exact pH is mea-
sured at the running temperature (10~) after diluting a small
portion 10-fold, i.e., to the final concentration in the gel.
c The quantities given are for a final volume of 1000 ml. The
terminating electrolytes thus obtained are used as prepared,
i.e., undiluted. The exact pH should be measured at the
operating temperature (10°).

Leading electrolyte, stock solution, 10 ml

Acrylamide solution, 10 ml
Bis solution, 10 ml
Riboflavin 5'-phosphate solution, 10 ml
Ammonium persulfate solution, 10 ml
Distilled water, 50 ml
The mixture is poured carefully into the column apparatus without the
introduction of air bubbles. This is best achieved by pouring down the
inside wall of the column. The lower outlet of the column is suitably
blocked until polymerization is complete. It is important to ensure that
the upper surface of the gel is fiat. The gel mixture is, therefore, overlayed
with about 5 ml of leading electrolyte (diluted 10-fold to yield the same
concentration as in the gel mixture). Photopolymerization is carded out
with the aid of suitable lamps. Although the reaction should be complete
[15] PREPARATIVE ISOTACHOPHORESIS 289

within 1 or 2 hr, storage of the gel at 10° overnight is recommended. In

order to prevent undue swelling and contracting, the gel should be cast
and stored at all times at the final running temperature, e.g., 10°.
Application of Sample. Shortly before use, the ove/'laying leading
electrolyte is removed from above the gel surface, and the sample is
carefully applied and allowed to diffuse into the gel matrix. The free space
above the gel is then filled with terminating electrolyte.
Separation Procedure. Since migration velocity is dependent on the
applied constant current, it is preferable to use as high a driving current as
possible, limited only by the cooling capabilities of the apparatus. In
practice, the absolute limit is generally 10 mA. Usually, 5-8 mA are
applied. Depending on the dimensions of the apparatus and the type of
electrolyte system selected, separation requires from 4 to 24 hr. Immedi-
ately before commencing the separation, the lower electrolyte region of
the equipment must be filled (air-bubble flee) with anolyte. This solution
is also used as elution buffer and can be varied somewhat to suit the
stability of the separands. However, a solution of 30 ml of 1 M H2SO4 with
8 g of Tris per liter (giving a pH of about 7) has been satisfactory in many
cases.
The elution flow rate must be set according to the rate of appearance
of the sample species from the gel. This will depend on the driving current
applied and the electrolyte system employed. In general, flow rates of the
order of 5-30 ml/hr are suitable. It is often useful to mix a small quantity
of dyestuff with the sample. This makes visualization of the position of
the stack possible. Again, the dyestuff to be used depends on the electro-
lyte system. However, it is of prime importance that the dye migrate
within the stack, not by zone electrophoresis in the leading or terminating
electrolytes. During the separation process, the dye should become con-
centrated and finally retain a fixed zone length during migration. If the
dyestuff zone becomes diffuse during the procedure, it is probably not
within the stack. In systems of low pH, amaranth red is useful. Dyes with
lower anionic mobility are fluorescein, bromophenol blue, and bromocre-
sol green. For cationic systems, Janus green has been useful.

Procedures in a Horizontal Slab of Granulated Gel

The method described in the preceding section requires a specially
designed column. Although the schematic representation in the figure
appears simple, the actual commercial equipment available requires con-
siderable care and patience for effective use. Since this equipment may
not be generally available, we suggest an alternative procedure that is less
290 ELECTROPHORESIS [15]
time-consuming and requires only minor additions to flatbed electropho-
resis equipment already present in many laboratories.

Apparatus
The basic requirement is a flatbed electrophoresis chamber with an
effective cooling plate of dimensions at least 25 x l0 cm. The electric
field is applied across the longer side with electrode tanks accommodating
at least 500 ml of anolyte and catholyte, respectively. The cooling plate
should be attached to circulating temperature-controlled water supplies,
with the option, via Y valves, of 50° for use in the casting procedure and of
10° for cooling during the separation. The power supply should deliver up
to 2000 V. A maximum current delivery of 20 mA is adequate.
A gel tray is required for casting the slab. This consists of a glass plate,
1 mm thick, cut to the required dimensions, but no larger than the cooling
plate. A silicone rubber gasket of cross section 5 x 5 mm is attached
around the rim of this glass plate.
Since the slab is exposed to the atmosphere during the procedure, it
may be useful to flush the chamber with carbon dioxide-free air, espe-
cially where high pH conditions are employed. The absorbance of carbon
dioxide into the system produces a long carbonate zone that can nega-
tively influence the quality of the separation through wastage of load
capacity.

Procedures
For these procedures, with a gel tray size of approximately 25 × 10
cm, the final gel slab thickness does not exceed 3 mm. For other dimen-
sions, slab thickness will have to be modified accordingly.
Gel Casting. The types of material suitable for casting granulated gels
in isotachophoresis should fulfill the same requirements as for isoelectric
focusing (see this volume [13]). In general, Sephadex G-75 SF or G-200
SF (Pharmacia) or Ultrodex (equivalent to G-75 SF, LKB) have been
found satisfactory. The gel (4 g) is suspended as a slurry in 100 ml of
diluted leading electrolyte. Table II lists some stock electrolytes, which
are diluted prior to mixing with the gel material. Of immediate importance
is the final leading electrolyte concentration in the gel. However, the gel is
initially applied as a slurry, and excess water is evaporated. Thus, the
actual dilution of the stock leading electrolyte will depend on the evapora-
tion limit of a particular batch of gel. This information should be quoted
by the manufacturers.
[15] PREPARATIVEISOTACHOPHORESIS 291

TABLE II
ELECTROLYTE SYSTEMS FOR ISOTACHOPHORETIC
SEPARATIONS IN GRANULATED SLAB GELSa

Approximate pH Leading ion Counterion

A. Leading electrolytes b
4 (anionic) HAc, 30 ml Tris, 10 g
6 (anionic) l M H3PO4, 300 ml Tris, 40 g
8.5 (anionic) l M HC1, 600 ml Tris, 140 g
5 (cationic) 1 M KOH, 600 ml HAc, 70 ml
B. Terminating electrolytes c
8.7 (anionic) fl-Alanine, 18 g Tris, 2.5 g
8.9 (anionic) EACA, 30 g Tris, 1.5 g
4.5 (cationic) Alanine, 18 g HAc, l l ml

a Abbreviations as in Table I.
b The quantities given are for a final stock volume of
1000 ml. Leading and counter species are mixed and
diluted to 1000 ml by the addition of distilled water.
The amount of stock electrolyte required for the gel
slurry is calculated as explained in the text. The ex-
act pH should be measured in the final dilution in gel
slurry at the running temperature (10°). Other leading
electrolyte systems can be adapted from Table I.
c The quantities given are for a final volume of 1000
ml. The terminating electrolytes are used undiluted.
The pH of the terminating electrolytes can be varied
by altering the concentration of counterion.

Equation (3) can be used to determine the amount of stock solution, in

milliliters, that should be diluted to a final volume of 100 ml with water
for addition to 4 g of gel material.
100 - [1.04 × evap limit (%)]
Volume of stock solution = 10 (3)

Filter paper strips 1 cm wide are cut to fit into the gel tray at each end,
i.e., strips 10 cm long. These strips are soaked in the diluted leading
electrolyte and are stacked to a height of 4 mm. At this stage, the tray with
paper strips is weighed. The tray is then placed on the cooling block of the
apparatus, ensuring that the latter is exactly level, thus providing an even
thickness of gel slab.
The vessel containing the gel slurry is weighed before and after pour-
ing into the tray. The difference gives the initial total weight of slurry
292 ELECTROPHORESIS [15]

applied. In order to assure relatively rapid evaporation of excess water

from the slurry, the block is heated to 50° by circulating water while a
gentle current of air from a fan is passed over the gel surface. At regular
intervals, the gel tray is removed and weighed. The evaporation step is
complete when the slurry has been reduced in weight by the evaporation
limit, e.g., 36%.
The circulating water supply at 10° is now allowed to pass through the
cooling block. The filter strips at the terminating end of the tray are
carefully removed and replaced by similar strips soaked in terminating
electrolyte. The ends of the gel slab are connected to the electrode vessels
by several layers of filter paper cut exactly to the width of the slab. The
leading end of the slab is connected by wicks soaked in leading electrolyte
at a concentration fourfold that of the final concentration in the gel. The
terminating end is connected by wicks soaked in terminating electrolyte.
It is advisable to cover the latter wicks with a layer of plastic film to
prevent drying, since the terminating region, in particular, can become
quite hot during the procedure.
Prerun. Since the sample should be applied in terminating electrolyte,
but the gel initially contains only leading electrolyte, a prerun must be
performed in order to allow terminating electrolyte to migrate into the gel,
a distance of approximately 5 cm. For this purpose, a current of up to 20
mA can be applied. The boundary between leading and terminating elec-
trolytes can be identified by colored (brown) impurities from the strips
that concentrate at this point. However, a drop of bromophenol blue may
be added to visualize the boundary.
Sample Application. A section of the gel is removed on the terminat-
ing side of the zone boundary. This section should be about 2 cm deep
across the whole width of the gel. The gel is scraped out with a spatula
and is resuspended in a maximum of 3 ml of sample solution (including the
volume of any spacers). This suspension is poured back into the cavity in
the gel and, owing to its consistency, should yield an even gel surface. For
convenience, sample application devices are available commercially.
Running Conditions. The isotachophoretic separation is carried out
under conditions of limiting current. Values in the range 4 to 6 mA are
suitable. Since the migration velocity is constant at constant current, this
latter value can be conveniently set to provide complete migration
through the slab overnight. The position of the trailing edge of the leading
electrolyte zone can be observed by the trace of dye added at the begin-
ning of the prerun.
Paper Prints. Before cutting the gel into sections for collecting the
separands, it is useful to obtain a paper print to establish the position of
[15] PREPARATIVEISOTACHOPHORESIS 293

the zones and estimate the quality of the separation. Since it is important
that straight zone boundaries be obtained, it is particularly useful to print
the whole surface of the gel. In this volume [13] printing procedures are
described that should be satisfactory also for isotachophoresis. Here, we
restrict ourselves to two simple and rapid methods of paper printing,
which have proved to be satisfactory.
1. Serva Blue W is a water-soluble dye that eliminates the need for
organic solvents. The paper print is fixed in a 10% trichloroacetic acid-
5% sulfosalicylic acid solution and stained in an aqueous solution (0.5 g
per liter of the dye). Destaining is performed with slightly acidified water.
2. Bromophenol blue provides a yellow stain in acid solution. The dye
is dissolved in a mixture (9: 1, v/v) of ethanol and glacial acetic acid (1%,
w/v). Destaining is performed with water.
Fractionation of the Gel. The simplest method for fractionation is the
use of a grid of at least 30 parallel blades oriented across the width of the
gel slab. Such grids are available commercially. However, for isota-
chophoresis it is useful to have grid widths of rather less than 5 mm (2 mm
is ideal), since the fractions are concentrated only in a part of the slab.
The gel is first fractionated by using the grid. The pH of each section can
then be measured with a surface contact pH electrode. The leading and
terminating zones have different pH values, and the protein stack can be
detected as that portion of the gel with steadily changing pH values.
After pH measurement, the gel is scraped out of the grid fraction with
a spatula and is transferred to small columns fitted with nylon sieves. The
fractions are equilibrated with one gel volume of a suitable elution buffer.
After the liquid has drained through the sieve, a second volume of buffer
is added to the gel. In this way, most of the protein is eluted from the
granulated gel. The fractions can then be assayed for total protein and for
the desired specific properties, such as enzymic activity.

Alternative Equipment for Preparative Isotachophoresis

In this chapter we have concentrated on the column apparatus as the
most widely employed method for preparative isotachophoresis. As can
be seen in Table III, compared with Tables IV and V, below, approxi-
mately 80% of the applications were performed with this type of equip-
ment. For normal research applications, however, the flatbed technique is
more convenient. Although precise details of all the methods presently
available cannot be presented here, it is useful to provide a short sum-
mary of possible alternatives.
294 ELECTROPHORESIS [ 15]

TABLE III
PREPARATIVE-SCALE PROTEIN SEPARATION IN POLYACRYLAMIDE GELS

Separands Source Apparatus Electrolytes a Reference b

Transferrin Human serum Glass tube L1, T1 1

Serum proteins Human serum Glass tube L2, T2 2
Histocompatibility Mouse lymphocytes Glass tube L3, T3 3
antigens
Fibrinogen breakdown Fibrinogen Column L2, T1 4
product
Membrane proteins Human erythrocytes Column L4, T4 5
Carcinoma antigen Human lung tumors Column L2, T4 6
Enterotoxin Escherichia coli Column L2, T4 7
HGPT Human erythrocytes Column L2, T2 8
CHE Human serum Column L4, T4 9
Membrane proteins Human erythrocytes Column NG c 10
Serum proteins Human serum Column L4, T4 11
Enterotoxin E. coli Column L2, T4 12
Serum proteins Human serum Column L2, T4 13
Growth hormone Human Glass tube Various 14
Enterutoxin Clostridiurn Glass tube L1, T1 15
perfringens
Antibodies Rabbit serum Glass tube NG 16
Growth hormone Human Glass tube Various 17
Crystallin Bovine lens Glass tube L5, T5 18
Immunoglobulins Myeloma lines Flatbed L2, T4 19
arMacroglobulin Rat serum Column L4, T4 20
Skin test antigens Blastomycin Column L6, T1 21
Enterotoxin E. coli Column L4, T4 22
Serum proteins Human serum Column L5, T5 23
Prolactin Canine pituitaries Column NG 24
Histoplasmin Human Histoplasma Column L6, T1 25
capsulatum
Inorganic pyrophos- Bakers' yeast Column L4, T4 26
phatase
Enterotoxin E. coli Column NG 27
Antibodies Rabbit serum Glass tube NG 28
Serum proteins Human serum Flatbed L2, T4 29
Tumor-specific surface Mice tumors Glass tube NG 30
antigen
Erythropoetin Human urine Column L 1, T6 31
NAD-dependent Methylotrophic bac- Column L2, T5 32
formate dehydro- teria
genase
Migration inhibition Human lymphocytes Column L2, T4 33
factor
Hibernating triggers Woodchuck plasma Column L6, T6 34
Albumin and hemo- Bovine blood Flatbed L4, T4 35
globin
[15] PREPARATIVE ISOTACHOPHORESIS 295

TABLE III (continued)

Separands Source Apparatus Electrolytes ~ Reference b

CSF proteins Human CSF Glass tube L7, T7 36

CSF proteins Human CSF Flatbed Various 37
LMW kininogen Human plasma Column L8, T8 38
High-density lipopro- Human plasma Column L9, T4 39
teins

a A key to the electrolyte systems is given in Table VI.

b Key to references:
1. D. B. Ramsden and L. Lewis, Protides Biol. Fluids 19, 521 (1972).
2. N. Catsimpoolas and J. Kenney, Biochim. Biophys. Acta 285, 287 (1972).
3. M. Hess and D. A. L. Davies, Eur. J. Biochem 41, 1 (1974).
4. I. Clemmensen and P. J. Svensen, Sci. Tools 20, 5 (1973).
5. T. C. BCg-Hansen, O. J. Bjerrum, and P. J. Svensen, Sci. Tools 21, 33 (1974).
6. M. J. Frost, G. T. Rogers, and K. D. Bagshawe, Br. J. Cancer 31, 379 (1975).
7. R. M611by, S. G. Hjalmarsson, and T. WadstrOm, FEBS Lett. 56, 30 (1975).
8. B. Bakay and W. L. Nyhan, Arch. Biochem.Biophys. 168, 26 (1975).
9. C.-H. Brogren, P. J. Svendsen, and T. C. BCg-Hansen, in "Progress in Isoelectric
Focusing and Isotachophoresis" (P. G. Righetti, ed.), p. 359. North-Holland Publ.,
Amsterdam, 1975.
10. T. C. BCg-Hansen, P. J. Svendsen, O. J. Bjerrum, C. S. Nielsen, and J. Ramlau,
Protides Biol. Fluids 22, 679 (1975).
11. T. C. BCg-Hansen, P. J. Svendsen, and O. J. Bjerrum, "Progress in Isoelectric
Focusing and Isotachophoresis" (P. G. Righetti, ed.), p. 347. North-Holland Publ.,
Amsterdam, 1975.
12. F. Dorner, J. Biol. Chem. 250, 8712 (1975).
13. S.-G. Hjalmarsson, Sci. Tools 22, 35 (1975).
14. A. Chrambach and J. S. Skyler, Protides o f Biol. Fluids 22, 701 (1975).
15. W. W. Yotis and N. Catsimpoolas, J. Appl. Bacteriol. 39, 147 (1975).
16. H. Brogren and G. Peltre, Ann. lmmunol. 126, 363 (1975).
17. G. Baumann and A. Chrambach, Proc. Natl. Acad. Sci. U.S.A. 73, 732 (1976).
18. F. S. M. van Kleef, M. Peeters, and H. J. Hoenders, Anal. Biochem. 77, 122 (1977).
19. A. Ziegler and G. K6hler, FEBS Lett. 71, 142 (1976).
20. F. Gauthier, N. Gutman, J. P. Muh, and H. Mouray, Anal. Biochem. 71, 181 (1976).
21. M. V. Lancaster and R. F. Sprouse, Infect. Immun. 13, 758 (1976).
22. F. Dorner, H. Jaschke, and W. St6ckl, J. Infect. Dis. 133, 142 (1976).
23. A. Kopwillem, W. G. Merriman, R. M. Cuddeback, A. J. K. Smolka, and M. Bier, J.
Chromatogr. 118, 35 (1976).
24. P. J. Knight, M. Gronow, and J. M. Hamilton, J. Endocrinol. 69, 127 (1976).
25. M. V. Lancaster and R. F. Sprouse, Anal. Biochem. 77, 158 (1977).
26. V. N. Kasho and S. M. Avaeva, Int. J. Biochem. 9, 51 (1978).
27. T. Wadstr6m, R. MOllby, B. Olsson, J. S6derholm, and C. J. Smyth, "Electrofocus-
ing and Isotachophoresis" (B. J. Radola and D. Graesslin, eds.), p. 443. de Gruyter,
Berlin, 1977.
28. C.-H. Brogren and G. Peltre, Scand. J. Immunol. 6, 685 (1977).
29. C.-H. Brogren, "Electrofocusing and Isotachophoresis" (B. J. Radola and D.
Graesslin, eds.), p. 549. de Gruyter, Berlin, 1977.
(continued)
296 ELECTROPHORESIS [ 15]

References to TABLE Ill (continued)

30. T. Natori, L. W. Law, and E. Appella, Cancer Res. 38, 359 (1978).
31. M. Puschman, W. Thorn, and Y. Yen, Res. Exp. Med. 173, 293 (1978).
32. Y. V. Rodionov, T. V. Avilova, E. V. Zakharova, L. S. Platonenkova, A. M.
Egorov, and I. V. Berezin, Biokhimiya 42, 1896 (1977).
33. L. H. Block, H. Jaksche, S. Bamberger, and G. Ruhenstroth-Bauer, J. Exp. Med.
147, 541 (1978).
34. P. R. Oeltgen, L. C. Bergmann, W. A. Spurrier, and S. B. Jones, Prep. Biochem. 8,
171 (1978).
35. F. Hampson and A. J. P. Martin, J. Chromatogr. 174, 61 (1979).
36. K. G. Kjellin and L. Hallander, J. Neurol. 221, 225 (1979).
37. L. Hallander and K. G. Kjellin, Anal. Chem. Syrup. Ser. 5, 245 (1980).
38. A. Adam, J. Damas, C. Schots, G. Heynen, and P. Franchimont, Anal. Chem.
Syrup. Ser. 5, 47 (1980).
39. G. B. Bon, G. Cazzolato, and P. Avogaro, J. Lipid Res. 22, 998 (1981).
c NG, Not given.

Small-Scale Preparative Isotachophoresis in Glass Tubes

This type of separation has been described in detail by C h r a m b a c h and
his group. 6 Isotachophoresis is p e r f o r m e d in glass tubes of 5- to 6-mm i.d.
containing p o l y a c r y l a m i d e gel. Separations can be carried out on m o s t
disc electrophoresis equipment, provided that cooling is adequate. The
gels are r e m o v e d as cylinders f r o m the tubes after completion of the
migration and sliced into fractions. One disadvantage of this method,
apart f r o m the previously mentioned d r a w b a c k s of glass, is the limited
load capacity. Handling of the gel is rather awkward, as it is with disc
electrophoresis in this type of equipment. Possibly, the newer type of
vertical slab c h a m b e r s could be useful.

Hollow Cylinder Technique

This variant, devised by H a m p s o n and Martin, j] is a logical develop-

ment of the glass tube method. The configuration of a vertical tube is
basically maintained, but with the following modifications: the diameter
of the tube is increased to 20 m m , and the gel is cast as a 1- to 2-mm-thick
cylinder around the outside wall. Cooling liquid is circulated through the
center of the tube. To prevent evaporation, the equipment is i m m e r s e d in
a cooled reservoir of nonaqueous, water-immiscible liquid, e.g., o-di-
chlorobenzene. The terminating electrolyte is floated at the top of the cylin-

11F. Hampson, in "Eiectrophoresis '79" (B. J. Radola, ed.), p. 583. de Gruyter, Berlin,
1980.
[15] PREPARATIVE ISOTACHOPHORESIS 297

der. The cooling capability of this apparatus is excellent. The possible

drawbacks of the technique, as with much self-built equipment, is compli-
cated handling. The toxicity of the cooling liquid may also be a hindering
factor.

Free-Flow Procedures
In the main, two groups have been working on this type of equipment
for several years. 12,13The apparatus is based on free-flow cell electropho-
resis, but with a modified chamber arrangement. The separation chamber
consists of a fiat rectangular cavity (0.5 mm thick) in a cooling block,
through which the electrolytes and sample are continuously pumped. An
electric field is applied across the cavity, at right angles to the direction of
laminar flow. This equipment has a potentially high load capacity, owing
to the variability of both current and flow rate. Furthermore, the absence
of stabilizing medium eliminates molecular sieving. The major advantage
of the technique, however, is the continuous nature of the separation; all
other procedures are batch processes. For industrial purposes, therefore,
free flow offers the greatest possibilities, and it is expected that commer-
cial apparatus for free-flow is0tachophoresis will become available
shortly.

Cooling
It will have become apparent that efficient cooling is essential for
suitable separation conditions and gel stability. Moreover, when separat-
ing heat-labile proteins, cooling is an obvious need. It is reasonable to
analyze the cooling systems of the two major types of equipment, column
and flatbed, to ascertain whether a major advantage pertains to one or the
other system. Although it is difficult to define the precise temperature
gradients and heat-exchange characteristics for these systems, the cooling
area as a function of gel volume can be compared, realizing that gel
volume is the decisive factor in load capacity and separability.
The first assumption here must be that the heat exchange per unit area
cooling in the column and flatbed are similar. In order to attain compara-
ble electrical conditions and load capacity, the length of gel slab is defined
identical to the length of gel column; i.e., the distance between electrodes
is similar, and the total gel volume is the same.

12 Z. Prusik, J. Stepanek, and V. Kasicka, in "Electrophoresis '79" (B. J. Radola, ed.),

p. 287. de Gruyter, Berlin, 1980.
13 H. Wagner and V. Mang, in "Analytical Isotachophoresis" (F. M. Everaerts, ed.), p. 41.
Elsevier, Amsterdam, 1980.
298 ELECTROPHORESIS [15]

TABLE IV
PREPARATIVE-SCALEPROTEIN SEPARATIONIN AGAROSEGELS

Separands Source Apparatus Electrolytes a Reference b

Urinary proteins Human urine NG ¢ LI, T1 1

Albumin and Human urine NG L1, T1 2
transferrin and CSF
Sweat proteins Human sweat Flatbed L1, T1 3
Sweat and Human sweat Flatbed LI, T1 4
urine proteins and urine
Serum proteins Bovine blood NG NG 5
Albumin and Bovine blood Hollow L4, T9 6
hemoglobin cylinder
Various proteins Diverse Glass tube Various 7
sources

a A key to the electrolyte systems is given in Table VI.

b Key to references:
1, V. Blaton, K. Uyttendaele, and H. Peeters, Acta Med. Acad. Sci. Hung. 31, 277
(1974).
2. K. Uyttendaele, M. de Groote, V. Blaton, H. Peeters, and F. Alexander, Protides
Biol. Fluids 22, 743 (1978).
3. K. Uyttendaele, V. Blaton, F. Alexander, H. Peeters, M. de Groote, N. Vinaimont-
Vandecasteele, and J. Chevalier, "Progress in Isoelectric Focusing and Iso-
tachophoresis" (P. G. Righetti, ed.), p. 341. North-Holland Publ., Amsterdam, 1975.
4. K. Uyttendaele, M. de Groote, V. Blaton, H. Peeters, and F. Alexander, J. Chroma-
togr. 132, 261 (1977).
5. A. J. P. Martin and F. Hampson, Br. UK Patent Appl. 2026546 (1980).
6. F. Hampson, in "Electrophoresis '79" (B. J. Radola, ed.), p. 287. de Gruyter, Berlin,
1980.
7. Z. Buz,'ls and A. Chrambach, Electrophoresis 3, 121 (1982).
c NG, Not given.

The flatbed apparatus with base dimensions of 25 × 10 cm, pro-

vides a cooling area of 250 cm z. At a gel thickness of d cm, gel volume
would be 250 x d cm 3. In the column configuration, the gel volume is also
250 x d cm 3, and the length of the column is 25 cm; column diameter is
then (40 × d/cr) 1/2. Given that the cooling area of the slab should be at least
as great as that of the column, i.e., cooling area of slab >- cooling area of
column, then, 250 cm 2 ~ 25~rV'(40d/cr) l/z and, therefore, d -< 0.8 cm.
Thus, where the gel slab is less than 8 mm thick, the cooling area per
unit volume gel is greater than that of the column apparatus. Normally, d
is of the order 3-4 mm, so that we can assume realistic cooling efficiency
compared with the column apparatus. The apparatus devised by Martin
and Hampson H offers even better cooling, since both sides of a "cylindri-
cal slab" are exposed to cooling fluid.
[15] PREPARATIVEISOTACHOPHORESIS 299

TABLE V
PREPARATIVE-SCALE PROTEIN SEPARATION IN GRANULATED GELS

Separands Source Apparatus Electrolytesa Referenceb

Antibodies Rabbit serum Flatbed L2 or L6 1

T5 or T4
Antibodies Rabbit serum Flatbed L2, T4 2
Plasma protein Human blood Column L5, T10 3
Plasma protein Human blood Column L5, T10 4
Serum proteins Human serum Flatbed L2 or L6 5
T5 or T4
Glutathione S-transferases Cat liver Flatbed L10, T4 6

a A key to the electrolyte systems is given in Table VI.

b Key to references:
1. G. Peltre and C.-H. Brogren, "Electrofocusing and Isotachophoresis" (B. J. Radola
and D. Graesslin, eds.), p. 577. de Gruyter, Berlin, 1977.
2. C.-H. Brogren and G. Peltre, "Electrofocusing and Isotachophoresis" (B. J. Radola
and D. Graesslin, eds.), p. 587. de Gruyter, Berlin, 1977.
3. M. Bier and A. Kopwillem, "Electrofocusing and Isotachophoresis" (B. J. Radola
and D. Graesslin, eds.), p. 567. de Gruyter, Berlin, 1977.
4. M. Bier, R. M. Cuddeback, and A. Kopwillem, J. Chromatogr. 132, 437 (1977).
5. A. Winter, H. Brogren, and T. Dobson, LKB Appl. Note 318 (1980).
6. R. V. Battersby and C. J. Holloway, Electrophoresis 3, 275 (1982).

Applications of Preparative-Scale Isotachophoresis for the Separation

and Purification of Proteins
Although the discussion has been limited to two types of configuration
o f apparatus, many variants of the electrolyte system and modifications
o f apparatus and running conditions have been described. To provide an
overview for general reference, a summary of a range of publications in
this field for the three major support media (polyacrylamide, agarose, and
granulated gels) is presented in Tables III, IV, and V, respectively. The
various electrolyte systems (where reported) are summarized in Table VI.

Predictable Areas of Application of Preparative-Scale Isotachophoresis

of Proteins
C o m p a r e d with preparative isoelectric focusing, and even more so
when c o m p a r e d with the chromatographic methods, there is too little
published work on isotachophoresis to conclude any specific applications
for which the latter technique is ideally suited. Indeed, most of the re-
ported applications do not exploit the special properties of the iso-
tachophoretic principle. Thus, for the guidance of those interested in this
 300 ELECTROPHORESIS [15]

TABLE VI
ELECTROLYTESYSTEMSEMPLOYEDIN THE VARIOUSLITERATURE
REFERENCESIN TABLESIII, IV, AND V

Electrolyte Approximate
system Ion Counterion pH

Leading
L1 Phosphate Tris 5.5
L2 Phosphate Tris 6.1
L3 Chloride Ammediol 8.4
L4 Phosphate Tris 8.1
L5 Cacodylate Tris 7.0
L6 Acetate Tris 4.4
L7 TES Tris 7.5
L8 Chloride Histidine 6.2
L9 Sulfate Tris 7.1
L10 Chloride Tris 7.0
Terminating
T1 Glycine Tris 8.6
T2 fl-Alanine Tris 8.0
T3 Phenolate Ammediol 7.8
T4 EACA Tris 8.9
T5 fl-Alanine Tris 8.9
T6 Glycine Tris 7.2
T7 EACA Ba(OH)2 9.5
T8 Diethylbarbiturate Histidine 7.2
T9 2-Aminopropanoic acid Tris Not given
T10 /3-Alanine Ba(OH)2 9.2

technique, it is worthwhile to discuss briefly the particular areas, which

we believe would be m o r e realistic.
It is not possible to provide definite figures for the capacity of the
s y s t e m described. At best, an e x t r e m e l y complex protein mixture of close
mobilities would be limited to less than 100 mg. F o r the isolation of a
specific protein, 1 g or m o r e has been applied. It is, therefore, a matter of
the actual sample, i.e., load capacity, which must be determined experi-
mentally.
A c o m p a r i s o n with preparative isoelectric focusing is inevitable, since
only these two m e m b e r s o f the electrophoresis family are worth consider-
ing f r o m the point of view of resolution capability for preparative-scale
work. In both methods, the sample is concentrated and diffusional effects
are actively countered. The slightly higher cost of isoelectric focusing, due
to the larger amounts of carrier ampholyte, is a decisive argument in f a v o r of
isotachophoresis. Certainly, in the flatbed configuration, the fractionation
[15] PREPARATIVEISOTACHOPHORESIS 301

of the gel is easier for the more widely separated focused zones than for
the protein stack in isotachophoresis. However, for those interested in
separating and purifying enzymes, isotachophoresis does have one intrin-
sic advantage over isoelectric focusing: we have frequently observed that
enzymes lose considerable activity when exposed for extended periods to
an environment at their isoelectric points. In isotachophoresis, it is possi-
ble to vary the electrolyte systems widely and thereby provide the condi-
tions under which an enzyme is most stable. Both methods are useful in
that the samples are concentrated rather than diluted during the separa-
tion process, a factor of importance when dealing with labile proteins.
[16] DETERGENT SOLUBILIZATION OF PROTEINS 305

[16] S o l u b i l i z a t i o n o f F u n c t i o n a l M e m b r a n e Proteins
By L E O N A R D M . H J E L M E L A N D a n d ANDREAS CHRAMBACH

Although the solubilization of functional membrane proteins is a cen-

tral task in modern membrane biochemistry, systematic experimentation
directed at achieving this goal has not been successful. Even less success
has attended attempts at optimizing procedures tailored to specific sys-
tems. There are, however, several approaches to both problems, and
these can be illustrated with specific examples.
The material outlined here is intended to present a systematic ap-
proach to the solubilization of membrane proteins. Specific details for the
initial investigations, directed at achieving solubilization, are outlined.
This is not the place for a discussion of the properties of detergents or
their many specialized uses in membrane biochemistry; several reviews
treat these topics in depth. 1-7 We believe that the initial trials presented
here for solubilizing functional membrane proteins are simple and that,
although guarantees cannot be given, the outlined approach will be suc-
cessful in many cases.

Choice of Criteria for Solubility

Before attempting solubilization, it is necessary to consider criteria
that can be used for distinguishing between soluble and insoluble mem-
brane proteins after treatment with detergents. Such criteria are by their
nature operational and refer to a defined set of conditions that must be
controlled throughout. The major criterion is that of retention of function
after centrifugation for 1 hr at 105,000 g. This process depends both on the
density and the temperature of the medium. Thus, a medium containing
50% glycerol (d -- 1.11 at 20°) may render some species soluble by reason
of nonsedimentability, whereas the same species may sediment in media
of lower density. Since density is also a function of temperature, this
variable should be kept constant. Another criterion for solubility involves
partitioning of species between the void volume and the included volume
l A. Helenius and K. Simons, Biochim. Biophys. Acta 415, 29 (1975).
2 C. Tanford, "The Hydrophobic Effect," 2nd Ed. Wiley, New York, 1980.
3 C. Tanford and J. A. Reynolds, Biochim. Biophys. Acta 457, 133 (1976).
4 A. Helenius, D. R. McCaslin, E. Fries, and C. Tanford, this series, Vol. 56, p. 734.
5 j. C. H. Steele, Jr., C. Tanford, and J. A. Reynolds, this series, Vol. 48, p. ll.
6 L. M. Hjelmeland and A. Chrambach, Electrophoresis 2, 1 (1981).
7 A. Tzagoloff and H. S. Penefsky, this series, Vol. 2'2, p. 219.

METHODS IN ENZYMOLOGY,VOL. 104 1SBN0-12-182004-1

306 TECHNIQUES FOR MEMBRANE PROTEINS [16]

of a gel filtration column. Again, the choice of gel filtration medium has an
obvious effect on how such proteins partition. The usual choice for such a
medium is one of the cross-linked agarose preparations, such as
Sepharose 6B (Pharmacia). Operationally, proteins that elute in the void
volume are considered to be insoluble, whereas proteins with larger elu-
tion volumes are considered to be soluble. If uncertainty exists concern-
ing the assignment of a peak to the void volume or the included volume, a
different gel filtration medium can be used for clarifying the point. In
general, it is of less importance whether one or the other criterion for
solubility is adopted than it is to decide on one of them, even though such
choice is arbitrary, and to apply the selected criterion consistently.

Assay of Soluble Activity

The other major consideration before beginning a solubilization trial
is the choice of a suitable assay for activity. The assay for solubilized
activity may be identical to that carried out with a particulate preparation,
the probable case for enzyme activity, but binding assays require addi-
tional consideration. Since many assays with particulate protein are car-
ried out by centrifugation, as for the separation of free ligand from bound
ligand, a method must be devised for separating the two when both are
soluble. The most convenient is the precipitation of membrane protein
and bound ligand with polyethylene glycol8 or ammonium sulfate. Alter-
natively, gel filtration may be used, especially when large differences in
molecular weight exist between the ligand and membrane protein. A third
protocol for binding consists of a filtration assay in which the protein-
ligand complex is bound to glass-fiber filters that are coated with poly-
ethyleneimine. 9 Since most proteins are negatively charged at physio-
logical pH, they bind to a positively charged filter; basic or neutral ligands
pass through. Filters and eluate can be assayed to determine bound and
free ligand, respectively. Details concerning the assay of solubilized pro-
teins are provided elsewhere/°

Choice of a Suitable Detergent

Although a great variety of detergents are commercially available,
many possess similar chemical structures. A review on the physical prop-
8 B. Desbuquois and G. D. Aurbach, Biochem. J. 126, 717 (1972).
9 R. F. Bruns, K. Lawson, and T. A. Pugsley, Anal. Biochem. 132, 74 (1983).
10 M. EI-Rafai, in "Receptor Biochemistry and Methodology" (J. C. Venter and L. Harri-
son, eds.), Vol. 1. Liss, New York, 1983. In press.
[16] DETERGENT SOLUBILIZATION OF PROTEINS 307

erties of detergents gives an exhaustive list of equivalent trade names

(Table II of Helenius et al.4). In particular, many trade names exist for the
popular Triton X-100 and Lubrol PX, but there is little if any evidence for
chemically or functionally distinct products. In our experience, a re-
stricted list of eight detergents (Fig. 1 and the table) should allow solubili-
zation of most functional membrane proteins. Except for digitonin all the
detergents are homogeneous, and all but deoxy-BIGCHAP H are commer-
cially available. Commercial preparations of digitonin contain only ap-
proximately 40% digitonin, the remainder being closely related steroidal
components. 12 Unfortunately, digitonin cannot be recovered from that
mixture by fractional crystallization and separates only by chromatogra-
phy on cellulose. It cannot be assumed, however, that any specific choice
from this list of eight will be successful. In general, digitonin, CHAPS,
and octyl glucoside have enjoyed popularity with difficult cases. In-
stances exist wherein the lesser known members of this list were the only
suitable choice for achieving solubilization and preserving function. Zwit-
tergent 3-14, for instance, was the detergent among those in the table most
capable of solubilizing 5'-nucleotidase, 13 whereas only deoxy-BIGCHAP
was effective in solubilizing cysteine S-conjugate N-acetyltransferase
from kidney in a form that could be fractionated by ion-exchange chroma-
tography. 14
Other factors, not directly related to solubilization, may affect the
choice of detergent. If detergent must be removed, e.g., then the deter-
gents with high values of the critical micelle concentration (CMC) and low
micelle molecular weights that can be dialyzed are suggested. 6 If absor-
bance at 280 nm is an important parameter, Triton must be excluded. If
separation techniques for exploiting predominantly or solely molecular
charge differences (charge fractionation) are to be employed, e.g., ion-
exchange chromatography or electrophoresis, charged detergents should
be avoided. If divalent cations are essential for function of the species in
question, sodium cholate with its carboxylic acid polar group should not
be used, since it forms insoluble complexes with divalent metals. Finally,
if precise physical data are to be obtained, a detergent with precisely
evaluated physical parameters, especially the partial specific volume,
must be used. The relevant physical properties of the eight detergents
listed in Fig. I are summarized in the table. Examination of the size of the
solubilized species also affects the choice of detergent, since it has been
suggested that nonionic detergents, such as Triton X-100 or Lubrol PX,
H L. M. Hjelmeland, W. A. Klee, and J. C. Osborne, Jr., Anal. Biochem. 130, 72 (1983).
12 R. Tschesche and G. Wulff, Tetrahedron 19, 621 (1963).
13 E. M. Baiyles, J. P. Luzio, and A. C. Newby, Biochem. Soc. Trans. 9, 140 (1981).
14 M. W. Duffel and W. B. Jakoby, Mol. Pharmacol. 21, 444 (1982).
308 TECHNIQUES FOR MEMBRANE PROTEINS [161

Structural Formula Chemical or Trade Name

0
II

HO Y '"0- Na + Sodium Cholate

H?'-'1-'-"OH
H

0
I+ 11
- ~Ng-0_ CHAPS

Deoxy-BiGCHAP

CH3

Digitonin

Zwittergent 3-14

Octyl Glucoside

OH

10-CH2 -CH2 1-OH Triton X-100

9-10

10 - CHz - CH2 1 - OH Lubrol PX

9-10

FIG . 1 . Structures and conventional names of detergents useful for the solubilization of
membrane proteins .
[16] DETERGENT SOLUBILIZATION OF PROTEINS 309

~. ~, ÷, ,÷

~oo

[...
a~

gh L)
eq ÷÷ I ÷1÷
O
',D

O~
.o ~h

L~

•"~ o

~:8 ~'°
310 TECHNIQUES FOR MEMBRANE PROTEINS [16]

do not disaggregate proteins nearly as well as the bile salts and their
derivatives (see below). Given those considerations, a promising start for
the initial solubilization trial is to use only octyl glucoside and CHAPS.

Choice of Initial Conditions of Buffer and Temperature

Since interaction between membrane-bound macromolecules can be
polar as well as nonpolar, the ionic strength of the solubilization medium
is a critical consideration. In the absence of a specific requirement for low
ionic strength, a high ionic strength should be used for the initial attempts
at solubilization. Concentrations between 0.1 and 0.5 M KCI are sug-
gested. The specific buffer that is used may also have an important effect.
To assure adequate buffer capacity, the concentration of buffer should be
at least 25 mM and the pH should be close to the pK. Finally, specific
considerations exist for the choice of specific buffer ions. Borate, for ex-
ample, is not suggested for use with glycoproteins, or with any system
requiring nucleotides, owing to its interactions with the cis-hydroxyl
groups of sugars. On the other hand, substitution of phosphate for KCI at
concentrations of 0.1-0.2 M is often capable of solubilizing many proteins
with detergent that cannot be brought into solution with the latter salt.~5
The origins of this peculiar effect of strong phosphate buffers are not well
understood although modifications of the structure of water, as well as
interaction with divalent cations, have been suggested in explanation.
Consideration of the experimental protocols that will follow the solubili-
zation step will influence the choice of buffers. For example, if ion-ex-
change chromatography is to follow, a buffer should be chosen that does
not bind to the ion exchanger. Basic buffers are suitable for anion-ex-
change and acidic buffers for cation-exchange chromatography. Buffers
with low ionic mobility should be chosen when electrophoresis is to fol-
low solubilization; this choice reduces net conductance and thereby re-
duces Joule heating and increases field strength.
In the absence of specific reasons to the contrary, initial trials should
be at 4°. Several "stabilizing" agents may be used in an attempt at pre-
serving the activity of proteins that are unstable under normal storage
conditions. This group of compounds will be discussed here, but it is
suggested that their use be avoided in the initial trials unless specific
reasons exist for their inclusion. Experiments designed to evaluate the
need for stabilizing additives correctly belong in the later stages, at which
time other parameters are optimized.

~5 A. C. Dey, R. Sheilagh, R. L. Rimsay, and I. R. Senciali, Anal. Biochem. 110, 373 (1981).
[16] DETERGENT SOLUBILIZATION OF PROTEINS 311

The Initial Solubilization Trial

The initial solubilization experiment serves to survey conditions that
lead to the maintenance of function of the desired protein. Particulate
protein preparations, obtained as either crude membrane fractions or
whole cells, are suspended in 50 mM buffer containing 0.15 M KC1, at a
protein concentration of 10 mg/ml and 4 °. Detergent stock solutions are
made in the same buffer at a concentration of 10% (w/v), except for
digitonin, which must be prepared at 4% (w/v) owing to limited solubility.
Detergent stock solutions, suspended protein, and buffer are mixed to a
final concentration of 5 mg of protein per milliliter, and a series of deter-
gent concentrations including 0.01, 0.03, 0.1, 0.3, 1.0, and 3.0% (w/v) are
prepared for each detergent to be tested. If only a restricted number of
detergents are to be evaluated initially, it is suggested that octyl glucoside
and CHAPS be considered first. As pointed out, however, it will be found
useful to examine all the detergents enumerated in Fig. 1.
The individual aliquots are stirred gently with a magnetic stirrer for 1
hr at 4°. It is important to avoid foaming and sonication, both of which
lead to protein denaturation. The individual preparations are centrifuged
at 105,000 g for 1 hr at 4 °. The supernatant liquid from each sample is
removed from the corresponding pellet, after which the pellet is resus-
pended in an equal volume of buffer with the identical detergent concen-
tration.
Assays for function and protein concentration are performed both on
the supernatant liquid and on the resuspended pellet for each centrifuged
sample. Care should be taken to resuspend each pellet completely; a
hand-held homogenizer or similar device may be helpful.
The results of this trial should be plotted as percentage of the particu-
late activity on the ordinate, against detergent concentration on the ab-
scissa, for both the solubilized supernatant liquid and the resuspended
pellet. It is also useful to plot the sum of these two curves on the same
graph.
Examination of the summed activities directly reveals whether the use
of any detergent leads to a progressive loss of the total activity as concen-
tration is increased. The detergent or detergents that yield the highest
activity in the supernatant liquids as well as the highest total activity,
should be chosen for further work.
Four general cases can be distinguished. In the first, detergents lead to
progressive solubilization of activity that is stable at high concentrations
of detergent. Figure 2 illustrates this for the solubilization of immuno-
globulin E (IgE) receptor with several detergents. 16 In the second case,
,6 B. Rivnay and H. Metzger, J. Biol. Chem. 257, 12800 (1982).
312 TECHNIQUES FOR MEMBRANE PROTEINS [16]

100

.-j
.Q

O
40
¢/)

20
I
0.1 0.3 1 3 10 30 100
Detergent Concemration (raM)
FIG. 2. Solubilization of immunoglobulin E-receptor from rat basophilic leukemia cells as
a function of the concentration of several detergents26 • 0 , Triton X-100; [] I-q,
CHAPS; A A, sodium cholate; O O, octyl glucoside.

soluble activity is seen to increase and then to decrease as detergent

concentration is raised, giving rise to an optimum value of activity at
some detergent concentration. This has been shown for the solubilization
of opiate receptor by CHAPS]7; the peak of soluble activity in this case is
presumably the result of progressive solubilization and inactivation at
higher detergent concentrations. In case 3, most or all of the activity
remains in the pellet.
If all eight detergents have proved to be unsuccessful, a number of
mixtures of detergents should be considered next. Specifically, one of the
four steroidal detergents might be mixed with either Triton X-100 or Lu-
brol PX on a 1 : 1 (w/w) basis. Such an approach has been useful for
solubilizing active fl-adrenergic receptors (digitonin and Triton X-100) 18
and cytochrome P-450 (cholate or CHAPS and Triton N-101). 19
In case 4, all activity is lost upon treatment with detergent. In this
instance, evaluation of the buffer and the protein : detergent ratio is indi-
cated as detailed in the two sections below.

Stabilization
If either very low activity or no activity at all is observed with any of
the eight detergents, systematic changes of the buffer with additives
known to stabilize solubilized proteins should be considered. Four com-

17 W. F. Simonds, G. Koski, R. A. Streaty, L. M. Hjelmeland, and W. A. Klee, Proc. Natl.

Acad. Sci. U.S.A. 77, 4623 (1980).
1~W. L. Strauss, G. Gahi, C. M. Frazer, and J. C. Venter, Arch. Biochem. Biophys. 196, 566
(1979).
19 M. Warner, M. Vella La Marca, and A. H. Neims, Drug Metab. Dispos. 6, 353 (1978).
[16] DETERGENT SOLUBILIZATION OF PROTEINS 313

mon classes of compounds that perform this function are polyols, reduc-
ing agents, chelating agents, and protease inhibitors. The following should
be considered: 25 and 50% glycerol (v/v); 1 mM dithiothreitol or 5 mM
mercaptoethanol; 1 mM EDTA; and phenylmethylsulfonyl fluoride
(PMSF) at 75/zg/ml, leupeptin at 20/zg/ml, and pepstatin (at acidic pH) at
20/zg/ml. Cytochrome P-450, for example, has an absolute requirement
for 20% glycerol for solubilization and is further protected by both
dithiothreitol and EDTA.
If no combination of detergents and protective reagents produces any
soluble activity, the problem will be a very difficult one. On the other
hand, if there is substantial recovery, albeit at low yields, the same mea-
sures that are outlined here will be useful in improving recovery in the
solubilization procedure.

Optimization of Protein-to-Detergent Ratio

The initial solubilization experiment is directed at finding a detergent
and buffer system that will serve as the basis for refinement. Many of the
proteins that have been solubilized appear to have well-defined detergent-
to-protein ratios for optimum solubilization, which the initial experiment
is unlikely to determine. The detergent-to-protein ratio, therefore, may be
a critical parameter for successful solubilization, although little attention
has been given to routine examination of this parameter. Since solubiliza-
tion is primarily achieved by dispersion of phospholipids, and since a
stoichiometric relationship appears to exist between phospholipids and
detergents in the formation of mixed micelles and solubilized complexes,
a single detergent concentration without knowledge of the protein con-
centration would not be expected to provide an optimum yield. Obvi-
ously, the value that should be presented is the optimum detergent-to-
lipid ratio; since the concentration relationship of protein and lipid in a
given membrane is relatively constant, and protein concentration is much
easier to measure, detergent-to-protein ratio is a reasonable substitute.
The optimum detergent-to-protein ratio may be experimentally evalu-
ated by solubilization at several different detergent concentrations for
three different protein concentrations. The detergent concentrations
should span the initial trial carried out at a protein concentration of 5 mg/
ml. In addition, experiments at higher concentrations of protein, possibly
7.5 and 10 mg/ml, should be performed with detergent concentrations that
correspond in multiples of 1.5 and 2 times those used for the experiment
with protein at 5 mg/ml. It seems to be less useful to work at concentra-
tions of 1 and 3 mg of protein per milliliter, since detergents with a high
CMC may not be present at solubilizing concentrations for a given deter-
 314 TECHNIQUES FOR MEMBRANE PROTEINS [16]

40O A B

E
Q.
o
300
f-

0
rn
a 200

5<
o
100

I I
ii I I I I
5 10 15 20 0 1.0 2.0
Detergent Concentration Detergent (w~
(mM) Protein ~w)
FIG. 3. Solubilization of [3H]DALAMID binding activity from NGI08 membranes by
CHAPS at three different initial protein concentrations2° as (A) a function of detergent
concentration, and (B) a function of the ratio of detergent to protein. Protein concentrations
(mg/ml) are: 11.1 (Q), 6.5 (A), and 4.3 ( I ) .

gent-to-protein ratio when the concentration of protein is low. This obser-

vation points to the fact that protein solubilization occurs at or near the
CMC for most detergents. Therefore, optimum detergent-to-protein ratios
for low concentrations of protein are expected to be slightly higher than
those for relatively high concentrations of protein.
Two plots of the results should be prepared. The first is a solubiliza-
tion curve for three different values of initial protein concentrations. The
ordinate would be solubilized activity, and the abscissa, detergent con-
centration. The second plot uses an abscissa showing the ratio of deter-
gent to protein (w/w). One of two types of curves can be expected in the
first plot. Either the percentage of particulate activity will increase and
plateau at a given concentration for each protein concentration, or the
soluble activity will rise and then fall to yield an optimum detergent con-
centration for each protein concentration. The second type of plot will
reflect these two cases as well. If soluble activity rises and then plateaus,
the selection of an appropriate detergent concentration can be based on
considerations of the size of the solubilized species (see below). If, on the
other hand, activity rises and falls as a function of detergent concentra-
tion, the plot of detergent-to-protein ratio will provide an optimum value
of this parameter. That value will in turn allow the choice of a suitable
detergent concentration for further study whenever protein concentration
[16] DETERGENT
SOLUBILIZATION
OFPROTEINS 315

Binding Lysis Solubilization Delipidation

Membranes H DetergentU Lysed U Lipid H Protein I

Boundto ! ~ Membranes[ ~ Protein Detergent
Membranes Detergent
+ +

I Detergent
Lipid I Lipid I
Detergent

I I I I
VeryLow CMC(0.1) (1-2) (10-20)
Detergentto ProteinRatio(w/w)
FIG. 4. Progressive solubilizationof whole membranes to form protein-detergentcom-
plexes as a functionof the ratioof detergentto protein. AdaptedfromHeleniusand Simons.1

is known. A representative example 2° of the second case is given in Fig. 3.

Panel A shows activity-detergent concentration profiles for three different
protein concentrations. In panel B the same data are plotted as a function
of the detergent-to-protein ratio. At the two higher protein concentra-
tions, a reasonable optimum of the detergent-to-protein ratio is observed
at approximately 0.4. At the lower initial protein concentration this value
is shifted to 0.6. Presumably, this discrepancy is due to a lowering of the
CHAPS concentration below the value needed for solubilization (approxi-
mately 0.15%), when at a protein concentration of 4.3 mg/ml the protein-
detergent ratio is 0.4. As noted, low protein concentrations may lead to
slightly higher values for detergent-to-protein ratios when detergents with
high values of the CMC are used. Since all these parameters are interre-
lated, it is obvious that buffer composition, pH, and temperature could be
optimized again at this stage. However, such repetition is probably un-
necessary.

Physical Characterization of the Solubilized Species

Once solubilization has been achieved, the further effects of change in
detergent concentration on the size and composition of solubilized com-
plexes should be examined. A schematic representation of the changes in
size and state of association of the species to be solubilized as a function
of detergent-to-protein ratio is presented in Fig. 4. At extremely low
20W. A. Klee and L. M. Hjelmeland,unpublisheddata.
316 TECHNIQUES FOR MEMBRANE PROTEINS [16]
concentrations of detergent, monomers partition into the membrane with-
out gross alterations in membrane structure. As the concentration of de-
tergent increases, the structure of the membrane is grossly changed, lead-
ing to lysis. Finally, at detergent-to-protein ratios of about 1:1, the
production of slowly sedimenting complexes occurs, which we define as
soluble. At this point, the species being generated are usually large, heter-
ogeneous complexes of lipids, detergent, and protein, with molecular
weights on the order of 0.5 to 1 million. An increase in the detergent-to-
protein ratio from 10 : 1 to 20 : 1 leads to the formation of protein-deter-
gent complexes that are free of lipid, and to mixed micelles of lipid and
detergent. Residual interactions between proteins, including those that
are artifactual, may not be dissociated at this point, nor may they be
capable of being dissociated by detergents that preserve biological func-
tion.
Electrostatic, and possibly hydrogen bonding, interactions also play
an important role in protein interactions and must be recognized and dealt
with by reagents directed at these forces. EquallY, the choice of detergent
may qualitatively affect the degree of protein interaction. It is now widely
accepted, for example, that bile salts and their derivatives are much more
efficient than the nonionic detergents in dissociating protein complexes.
Within the class of bile salts and their derivatives, more subtle effects can
be obtained by changing polar groups. With cholic acid as detergent, for
example, the molecular weight of cytochrome P-450 is of the order of
500,000, whereas, with the zwitterionic cholate derivative CHAPS, the
weight of this complex is only about 1 0 0 , 0 0 0 . 21 In this respect, the com-
bining of different detergents also appears to be useful in disaggregating
complexes. Specifically, the mixing of steroidal detergents and nonionics,
as has been noted, may lead to reduction in the total size of protein-
detergent complexes, an effect observed for adenylate cyclase22 and cyto-
chrome P-450.19
Perhaps the simplest means of determining an approximate molecular
weight is by gel filtration. 23 Modern gel filtration media are stable to each
of the detergents listed in the table and the fundamental protocol for
carrying out such a study for solubilized proteins is the same as that for a
soluble protein. It is, however, important to realize that the molecular
weight estimate yielded by this method includes all bound lipid and deter-
gent. Since many well-characterized detergent protein complexes contain
up to 50% of their total weight as detergent, this is a significant contribu-

21 L. M. Hjelmeland, D. W. Nebert, and J. C. Osborne, Jr., Anal. Biochem. 130, 72 (1983).

zz A. C. Newby and A. Chrambach, Biochem. J. 177, 623 (1978).
23 C. Tanford, Y. Nozaki, J. A. Reynolds, and S. Makino, Biochemistry 13, 2369 (1974).
[16] DETERGENT SOLUBILIZATION OF PROTEINS 317

tion. The matter is often further confused by calibration of the gel filtra-
tion column with soluble proteins in the presence of detergents. Many
soluble proteins may bind very small amounts or no detergent at all; i.e.,
their molecular weight is unaffected by the presence of detergents. More
fundamental problems exist as well. The axial ratios and partial specific
volumes of membrane detergent complexes are usually quite different
from those of soluble proteins, and assumptions based on the equality of
such parameters between standards and unknowns are prone to system-
atic errors. The magnitude of the errors is unknown. It would seem appro-
priate that relative values produced by different detergents be assessed in
order to determine a detergent that yields a functional species of minimum
size.
The other general method for measuring the size of detergent-protein
complexes is gel electrophoresis. 6 Although more laborious in some re-
spects, several advantages apply to it. Several conditions can be exam-
ined at the same time in several different gel tubes and, since the physical
dimension of such gels is small, a marked economy of sample is achieved.
The resolving power of gel electrophoresis is also somewhat higher than
that obtained with standard gel filtration columns. In principle, such
studies are carded out by examining the relative mobility or the desired
activity in gels of several different total concentrations of acrylamide. The
results are then plotted as the logarithm of the Rf versus the percentage
gel concentration, the familiar Ferguson plot. 24

Comparison of Particulate and Solubilized Function

The final task in assessing the success of solubilization is to examine
the function of the solubilized species relative to its particulate origin.
Standard methods for evaluation of kinetic parameters should be used if
the activity in question is catalytic, whereas binding analyses such as the
Scatchard plot would evaluate the function of receptors. The need for
such evaluation is emphasized by solubilization of prolactin receptors by
Triton X-100, which results in changes of both the binding capacity and
the affinity of the ligand. 25
The use of detergents is associated with a number of special problems.
It is clear that the binding of ligand to macromolecules, whether they are
enzymes or receptors, can be affected by detergents. Detergents may bind
directly at the active site, thereby serving as a competitive ligand, or free
detergent may trap ligand and thereby appear as a high-capacity, low-
affinity site. A final consideration is the effect of detergent on the macro-

24 A. C. Newby, M. Rodbell, and A. Chrambach, Arch. Biochem. Biophys. 190, 109 (1978).
z5 R. P. Shiu and H. G. Friesen, J. Biol. Chem. 249, 7902 (1974).
318 TECHNIQUESFOR MEMBRANEPROTEINS [17]

molecule itself, either in dissociating subunits or in changing the confor-

mation of the macromolecule. 26

Comments
The solubilization of membrane proteins with maintenance of function
is frequently, but not always, feasible with presently available reagents
and relatively simple procedures. An outline of procedural steps aimed at
designing an appropriate program for solubilization has been presented.

26 D. S. Liscia, T. Alhadi, and B. K. Vonderhaar, J. Biol. Chem. 257, 9401 (1982).

[17] S e p a r a t i n g D e t e r g e n t f r o m P r o t e i n s
By A N N A J. F U R T H , HILARY BOLTON,
J E N N I F E R POTTER, a n d JOHN D . PRIDDLE

The Need for Detergent Separation. Separating unbound detergent

from hydrophobic protein may become necessary at three stages in the
purification of a detergent-solubilized protein. In the initial extraction
from membrane or other lipoprotein particle, the protein is integrated into
a detergent micelle; here its hydrophobic surfaces, previously in contact
with lipid or other hydrophobic protein, become occupied by detergent.
This step needs detergent at high concentration, both to maximize solubi-
lization and to reduce the danger of micelle sharing by unrelated pro-
tein.l,2 Excess detergent, together with unwanted phospholipid, can then
be removed in the first detergent-separation step.
The second step arises because detergents used for the initial extrac-
tion are often bulky molecules with low critical micelle concentration
(CMC) values and high aggregation numbers (see the table). This means
that they tend to be less suitable for later stages of the project. Low CMC
values, for instance, are undesirable when the micellar protein solution is
concentrated (as for enzymic or immunological work), whereas high ag-
gregation numbers give large micelles, where detergent rather than pro-
tein is the major component and may dominate the properties of the
micelle. Therefore a second detergent-separation step is often needed, so
that protein can be transferred from the extraction detergent to one form-
ing smaller micelles.
1 A. Helenius, D. R. McCaslin, E. Fries, and C. Tanford, this series, Vol. 56, p. 734.
z C. Tanford and J. A. Reynolds, Biochim. Biophys. Acta 457, 133 (1976).

Copyright© 1984by AcademicPress, Inc.

METHODSIN ENZYMOLOGY,VOL. 104 All rightsof reproductionin any formreserved.
ISBN 0-12-182004-1
318 TECHNIQUESFOR MEMBRANEPROTEINS [17]

molecule itself, either in dissociating subunits or in changing the confor-

mation of the macromolecule. 26

Comments
The solubilization of membrane proteins with maintenance of function
is frequently, but not always, feasible with presently available reagents
and relatively simple procedures. An outline of procedural steps aimed at
designing an appropriate program for solubilization has been presented.

26 D. S. Liscia, T. Alhadi, and B. K. Vonderhaar, J. Biol. Chem. 257, 9401 (1982).

[17] S e p a r a t i n g D e t e r g e n t f r o m P r o t e i n s
By A N N A J. F U R T H , HILARY BOLTON,
J E N N I F E R POTTER, a n d JOHN D . PRIDDLE

The Need for Detergent Separation. Separating unbound detergent

from hydrophobic protein may become necessary at three stages in the
purification of a detergent-solubilized protein. In the initial extraction
from membrane or other lipoprotein particle, the protein is integrated into
a detergent micelle; here its hydrophobic surfaces, previously in contact
with lipid or other hydrophobic protein, become occupied by detergent.
This step needs detergent at high concentration, both to maximize solubi-
lization and to reduce the danger of micelle sharing by unrelated pro-
tein.l,2 Excess detergent, together with unwanted phospholipid, can then
be removed in the first detergent-separation step.
The second step arises because detergents used for the initial extrac-
tion are often bulky molecules with low critical micelle concentration
(CMC) values and high aggregation numbers (see the table). This means
that they tend to be less suitable for later stages of the project. Low CMC
values, for instance, are undesirable when the micellar protein solution is
concentrated (as for enzymic or immunological work), whereas high ag-
gregation numbers give large micelles, where detergent rather than pro-
tein is the major component and may dominate the properties of the
micelle. Therefore a second detergent-separation step is often needed, so
that protein can be transferred from the extraction detergent to one form-
ing smaller micelles.
1 A. Helenius, D. R. McCaslin, E. Fries, and C. Tanford, this series, Vol. 56, p. 734.
z C. Tanford and J. A. Reynolds, Biochim. Biophys. Acta 457, 133 (1976).

Copyright© 1984by AcademicPress, Inc.

METHODSIN ENZYMOLOGY,VOL. 104 All rightsof reproductionin any formreserved.
ISBN 0-12-182004-1
[17] DETERGENTREMOVAL 319

Detergents with high CMC values are also preferable when working
with liposomes in which purified protein is inserted into a closed phospho-
lipid bilayer (vesicle) of known composition. In such "reconstitution"
procedures, removal of detergent by dialyis or dilution is often the critical
step (see also this volume [19]). Because of kinetic factors, rapid removal,
and hence a high CMC value, may be crucial for determining vesicle size
and stability. This subject has been reviewed elsewhere, 3 and here we
consider it only as a third detergent-separation step, in which protein-
bound detergent is first exchanged for phospholipid and then separated
from vesicle-incorporated protein.
Purification Strategy for a Hydrophobic Protein, Incorporating Deter-
gent-Separation Steps. The overall procedure, taking hydrophobic pro-
tein from crude extract to small micelle and possibly on to liposome, is
exemplified here mainly from work in this laboratory on lactase, a mem-
brane-bound disaccharase from the intestinal brush border. The first de-
tergent-separation step is carried out on the crude membrane extract and
involves gel chromatography in buffer of high detergent concentration
(1% Triton X-100 or Emulphogen BC-720) to remove unbound detergent
and phospholipid. Lactase-active fractions may be further purified at this
stage, e.g., by ion-exchange chromatography. Since this does not involve
detergent separation, we move here directly to the second detergent-
separation step, where micellar lactase is transferred from Triton to de-
oxycholate. The exchange involves a second gel filtrationof the partially
purified membrane extract, this time in 2% deoxycholate. In this form the
protein is readily concentrated, and its properties can be investigated as a
concentrated micellar solution or, following a third detergent-separation
step, as a vesicle-incorporated protein.
A problem not discussed here is that of removing the final traces of
tightly bound detergent from a purified protein. Given that undesirable
detergents from the extraction step can be removed by exchange with
other detergents, and that fully detergent-free hydrophobic proteins tend
to be insoluble--or at best, highly aggregated--it may be preferable to
retain a belt of solubilizing monomeric detergent for physicochemical
studies or, alternatively, to replace detergent with phospholipid for recon-
stitution studies; either procedure circumvents the problem of total re-
moval of detergent.
Choice of Detergent for Pilot Experiments on Detergent Separation.
Nonionic detergents such as Triton X-100 and Emulphogen BC-720 are
often used for the initial extraction. Where the removal of such detergents
needs to be monitored--as in the pilot experiments described here--we

3 E. Racker, this series, Vol. 55, p. 699.

320 TECHNIQUES FOR MEMBRANE PROTEINS [17]

~A~ = 22). But for routine

use Triton X-100 with its high UV absorbance ~,-,2s0
preparations with established procedures, we use Emulphogen BC-720, a
nonabsorbing detergent with similar hydrodynamic properties (see the
table). This permits protein concentration to be estimated from absorb-
ance at 280 nm. Emulphogen absorbs very little above 235 nm and has an
Al~
230 value of 0.30.

I. Micellar Dimensions
Stokes Radius. Protein-free and protein-containing micelles can be
separated from one another, from detergent monomers, or from phospho-
lipid vesicles by methods based on differences in charge, hydrophobicity,
density, or micellar dimensions. 4 Here we shall concentrate on gel filtra-
tion, which has the advantage of high protein recoveries and the possibil-
ity of monitoring separation clearly. Ultrafiltration is also worthy of note.
Both methods exploit differences in effective micellar dimensions; these
take into account the shape and hydration of the particle, as well as the
summed molecular weights of its components. Effective particle radius or
Stokes radius (a) is related to diffusion coefficient (D) by the equation D =
kT/6w~a (where k is Boltzmann constant, T absolute temperature, and
viscosity of water). In practice, Stokes radius is usually determined indi-
rectly, by comparing the hydrodynamic behavior of standard and un-
known particles on gel filtration. Elution volume can be related graphi-
cally5,6 to Stokes radius through calculation of the partition or distribution
coefficient KD (which approximates to Kav, where gav is Vs~mp~e Vvoid/Vtotal
-

-- Ovoid).
The important point is that Stokes radius is an empirical value, influ-
enced by molecular mass, hydration, and shape and measured from be-
havior under specified conditions of particles of unknown shape and hy-
dration. Where the particle is a water-soluble protein, small changes in
ionic strength or the presence of other molecules have comparatively little
effect on Stokes radius and hence on gel filtration behavior. With deter-
gent-solubilized proteins the situation is different; we have a fluid aggre-
gate of protein and detergent in which particle dimensions may be
strongly influenced by small environmental changes. Therefore to sepa-
rate protein-free and protein-containing micelles successfully and repro-
ducibly, conditions must be carefully standardized.
Size of Pure Detergent Micelles. The table gives the molecular proper-
ties of a range of detergents, using the traditional classification into ionic
4 A. J. Furth, Anal. Biochem. 109, 207 (1980).
5 S. Clarke, J. Biol. Chem. 250, 5459 (1975).
6 D. Snary, P. Goodfellow, W. F. Bodmer, and M. J. Crumpton, Nature (London) 258, 240
(1975).
[17] DETERGENT REMOVAL 321

..=
¢;

~o

0
.0
Q)

0
z
0

C~

x I 2x x xxx x xx
• ~ ~
,.J
Z
0 0

0
q~ ?
(0

0
0
O ~
-r~

0 C~

r..) ~

~,-, o
•ff:~u ~.~.-
Q;

_ o
0
•

0
(.;

0
'r=
o
322 TECHNIQUES FOR MEMBRANE PROTEINS [17]
and nonionic. In practice, this grouping may be misleading, since not all
nonionic detergents have the low CMC values characteristic of this group,
nor do ionic detergents invariably have low aggregation numbers. More
important, the blanket term ionic fails to emphasize the difference be-
tween denaturing detergents (such as sodium dodecyl sulfate) and gentler
detergents, like the bile salts, capable of solubilizing hydrophobic proteins
in native form. It is these that feature in protein purification, as described
here. 7
Detergent properties affect micellar dimensions in several ways. The
contribution of molecular mass can be expressed in terms of aggregation
number. This is the number of monomer detergent molecules that can be
accommodated within a single micelle, and it tends to be low for small
polar molecules like the bile salts, and several orders of magnitude higher
for bulky detergents like Triton (see the table). Aggregation number may
rise sharply with ionic strength, e.g., from 2.2 to 22 for deoxycholate and
from 2.8 to 4.8 for cholate, when salt concentration is increased from I0
mM to 150 mM. 8
Data on the effect of solution variables on aggregation number are still
scanty. High concentrations of counterions, for example, tend to reduce
mutual repulsion between polar "heads," allowing monomers to close
ranks and so reduce micellar dimensions; the molecular weight of a Triton
micelle falls from 95,000 to 86,000 (as determined by sedimentation equi-
librium) when Tris-HC1 is replaced by a phosphate buffer of higher ionic
strength. 9 Conversely, raising the temperature or adding phospholipid
tends to increase Stokes radius, 8 e.g., from 41 A in the pure detergent to
54/~ in mixed micelles, with detergent-to-phospholipid molar ratios of
3 : 1. Whatever the mechanism, it is clear that changes in pH, tempera-
ture, and ionic strength, and the presence of phospholipid impurities from
the initial membrane extraction, may all affect the dimensions of deter-
gent micelles, and hence their removal by gel filtration. For reproducible
results, all these variables must be carefully defined.
The quickest way to separate unbound detergent from proteins is to
dilute below the detergent CMC, so that protein-free micelles disperse
into monomers, readily distinguishable from protein-detergent micelles
by size. This requires information on CMC value, which is effectively the
highest concentration of detergent monomer attainable. In practice it rep-
resents a narrow range of values rather than a single concentration and

7 If bile salts also c a u s e denaturation, as with deoxycholate,~ a popular alternative h a s been

diiodosalicylate.
s H. H. Paradies, J. Phys. Chem. 84, 599 (1980).
9 R. J. R o b s o n and E. A. Dennis, Biochim. Biophys. Acta 5 ~ , 513 (1978).
[17] DETERGENTREMOVAL 323

fluctuates widely with changes in ionic strength, pH, and temperatm'e.

Again, for reproducible results, conditions must be carefully defined.
It is immediately apparent from the table that detergents commonly
used in the initial extraction may have CMC values so low that monomers
are present in only very small amounts. To separate such detergents in
monomer form requires the use of special micelle-dispersing tricks, to be
described later.
Size of Protein-Detergent Micelles. To separate protein-free and pro-
tein-containing micelles by size, it would clearly be an advantage to pre-
dict the dimensions of protein-detergent micelles. These depend partly on
the number of bound detergent molecules and partly on the aggregation
state of the protein. Both parameters may change as protein purification
proceeds. In the high detergent concentrations of the initial extraction,
each protein molecule may be associated with a complete detergent mi-
celle; at lower concentrations, only a few detergent molecules may re-
main, bound to discrete hydrophobic sites on the protein. Similarly, as
lipid and other proteins are removed, the ratio of protein to detergent
rises, again altering self-association and detergent-binding properties of
the protein. This may account for some of the frustratingly irreproducible
behavior of hydrophobic proteins 1° and makes it crucial to define not just
detergent concentration, but the ratio of detergent to protein. Fortunately
our chapter is restricted to the removal of protein-free micelles and does
not concern those situations where small differences in protein micellar
size are critical, as in the fractionation of mixtures of detergent-solubi-
lized proteins.

II. S e p a r a t i n g P r o t e i n a n d E x c e s s D e t e r g e n t b y G e l F i l t r a t i o n of C r u d e
Membrane Extracts
Strategy. This method applies particularly to our first detergent-sepa-
ration step, the removal of excess detergent--together with phospho-
lipidmafter the initial extraction in high-detergent buffer. We exemplify it
by the chromatography of Triton-solubilized brush border membranes on
Sephacryl S-400, eluting with high-detergent buffer (1% Triton), and as-
saying fractions for the required enzyme (lactase). Detergent concentra-
tion must be kept high during chromatography, to prevent micelle-sharing
of unrelated proteins in the crude extract.
Before starting, the likely separation between protein-free and pro-
tein-containing micelles may be prejudged from a preliminary calibration
run, using pure detergent (2% Triton) as the starting sample and low-

to R. D. C. Mcnair and A. J. Kenny, Biochem. J. 179, 379 (1979).

324 TECHNIQUES FOR MEMBRANE PROTEINS [17]

1.0 Vo vt

0
O0
¢N micelles

8 0.5 I monomers

<

o ' 200 ' 460

Elution volume (ml)
FIc. 1. Sephacryl S-400 chromatography of Triton X-100 micelles after partial dispersion
into monomers by addition of deoxycholate. The starting sample contains 1% Triton and 2%
deoxycholate; the elution buffer contains 0.02% Triton and 0.06% deoxycholate.

detergent buffer (0.02% Triton) for elution. Detergent concentration dur-

ing elution must be kept around the CMC value to prevent trailing of the
Triton micelle peak.
Method. A column of Sephacryl (approximately 2.6 by 64 cm) is eluted
at 4 ° with constant flow rate (20 ml/hr), using an LKB Varioperpex II
pump, and monitoring eluent at 278 nm. Column void volume (v0) is
determined from the elution position of high-molecular-weight Triton
aggregates (prepared by incubating a 2% Triton sample at 37° for 5 min).
The vt value is determined from the elution position of sodium chloride,
applied as a 0.2 M solution, and monitored with a conductivity meter.
(With e-DNP lysine, ot values are anomalously high, suggesting adsorp-
tion.) The elution position of pure Triton micelles is determined from a
calibration run, applying 2 ml of a 2% (w/v) solution of Triton X-100 in 10
mM Veronal buffer at pH 8.2. Elution buffer is I0 mM Veronal-0.02%
Triton. Identical chromatographic arrangements are maintained for the
micellar protein solution, applying 2 ml of a brush border membrane
extract in 10 mM Veronal buffer-2% Triton X-100. Elution buffer is 10
mM Veronal with only 1% Triton. Elution position of protein-containing
micelles of the required enzyme, here lactase, is determined by enzyme
assay. 1l
Result. The elution position of pure Triton micelles is shown in Fig. 1.
(the production of monomers is described in Section IV of this chapter).
When the same column is used to remove excess detergent from the crude
membrane extract, lactase-Triton micelles elute near the void volume,
leaving a second, clearly separated UV-absorbing peak of mixed phos-
11 M. W. Ho, S. Povey, and D. Swallow, A m . J. H u m . Genet. 34, 650 (1982).
[17] DETERGENTREMOVAL 325

pholipid-Triton micelles. The elution position of this second peak encom-

passes that of pure Triton micelles, but is considerably broader, owing to
the combined effect of phospholipid-induced swelling 9 and resin adsorp-
tion. 12
In a similar procedure, Helenius and Simons 13 have used tritiated
Triton to demonstrate removal of unbound detergent on Sepharose 6B.
Clearly, the choice of gel filtration medium depends on protein molecular
weight and aggregation state and on the detergent used. For example,
monomeric detergent-solubilized lactase has an apparent molecular
weight of 260,000 in Triton and of 160,000 in deoxycholate, 14 whereas
phage coat protein (dimer molecular weight 10,000) is excluded from
Sephadex G-200 when solubilized in Triton or Brij 96, but is well included
when solubilized in sodium dodecyl sulfate) 5

III. Transferring Proteins from Triton to Bile Salt Micelles,

Using Gel Filtration
Micelle-Dispersal Techniques. Proteins are difficult to concentrate
when solubilized in the bulky detergents often used for membrane extrac-
tion. Dialyzable monomers form only a small proportion of the detergent
population, and the micelles, even when protein free, are too large to pass
through small-pore dialysis membranes. Therefore a concentration tech-
nique such as ultrafiltration tends to produce only viscous gels, in which
both protein and detergent have been simultaneously concentrated.
However, by adding such highly polar molecules as ethanedio116 or
bile salt, 17 large micelles of Triton and similar detergent may be rapidly
dispersed into monomers. Figure 1 shows that gel filtration may be used
to demonstrate the simultaneous presence of micelles and monomers in a
partially dispersed Triton solution. Dispersal is brought about by mixing
1% Triton with 2% deoxycholate and eluting with buffer containing 0.02%
Triton and 0.06% deoxycholate. Complete dispersal of micelles to mono-
mers can be achieved by raising the ratio of bile salt to Triton, e.g., by
eluting in 2% deoxycholate.
D e t e r g e n t E x c h a n g e within a Protein-Containing Micelle. The same
micelle-dispersing techniques can be used with protein-containing Triton
micelles, giving two desirable results. First, Triton is dispersed into
12C. Huang, Biochemistry 8, 344 (1969).
13A. Helenius and K. Simons, J. Biol. Chem. 2.47, 3656 (1972).
14H. Bolton, A. J. Furth, M. W. Ho, and J. Potter, Biochem. Soc. Trans. (in press).
15See Fig. 2 in S. Makino, J. L. Woolford,C. Tanford, and R. E. Webster, J. Biol. Chem.
250, 4327 (1975).
16C. E. Frasch, Dialog, February 1-3. Amicon, Lexington, Massachusetts.
17H. BoRonand A. J. Furth, unpublished results, 1982.
326 TECHNIQUES FOR MEMBRANE PROTEINS [17]

monomers, and can readily be removed by gel filtration or dialysis. At the

same time, deoxycholate replaces Triton at hydrophobic binding sites on
the protein surface, leaving protein as the major component in a small
mixed micelle. For bile salt detergents of low aggregation number, such
micelles contain only a few detergent molecules and allow the protein to
remain soluble yet with minimum detergent presence to interfere with
physicochemical or enzymic characterization. Finally, this micelle-dis-
persal method has the great advantage of producing a hydrophobic pro-
tein solution amendable to concentration by ultrafiltration.
The method can be demonstrated with lactase using a partially purified
solution in I% Triton, previously freed of unbound Triton chromato-
graphically, as described in Section II. When lactase-active fractions from
this first chromatography are applied to the same column, but elated in
2% deoxycholate, the micellar Mr drops from 260,000 to 160,000. Within
the resolving power of the column, this is consistent with a lactase mono-
mer (Mr 160,00018) transferring from a Triton micelle to one of deoxycho-
late. At the same time, displaced Triton monomers can be seen eluting
near the vt position.19
Method. A lactase-Triton micellar solution in 10 mM Veronal at pH
8.2 and 1% Triton is chromatographed as described under Method in
Section II.1~
It is not necessary to add deoxycholate directly to the starting sample
provided the elution buffer is high in deoxycholate (2% deoxycholate in 10
mM Veronal at pH 8.2). To determine the shift in micellar lactase Mr, the
column may be calibrated in detergent-free buffer, using standard pro-
teins.

IV. Separation of Phospholipid Vesicles and Detergent Micelles by

Gel Filtration
Transferring Protein from Micelle to Vesicle. Mixed micelles of pro-
tein in bile salt detergent, prepared as above, are an excellent starting
point for incorporating protein into phospholipid vesicles by Racker's
cholate dilution method. 2° Mixed phospholipid-protein-detergent mi-
celles are formed by incubating together concentrated solutions of deter-
gent-solubilized phospholipid and detergent-solubilized protein. The de-
tergent concentration is then rapidly reduced by dilution, allowing
is H. Skovberg, H. Sj6strom, and O. Noren, Eur. J. Biochem. 114, 653 (1981).
19 Compared to Fig. 1 there is an increase in Triton monomer elution volume. This is
observed frequently after a column has been used for several consecutive runs, and it
appears to be caused by adsorption to Sephacryl. The trend could be reversed and the
original elution position reproduced by unpacking the column and suspending the resin
overnight in a large volume of detergent-free buffer before repacking.
2o E. Racker, T. F. Chien, and A. Kandrach, FEBS Left. 57, 14 (1975).
[ 17] DETERGENTREMOVAL 327

1.0
Vo L-TX position Vt
E
t-
O
oo
1 l 1 e-

L-DOC /~ nomers
~ O.5

<
..J

0 2C)0 400
Elution volume (ml)
FIG. 2. Use of Sephacryl S-400 chromatography to transfer micellar lactase from Triton
to deoxycholate. Elution buffer is 2% deoxycholate in 10 mM Veronal buffer, pH 8.2. The
starting sample is a partially purified membrane extract, freed of unbound detergent by a
previous run on the same column (eluted in 1% Triton-10 mM Veronal). The elution
position of lactase-Triton micelles in a previous run is indicated by the arrow; L-DOC
indicates the elution position of laetase-DOC micelles, and TX indicates that of displaced
Triton monomers, both in the present run. Solid line represents absorbance at 278 nm, and
dashed line represents lactase activity.

spontaneous formation of phospholipid vesicles into which protein is in-

serted via the hydrophobic domain. To separate detergent from vesicles
and at the same time estimate Stokes radius of the vesicles, the dilution
step may be followed immediately by gel filtration. On Sepharose CL-4B,
for example, vesicle-incorporated lactase elutes near the void volume
(KD = 0.02), leavingtraces of micellar lactase well included (KD = 0.3). 21
Method for Vesicle Formation. 21A sample (40/zl) of phosphatidylcho-
line (type V from egg white, supplied by Sigma as a 1 mg/10/~I solution in
chloroform-methanol) is pipetted into a 5-ml vial, dried in a stream of
nitrogen, and twice washed by dissolving in 0.5 ml of diethyl ether. Then
0.2 ml of deoxycholate (25 mg/ml in l0 mM Tris-HC1 at pH 8.5) is added.
The mixture is stirred with a Vortex mixer to yield a clear yellow solution.
After standing at 20° for 60 min, the solution is rapidly diluted to 4 ml with
detergent-free buffer. Where dilution is to be followed by gel filtration,
this buffer should be of high ionic strength, e.g., 10 mM Tris-HCl at pH
8.5 containing 0.5 M NaC1. When protein is to be incorporated, the 0.2 ml
of deoxycholate solution is an aliquot of protein-containing eluate, from
the detergent-exchange chromatographic step described in Section III.
V. Removing Detergent by Ultrafiltration
Detergents with high CMC value may be separated from protein-con-
taining micelles by simple dialysis. The main requirement is that a signifi-
21 A. J. Furth and J. D. Priddle, unpublished results, 1980.
 328 TECHNIQUES FOR MEMBRANE PROTEINS [17]
cant proportion of detergent be present as monomer rather than micelle.
However, this proportion may be artificially boosted, using either dilution
or micelle dispersers. Dilution involves reducing the concentration to well
below the CMC, by adding detergent-free buffer, then rapidly concentrat-
ing on ultrafiltration, using a membrane of low cutoff point. Alternatively,
a suitable micelle-dispersing agent is ethylene glycol. At a concentration
of 30%, ethylene glycol allows 48% of a 0.5% Triton solution to be re-
moved through a Diaflo PM-10 membrane, using an Amicon Model 52
stirred ultrafiltration cell. 16 Lower concentrations (25%) have been used
to dispel I% Triton and thereby separate unbound detergent from pepti-
dase-containing micelles. 22
Given the range of ultrafiltration membranes now available, it should
also be possible to separate detergents of low aggregation number as
micelles rather than as monomers. However, repeated exposure to con-
centrated (1%) detergent solutions may cause damage to membranes.
Furthermore, most detergent micelles are ellipsoidal, so that surprisingly
large micellar proteins may pass through small pores "head on."

VI. Comments
Because the protein-detergent micelle is a fluid, noncovalently linked
aggregate, its size and shape are much more susceptible to environmental
changes than those of a simple, water-soluble protein. Therefore, to
achieve reproducible results with detergent-solubilized proteins, it is im-
perative to standardize such parameters as ionic strength, detergent con-
centration, and protein concentration.
Protein detergent micelles may be separated from unbound detergent
by the methods described here. Also suggested is a protocol that enables
membrane proteins to be extracted (without the problem of micelle
sharing) into bulky detergent micelles, then to be transferred to small
micelles (with the aid of bile salts or similar detergent), and, finally, to be
transferred from micelle to phospholipid vesicle. Gel filtration may be
used at each of the three detergent-separation steps, both to remove
unwanted detergent and to estimate the Stokes radius of the protein-
containing particle.

Acknowledgments
This work was supported in part by grants to M. W. Ho from NuffieldFoundation and
Medical Research Council.
22B. Svensson, M. Danielsen, M. Staun, L. Jeppesen, O. Noren, and H. Sj6strom, Eur. J.
Biochem. 90, 489 (1978).
[18] , PURIFICATION OF INTEGRAL MEMBRANE PROTEINS 329

[18] Purification of I n t e g r a l M e m b r a n e P r o t e i n s
By Jos VAN RENSWOUDE and CHRISTOPH KEMPF

The study of structure and function of biological membranes has been

hampered by the lack of efficient methods for separating the functional
components of the membrane, most of which are amphipathic in nature.
Although this chapter focuses on some of the methods available for the
isolation of integral membrane proteins, it will be obvious that no step-by-
step outline can be presented.
Integral membrane proteins behave as amphiphilesl,2; they possess
hydrophilic domains, formed by ionic and polar residues (including carbo-
hydrates), which are exposed to an aqueous environment at the surface of
the membrane, and hydrophobic domains, rich in apolar residues, which
are buried within the hydrophobic core of the membrane's lipid bilayer.
The topological distribution of the hydrophobic and hydrophilic domains
of a protein determines its arrangement in the membrane. Some of these
proteins are situated in the bilayer so that the hydrophilic portion is ex-
posed at only one surface of the membrane, whereas others span the
membrane, allowing hydrophilic domains to interact with the aqueous
phase on both sides of the membrane. Endoplasmic reticulum cyto-
chrome b5 and erythrocyte acetylcholinesterase are probably organized in
the first way. There are several examples of proteins that span the mem-
brane: the major erythrocyte sialoglycoprotein, glycophorin; the erythro-
cyte anion transport protein; bacteriorhodopsin; some membrane trans-
port proteins; and viral envelope glycoproteins. It is conceivable that a
third type of membrane protein may be totally immersed within the hy-
drophobic phase of the bilayer and have totally hydrophobic surfaces.
Some of the proteolipids from myelin and mitochondria that are soluble in
organic solvents may be of this class.
It is the amphiphilic character that presents great difficulty in isolation
and purification. In the intact membrane, the proteins are associated with
a phospholipid bilayer and, at some stage, must be separated from the
phospholipid. Once dissociated, they will exhibit the same preferential
interactions that cause them to be located in a hydrophobic environment
in the first place, and their amphiphilic properties tend to make them
unstable in both aqueous and organic solvents. In aqueous media, inter-
molecular self-association at the hydrophobic surfaces of the protein is

1 j. L e n a r d an d S. J. Singer, Proc. Natl. Acad. Sci. U.S.A. 56, 1828 (1966).

2 O. F. H. W a l l a c h and P. H. Zahler, Proc. Natl. Acad. Sci. U.S.A. 56, 1552 (1966).

Copyright© 1984by AcademicPress, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproductionin any form reserved.
ISBN 0-12-182004-1
330 TECHNIQUES FOR MEMBRANE PROTEINS [18]

favored, leading to the formation of aggregates that are insoluble in most

instances. In organic solvents, on the other hand, the polar domains will
tend to self-associate in order to gain maximum protection by exposure of
hydrophobic domains to the solvent.
In devising an approach to the isolation of a specific membrane pro-
tein, it is important to define the purpose of isolating the molecule. When
the aim is structural and chemical characterization, without regard for
biological activity, a relatively wide range of isolation procedures is avail-
able, including those that result in irreversible denaturation of the protein;
such procedures form the bulk of those described. If, however, the biolog-
ical activity of the isolated protein is to be studied, the choice is limited to
techniques that do not cause irreversible denaturation; examples are few.
Since methods for the isolation of integral membrane proteins have
not yet reached the degree of discrimination of those available for separat-
ing soluble proteins, it is often advantageous to "simplify" the protein
source as much as possible before solubilization and fractionation. Unless
highly selective methods such as affinity chromatography can be applied,
it is worth making the effort to ensure that the population of cells used as
starting material is as homogeneous as possible and that the membrane
fractions obtained from these cells contain the membrane of interest in the
highest purity achievable. A useful preliminary step in isolating mem-
brane proteins is the removal of contaminating cytosolic and peripheral
membrane proteins from the membrane preparation. A number of proce-
dures have been used to solubilize peripheral membrane proteins. These
procedures include the use of chelating agents, salt in high and low con-
centration, and variations of the pH from acid to basic.
Here, we outline an approach to the isolation of an integral membrane
protein and note some of the available methods. It is necessary to stress
that isolation of these proteins, particularly isolation with residual biologi-
cal activity, is a demanding process that is not always successful.

Membrane Isolation and Purification

The (sub)cellular membranes of eukaryotic and prokaryotic cells con-
stitute the sources of membrane proteins. The usual first step in isolation
of a membrane protein is fractionation of the source tissues or cells to
obtain highly purified subcellular membranes. A multitude of cell fraction-
ation procedures are available, many of which are summarized in this
series and elsewhere. 3,4 There are a few instances in which relatively pure

3 This series, Vol. 31.

4 G. D. Birnie (ed.). "Subcellular Components, Preparation and Fractionation." But-
terworth, London, and University Park Press, Baltimore, Maryland, 1972.
[18] PURIFICATION OF INTEGRAL MEMBRANE PROTEINS 331

membranes can be obtained simply. The erythrocyte membrane is readily

available after converting red cells to red cell ghosts, and vesicles re-
leased by secretory cells and plasma membrane vesicles, obtained by
forced blebbing, 5 are also convenient starting materials. Even at this first
step it is important to minimize the risk of degradation of the proteins of
interest by endogenous or exogenous proteolytic activity. This can be
achieved by working quickly, in the cold whenever possible, and by add-
ing protease inhibitors such as phenylmethylsulfonylfluoride (PMSF), N-
tosyl-L-phenylalanylchloromethylketone (TCPK), and leupeptin.
After the desired membrane is obtained, soluble proteins that have in
some way remained associated with the membrane and peripheral mem-
brane proteins should be removed. This is usually achieved by exposing
the membranes to one or more of the following agents or conditions:
1. KCI or NaC1 in relatively high concentrations (0.15-3.0 M). This
treatment will result in greatly decreased electrostatic interaction be-
tween proteins and charged lipids. It should be noted that certain endoge-
nous proteases are activated under conditions of high ionic strength, so
that adequate precautions (inhibitors, low temperature) should be taken.
2. Washing with buffers of acid or basic pH or with sodium car-
bonate .6
3. Chelating agents such as EDTA and EGTA (up to 10 mM). These
substances destabilize membranes by complexing Mg 2÷ and Ca2+.
4. Chaotropic ions (I-, Br-, C104-, SCN-), in high concentrations (2-
4 M), act by disordering the structure of water. They disrupt hydrophobic
bonds near the surface of membrane structures and promote the transfer
of hydrophobic groups from an apolar environment to the aqueous phase.
5. The phenolic compound lithium 3,5-diiodosalicylate 7 presumably
acts by virtue of detergent-like properties. It has relatively low chemical
reactivity and is easily removed by dialysis. At concentrations above 1 M,
membrane structure generally breaks down completely.
6. Denaturing agents such as urea and guanidine hydrochloride break
noncovalent bonds if used at high concentrations (6-10 M). They are
often used in combination with reducing agents (2-mercaptoethanol or
dithiothreitol) that would cleave disulfide bonds. Use of urea and guani-
dine renders mixtures of denatured, dissociated polypeptide chains.
7. Protein-modifying reagents, such as p-chloromercuribenzoate, p-
chloromercuribenzene sulfonate, and acid anhydrides (e.g., succinic or
maleic anhydrides).

5 D. W. Tank, E.-S. Wu, and W. W. Webb, J. Cell Biol. 92, 207 (1982).
6 y . Fujiki, A. L. Hubbard, S. Fowler, and P. B. Lazarow, J. Cell Biol. 93, 97 (1982).
7 V. T. Marchesi and E. P. Andrews, Science 174, 1247 (1971).
332 TECHNIQUES FOR MEMBRANE PROTEINS [18]
All of the above-mentioned reagents are able to denature protein,
leading to irreversible loss of biological activity. In order to warrant selec-
tive removal of peripheral membrane proteins and membrane-associated
soluble proteins, while leaving integral membrane proteins in place, the
conditions in the purification procedure should be chosen such that the
lipid backbone of the membrane remains intact. After application of the
above-mentioned protocols, recovery of the membranes by centrifugation
allows the extent of loss of lipid bilayer structure to be easily assessed as a
loss of residue after centrifugation. A large decrease in turbidity of the
membrane suspension upon addition of a reagent may also be taken as
indicative of extensive lipid bilayer collapse. The process of membrane
purification and all subsequent steps in the isolation of a membrane pro-
tein can be conveniently monitored by conventional sodium dodecyl sul-
fate-polyacrylamide electrophoresis. Once membranes have been "puri-
fied" using one or more of the above procedures, they may be solubilized
to liberate their constituents, lipids as well as integral membrane proteins.

Solubilization
Solubilization, by definition, involves the disintegration of the lipid
bilayer and can be achieved in a number of different ways, of which only a
few are suitable for our purposes. Solubilization for the purpose of isolat-
ing membrane proteins is usually accomplished by treating the mem-
branes with detergents or by extracting them with organic solvents.

Detergents
The principles of detergent solubilization are discussed in detail in this
volume, s The detergents used in solubilization can be divided into three
classes: (1) nonionic detergents, such as octylglucoside and the poly-
oxyethylenes (e.g., Triton), which form relatively large micelles (Mr
50,000-100,000) and have relatively low CMC values (10 -4 to 10 -5 M);
(2) zwitterionic detergents like 3-(3'-cholamidopropyl)dimethylammonio-
1-propanesulfonate (CHAPS) and sulfobetaine; and ionic detergents, such
as cetylammonium bromide and sodium dodecyl sulfate (SDS), which
possess strongly acidic or basic polar head groups (e.g., sulfate or a
quaternary nitrogen). These detergents arrange into relatively small mi-
celles ( M r 10,000-20,000) and have relatively high CMC values (10 -2 to
10 -3 M); (3) bile salts, such as cholate, taurocholate, and deoxycholate,
which form small aggregates consisting of 2-8 monomers, at monomer
s L. M. Hjelmeland and A. Chrambach, this volume [16].
[18] PURIFICATION OF INTEGRAL MEMBRANE PROTEINS 333

concentrations of 10 -2 to 10 -3 M. The choice of a detergent for solubiliza-

tion will be determined by several factors. If, for instance, preservation of
the biological activity of the membrane protein is desired, nonionic deter-
gents such as Triton X-100 are preferable. Nonionic detergents have the
added advantage of not interfering with subsequent separation procedures
based on charged groups on the protein, e.g., ion-exchange chromatogra-
phy. A disadvantage in their use is that they are hard to remove from the
solubilization mixture because of their low CMC. 9 Removal of detergent
may be desirable if a membrane protein is to be reconstituted.I° The use of
certain resins, e.g., SM-2 Biobeads, that strongly bind nonionic am-
phiphiles, will improve the speed and completeness of their removal.11 In
the absence of a specific interest in maintaining biological activity of the
membrane protein, solubilization of the source membrane in such ionic
detergents as the bile salts may be the method of choice; one limitation in
their use is that they form very large, precipitating aggregates at low pH.
They should be used at a pH greater than 7.8.

Extraction with Organic Solvents

Some membrane proteins (proteolipids) are soluble in organic sol-

vents, especially if they are neutralized by suitable counterions. Much of
this work has been discussed in this series. 12,13This property can be used,
in principle, for purification by allowing the proteins to partition into the
organic phase of a water-organic solvent two-phase system. For any
given protein, the overall hydrophobicity of the polypeptide chain and the
presence of such polar side groups as sugars will determine the extent of
partitioning between organic and aqueous phases. Obviously, heavily gly-
cosylated membrane proteins will tend to partition into the aqueous
phase. 14,15 A number of organic solvents have been used for this pur-
pose, 12,13 the more common of which are noted below.
1. n-Butanol was originally described as a solvent for extraction of
erythrocyte membrane proteins and glycoproteins. Butanol extraction is a
relatively mild procedure that tends to allow preservation of antigenic
properties and, sometimes, enzymic function22,13 Most membrane pro-
9 A. J. Furth, H. Bolton, J. Potter, and J. D. Priddle, this volume [17].
~0R. D. Klausner, J. van Renswoude,and B. Rivnay,this volume [19].
IIp. W. Holloway,Anal. Biochem. 53, 304 (1973).
~ H. S. Penefskyand A. Tzagoloff,this series, Vol. 22, p. 204.
~3A. Tzagoloffand H. S. Penefsky,this series, Vol. 22, p. 219.
14D. J. Anstee and M. J. A. Tanner, Biochem. J. 138, 381 (1974).
15D. J. Anstee and M. J. A. Tanner, Eur. J. Biochem. 45, 31 (1974).
334 TECHNIQUES FOR MEMBRANE PROTEINS [18]
teins partition into the aqueous phase of the butanol-water system,
whereas lipids find the butanol phase.
2. n-Pentanol is less selective than butanol in that polar lipids will,
together with membrane proteins, partition into the aqueous phase. Pen-
tanol extraction may be considered if lipid-protein interactions are to be
preserved, e.g., in the case of membrane enzymes that require a lipid
environment for activity.
3. Aqueous phenol, 50% or greater by volume, is used at either neu-
tral or acid (acetic acid, 25-33%) pH. J6 Lipids and a large proportion of
membrane protein partitions into the organic phase, whereas glycopro-
teins are predominantly found in the water layer.
4. Pyridine, used in pyridine-water (2 : l, v/v) mixtures, is reputed to
be a good solubilization reagent.17 At low temperatures, it has been suc-
cessfully applied in solubilizing relatively labile membrane components
with preservation of biological activity.
5. Chloroform-methanol mixtures (chloroform-methanol-water,
18:9: 1) have been used extensively as lipid extraction solvent sys-
tems) 8,19Membrane proteins generally copartition with the lipid into the
organic phase, whereas most membrane glycoproteins are recoverable
from the aqueous phase.
6. Both acetic and formic acid interact strongly with hydrogen bonds;
both act as proton donors to the carbonyl moiety of the amide group,
competing with the amide proton. Both solvents also perturb hydrophobic
interactions between nonpolar residues of the protein. The result is dena-
turation, aggregation in aqueous solvents, and even covalent modification
(formylation) of membrane proteins.2° Acetic acid and formic acid should
be avoided if biological activity of membrane proteins is to be maintained.
7. 2-Chloroethanol is a weakly protic solvent with relatively high solu-
bilizing power, presumably due to its content of HCI (2-chloroethanol is
unstable and slowly decomposes to form HCI). It does not, to any appre-
ciable extent, affect the hydrophobic domains of membrane proteins, and
it may represent one of the more suitable solvents for the study of
protein-lipid interactions. 2~,22 2-Chloroethanol is generally used at a

16 K. Takayama, D. H. McLennan, A. Tzagoloff, and C. D. Stoner, Arch. Biochem.

Biophys. 114, 223 (1964).
17 O. Blumenfeld, Biochem. Biophys. Res. Commun. 30, 200 (1968).
is E. G. Bligh and W. J. Dyer, Can. J. Biochem. Physiol. 37, 911 (1959).
19j. Folch, M. Lees, and G. H. Sloane Stanley, J. Biol. Chem. 226, 496 (1957).
2o W. Menke and H. G. Ruppel, Z. NaturJbrsch. B Anorg. Chem. Org. Chem. 26B, 825
(1971).
21 p. Zahler and D. F. H. Wallach, Biochim. Biophys. Acta 135, 371 (1967).
22 p. Zahler and R. E. Weibel, Biochim. Biophys. Acta 219, 320 (1970).
[18] PURIFICATION OF INTEGRAL MEMBRANE PROTEINS 335

concentration of 90%, by volume. It allows full recovery of lipid-free

membrane proteins in a soluble, presumably largely monomeric, form,
although biological activity may be lost.
8. Other solvents (Methyl Cellosolve, Ethyl Cellosolve, dimethyl-
formamide, formamide, N-methylpyrrolidone, hexafluoroacetone, and
diethylene glycol monobutyl ether) have been used, mainly under acid
conditions, in extracting membrane proteins. Their application, however,
has been limited to specific examples. 23-25
Organic solvent extraction procedures may be useful in attempts to
isolate some of the integral membrane (glyco)proteins. Organic solvents,
however, tend to result in at least some degree of protein denaturation;
their use should, as a rule of thumb, be considered mainly in those in-
stances in which biological activity need not be preserved.

Fractionation of Membrane Proteins

The choice of methods for fractionating a mixture of membrane pro-
teins depends, in part, on the procedure used in solubilization. After
solubilization, a membrane protein exists either in a mixed micelle with
detergent, dissolved in an organic solvent, or in an aqueous environment,
usually in an aggregated state. Upon solubilization with a detergent, the
resultant mixture includes mixed lipid-detergent micelles, mixed protein-
detergent micelles, and mixed lipid-protein-detergent micelles. Protein-
containing mixed micelles can be conveniently separated from nonpro-
tein-containing ones by density gradient centrifugation, since the buoyant
density of the latter is usually lower than that of the protein-containing
species. 26Protein-containing micelles can then be subjected to a variety of
fractionation methods (see below). Membrane proteins that are dissolved
in organic solvents cannot, in general, be fractionated as such; the or-
ganic solvent should be removed by evaporation and the proteins subse-
quently taken up in a solution of detergent. Similarly, aqueous dispersions
of aggregated membrane proteins should be solubilized in detergent be-
fore additional fractionation can be undertaken. All chromatographic
procedures noted below, except gel filtration, can be carried out in a
batch-wise manner, with often significant gain in speed.
Phase Separation. Solubilization of membrane proteins with a number
of detergents leads to formation of protein-detergent micelles. Several

23 B. Kohl and H. Sandermann, FEBS Lett. 80, 408 (1977).

24 R. L. Juliano, Biochim. Biophys. Acta 266, 301 (1972).
25 p. H. Zahler, D. F. H. Wallach, and E. F. Luescher, Protides Biol. Fluids 15, 67 (1967).
J. Yu, D. A. Fischman, and T. L. Steck, J. Supramol. Struct. 1, 233 (1973).
336 TECHNIQUES FOR MEMBRANE PROTEINS [18]
properties of these mixed micelles, including size and hydrophilicity, are
a property of the detergent used as well as of the solubilized protein. The
common detergent Triton forms clear micellar solutions in a given range
of temperature. Upon warming, the solution undergoes phase separation
resulting in two clear phases, one depleted and the other enriched in
detergent. The phase separation temperature is dependent on the deter-
gent structure (Triton X-100 at approximately 64°; Triton X-114 at approx-
imately 20°). Utilizing this property, integral membrane proteins have
been separated from the bulk of the erythrocyte membrane proteins with
Triton X-114. 27 The method seems to be efficient in separating integral
membrane proteins from soluble and extrinsic proteins. Differences be-
tween detergent micelles and protein-detergent micelles with regard to
their partitioning in a two-phase system have also been used; the light-
harvesting chlorophyll a/b protein was isolated with Triton X-100 in com-
bination with the aqueous two-phase system dextran-polyethylene gly-
col. 28 Phase separations of this sort have been described in detail by
Albertsson, 29-31 and certain variations in the technique are described in
this volume. 32
Gel Filtration. An early step in the purification of membrane proteins
will usually be gel filtration, which achieves at least partial purification.
Proteins originating from a wide variety of membranes have been purified
by this means. 33-35 The large size of nonionic detergent micelles usually
necessitates separation on agarose rather than on dextran or acrylamide
media, and the viscosity of the detergent-containing solutions tends to
make the separation rather slow. A new generation of gel matrices based
on a hydrophilic vinyl polymer (Fractogel, Merck) may prove to be
superior.
Ion-Exchange Chromatography. Ion-exchange chromatography can
be used for membrane proteins in nonionic or zwitterionic detergents.
Some proteins have the unfortunate characteristic of binding to the resin
in an apparently irreversible fashion. Several proteins have been purified
by means of ion-exchange chromatography on DEAE-cellulose, e.g.,

27C. Bordier,J. Biol. Chem. 256, 1604(1981).

28 P.-A. Albertsson and B. Andersson,J. Chromatogr. 215, 131 (1981).
P.-A. Albertsson, "Partitionof Cell Particles and Macromolecules."Wiley,New York,
1971.
3oP.-A. Albertsson, Endeavour 1, 69 (1977).
31P.-A. Albertsson, J. Chromatogr. 159, 111 (1978).
32K. C. Ingham, this volume [20]; G. Johansson,this volume [21].
33M. J. A. Tanner, R. G. Jenkins, D. J. Anstee, and J. R. Clamp, Biochem. J. 155, 701
(1976).
34M. Sone, M. Yoshida, H. Hirata, and Y. Kagawa,J. Biol. Chem. 250, 7917 (1975).
3~K. Kameyama,T. Nakae, and T. Takagi,Biochim. Biophys. Acta 706, 19 (1982).
[18] PURIFICATION OF INTEGRAL MEMBRANE PROTEINS 337

5'-nucleotidase from liver plasma membranes (solubilized in sulfobetaine-

14, a zwitterionic detergent), 36 and the erythrocyte glucose transport pro-
tein. 37
Affinity Chromatography. Affinity chromatography involves the selec-
tive adsorption of a protein to a matrix bearing molecules that will specifi-
cally bind to the protein. The bound protein can then be eluted from the
resin by disruption of the specific interaction. 3a Since most membrane
proteins are glycosylated, lectins coupled to a resin have found applica-
tion. The dissociation of the lectin-sugar linkage often can be released by
simple sugars under mild conditions. One limitation of the method is that
most immobilized lectins are stable only in nonionic detergents. Cationic
and zwitterionic detergents, for example, inhibit concanavalin A and soy-
bean agglutinin, whereas SDS interferes with the binding capacity of most
lectins. 39 A serious complication in the use of immobilized lectins is that
the glycoproteins may be heterogeneous in their glycoconjugate content
and, therefore, will not bind uniformly to any given lectin; such proteins
would be recovered in low yields. 4° Other affinity methods also have been
successful for the purification of membrane receptors; instead of lectins,
receptor-specific ligands have been coupled to a resin matrix. Examples
include the insulin receptor, purified on insulin-agarose4~ ; the acetylcho-
line receptor, purified with the help of quaternary ammonium ligands as
analogs of cholinergic compounds42 ; and the fl-adrenergic receptor, puri-
fied from digitonin-solubilized frog erythrocyte membranes by chroma-
tography on a Sepharose-alprenolol column. 43 For proteins with very
great affinity for the ligand, the dissociation step may prove to be too
difficult to be practical. An affinity ligand should be chosen with this
constraint in mind so as to achieve freely reversible binding.
The development of the monoclonal antibody technique to individual
antigens of complex biological mixtures has already greatly facilitated the
isolation of membrane proteins and will continue to do so. Several have
been isolated by monoclonal antibody affinity chromatography, including
the Ia antigens, 44 HL-A antigen, 45 and the human transferrin receptor. 46

36 E. M. Baiyles, A. C. Newby, K. Siddle, and J. P. Luzio, Biochem. J. 203, 245 (1982).

37 M. Kasahara and P. C. Hinkle, J. Biol. Chem. 252, 7384 (1977).
A detailed discussion of the method is presented in Vol. 34 of this series and is updated in
this volume [1]-[4].
39 R. Lotan, G. Beattie, W. Hubbell, and G. L. Nicoison, Biochemistry 16, 1787 (1977).
4o M. J. A. Tanner and D. J. Anstee, Biochem. J. 152, 265 (1976).
41 p. Cuatrecasas, Proc. Natl. Acad. Sci. U.S.A. 69, 1277 (1972).
42 j. O. Dolly and E. A. Barnard, Biochemistry 16, 5053 (1977).
43 R. G. L. Shorr, S. L. Heald, P. W. Jeffs, T. N. Lavin, M. W. Strohsacker, R. J. Lefko-
witz, and M. C. Caron, Proc. Natl. Acad. Sci. U.S.A. 79, 2778 (1982).
44 W. R. McMaster and A. F. Williams, Immunol. Rev. 47, 117 (1979).
338 TECHNIQUES FOR MEMBRANE PROTEINS [18]

The use of monoclonal antibodies for general protein purification is pre-

sented in this volume [24].
The choice of detergent for membrane protein solubilization in immu-
noaffinity chromatography is restricted to the use of nondenaturing spe-
cies. The dissociation step in immunoaffinity chromatography often may
require such harsh methods as low pH or chaotropic agents, all undesir-
able if a functional protein is to be recovered. An electrophoretic desorp-
tion method 47 for dissociating membrane proteins from immobilized
high-affinity ligands such as antibodies is a relatively mild procedure
through which the use of immunoabsorbents in membrane protein isola-
tion may be extended. Immunoaffinity chromatography has also been
used in combination with covalent chemical modification of membrane
proteins. Rothman and Linder48 labeled intact cells, isolated the washed
membranes, and purified the chemically modified proteins with antibodies
against haptens on the labeling reagents. Isolated platelet membrane pro-
teins were labeled in the carbohydrate moiety with 2,4-dinitrophenyl-
alanine hydrazide, and erythrocyte membrane proteins were labeled with
diazodiiodoarsanilic acid (a tyrosine-specific reagent); antidinitrophenyl
and antiarsanilic acid antibodies, respectively, served in the immobilized
phase.
Covalent Chromatography. Covalent chromatography allows the se-
lective absorption of proteins from a mixture to an insoluble matrix by
means of covalent, reversible reactions between reactive groups on the
protein and immobilized chemical reagents on the resin. Most commonly,
thiol groups on the protein and the matrix are used to form a disulfide that
can subsequently be cleaved from the column with a low-molecular-
weight thiol. Major erythrocyte membrane proteins, glycophorin and
band III, have been purified by this approach with an organic mercurial
linked to agarose. 49 The selectivity of the method is not high because of
the frequency of free SH groups on proteins. Nevertheless, useful purifi-
cation can be achieved.
Hydrophobic Interaction Chromatography and HPLC. Some at-
tempts have been made to separate integral membrane proteins on the
basis of their variation in hydrophobicity. If proteins are applied to chro-
matographic media that themselves have hydrophobic characteristics,

45 p. Parham, J. Biol. Chem. 254, 8709 (1979).

I. S. Trowbridge and M. B. Omary, Proc. Natl. Acad. Sci. U.S.A. 78, 3039 (1981).
47 M. R. A. Morgan, P. J. Brown, M. J. Leyland, and P. D. G. Dean, FEBS Lett. 87, 239
(1978).
48 A. Rothman and S. Linder, Biochim. Biophys. Acta 641, 114 (1981).
49 M. L. Lukacovic, M. B. Feinstein, R. I. Shaafi, and S. Perrie, Biochemistry 20, 3145
(1981).
[18] PURIFICATION OF INTEGRAL MEMBRANE PROTEINS 339

e.g., octyl- or phenylagarose: °,5~ separation on the basis of hydrophobic

interactions may be possible (see this volume [3]). Mitochondrial proteins
have been partially resolved on columns containing hydrophobic acryl-
amide derivatives: 2 and attempts have been made to purify erythrocyte
membrane proteins on a column of N-(3-carboxypropionyl) aminodecyl-
agarose, 53 a compound having ionic as well as hydrophobic characteris-
tics. The expression of hydrophobicity of such a resin can be controlled
by pH.
Methods for separating membrane proteins by means of reversed-
phase HPLC are in an early stage of development but are covered in this
volume [9]. The molecular weight of porin, a protein of the outer mem-
brane of Escherichia coli, has been determined54by this method, although
the protein was previously purified by conventional chromatographic pro-
cedures. Purification of two membrane proteins from T cells, H-2 and
Lyt-2, made use of reversed-phase HPLC (Vydac RP-18 column) as well
as gel permeation HPLC (BioGel TSK column) with a Nonidet P-40-
containing elution buffer) 5 However, the relatively small pore size (100
/~) led to low yields in protein recovery. Much better results in fractionat-
ing T-cell membrane proteins were obtained when the membranes were
solubilized in trifluoroacetic acid (0.1%) and acetonitrile (10%) and then
separated on a wide-pore (300 ~) reversed-phase C~8 column eluted with
an acetonitrile gradient (V. L. Alvarez, personal communication).
Other Methods. The qualitative protein composition of membranes is
most often determined by SDS-polyacrylamide electrophoresis. Proteins
have been successfully fractionated on an analytical scale from erythro-
cyte and platelet membranes by preparative detergent electrophoresis on
a 2 x 6 cm acrylamide gel) 6 The capacity of such gels is moderate; up
to 60 mg of proteins was separated by the method. Inherent to this
preparative method is the exposure of proteins to SDS and the high risk of
denaturation. Additional limitations of the method include potential co-
migration of nonidentical protein species, since the separation is mainly
based on the molecular weight of the proteins. This electrophoretic ap-
proach nonetheless has the advantage of speed and simplicity.

~0 j. Rosengren, S. Pahlman, M. Glad, and S. Hjert6n, Biochim. Biophys. Acta 412, 51

(1975).
51 S. D. Carson and W. H. Konigsberg, Anal. Biochem. 116, 398 (1981).
52 H. Weiss and T. Buecher, Eur. J. Biochem. 17, 561 (1970).
53 R. J. Simmonds and R. J. Yon, Biochem. J. 157, 153 (1976).
K. Kameyama, T. Nakae, and T. Tagaki, Biochim. Biophys. Acta 706, 19 (1982).
55 V. L. Alvarez and M. Mage, Fed. Proc. Fed. Am. Soc. Exp. Biol. 41, 838 (1982).
W. L. Nichols, D. A. Grastineau, and K. G. Mann, Biochim. Biophys. Acta 554, 293
(1979).
340 TECHNIQUES FOR MEMBRANE PROTEINS [19]

[19] R e c o n s t i t u t i o n o f M e m b r a n e P r o t e i n s
By R I C H A R D D . K L A U S N E R , J o s VAN R E N S W O U D E ,
and BENJAMIN RIVNAY

Reconstitution of membrane proteins continues to be a crucial step in

studying the function and structure of these molecules. The word recon-
stitution is poorly defined but, in general, refers to the reincorporation of
a solubilized membrane protein into a natural or artificial membrane. The
major virtue of this technique is realized only when functional reconstitu-
tion of activity is accomplished. The necessity for reconstitution arises
because many membrane proteins express their full activity only when
correctly oriented and inserted in a lipid bilayer. In order to purify a
membrane protein to any degree, it is necessary first to remove it from its
natural membrane. Thus, it becomes necessary to develop a reconstitu-
tion scheme in order to study the function of these purified (or partially
purified) components. The last review of methods for membrane protein
reconstitution to appear in this series provided a thorough overview of the
field in 1979,1 with concentration primarily on membrane enzymes. Dur-
ing the past several years, increasing attention to another class of mem-
brane proteins, receptors, has provided a number of approaches to the
reconstitution of integral membrane receptors.
The vast majority of reconstitution procedures can be summarized as
involving (1) solubilization of the protein with a suitable detergent (this
volume [16] and [18]); (2) mixing the solubilized protein with either lipid-
detergent micelles or preformed lipid or natural membranes; and (3) re-
moving the detergent (see also this volume [17]). The result, if successful,
is the integral incorporation of the protein into the lipid bilayer. This
summary gives little feeling for the great complexity and variety of ap-
proaches. The specific detergent and the precise conditions of the solubili-
zation, the membrane or lipid preparation, the order of mixing, and the
technique of detergent removal have been varied freely, and only recently
has there begun to be a semblance of rational order to the process of
successful reconstitution.

Detergent, Lipid, and the Denaturation of Proteins

The loss of function during reconstitution can be broadly assumed to
be the result of denaturation, which, in turn, is the result of the detergent
' E. Racker, this series, Vol. 55, p. 699.

Copyright © 1984by Academic Press, Inc.

METHODS IN ENZYMOLOGY,VOL. 104 All rightsof reproduction in any formreserved.
ISBN 0-12-18201)4-1
[19] RECONSTITUTION OF MEMBRANE PROTEINS 341

solubilization. The functions that can be lost range from the ability to
insert into a lipid bilayer, ligand-binding activity, enzyme activity, and ion
fluxes, among others. By trial and error, optimal detergents have been
found that minimize the denaturation of specific membrane proteins when
used for solubilization (see this volume [16]). Examples include digitonin
for the fl-adrenergic receptor, 2 CHAPS for the IgE receptor3 and the
opiate receptor, 4 cholate for the acetylcholine receptor, 5 and octyl gluco-
side for rhodopsin. 6 It has been known for some time that phospholipid,
present during solubilization and reconstitution, protects against dena-
turation of integral membrane enzymes. This has been extended to recep-
tors as well. 7 In an attempt to bring order to the chaos of detergent-
induced denaturation, Rivnay and Metzger 3 have proposed a single
parameter, p, which defines the relationship of lipids and detergent. This
parameter is defined by Eq. (1) in which CMCeff is the critical micelle
concentration under the particular experimental conditions and will de-
pend on salt and lipid concentrations.
p = [(detergent) - CMCeff]/(phospholipid) (1)
The function p is similar to the parameter Refr, the effective ratio,
derived by Jackson et al. s [Eq. (2)].
R,f~ = [(detergent) - 0.22]/(phosphatidylcholine) (2)
An analysis of the literature on reconstitution reveals that the likeli-
ness of detergent-induced denaturation rises as p increases. 9 The p value
is minimized by using a low detergent concentration while maintaining a
high phospholipid concentration. On the other hand, if 19is too small, the
protein will not be solubilized. Experience with the acetylcholine receptor
suggests that p > 1.5 is required for solubilization. In general, every
molecule of protein is not solubilized; for the acetylcholine receptor,
cholate solubilizes only 70% of the receptor. The sensitivity of a receptor
to denaturation during the solubilization and reconstitution process de-
pends on the criteria chosen to assess denaturation. This is well illustrated
for the acetylcholine receptor. Agonist-induced Na + flux appears to be
2 R. Neubig, E. Krodel, N. Boyd, and J. Cohen, Proc. Natl. Acad. Sci. U.S.A. 76, 690
(1979).
3 B. Rivnay and H. Metzger, J. Biol. Chem. 257, 12800 (1982).
4 W. Simonds, G. Koski, R. A. Streoty, L. Hjelmeland, and W. A. Klee, Proc. Natl. Acad.
Sci. U.S.A. 77, 4623 (1980).
R. Anholt, J. Lindstrom, and M. Montal, J. Biol. Chem. 256, 4377 (1981).
G. W. Stubbs, H. G. Smith, and B. J. Litman, Biochim. Biophys. Acta 425, 46 (1976).
7 M. Epstein and E. Racker, J. Biol. Chem. 253, 6660 (1978).
8 M. L. Jackson and B. J. Litman, Biochemistry 21, 5601 (1982).
9 R. D. Klausner, J. van Renswoude, R. Blumenthal, and B. Rivnay (in press).
342 TECHNIQUES FOR MEMBRANE PROTEINS [19]

most sensitive to functional loss, and high p values, owing to either excess
detergent or the absence of lipids, lead to loss of this activity.l° Less than
10% loss was reported with cholate at p = 2, whereas no channel activity
remained at a p of 40 or more. From several studies, the optimum p for
this receptor appears to be 1.5 < p < 10 for solubilization and p = 2 for
the reconstitution. Agonist binding is less readily lost, and the ability to
insert into a lipid bilayer is the function that appears most resistant to
denaturation. Studies with the receptor for IgE confirm and extend
these findings.3 Solubilization of 90% of the receptor is achieved at p = 2.
Again, the presence of lipids, which keeps p low, protects against dena-
turation. The situation is more complex than might be implied by the p
value alone. Thus, the IgE receptor is relatively better protected from
denaturation by inclusion of natural lipids from the cells from which the
receptor is isolated than is evident when soybean lecithin is used.

Specific Examples of Receptor Reconstitution

Virtually every reconstitution system that has been reported is differ-
ent. It is, therefore, difficult to assign specific and rational explanations
for each of the details unique to each study. However, specific examples
of successful reconstitutions are useful as menus, if not exact recipes, for
the general procedures that are available. It is the purpose of this section
to outline the details of specific examples of receptor reconstitution so as
to allow adaptation of this group of experimental protocols to the needs of
other investigators.

Acetylcholine Receptor
Anholt et al. reported a well-documented reconstitution of this recep-
tor into lipid vesicles. 11 The vesicles demonstrate activation and desensi-
tization and distinguish between agonists and antagonists. Numerous
other groups have reported successful reconstitution of this receptor. 12-~6

10 R. L. Huganir, M. A. Schell, and E. Racker, FEBS Lett. 108, 105 (1979).

u R. Anholt, D. R. Fredkin, T. Deernick, M. Ellisman, M. Montal, and J. Lindstrom,
J. Biol. Chem. 257, 7127 (1982).
12 W. Wu and M. A. Raftery, Biochem. Biophys. Res. Commun. 89, 26 (1979).
13j. Lindstrom, R. Anholt, B. Einarson, A. Engel, M. Osame, and M. Montal, J. Biol.
Chem. 255, 8340 (1980).
14L-P. Changeux, J. Heidmann, J.-L. Popot, and A. Sobel, FEBS Len. 105, 181 (1979).
15L M. Gonzalez-Ros, A. Paraschos, and M. Marring-Carrion, Proc. Natl. Acad. Sci.
U.S.A. 77, 1796 (1980).
16R. Anholt, Trends Biochem. Sci. (Pers. Ed.) 6, 288 (1981).
[19] RECONSTITUTION OF MEMBRANE PROTEINS 343

Membranes enriched in the acetylcholine receptor may be prepared

from Torpedo californica electroplax organs.17 Such membranes consist
predominantly of receptor subunits but include as well an extrinsic mem-
brane protein of Mr = 43,000. The latter protein, not considered part of
the receptor system, can be removed by alkaline extraction.IS The resul-
tant preparation is then available for solubilization and reconstitution. 6
Solubilization. Membranes are solubilized in 2% sodium cholate in the
presence of crude soybean lecithin (L-ot-phosphatidylcholine type II,
Sigma) at a lipid concentration of 5 mg/ml in l0 mM sodium phosphate at
pH 7.4 containing 0.1 M NaCl. Lipids are prepared as a stock solution
containing 150 mg/ml by dispersion in distilled water with a bath sonica-
tot. Sonication is performed under a stream of argon to reduce lipid
oxidation. The cholate-membrane suspension is gently shaken at 4° for
18-24 hr. At the end of this solubilization period, the mixture is centri-
fuged at 165,000 g for 30 min and the insoluble material is discarded. The
process results in 65% solubilization of receptor as shown by 125I-labeled
o~-bungarotoxin-binding activity. Raising the cholate concentration above
3% does not enhance receptor solubilization but does lead to diminished
functional activity after reconstitution.
Reconstitution. After solubilization, soybean lipids (from the 150 mg/
ml dispersion) are added to the extract to a final lipid concentration of 25
mg/ml. At this point, the concentration of the receptor is 1-2/~M. Cholate
is removed by dialysis for 16-18 hr against 500 volumes of 100 mM NaC1,
10 mM NAN3, and 10 mM sodium phosphate at pH 7.4. This step is
followed by 16-18 hr of dialysis against 500 volumes of 145 mM sucrose,
I0 mM NAN3, and 10 mM sodium phosphate at pH 7.4.
The vesicles formed have a mean diameter of 520/~. Two maneuvers
can increase their size. The vesicles can be frozen at - 2 0 °, followed by
thawing at room temperature and resulting in vesicles with an average
diameter of 620 .~. If cholesterol is incorporated during reconstitution at a
ratio (cholesterol : phospholipid, w/w) of I : 4, the vesicles have an aver-
age diameter of 760 ~ after a single cycle of freeze-thawing.

Receptor for Immunoglobulin E (IgE)

A number of variables have been evaluated as to their effect on recon-
stitution, 3 success being measured by the incorporation of the receptor
into lipid vesicles.

17j. Elliott, S. G. Blanchard, W. Wu, J. Miller, C. D. Strader, P. Hartig, H.-P. Moore, J.

Racs, and M. A. Raftery, Biochem. J. 185, 667 (1980).
18 R. Neubig, E. Krodel, N. Boyd, and J. Cohen, Proc. Natl. Acad. Sci. U.S.A. 76, 690
(1979).
344 TECHNIQUES FOR MEMBRANE PROTEINS [19]

Membranes are prepared from rat basophilic leukemia cells by first

washing the cells in buffer containing I20 mM NaC1, 15 mM Tris-HC1 at
pH 7.6, 5 mM KC1, 0.5 mM MgCI2, and 0.5 mM CaCl2. The cells are
swollen in a 1 : 3 dilution of this buffer with water, including 1 mM phenyl-
methylsulfonyl fluoride (PMSF), 20 mM iodoacetamide, and 10 mg of
DNase I per milliliter for 3-5 min. They are then sonicated for 1 min in a
cuphorn sonicator at 0 ° and 60% of maximum output. The suspension is
centrifuged at 31,000 g, and the pelleted membranes are suspended in full-
strength buffer by brief sonication.
Lipids are prepared from the rat basophilic leukemia cells grown as
tumors. Tumor (10 g) is homogenized in 25 ml of 0.16 M sodium borate at
pH 8.0 containing 0.2 M NaCl and mixed with 4 volumes of chloroform-
methanol (2 : l, v/v). The upper layer and interface formed after centrifu-
gation for 20 min at 3000 rpm are extracted with chloroform and combined
with the initial lower layer. Liposomes are made by first drying the chlo-
roform solution containing 20 mg of phospholipids under nitrogen, leaving
a lipid film that is subjected to lyophilization overnight. Vesicles are
formed by swelling the lipid in 1 ml of the Tris-saline buffer (described
above) for 15 min at room temperature. The suspension is sonicated for 15
min with a microtip (Model W225R, Heat Systems Ultrasonics, Inc.) at an
output of 50% of maximum.
Reconstitution is achieved by mixing membranes, tumor lipids as lipo-
somes, and CHAPS. Total phospholipid is 6 mM, of which approximately
4% is derived from the receptor-containing membranes (determined by
inorganic phosphate). The detergent concentration is 12.2 mM, which
yields a p value of 1.25. The mixture can be dialyzed either immediately
or after incubation at 2° for periods of up to 18 hr against 500 volumes
of the Tris-saline buffer using 3 to 5 changes over 3-4 days. The result
is the incorporation of about 90% of the receptor into vesicles as deter-
mined by banding on 3 to 50% sucrose gradients run for about 40 hr in an
SW 50.1 rotor at gay = 189,000 in the cold.

fl-Adrenergic Receptor
For the fl-adrenergic receptor from rat erythrocyte membranes, 2 the
membranes are washed and suspended at i0 mg of protein per milliliter in
20 mM HEPES at pH 8.0 containing 2 mM EDTA. Digitonin, 70 mg/ml, is
added to give a detergent:protein ratio of 3.5 : l (w/w). After stirring at 2 °
for 1 hr, the solution is diluted with an equal volume of saturated ammo-
nium sulfate. After 30 min at 2°, the material is centrifuged at 50,000 rpm
(50 Ti rotor) for 30 min, leaving the solubilized receptor in the supernatant
[19] RECONSTITUTION OF MEMBRANE PROTEINS 345

fluid. The procedure results in 50% solubilization with a twofold purifica-

tion of the receptor.
Dimyristoylphosphatidylcholine vesicles are sonicated at a lipid con-
centration of 150 mM to form small unilamellar vesicles. They are added
to the membrane extract (all in 55 mM HEPES at pH 8.0-5.5 mM EDTA-
10.6 mM MGC12-98 mM NaCI-0.89 M NH4SO4) to give a final lipid con-
centration of 5 mM. The mixture is chromatographed on Sephadex G-50
in 50 mM Tris-HC1 at pH 7.5-5 mM MgC12-100 mM NaCI in the absence
of detergent. The early void volume fractions are collected. These frac-
tions are turbid and contain 50-65% of the receptors, total protein, and
phospholipid, but only 10-20% of the digitonin. The detergent concentra-
tion at this point is 0.5-1 mM, still higher than the CMC for digitonin
(0.08-0.3 mM). This fraction is further treated by rate zonal centrifuga-
tion on a linear sucrose density gradient. Reconstituted vesicles are found
in a turbid band at 27% sucrose. This band contains 50% of the receptor
activity, 6/zmol of lipid per milligram of protein, and no detectable digito-
nin (the limit of detection is 40 ~M). Electron microscopic studies reveal
unilamellar vesicles, 500-900/~ in diameter. The reconstituted vesicles
display ligand-binding properties similar to those found for the intact
receptor in native membranes. Interestingly, the agonist, iodohydroxy-
benzylpindolol, which binds to the native receptor, does not bind to the
digitonin-solubilized receptor. However, full binding activity is regained
upon reconstitution and detergent removal.
One of the most exciting aspects of reconstitution studies with this
receptor involves that of the interacting components of the adrenergic
receptor-cyclase system.~9 Separate solubilization of the receptor and the
nucleotide binding subunit (G) with subsequent reconstitution into phos-
pholipid vesicles has been described.
Receptor Solubilization. Washed turkey erythrocyte membranes are
treated with 50 mM potassium phosphate at pH 11.9 for 20 min at 4°, a
procedure that does not affect the receptor, but destroys G protein activ-
ity. The membranes are centrifuged at 27,000 g for 15 min and suspended
in 20 mM MOPS at pH 7.0 and 1 mM mercaptoethanol at a final protein
concentration of 1 mg/ml. Phospholipid vesicles are prepared by sonica-
tion of 10 mg of soybean lecithin per ml of 10 mM Tris-HC1 (pH 7.5)-2
mM EDTA. For each 1 ml of suspended membranes, 0.2 ml of vesicle
suspension is added. The mixture is incubated at 4° in 10 mM MgC12.
Upon addition of 20 ~M i soprenaline, an adrenergic ligand, the incubation
temperature is raised and maintained at 30° for 10 min, and the mixture is

19 y . Citri and M. S c h r a m m , Nature (London) 282, 297 (1980).

346 TECHNIQUES FOR MEMBRANE PROTEINS [19]

centrifuged for 10 min at 18,000 g. The membrane pellet is suspended in

20 mM MOPS (pH 7.0)-0.5 M sucrose-0.2 mM EDTA-0.1 mM mercap-
toethanol-0.01 mM PMSF and then solubilized by the addition of an
equal volume of sodium deoxycholate (12 mg/ml). Insoluble material is
removed by centrifugation for 30 min at 200,000 g.
G Protein Solubilization. Turkey erythrocyte membranes are washed
in 20 mM HEPES (pH 8.0)-2 mM mercaptoethanol-1 mM MgCI2-0.2
mM EDTA-0.02 mM PMSF and resuspended at a protein concentration
of 10 mg/ml. An equal volume of sodium cholate (20 mg/ml) is added, and
the mixture is incubated at 4 ° for 30 min. Insoluble material is removed by
centrifugation for 30 min at 200,000 g. The supernatant liquid contains
solubilized G protein but is without receptor activity.
Reconstitution. Phospholipid vesicles are prepared by sonication (5
min) of 10 mg of lecithin and 1 mg of stearylamine in 1 ml of 10 mM Tris-
HCI at pH 7.5 containing 1 mM EDTA. The two solubilized proteins are
mixed in desired proportions and treated with SM-2 resin (Bio-Rad) to
remove detergent. This is accomplished by mixing 1.2 g of wet resin per
milliliter of mixture, followed by shaking for I hr. The resin is removed by
centrifugation, and the preformed vesicles are mixed with the supernatant
fluid to give 10 mg of phospholipid per milliliter of supernatant liquid. The
suspension is incubated briefly at 4 °, and the resultant reconstituted vesi-
cles are centrifuged at 20,000 g for 15 min.
Insertion of Reconstituted Vesicles into Biological Membranes. Re-
constituted proteoliposomes 2° can be reinserted into biological mem-
branes in a manner such that the inserted receptors will couple function-
ally with integral membrane enzymes. 21 The reconstituted vesicles that
have been described are centrifuged at 18,000 g for 10 rain.
Friend erythroleukemia cells are suspended at 5 × 106 cells per millili-
ter in a solution of 135 mM NaCI, 5 mM KC1, 0.8 mM MgCI2, and 20 mM
Tris-HC1 at pH 7.4. One milliliter is layered over the vesicles, and the
cells are allowed to sediment at room temperature. The buffer is removed,
and the pellet of cells and vesicles is mixed. A solution of polyethylene
glycol, 0.5 ml at 37°, is added and mixed. This solution contains 1040 mg
of PEG 6000 per milliliter of the cell buffer, modified by containing 5 mM
glucose, 4 mM MgCI2, 2 mM ATP, and 0.1 mM EDTA. The modified
buffer is added in progressively increasing volumes after 100 sec and
every 2 min thereafter, employing the following additions: 0.2, 0.3, 0.5,
1.5, 3.5, and 7 ml. The resultant mixture is stored at 2° after centrffugation

2o M. Schramm, Proc. Natl. Acad. Sci. U.S.A. 76, 1174 (1979).

2i S. Eimerl, G. Neufeld, M. Korner, and M. Schramm, Proc. Natl. Acad. Sci. U.S.A. 72,
760 (1980).
[19] RECONSTITUTION OF MEMBRANE PROTEINS 347

and suspension in the buffer. The procedure is reported to allow the

implantation of the receptor and G protein into the cell membrane.

Insulin Receptor
The insulin receptor may be solubilized from a preparation of turkey
erythrocyte membranes by treatment of 2.9 mg per milliliter of membrane
protein with 1% octyl fl-glucoside in 30 mM NaC1, 10 mM sucrose, 1 mM
EDTA, and 85 mM Tris-HC1 at pH 7.8. 22The mixture is stirred for 15 min
at room temperature, and insoluble membrane components are removed by
centrifugation for 60 min at 4° and 104,000 g. The solubilized receptor in
the liquid phase is mixed with a 20-fold excess of phospholipid in 2% octyl
glucoside. Two types of phospholipid mixtures have been used: soybean
phosphatidylcholine-bovine brain phosphatidylserine (4 : 1, w/w) and di-
myristoylphosphatidylcholine-bovine brain phosphatidylserine (4: 1,
w/w). Each mixture of lipid, detergent, and solubilized protein can be
purified with Sephadex G-50 equilibrated in the absence of detergent at
a temperature that is above the phase transition temperature of the lipid
system used. Material eluting in the void volume of the eluate is dialyzed
for 48 hr. The vesicles are centrifuged at 104,000 g for 1 hr at 4°, sus-
pended in buffer, and centrifuged in a 2 to 30% continuous sucrose gradi-
ent until equilibrium is reached. Vesicles of either lipid composition band
at 1.071 g/ml. Both types of vesicle preparations consist almost entirely of
unilamellar structures about 1000 A in diameter. These vesicles contain the
insulin receptor as demonstrated by the specific binding of iodoinsulin.
Although both types of lipid vesicles allow successful receptor reconstitu-
tion, the affinity of the reconstituted receptor for insulin is markedly
different for each lipid system.

22 R. J. Gould, B. H. Ginsberg, and A. A. Spector, J. Biol. Chem. 257, 477 (1982).

[20] POLYETHYLENE GLYCOL FRACTIONATION 351

[20] P r o t e i n P r e c i p i t a t i o n with P o l y e t h y l e n e Glycol

B y KENNETH C. INGHAM

The use of nonionic water-soluble polymers, in particular polyethyl-

ene glycol (PEG), for fractional precipitation of proteins was introduced
by Poison et al. 1 and discussed by Fried and Chun in this series. 2 The
intervening years have provided an improved understanding of the molec-
ular basis of the protein-precipitating action of PEG and additional docu-
mentation of the unique advantages of this polymer over other reagents
used for this purpose. Although much of the literature on this subject
deals with purification of proteins from blood plasma, 3 the approach is
applicable to any complex mixture. The principles involved have been
further clarified by studies with purified proteins, and the purpose of this
chapter is to summarize briefly these principles with emphasis on practi-
cal information enabling the reader to assess the potential applicability of
this technique to specific separation problems.

Advantages of Polyethylene Glycol

The advantages of PEG as a fractional precipitating agent stem primar-
fly from its well-known benign chemical properties. Unlike ethanol and
other organic precipitating agents, PEG has little tendency to denature or
otherwise interact with proteins even when present at high concentrations
and elevated temperatures. Careful experiments designed to test this prin-
ciple revealed that PEG 4004 and PEG 40004 at concentrations up to 30%
(w/v) had no detectable effect on the circular dichroic spectrum or thermal
denaturation temperature of ribonuclease) The low heat of solution and
the relative insensitivity of PEG-precipitation curves to minor variations
in temperature eliminate the need for controlling temperature during re-
agent addition. Another advantage of PEG over ethanol or ammonium
sulfate is the shorter time required for the precipitated proteins to equili-

i A. Poison, G. M. Potgieter, J. F. Largier, G. E. F. Mears, and F. J. Joubert, Biochem.

Biophys. Acta 82, 463 (1964).
2 M. Fried and P. W. Chun, this series, Vol. 22, p. 238.
Y. L. Hao, K. C. Ingham, and M. Wickerhauser, in "Methods of Protein Fractionation"
(J. M. Curling, ed.), p. 57. Academic Press, New York, 1980.
4 PEG = poly(ethylene glycol), poly(ethylene oxide), polyoxyethylene. Chemical formula:
HOCH2CHz(CH2CH20)nCHzCH2OH. PEG 400 and PEG 4000 signify heterogeneous mix-
tures having nominal average molecular weights of 400 and 4000, respectively.
5 D. H. Atha and K. C. Ingham, J. Biol. Chem. 256, 12108 (1981).

Copyright © 1984 by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
352 OTHER SEPARATION SYSTEMS [20]

brate and achieve a physical state suitable for large-scale centrifugation.

The advantages of PEG in facilitating the growth of protein crystals is well
documented. 6,7

Mechanism of Action
Careful measurements with a variety of purified proteins indicate that
their solubilities decrease exponentially with increasing concentration of
PEG according to Eq. (1)
log S = log So - tiC (1)
where S is the solubility in the presence of PEG at concentration C (%
w/v) and So is the a p p a r e n t intrinsic solubility obtained by extrapolation
to zero PEG. 5 Plots of log S vs [PEG] exhibit striking linearity over a wide
range of protein concentration, the slope for a given protein being rela-
tively insensitive to pH and ionic strength but markedly dependent upon
the size of the PEG up to about 6000 daltons. The slopes also tend to
increase with increasing size of the protein, reinforcing the popular notion
of a steric exclusion mechanism whereby proteins are concentrated in the
extrapolymer space, eventually exceeding their solubility limit under the
given solution conditions. Although a quantitative explanation of this
behavior is yet to come, it is clear that, in the absence of specific
interactions, the sequence of precipitation of several proteins in a mixture
will depend primarily on the ratios of their initial concentrations relative
to their respective solubilities in the absence of PEG. Thus, even though
larger proteins have steeper slopes, a large protein initially present at high
concentration could precipitate later than a small one present at low con-
centration if the intrinsic solubility of the latter is much less than that of
the former. Manipulation of the solution conditions is expected to im-
prove the separation of a given pair of proteins to the extent that their
intrinsic solubilities diverge.

Which PEG to Use?

Most workers use material with a nominal average molecular weight in
the 4000-6000 range. Polymers larger than this offer no advantage, since
their solutions are more viscous and the precipitation curves are not much
different from those obtained with PEG 6000.1,5 Decreasing the molecular
weight below 4000 spreads the precipitation of a mixture over a broader

6 W. B. Jakoby, this series, Vol. 22, p. 248.

7 A. McPherson, Jr., J. Biol. Chem. 251, 6300 (1976).
[20] POLYETHYLENE GLYCOL FRACTIONATION 353

range of PEG concentrations. The improved resolution that might be thus

anticipated is partially offset by the shallower slopes obtained for individ-
ual proteins. Nevertheless, Honig and Kula 8 found the degree of purifica-
tion of y-glucosidase from yeast extract to be about twofold greater with
PEG 400 than with PEG 4000 or 6000. That PEG 400 is a liquid at room
temperature whose solutions are substantially less viscous than those of
the higher polymers, coupled with the potentially greater ease of remov-
ing it by molecular sieve methods, indicates a need for further com-
parisons.

The Analytical Precipitation Curve

The following simple experiment is designed to quickly overcome ig-
norance about the amount of PEG required to precipitate a given pro-
tein(s) from a complex mixture. The scale of this experiment is dictated
by the sensitivity of the assay employed; the availability of a radiolabeled
tracer is a definite advantage. One dispenses a fixed amount (0.1-0.5 ml)
of the mixture into a series of tubes (preferably in duplicate) to each of
which is subsequently added an equal volume of buffer containing in-
creasing amounts of PEG to produce a final concentration of 25-30% in
the most concentrated tubes. It is important to buffer the PEG stock
solutions to avoid PEG-induced changes in pH. 5,9 The increment in PEG
concentration is arbitrary, but 3% (w/v) is adequate for initial screening.
The vigor with which one mixes these solutions depends on the extent to
which the desired protein(s) can withstand mechanical stress; gentle agi-
tation on a vortex mixer is one approach. After 0.5-1.0 hr of incubation at
room temperature or on ice, the samples are centrifuged and the percent-
age of the desired activity remaining in the supernatant liquid is deter-
mined. Inspection of the resulting "analytical precipitation curve" pro-
vides an estimate of the maximum concentration of PEG that can be
added at one time without precipitating the protein of interest as well as
the minimum concentration required to bring it out of solution, parame-
ters that can then be more precisely defined with a second experiment that
focuses on the relevant concentration range. With luck, the curve will fall
either far to the left or far to the right on the PEG axis, defining a simple
one-step method for removing a large portion of unwanted macromole-
cules and/or concentrating the desired activity prior to further processing
by other methods. Otherwise, it may be necessary to obtain a "PEG cut"

s W. Honig and M.-R. Kula, Anal. Biochem. 72, 502 (1976).

9 G. Eichele, D. Karabelnik, R. Halonbrenner, J. N. Jansonius, and P. Christen, J. Biol.
Chem. 253, 5239 (1978).
354 OTHER SEPARATION SYSTEMS [20]

via two precipitation steps utilizing in turn the m a x i m u m and minimum

concentration of PEG referred to above.
It is always possible to manipulate the precipitation curve horizontally
along the PEG axis by varying solution conditions. For screening pur-
poses, it is expedient to choose a fixed concentration of PEG that causes
approximately 50% precipitation of the desired protein under a given set
of solution conditions in order to determine rapidly the extent to which
altering those conditions might enhance or inhibit precipitation. The most
gratifying result of this approach would be to identify substances or condi-
tions that selectively influence the solubility of the desired protein. This
concept is further developed in the following section.

Influence of Protein-Protein and Protein-Ligand Interactions

Studies with purified self-associating and heteroassociating proteins
have shed some light on the role of protein-protein interactions on solu-
bility in the presence of PEG. 5,1°-12 Based on the above-mentioned ex-
cluded volume considerations, one predicts that conditions that foster
protein association should enhance precipitation because of the larger
size of the complexes, whereas those that inhibit association would have
the opposite effect. This is the case with almost all systems that have been
examined. Of particular relevance in the present context was the observa-
tion 1° that bovine liver glutamate dehydrogenase at 2.8 mg/ml in 0.2 M
potassium phosphate at pH 7.0, conditions known to promote extensive
self-association, was quantitatively precipitated by PEG 4000 at concen-
trations above 15% (w/v). Such precipitation was completely inhibited,
even at higher concentration of PEG, by the combined presence of 10-3 M
NADH and GTP, cofactors known to reverse the self-association. It re-
mains to be seen whether this ligand-specific manipulation of the solubil-
ity could be exploited in a fractionation scheme. Similar effects were
observed with chymotrypsin, chymotrypsinogen, and fl-lactoglobulin A,
in which cases self-association was manipulated by varying pH and ionic
strength, parameters likely to be less selective. Nonspecific electrostatic
interactions between oppositely charged proteins such as albumin and
lysozyme can also have profound effects on solubility that are most pro-
nounced at low ionic strength at a pH between the pI of each of the two
proteins.l~ While such interactions are frequently viewed as a nuisance, to

lo S. I. Miekka and K. C. Ingham, Arch. Biochem. Biophys. 191, 525 (1978).

11 S. I. Miekka and K. C. Ingham, Arch. Biochem. Biophys. 203, 630 (1980).
lz j. Wilf and A. P. Minton, Biochim. Biophys. Acta 670, 316 (1981).
[20] POLYETHYLENE GLYCOL FRACTIONATION 355

be minimized by maintaining near-physiological ionic strength, the possi-

bility of using them to advantage in a purification scheme should be kept
in mind.
A more specific type of heteroassociation of the type that might be
exploited in purification is the functional interaction between human
plasma fibronectin and denatured collagen, i.e., gelatin. The precipitation
curve for the plasma protein in phosphate-buffered saline shifted from
11% PEG to less than 3% PEG upon addition of gelatin, which by itself
was not precipitated by PEG under these conditions. ~3Since the complex
between the two proteins is very stable, even at high ionic strength, it
should be possible to precipitate fibronectin selectively from a complex
mixture by this method. The contaminating gelatin could then be re-
moved, e.g., by ion-exchange chromatography in the presence of urea.
Although the advantage of this approach over affinity chromatography on
immobilized gelatin is debatable, the example serves as an additional
illustration of the application of bioaffinity principles to fractional precipi-
tation. Any substance that interacts specifically with the desired protein
has the potential to alter its solubility selectively and should thus be
tested. Enzymes are ideal candidates for this approach, since they often
interact with one or more effectors or cofactors, sometimes with large
changes in the state of association.

Methods of Removing PEG

In many applications, PEG is used early in the purification scheme and
is removed during subsequent chromatographic steps on ion-exchange or
affinity columns to which PEG has no tendency to adsorb. A word of
caution is in order regarding the application of PEG-containing solutions
to some exclusion columns, the performance of which can be significantly
altered owing to osmotic effects of the polymer. 14Alternative approaches
to removing PEG include ultrafiltration 15,16and salt-induced phase separa-
tion 17 as reviewed. 18The latter method is particularly useful for solutions
containing relatively high concentrations of PEG and has the potential
advantage that the protein may be concentrated in a low-volume, salt-

13 K. C. Ingham, S. A. Brew, and S. I. Miekka, Mol. Immunol. 20, 287 (1983).

14 K. Helising, J. Chromatogr. 36, 170 (1968).
15 T. F. Busby and K. C. Ingham, J. Biochem. Biophys. Methods 2~ 191 (1980).
16 K. C. Ingham, T. F. Busby, Y. Sahlestrom, and F. Castino, in "Ultrafiltration Membranes
and Applications" (A. R. Cooper, ed.), p. 141. Plenum, New York, 1980.
17 T° F. Busby and K. C. Ingham, Vox Sang. 39, 93 (1980).
18 K. C. Ingham and T. F. Busby, Chem. Eng. Commun. 7, 315 (1980).
356 OTHER SEPARATION SYSTEMS [21]

rich phase. For many research purposes it is probably unnecessary to

remove all traces of polymer from the final product, since it is optically
transparent 19 and helps prevent loss of protein by adsorption on glass.

Summary
Polyethylene glycol is a nondenaturing water-soluble polymer whose
ability to precipitate protein from aqueous solution can be qualitatively
understood in terms of an excluded volume mechanism. The increment in
PEG concentration required to effect a given reduction in solubility is
unique for a given protein-polymer pair, being insensitive to solution
conditions and primarily dependent on the size of the protein and poly-
mer. Selective manipulation of the solubility of specific proteins through
control of their state of association or ligand environment can potentially
remove some of the empiricism otherwise involved in fractional precipita-
tion. Adequate methods for removing the polymer are available.

19 The low level of U V a b s o r b a n c e frequently found in s o m e P E G preparations is not

inherent to the p o l y m e r but is due to a small a m o u n t of antioxidant s o m e t i m e s added by
the m a n u f a c t u r e r .

[21] Affinity P a r t i t i o n i n g
By GOTE JOHANSSON

Partition of enzymes and other proteins between two liquid aqueous

phases can be strongly influenced by specific or group-specific ligands
bound to a water-soluble polymer included in the two-phase system. The
two nonmiscible phases are obtained by mixing water solutions of two
polymers. Several pairs of polymers can be used, 1 but the most popular
system has been the one containing dextran and polyethylene glycol
(PEG). To obtain two phases, the concentrations of the two polymers
must exceed certain values that depend on the molecular weights of the
polymers and the temperature. The composition of the phases can be
found in the phase diagrams elaborated by Albertsson3 The partition of
proteins between the phases depends on a number of factors that include
P.-A. Albertsson, "Partition of Cell Particles and M a c r o m o l e c u l e s . " Wiley, N e w York,
1971.

Copyright© 1984by AcademicPress, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All righls of reproduction in any form reserved.
ISBN 0-12-182004-1
356 OTHER SEPARATION SYSTEMS [21]

rich phase. For many research purposes it is probably unnecessary to

remove all traces of polymer from the final product, since it is optically
transparent 19 and helps prevent loss of protein by adsorption on glass.

Summary
Polyethylene glycol is a nondenaturing water-soluble polymer whose
ability to precipitate protein from aqueous solution can be qualitatively
understood in terms of an excluded volume mechanism. The increment in
PEG concentration required to effect a given reduction in solubility is
unique for a given protein-polymer pair, being insensitive to solution
conditions and primarily dependent on the size of the protein and poly-
mer. Selective manipulation of the solubility of specific proteins through
control of their state of association or ligand environment can potentially
remove some of the empiricism otherwise involved in fractional precipita-
tion. Adequate methods for removing the polymer are available.

19 The low level of U V a b s o r b a n c e frequently found in s o m e P E G preparations is not

inherent to the p o l y m e r but is due to a small a m o u n t of antioxidant s o m e t i m e s added by
the m a n u f a c t u r e r .

[21] Affinity P a r t i t i o n i n g
By GOTE JOHANSSON

Partition of enzymes and other proteins between two liquid aqueous

phases can be strongly influenced by specific or group-specific ligands
bound to a water-soluble polymer included in the two-phase system. The
two nonmiscible phases are obtained by mixing water solutions of two
polymers. Several pairs of polymers can be used, 1 but the most popular
system has been the one containing dextran and polyethylene glycol
(PEG). To obtain two phases, the concentrations of the two polymers
must exceed certain values that depend on the molecular weights of the
polymers and the temperature. The composition of the phases can be
found in the phase diagrams elaborated by Albertsson3 The partition of
proteins between the phases depends on a number of factors that include
P.-A. Albertsson, "Partition of Cell Particles and M a c r o m o l e c u l e s . " Wiley, N e w York,
1971.

Copyright© 1984by AcademicPress, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All righls of reproduction in any form reserved.
ISBN 0-12-182004-1
[21] AFFINITY PARTITIONING 357

the concentrations of polymers, the type and concentration of salt in-

cluded, the pH, the nature of polymers carrying charged groups, and the
temperature. 1,2 This large number of variables allows the design of a
system in which most proteins are concentrated in one phase. An enzyme
can then be extracted into the opposite phase by introducing a specific
ligand (for the target enzyme) that is bound to the polymer concentrated
in this phase, i.e., affinity partitioning. For large-scale extractions it is
necessary to choose ligand-polymers that can be prepared easily in suffi-
cient quantities and at moderate cost. Polymer derivatives of reactive
triazine dyes fulfill these requirements and can be used for extraction of a
number of dehydrogenases and kinases. 3-5 A number of other ligand-
polymers, useful for affinity partitioning, and methods for their prepara-
tions have been summarized elsewhere. 6,7

Synthesis of Triazine Dye-Polyethylene Glycol

Polyethylene glycol (Mr = 6000-7500), 100 g, is dissolved in 200 ml of
water. The solution is heated in a water bath to 80°, and 10 g of triazine
dye, e.g., Cibacron Blue F3G-A, is added followed by 2 g of LiOH. The
LiOH can be replaced by NaOH, but this results in a lower yield. The
mixture is maintained at 80 ° for 2 hr, preferably under stirring. Solid
NaH2PO4 is added until a sample, diluted with water, reaches pH 7.0 to
7.5. Excess water is evaporated, and the residue is extracted with five
300-ml portions of hot chloroform. The pooled extracts are dried with
anhydrous NaESO4 and filtered through a glass-fiber paper (Whatman
GF/A). The solvent is recovered by distillation, and the remaining poly-
mer is dissolved in 2 liters of water. Washed DEAE-cellulose (Whatman
DE-52) at about pH 7 is added, and the mixture is mechanically stirred
until nearly all colored material is bound to the ion exchanger. This is
collected by suction filtration and is washed several times with water to
remove unreacted polymer. The dye-polymer is eluted by slowly passing
a 2 M KCI solution through the exchanger bed; residual dye remains
bound to the bed. The polymer is extracted from the KCl-containing
eluate with chloroform. The chloroform layer is dried with NaESO4 and
filtered, and the solvent is removed by evaporation. The product consists
of PEG covalently bound to the triazine dye in a 1 : 1 molar ratio.

2 G. Johansson, Mol. Cell. Biochem. 4, 169 (1974).

3 M.-R. Kula, G. Johansson, and A. F. Biickmann, Biochem. Soc. Trans. 7, 1 (1979).
4 G. Kopperschl~iger and G. Johansson, Anal. Biochem. 124, 117 (1982).
5 G. Kopperschl~iger, G. Lorenz, and E. Usbeck, J. Chromatogr. 259, 97 (1983).
6 I. N. Topchieva, Russ. Chem. Rev. (Engl. Transl.) 49, 260 (1980).
7 A. F. Btickmann, M. Morr, and G. Johansson, Makromol. Chem. 182, 1379 (1981).
358 OTHERSEPARATIONSYSTEMS [21]
Preparation of Two-Phase Systems
The two-phase systems are prepared from stock solutions of dextran,
20-30% (w/w) and PEG, 20-40% (w/w). Concentration of the dextran
stock solution is checked by polarimetry because of the varying water
content of solid dextran (2-5%). About 5 g of the dextran solution is
weighed and diluted to 25 ml with water. Optical rotation is measured
using sodium light at 20°. The concentration of dextran, Cdx, as a percent-
age (w/w) is calculated from the angle of rotation, a (in degrees), the
weight of stock solution, m (in grams), and the length of the light pathway,
(in dm), using Eq. (1).
a25
Cdx - m199----~100 (I)

Stock solutions of ligand-polymer are used at 10% (w/w) or less because

of their high viscosity. Calculated amounts of the polymer solutions are
weighed, and salt solutions, buffer, and water are added, leaving suffi-
cient volume for the protein solution. After mixing and attaining constant
temperature, the protein solution at the same temperature is added. The
phases are carefully mixed by inverting the vessel 10-20 times. For sys-
tems up to 100 ml in final volume, the total mixing time is 15-30 sec. The
two phases separate within 15 min by gravity or in less than 1 min by
centrifugation at low speed.

Analysis
Samples of known volumes (25-500 ~1) are withdrawn from each
phase for analysis. Enzymic activity can be determined directly, since the
phase-forming polymers do not usually interfere with the assay at low
concentrations. The presence of ligand-polymer may result in interfer-
ence by competition with substrate binding; increasing the concentration
of substrate should be tried if inhibition is noted. Several methods for
determination of proteins are strongly influenced by the polymers. The
method based on Coomassie Brilliant Blue G 8 has, however, been found
suitable. Because of the high viscosity, particularly of the lower phase,
great care must be taken in transferring samples of correct volume and
mixing them thoroughly with the assay solutions. Reproducible volumes
are obtained by filling and emptying the plastic tip of automatic pipettes
several times with the assay solution.

Factors Determining the Partition

The partition of a substance between the two phases is described by a
partition coefficient, K [Eq. (2)].
s M. M. Bradford, Anal. Biochem. 72, 248 (1976).
[21] AFFINITY PARTITIONING 359
concentration in upper phase
K = concentration in lower phase (2)
For a high degree of purification, the target enzyme should have a large K
value (>3) when the polymer-bound ligand is in the upper phase, whereas
other proteins should have low partition coefficients (K < 0.1). The fac-
tors determining the partition of proteins that are without affinity for the
ligand include the following.
1. Molecular weight of dextran; K increases with Mr.
2. Molecular weight of PEG; K decreases with increasing Mr.
3. Increasing polymer concentrations result in more extreme, usually
decreasing, K values.
4. Salt and pH. When a negative protein (pH > pI) is partitioned, its
K depends on the type of salt included in the system. K decreases in the
series phosphate > sulfate > fluoride > acetate > chloride > bromide >
iodide > perchlorate, as well as lithium > ammonium > sodium > potas-
sium. 9 When pH < pI the partition is influenced in the opposite manner.
The effect of salt increases with the net charge of the protein. The parti-
tion of proteins depends strongly on the concentration of salt in the range
up to 50 mM and is dependent on the amount of protein. K is nearly
constant at higher salt concentrations. 9
5. Negatively charged PEG, e.g., PEG sulfonate, 1° repels negatively
charged proteins toward the lower phase and attracts positive proteins
into the upper phase. Positively charged PEG, e.g., trimethylamino-PEG,
has the opposite effect. Charged PEG has a much stronger influence on
the K value than does salt, but the effect is counteracted by salts at low
concentration (5-25 mM). H
6. Temperature. In most instances, increasing temperature increases
the partition coefficient.

Factors Determining the Affinity Partitioning Effect

Affinity partitioning is measured as the increase in logarithmic parti-
tion coefficient, hlog K, of an enzyme caused by introducing the ligand-
polymer into the system [Eq. (3)].
Alog K = log K (with ligand) - log K (without ligand) (3)
An enzyme can easily be extracted if hlog K -> 2. When triazine dyes are
used as ligands bound to PEG, hlog K depends on the following factors.
9 G. Johansson, Acta Chem. Scand. Ser. B 28, 873 (1974).
10 G. Johansson, A. Hartman, and P.-/~. Albertsson, Eur. J. Biochem. 33, 379 (1973).
~t G. Johansson and A. Hartman, in "Proceedings of International Solvent Extraction Con-
ference, Lyon 1974" (G. V. Jeffreys, ed.), p. 927. The Society of Chemical Industry,
London, 1974.
360 OTHER SEPARATION SYSTEMS [2 1]

1. The amount of ligand-polymer in relation to ligand-binding enzyme.

With increasing concentration of ligand-polymer, Alog K approaches a
saturation value, Alog Kmax.
2. Increasing concentrations of dextran and PEG give larger Alog Kmax
(Fig. 1).
3. Salts decrease the affinity-partitioning effect. The reduction de-
pends on salt concentration, the nature of the salt, and the specific ligand
and enzyme. For phosphofructokinase extracted with Cibacron Blue
F3G-A, sodium phosphate (pH 7.0) and sodium formate (pH 7.0) have
only marginal influence on Alog K at 135 and 250 mM salt, respectively,
compared with low concentration (I0 mM) of phosphate buffer. 12Sodium
perchlorate and potassium phthalate (pH 7.0), on the other hand, have
greater influence and Alog K is reduced to half its initial value at 60 mM
salt. Similar effects have been observed in the case of glucose-6-phos-
phate dehydrogenase.13
4. Alog K depends strongly on pH. Working at low pH, the number of
extractable enzymes increases and Alog K values are larger (see the
table).
5. The Mr of the polymer can influence Alog Kmax ,12 but only limited
effects have been found.
6. It has been observed that Alog K can decrease markedly with in-
creasing temperature from about - 2 to 60°. Partition can often be carried
out at room temperature, since many enzymes are stabilized by the
polymers.
7. Among the triazine dyes there are large differences in the interac-
tion with a given enzyme. A suitable ligand can therefore be selected by
screening a number of dyes, 3,5 e.g., the Procion dyes Blue MX-R, Red
MX-2B, and Yellow MX-4R.
To find optimal conditions for purification by affinity partitioning, each
of the described parameters should be systematically varied and analyzed
by following the partition of both the desired enzyme and the mass of
protein.

Strategy of Selecting a Two-Phase System

1. A 4-g s y s t e m , c o n t a i n i n g 7 % d e x t r a n (Mr = 500,000), 5 % P E G

(Mr = 6000-7500), and 25 mM buffer (sodium phosphate at pH 7.0, or Tris

12G. Johansson, G. Kopperschl~iger, and P.-/~. Albertsson, Eur. J. Biochem. 131, 589
(1983).
13K. H. Kroner, A. Cordes, A. Schelper, M. Morr, A. F. B0ckmann, and M.-R. Kula, in
"Aftinity Chromatography and Related Techniques" (T. C. J. Gribnan, J. Visser, and
R. J. F. Nivard, eds.), p. 491. Elsevier, Amsterdam, 1982.
[21] AFFINITY PARTITIONING 361

I I I
2 3 4 5 6 7 8 9 lb
% PEG

% Dextran
FIG. 1. Partition of phosphofructokinase when an extract of bakers' yeast is included in
systems containing increasing concentrations of dextran (Mr = 70,000) and polyethylene
glycol (PEG) (Mr = 38,000), with (0) and without (©) Cibacron Blue F3G-A PEG (Mr =
6500), 6.25% of total PEG. The system also contains 25 mM sodium phosphate buffer, 5 mM
2-mercaptoethanol, and 0.25 mM EDTA, pH 7.0; temperature, 0°.

acetate at pH 7.5) is prepared by mixing 1.4 g of 20% (w/w) dextran, 0.5 g

of 40% (w/w) PEG, 0.75 ml of 100 mM buffer, and 0.35 ml of water in a
graduated centrifuge tube. The mixture is maintained at a constant tem-
perature (0 or 25°), and 1 ml of protein solution, containing 25 mM buffer,
is added. After mixing for 15 sec, the tube is centrifuged at the same
temperature for 2 min at I000 g The volumes of upper and lower phases
are measured, and samples are removed for analysis. Before determining

EFFECT OF pH ON ALOG K OF SOME

GLYCOLYTIC ENZYMES FROM BAKERS' YEASTa

ALog K

pH Enzyme I b Enzyme 2b Enzyme 3b

4 2.6 2.3 1.1

5 2.5 2.6 0.5
6 1.4 0.9 0.2
7 0.7 0.1 0
8 0.3 0 0

a Two-phase system: 7.5% dextran (Mr =

70,000), 5% PEG (Mr = 38,000), 12.5 mM so-
dium phosphate at 0°. In the ligand-containing
system, one-fifth of the PEG was replaced
with Cibacron Blue F3G-A PEG (Mr = 6500).
b Enzymes 1, 2, and 3 are, respectively, glyc-
eraldehydephosphate dehydrogenase, 3-phos-
phoglycerate kinase, and enolase.
362 OTHERSEPARATIONSYSTEMS [21]

protein the phases are diluted fivefold or more with water. In this two-
phase system proteins are generally concentrated in the lower phase.
2. Extraction curves are prepared by replacing increasing amounts of
PEG with ligand-PEG, e.g., 1/500, 1/250, 1/100, 1/50, 1/25, and 1/10. The
curve of Alog K versus ligand-PEG in total PEG provides a guide for the
amount of ligand-PEG necessary for maximal extraction effect. Straight
lines are obtained when inversed plots (1/Alog K versus 1/{concentration
of ligand}) are used, 12 which allows the determination of a saturation
value, a Vmax, from only a few measurements. A selective ligand is sought
by testing several triazine dyes, e.g., the several Procion dyes produced
by ICI and obtainable from Sigma Chemical Co. (St. Louis, MO) or Serva
Feinbiochemia (Heidelberg).
3. The Alog Kmax value can be further increased by using higher con-
centrations of dextran and PEG, as shown in Fig. 1, and retaining a
constant ligand-PEG:PEG ratio. Since the partition coefficient of the
enzyme in the presence of ligand-PEG usually goes through a maximum
when the polymer concentration increases, this curve is used to choose a
polymer composition that provides a reasonable recovery of the enzyme
in the upper phase.
4. The selectivity of the extraction or Alog Kmax may additionally be
improved by changing the pH or by introducing salt to the system chosen
under 3. Decreasing pH results in less specific extraction but in larger
Alog K values; lower salt concentrations have the same influence.
5. The percentage of enzyme extracted into the upper phase can be
chosen freely within practical limits by adjusting the ratio between the
volumes of the two phases. The volume ratio, V (volume of upper
phase : volume of lower phase), that yields the percentage, P, of the en-
zyme in the upper phase, is calculated by Eq. 4, where K is the partition
coefficient of the enzyme.
V = P/[(100 - P)K] (4)
A study of this type has been made with phosphofructokinase from
bakers' yeast, ~2and affinity partitioning has been applied to the isolation
of this enzyme. 4

Repeated Extractions
The effectiveness of purification can be enhanced by using repeated
extractions that might include the following steps.
1. Preliminary extraction with an upper phase not containing ligand-
PEG for removal of material with relatively high K values.
[21] AFFINITY PARTITIONING 363

2. Several extractions with ligand-containing upper phase to improve

the recovery of enzyme from a lower phase.
3. Preliminary extraction with an upper phase containing a ligand that
can bind several enzymes but n o t the target enzyme.
4. Repeated washing of ligand-containing upper phase, after the affin-
ity-partitioning step, with a fresh lower phase to remove coextracted
proteins that are less strongly extracted by the ligand-PEG.
In several instances Alog K has been observed to increase by as much
as 0.5 unit when the bulk of the protein is removed.

Preparative Extractions
Preparative affinity partitioning with systems selected by the above
procedure can easily be carried out on a large scale. Two-phase systems
of up to several liters can be mixed in separatory funnels; the phases will
settle within 30-60 min. Very rapid extraction (-<5 min) is achieved by
carrying out the partition in large centrifuge bottles that are used in pre-
parative centrifuges. The phases are collected by siphoning. Larger quan-
tities of partitioning polymers may be mixed with a motor-driven stirrer of
the paddle type at low speed in a vessel with an outlet in its lower part;
such separation may, however, require several hours. For rapid separa-
tion of very large volumes of the two phases, standard centrifugal separa-
tors have been used. 14

Removal of Polymers
Enzyme can be separated from polymer by dissolving a mixture of
solid potassium phosphates (yielding a pH of 7 to 9) in the recovered
upper phase; 15 g or more are required per 100 g of solution. The result is
another two-phase system with PEG and ligand-PEG in the viscous upper
phase and the enzyme in the lower salt-containing phase.
Another possibility is to dilute the original upper phase with several
volumes of buffer, e.g., 5 mM sodium phosphate at pH 7 to 9, and then
bind the enzyme to a bed of DEAE-cellulose (or CM-cellulose for more
basic proteins). Nonbound protein and PEG are washed away, and the
enzyme is eluted with a salt gradient. The bound ligand-PEG may be
eluted by increasing the concentrations of salt.

14 K. H. Kroner, H. Hustedt, S. Granada, and M.-R. Kula, Biotechnol. Bioeng. 20, 1967
(1978).
364 O T H E R S E P A R A T I O N SYSTEMS [22]

Comments

The amount of protein that can be included in the system is limited by

the solubility in the two phases, but 50 g of protein per kilogram (final
weight) of partition system can often be used. Since a solid matrix is not
involved, nonspecific interactions and irreversible binding of "multiple
attachment" type are eliminated. The recovered ligand-PEG may be re-
used, resulting in a process economical for large-scale operations.

[22] A f f i n i t y P r e c i p i t a t i o n o f D e h y d r o g e n a s e s
By PER-OLOF LARSSON, SUSANNE FLYGARE, and KLAUS MOSBACH

Affinity precipitation of enzymes is a novel technique closely related

to affinity chromatography and immunoprecipitation. The technique is
applicable to oligomeric enzymes and may be used for purification, L2 for
analysis, ~ and possibly for studies of molecular architecture and enzyme
subunit arrangement. ~,2
The first step in enzyme precipitation is the mixing of the enzyme with
a bifunctional ligand composed of two ligand entities connected by a
spacer. If the spacer is sufficiently long to bridge the distance between the
respective binding sites of two enzyme molecules, and if the bioaffinity
between the ligand and the enzyme is sufficiently strong, the two enzyme
molecules will be linked to each other. If, in addition, the enzyme is
oligomeric, more than one Bis-ligand can bind to each enzyme molecule,
and it is conceivable that a network of enzymes and Bis-ligands will form.
When this network reaches a sufficient size, it will no longer be retained in
solution and will consequently precipitate.
Clearly a critical aspect of affinity precipitation is the length of the
spacer connecting the two ligand entities. If the ligand binding is too short
it may not be able to span the distance between two enzymes, a situation
that might occur if the sites are deeply hidden beneath the contour line of
the enzyme. On the other hand, a spacer that is too long may lead to
intramolecular cross-linking, a situation bearing interest on questions
concerning subunit arrangement, but giving no precipitation. For pre-
cipitation of dehydrogenases a Bis-NAD with a 17-/~ spacer length,
N2, N2' adipodihydrazido-bis(N6-carbonylmethyl-NAD),1 has functioned
i P.-O. Larsson and K. Mosbach, FEBS Lett. 98, 333 (1979).
2 S. Flygare, T. Griffin, P.-O. Larsson, and K. Mosbach, Anal. Biochem. 733, 409 (1983).

Copyright © 1984by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
364 O T H E R S E P A R A T I O N SYSTEMS [22]

Comments

The amount of protein that can be included in the system is limited by

the solubility in the two phases, but 50 g of protein per kilogram (final
weight) of partition system can often be used. Since a solid matrix is not
involved, nonspecific interactions and irreversible binding of "multiple
attachment" type are eliminated. The recovered ligand-PEG may be re-
used, resulting in a process economical for large-scale operations.

[22] A f f i n i t y P r e c i p i t a t i o n o f D e h y d r o g e n a s e s
By PER-OLOF LARSSON, SUSANNE FLYGARE, and KLAUS MOSBACH

Affinity precipitation of enzymes is a novel technique closely related

to affinity chromatography and immunoprecipitation. The technique is
applicable to oligomeric enzymes and may be used for purification, L2 for
analysis, ~ and possibly for studies of molecular architecture and enzyme
subunit arrangement. ~,2
The first step in enzyme precipitation is the mixing of the enzyme with
a bifunctional ligand composed of two ligand entities connected by a
spacer. If the spacer is sufficiently long to bridge the distance between the
respective binding sites of two enzyme molecules, and if the bioaffinity
between the ligand and the enzyme is sufficiently strong, the two enzyme
molecules will be linked to each other. If, in addition, the enzyme is
oligomeric, more than one Bis-ligand can bind to each enzyme molecule,
and it is conceivable that a network of enzymes and Bis-ligands will form.
When this network reaches a sufficient size, it will no longer be retained in
solution and will consequently precipitate.
Clearly a critical aspect of affinity precipitation is the length of the
spacer connecting the two ligand entities. If the ligand binding is too short
it may not be able to span the distance between two enzymes, a situation
that might occur if the sites are deeply hidden beneath the contour line of
the enzyme. On the other hand, a spacer that is too long may lead to
intramolecular cross-linking, a situation bearing interest on questions
concerning subunit arrangement, but giving no precipitation. For pre-
cipitation of dehydrogenases a Bis-NAD with a 17-/~ spacer length,
N2, N2' adipodihydrazido-bis(N6-carbonylmethyl-NAD),1 has functioned
i P.-O. Larsson and K. Mosbach, FEBS Lett. 98, 333 (1979).
2 S. Flygare, T. Griffin, P.-O. Larsson, and K. Mosbach, Anal. Biochem. 733, 409 (1983).

Copyright © 1984by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
[22] AFFINITY PRECIPITATION 365

If --

' I ' , ' I I h - -

1 2 3 10
NAD EQUNB.LENTS/ SUBUNIT
FIG. 1. Precipitation yield as a function of the ratio NAD equivalents per enzyme sub-
units after 16 hr. © ©, Lactate dehydrogenase; • O, glutamate dehydrogenase.

well. Two other spacer lengths, 7 and 32/~, have also been used with
some systems.
In order to improve the selectivity of the affinity precipitation process
and to increase the effective binding strength, ternary complex formation
may be used. For example, Bis-NAD together with pyruvate or oxalate
will form strong ternary complexes with lactate dehydrogenase, causing it
to precipitate. In fact, in the absence of oxalate or pyruvate precipitation
does not occur, an observation that is readily understood when the rather
weak binary enzyme-coenzyme complex is considered (lactate dehydro-
genase Kd for NAD = 3 x 10-4 M). 3
A major point in affinity precipitation is the ratio between ligand entit-
ies and enzyme subunits. Maximum precipitation occurs at a ratio of
unity. At lower ratios not all enzymes will be engaged in complexes, and
at higher ratios only one end of the Bis-ligand may become attached to the
enzyme. Moderate deviation from unity will be of less importance if the
enzyme is an oligomer with a high subunit number. A comparison has
been made between lactate dehydrogenase and glutamate dehydrogenase
with respect to precipitation yield as a function of the ligand : subunit ratio
(Fig. 1). Glutamate dehydrogenase is a hexamer, known to aggregate
spontaneously when present in high concentrations, whereas lactate de-
hydrogenase is a tetramer without such properties. The precipitation yield
for lactate dehydrogenase is satisfactory for ratios between 0.8 and 2.5
and almost quantitative at a ratio of 1.3. For glutamate dehydrogenase, on
the other hand, a very wide range of ratios gives almost quantitative
3 R. A. Stinson and J. J. Holbrook, Biochem. J. 131, 719 (1973).
366 OTHER SEPARATION SYSTEMS [22]

precipitation (0.3-10). The observation that the precipitation yield is very

good, even at a ratio as low as 0.16, suggests that glutamate dehydro-
genase under the prevailing conditions is an oligomer with at least 12
subunits.
In the following sections, procedures are given for affinity precipita-
tion of lactate dehydrogenase from a crude extract and for glutamate
dehydrogenase in a model experiment. Affinity precipitation within a gel
is also described.

The Bis-Functional NAD-Derivative, Bis-NAD

The Bis-NAD derivative used below, N2, Nz'-adipodihydrazido-bis-(N 6-
carbonylmethyl--NAD), has a 17-/~ spacer joining the two NAD entities
to each other by their N 6 amino groups [NADmCH2CONHNHCO
(CH2)4CONHNHCOCHz--NAD]. The reagent is commercially available
(Sigma Chemical Co., St. Louis, MO) or can be prepared as has been
describedJ

Affinity Precipitation of Lactate Dehydrogenase 2

A homogenate of ox heart, subjected to centrifugation for 30 rain at
20,000 g and dialyzed overnight at 4° against 0.05 M sodium phosphate at
pH 7.5, serves as the source of lactate dehydrogenase for illustrating the
procedure.

Optimum Bis-NAD Concentration (Pilot Affinity Precipitation)

To maximize yield in the affinity precipitation step, optimum Bis-NAD
concentration is first determined in a small-scale affinity precipitation
experiment. To a sample of the crude extract (2 ml), 0.25 ml of 0.1 M
sodium oxalate and 0.25 ml of Bis-NAD are added. The concentrations of
the Bis-NAD are chosen so that the nominal ratio between NAD equiva-
lents and lactate dehydrogenase subunits are 1 : 4, 1 : 2, 1 : 1, 2 : 1, 3 : 1,
and 4: 1; the nominal enzyme concentration is estimated from activity
measurements. Precipitation occurs almost immediately; after 30 rain the
precipitate is collected by centrifugation at 10,000 g for 10 rain. Only the
supernatant fluid need be analyzed for enzyme activity with the assump-
tion that the tube with the lowest residual activity represents the best ratio
for affinity precipitation. These conditions are then applied on a prepara-
tive scale.
[22] AFFINITY PRECIPITATION 367

Preparative Affinity Precipitation

When 200 ml of extract was treated at 4° with 25 ml of 0.1 M sodium
oxalate and 20 ml of 130/~M Bis-NAD, as determined by pilot affinity
precipitation, a precipitate formed that was collected by centrifugation
after 20 hr. The precipitate was dissolved in 10 ml of sodium phosphate at
pH 7.5 containing 0.6 mM NADH and dialyzed against the phosphate
buffer. This solution represents a 50-fold increase in specific activity of
lactate dehydrogenase. The recovery is about 90%, and the purity, as
judged by electrophoresis, is greater than 95%.

Comments
Oxalate is the preferred third component when affinity-precipitating
lactate dehydrogenase, although pyruvate can also be used. 2 A concentra-
tion of 10 mM oxalate was chosen since it is sufficient to promote com-
plete precipitation under the conditions used. When the enzyme concen-
tration is low, a higher oxalate concentration is recommended to further
enhance complex formation. However, a high oxalate concentration has
the drawback of interfering with activity measurements.
It is advisable to carry out a pilot affinity precipitation prior to the
preparative affinity precipitation. Activity measurements of the crude ex-
tract usually give an underestimation of the lactate dehydrogenase con-
centration by a factor of 1.5-3, possible owing to the presence of inhibi-
tory substances in crude extracts. Addition of Bis-NAD, based only on
determination of enzyme activity, would give a low yield of precipi-
tate as is evident from the data in Fig. 1. Although the pilot trial may seem
a cumbersome extra step, it can be carried out rapidly. The precipitation
yield for the best ratio is, at near maximum, 85% after only 20 min,
justifying the 30 min allowed for the procedure. The actual maximum
yield is reached after 2 hr. Thus, the 20 hr used for precipitation in the
preparative procedure may be shortened considerably if convenient.

M o d e l Affinity Precipitation of G l u t a m a t e D e h y d r o g e n a s e 2

Procedure
To glutamate dehydrogenase4 (1.5 ml, 1.8 mg/ml) in 0.05 M sodium
phosphate at pH 7.5, is added 0.25 ml of 0.8 M glutarate followed by 0.25
ml of 0.12 mM Bis-NAD. Upon gentle mixing, the solution rapidly be-

4 Bovine liver, type I, Sigma, St Louis, MO.

368 OTHER SEPARATION SYSTEMS [22]

® ®

® ®
FxG. 2. Affinity precipitation in agarose gel. The center well contained 30/zg beef heart
lactate dehydrogenase. For the peripheral wells 1 to 3, 1.5 nmol bis-NAD; well 5, 3 nmol
NAD; and wells 4 and 6, buffer only. Reproduced, with permission, from Larsson and
Mosbach.

comes opaque; a precipitate is formed after I0 min; it is allowed to settle

at 4 ° overnight. The precipitate is dissolved in 0.5 ml of 10 mM NADH.

Comments
Affinity precipitation of glutamate dehydrogenase is not particularly
dependent on a correct ligand: subunit ratio as is the case for lactate
dehydrogenase (Fig. 1). Affinity precipitates formed near a ratio of unity
are extensively cross-linked, judged from the observation that a rather
high N A D H concentration is needed to achieve a reasonably fast dissolu-
tion. Affinity precipitation has been carried out at lower glutamate dehy-
drogenase concentrations (0.1 mg/ml), also with a precipitation yield of
over 90%.

Affinity Precipitation in Gels

Procedure
Agarose gels containing 0.8% agarose, 0.3 M sodium pyruvate, and
0.05 M sodium phosphate at pH 7.5 are cast on microscope glass slides.
Wells are punched out with a die as shown in Fig. 2. The center well is
filled with 15/zl of lactate dehydrogenase solution (5-50/xg of enzyme).
The peripheral wells are filled with 7 /zl of Bis-NAD solution (0.5-10
nmol), with NAD solution (0.5-10 nmol), or with buffer. The slides are
placed in a moist chamber at room temperature and allowed to develop
[22] AFFINITY PRECIPITATION 369

for 1.5-30 hr, typically 16 hr. Bands of precipitated proteins may be

developed subsequently with Amido Black. 5

Comments
Affinity precipitation of lactate dehydrogenase in agarose gels closely
resembles the corresponding immunoprecipitation technique. 5 Pyruvate
has been used as the third component to enhance complex formation, and
oxalate would probably work even better, judging from the experience
with affinity precipitation from solution. 2 Since bands of precipitate are
directly visible only when a high concentration of enzyme is used, a
protein stain must usually be applied. Figure 2 shows that precipitation
occurs only where lactate dehydrogenase diffuses toward Bis-NAD.

Other Systems
Yeast alcohol dehydrogenase has also been affinity precipitated. 2 For
precipitation to occur (in the presence of pyrazole) 0.2 M sodium chloride
must also be added. The precipitation was shown not to be a salting-out
effect. However, the process was slow and of low yield.
Clearly, precipitation may be difficult to achieve also in other systems.
An alternative approach would then be to apply the nonprecipitating mix-
ture to a gel permeation column (e.g., a Sephacryl S-400 column) and
subsequently collect the desired enzyme in the high-molecular-weight
fractions.
Similar to the affinity precipitation described here is a report on the
synthesis of N,N'-bis-3(dihydroxYboronylbenzene)adipamide and the
demonstration of its ability to agglutinate red blood cells by complexing
with cell surface carbohydrates. 6

5 j. Clausen, in "Laboratory Techniques in Biochemistry and Molecular Biology" (T. S.

Work and E. Work, eds.), Vol. 1, p. 397. North-Holland Publ., Amsterdam, 1969.
6 T. J. Burnett, H. C. Pcebles, and J. H. Hageman, Biochem. Biophys. Res. Commun. 96,
157 (1980).
370 OTHER SEPARATION SYSTEMS [23]

[23] P r o t e i n C r y s t a l l i z a t i o n : T h e G r o w t h o f L a r g e - S c a l e
Single Crystals
By GARY L. GILLILAND and DAVID R. DAVIES

The establishment of the conditions necessary for growing large single

crystals of a protein has always seemed to the uninitiated very much like
black magic. Since these conditions can only be established empirically,
and since there are many variables, each of which can be changed by
small increments, it would appear at first sight that the number of possible
combinations is close to infinite. We suggest that an examination of the
factors (pH, temperature, precipitant) that have been used to produce
large single protein crystals is useful in restricting the number of experi-
ments necessary to find suitable growth conditions. It is also suggested
that an analysis of the procedures that have been used successfully in
crystallizing proteins indicates that certain methods are more practical in
the laboratory than others. We shall outline the general procedures that
have been used in the search for large single crystals of a size (>0.2 mm)
suitable for use in an X-ray diffraction investigation and shall describe in
detail the three with the most general utility for screening crystal growth
conditions.

Factors Important in Crystallization

The growth of crystals can be separated into the distinct phases of
nucleation, postnucleation growth, and cessation of growth.1 Nucleation
and postnucleation growth are driven by supersaturation of the protein in
solution. Crystal nuclei will form only at a critical point of supersaturation
in an appropriate environment, e.g., ionic strength, pH, temperature.
After nucleation a crystal will continue to grow as long as a state of
supersaturation exists. Growth stops when the protein environment is
altered or when errors in the crystal lattice become too great to permit
continued growth.
The usual procedure for growing large single crystals of a protein is to
approach the point of supersaturation slowly in order to reduce the num-
ber of nucleation sites. This can be achieved by gradually altering the
ionic strength, pH, temperature, or dielectric constant of a protein solu-

I Z. Kam, H. B. Shore, and G. Feher, J. Mol. Biol. 123, 539 (1978).

METHODS IN ENZYMOLOGY,VOL. 104 ISBN 0-12-182004-1

[23] PROTEIN CRYSTALLIZATION 371

tion. The many physical methods that have been employed for growing
protein crystals have been reviewed by McPherson z,3 and Blundell and
Johnson. 4

Protein Concentration
Proteins have been crystallized from solutions containing from one to
several hundred milligrams per milliliter. However, for crystallization
trials we would recommend a concentration of 10-20 mg/ml if possible.
Some proteins that are relatively insoluble under the initial conditions of
concentration may be rendered more soluble by changing the pH or ionic
strength.

Preliminary Analysis of Precipitation Conditions

In order to obtain some indication of the quantities of precipitant that
should be used, a preliminary examination should be made of the solubil-
ity of the protein as a function of precipitant concentration. This can be
carried out in a depression slide with small quantities of protein solution
(-10/.d) by adding the precipitant in small aliquots, sealing the depression
with a coverslip, and observing the droplet for the next 15 min or so with a
low-power microscope for signs of precipitation. The quantity of precipi-
tant necessary to effect precipitation should be noted, and this concentra-
tion gradually be approached in the subsequent crystallization experi-
ments.
This preliminary examination should be carried out over a broad range
of pH, since there can be a large variation in solubility at different pH
values. In many cases, a protein is least soluble at its isoelectric point, so
that a prior investigation by techniques such as isoelectric focusing can be
used to suggest conditions for certain types of crystallization (in particu-
lar, low ionic strength). For many enzymes it is desirable to attempt
crystallization at the pH at which the enzyme is fully active.
Temperature is another important factor to consider. In general, labo-
ratories are not equipped to allow crystallization trials to be carried out
over a large range of temperatures, so that this work is usually done at
room temperature or in a cold room (4-6°). The solubilities of individual
proteins vary markedly with temperature. 4,5

2 A. McPherson, Jr., Methods Biochem. Anal. 23, 249 (1976).

3 A. McPherson, "Preparation and Analysis of Protein Crystals." Wiley, New York, 1982.
4 T. L. Blundell and L. Johnson, "Protein Crystallography." Academic Press, New York,
1976.
5 A. A. Green, J. Biol. Chem. 93, 495 (1931).
372 OTHER SEPARATION SYSTEMS [23]

Choice of Precipitant
Examination of the table indicates clearly that certain precipitants
have been much more successful than others in inducing crystallization.
Whether this reflects a genuine preferred ability of these materials to
produce crystals, or merely reflects the fashions of crystallizers, is not
always clear, but it would appear that the aspiring crystallizer would be
well advised to begin with those precipitants that have proved to be the
most successful. In particular, ammonium sulfate, polyethylene glycol
(low ionic strength), and 2-methyl-2,4-pentanediol seem to be the precipi-
tants of choice for an initial investigation.

General Methods

Direct Addition of Precipitants

Direct addition or batch procedures involve the use of salts, acids,
bases, water, metal ions, or other effector molecules of the protein solu-
tion to induce crystallization. After the addition of the precipitant, if the
solution is left undisturbed, nucleation and crystal growth will often oc-
cur. Nearly all of the early crystallizations were carried out by this
method, but when used for exploring crystallization conditions, it does
require a great deal of protein. If small volumes are used to conserve a
precious supply of protein, it becomes difficult to add the precise amounts
that are necessary for reproducibility of crystallization.
Nevertheless, when the crystallization conditions are known, and es-
pecially when they involve simply adjusting the pH or ionic strength of
the protein solution, the batch procedure can often be reproducibly used
on a microscale to produce large crystals. 6,7

Dialysis
Traditionally, dialysis through a semipermeable membrane has been a
favored technique. Dialysis is particularly applicable for proteins that
crystallize at low ionic strength. The method is simple to use, and the
point of supersaturation can be approached gradually. However, a dialy-
sis bag requires relatively large amounts of material, and consequently
this method has now been superseded by the Zeppezauers,9 procedure of
6 M. Dobler, S. D. Dover, K. Laves, A. Binder, and H. Zuber, J. Mol. Biol. 71, 785 (1972).
7 R. R. Bott, M. A. Navia, and J. L. Smith, J. Biol. Chem. 257, 9883 (1982).
s M. Zeppezauer, H. Eklund, and E. S. Zeppezauer, Arch. Biochem. Biophys. 126, 564
(1968).
9 M. Zeppezauer, this series, Vol. 22, p. 253.
[23] PROTEIN CRYSTALLIZATION 373

P R E C I P I T A T I N G A G E N T S U S E D T O I N D U C E CRYSTALLIZATIONa

Crystal Crystal Overall

Precipitating agent forms/subgroup forms/group ranking

Salts
Ammonium acetate -- 1 20
Ammonium citrate -- 2 19
Ammonium nitrate -- 1 20
Ammonium sulfate 296 361 1
+ Acetate b 2 -- --
+ Acetone 1 -- --
+ Ammonium chloride 1 w _
+ Ammonium phosphate 1 -- --
+ Cesium chloride 12 -- --
+ Citrate ~ 4 -- --
+ Dimethylformamide 1 -- --
+ Dimethyl sulfoxide 1 -- --
+ Dioxane 3 -- --
+ Isocitrate b 1 -- --
+ Lithium nitrate 1 u _
+ Phosphate b 3 ~ --
+ Potassium phosphate 6 ~ --
+ Sodium chloride 13 ~ --
+ Sodium citrate 1 -- --
+ Sodium formate 1 -- --
+ Sodium phosphate 3 -- --
+ Cesium chloride 1 -- --
+ Sodium potassium phosphate 6 -- --
+ Tris b 3 -- --
Citrate b -- 4 17
Cadmium sulfate -- 4 17
Lithium chloride -- 1 20
Lithium sulfate -- 8 14
Magnesium chloride -- i 20
Magnesium sulfate -- 9 13
Phosphate b -- 15 9
Potassium borate -- 1 20
Potassium nitrate -- 1 20
Potassium phosphate -- 11 11
Potassium sodium phosphate 15 16 8
+ Ammonium sulfate 1 -- --
Potassium tartrate -- l 20
Sodium chloride 23 25 7
+ Sodium citrate 1 -- --
+ Sodium potassium phosphate 1 -- --
Sodium citrate -- 10 12
Sodium iodide -- 1 20
Sodium nitrate 4 5 16
+ Acetone 1 -- --

(contmued)
374 O T H E R S E P A R A T I O N SYSTEMS [23]

TABLE (continued)

Crystal Crystal Overall

Precipitating agent forms/subgroup forms/group ranking

Sodium phosphate -- 4 17
Sodium sulfate 2 3 18
+ Acetate b 1 -- --

Organic precipitating agents

Acetone 3 4 17
+ Salt 1 --
Dioxane -- 7 15
Dimethylsulfoxide -- 2 19
+ Ammonium sulfate 1 -- --
+ Calcium acetate 1 -- --
tert-Butanol -- 7 15
Ethanol 32 33 5
+ Ammonium sulfate 1 -- --
Glucose -- 1 20
Glutamic acid -- 1 20
Isopropanol -- 14 l0
Methanol -- 7 15
2-Methyl-2,4-pentanediol 70 77 4
+ Dimethyl sulfoxide 1 -- --
+ Magnesium acetate 1 --
+ Magnesium chloride 2 -- --
+ Methanol 1 -- --
+ Sodium chloride 2 --
1,3-Propanediol -- 1 20
n-Propanol -- 5 16

Polyethylene glycol
P o l y e t h y l e n e g l y c o l ( u n s p e c i f i e d Mr) 10 11 11
+ Ammonium acetate 1 -- --
P o l y e t h y l e n e g l y c o l 4 0 0 ( a v g Mr) - - 5 16
P o l y e t h y l e n e g l y c o l 1000 ( a v g Mr) -- I 20
P o l y e t h y l e n e g l y c o l 2 0 0 0 ( a v g Mr) -- 3 18
P o l y e t h y l e n e g l y c o l 4 0 0 0 ( a v g Mr) 29 30 6
+ Ammonium sulfate 1 -- --
P o l y e t h y l e n e g l y c o l 6 0 0 0 ( o r 8000)
( a v g Mr) 69 78 3
+ Calcium acetate 1 -- --
+ Lithium chloride 1 -- --
+ Sodium chloride 5 -- --
+ Potassium chloride 1 --
+ Phosphate b 1 -- --
P o l y e t h y l e n e g l y c o l 2 0 , 0 0 0 ( a v g Mr) -- 3 18

Low ionic strength c -- 79 2

[23] PROTEIN CRYSTALLIZATION 375

microdialysis into a capillary. Although not simple, the Z e p p e z a u e r

m e t h o d has several advantages. It requires very little material, and pre-
cipitated protein can be dissolved readily for another trial. H o w e v e r , the
dialysis capillaries are not e a s y to assemble, and, without practice, it is
difficult to load the protein solution into the capillary without introducing
air bubbles. Crystals, w h e n formed in the capillary, are difficult to ob-
serve and s o m e t i m e s difficult to extract. Although there are several modi-
fications o f this p r o c e d u r e that use different types of dialysis cells to try to
avoid the p r o b l e m s mentioned, 4,a-~/the Z e p p e z a u e r procedure remains
the m o s t popular. A description o f the method is presented in detail.

Vapor Diffusion
V a p o r diffusion is p e r h a p s the m o s t popular of the methods currently
in use. It is effective in concentrating a protein solution containing a
nonvolatile precipitant, or for diffusion of a volatile precipitant into the
protein solution.
In the first case, the protein solution is p r e p a r e d in a drop containing
an a p p r o p r i a t e a m o u n t of precipitant, e.g., a m m o n i u m sulfate or polyeth-
ylene glycol, with the intent of maintaining the protein below supersatura-
tion. The drop is then equilibrated in a sealed c h a m b e r with a larger
volume o f a solution o f the precipitant at a concentration that is just at, or
slightly above, the concentration required for supersaturation. The pro-
tein solution then gradually a p p r o a c h e s the supersaturation point as w a t e r
e v a p o r a t e s f r o m the drop and disperses into the reservoir.

10B. H. Weber and P. E. Goodkin, Arch. Biochem. Biophys. 141, 489 (1970).
it V. Lagerkvist, L. Reyno, O. Lindquist, and E. Andersson, J. Biol. Chem. 247, 3897
(1972).
t2 I. Rayment, J. Appl. Crystallogr. 14, 153 (1981).

a This table was compiled by one of us (G. L. G.) from a data base consisting of data
extracted from published crystallization procedures in which production of crystals
suitable for X-ray diffraction studies was described. Included in this table are data for
more than 500 macromolecules with more than 800 different crystal forms. The precipi-
tating agents included in this table have been divided into groups according to the
predominant additive. The subgroups define those cases in which additional substances
have been added to induce crystallization. The subgroup elements have been tabulated
if the concentration of additional salts was >0.2 M or if the concentration of organic
agents was >5% (v/v).
b Unspecified counterion.
¢ Included in this category are many proteins whose crystallization has proved to be very
sensitive to the presence of small amounts of divalent metal ions, ligands, products, or
substrates or other effector molecules.
376 OTHER SEPARATION SYSTEMS [23]

In the second case, in which a volatile precipitant is used, a similar

procedure can be followed, although it is not necessary to add precipitant
directly to the drop. Equilibration will bring the protein solution gradually
to supersaturation through a transfer of precipitant from the larger volume
to the drop.
There are two principal techniques that utilize vapor diffusion. They
are the "sandwich box" technique 2"4,13-21 and the "hanging drop"
method. 3,22 Despite a basic similarity, each has advantages for certain
types of crystallization. In the "sandwich box" procedure, the crystalli-
zation drop is placed in a well of a depression slide; several slides may be
enclosed in a sealed box with a larger volume of appropriate precipitant
solution. The advantages of the procedure include the ability to use large
drops (> 100/.d), ease of preparation, and ease of observation. However,
only a limited number of samples can be prepared per box, all must be
equilibrated against the same reservoir solution, and, if one sample has to
be removed from the box, the others will be disturbed.
In the "hanging drop" method, a drop of protein solution is suspended
on the lower face of a microscope slide coverslip placed above equilibrat-
ing solution in the well of a 24-well tissue culture plate. Since each well
can contain a different equilibrating solution, the method can be utilized
to explore a variety of conditions in a compact manner. The drops can be
easily observed, and the plates stack together in compact form for con-
venient storage. There are disadvantages; the drop size is limited to ap-
proximately 20/zl and larger drops will fall off, and it is difficult to use the
procedure with many organic precipitants because the droplets spread out
owing to low surface tension.

Crystall&ation by Concentration
Gradual concentration of a protein solution to produce crystals is a
method that has been used infrequently in comparison with those meth-
13 B. F. C. Clark, B. P. Doctor, K. C. Holmes, A. Klug, K. A. Marcker, S. J. Morris, and
H. H. Paradies, Nature (London) 219, 1222 (1968).
14 S. H. Kim and A. Rich, Science 162, 1381 (1968).
15 A. Hampel, M. Labanauskas, P. G. Connors, L. Kirkegard, V. L. Rajbhandary, P. B.
Sigler, and R. M. Bock, Science 161, 1384 (1968).
16 F. Cramer, F. Van Den Haar, W. Saenger, and E. Schlimme, Angew. Chem. 80, 969
(1968).
17j. R. Fresco, R. D. Blake, and R. Langridge, Nature (London) 220, 5174 (1968).
18 C. D. Johnson, K. Adolph, J. J. Rosa, M. D. Hall, and P. B. Sigler, Nature (London) 226,
1246 (1970).
19 D. R. Davies and B. P. Doctor, "Procedures in Nucleic Acid Research," Vol. 2. Harper,
New York, 1971.
20 D. R. Davies and D. M. Segal, this series, Vol. 22, p. 266.
[23l PROTEIN CRYSTALLIZATION 377

ods that reduce the solubility of the protein by the addition of precipi-
tants. Nevertheless, it has been used, and in its simplest form allows
evaporation to increase protein concentration. 23 Other techniques of con-
centration include a collodion bag apparatus 24 and dialysis against poly-
ethylene glycol 20,000 or Lyphogel. 2,3

Crystallization by Varying the Temperature

Another infrequently used procedure lowers the solubility by gradu-
ally raising or lowering the temperature. Baker and Dodson 25 described a
method for crystallizing insulin by raising the temperature of the protein
solution to 55 ° and then allowing it to cool slowly in a Dewar flask. The
addition of insulation about the flask can further slow down the cooling
process .4
Jakoby ~6 has presented a general method of growing small crystals
(about 1-10/zm) that depends on the differential solubility with respect to
temperature of protein solutions in the presence of ammonium sulfate.
Protein, precipitated with ammonium sulfate, is extracted with solutions
of decreasing concentrations of ammonium sulfate at or near 0°. The
resultant extracts are allowed to warm to room temperature when
crystallization usually results. The crystals, although very Small, could be
used as seeds for growing larger crystals.

The Use of Detergents in Crystallizing Membrane Proteins

Spectacular advances have been made in crystallizing integral mem-
brane proteins. 27-29The methods use small amounts of nonionic or zwitte-
rionic detergents such as octyl fl-glucopyranoside and N,N-dodecyldi-
methylamine N-oxide in order to inhibit micelle formation and permit
crystallization. Three different proteins, bacteriorhodopsin from Halo-
bacterium halobium, 27 the photosynthetic reaction center from Rhodo-
pseudomonas viridis, 28and porin from Escherichia coli, 29have been crys-
tallized using vapor diffusion methods with salts or polyethylene glycol as

21 S.-H. Kim and G. H. Quigley, this series, Vol. 59, p. 3.

22 A. Wlodawer and K. O. Hodgson, Proc. Natl. Acad. Sci. U.S.A. 72, 398 (1975).
23 N. Camerman, T. Hofmann, S. Jones, and S. C. Nyburg, J. Mol. Biol. 44, 569 (1969).
24 D. C. Richardson, J. C. Bier, and J. S. Richardson, J. Biol. Chem. 247, 6368 (1972).
25 E. N. Baker and G. Dodson, J. Mol. Biol. 54, 605 (1970).
26 W. B. Jakoby, this series, Vol. 22, p. 248.
27 H. Michel and D. Oesterhelt, Proc. Natl. Acad. Sci. U.S.A. 77, 1283 (1980).
28 H. Michel, J. Mol. Biol. 158, 567 (1982).
29 R. M. Garavito, J. Jenkins, J. N. Jansonius, R. Karlsson, and J. P. Rosenbusch, J. Mol.
Biol. 164, 313 (1983).
378 OTHER SEPARATION SYSTEMS [23]

the precipitant with protein that had been solubilized with 0.5-1.0% de-
tergent.

Crystallization Procedures

The Hanging Drop Procedure 3,z2

Equipment
A tissue culture tray with 24 wells (Linbro, Catalog No. 76-033-05)
Plastic or siliconized glass, 22 x 22 mm, microscope coverslips.
Round coverslips, 22 mm in diameter, may also be used. Plastic
coverslips are cleaned by rinsing in ethanol followed with distilled
water
Stopcock grease
Procedure
1. Apply the stopcock grease around the rim of those wells in the
tissue culture tray that will be used for crystallization attempts.
2. Prepare the precipitant solution and add approximately 1 ml at the
appropriate concentration to each well.
3. Mix some precipitant with the protein solution and centrifuge in a
small-volume bench-top centrifuge to remove any precipitate or
debris. It is common to mix equal volumes of the well solution with
the protein solution. Where precipitation is not a problem this can
be done on the coverslip.
4. Place 5-20/zl of the protein-precipitant mixture on the coverslip,
invert the coverslip, and seal it over the well chamber.

The Sandwich Box Procedure 2-4,3-21

Equipment
The "sandwich box," a plastic or glass container (approximately
11 x 11 × 3 cm) that can easily be sealed
Glass microculture slides containing one or more depressions. These
are usually siliconized to reduce spreading of the droplet
Stopcock grease
Procedure
1. Add a suitable volume (5-25 ml) of precipitant solution to the bot-
tom of the box. Alternatively, the solution may be placed in a
beaker or other suitable vessel.
[23] PROTEIN CRYSTALLIZATION 379

2. Place the slides in the sandwich box on top of a su.pport such as an

inverted petri dish to raise it above the level of the surrounding
precipitant solution.
3. Mix some precipitant with the protein solution and centrifuge to
remove any precipitate or debris. As in the hanging drop method, it
is common to mix equal volumes of precipitant in the box with the
protein solutions.
4. Pipette 10-100 pl of the protein-precipitant mixture into the de-
pression slide. Seal the sandwich box.

The Zeppezauer Microdialysis Procedure 8'9

Equipment
Glass capillary, usually about 30 mm in length, with outer diameter of
- 8 mm and an inner diameter of 0.5-2.0 mm. The ends of the
capillary should be flattened and rounded to remove sharp edges
that might damage the dialysis membrane. It is recommended that
the glass capillaries be siliconized
Dialysis membrane, 2-3 cm 2 square. The membrane should have the
appropriate molecular weight cutoff and be prepared according to
the manufacturer's or one's own specifications
Tygon or PVC tubing is cut into -1.5-cm pieces with notches of - 0 . 5
cm cut into one end. The tubing should have an inner diameter
slightly less ( - 7 mm) than the diameter of the capillary to ensure a
snug fit
Parafilm
Scintillation vial or other appropriate vessel that might be sealed and
will allow observation inside the capillary

Procedure
I. Assemble the dialysis cell by covering one end of the glass capillary
with the semipermeable membrane, attaching it securely with the
Tygon or PVC tubing.
2. Transfer the protein solution into the dialysis cell using a syringe or
small plastic tubing. Care must be taken to ensure that the mem-
brane is not ruptured and that air bubbles are not trapped between
the protein solution and the dialysis membrane.
3. Seal the open end of the dialysis cell with Parafilm.
4. Place the dialysis cell into the scintillation vial, which contains
about 5 ml of the equilibration solution. Ensure that there are no
bubbles between the outside of the dialysis membrane and the
equilibration solution. Seal the vial.
380 OTHERSEPARATIONSYSTEMS [23]

Seeding

If crystals can be grown but remain too small for X-ray analysis in
spite of attempts to make them larger by increasing the volume of the drop
or increasing the protein concentration, seeding may be helpful. This may
be done microscopically or macroscopically.
Microseeding 4 involves crushing a crystal or crystals followed by se-
rial dilution of the suspension in a liquid that will not dissolve the crystals.
Serial solutions are generally by factors of 10, and each is tested by adding
a small amount, about 1 /xl, to separate crystallizing droplets until the
fewest crystals are generated. For this method, the protein solution is
usually equilibrated to a point at or just below normal crystallization
conditions.
In macroseeding, on the other hand, an intact single seed crystal is
transferred to a fresh protein solution that has been brought very close to
the original crystallization conditions. 3°,31 The seed crystal is first washed
several times in a solution that will dissolve potential nucleation sites and
generate "fresh" surfaces on the crystal that will not inhibit further
growth. It has been possible in this way to obtain large single crystals
from proteins that, in the absence of seeding, were of inadequate size.

Final Considerations
As a general strategy for attacking the problem of crystallizing a spe-
cific protein, we recommend that the initial step be the testing of these
precipitating agents most commonly used to grow protein crystals (see the
table) at several pH values and at both room temperature and 4 °. If crys-
tallization does not occur in the presence of buffer and precipitating agent
alone, occasionally the addition to the protein solution of small quantities
of monovalent or divalent cations, ligands, substrates or products, sub-
strate or product analogs, alcohols, and other organic molecules may aid
in inducing crystallization. For compounds that bind to the protein, con-
centrations may range from stoichiometric amounts to 10 mM. For mono-
valent and divalent cation additives, the concentrations can range from 1
to 10 mM. For alcohols and other organic molecules the concentra-
tions may range from only trace amounts to 5-10% (v/v). Mixtures of
polyethylene glycol and salt, and even of polyethylene glycol and alcohols
(> 10%), have been reported to be effective. 32,33

30 C. Thaller, L. H. Weaver, G. Eichele, E. Wilson, R. Karlsson, and J. N. Jansonius, J.

Mol. Biol. 147, 465 (1981).
31 j. D. G. Smit and K. H. Winterhalter, J. Mol. Biol. 146, 641 (1981).
32 R. J. Collier, E. M. Westbrook, D. B. McKay, and D. Eisenberg, J. Biol. Chem. 257,
5283 (1982).
33 R. J. Collier and D. B. McKay, J. Mol. Biol. L~7, 413 (1982).
[24] IMMUNOSORBENTSEPARATIONS 381

Although the total number of possible experiments is infinite, crystals

are often obtained after relatively few attempts. If a protein proves to be
difficult to crystallize, however, the number of trials necessary may be-
come too large to be practically feasible. One remedy to this situation has
been proposed by Carter and Carter, 34 who have devised a statistically
effective sampling method using an incomplete factorial procedure poten-
tially to reduce the number of trials. Despite the existence of stubborn
proteins that fail to crystallize, other proteins occasionally will crystallize
over a wide range of conditions, with crystals appearing in the first few
attempts. The usual case lies between the extremes.

34 C . W . C a r t e r , J r . a n d C . W . C a r t e r , J. Biol. Chem. 2 5 4 , 12219 (1979).

[24] I m m u n o s o r b e n t S e p a r a t i o n s
By GARY J. CALTON

Immunosorbent separation is based on the principle of molecular rec-

ognition, defined as the formation of a complex between two specific
molecules and exemplified by the immunological complex between anti-
gen and antibody. The application of chromatographic methods to im-
munosorbent separation can lead to spectacular one-step purification
schemes for proteins from such complex mixtures as urine, cell extracts,
and microbial fermentation broths. The reasons for the rising popularity
of this process include simplicity, a high degree of specificity, selective
affinities, rapid, single-step isolation, and minimal contamination by non-
specific proteins.

Preparation of Affinity Reagents

Immunosorbent chromatography reagents are prepared by immobiliz-
ing a specific antibody by means of an immobilization reagent to an appro-
priate chromatography matrix. The use of monoclonal antibodies for af-
finity reagents is superior in all cases to polyclonal antibodies from animal
sera. The preparation of monoclonal antibodies cannot be covered here
and the reader is referred to the excellent reviews that are available. ,-3

1 j. W. Goding, J. lmmunol. Methods 39, 285 (1980).

2 R. H. Kennet, T. J. McKearn, and K. B. Bechtol, "Monoclonal Antibody Hybridomas: A
New Dimension in Biological Analysis." Plenum, New York, 1980.
3 j. j. Langone and H. Van Vunakis, this series, Vol. 92.
Copyright © 1984by Academic Press, Inc.
METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
[24] IMMUNOSORBENTSEPARATIONS 381

Although the total number of possible experiments is infinite, crystals

are often obtained after relatively few attempts. If a protein proves to be
difficult to crystallize, however, the number of trials necessary may be-
come too large to be practically feasible. One remedy to this situation has
been proposed by Carter and Carter, 34 who have devised a statistically
effective sampling method using an incomplete factorial procedure poten-
tially to reduce the number of trials. Despite the existence of stubborn
proteins that fail to crystallize, other proteins occasionally will crystallize
over a wide range of conditions, with crystals appearing in the first few
attempts. The usual case lies between the extremes.

34 C . W . C a r t e r , J r . a n d C . W . C a r t e r , J. Biol. Chem. 2 5 4 , 12219 (1979).

[24] I m m u n o s o r b e n t S e p a r a t i o n s
By GARY J. CALTON

Immunosorbent separation is based on the principle of molecular rec-

ognition, defined as the formation of a complex between two specific
molecules and exemplified by the immunological complex between anti-
gen and antibody. The application of chromatographic methods to im-
munosorbent separation can lead to spectacular one-step purification
schemes for proteins from such complex mixtures as urine, cell extracts,
and microbial fermentation broths. The reasons for the rising popularity
of this process include simplicity, a high degree of specificity, selective
affinities, rapid, single-step isolation, and minimal contamination by non-
specific proteins.

Preparation of Affinity Reagents

Immunosorbent chromatography reagents are prepared by immobiliz-
ing a specific antibody by means of an immobilization reagent to an appro-
priate chromatography matrix. The use of monoclonal antibodies for af-
finity reagents is superior in all cases to polyclonal antibodies from animal
sera. The preparation of monoclonal antibodies cannot be covered here
and the reader is referred to the excellent reviews that are available. ,-3

1 j. W. Goding, J. lmmunol. Methods 39, 285 (1980).

2 R. H. Kennet, T. J. McKearn, and K. B. Bechtol, "Monoclonal Antibody Hybridomas: A
New Dimension in Biological Analysis." Plenum, New York, 1980.
3 j. j. Langone and H. Van Vunakis, this series, Vol. 92.
Copyright © 1984by Academic Press, Inc.
METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
382 OTHER SEPARATION SYSTEMS [24]

Antigen Preparation and Testing

In order to obtain a monoclonal antibody it is best to have a supply of
homogeneous antigen, which can be obtained by use of alternate meth-
ods of isolation, e.g., by chromatography or electrophoresis. The purified
antigen is then used to immunize an appropriate strain of animal in which
the monoclonal antibodies are to be produced. Purified antigen is also best
for immunization if animal antiserum, i.e., polyclonal antibody, is to be
used. Similarly, homogeneous antigen is useful as a reagent for an
enzyme-linked immunosorbent assay (ELISA) 4 or radioimmunoassay
(RIA) 5 for identification of hybridomas producing monoclonal antibodies
that react with the antigen. The quantities of purified antigen needed for
immunization are quite small; 10/~g with an appropriate adjuvant may be
sufficient. Tissue culture methods can be used for immunization of spleen
cells in vitro and require even less antigen (10-100 ng). 6
It is possible to use crude preparations of the antigen in order to obtain
monoclonal antibodies and for the detection of the specific hybridomas
that produce monoclonal antibodies to the antigen of interest. One
scheme for using crude mixtures as the starting point for monoclonal
antibodies 7 is the following:
Separation of the crude material is attempted on an ion-exchange resin
and a purified fraction is obtained. The course of purification is deter-
mined, in this instance, by a pharmacological assay of each of the eluted
fractions. A second and completely different method of ion exchange for
separation of the crude mixture is also carried out and the chromato-
graphic fractions obtained are again assayed pharmacologically. The
chromatographic fractions from each of the two separations having the
highest specific activity in the pharmacological assay are then used to
develop an enzyme-linked immunosorbent assay. The protein of interest,
which is enriched in each of these fractions, is used as the initial coat for
an ELISA plate. Antibodies produced by murine hybridomas may then be
incubated with each of the enriched chromatographic fractions, after
which enzyme-labeled rabbit antimouse antibody is incubated in the
ELISA. Following workup of the ELISA, the data are surveyed with the
intent of determining which of the antibodies reacted with each of the
chromatographic fractions having highest specific activity. The likelihood

4 A. Voiler, D. E. Biddwell, and A. Bartlett, "The Enzyme Linked Immunosorbent Assay

(ELISA)." Dynatech Laboratories, Inc., Alexandria, Virginia, 1979.
5 H. Van Vunakis and J. J. Langone, this series, Vol. 70, p. 201.
6 R. A. Luben and M. A. Mohler, Mol. Irnmunol. 17, 635 (1980).
7 p. K. Gaur, R. L. Anthony, T. S. Cody, G. J. Calton, and J. W. Burnett, Proc. Soc. Exp.
Biol. Med. 167, 374 (1981).
[24] IMMUNOSORBENTSEPARATIONS 383

of having the same ratio of contaminating proteins present in peak frac-

tions from two diverse methods of chromatography is relatively low.
Those wells having cells producing antibody to the active fractions are
then cloned and the testing procedure repeated in order to determine
which clones are producing specific antibody. The monoclonal antibodies
that are obtained are tested for neutralization of the pharmacological
activity as a check for the antibody actually bearing the desired activity.
This scheme provides a rational method for selection of hybridomas
producing antibody to specific substances, even if the substance has not
been purified. The method is valid, of course, only when pharmacological
or enzyme activity can be tested. Direct neutralization of such activity
may also be used to determine the presence of monoclonal antibodies
against a specific protein in a mixture, although such assays are often
more time consuming and difficult than the above procedure. Alterna-
tively, analytical procedures such as electrophoresis or isoelectric focus-
ing may be used; the resulting bands are identified by tagging the mono-
clonal antibody of interest with a radioactive isotope and following the
reactivity with the proteins in the gel by means of a nitrocellulose blot)
Once a group of antibodies has been isolated, evaluation of their useful-
ness for immunosorbent chromatography can begin.

Specificity
A high degree of specificity is one of the major advantages in using
monoclonal as compared to polyclonal antibodies. For example, mono-
clonal antibodies may be selected for a specific site on a protein molecule
even if that site is not the most antigenic one. This differentiation can be
fine tuned to allow recognition of peptides that differ by as little as one
amino acid. This degree of selectivity also allows the isolation of individ-
ual subunits of a protein, thereby obtaining undegraded or partially de-
graded material. For instance, urokinase, which consists of two chains
connected by a disulfide bridge, has been isolated as the preferred high-
molecular-weight species 9 by use of a monoclonal antibody. This route
was used because the light chain of urokinase is readily susceptible to
proteolytic degradation. In urine and tissue culture fluids, both proteolyt-
ically degraded and nondegraded urokinases exist. The separation of the
two species is difficult by alternate methods because the light chain is
degraded most rapidly by proteases. It was possible, by selection of a
monoclonal antibody for the light chain, to isolate only high-molecular-

s H. Towbin, T. Staehelin, and J. Gordon, Proc. Natl. Acad. Sci. U.S.A. 76, 4350 (1979).
See also this volume [33].
9 D. A. Vetterlein and G. J. Calton, Thromb. Haemostasis 49, 24 (1983).
384 OTHER SEPARATION SYSTEMS [24]

weight urokinase consisting of both chains. This method provides a

means of obtaining homogeneous, high-molecular-weight urokinase in a
single step from such diverse, relatively complex starting materials as
urine or tissue culture fluids. It also allows the isolation of materials that
differ in pharmacological activity in vitro and which are extremely diffi-
cult to separate by alternate means.

Stability
Stability of the monoclonal antibody must be assessed if immunosor-
bent chromatography is to be used for separating mixtures that may con-
tain proteolytic enzymes, as with crude protein extracts. Monoclonal
antibodies from the same fusion may have widely differing stabilities to
proteolytic degradation. If the stability of the antibody is not determined
prior to immobilization, the number of times that an immunosorbent chro-
matography column can be reused may be severely reduced. Stability to
proteolytic degradation may be determined by subjecting each of the
monoclonal antibodies to each of the proteases present in the mixture to
be separated, if they are known. If the identity of the protease is un-
known, pepsin and papain should be used to examine the stability of the
monoclonal antibody to proteolysis. Gel electrophoresis of the mixture
after proteolysis will easily distinguish between stable and unstable anti-
bodies by examination of the molecular weights of the fragments obtained
in relation to the number of hours that the mixture was incubated with the
protease.

Affinity
Of greatest importance for any chromatographic scheme is the affinity
of the antibody for the protein of interest. It is possible to identify mono-
clonal antibodies with specific affinities that will dissociate from their
antibody-antigen complex under conditions appropriate to maximum sta-
bility for the protein to be isolated. Most of the work with immunosorbent
chromatography has used extremes of pH or chaotropic agents for disso-
ciation of the antibody-antigen complex. 1° Such extreme conditions are
unnecessary, because antibodies can be obtained that dissociate under
specified circumstances. To accomplish this end, a previously developed
ELISA can be used for antigen determination. The ELISA may be con-
ducted under standard conditions by coating a 96-well plate with antigen
followed by application of an excess of monoclonal antibody. A series of
dissociation buffers, in which the antigen is stable, are added to the wells

10M. Wilchek, this volume [1].

[24] IMMUNOSORBENTSEPARATIONS 385

containing the antigen-monoclonal antibody complex. Sufficient time for

dissociation of the complex is allowed, from 1 to 2 hr, and the degree of
dissociation is evaluated by addition of enzyme-labeled antiserum. The
data obtained from this procedure will indicate the degree of dissociation
of the antibody-antigen complex. Thus, one can search each of the clones
for specific dissociation conditions. In this manner, it has been possible to
select monoclonal antibodies that dissociate from their antibody-antigen
complex under conditions as mild as 20 mM potassium phosphate and 750
mM sodium chloride at pH 7.11 Under very mild conditions, a matrix for
immobilization must be chosen that retains only slight quantities of non-
specifically adsorbed protein, because washing of the column must be
limited, rapid, and gentle.

Polyclonal vs Monoclonal Antibodies

As noted, monoclonal antibodies are superior to polyclonal antibodies
for immunosorbent chromatography. However, polyclonal antibodies can
be used for this procedure with excellent results. 12-14Basic differences in
the use of monoclonal and polyclonal antibodies involve specificity, affin-
ity, and capacity. In each area monoclonal antibodies possess significant
advantages over the use of polyclonal antibodies. On the other hand,
polyclonal antibodies are simpler for most investigators to develop be-
cause the procedures are clearly defined 15 whereas production of mono-
clonal antibodies requires a significant investment in equipment, time,
and skills.
The specificity of monoclonal antibodies is much higher than that of
polyclonal antibodies because a binding protein recognizing only a single
site is produced by an appropriately cloned hybridoma. This is in contrast
with the situation with polyclonal antibodies, in which each of the spleen
cells in the animal is capable of producing an antibody, thereby providing
a vast array of antibodies directed against many sites on the protein to
which the animal has been exposed. Cross-reactivity with such animal
sera is recognized.
Within an animal serum, both high- and low-affinity antibodies exist.
In immunosorbent chromatography, antibodies with very low affinity
tend to allow the antigen to bleed, i.e., to wash slowly from the column

u C. S. Cobbs, P. K. Gaur, A. J. Russo, J. E. Warnick, G. J. Calton, and J. W. Burnett,

Toxicon, 21, 385 (1983).
12 j. W. Eveleigh and D. E. Levy, J. Solid Phase Biochem. 2, 45 (1977).
13 D. Vettedein, T. E. Bell, P. L. Young, and R. Roblin, J. Biol. Chem. 225, 3665 (1980).
14 See this series, Vol. 34, [90-94].
i5 p. H. Maurer and H. J. Callahan, this series, Vol. 70, p. 49.
386 OTHERSEPARATIONSYSTEMS [24]

regardless of eluant, whereas antibodies with very high affinity are virtu-
ally impossible to dissociate from their antibody-antigen complex with-
out resorting to denaturing agents. A similar situation holds with mono-
clonals, but with them it is possible to select the degree of affinity that is
desired for chromatographic purposes.
The degree of affinity of the antibody also affects capacity of the
immunosorbent chromatography reagent. With monoclonal antibodies, a
defined moderate-level affinity can be obtained and recognized (by the
buffer/elution experiments suggested above) and used for immunosorbent
chromatography. However, because animal antisera contain antibodies
with a large range of binding affinities, the antibody-antigen complex at
these sites is normally not dissociable, with capacity decreasing severely
upon initial use of the reagent. With polyclonal antibodies, 5 to 10 runs are
required to obtain a level at which antigen is released under desirable
conditions. Low-affinity antibodies, present in animal antisera, preclude
isolation of the antigen of interest in high yields because the washing step
required to remove extraneously adsorbed protein on the matrix results in
bleeding of the antigen of interest due to dissociation of the weakly bound
antigen.

Immobilization
Once appropriate monoclonal antibodies have been obtained, a num-
ber of methods of immobilization are available to the investigator.16 Be-
cause the monoclonal antibodies differ from one another, immobilization
conditions must be tailored to the specific monoclonal antibody. For opti-
mum results in immobilization, it is recommended that a series of mono-
clonal antibodies be used, all of which react on ELISA with the antigen of
interest. The antibodies may be from one or more fusions, but it is not
unusual to find that not more than 1 in 10 monoclonal antibodies is suit-
able for use as an immunosorbent reagent. A major reason for this con-
cerns the conformational constraints that may occur when the mono-
clonal antibody is immobilized. An erroneous impression of suitability for
immunosorbent chromatography may be obtained from the ELISA. If the
monoclonal antibody is complexed with the antigen that has been immobi-
lized on a polystyrene ELISA plate, absolute freedom of conformation of
the antibody is possible. When the monoclonal antibody is immobilized
on a polymer matrix, its structure may be deformed, active sites may be
bound, or the Fab portion may be in a position in which it is not possible
to form an antibody-antigen complex. These observations explain the

16 K. Mosbach, this series, Vol. 44.

[24] IMMUNOSORBENTSEPARATIONS 387

discrepancy that is often found between reactivity in an ELISA and

results with a monoclonal antibody in immunosorbent chromatography.

Chromatography

The chromatographic step itself is carried out by the techniques of

standard affinity chromatography.~7 The antigen should be charged onto
the column in a buffer system adequate for protein stability, while at the
same time being sufficiently dilute to avoid dissociation of the antibody-
antigen complex. Such conditions are determined by a study of the pro-
tein itself and from the ELISA/buffer dissociation experiments that have
been described. The concentration of the antigen in the solution is of little
importance because the concentration of the antigen is normally dilute
when compared to the quantity of antibody immobilized on the column
(2-10 mg/ml of matrix). Because the antibody will be in large excess as
the chromatographic front advances, the equilibrium will favor formation
of the antibody-antigen complex. Antibody-antigen complexes are rap-
idly formed under chromatographic conditions and flow rates as high as 20
bed volumes per hour have been used. The normal limitation imposed
upon the flow rate is due to the polymer matrix on which the monoclonal
antibody has been immobilized, rather than to the formation of the anti-
body-antigen complex. For laboratory isolations this is not a hindrance,
but in large-scale purifications the flow rate becomes a limiting factor.
Washing conditions are based upon the previously described ELISA,
which determines the dissociation conditions to be used. Such conditions
should be sufficiently mild that dissociation of the antibody-antigen com-
plex will not occur, but sufficiently strong that the column will be freed of
nonspecifically adsorbed proteins. Once extraneous proteins have been
desorbed, as determined by adsorbance or other suitable analytical meth-
ods, dissociation buffers are used as indicated by the ELISA buffer assay.
It is not unusual that either stronger or weaker buffers than those selected
by ELISA be required for dissociation of the antibody-antigen complex
because immobilization of the antibody may alter the binding affinity of
the antibody.
The method of immunosorbent chromatography, particularly with the
use of monoclonal antibodies, is rapid, specific, and gentle, and the affin-
ity reagents can be used repeatedly.

17W. B. Jakoby and M. Wilchek, this series, Vol. 34.

[25] H I G H - P U R I T Y LABORATORY W A T E R 391

[25] P r e p a r a t i o n of H i g h - P u r i t y L a b o r a t o r y W a t e r
By GARY C. GANZI

High-purity water is critically important for separations, analyses, and

biological preparations.

Water Impurities
There is ample evidence that research results may be affected signifi-
cantly by levels of water impurities that are minute in comparison to
levels found in typical potable supplies. Contaminants may occur natu-
rally, be added by potable treatment processes, or be leached from distri-
bution systems. Levels of contaminants in source waters may vary sea-
sonally and, at any given location in the distribution system, may vary
daily. Such impurities are commonly categorized as follows.
Organic and inorganic particulates are found in all natural waters.
They are also added by sloughage from distribution system components
and can be formed by the interaction and precipitation of dissolved con-
stituents. Particles may clog precision equipment and shield bacteria from
disinfection. Colloidal materials (particles smaller than about 0.1 ~m in
diameter), can interact with biological systems as a result of their inherent
electrostatic charge. Particulate concentrations in potable water are com-
monly in the part per million range.
Concentrations of dissolved organic compounds (expressed as total
organic carbon) are in the part per million range in many potable water
supplies. In most surface water supplies, the primary process of synthesis
of dissolved organics is the natural biodecomposition of living matter. For
example, tannins, lignins, and humic acids are formed by the decomposi-
tion of vegetation. Other sources are agricultural runoff and compounds
added during treatment of municipal water. Pesticides, herbicides, deter-
gents, and polyelectrolytes are commonly found in the part per billion
range of concentrations. In addition, a wide variety of organic compounds
are often synthesized in water supplies by the interaction of organic and
inorganic impurities with oxidizing additives used for microbial control.
Dissolved inorganics are primarily ionic in nature and are found in
relatively large concentrations in most water supplies. Typical concentra-
tions of major constituents are shown in Table I. Trace inorganic constitu-
ents, i.e., less than about one part per million, commonly include metals
and chlorine-containing sanitizing agents, which even at parts per billion

Copyright © 1984by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182(104-1
 392 RELATED TECHNIQUES [25]

TABLE I
MAJOR INORGANIC CONSTITUENT CONCENTRATIONS OF TYPICAL POTABLE W A T E R SUPPLIES

Concentrationof inorganicconstituents(partsper million)

New San
York Los Palo St. Suffcrn, Juan,
Constituent City~ Angelesb Chicagoc Bostond Altoe Dallasy Louisg N.J.h p.R.i

Calcium 17 81 36 9 68 26 -- 43 38
Magnesium 6 16 10 5 19 4 45 11 6
Sodium 10 22 5 11 69 17 44 61 14
Potassium 2 5 1 1 11 -- -- 2 6
Bicarbonate 59 190 64 8 429 61 26 137 73
Sulfate 16 83 25 5 7 38 152 35 50
Chloride 24 56 11 11 35 18 24 100 30
Nitrate 1 4 0.2 0.1 2 1 5 3.5 6
Silica 5 22 3 -- 106 -- -- 15 30
Oxygen 11 -- 12 10 . . . . .

a Groton Supply from representative sites: average values for 1978.

b Municipal Supply, February, 1982.
c City of Chicago, South Water District, February, 1976.
a Weston Reservoir, 1980 average.
e Municipal Supply, December, 1981.
¢ Dallas Utilities East Side Plant, June, 1979.
Municipal Supply, May, 1972.
h Municipal Well Source, January, 1982.
i Municipal Supply, November, 1981.

concentrations, e.g., the h e a v y metals, may affect both biological and

biochemical systems.
Microorganisms a r e f o u n d in m o s t w a t e r s o u r c e s a n d t e n d to p r o l i f e r -
a t e in d i s t r i b u t i o n s y s t e m s . I n a d d i t i o n to b e i n g a d i r e c t s o u r c e o f c o n t a m -
ination, e n z y m e s p r o d u c e d by them may interact with biochemical com-
pounds and reduce product purity. S o m e endotoxins from gram-negative
b a c t e r i a a r e k n o w n to a f f e c t b i o l o g i c a l s y s t e m s in c o n c e n t r a t i o n s in t h e
p a r t p e r trillion r a n g e .

Water Treatment Systems

A wide variety of water treatment systems are available to the investi-
gator that provide water suitable for laboratory use. Each system has its
advantages and limitations and, since no individual unit process is capable
of removing all contaminants, water treatment equipment is often in-
stalled in sequential arrangements to provide adequate purity. S o m e of
[25l HIGH-PURITY LABORATORY WATER 393

TABLE II
RETENTION EFFICIENCIES OF CARTRIDGE-TYPE DEPTH FILTERSa

Weight percentage
retention of
Cartridge description AC fine dust b

Porous cellulose (5-/,m rating) 57

Spun polyethylene (5-/~m rating) 60
Polypropylene wound fibers (10-tzm rating) 72
Polypropylene wound fibers (5-tzm rating) 79
Cotton wound fibers (10-/.~m rating) 81
Cotton wound fibers (5-/zm rating) 85

a Data obtained by F. Badmington, Millipore Corp., Ashby

Road, Bedford, MA.
b Purchased from the AC Spark Plug Division of General Motors
Corp. The cartridges were challenged with 200 ppm dust in
water suspension. Dust particle size distribution as specified by
the vendor was as follows:

Particle diameter range (l~m) Weight percentage

0-5 39 -+ 2
5-10 18 - 3
10-20 16 -+ 3
20-40 18 -+ 3
40-80 9 -+ 3

the more common laboratory water treatment processes are discussed

here.
Depth Filtration. Depth filtration is primarily effected by passing wa-
ter through beds of granular material such as sand or through cartridges
constructed of closely wound polymeric fibers or packed porous matrices.
Depth filters primarily remove particles larger than 1 /xm in diameter.
Removal is more efficient as particle size increases. Typical removal effi-
ciencies of common cartridge-type depth filters are shown in Table II. The
predominant mechanism of particle removal is physical exclusion. Sur-
face modification of the filter matrix can often result in higher removal
efficiencies by adsorption mechanisms. In some cases, depth filtration is
preceded by chemical reagent addition. This can cause agglomeration of
colloids or precipitation of dissolved materials. The particles formed are
then large enough to be removed by depth filters. Most depth filters are
discarded and replaced when plugging of the filter matrix causes exces-
sive pressure losses in the purification system. However, beds of sand
394 RELATED TECHNIQUES [25]

100 I I ~ i I I I l I I i

I /1 I I - 2

a
UJ

0
D
<C
/
I I I I I I I I I I
0.05 0.1 0.5 1.0

MONOCHLORAMINE AT EQUILIBRIUM
(ppm)

FIG. I. Activated carbon adsorption isotherms of aqueous monochloramine solutions at

25° and pH 7.5. Curves: 1, bituminous coal based; 2, coconut shell based; 3, petroleum
based; 4, synthetic polymer based. Data obtained by K. Siu, Continental Water Systems
Corp., Ashby Road, Bedford, MA.

and similar media can be flushed clean of particles by periodic high veloc-
ity backwash cycles.
Activated Carbon Treatment. Activated carbons are treated carbona-
ceous materials with extremely high porosity and surface areas. The
structure of activated carbon allows physical adsorption and catalytic
decomposition of some organic contaminants and chlorine-based sanitiz-
ing agents. Granular or powdered carbons are available in a wide variety
of pore size distributions. As a result, their ability to adsorb specific
contaminants varies widely. Adsorption isotherms for monochloramine, a
common potable water-sanitizing agent, are shown in Fig. 1. In general, a
carbon with a predominance of large pores performs more efficiently on
tannins, lignins, humic acids, and colloidal particles. A high-surface-
area carbon with a predominance of small pores removes lower molecular
weight compounds more efficiently. For this reason, many ~ligh-perfor-
mance water purification systems utilize a large-pore carbon bed for pri-
mary organic removal, followed by a small-pore polishing bed to remove
low-molecular-weight compounds. Dissolved organic carbon levels usu-
ally can be reduced to concentrations below 50 parts per billion in this
[25] HIGH-PURITY LABORATORY WATER 395

way. Carbon filters are discarded and replaced when their effectiveness
diminishes, although commercial regeneration services are available.
Ultrafiltration. Ultrafilters are available primarily as polymeric mem-
branes in flat sheet or tubule form. When contaminated water is driven
through them under pressure, their extremely small pores efficiently, but
not entirely, prevent passage of particles, colloids, microorganisms, and
bacterial endotoxins. Typical retention efficiencies are shown in Fig. 2.
Ultrafilters range in nominal pore size. For water treatment, they are
commonly specified to exclude 90-95% of contaminants of 10,000
daltons. Despite their small pore size ratings, water flux through ultrafil-
ters is usually high at relatively low transmembrane pressure differentials
(170-340 kPa). In order to reduce membrane fouling and loss of flux,
ultrafiltration equipment is designed to optimize the rate of diffusion of
excluded impurities away from the membrane surface. When membrane
fouling does occur, ultrafiltration modules can be chemically cleaned ei-
ther in place or by a commercially available service.
Ion Exchange. Ion exchangers contain chemically bound ionic groups

100
¢ f
80

z
o so
I,-
z
t,t.i
,411 /
I-
i,u
a,.

/
20

J .y,//'
102 103 104 105 10 6 107

IMPURITY MOLECULAR WEIGHT

FIG. 2. Retention efficiencies for ultralilters of different pore size as a function of impu-
rity molecular weight. Curves: 1, 1 x 103 dalton rating; 2, 1 x 104 dalton rating; 3, 2.5 x 104
dalton rating; 4, 1 x 105 dalton rating; 5, 1 x 106 dalton rating. Data were obtalnccl from
Millipore Corp., Ashby Road, Bedford, MA.
396 RELATED TECHNIQUES [25]

T A B L E 111
TYPICAL METAL PURITY LEVELS ATTAINABLE BY USE OF
ION EXCHANGE AND DISTILLATIONa

Ion exchange
Two-stage Single-pass followed by
Metal T w o - s t a g e still q u a r t z still ion e x c h a n g e distillation Milli-Q b

A1 10 0.5 -- 0.1 <0.05

Ag 1 -- -- -- 0.01
B 0.01 -- -- -- 3
Ca 50 0.07 0.2 0.03 <0.05
Cr -- -- 0.02 -- <0.1
Cu 50 -- -- -- <0.05
Fe 0.1 -- 0.02 -- <0.01
Mg 80 0.05 2 0.01 0.03
Mn 0.01 -- 0.02 -- <0.01
Na 1 -- -- -- 0.07
Ni -- -- 0.002 -- <0.01
Pb 50 -- 0.02 -- <0.05
Si 50 5 -- 1 1
Sn 50 -- -- -- <0.1
Zn 10 -- 0.06 -- 0.03

a A . R. K n o t t , At. Absorpt. Newsl. 14, N o . 5, 126 (1975). D a t a a r e g i v e n in p a r t s p e r

billion.
b T r a d e m a r k o f M i l l i p o r e C o r p . T h e Milli-Q c o n t a i n s a c t i v a t e d c a r b o n a n d i o n - e x c h a n g e
b e d s f o l l o w e d b y a m i c r o p o r o u s filter.

in equilibrium with free replaceable ions of opposite charge. They are

primarily available in the form of polymeric beads. The most common
laboratory configuration are beds consisting of a mixture of cation- and
anion-exchange resins. Water is deionized as it passes through these beds
by the exchange of ionized contaminants with equal equivalents of hydro-
gen and hydroxide ions. Since ion-exchange beds are extremely efficient
deionizers, purification levels approaching the theoretical electrical resis-
tivity of water (18.3 ~-cm x 106) are routinely obtained. Typical purity
levels attainable are shown in Table III.
Ion-exchange resins are available with variations in charge density,
porosity, and chemical structure and, therefore, have a wide range of
properties and purification capabilities. Strongly ionized resins are uti-
lized for high degrees of deionization and the removal of dissolved silica
and carbon dioxide. They are most effective in the absence of compounds
that may irreversibly foul them, e.g., dissolved iron or fulvic acid. Macro-
porous resins are utilized to deionize in the presence of colloids. Weakly
[25] HIGH-PURITY LABORATORY WATER 397

ionized resins are available for pretreatment and partial deionization of

highly fouling feeds.
Ion-exchange resins can also be used as adsorbents for removal of
colloids and dissolved organics. In addition, tailored resins are available
for the preferential reaction and reduction of concentrations of specific
ions, particularly heavy metals, to extremely low levels.
Since ion exchangers remove ionic impurities by chemical substitu-
tion, their ion removal capacity is limited. For applications involving high
ionic loads, where resin disposal is not economical, resins are regenerated
by use of special equipment on site or by a commercially available regen-
eration resin exchange service. For applications involving low ionic
loads, exhausted resins are often discarded.
The proper use of ion exchangers depends on a thorough understand-
ing of the multicomponent chemical equilibria involved. The reversible
nature of ion-exchange reactions can result in leakage of trace quantities
of ions (e.g., sodium, silicate, and chloride) by elution with ions (e.g.,
calcium and sulfate) in the feed water that associate more strongly with
the resins.
For applications that indicate the use of disposable resins, leakage of
ions is minimized with resins of maximum acidity and basicity (such anion
resins are called type I resins). To minimize residual ions from resin
synthesis, a multistep chemical conditioning process is performed using
large excesses of acid and caustic. Cation and type I anion resins condi-
tioned in this manner are called nuclear grade resins.
In cases where resins are regenerated, the extent of ionic leakage also
depends on the service history of the resin, the thoroughness of regenera-
tion, and the ease of contaminant removal from the resin. The advantage
of the high basicity of type I anion resin becomes a disadvantage during
regeneration because of its high affinity for contaminants. For example,
relatively small concentrations of chloride ion in the regeneration caustic
can cause significant chloride leakage upon subsequent deionization with
type I resins, l Higher water purity can often be obtained with resins
having improved regeneration characteristics, such as the dimethyl-
ethanolamine-based type II anion resins. 2These resins have a higher affin-
ity for hydroxide ion than the trimethylamine-based type I resins. Con-

F. X. McGarvey, S. M. Ziarkowsky, E. W. Hauser, and M. C. Gottlieb, "Effect of Caustic

Quality on the Performance of Strong Base Anion Exchangers." Sybron Chemical Divi-
sion, Birmingham, NJ, IWC-81-31.
2 M. C. Gottlieb and G. P. Simon, "Type II Strongly Basic Anion Exchange Resins for Two
Bed Demineralizers." Presented at the 13th Liberty Bell Corrosion Conference, Septem-
ber 18, 1975, Philadelphia, PA.
398 RELATED TECHNIQUES [25]

TABLE IV
TYPICAL SEPARATION EFFIC1ENCIES OF
CELLULOSE ACETATE REVERSE
OSMOSIS MEMBRANESa

Solute Percentage separation

LiF 98.5
LiCI 94.2
LiBr 93.2
LiNO3 88.9
NaCI 94.4
KF 94.8
KCI 94.5
KBr 93.7
LCIO3 91.2
KNO3 86.5
KCIO4 86.3
RbCI 94.7
RbBr 94.5
CsCI 95.0
CsBr 94.5
MgC12 98.0
MgBr2 96.9
Mg(NO3)z 96.1
CaCI2 96.4
SrC12 95.8
BaCI2 96.3

a T. Matsuura, L. Pageau, and S. Sourirajan,

J. Appl. Polym. Sci. 19, 179 (1975).

taminants are more readily eluted from type II resins, and they are less
prone to recontamination by reagents and rinse water during regenera-
tion.
Distillation. Equipment for distillation purifies by vaporizing feed wa-
ter and condensing the vapor as product. Impurity removal efficiencies
generally increase as the impurity vapor pressure decreases below that of
water. Therefore, distillation efficiently removes particles, microorga-
nisms, macromolecules, and ionic contaminants. Product purity generally
can be improved by increasing the number of consecutive times the prod-
uct is distilled, by increasing the ratio of feed to product water, and by
using stills with an inert construction material, such as quartz. Typical
product quality from stills is listed in Table III.
Reverse Osmosis. Reverse osmosis equipment makes use of mem-
branes, usually in the form of spirally wrapped flat sheets or hollow fibers,
[25] HIGH-PURITY LABORATORYWATER 399

that are highly permeable to water and exclude most impurities (primarily
on the basis of molecular size and charge). Such systems efficiently re-
move particles, microorganisms, macromolecules, and most dissolved
organic materials. Ion rejections generally vary from 90 to 99%. Since
product flow per unit membrane area is a function of the transmembrane
pressure differential, most reverse osmosis equipment utilizes a high-
pressure feed pump to increase productivity. Typical removal efficiencies
are presented in Table IV.
Ultraviolet Irradiation. Irradiation of water can destroy microorga-
nisms and, depending on intensity and frequency, decompose a number of
organic materials. Typical capabilities are shown in Table V. Irradiation
equipment is designed to maximize energy input density and to prevent
coating of the ultraviolet lamps with water impurities.

TABLE V
INCIDENT ULTRAVIOLETENERGIESAT 2537 A
NECESSARY FOR MICROORGANISMINHIBITIONa

Energy required
for inhibition
(microwatt -
sec/cm2)

Organism 90% 100%

Bacteria
Bacillus subtilis 5,800 11,000
B. subtilis (spores) 11,600 22,000
Escherichia coli 3,000 6,600
Micrococcus candidus 6,050 12,300
Pseudomonas aeruginosa 5,500 10,500
Pseudomonas fluorescens 3,500 6,600
Staphylococcus albus 1,840 5,720
Staphylococcus aureus 2,600 6,600
Streptococcus lactis 6,150 8,800
Yeast
Saccharomyces ellipsoideus 6,000 13,200
Saccharomyces cerevisiae 6,000 13,200
Mold spores
Aspergillus glaucus 44,000 88,000
Aspergillus flavus 60,000 99,000
Aspergillus niger 132,000 330,000

R. Nagy, Am. Ind. Hyg. Assoc. J. 25, 274 (1964).

400 RELATED TECHNIQUES [25]

J~ .... ~)" A

o o ~ ~ ' ~ ~

8 . >. ~*~ ~

u U ~
te
C~

Z
I- p '=
< u ~

•~ '= o

0 0 J 0

'~ o~ .~' ~ "~

I-

[-' < 2 2 ~

p
.t 6 6~

<
L)
<

0 ~ r2 0 0
u~
o
~ o~
<

r.

,.0

.~ Q.)

"= = 0

. o~
P~ c)
[25] HIGH-PURITY LABORATORYWATER 401

Microporous Membrane Filtration. Microporous membrane filters are

polymeric filters of extremely well-defined pore size that quantitatively
screen particles and microorganisms larger than their rated size. Micropo-
rous membranes also remove particles smaller than their rating by a depth
filtration mechanism (although removal for these smaller particles is not
absolute). Membrane filters are available primarily in the form of flat
sheets or pleated cartridges in disposable configurations, designed to
maximize membrane surface area per unit volume. Since microporous
filters tend irreversibly to plug in the presence of high particulate loads,
their predominant use in water purification is as a final polishing and
microbiological barrier.

Process Considerations
Choosing the appropriate combination of purification steps for produc-
tion of laboratory-grade water depends not only on the desired product
water quality and quantity, but also on the physical and chemical limita-
tions of the purification equipment. For example, several of the processes
remove certain impurities efficiently, but may themselves add trace quan-
tities of other impurities. For this reason, the order of the purification
steps must be carefully selected. In addition, the active agent in the equip-
ment can sometimes be chemically degraded, fouled, or plugged by spe-
cific water impurities, necessitating the use of pretreatment steps to main-
tain equipment performance. Typical removal capabilities and common
limitations of each purification process are presented in Table VI. The
schematics in Fig. 3 show some common unit operation sequences for the
production of laboratory-grade water.

Practical Considerations
Practical considerations in the choice of water purification equipment
should include not only initial costs, operating and maintenance costs,
and laboratory space required but such other factors as may influence
water quality. Equipment that is unreliable, unsuited to feed water condi-
tions, or requires frequent cleaning may produce a product of poor qual-
ity. The capability of the purification equipment in providing the needed
quantity of water on demand is a distinct advantage, since ultrapure water
tends to become contaminated when stored. Clearly, the ability of the
equipment manufacturer to provide consistent, high-quality, and properly
treated membranes, resins, and wetted materials of construction and to
provide the user with the information required to maintain the equipment
is a critical factor in the ability to produce high-quality water.
402 RELATED TECHNIQUES [25]

FEED

B ~ ~ 8

J I~" ~ PRODUCT

DRAIN

FEED

C PRODUCT

DRAIN DRAIN

FIG. 3. Schematics of common unit operation sequences for the production of labora-
tory-grade water: 1, depth filter; 2, activated carbon bed; 3, ion exchanger; 4, microporous
filter; 5, faucet; 6, still; 7, storage tank; 8, ultraviolet sterilizer; 9, reverse osmosis system;
I0, ultrafilter. (A) High-purity water using service ion exchange. (B) High-purity water
production using distillation. (C) Ultrapure water production using reverse osmosis, two-
stage carbon adsorption, and ultra_filtration. Depending on feed water quality, the systems
shown may require pretreatment by processes such as depth filtration, ultrafiltration, soften-
ing, and pH adjustment.

Ideally, the investigator has been able to identify the water contami-
nants that may affect experimental results and can therefore specify
equipment that provides water of the required purity. Since this is usually
not the case, the investigator can review published water quality stan-
dards and determine which equipment manufacturers meet or exceed
them. Table VII summarizes the purity specifications established by the
American Society for Testing and Materials (ASTM), the College of
American Pathologists (CAP), and the National Committee for Clinical
Laboratory Standards (NCCLS).
Standard methods for the examination of water and waste water also
include a "distilled water suitability test" that determines whether puri-
[25] HIGH-PURITY LABORATORY WATER 403

TABLE VII
LABORATORY WATER PURITY STANDARDS a

Type Type Type Type

Standard Society I II III IV

Electrical resistivity, minimum, ASTM 16.66 1.0 1.0 0.2

ft-cm x 106 a t 25 ° CAP 10.00 2.0 0. l --
NCCLS 10.00 2.0 0. l --
p H a t 25 ° ASTM -- -- 6.2-7.5 5.0-8.0
CAP -- -- 5.0-8.0 --
NCCLS w -- 5.0-8.0 --
Color-retention time of potassium ASTM 60 60 10 10
permanganate (rain) CAP 60 60 60 --
NCCLS 60 10 -- --
Total bacteria (colonies/ml) ASTM . . . .
CAP 10 104 -- --
NCCLS 10 103 -- --
Particulate matter, maximum size ASTM . . . .
(larger than 0 . 2 v.m) CAP None -- -- --
NCCLS None -- -- --
Silica (parts per million, SiO2, ASTM . . . .
maximum) CAP 0.05 0.1 1.0 --
NCCLS 0.05 0.1 1.0 --

reagent grade water; t y p e II, a n a l y t i c a l grade water; t y p e s III a n d I V , general

T y p e I,
laboratory grade water.

fled water inhibits or promotes the growth of bacteria. 3 Other evaluations

report the growth of cells in tissue culture and evaluations of cell yields
as a function of water quality.
The technological advances in biology and biochemistry and improved
sensitivity of analytical methods have increased the need for water of high
purity. The requirement of highly purified water should be recognized as a
necessary component of a successful research program.

3 B. L. Green and W . L i t s k y , J. Food Prot. 8, 6 5 4 (1979).

404 RELATED TECHNIQUES [26]

[26] B u f f e r s for E n z y m e s
By JOHN S. BLANCHARD

The desirability of maintaining a stable pH during an enzyme-cata-

lyzed reaction has long been realized, l Since 1900, when phosphate was
first used as a means of maintaining a stable pH in enzyme studies, 2 a
large number of inorganic compounds have served as buffers, including
phosphate (pK2 7.2), cacodylate (pK 6.3), borate (pK 9.2), and bicarbon-
ate (pK 6.4 and 10.0). The use of weak organic acids and bases extended
the range of pH beyond that which could be obtained with inorganic
buffers. Unfortunately, many buffers have intrinsic handicaps associated
with their use in biochemical systems, e.g., interactions with substrates or
enzymes, resulting in reduced activity. Because buffers are generally
present in much higher concentration than any other component in reac-
tion mixtures, interactions of any sort can seriously influence the interpre-
tation of enzymological data. In this context, the pioneering development
by Good and his colleagues3 in preparing a series of N-substituted taurine
and glycine buffers with pK values in the region of most interest to bio-
chemists has allowed the systematic evaluation of buffers and specific
buffer effects. Since the first report by Good, a number of other
zwitterionic, amino-containing, sulfonic acid buffers with pK values
ranging from 6.1 to 10.4 have been developed 4,5and are commercially avail-
able. A comprehensive discussion of buffers for biological research is
available, 6 and should be consulted for more complete information.

Buffer Selection
Purity and availability are of major practical importance in the choice
of a buffer. Most, if not all, of the buffers listed in Table I meet these
standards and are, as well, reasonably inexpensive. Many are routinely
analyzed commercially for their heavy-metal content.
1 This series, Vol. 22, p. 3, and Vol. 24, p. 53.
2 A. Fernbach and L. Hubert, C. R. Hebd. Seances Acad. Sci. 131, 293 (1900).
3 N. E. Good, G. D. Winget, W. Winter, T. N. Connolly, S. Izawa, and R. M. M. Singh,
Biochemistry 5, 467 (1966).
4 M. A. Jermyn, Aust. J. Chem. 20, 183 (1967).
5 W. J. Ferguson, K. I. Braunschweiger, W. R. Braunschweiger, J. R. Smith, J. J. McCor-
mick, C. C. Wasmann, N. P. Jarvis, D. H. Bell, and N. E. Good, Anal. Biochem. 1114, 300
(1980).
6 D. D. Perrin and B. Dempsey, "Buffers for pH and Metal Ion Control." Chapman & Hall,
London, 1974.

Copyright© 1984by AcademicPress, Inc.

METHODS1N ENZYMOLOGY,VOL. 104 All rightsof reproductionin any formreserved.
ISBN 0-12-182004-1
[26] BUFFERS FOR ENZYMES 405

TABLE l
SELECTED BUFFERS AND THEIR pK VALUES AT 25 °

Trivial name Buffer name pKaa dpKa/dt b

Phosphate (pK0 -- 2.15 0.0044

Malate (pK0 -- 3.40 --
Formate -- 3.75 0.0
Succinate (pK0 -- 4.21 -0.0018
C i t r a t e (pK2) -- 4.76 -0.0016
Acetate -- 4.76 0.0002
Malate -- 5.13 --
Pyridine -- 5.23 -0.014
Succinate (pK2) -- 5.64 0.0
MES 2-(N-Morpholino)ethanesulfonic acid 6.10 -0.011
Cacodylate Dimethylarsinic acid 6.27 --
Dimethylglutarate 3,3-Dimethylglutarate (pK2) 6.34 0.0060
Carbonate (pK1) -- 6.35 -0.0055
Citrate (pK3) -- 6.40 0.0
BIS-Tris [Bis-(2-hydroxyethyl)imino]tris(hydroxy- 6.46 0.0
methyl) methane
ADA N-2-Acetamidoiminodiacetic acid 6.59 -0.011
Pyrophosphate -- 6.60 --
EDPS (pK0 N,N'-Bis(3-sulfopropyl)ethylenediamine 6.65 --
Bis-Tris propane 1,3-Bis[tris(hydroxymethyl)methyl- 6.80 --
amino]propane
PIPES Piperazine-N,N'-bis(2-ethanesulfonic acid) 6.76 -0.0085
ACES N-2-Acetamido-2-aminoethanesulfonic acid 6.78 -0.020
MOPSO 3-(N-Morpholino)-2-hydroxypropanesulfonic acid 6.95 -0.015
Imidazole -- 6.95 -0.020
BES N,N-Bis-(2-hydroxyethyl)2-aminoethanesulfonic 7.09 -0.016
acid
MOPS 3-(N-Morpholino)propanesulfonic acid 7.20 0.015
Phosphate (pK2) -- 7.20 -0.0028
EMTA 3,6-Endomethylene- 1,2,3,6-tetrahydrophthalic acid 7.23 --
TES 2-[Tris(hydroxymethyl)methylamino]ethane- 7.40 -0.020
sulfonic acid
HEPES N-2-Hydroxyethylpiperazine-N'-2-ethanesulfonic 7.48 -0.014
acid
DIPSO 3-[N-Bis(hydroxyethyl)amino]-2-hydroxy- 7.60 -0.015
propanesulfonic acid
TEA Triethanolamine 7.76 -0.020
POPSO Piperazine-N,N'-bis(2-hydroxypropanesulfonic 7.85 -0.013
acid)
EPPS, HEPPS N-2-Hydroxyethylpiperazine-N '-3- 8.00 --
propanesulfonic acid
Tris Tris(hydroxymethyl)aminomethane 8.06 -0.028
Tricine N-[Tris(hydroxymethyl)methyl]glycine 8.05 -0.021

(continued)
406 RELATED TECHNIQUES [26]

TABLE I (continued)

Trivial name Buffer name pKa ~ dpga/dt b

Glycinamide -- 8.06 -0.029

PIPPS 1,4-Bis(3-sulfopropyl)piperazine 8.10 --
Glycylglycine -- 8.25 -0.025
Bicine N,N-Bis(2-hydroxyethyl)glycine 8.26 -0.018
TAPS 3-{[Tris(hydroxymethyl)methyl]amino} 8.40 0.018
propanesulfonic acid
Morpholine -- 8.49 --
PIPBS 1,4-Bis(4-sulfobutyl)piperazine 8.60 --
AES 2-Aminoethylsulfonic acid, taurine 9.06 -0.022
Borate -- 9.23 -0.008
Ammonia -- 9.25 -0.031
Ethanolamine -- 9.50 -0.029
CHES Cyclohexylaminoethanesulfonicacid 9.55 0.029
Glycine (pK2) -- 9.78 -0.025
EDPS N,N'-Bis(3-sulfopropyl)ethylenediamine 9.80 --
APS 3-Aminopropanesulfonic acid 9.89 --
Carbonate (pK2) -- 10.33 -0.009
CAPS 3-(Cyclohexylamino)propanesulfonic acid 10.40 0.032
Piperidine -- 11.12 --
Phosphate (pK3) -- 12.33 -0.026

These data are compiled from references 3-6.

b See footnote 10.

Because buffering capacity is maximal at the pK, the compound cho-

sen should have a pK in the range of the desired pH. If the pH optimum of
an enzymic reaction is not known, it is desirable to use a related series of
buffers and measure reaction rates at several pH values; pH 6.0 (MES),
pH 7.0 (PIPES), pH 8.0 (EPPS), and pH 9.0 (TAPS or CHES) provide a
useful range. Once the pH optimum is approximated experimentally,
structurally unrelated buffers with a similar pK should be tested. Thus,
the use at pH 7.5 of TES (pK 7.5), phosphate (pK 7.2), and triethanol-
amine (pK 7.7) would serve to alert the investigator to interactions be-
tween buffer and reaction components.
The problem of specific buffer interaction with other reaction compo-
nents is most prevalent when inorganic buffers are used. Phosphate, for
example, inhibits many kinases and dehydrogenases as well as enzymes
with phosphate esters as substrates. Phosphate also inhibits carboxypep-
tidase, fumarase, and urease. Borate forms covalent complexes with g e m -
diols, including such important biological molecules as mono- and oligo-
saccharides, the ribose moieties of nucleic acids and pyridine nucleotides,
and such polyols as glycerol among other metabolic intermediates. Bicar-
[26] BUFFERS FOR ENZYMES 407

bonate, because it is in equilibrium with CO2, requires closed systems,

and since equilibration of CO2 and bicarbonate is slow, carbonic anhy-
drase is often added to such buffers. Buffers containing such primary
amines as Tris form Schiff bases with aldehydes and ketones.
Perhaps the most serious problems encountered when using inorganic,
primary amine or carboxylic acid buffers are their propensity for forming
coordination complexes with di- and trivalent metal ions. 7-9 Complex-
ation of a metal ion by the buffer results in proton release with consequent
decrease in pH. This is not as serious a problem as the formation of
insoluble precipitates upon complexation, or the chelation of a metal
required for enzymic activity (e.g., Mg2÷ for kinases, Cu 2÷ or Fe 2÷ for
hydroxylases). The Good buffers generally have low or known metal-
binding capabilities 3 and are favored for use in studies of metal-requiring
enzymes.
Once a noninteracting buffer of appropriate pK is found, a decision as
to the concentration to be used must be made. In general, the lowest
possible concentration of buffer should be used in order to avoid nonspe-
cific ionic strength effects. The most straightforward way to accomplish
this is to set up a reaction mixture with a low (10 mM) concentration of
buffer. The pH is checked before the addition of enzyme, and a suitable
period of time after addition of the enzyme, possibly the standard reaction
time or after equilibrium is reached. If a significant change in pH has
occurred (-+0.05 pH), a higher concentration of buffer must be used,
perhaps 20 or 50 mM. pH stability is particularly important in the mea-
surement of systems in which protons are generated or consumed stoi-
chiometrically with substrate utilization.
Stock buffer solutions should be prepared in glass containers (plastic
containers may leach UV-absorbing plasticizers) at a temperature close to
the working temperature. The pK, and therefore pH, of buffers, particu-
larly amine buffers, are sensitive to temperature (see Table I1°). Tris, for
example, when titrated to pH 8.06 at 25° will have a pH of 8.85 at 0°. The
Good buffers generally have only a small inverse dependence of pK on
temperature, whereas the pK of carboxylic acid buffers are even less
temperature sensitive. Since the dilution of a more concentrated stock
buffer solution will change the pH, as will the addition of salts, the pH of a
7 R. M. Smith and A. E. Martell, "Critical Stability Constants," Vol. 2. Plenum, New
York, 1975.
8 L. G. Sillen and A. E. Martell, "Stability Constants of Metal-Ion Complexes," Spec.
Publ.--Chem. Soc. n17 (1964).
9 L. G. Sillen and A. E. Martell, "Stability Constants of Metal-Ion Complexes. Supple-
ment" Spec. Publ.--Chem. Soc. n25 (1971).
~0The change in pK per degree centigrade. These values can be used to correct the pKs in
Table I to higher or lower temperatures.
408 RELATEDTECHNIQUES [26]

reaction should be checked, with all components added, by direct mea-

surement of the pH with a micro-combination electrode. By this means
guesswork is removed in the reporting of experimental conditions. Titra-
tion of the buffer at the approximate working concentration and tempera-
ture is an important consideration.
Since many enzyme rate studies rely on spectroscopic measurement
of substrates utilized or products formed, ideally the buffer used should
be transparent in the spectral region of interest. Although all the buffers in
Table I are nonabsorbing in the visible region, several have appreciable
absorption in the UV. ADA and ACES absorb strongly below 260 nm, but
most of the other Good buffers can be used at and above 240 nm. A series
of buffers that are transparent in the visible range and down to 240 nm
have been developed H for spectrophotometric determination of pKa.

Factors in Buffer Preparation

After selection of the buffer, it may be prepared by titrating the crys-
talline-free acid to the desired pH. The pH meter used should be cali-
brated with two or more commercially available pH standards that
bracket the desired pH region. If the effects of monovalent cations are
being investigated, mineral countercations can be avoided by titrating the
buffer with tetramethylammonium hydroxide. It may be desirable to filter
the titrated buffer through a sterile ultrafiltration device to prevent bacte-
rial or fungal growth, especially with solutions in the pH 6-8 range; this is
particularly important if large quantities are prepared for extended stor-
age. It may also be desirable to add a small amount of EDTA (10/zM) to
chelate heavy-metal ions which may be present.
The purification of proteins or enzymes requires large amounts of
buffered solutions for homogenization, chromatographic separations, and
dialysis. Cost, then, becomes a significant concern in buffer selection, and
the relatively inexpensive inorganic buffers and Tris are advantageous.
Tris, in particular, has been used extensively in enzyme purification,
although it has disadvantages: it is not a good buffer below pH 7.5, has a
highly temperature-dependent pK, and, since it is a primary amine, inter-
feres with the Bradford dye-binding protein assay/2 Glycine and other
primary amines interfere similarly in that assay whereas Bicine, HEPES,
and EPPS all give false-positive colors with the Lowry assay.
The choice of suitable buffers for use in protein purification will de-
pend on the pH required for maximal stability and resolution of the pro-

1, D. D. Perrin, Aust. J. Chem. 16, 572 (1963).

12 M. M. Bradford, Anal. Biochem. 22, 248 (1976).
[26] BUFFERS FOR ENZYMES 409

tein of interest, and the particular method employed. Although any buffer
is suitable for gel permeation chromatography, absorption and elution of
proteins from hydroxyapatite is generally performed in phosphate solu-
tions. Cationic buffers, e.g., Tris, are used in anion-exchange chromatog-
raphy, whereas cation-exchange chromatography requires the use of an-
ionic buffers such as phosphates. Buffer exchange, after chromatography,
is readily accomplished by passage of the protein solution through a
Sephadex G-10 or G-25 column equilibrated with the desired buffer.
Some buffers contain volatile components, thereby allowing their re-
moval by lyophilization. A number of such volatile buffer systems have
been described, 6 and are used extensively for preparative ion-exchange
chromatography.

Metal Ions
Specific reaction mixtures may require that mono-, di-, or trivalent
metal ions be completely absent from the reaction; in other work, a satu-
rating or intermediate concentration of such ions may be required. If a
specific monovalent cation requirement is being investigated, the buffer
should be titrated with tetramethylammoniumhydroxide. The Kmfor acti-
vation of the enzyme can be determined by subsequent addition of the
appropriate mineral cation salt to the assay system. If other enzymic
parameters of a mineral cation-stimulated enzyme are being investigated,
then a relatively high concentration of buffer (50-I00 mM) containing the
desired cation as counterion, e.g., K ÷ with pyruvate kinase, would be
used. This method generally ensures saturation with the monovalent cat-
ion and prevents the need for adding additional salts to the reaction mix-
ture with concomitant increase in ionic strength.
A number of assays require the control of di- and trivalent cations. If
such metals must be absent from the reaction, effective "removal" is
generally accomplished by addition of a chelating agent (0.1-5.0 mM). A
number of polydentate chelators serve this purpose, including citric acid,
tripolyphosphoric acid, nitrilotriacetic acid (NTA), ethylene glycol bis(fl-
aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA), ethylenedi-
aminetetraacetic acid (EDTA), and diethylenetriaminepentaacetic acid
(DTPA). The stability constants of these chelators with numerous metal
ions have been reported9,1° and are shown in Table II for several of bio-
chemical interest.
The maintenance of fixed concentrations of necessary metal ions is
more difficult. The above-mentioned polydentate chelators have been
used as metal ion buffers to ensure constant, but low, levels of free metal
ions in nutrient media. Metal buffering is particularly important when salt
410 RELATED TECHNIQUES [26]

T A B L E II
STABILITY CONSTANTSa OF METAL IoN-CHELATE COMPLEXES

Chelator

Metal ion Citrate PPPi NTA EGTA EDTA DTPA

Ca 2+ 3.6 5.2 6.5 11.0 10.6 10.6

Cd 2+ 3.8 6.6 10.0 16.1 16.6 19.1
Co 2+ 5.0 6.9 10.8 12.4 16.5 19.0
Cu 2+ 5.9 8.3 13.1 17.7 18.9 21.1
Fe 3+ 11.4 -- 15.9 20.5 25.1 28.6
Mg 2÷ 3.4 5.8 5.5 5.2 8.7 9.3
Mn 2+ 3.7 7.2 7.4 12.2 14.1 15.1
Ni 2÷ 5.4 6.8 11.5 12.7 18.7 20.2
Zn 2÷ 5.0 7.5 10.4 12.7 16.7 18.7

a Expressed as log K, where K = (ML,)/(M)(L)" for the reaction

M + nL ~ ML, where n = I. Data from Smith and Martell 7 and
Sillen and Martell. 8

solubility is low or when the free hydrated metal ion is susceptible to

hydrolysis, e.g., the reaction, M n÷ + H 2 0 ~ M(OH) ~n-l~÷ + H + . 13

Broad-Range Buffers
Attempts have been made to formulate buffers that can span pH re-
gions of 5 or more pH units. One method employs a mixture of buffers to
achieve a buffeting capacity that is approximately constant across the pH
region of interest. A second, more common, approach already noted is to
use a series of structurally related individual buffers, e.g., the Good
buffers, with pK values evenly spaced across the pH range. Each ap-
proach has limitations of which the investigator should be aware. First,
each component of the buffer mixture will be buffering optimally over a
relatively small pH region (pH -- pK -+ 1), outside of which it will play no
role. A specific ionized form of such a nonparticipatory buffer may be
inhibitory across the pH region in which that form is present. Further-
more, the presence of extraneous nonbuffering components will increase
the ionic strength of the solution with possible undesirable effects. A
thorough and detailed account of buffer mixtures that provide a wide

13 For an extensive list of pK values of hydrolysis, see Perrin and Dempsey. 6 For a thorough
discussion of metal-ion control in the study of metal-activated enzymes, in particular the
phosphotranferases, see J. M. Morrison, this series, Vol. 63, p. 257.
[26] BUFFERS FOR ENZYMES 411

range of buffering capacity with constant ionic strength is available 14and

has been used successfully.15
When using a series of individual buffers, it is important to overlap the
working pH with buffers of different pK values to ensure that specific
buffer interactions are absent. A hypothetical pH profile from pH 6-10
could be determined, for example, with the following series of buffers at
the indicated pH values: MES (pH 6.0-6.7), PIPES (pH 6.5-7.4), HEPES
(pH 7.2-8.2), TAPS (pH 8.0-9.0), CHES (pH 8.9-9.9), and CAPS (pH
9.8-10.0). This procedure has been found to be useful and results in
smooth pH profiles of kinetic parameters ~6,17without breaks indicative of
specific buffer effects.

Effect of Organic Solvents

Apparent pH values have been determined18,19for a large number of
buffers in various nonaqueous solvent mixtures. These studies have re-
sulted in the formulation of a pH* scale, formally equivalent to the pH
scale, for nonaqueous solvent mixtures. Simon2° reported pH* values of
over 1000 organic compounds in a number of organic solvents, including
Methyl Cellosolve, acetone, acetonitrile, dimethylformamide, and diox-
ane. For acid-base equilibria, pK* values for charge-generating ioniza-
tions (neutral acids) were found to be higher than the corresponding
pKH20 v a l u e s , whereas those in which a charge was not generated (cat-
ionic acids) had pK* values lower than pKH2o.
Similar effects of added organic solvents are expected for the ioniza-
tion behavior of functional groups on the enzymes that participate in
catalysis and binding. Thus, the addition of nonaqueous solvents should
elevate the pK of a neutral acid (carboxyl, sulfhydryl, tyrosyl, or metal-
coordinated water), but have little effect on the pK of a cationic acid
(amino or imidazole). The experimental protocol and the choice of suit-
able buffers for use in such experiments has been outlined,21 and has been
successfully applied to the identification of groups involved in cataly-

14 K. J. Ellis and J. F. Morrison, this series, Vol. 87, p. 405.

15 j. W. Williams and J. F. Morrison, Biochemistry 20, 6024 (1981).
16 p. F. Cook, G. L. Kenyon, and W. W. Cleland, Biochemistry 20, 1204 (1981).
17 R. E. Viola and W. W. Cleland, Biochemistry 17, 4111 (1978).
18 R. G. Bates, M. Paabo, and R. A. Robinson, J. Phys. Chem. 67, 1833 (1963).
19 C. L. deLigny, P. F. M. Luykx, M. Renbach, and A. A. Wiereke, Recl. Tray. Chim. Pays-
Bas 79, 699 (1960).
2o W. Simon, Angew. Chem. Int. Ed. Engl. 3, 661 (1964).
21 W. W. Cleland, Adv. Enzymol. Relat. Areas Mol. Biol. 45, 273 (1977).
412 RELATEDTECHNIQUES [26]
sis. 16'17 This technique, in conjunction with the measurement of the tem-
perature dependence of the pK of a functional group, should allow identifi-
cation of enzymic groups to be made.

Buffers at Low Temperature

Buffers suitable for the study of enzyme-catalyzed reactions at sub-
zero temperatures in mixed, organic cryosolvents have been investi-
gated, z2,23The pH dependence of nine buffers, spanning a range of pH 3-
10 as a function of cryosolvent mixture and temperature, has been
examined for which a linear dependence of pH* on 1/T (°K) was found. 22
By this means, it has been possible to predict the pH* of a buffered
solution in a specific cryosolvent at a specific subzero temperature. The
practical problems of low buffer solubility at low temperature and high
percentages of organic solvent remain to be solved. 24

Use of Deuterated Buffer

Deuterated buffer solutions are increasingly being used in the investi-
gation of deuterium isotope effects and proton NMR behavior of enzyme-
catalyzed reactions. The preparation of deuterated buffers requires the
titration of the free acid, dissolved in D20, with KOH(D) to the required
pD. Note that the pD of the buffer is 0.4 pH unit higher than the reading
on the pH meter. If highly enriched deuterated buffers (>95 atom % D)
are required, it is advisable to exchange the protons in both the buffer and
the base prior to titration by dissolving each in a small amount of D20,
and removing the solvent by rotary evaporation or lyophilization. Since
D20 solutions are hydroscopic, they should be stored in a tightly stop-
pered container, preferably in a desiccator (without desiccant).
Deuterated buffer solutions have been used for the enzymic synthesis
of stereospecifically deuterated compounds and in studies of either pri-
mary deuterium kinetic isotope effects or solvent isotope effects. In the
former case, D20 is usually a substrate in the reaction25,26 and solvent-
derived deuterium is thus incorporated into products to give isotope
effects, as shown for a number of other enzymes, z7 Deuterated buffer
solutions used for the study of isotope effects should be measured at

22 G. H. B. Hoa, P. Douzou, and A. M. Michelson, Biochim. Biophys. Acta 182, 334 (1969).
23 G. H. B. Hoa and P. Douzou, J. Biol. Chem. 248, 4649 (1973).
A. L. Fink and M. A. Geeves, this series, Vol. 63, p. 336.
25 j. S. Blanchard and W. W. Cleland, Biochemistry 19, 4506 (1980).
~6 T. Y. S. Shen and E. W. Westhead, Biochemistry 12, 3333 (1973).
27 T. B. Dougherty, V. R. Williams, and E. S. Younathan, Biochemistry 11, 2493 (1972).
[26] BUFFERS FOR ENZYMES 413

' ' 'I/' ' 2'

~ ~I'~ ,,I I,
1

P,P~S ,I]~

,oP~ I~ ,,l~i,
I,
TES

HEPES
,I I, , 1212I,
EPPS
2112,,
TAPS
II It

AES

CHES
I, 12
APS
I,
CAPS
II

ADA

TEA II,
TRIS
31 I,
I TRICINE tI 3

ii ' BICINE , I 2 ,
,7o',~5 ~' 60 s'o ' 40 ' 3'o ' 2b
PPM
FIG. 1. Natural abundance ~sC NMR line spectra of "good" buffers. The spectra were
obtained at 50 MHz, and contained 300 mM buffer (titrated to their respective pKs) with 50
mM dioxane (D, 67.4 ppm) as internal standard. Integrated intensities are shown next to the
upfield 13C lines; downfield carbonyl tsC resonances show reduced intensities due to longer
relaxation times.
414 RELATED TECHNIQUES [26]

several different mole fractions of deuterated solvents to ensure linearity

with deuterium content. Solvent isotope effects, 28 on the other hand, are
exhibited when deuterium from the solvent is not incorporated into
substrates or products, e.g., with the pyridine nucleotide dehydro-
genases. 29,3° Such effects may be small and may be due to true rate ef-
fects, different solvation effects in D20, or perturbations in enzymic pK
values when in deuterated solvents.

Buffers for N M R Studies

There has been increasing interest in NMR approaches to the study of
enzyme-catalyzed reactions, both in vitro and in vivo. The choice of a
specific buffer will depend on the experimental design and the nucleus
being used as probe, as well as factors considered earlier. Proton N M R
investigations are conducted in D20 and interference by buffer reso-
nances can be avoided by using inorganic buffers or those with only
exchangeable protons. However, such buffers are the same as those dis-
cussed above as being associated with a number of specific buffer effects,
and they are generally of limited usefulness. For NMR investigations of
19F and 31p, any of the buffers in Table I are suitable, since they do not
contain such nuclei.
13C N M R investigations are the most broadly applicable to enzyme-
catalyzed reactions and to studies of metabolism, 3~as they do not rely on
the presence of phosphorus or fluorine in the compound, or require the
use of highly enriched, deuterated solvent. Since ~3C NMR is so broadly
applicable to mechanistic and metabolic investigations, and since the
" G o o d " buffers are generally superior for enzyme work, the analysis of
the ~3C NMR chemical shifts of 18 " G o o d " buffers have been determined,
and are presented in Fig. I. These data may be valuable to investigators
planning to use ~3C NMR as a tool in the study of enzyme-catalyzed
reactions by allowing buffers to be chosen whose 13C resonances will not
interfere with the resonances of substrates, products, or the enzyme it-
self.

2s R. L. Schowen, Prog. Phys. Org. Chem. 9, 275 (1972).

29 j. Schmidt, J. Chen, M. DeTraglia, D. Minkel, and J. T. McFarland, J. Am. Chem. Soc.
101, 3634 (1979).
30 K. M. Welsh, D. J. Creighton, and J. P. Klinman, Biochemistry 19, 2005 (1980).
31 "Biological Applications of Magnetic Resonance" (R. G. Shulman, ed.). Academic Press,
New York, 1979.
[27] PROTEINASSAY 415

[27] P r o t e i n A s s a y for Dilute Solutions

By ENRICO CABIB and ITZHACK POLACHECK

Protein solutions of concentration lower than 25-50/zg/ml require a

concentration step before assay by Lowry's method, l In the procedure
described below, 2 trichloroacetic acid is used to concentrate proteins by
quantitative precipitation in the presence of yeast-soluble ribonucleic acid
as a carrier. Many substances that interfere with the colorimetric assay3
remain in solution and are thereby eliminated.

Assay Method
Principle. Protein is precipitatcd with 10% trichloroaccticacid, in the
presence of ribonucIeic acid as a carficr.The precipitatedprotein is dis-
solved in sodium hydroxide and measured by a slight modification of
Lowry's procedure.
Reagents
Yeast-soluble ribonucleic acid, 5 mg/ml. Each batch of ribonucleic
acid should be checked, and those producing a blank above 0.15, as
measured against water, are rejected. The product of Calbiochem-
Behring has been satisfactory
Trichloroacetic acid, 100 g per 100 ml of solution
Sodium hydroxide, 0.1 M
Sodium carbonate, 3% (w/v) in 0.1 M sodium hydroxide
Sodium tartrate, 4% (w/v)
Cupric sulfate (CuSO4 • 5 H20), 2% (w/v)
Lowry reagent: Mix 9.6 ml of sodium carbonate-sodium hydroxide
reagent with 0.2 ml each of sodium tartrate and copper sulfate
Folin reagent (phenol reagent solution 2 N; Fisher Scientific Com-
pany)
Bovine serum albumin, 1 mg/ml
Procedure. To 1 ml of sample, containing between 5 and 25 /~g of
protein, in a 10 × 75 mm Pyrex tube (DuPont, No. 00100), 25/zl of soluble
RNA is added, followed by 0.11 ml of trichloroacetic acid. The tube

O. H. L o w r y , N. J. R o s e b r o u g h , A. L. Farr, and R. J. Randall, J. Biol. Chem. 193, 265

(1951).
z I. P o l a c h e c k and E. Cabib, Anal. Biochem. 117, 311 (1981).
3 G. L. Peterson, Anal. Biochem. 100, 201 (1979).

Copyright © 1984by AcademicPress, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
416 RELATED TECHNIQUES [28]

contents are mixed by vortexing. After standing in ice for 45 min, the
tubes are centrifuged for l0 min at 27,000 g in a Sorvall RC2-B centrifuge
with an SS34 rotor. The supernatant fluid is aspirated, and the pellet is
suspended, by vortexing, in 0.5 ml of 0.1 M sodium hydroxide. If dissolu-
tion is not complete the tubes are placed in a boiling water bath for 5 min
and then cooled in tap water. After addition of 0.5 ml of Lowry reagent,
the tubes are incubated for l0 min at 37°. Folin reagent, 50/zl, is added
with immediate vortexing of each tube. After an additional 15 min at 37°,
the absorbance at 750 nm is measured. A series of standards containing
between 5 and 25/.tg of bovine serum albumin is processed and measured
at the same time.
Comments. The standards yield a straight line up to at least 25/zg of
protein, compared to the convex curve observed in this range with the
direct Lowry assay. The presence of several substances that interfere
with the Lowry procedure 2 [examples: 0.5% digitonin, 0.5% Zwittergent
3-08 or 3-16 (Calbiochem), 1% sodium dodecyl sulfate, or 0.05 M Tris-
chloride] does not alter the results. When Tris is present, however, it is
necessary to wash the trichloroacetic acid precipitate once with 1 ml of
cold 10% trichloroacetic acid. When 0.1% Triton X-100 is present in the
sample, absorbance values are about 55% of the normal.
Similar procedures have been published 4,5 in which deoxycholate is
used as a coprecipitating carrier. Deoxycholate, however, does not yield
a precipitate with trichloroacetic acid in the presence of certain deter-
gents, such as sodium dodecyl sulfate, digitonin, and sulfobetaines.

4 A. Bensadoun and D. Weinstein, Anal. Biochem. 70, 241 (1976).

G. L. Peterson, Anal. Biochem. 83, 346 (1977).

[28] E n z y m e L o c a l i z a t i o n in G e l s
By MARY J. HEEB and OTHMAR GABRIEL

This chapter is an update of an earlier one in this series I and provides

general information for the location of functionally active enzymes after
separation by electrophoresis or electrofocusing in various anticonvec-
tion media such as agarose, starch, or polyacrylamide gels. The scope of
this chapter is limited to illustrating recent developments in the detection
i O. Gabriel, this series, Vol. 22, p. 578.

Copyright © 1984by AcademicPress, Inc.

METHODS IN ENZYMOLOGY,VOL. 104 All fightsof reproductionin any formreserved.
ISBN 0-12-182004-1
416 RELATED TECHNIQUES [28]

contents are mixed by vortexing. After standing in ice for 45 min, the
tubes are centrifuged for l0 min at 27,000 g in a Sorvall RC2-B centrifuge
with an SS34 rotor. The supernatant fluid is aspirated, and the pellet is
suspended, by vortexing, in 0.5 ml of 0.1 M sodium hydroxide. If dissolu-
tion is not complete the tubes are placed in a boiling water bath for 5 min
and then cooled in tap water. After addition of 0.5 ml of Lowry reagent,
the tubes are incubated for l0 min at 37°. Folin reagent, 50/zl, is added
with immediate vortexing of each tube. After an additional 15 min at 37°,
the absorbance at 750 nm is measured. A series of standards containing
between 5 and 25/.tg of bovine serum albumin is processed and measured
at the same time.
Comments. The standards yield a straight line up to at least 25/zg of
protein, compared to the convex curve observed in this range with the
direct Lowry assay. The presence of several substances that interfere
with the Lowry procedure 2 [examples: 0.5% digitonin, 0.5% Zwittergent
3-08 or 3-16 (Calbiochem), 1% sodium dodecyl sulfate, or 0.05 M Tris-
chloride] does not alter the results. When Tris is present, however, it is
necessary to wash the trichloroacetic acid precipitate once with 1 ml of
cold 10% trichloroacetic acid. When 0.1% Triton X-100 is present in the
sample, absorbance values are about 55% of the normal.
Similar procedures have been published 4,5 in which deoxycholate is
used as a coprecipitating carrier. Deoxycholate, however, does not yield
a precipitate with trichloroacetic acid in the presence of certain deter-
gents, such as sodium dodecyl sulfate, digitonin, and sulfobetaines.

4 A. Bensadoun and D. Weinstein, Anal. Biochem. 70, 241 (1976).

G. L. Peterson, Anal. Biochem. 83, 346 (1977).

[28] E n z y m e L o c a l i z a t i o n in G e l s
By MARY J. HEEB and OTHMAR GABRIEL

This chapter is an update of an earlier one in this series I and provides

general information for the location of functionally active enzymes after
separation by electrophoresis or electrofocusing in various anticonvec-
tion media such as agarose, starch, or polyacrylamide gels. The scope of
this chapter is limited to illustrating recent developments in the detection
i O. Gabriel, this series, Vol. 22, p. 578.

Copyright © 1984by AcademicPress, Inc.

METHODS IN ENZYMOLOGY,VOL. 104 All fightsof reproductionin any formreserved.
ISBN 0-12-182004-1
[28] ENZYME LOCALIZATION IN GELS 417

of enzymes in situ as perceived by the authors. Although no attempt is

made to provide a comprehensive review of the literature, many different
enzymes are covered in detail and references to others are presented.
Several publications contain valuable information on the localization of
enzymes, including reviews that provide extensive lists of reagents and
methods for enzyme detection. 2,3 A list of references for a wide variety of
enzymes 4 and a general review5 have been published.
Some of the more recent trends in localizing enzymes on gels include
the use of renaturing techniques, as applied to enzymes that have been
subjected to electrophoresis in detergent-containing gels. 6 A number of
potentially carcinogenic reagents, such as benzidine and o-dianisidine,
have been replaced in some assay systems with eugenol or tetrabase. 7,8
Fluorophores attached to substrates (e.g., umbelliferyl or dansyl deriva-
tives) have been used extensively, providing sensitive methods whereby
the progress of reaction can be observed directly. Ultrathin agar overlays
on polyester sheets have been used for more sensitive sandwich-type
incubations 9 and are particularly useful when applied after ultrathin-layer
isoelectric focusing. A procedure has been described by which aliquots of
a large number of column fractions may be subjected to electrophoresis at
the same time and then stained for several different enzymes.~°

Retention of Enzymic Activity

For any of the enzyme detection methods that will be discussed, a

major consideration should be the preservation of the maximum level of
enzymic activity throughout the process. Special attention must be paid
to the possible detrimental effect of reagents used in preparation of the
gel: oxidizing agents (ammonium persulfate), incomplete reaction of
polymerizing agents, buffer ions, and a pH incompatible with the enzyme.
Appropriate precautions should be taken to limit inactivation of the en-

2 M. J. Siciliano and C. R. Shaw, in "Chromatographic and Electrophoretic Techniques" (I.

Smith, ed.), Voi. 2, p. 185. Heinemann, London, 1976.
3 H. Harris and D. A. Hopkinson, "Handbook of Enzyme Electrophoresis in Human
Genetics." North-Holland Publ., Amsterdam, 1976.
4 B. D. Hames and D. Rickwood, "Gel Electrophoresis of Proteins: A Practical Ap-
proach," IRL Press, London and Washington D.C., 1981.
5 W. Ostrowski, J. Chromatogr. Libr. 1811, 287 (1983).
6 R. E. Manrow and R. P. Dottin, Anal. Biochem. 120, 181 (1982).
7 E. H. Liu and D. M. Gibson, Anal. Biochem. 79, 597 (1977).
a B. Lomholt, Anal. Biochem. 65, 569 (1975).
9 M. Hofelmann, R. Kittsteiner-Ebede, and P. Schrier, Anal. Biochem. 128, 217 (1983).
10 p. H. Odense, C. Annand, and J. Barlow, Anal. Biochem. 108, 257 (1980).
418 RELATED TECHNIQUES [28]

zyme. For example, prior electrophoresis of the gel alone, II followed by

tests for the absence of persulfate ions, 12 and inclusion of sulfhydryl
group-containing components into buffer systems as protection against
oxidation 13were reported in many instances to result in retention of other-
wise labile activity. Direct addition of sulfhydryl reagents to polymeriza-
tion mixtures, however, is not useful, since polymerization will be pre-
vented or delayed.
With isoelectric focusing gels, it may be necessary to remove the
ampholytes at the end of the procedure by soaking in an appropriate
buffer prior to application of enzyme detection methods.14 If the ampho-
lytes or the pH of the gel do not interfere with enzyme detection, omission
of this step will reduce diffusion of bands.
A number of monomeric enzymes and a few enzymes containing iden-
tical subunits have been renatured after electrophoresis in sodium dode-
cyl sulfate (SDS)-containing gels 15; the extent of renaturation often de-
pends on the commercial source 16of the SDS. 17In some cases it has been
sufficient to soak gels in the appropriate buffer, but for enzymes that are
tolerant of the procedure, SDS can be removed more quickly, with
greater recovery of enzymic activity, by incubation in buffered 25% iso-
propano118 (see example 3 in the last section of this chapter). A combina-
tion of nonionic detergent, high salt, glycerol, substrate, and a sulfhydryl
reducing agent has been used successfully to renature a number of en-
zymes in SDS gels or O'Farrell gels 6 (see example 4 in the last section of
this chapter). Another approach is to subject an SDS gel to isoelectric
focusing in urea and nonionic detergent in a second dimension; since SDS
is removed in the second dimension, many proteins can be renatured from
urea.19.2° Alternatively, for some separations the isoelectric focusing step
might be sufficient in a single dimension. An advantage to renaturing
enzymes within gels is that part of the activity may be physically bound to
the gel. 15

n O. Gabriel, this series, Vol. 22, p. 565.

12 A. Bennick, Anal. Biochem. 26, 453 (1968).
13 j. M. Brewer, Science (Washington, D.C.) 156, 256 (1967).
14 H. Mukasa, A. Shimamura, and H. Tsumori, Anal. Biochem. 123, 276 (1982).
15 S. A. Lacks and S. S. Springhorn, J. Biol. Chem. 255, 7467 (1980).
16 Inhibition of renaturation is correlated with the amount of hexadexyl sulfate and tetrade-
cyl sulfate present as contaminants.
17 S. A. Lacks, S. S. Springhorn, A. L. Rosenthal, Anal. Biochem. 100, 357 (1979).
is A. Blank, R. H. Sugiyama, and C. A. Dekker, Anal. Biochem. 120, 267 (1982).
19 R. E. Manrow and R. P. Dottin, Proc. Natl. Acad. Sci. U.S.A. 77, 730 (1980).
2o R. P. Dottin, R. E. Manrow, B. R. Fishel, S. L. Aukerman, and J. L. Culleton, this series,
Vol. 68, p. 513.
[28] ENZYME LOCALIZATION IN GELS 419

Localization of Enzymes
It is essential to obtain as much information as possible about the
properties of the enzyme before attempts are made to locate it in gels. The
localization must follow separation as quickly as possible, and several
approaches can be tried.
Elution of Enzyme from the Gel. Cutting of the gel into segments
followed by elution with appropriate buffer and assay of the eluted en-
zyme by conventional methods is tedious and time consuming. Neverthe-
less, much valuable information is gained concerning the properties of the
enzyme and its chances of remaining enzymically active during the sepa-
ration and elution procedure. Elution of active enzyme will depend not
only on prevention of protein denaturation, but also on recovery from the
gel. Recovery depends in part on the molecular weight of the enzyme
protein and the extent of cross-linkage used for the electrophoretic sepa-
ration. It should be kept in mind that individual bands of separated com-
ponents contain only microgram quantities of protein so that stabilization
of such minute amounts, once eluted, may be necessary. Addition of
carrier proteins or of polyhydroxyalcohols, such as glycerol, can be effec-
tive.
Staining for Protein prior to Elution. Most of the commonly used
protein stains result in irreversible denaturation of enzymes, and it is not
wise to attempt to detect enzyme activity after using a protein stain.
However, the detection of proteins under relatively nondenaturing condi-
tions such as UV-scanning, limited tannic acid treatment, 2~ or binding of
fluorophore, 1 followed by elution and localization of enzymic activity,
represents a compromise for detection of protein as well as enzymic
activity. The number of gel slices that must be eluted and assayed can be
greatly reduced by this approach.
Staining for Enzymic Activity in Situ after Electrophoretic Separation.
The general principles for the localization of enzymes in situ can be di-
vided into several major groups according to the methods and type of
reagents involved. Most, if not all, of these techniques were originally
described by histochemists.
1. Simultaneous capture. This method is widely used and is applied to
enzymes that remain active in the presence of the staining reagents. The
substrate is converted by the enzyme to a product that, in turn, couples
with a reagent present in the incubation mixture to form an insoluble
colored product.

21 K. Aoki, S. Kajiwara, R. Shinke, and H. Nishira, Anal. Biochem. 95, 575 (1979).
420 RELATEDTECHNIQUES [28]

2. Postincubation coupling. Incubation of substrate with enzyme

results in a product, and this step is followed by addition of reagent to
yield colored material. Diffusion of product during the first incubation
period results in broadening of bands.
3. Autochromic methods. Direct visualization of the progress of reac-
tion is made possible when there are changes in the optical properties of
either substrate or product. An advantage of the method is direct observa-
tion of the progress of the reaction as well as allowing an estimate of the
extent of diffusion during the incubation period.
4. Sandwich-type incubation. The method uses auxiliary indicator en-
zymes or high-molecular-weight substrates that require the overlay of the
gel with another gel, paper, or other suitable matrix containing the neces-
sary components. Incubation of the separating gel with the "indicator
matrix" permits localization of enzymic activity.
5. Copolymerization of substrate in the gel. This method is possible
and desirable for certain high-molecular-weight substrates, such as nu-
cleic acid, starch, gelatin, or pectin. It is often necessary to prevent the
enzyme from acting on the substrate during electrophoresis by deletion of
a cofactor, inclusion of an inhibitor or chelating agent, use of a less than
optimal pH for enzymic activity, or other means. Conditions can be re-
stored to near optimal during the subsequent incubation period in order to
detect bands of enzyme activity.
For most of the general methods just described, it is advisable to mark
the position of the enzyme with a small piece of fine wire, or other means,
at the earliest time at which bands are detectable and the least diffuse.
The enzyme detection step can then be followed with a general protein
stain.

Limitations and Problems Encountered with in Situ

Localization of Enzymes
Generally, diffusion of both the enzyme and its reaction products will
cause reuniting and broadening of separated components. It is imperative
to minimize these factors by variation of incubation and reaction condi-
tions to optimize staining while minimizing diffusion. Considering the
complex kinetic events in a gel during staining, it requires extensive ex-
perimental evidence to establish the rate-limiting factors in a specific
staining procedure. For most systems, it is unrealistic to expect that
quantitative conclusions can be derived from the staining intensity, and
inx,estigators should use extreme caution in interpreting densitometric
data from enzyme activity stains in a quantitative manner.
[28] ENZYME LOCALIZATION IN GELS 421

A common problem encountered with in situ staining techniques con-

cerns the issue of specificity of the procedure. It is important to establish
the same pattern of substrate specificity for the enzyme in a gel as
that established using conventional solution methods. A number of
control incubations (absence of substrate, presence of specific inhibitors)
will eliminate artifactual staining, such as the "nothing dehydro-
genases." 1,22,23

Locating Specific Enzymes

A number of enzymes are presented in tabular form with appropriate
references to guide the reader to recent literature that provides detailed
description for localizing enzymes in situ. A division into major groups of
enzymes, such as oxidoreductases, transferases, and the like, was
thought to be useful since analogous functional enzyme properties will
lend themselves to the use of similar techniques for detection. The meth-
ods selected for inclusion are primarily those published in readily avail-
able journals since 1974 and are ones that are generally reproducible
under standard laboratory conditions. A similar compilation in Vol. 22 of
this series I should be consulted for methods applied to a number of other
enzymes. The methods are selected to demonstrate the principles for
detecting specific enzymes but can, with slight modifications, be applied
to the localization of similar enzymes. The included methods are based on
the detection of enzyme activity; immunological techniques are excluded.
Many valuable contributions are not included in the chapter, and there is
no intent to be totally comprehensive.
Abbreviations used in Tables I-V are phenazine methosulfate, PMS;
nitro blue tetrazolium, NBT; 3-(4,5-dimethylthiazoyl-2)-2,5-diphenyl-
tetrazolium bromide, MTT; p-iodonitrotetrazolium violet, INT; 2,3,5-
triphenyltetrazolium chloride, TTC; and 4-methylumbelliferyl, 4-MU.

Oxidoreductases (Table I)
Many enzymes in this group use NAD or NADP as a cosubstrate.
Oxidation of one substrate, then, results in formation of a reduced pyri-
dine nucleotide. The transfer of reducing power to various tetrazolium
dyes can be mediated, in turn, by phenazine methosulfate to yield a
deeply colored, insoluble formazan. The advantage of the technique is
that only a single incubation containing all components is required with
J. L. O'Conner, D. P. Edwards, and E. D. Bransome, Anal. Biochem. 78, 205 (1977).
23 p. H. Springell and T. A. Lynch, Anal. Biochem. 74, 251 (1976).
422 RELATED TECHNIQUES [28]

~ q~l " ~ ~

. . . . . . . . . . . . . . . ~ .~ !~ . . . .

II) ~1
£.1 ca ,1~

+ ~ ~
0

~ =a~ + ~ Zo o . _

Z Z Z . o = ~ =

= =i~= ~~ ~.,,_~

¢0

0 0 ~ ~ 0 ~ 0 ~''~ ~ ' , . t = ¢~
ua~
•~ ".~ ..~ .,~, .,..,
m

X
©
¢}

,.1=
N =I
•o g

a,, " 0

6
Z
~a "2.
[28] ENZYME LOCALIZATION IN GELS 423

u .~

.>_, c~ u
,-:- ,,-A
e',l u'~

0
oo
(-q
N

,.o

0
,.¢=
o.E
E
...~ ,.13
:'~ "1::

.~ ,-~

• ,~ ,a~

• . ~ ~ ~ ~ ~,~l~O ~ ~'~ t ' ~ ~" ,

~'~ d ~ . ~ • ~ ~ ~t.-~:

~ - ~ .~. o ~'~ ~ ~-~

.~

• "~,.~ ~-~

oo E.m = ~ ~: o = ~ ~-.. ~ ~ ..m.o o~ G~

0"~

¢'N
0 ;~ ¢: . . .

• .
424 RELATED TECHNIQUES [28]

the formation of an insoluble product that is not subject to diffusion. The

disadvantage of the method is sensitivity to light and oxygen 24 as well as
the necessity of carefully controlled reaction conditions and of appropri-
ate controls (such as omission of substrate) to avoid erroneous interpreta-
tion. t,22,23

Transferases (Table II)

Transferases catalyze the transfer of a group (such as methyl or glyco-
syl) from one compound to another. These enzymes are frequently de-
tected by coupling the primary reaction product with auxiliary enzyme(s)
which, in turn, can be coupled to tetrazolium dyes or other indicators.
Consideration must be given to the slow penetration of the coupling en-
zymes into the gel. For this reason, many of the indicator systems de-
scribed are applied peripherally in agar or paper overlays in a sandwich
type of incubation. A device for recovering expensive coupling enzymes
has been described. 25 Some of the more recent methods for the detection
of transferases use different principles. For example, 1-chloro-2,4-dinitro-
benzene is used as an acceptor for glutathione with glutathione S-trans-
ferase26; phosphorylase b is copolymerized in the separating gel as a
coupling enzyme for the detection of oligo-1,4 --~ 1,6-glucosyltransferase
(branching enzyme)27; and precipitation of pyrophosphate with manga-
nese ion is used for detection of hypoxanthine-guanine phosphoribo-
syltransferasefl 8

Hydrolases (Table III)

Hydrolases catalyze the addition of water to various bonds, resulting
in cleavage. They compose a large group of enzymes, many of which are
very stable and, therefore, amenable to a wide variety of detection meth-
ods. Several techniques use substrates such as umbelliferylfl9,3° p-ni-
trophenyl, 31 or other chromogenic derivatives that change their optical
properties as a consequence of the enzymic reaction. Enzymes that re-
lease inorganic phosphate, pyrophosphate, or carbon dioxide as products
can be visualized by precipitation with Ca 2÷ ions. 32 Enzymes that are

W. Worsfold, M. J. Marshall, and E. B. Ellis, Anal. Biochem. 79, 152 (1977).

T. Yamashita and H. Utoh, Anal. Biochem. 84, 304 (1978).
26 p. G. Board, Anal. Biochem. 105, 147 (1980).
27 K. Sato and K. Sato, Anal. Biochem. 108, 16 (1980).
28 B. Vasquez and A. L. Bieber, Anal. Biochem. 84, 504 (1978).
P. L. Chang, S. R. Ballantyne, and R. G. Davidson, Anal. Biochem. 97, 36 (1979).
30 p. M. Coates, M. A. Mestriner, and D. A. Hopkinson, Ann. Hum. Genet. 39, 1 (1975).
31 M. E. Hodes, M. Crisp, and E. Gelb, Anal. Biochem. $0, 239 (1977).
3z H. G. Nimmo and G. A. Nimmo, Anal. Biochem. 121, 17 (1982).
[28] ENZYME LOCALIZATION IN GELS 425

~o

:= .a- ..~
i=. .. ~,=

~a

°~

.~ ~ . ~ ~a'rD o ~ * ~ o .~

~'- ~=~ Z= ~
.g <

ID
~2 "v:l

o " " I I

O
e~
:a ,~ ~

,A

~ ~i~ ~ i
~ "2.
e4
e,l
~. ~ ,4." ,e.
426 RELATED TECHNIQUES [28]

o~ .,.2

nl
_= . g :~
== = , z 5~e

q
~a
i b~
m

~g 8Z a =o. o

-~,
I I
ca ,= o N
N m,.~
0

.a
<
b.

"7' .

~2
[.. = "= 8 "~
"gr, ~ "

< Z ~ D

g
g ~
= "~.
E
0
•g =
= o e
to <

t.: r.:
I~ t'~
e4 e4 e.i e-i ~4 e.i e.i
[28] ENZYME LOCALIZATION IN GELS 427

=,:
.~2
o • 0

~c 6.

0
=] >-
02
"10 .r,

• -~ r-- ~ ~o,~
oo 0.Q,
O~ ~ ~ O~ ~ ~O'~...J ~ "-w~

. ~ ~ ~ c~ ~ ~ ~ ,,d o

• I~ - ~ ~'~ ~ ~ ~ ¢~ ~ ~ , ~ ; ,~, ,.~

~" . o,0 ......, ".-." ,.~.: w,~,Qg'~ ~ ~ ~ .,w ~ 0

--..~.~ ~,~ ~ ~ ~ ,..-.,~ .,~ ,='~ ~ . ~ ~,.-.

~g~z~.~

~ . ~ ~

go ~..~a.~ ~ . ~ ~ ~ = ~ ~-~ . o . ~ .
~ ~ z ~ ~ ~:~ > ~ ~ . ~ ~ ~ ~:~ ~ ~. ~ ~.
428 RELATED TECHNIQUES [28]

~ " ~ .~.~ ~i ~. ~ ~:~

:~ .~ :" .~ :" :" ~ :" :~ ~ .. .. :" :~ ~ :" ~ :" :~ ~ :-~ : " :" .-- :" O

¢,~ e-

=02~
"n .~-= :~,o .~ ~
"u .=
-

0 •~ ~: g ~
+ ~:~.~ ~, " ~ #~ ,-~ ~ ~ ¢~
~ ~'~'~ ~" ¢~ o ~ "-~

-g
o
~ oN
>
.~ .~" ~

r~

oe.

Z
 [28] ENZYME LOCALIZATION IN GELS 429

6 ~

..~ o o o a: .~ ~ ..

~..~= ~ .o ~ r ~ ~ ~ ~ = o =
.~. ~ ~- ~ .~ ~ .~ .~.~
.,~ ~
..~'~ .-~ ~ ~ ~ ~ =~ ~ ..
0
~ o o o~ o~ - o

6
,, ~ . ' ~ <
~ - ,
e~
O
O

E'~- ~.~

<, e a 1 _ = ~
~~ ~
4 ~ m N ~

.-~

o
0 ,m
o

~ No) [-

~D

e-i e4 ~ ~." .4,4

,,4
 430 RELATED TECHNIQUES [28]

e,,a
o >
c-

o
e~
[-
o
~o
~d
0

<
==
>

~o
~a

N~, N~
.r.
~a
~ ~=. ~ z ~ z < ~_~ ~
•
~- ~ ~ ; ~"
~
.
~: 0
ell ~ '~"

¢0 .~ ~. . - ~ - ~ ~ .~oo~ •

<

.~=.= •~ o~ ~
.g,,a <-= ~ ~ ~
.= .~ --.~ ~ . ~ _~--~ ~.-~

~ ~-~'~ ~ . ~ ~ - ~ • .~,

g
~ ~ 0 '~ ~ • ~ "

~.~:>~4~:/~ ~'~ ,.~ ~ .

~a
.. ~ ~

~ 0 0 , . . •

Z
¢o
~a
 [28] ENZYME LOCALIZATION IN GELS 431

E
.E .=2
0
m

o
0

I"..

o
,-Z

E
,-k ,.-Z r ~

t'-.

<
E

,-A
• t'q

.~® ~2~.~.~ ~ ~
¢.I oo ,,~ . ~ ~ ~
o~ ~ ~ ~ .~ ~ ~,,~ b~ , . - . ~
"-'" ~ Z ~g'~" ,o~
r'-- ~ .~ "~ .~ v,,,I ~ .-~

~ ~ 0 • ._ ~
.,~
~i = -~ ,.x.~,.~ .~
i _ = o~ ~.~,, ~ ~-~ ~ ~ ~ ~ " ~ ~,~ • ~ =. . ,~..,~

~I ~ o ~.'~ ~ ~l'~r-~.~. ~'~ ~'~

~ . ~ ~ ~ ~r~ ~ •

,~<~
~e ~ .=~. .~ ~ .~ ~ ~ ,,~.~
~..,..~ ~-~ ~ ~ . .~ ~

~'~ ~ . ~ , ~ .~fi ~ • . o
432 RELATED TECHNIQUES [28]

..~ ~ ~ ,,.q --~'-'~ ~ ~ ~

0
,.1=
•.= :-~ .- ."a :.~ .. :=., i~ .... :=

•~ ~
e~
~'~ .~ .~

0
ca

.. ~ -~'= ~. ~=
o ~. .~=~ .~'~.~'
e~

='~ e'~eq
._ ~ =o ~8 .~

ua~
,-1
m~
.a

.g

t~

. . . . e-i e,,i r-,i

44 "4 4 ,4-/,4 "4 ,4,~- ,4
[28] ENZYME LOCALIZATION IN GELS 433

6
t~

,a.

~ ' ~ ~.
J::

m
._~..~ .~
• I:1

r,.)~
qD ..-.. ~ l"-. ~ ~ ~v.~ . ~ I.~. t/~ ~

•~. .~~

,.~ .~_ ~ . ~ ~ . 1 ~ o ~,

.~° ~ ~ . ~ . ~ ~. ~ ~ ~ .~ ~. ~
.~'~
• ..~ @

_~-,
o . .~ . ..~ .~ . . ~..
434 RELATED TECHNIQUES [')8]

involved in ammonia metabolism can be coupled with glutamate dehydro-

genase, a pyridine nucleotide-requiring reaction. 33 Hydrolases such as
nucleases, 17,34proteases,15 lipase and amylase,15 have been detected after
renaturation in detergent-containing gels.

Lyases (Table IV)

Lyases catalyze the cleavage of bonds by means other than hydroly-
sis. In most cases they are detected on gels by reason of the chemical
properties of the product of the enzymic reaction that is coupled with a
chromogenic reagent.

Isomerases and Ligases (Table V)

Since only a few examples of detection methods for these classes are
available, both groups of enzymes are presented in the same table. Isom-
erases catalyze reactions resulting in a molecular rearrangement. They
are usually detected on gels by coupling with an enzyme that requires a
pyridine nucleotide. Ligases catalyze the joining of two molecules with
concomitant hydrolysis of a nucleoside triphosphate.

Selected Examples of Detailed Detection Methods

For many detection methods, economy of reagents may be desirable.
At least 50 ml of reagent is generally needed to immerse a slab gel, but it is
sufficient to wet the cut surface with as little as 6 ml of reagent. Tube gels
can be immersed in reagent in test tubes of narrow diameter. For sand-
wich-type incubations, ultrathin layer techniques have been described. 9
Example 1. ct-Glycerophosphate Dehydrogenase (EC 1.1.1.8, Glyc-
erol-3-phosphate Dehydrogenase). Simultaneous Capture Technique (i).
The method 35 illustrates the use of tetrazolium salts for the detection of
enzymes that produce reduced pyridine nucleotides. Similar techniques
can be used for other dehydrogenases or for enzymes that can be coupled
to dehydrogenases. The reaction must be carried out in the dark, with
careful control of pH and temperature. Reagents are prepared just before
use. As previously noted, control incubations without substrate are es-
sential. If the stained gel is to be preserved, excess reagent must be
removed to decrease background discoloration.
For a horizontal starch gel, 50 ml of staining reagent contains 25 mg of
NAD, 15 mg of nitro blue tetrazolium (NBT), 1 mg of phenazine metho-

33 R. L. Nelson, S. Povey, D. A. Hopkinson, and H. Harris, Biochem. Genet. 15, 1023

(1977).
34 j. Huet, A. Sentenac, and P. Fromageot, FEBS Lett. 94, 28 (1978).
35 F. Leibenguth, Biochem. Genet. 13, 263 (1974).
[28] ENZYME LOCALIZATION IN GELS 435

,d

O
,.3
-b

,..r.

e~ E
o
~ z "~ d o

od

>
u3
.1
,.3
<
<

r..)
m

r/l

ID

~9
,.~ .

b.

6
Z M

04 ,~
u3
436 RELATED TECHNIQUES [28]

sulfate, 5 ml of 1 M sodium a-glycerophosphate at pH 7, 10 ml of 0.2 M

Tris-HCl at pH 8.0, and 35 ml of water. Pour the solution over the cut
surface of the gel and incubate in the dark at 37 ° until dark blue bands
appear. Rinse in water and fix in ethanol-acetic acid-glycerol-water
(5:2:1:4).
Example 2. Argininosuccinase (EC 4.3.2.1, Argininosuccinate Lyase).
Simultaneous Capture Technique (i). The method 33 illustrates the use of
coupling enzymes that require pyridine nucleotides as cosubstrates and
also the use of a negative stain. As coupling enzyme, glutamate dehydro-
genase is used, an enzyme that is equally useful for the detection of other
enzymes that catalyze the production of ammonia, or that can be coupled
to such enzymes, e.g., cytidine deaminase, adenosine deaminase, adeno-
sine monophosphate deaminase, arginase, D-amino acid oxidase, D-aspar-
tate oxidase, and urease.
The stain solution consists of 50 mg of barium arginosuccinate, 1 mg
(40 units) of arginase (Sigma), 2 mg (20 units) of urease (Sigma type VI),
25 mg of a-ketoglutarate, 10 mg of NADH, and 50/zl of glutamate dehy-
drogenase (Sigma, 500 units/ml) in 5 ml of 0.1 M Tris-HCl at pH 7.6. The
solution is applied to a Whatman 3 MM paper overlay on the cut surface
of a starch slab gel, incubated at 37°, and is examined for nonfluorescent,
i.e., quenched, zones under an UV lamp. The paper is photographed
before the zones become diffuse. Glutamate dehydrogenase and lactate
dehydrogenase bands may appear in some electrophoresed extracts.
These bands are seen on the gel before they appear on the paper, in
contrast to arginosuccinase bands. Bands from glutamate dehydrogenase
can also be located by spraying ammonia on the paper just after the
primary stain has developed. The paper will fade rapidly, but bands of this
enzyme will be visible on the gel within 5 min.
Example 3. Nucleolytic Enzymes (EC 3.1). Postincubation Capture (ii)
and Copolymerization of Substrate in the Running Gel (v). One type of
renaturing technique is described in the following example.18 Separating
SDS slab gels are prepared by copolymerizing either rRNA (about 7 A260
units/ml or 0.3 mg/ml) or DNA (0.3 mg/ml) in the gels. Samples are treated
by heating for 2 min at 100° in 2% SDS, 10% glycerol, 0.005% bro-
mophenol blue, and 0.0625 M Tris-HCl at pH 6.8. For detection of RNase
after electrophoresis, the gels are first incubated in 0.25 liter of 0.01 M
Tris-HC1 at pH 7.4, containing 25% isopropanol, for 30 min, with one
change after 15 min, followed by incubation in buffer alone for 60 min with
two changes at 20-min intervals. For digestion of RNA, the gel is incu-
bated for 90 min at 37° in 0.1 M Tris-HC1 at pH 7.4. For staining, the gel is
incubated for 10 min in 0.01 M Tris-HC1, followed by another 10 min in
150 ml of 0.2% toluidine blue in the same buffer. The gel is washed for 60
[28] ENZYME LOCALIZATION IN GELS 437

rain in the Tris buffer, with changes at 10 and 30 min, and is then ready for
photography. All incubations are performed on a gently rotating shaker.
For detection of DNase, the time of the first two incubations is increased
to 60 min, with changes each 20 min. When the purpose is digestion,
incubation is increased to 14 hr, and 10 mM MgC12 and 5 mM CaCl2 are
included in the buffer.
The use of isopropanol in the renaturing buffer provides a more effec-
tive means for removal of SDS and allows greater recovery of nucleolytic
enzyme activity. Variations in recovery of activity due to the use of
different SDS preparations are minimized in many instances.16 It is rec-
ommended that the optimal isopropanol concentration and exposure time
for a particular enzyme, with a given source of SDS, be individually
adjusted.
Example 4. UDPGpyrophosphorylase (EC 2.7.7.9, Glucose-l-phos-
phate Uridylyltransferase). Simultaneous Capture (i). An alternative ap-
proach to renaturation is described in the following technique, 6 which was
applied to both SDS and O'Farrell gels (isoelectric focusing in urea in the
first direction followed by SDS gels in a second dimension). The success
of the method relies on a renaturing buffer containing nonionic detergent,
high salt, glycerol, substrate, and in many cases, sulfhydryl reducing
agents. When sulfhydryl reagents are used, they are omitted from the final
buffer changes to avoid interference with enzyme staining reagents. The
method has been applied with variations to adenylate kinase and to the
homomeric enzymes alkaline phosphatase and creatine kinase. It was not
possible to renature lactate dehydrogenase in SDS gels by this technique,
although this enzyme, as well as alcohol dehydrogenase, was successfully
renatured after treatment with SDS followed by isoelectric focusing in
urea gel. 2°
After electrophoresis, SDS gels or O'Farrell gels containing UDPG
pyrophosphorylase are incubated in 25 mM Tricine-HC1 at pH 7.5, con-
taining 10% glycerol, 200 m M NaCl, 1 mM uridine, and 2.5% Nonidet
P-40. Tube gels are incubated at room temperature for 4 hr with 4
changes of I0 volumes each; slab gels are incubated for 2.5 hr with 5
changes of 5 volumes each. The gels are then stained in a solution contain-
ing 0.08 unit/ml phosphoglucomutase, 0.35 unit/ml glucose-6-phosphate
dehydrogenase, 1 mM UDPG, 2 mM sodium pyrophosphate, 1.6 mM
NADP, 10/zM glucose 1,6-diphosphate, 1 mM EDTA, 4 mM MgC12, 85
mM Tricine-HC1 at pH 7.6, 400 p~g/ml INT, and 40 /zg/ml phenazine
ethosulfate. The reaction mixture is incubated in the dark at room temper-
ature until blue bands appear, generally a matter of several hours.
Example 5. Aspartate Transcarbamylase (EC 2.1.3.2, Aspartate Car-
bamoyltransferase). Simultaneous Capture (i) and Postincubation Cap-
438 RELATED TECHNIQUES [28]

ture (ii). This method 36 illustrates the detection of enzymes that release
phosphate, pyrophosphate, or CO2 by use of calcium ion as precipitant,
with additional enhancement of visualization by Alizarin Red. Other
enzymes that may be detected by these means are ornithine trans-
carbamylase, isocitrate dehydrogenase, UDPGpyrophorylase, xanthine-
guanine phosphoribosyltransferase, ribonuclease, and a variety of decar-
boxylases, synthases, phosphatases, and pyrophosphatases. Appropriate
variations in assay conditions for a number of these enzymes have been
presented. 32
Gels are incubated in 0.1 M glycine-NaOH, pH 10, containing 1 mM
aspartate, 0.2 mM carbamoyl phosphate, and 2 mM CaCI2. An opalescent
precipitate appears after approximately 20 min of incubation at room
temperature and can be viewed or photographed against a dark back°
ground. If desired, the calcium phosphate thus formed can be better visu-
alized, or scanned, after staining with 0.001% Alizarin Red S and
0.0005% AIC13 in 0.1 M potassium phosphate at p H l l for 12-24 hr.
Example 6. Proteases or Lipases (EC 3.1). Postincubation Capture (ii)
or Simultaneous Capture (i) in a Sandwich-Type Indicator Gel (iv). The
method 9 involves the use of ultrathin-layer agar gels as overlays for
sandwich-type incubations. The ultrathin technique is particularly suit-
able when applied to ultrathin-layer isoelectrofocused gels, thereby re-
ducing the problem of diffusion. For many other applications, a thicker
block of agar or acrylamide, prepared in a similar manner to the electro-
phoresis gel, may be substituted. 37
For the detection of lipase, 0.4 g of agar is dissolved with heating and
stirring in 20 ml of 0.1 M sodium succinate at pH 6. The mixture is cooled
to 60-65 °, and 4 mg of rhodamine B and 0.5 g of trioleine are added and
emulsified for 1 min with an Ultra-Turrax homogenizer.
For detection of protease, l g of agar is dissolved in 25 ml of 0.3 M
Tris-HCl at pH 7, and 0.5 g of casein is dissolved in 25 ml of the Tris
buffer, each with heating and stirring. The two warm solutions are then
mixed. The agar containing substrate (a measured volume, calculated to
be 30% in excess) is poured onto Mylar (Technoplast, Cologne) polyester
film of 100 /~m thickness (300 x 125 mm), which is supported on a
warmed glass base plate (40°). Greased polyester spacers of 150/~m are
used along both edges of the film. The surface of the agar is covered with a
polyester film that has been pretreated with 6 N NaOH for 15 min, rinsed,
and dried. The film is pressed with a loaded glass plate. After the agar

as j. E. Grayson, R. J. Yon, and P. J. Butterworth, Biochem. J. 183, 239 (1979).

37 D. Every, Anal. Biochem. 116, 519 (1981).
[29] RAPID PROTEIN STAIN 439

solidifies, the glass plate is removed and the film is applied to the electro-
focused gel. The sandwich that is formed is incubated at 40° for 2-5 min.
For detection of lipase, light pink bands (fluorescent at 366 nm) appear
on a deep red gel. For detection of proteases, the agar gel is immersed in
10% TCA; clear bands appear on an opalescent gel.
Example 7. a-Glucosidase (EC 3.2.1.20). Autochromic Method (iii)
with Sandwich-Type Incubation (iv). The autochromic method, by which
spectral changes in either the substrate or product can be directly de-
tected, is obviously preferred when applicable. It is basically simple and
allows rapid detection, thereby reducing the problem of diffusion. Tables
I through V include a number of examples in which chromogenic deriva-
tives are used as substrates. Umbelliferyl derivatives, in particular, have
been used to detect both a- and fl-glucosidases,2,38 arylsulfatase,29,39 ester-
a s e s , 22 and phosphatases. 4°
Whatman No. 3 filter paper is soaked in 0.1 M sodium citrate at pH 4,
containing 4-methyl umbelliferyl t~-o-glucopyranoside (0.5 mg/ml). After
electrophoresis of the enzyme, the paper is smoothed onto the cut surface
of a starch gel and covered with Saran Wrap. The gel is viewed under UV
light during incubation at room temperature for 15-60 min. Once bands
appear, fluorescence may be enhanced by spraying with ammonia.

3s D. M. Swallow, G. Corney, H. Harris, and R. Hirschhorn, Ann. Hum. Genet. 38, 391
(1975).
39 p. Manowitz, L. Goldstein, and F. Bellomo, Anal. Biochem. 89, 423 (1978).
40 D. M. Hawley, K. C. Tsou, and M. E. Hodes, Anal. Biochem. 117, 18 0981).

[29] G e l P r o t e i n Stains: A R a p i d P r o c e d u r e 1
By A. H. REISNER

The triphenylmethane textile dye Coomassie Brilliant Blue R250 (Acid

Blue 83) was first used by Fazekas de St. Groth et al. 2 to stain proteins
on cellulose acetate. In a solvent system of methanol-acetic acid-water
(5 : 1 : 5), proteins in polyacrylamide gels (PAG) were stained with it, 3 but
the procedure caused marked shrinkage of the gel and a high background
that required destaining and thereby possible loss of some bands. The use
i This chapter is dedicated to the memory of P. Nemes.
2 Fazekas de St. Groth, R. G. Webster, and A. Datyner, Biochim. Biophys. Acta 71, 377
(1963).
3 T. S. Meyer and B. L. Lambert, Biochim. Biophys. Acta 107, 144 (1965).

Copyright © 1984 by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
[29] RAPID PROTEIN STAIN 439

solidifies, the glass plate is removed and the film is applied to the electro-
focused gel. The sandwich that is formed is incubated at 40° for 2-5 min.
For detection of lipase, light pink bands (fluorescent at 366 nm) appear
on a deep red gel. For detection of proteases, the agar gel is immersed in
10% TCA; clear bands appear on an opalescent gel.
Example 7. a-Glucosidase (EC 3.2.1.20). Autochromic Method (iii)
with Sandwich-Type Incubation (iv). The autochromic method, by which
spectral changes in either the substrate or product can be directly de-
tected, is obviously preferred when applicable. It is basically simple and
allows rapid detection, thereby reducing the problem of diffusion. Tables
I through V include a number of examples in which chromogenic deriva-
tives are used as substrates. Umbelliferyl derivatives, in particular, have
been used to detect both a- and fl-glucosidases,2,38 arylsulfatase,29,39 ester-
a s e s , 22 and phosphatases. 4°
Whatman No. 3 filter paper is soaked in 0.1 M sodium citrate at pH 4,
containing 4-methyl umbelliferyl t~-o-glucopyranoside (0.5 mg/ml). After
electrophoresis of the enzyme, the paper is smoothed onto the cut surface
of a starch gel and covered with Saran Wrap. The gel is viewed under UV
light during incubation at room temperature for 15-60 min. Once bands
appear, fluorescence may be enhanced by spraying with ammonia.

3s D. M. Swallow, G. Corney, H. Harris, and R. Hirschhorn, Ann. Hum. Genet. 38, 391
(1975).
39 p. Manowitz, L. Goldstein, and F. Bellomo, Anal. Biochem. 89, 423 (1978).
40 D. M. Hawley, K. C. Tsou, and M. E. Hodes, Anal. Biochem. 117, 18 0981).

[29] G e l P r o t e i n Stains: A R a p i d P r o c e d u r e 1
By A. H. REISNER

The triphenylmethane textile dye Coomassie Brilliant Blue R250 (Acid

Blue 83) was first used by Fazekas de St. Groth et al. 2 to stain proteins
on cellulose acetate. In a solvent system of methanol-acetic acid-water
(5 : 1 : 5), proteins in polyacrylamide gels (PAG) were stained with it, 3 but
the procedure caused marked shrinkage of the gel and a high background
that required destaining and thereby possible loss of some bands. The use
i This chapter is dedicated to the memory of P. Nemes.
2 Fazekas de St. Groth, R. G. Webster, and A. Datyner, Biochim. Biophys. Acta 71, 377
(1963).
3 T. S. Meyer and B. L. Lambert, Biochim. Biophys. Acta 107, 144 (1965).

Copyright © 1984 by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
440 RELATED TECHNIQUES [29]

of 12.5% trichloroacetic acid (TCA) 4,5 as a solvent reduced these draw-

backs, but background staining remained a problem.
The relatively low solubility of the methyl-substituted G form of the
dye (Coomassie Brilliant Blue G250, Acid Blue 90, Xylene Brilliant
Cyanin G) in 12.5% TCA led to its substitution for the R form, 6 and a
concomitant reduction in background was obtained because the stain did
not penetrate the gel. If, on the other hand 6% (w/v) perchloric acid
replaced TCA as the solvent for the G form, the dye was highly soluble
but rendered into its leuco form (orange brown). 7 Furthermore, when
bound to proteins in this environment the dye was transformed to its
intense blue form. Thus, the dye fully penetrated the gel, producing a pale
orange background in which protein zones appeared blue. Finally, the
Ampholines (LKB) used in isoelectric focusing tended not to bind the
dye, making the system particularly useful for staining proteins after iso-
electric focusing.
The method per se is not suitable for gels in which sodium dodecyl
sulfate is a component.

Reagent
Coomassie Brilliant Blue G250, HC104 solution. Dilute 100 ml of 70%
HCIO4 (density 1.70 g/ml) to 2000 ml with water and add 0.8 g of the
dye. Stir at room temperature for 1 hr and filter through Whatman
No. 1 paper. Follow by filtration through a 0.45/zm Millipore mem-
brane. The solution is stable indefinitely at room temperature. Some
batches of the dye cause a deeper orange background coloration than
others. If the background is darker than desired, dilute the stock
solution with 6% (w/v) HC104.

Procedure
Gels to be stained are placed into a volume of the stain such that the
ratio of water in the gel to volume of stain is approximately 3 : 5. The gels
can be developed between 20 and 37°. Dense protein bands are observed
within 10 sec at room temperature. After 10 min, most bands can be seen
and all the bands that are going to be observed are evident within 90 rain.
At 37°, the required time is about half that at room temperature. About 8-
10 hr at 37° (overnight at room temperature) is required for the dye to
4 A. Chrambach, R. A. Reisfeld, M. Wyckoff, and J. Zaccari, Anal. Biochem. 20, 150 (1967).
5 D. Rodbard and A. Chrambach, Anal. Biochem. 40, 95 (1967).
6 W. Diezel, G. Kopperschlager, and E. Hoffman, Anal. Biochem. 48, 617 (1972).
7 A. H. Reisner, P. Nemes, and C. Bucholtz, Anal. Biochem. 64, 509 (1975).
[30] SILVER STAINS 441

penetrate completely a 4-mm-thick gel. It has been shown s that the sensi-
tivity can be increased about threefold when the gels, after staining, are
placed in 5% (v/v) acetic acid; under these conditions the background
color changes to pale blue.
When dealing with large numbers of gel slabs, it may be convenient to
stain them in polythene bags. Gels may be placed into the bags using an
appropriately shaped spatula, after which stain is added and the bag is
sealed by heat. After 45 min at 37 °, the gel should be transferred to a new
polythene bag and about 0.5 ml o f 6% (w/v) HCIO4 containing 0.005%
(w/v) Coomassie Brilliant Blue G250 added prior to sealing the bag. The
gel can be stored for many months and photographed if air spaces are
eliminated between the gel-polythene interface. To photograph the gels it
is useful, though not essential, to use a medium red filter such as a Wrat-
ten Series A together with a fine-grained panchromatic film such as Ilford
Pan F or Kodak Panatomic X.
A more sensitive but relatively elaborate procedure using Coomassie
Brilliant Blue G250 to stain isoelectric focused gels has been described by
Vesterberg et al. 9

8 I. B. Holbrook and A. G. Laver, Anal. Biochem. 75, 634 (1976).

9 O. Vesterberg, L. Hansen, and A. Sjosten, Biochim. Biophys. Acta 491, 160 (1977).

[30] G e l P r o t e i n S t a i n s : S i l v e r S t a i n
B y CARL R. MERmL, DAVID GOLDMAN, and
MARGARET L. VAN KEUREN

Silver Staining
Applications of silver-based histologicaP -6 and photographic tech-
niques 7-9 to the detection of proteins and other biopolymers, separated on
t L. Keranyi and R. Gallyas, Clin. Chem. Acta 38, 465 (1972).
P. Verheechi, J. Neurol. 209, 59 (1975).
3 D. Karcher, A. Lwenthal, and G. Van Soon, Acta Neurol. Belg. 79~ 335 (1979).
4 R. C. Switzer, C. R. Merril, and S. Shifrin, Anal. Biochem. 98, 231 (1979).
5 H. R. Hubell, L. I. Rothblum, and T. C. Hsu, Cell Biol. Int. Rep. 3, 615 (1979).
6 C. R. Merril, R. C. Switzer, and M. L. Van Keuren, Proc. Natl. Acad. Sci. U.S.A. 76,
4335 (1979).
7 C. R. Merril, M. L. Dunau, and D. Goldman, Anal. Biochem. 110, 201 (1981).
8 C. R. Merril, D. Goldman, S. A. Sedman, and M. H. Ebert, Science 211, 1437 (1981).
9 C. R. Merril, D. Goldman, and M. L. Van Keuren, Electrophoresis 3, 17 (1982).
J

METHODS IN ENZYMOLOGY, VOL. 104 ISBN 0-12-182004-1

[30] SILVER STAINS 441

penetrate completely a 4-mm-thick gel. It has been shown s that the sensi-
tivity can be increased about threefold when the gels, after staining, are
placed in 5% (v/v) acetic acid; under these conditions the background
color changes to pale blue.
When dealing with large numbers of gel slabs, it may be convenient to
stain them in polythene bags. Gels may be placed into the bags using an
appropriately shaped spatula, after which stain is added and the bag is
sealed by heat. After 45 min at 37 °, the gel should be transferred to a new
polythene bag and about 0.5 ml o f 6% (w/v) HCIO4 containing 0.005%
(w/v) Coomassie Brilliant Blue G250 added prior to sealing the bag. The
gel can be stored for many months and photographed if air spaces are
eliminated between the gel-polythene interface. To photograph the gels it
is useful, though not essential, to use a medium red filter such as a Wrat-
ten Series A together with a fine-grained panchromatic film such as Ilford
Pan F or Kodak Panatomic X.
A more sensitive but relatively elaborate procedure using Coomassie
Brilliant Blue G250 to stain isoelectric focused gels has been described by
Vesterberg et al. 9

8 I. B. Holbrook and A. G. Laver, Anal. Biochem. 75, 634 (1976).

9 O. Vesterberg, L. Hansen, and A. Sjosten, Biochim. Biophys. Acta 491, 160 (1977).

[30] G e l P r o t e i n S t a i n s : S i l v e r S t a i n
B y CARL R. MERmL, DAVID GOLDMAN, and
MARGARET L. VAN KEUREN

Silver Staining
Applications of silver-based histologicaP -6 and photographic tech-
niques 7-9 to the detection of proteins and other biopolymers, separated on
t L. Keranyi and R. Gallyas, Clin. Chem. Acta 38, 465 (1972).
P. Verheechi, J. Neurol. 209, 59 (1975).
3 D. Karcher, A. Lwenthal, and G. Van Soon, Acta Neurol. Belg. 79~ 335 (1979).
4 R. C. Switzer, C. R. Merril, and S. Shifrin, Anal. Biochem. 98, 231 (1979).
5 H. R. Hubell, L. I. Rothblum, and T. C. Hsu, Cell Biol. Int. Rep. 3, 615 (1979).
6 C. R. Merril, R. C. Switzer, and M. L. Van Keuren, Proc. Natl. Acad. Sci. U.S.A. 76,
4335 (1979).
7 C. R. Merril, M. L. Dunau, and D. Goldman, Anal. Biochem. 110, 201 (1981).
8 C. R. Merril, D. Goldman, S. A. Sedman, and M. H. Ebert, Science 211, 1437 (1981).
9 C. R. Merril, D. Goldman, and M. L. Van Keuren, Electrophoresis 3, 17 (1982).
J

METHODS IN ENZYMOLOGY, VOL. 104 ISBN 0-12-182004-1

442 RELATED TECHNIQUES [30]
gels, have resulted in highly sensitive protein stains. These silver stains
are 100 times more sensitive for protein than the commonly used
Coomassie Blue. 4,6-9
In photography, photoreduced metallic silver on the "light-exposed"
crystals acts as a catalyst, resulting in a differential reduction of silver
halide crystals during chemical image development. Chemical image de-
velopment usually relies on the use of organic reducing agents in alkaline
solutions.
Histological silver stain development closely paralleled the evolution
of photographic methods. However, the chemical basis of the selective
reduction of ionic to metallic silver in most histological stains remains
unknown. In some cases it may be catalyzed by reducing agents in the
tissue. Reduction of acidic solutions of silver nitrate has been observed in
tissue known to contain significant amounts of ascorbic acid. l° In other
tissues, differences achieved with silver staining appear to be based on
physical interface phenomena.l° Although aldehydes were thought to play
a general role, neither aldehyde-creating nor aldehyde-blocking reagents
appreciably affect silver staining in some tissues, l°

A Negative-Image Silver Stain

Detection of proteins that have been separated by electrophoresis in
gels with silver stains is probably due primarily to physical effects. The
ability to generate a negative image in a gel containing separated proteins
strongly supports this suggestion.
Procedure. All steps prior to the final washing (step 5) must be con-
ducted in the dark. All solutions used should be in a quantity that is 10-
fold larger than the volume of the gel.
1. Fix the gel in 50% methanol-12% acetic acid (v/v) for 30 min and
wash for 10 min with deionized water.
2. Soak the washed gel in 0.2 M silver nitrate solution in the dark for
20 min, and briefly rinse with deionized water.
3. After rinsing, soak the gel in photographic developer (Kodak D76
in 1 : 5 dilution) for 10 min.
4. Rinse for 5 min with photographic fixer (Kodak general-purpose
fixer, used without dilution).
5. Wash with water for three 20-min periods.
The negative image obtained shows clear regions in those portions of
the gel containing protein whereas the background region is a brownish
10 H. S. W. Thompson and R. D. Hunt, in "Selected Histochemical and Histopathological
Methods," p. 800. Thomas, Springfield, Illinois, 1966.
[30] SILVER STAINS 443

gray. The production of such a negative image indicates that the protein
has affected the reducibility of the silver in the region of the gel occupied
by the protein. This image may be reversed by exposure to light during
development or by the assistance of chemical reversal procedures .7-9 The
negative silver stain is 10-fold less sensitive then the positive silver stains.

A Positive-Image Silver Stain

As positive-image stains are easier to analyze than negative-image
stains, a highly sensitive, simply performed reversal stain was developed
by utilizing photochemical techniques. 7-9 In the most sensitive of these
stains, image reversal is facilitated by the use of potassium dichromate. 8,9
Procedure. All solutions used should be in a quantity that is 10-fold
larger than the volume of the gel.
1. Gels may be fixed in either 20% (w/v) trichloroacetic acid or 50%
methanol-12% acetic acid (v/v) for 30 min (gels thinner than 0.5 mm
should be fixed only in trichloroacetic acid).
2. Wash gels twice for 15 min each with 10% ethanol or methanol and
5% acetic acid in water. This step allows the gel to swell to normal size
and assists in removing contaminating buffers and ions that might other-
wise reduce the sensitivity of the stain. Gels thicker than 1 mm require
additional washing.
3. Soak gels for 15 min in 3.4 mM potassium dichromate containing
3.2 mM nitric acid. The amount of nitric acid added should be just suffi-
cient to retain the potassium dichromate in solution. Addition of excess
nitric acid reduces the sensitivity of the stain.
4. Soak for 20 min in 12 mM silver nitrate.
5. The silver nitrate is discarded and a solution of 0.28 M sodium
carbonate containing 0.5 ml of formaldehyde (a commercially available
solution of 37% formaldehyde) per liter is added. A precipitate of silver
salts will form rapidly; to prevent the precipitate from adsorbing to the
surface of the gel, the carbonate-formaldehyde solution should be
changed at least twice. The sodium carbonate in this step is used to make
the gel alkaline so that the formaldehyde can reduce ionic silver to the
metallic form. Formic acid formed in the reaction is also buffered by the
sodium carbonate. For maximum sensitivity, image development is al-
lowed to continue until a yellowish background appears: a period of 15-20
min is required with a 1-mm-thick gel; longer periods are necessary with
thicker gels. 9 When the image is sufficiently developed, the process is
stopped by placing the gel in 3% (v/v) acetic acid for 5 min. Gels should be
washed at least twice, 20 min each time, with water before storage.
444 RELATED TECHNIQUES [30]

6. Gels may be stored indefinitely in water or, at any time thereafter,

soaked in 3% glycerol for 5 min and dried between dialysis membranes
under reduced pressure at 80-82 ° for 3 hr. This results in a transparency
that is relatively permanent and easy to store. If the gel has been properly
washed, the image is intensified upon drying. For autoradiography or
fluorography, gels are soaked in 3% glycerol for 5 min and dried onto
Whatman 3 MM filter paper under reduced pressure at 80-82 ° for 3 hr.

A Modified Positive-Image Siloer Stain

In some applications, it is desirable to combine the potassium dichro-
mate and silver nitrate in a single solution, particularly when a 7% or
lower concentration of acrylamide gel is used, or with applications of
agarose gels. The first two steps are performed as specified (note that
agarose is more efficiently fixed with trichloroacetic acid). In step 3, the
gel is placed in a solution that is 19.5 mM silver nitrate, 1.34 mM potas-
sium dichromate, and 13.5 mM sulfuric acid for 20 min. The gel image is
then developed normally as described in steps 5 and 6.

Recycling for Increased Sensitivity

Recycling is accomplished by staining as described in the positive
silver stain procedure followed by two additional 15-min rinses in 10 gel
volumes of 3% (v/v) acetic acid and repeating or recycling through steps 4
and 5. This procedure results in an intensification of the stain because the
image intensity in the positive image stain, i.e., without recycling, is
limited by diffusion of silver from the gel during step 5. By recycling the
gel through steps 4 and 5, silver ions can be replaced in the gel and
additional staining intensity can be achieved. 9 Acidification with acetic
acid is required in the recycling procedure to prevent the nonspecific
reduction of silver that would occur if an alkaline gel, containing formal-
dehyde, were placed in a silver nitrate solution in step 4. Recycling can be
repeated several times to intensify minor spots; however, the background
also darkens and may become a problem.

Other Image Intensification and Destaining Procedures

Since the gel silver image is similar to a photograph, photographic
image intensification and destaining procedures may be adapted from
photographic formulas.~l Modification of chemical concentrations and
timing of the procedure are usually necessary, since most polyacrylamide
11 E. J. Wall, F. I. Jordan, and J. S. Carrol, in "Photographic Facts and Formulas," p. 168.
American Photographic Book Publishing Co., New York, 1976.
[30] SILVER STAINS 445

gels are thicker and have diffusion properties different from those of
photographic emulsions. For destaining silver images, we have found the
following method most useful.
1. Dissolve 37 g of sodium chloride and 37 g of cupric sulfate in 850 ml
of deionized water. Add concentrated ammonium hydroxide until all the
precipitate is dissolved and a deep blue solution is achieved before adjust-
ing the volume to 1 liter.
2. Dissolve sodium thiosulfate, 436 g, in 1 liter of water.
3. Just prior to use, equal volumes of the reagents prepared in steps 1
and 2 are mixed and used directly if total destaining is required. They may
be diluted (1 : 10 or 1 : 100) with water if light silver deposits are to be
removed or if the image is to be lightened only slightly. It is advisable to
photograph the gels prior to and during destaining to preserve a transient
of the image, since it is difficult to stop destaining at a precise point. Gels
may be restained after destaining by washing the gel three times in deion-
ized water, 10 min each time, and then repeating steps 2 through 5 of the
positive-image staining procedure. If the gel is insufficiently washed prior
to restaining, silver will be reduced within the gel in step 4.

Sensitivity, Quantitation, and Protein Detection

Proteins have been detected with silver stains at concentrations as low
as 20 pg/mm2. 9 The positive-image silver stain procedure described above
has been shown, with 8 purified proteins, to be linear over a 40-fold
range in concentration, between 50 pg/mm 2 to 2 ng/mm2. At concentra-
tions greater than 2 ng/mm2, the stain becomes nonlinear as spot densities
reach saturation. 9 The dynamic range may be extended by recording the
image during development. It should be noted that the relationship be-
tween density of silver staining and the concentration of protein is charac-
teristic for each protein. 9 Quantitative use of the silver stain is possible if
constitutive or marker proteins are present on each gel, so that densities
can be normalized. Care must be taken to work within the linear range of
the stain. There are some proteins that will stain with Coomassie Blue,
but will not stain with the positive-image silver stain unless the recycling
procedure is employed. 9
Some of the histological silver stains that were developed for subcellu-
lar organelles or cellular structures have proved to be useful for staining
specific proteins separated on polyacrylamide gels. One of these stains
neurofilament protein, 12and another primarily stains nucleolar proteins. 13

12 p. Gambetti, L. Autilio-Gambetti, and S. C. Papasozonenos, Science 213, 1521 (1981).

~3 H. R. Hubbell, L. I. Rothblum, and T. C. Hsu, Cell Biol. Int. Rep. 3, 615 (1979).
446 RELATED TECHNIQUES [30]

Colored Protein Images with Silver Stains

Most silver stains produce some colored bands or spots. Such colora-
tion has also been observed in silver-based photographic processes. Color
produced by this means was dependent on three variables: silver grain
size, the refractive index of the gel or emulsion, and the distribution of
silver grains in the gel. In general, the smaller grains transmit reddish or
yellow-red light. Larger grains give bluish colors, and the very large
grains produce black images. In a study of human cerebrospinal fluid
proteins, utilizing a histochemical silver stain to stain proteins separated
by two-dimensional electrophoresis, some lipoproteins stained blue
whereas some glycoproteins appeared as yellowish-brown and red
spots. ~4 By modifying silver stain procedures, color effects can be en-
hanced ~5although saturation and negative staining effects can usually be
accentuated by such modifications, making quantitation more difficult.

Comments
Although silver stains for the detection of protein are highly sensitive
and fairly easy to perform, they are not without problems. The major loss
of sensitivity in silver staining is due to inadequate water purity. Deion-
ized water with a conductivity of less than 1 /xmho is required in all
reagents, including wash and fixing solutions. The second major cause of
difficulty is usually inadequate fixation of proteins prior to staining. Poly-
acrylamide gels thinner than 0.5 mm, and all agarose gels, require fixation
in 20% trichloroacetic acid. Gels thicker than 1 mm require additional
washing prior to staining. Occasional surface artifacts, caused by silver
carbonate adsorption, will mar a gel. They can be minimized by rapidly
changing the sodium carbonate-formaldehyde solution during initial im-
age development and by handling gels carefully. Pressure, fingerprints,
and surface drying also cause surface artifacts. When sodium dodecyl
sulfate gels (containing samples denatured in a solution containing mer-
captoethanol) are stained to the point that the background is a yellowish
brown, two horizontal lines will be observed at 60,000 and 67,000 daltons.
These artifacts are caused by the mercaptoethanol and can be eliminated
by reducing the amount employed.
Quenching may be observed if gels stained with silver are to be used
for autoradiography or fluorography. The histological silver stain 4A6 al-

14 D. Goldman, C. R. Merril, and M. H. Ebert, Clin. Chem. 26, 1317 (1980).

~5 W. Sammons, L. D. Adams, and E. E. Nishizawa, Electrophoresis 2, 135 (1981).
t6 M. L. Van Keuren, D. Goldman, and C. R. Merril, Anal. Biochem. 116, 248 (1981).
[31] GLYCOPROTEINSTAXNS 447

most completely quenches the detection of 3H-labeled compounds; how-

ever, the photochemical positive-image silver stain described here causes
less quenching with 3H-labeled proteins; quenching with ~4C-labeled pro-
teins is barely perceptible.16 Fluorographic detection of 3H-labeled pro-
teins can be restored almost completely by destaining the gel prior to
fluorography.~6

[31] Gel P r o t e i n Stains: Glycoproteins

By JOHN E. GANDER

Procedures for identifying glycoproteins on gels are available ~,2 and

have been improved recently. The methods currently available are di-
vided into the following categories: (1) thymol-HESO4 method, (2) peri-
odic acid-Schiff base method, and (3) fluorescein isothiocyanate-labeled
lectin or -antibody method.

Thymol-H2S04 Method
This procedure is useful for the location of glycoproteins containing at
least 50 ng of carbohydrate.2 Glycoproteins bearing hexosyl, hexurono-
syl, or pentosyl residues react with H2SO4 to form furfural derivatives,
which, in turn, react with thymol to form a chromogen. The chromogen is
stable for only a few hours at ambient temperature. Furfural derivatives
are not formed when 2-deoxy- or 2-acetamido-2-deoxyhexosaminylresi-
dues are allowed to react with H2804. Gels must be washed free of low-
molecular-weight contaminants before they are treated with the acid.
Since protein zones become purple upon treatment with concentrated
H2SO4 in the presence of glycine, this amino acid must be removed if Tris-
glycine is used as a buffer system; the chromogen is presumably derived
from the reaction of protein-bound tryptophan with the glyoxylic acid
formed when HzSO4 interacts with glycine,z
The method cannot be used with gels containing glycosyl residues,
e.g., agarose; such residues would be dehydrated to furfural derivatives
by HzSO4.
Procedure. Fractionate protein(s) on 5-10% polyacrylamide gels in
tubes 5 mm in diameter using any of the usual buffer systems including

i K . B u r r i d g e , t h i s s e r i e s , V o l . 50, p. 54.
2 D . R a u c h s e n , Anal. Biochem. 99, 4 7 4 (1979).

Copyright © 1984 by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
[31] GLYCOPROTEINSTAXNS 447

most completely quenches the detection of 3H-labeled compounds; how-

ever, the photochemical positive-image silver stain described here causes
less quenching with 3H-labeled proteins; quenching with ~4C-labeled pro-
teins is barely perceptible.16 Fluorographic detection of 3H-labeled pro-
teins can be restored almost completely by destaining the gel prior to
fluorography.~6

[31] Gel P r o t e i n Stains: Glycoproteins

By JOHN E. GANDER

Procedures for identifying glycoproteins on gels are available ~,2 and

have been improved recently. The methods currently available are di-
vided into the following categories: (1) thymol-HESO4 method, (2) peri-
odic acid-Schiff base method, and (3) fluorescein isothiocyanate-labeled
lectin or -antibody method.

Thymol-H2S04 Method
This procedure is useful for the location of glycoproteins containing at
least 50 ng of carbohydrate.2 Glycoproteins bearing hexosyl, hexurono-
syl, or pentosyl residues react with H2SO4 to form furfural derivatives,
which, in turn, react with thymol to form a chromogen. The chromogen is
stable for only a few hours at ambient temperature. Furfural derivatives
are not formed when 2-deoxy- or 2-acetamido-2-deoxyhexosaminylresi-
dues are allowed to react with H2804. Gels must be washed free of low-
molecular-weight contaminants before they are treated with the acid.
Since protein zones become purple upon treatment with concentrated
H2SO4 in the presence of glycine, this amino acid must be removed if Tris-
glycine is used as a buffer system; the chromogen is presumably derived
from the reaction of protein-bound tryptophan with the glyoxylic acid
formed when HzSO4 interacts with glycine,z
The method cannot be used with gels containing glycosyl residues,
e.g., agarose; such residues would be dehydrated to furfural derivatives
by HzSO4.
Procedure. Fractionate protein(s) on 5-10% polyacrylamide gels in
tubes 5 mm in diameter using any of the usual buffer systems including

i K . B u r r i d g e , t h i s s e r i e s , V o l . 50, p. 54.
2 D . R a u c h s e n , Anal. Biochem. 99, 4 7 4 (1979).

Copyright © 1984 by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
448 RELATEDTECHNIQUES [31]

Tris-glycine, ampholytes, or sodium dodecyl sulfate (SDS). 2 After elec-

trophoresis, wash the gels twice for at least 2 hr in borosilicate tubes with
isopropanol-acetic acid-H20 (25 : 10 : 65) to fix the proteins and to re-
move low-molecular-weight substances. Additional washes may be nec-
essary if the protein samples contain large concentrations of sucrose or
other soluble carbohydrates that can react with H2SO4 to form a furfural
derivative. A final wash for 2 hr in the same solvent containing 0.2%
thymol (w/v) results in formation of a stable gel. After washing with
thymol, decant the liquid and allow the gels to drain. Add a solution of
concentrated H2SO4-absolute ethanol (80 : 20) at ambient temperature to
each gel and cap the tubes. At least 10 ml of reagent per milliliter of gel
should be used. Gently shake the tubes held in a horizontal position at 35°
for 2.5 hr or until the opalescent appearance of the gels just disappears.
Zones containing glycoproteins stain red whereas the background is yel-
low. Other proteins do not form visible zones when treated in this
manner.

Periodic Acid-Schiff Base Method

This procedure is another generally useful technique for locating gly-
coproteins on gels. Periodic acid oxidizes and cleaves gem secondary
alcohols of glycosyl residues to dialdehydes; the aldehydes are then al-
lowed to react with fuchsin, 3,4 Alcian b l u e : or dansyl hydrazine 6 to form a
Schiff base. A lower limit of 2-3 ~g of carbohydrate can be detected using
fuchsin or Alcian blue 7 compared with 40 ng using dansyl hydrazine. 6
Neutral glycosyl residues that are substituted at C-3 and 2-deoxyglycosyl
or 2-acetamido-2-deoxyglycosaminyl residues substituted at either C-3 or
C-4, or both positions, will not be oxidized by periodic acid.
The method is not as sensitive as the H2SO4-thymol method when
fuchsin or Alcian blue are used as the base but is comparable when dansyl
hydrazine is used. However, with dansyl hydrazine, it is necessary to
modify the Eckhardt et ai. procedure 6 if glycoproteins are separated on
agarose gels. 8
Procedure with Dansyl Hydrazine as the Base. Polymerization of
acrylamide to polyacrylamide must be catalyzed by a nonfluorescent re-
agent such as ammonium persulfate.
3 R. M. Zaccharias, T. E. Zell, J. H. Morrison, and J. J. Woodlock, Anal. Biochem. 30, 148
(1969).
4 G. Fairbanks, T. L. Steck, and D. F. H. Wallach, Biochemistry 10, 2606 (1971).
5 A. H. Wardi and G. A. Michos, Anal. Biochem. 49, 607 (1972).
6 A. E. Eckhardt, C. E. Hayes, and I. E. Goldstein, Anal. Biochem. 73, 192 (1976).
7 R. A. Kapitany and E. J. Zebrowski, Anal. Biochem. 56, 361 (1973).
s M. Furlan, B. A. Perret, and E. A. Beck, Anal. Biochem. 96, 208 (1979).
[31] GLYCOPROTEINSTAINS 449

Dissolve proteins in 10% glycerol (v/v) and 0.005% methylene green

(w/v), layer on the stacking gel, and electrophorese in 7.5% poly-
acrylamide gels. Fix the gels overnight in ethanol-glacial acetic acid-
water (40:5 : 55). Treat the gels with 20/zl of 0.7% paraperiodic acid (w/v)
in acetic acid-water (95 : 5) for 2 hr. Rinse gels with water, treat with 0.5%
sodium metabisulfite in 5% acetic acid until the gels are colorless (1-1.5
hr), and rinse repeatedly with water. Transfer the gels to tubes of appro-
priate size and add equal volumes of acidic dimethyl sulfoxide (0.6 ml of
12 N HC1 per liter of dimethyl sulfoxide) and a freshly prepared solution
containing 2 mg of dansyl hydrazine per milliliter of dimethyl sulfoxide.
Stopper tubes, mix the solutions, and maintain at 60° for 2 hr. Decant the
solution and add sufficient NaBH4 in dimethyl sulfoxide (0.2 mg of NaBH4
per milliliter of dimethyl sulfoxide) to cover the gels. After 30 min at 25°,
decant the liquid and rinse the gels with water. Treat the gels with acetic
acid-water (99 : 1) repeatedly until the background is colorless when illu-
minated with ultraviolet light (hrnax 366 nm). If a 4 x 5 press-type camera
equipped with a Polaroid 545 film adapter and Polaroid type 57 (ASA
3200) film is used, an exposure time of 2.5 sec at f 5.6 with a Kodak
Wratten 16 filter should provide adequate photographic record of the gel.
Controls in which the periodate oxidation step is omitted will disclose
proteins with natural fluorescence or those that noncovaiently bind dansyl
hydrazine.

Fluorescein Isothiocyanate-Labeled Lectin Method

This procedure has specificity toward the carbohydrate region of gly-
coproteins to a degree dictated by the specific lectin. Fluorescein-labeled
concanavalin A reacts with numerous glycoproteins because many con-
tain residues, such as D-mannopyranosyl, for which concanavalin A has
an affinity. In contrast, fluorescein-labeled monoclonal antibody, with its
antigenic determinants directed toward a unique functional group on a
specific glycoprotein, would provide a very sensitive and specific probe
for that glycoprotein.
Furlan et al. have described a method for staining glycoproteins, frac-
tionated on either polyacrylamide or agarose gels, by using fluorescein
isothiocyanate conjugated to Ricinus communis agglutinin, to con-
canavalin A, or to immunoglobulins. 8 Less than 100 ng of hexosyl resi-
dues bound to protein is detectable.
Procedure. Conjugate lectin or antibody (5 mg/ml) by allowing to react
with fluorescein isothiocyanate (100/.~g/ml final concentration) in 0.15 M
Na2HPO4 (pH 9.5) at 25° for 20 hr. 9 Dialyze the reaction mixture against
9 T. H. The and T. E. Feltkamp, Immunology 18, 865 (1970).
450 RELATEDTECHNIQUES [31]
several changes of 0.1 M NaC1-0.05 M Tris-HCl-1 mM CaCI2 at pH 7.0.
The A280 : A485ratio should be about 0.9 depending on the molar extinction
coefficient at 280 nm of the protein.
Glycoprotein separations have been carried out on polyacrylamide
gels (5% acrylamide containing 5% bisacrylamide) made in 0.2% SDS-0.1
M Tris-HC1-6 M urea at pH 7.48; or 2% polyacrylamide-0.5% agarose
buffered with 0.075 M sodium 5,5'-diethylbarbiturate-0.01 M ethylene-
diaminetetraacetic acid at pH 8.6 and containing 6 M urea and 0.2%
SDS 1°; or on I% agarose gels in a buffer containing 0.075 M sodium 5,5'-
diethylbarbiturate-0.01 M ethylenediaminetetraacetate-0.2% SDS at
pH 8.6.11 Remove the SDS by washing the gels for 20 hr in methanol-
acetic acid-water (10 : 3 : 27) and in 5% acetic acid. Wash the fxed gel for
4 hr during each of four changes of 0.1 M NaC1-0.05 M Tris-HCl (pH 7.0)
containing 1 mM CaC12 and 1 mM MnC12. Dialyze the fluorescein
isothiocyanate-lectin or fluorescein isothiocyanate-antibody, conjugate
against the NaC1-Tris-CaCI2-MnC12 buffer described above, and adjust
the concentration of the fluorescein-labeled protein to 1 mg/ml. Treat the
gels with fluorescein-labeled protein solution for 12 hr at 25°. Follow by
destaining in the same buffer for 2 days. The stained bands remaining are
noted as being visible for up to 1 month at either 25 or 40.8 The staining
solution can be used at least three times without loss of significant staining
intensity.
Intense background staining of agarose gels was obtained with galac-
tose-specific Ricinus lectins and the fuorescent substances could not be
removed by washing with the Tris-NaCI-CaCl2 buffer. However, buffer
containing 0.1 M D-galactose did remove the background fluorescence.
No difficulty should be encountered in destaining polyacrylamide or
agarose gels treated with fluorescein-labeled concanavalin A.

Comments
Staining techniques that couple the action of enzymes to the formation
of chromogenic or fluorogenic products have been used to locate treha-
lase, 12 invertases, 13 and glycoproteins that react with sialic acid-contain-
ing lectins 14or concanavalin A.15 The latter method depends on the forma-

l0 B. A. Perret, M. Furlan, and E. A. Beck, Biochim. Biophys. Acta 578, 164 (1979).
11 E. A. Beck, L. Tranqui-Pouit, A. Chapel, B. A. Perret, M. Furlan, G. Hudry-Clergeon,
and M. Suscillon, Biochim. Biophys. Acta 578, 155 (1979).
12 K. A. Killick and L.-W. Wang, Anal. Biochem. 106, 367 (1980).
13 p. Babczinski, Anal. Biochem. 105, 328 (1980).
14 K. Yamada and S. Shimizu, Histochem. J. 11, 457 (1979).
15 K. Yamada and S. Shimizu, Histochem. 47, 159 (1976).
[32] PHOSPHOPROTEINSTAINS 451

tion of a stain when horseradish peroxidase, which is covalently attached

to a lectin or an antibody, catalyzes the oxidation of diaminobenzidine.13
Most of the noted techniques are less than completely satisfactory.
For instance, as sensitivity is increased, interference by background sub-
stances in gels also increases. Adequate controls must be included for
each staining series because any one preparation of glycoproteins under
investigation may contain substances that lead to artifacts.

[32] G e l P r o t e i n Stains: P h o s p h o p r o t e i n s
By JOHN A. CUTTING

Phosphoproteins may be detected on polyacrylamide gel electrophero-

grams by use of either the radioactivity of intrinsically incorporated 32Pi,
or a stain that has some specificity for the phosphate moiety. The use of
3Zp-labeled samples has the advantage of greater sensitivity over direct
staining. However, only de novo incorporated phosphate is thereby de-
tected. There are several reasons why phosphoproteins of interest may
not become labeled during incubation of a biosystem with a 32p-containing
substrate: the protein is not being synthesized or presynthesized mole-
cules do not have available sites for additional phosphorylation; the pro-
tein may be subject to avid phosphatases; or a necessary, specific kinase
may be unavailable. Each situation precludes detection of the protein. In
contrast, all phosphate linked to a particular amino acid in a phosphopro-
tein is available for detection by a suitable dye method.

Method I. The Entrapment of Liberated Phosphate (ELP) 1

This is the most specific of the available phosphoprotein staining
methods. Moreover, it is easy to perform a test for the authenticity of a
stained band (see Controls). A modification permits detection of enzymes
that release inorganic phosphate from their substrates (see below). The
method depends on entrapment within the gel of insoluble calcium phos-
phate formed during the alkaline hydrolysis of susceptible protein phos-
phoester bonds in the presence of calcium chloride. Subsequently, en-
trapped phosphate is allowed to react with a modified Fiske-SubbaRow
reagent, and detection of the resulting insoluble blue complex is enhanced
i j . A . C u t t i n g a n d T. F . R o t h , Anal. Biochem. 54, 386 (1973).

Copyright © 1984by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
[32] PHOSPHOPROTEINSTAINS 451

tion of a stain when horseradish peroxidase, which is covalently attached

to a lectin or an antibody, catalyzes the oxidation of diaminobenzidine.13
Most of the noted techniques are less than completely satisfactory.
For instance, as sensitivity is increased, interference by background sub-
stances in gels also increases. Adequate controls must be included for
each staining series because any one preparation of glycoproteins under
investigation may contain substances that lead to artifacts.

[32] G e l P r o t e i n Stains: P h o s p h o p r o t e i n s
By JOHN A. CUTTING

Phosphoproteins may be detected on polyacrylamide gel electrophero-

grams by use of either the radioactivity of intrinsically incorporated 32Pi,
or a stain that has some specificity for the phosphate moiety. The use of
3Zp-labeled samples has the advantage of greater sensitivity over direct
staining. However, only de novo incorporated phosphate is thereby de-
tected. There are several reasons why phosphoproteins of interest may
not become labeled during incubation of a biosystem with a 32p-containing
substrate: the protein is not being synthesized or presynthesized mole-
cules do not have available sites for additional phosphorylation; the pro-
tein may be subject to avid phosphatases; or a necessary, specific kinase
may be unavailable. Each situation precludes detection of the protein. In
contrast, all phosphate linked to a particular amino acid in a phosphopro-
tein is available for detection by a suitable dye method.

Method I. The Entrapment of Liberated Phosphate (ELP) 1

This is the most specific of the available phosphoprotein staining
methods. Moreover, it is easy to perform a test for the authenticity of a
stained band (see Controls). A modification permits detection of enzymes
that release inorganic phosphate from their substrates (see below). The
method depends on entrapment within the gel of insoluble calcium phos-
phate formed during the alkaline hydrolysis of susceptible protein phos-
phoester bonds in the presence of calcium chloride. Subsequently, en-
trapped phosphate is allowed to react with a modified Fiske-SubbaRow
reagent, and detection of the resulting insoluble blue complex is enhanced
i j . A . C u t t i n g a n d T. F . R o t h , Anal. Biochem. 54, 386 (1973).

Copyright © 1984by Academic Press, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
 452 RELATED TECHNIQUES [32]
by staining with methyl green. The location of a phosphoprotein on the gel
results in a bright green band. Sensitivity is 1 nmol of phosphate.
Reagents
A. 10% (w/v) sulfosalicylic acid (SSA) in deionized water
B. 10% (w/v) SSA in 25% (v/v) 2-propanol in deionized water
C. 0.5 M CaC12 in 10% (w/v) SSA
D. 0.5 N NaOH
E. 1% (w/v) ammonium molybdate in deionized water
F. 1% (w/v) ammonium molybdate in 1 N HNO3
G. 0.5% (w/v) methyl green (CI No. 42590) in 7% (v/v) aqueous acetic
acid
H. 7% (v/v) aqueous acetic acid
All reagents must be free of phosphate, and glassware should be
washed with a phosphate-free detergent. Suspect reagents and glassware
can be checked by the addition of the Fiske-SubbaRow reagent (4.2%
ammonium molybdate in 4.5 N HC1), which will present as a blue color in
the presence of phosphate. Acceptable glassware should then be thor-
oughly rinsed with deionized water.
Procedure. For optimal results, the volumes of reagent solutions em-
ployed at each step should be 15 times the volume of the gel being stained.
1. If the sample or gel system contains a phosphate buffer or other
low-molecular-weight phosphates, the gel must be fixed in reagent A.
Gels containing SDS must be fixed in reagent B. In both instances gels
should be treated for 12 hr, with several changes of the reagent, to permit
diffusion from the gel of these interfering substances.
2. All gels: Fix in reagent C for 1 hr.
3. Rinse the gel thoroughly in deionized water to avoid carryover of
surface CaC12, which will otherwise cause excessive background staining.
4. Transfer the gel to reagent D at 60° for 30 min.
5. Rinse the gel twice for 10 min each in reagent E.
6. Place the gel in reagent F for 30 min.
7. Stain the gel with reagent G for 30 min.
8. Destain with reagent A. This process is hastened at 60% 2
9. Store gels in reagent H.
Controls. Protein itself may be stained, e.g., with Coomassie Blue, in
a parallel-run gel. Separate gels or lanes should contain a known phospho-

2 D. H. Ohlendorf, G. R. Barbarash, A. Trout, C. Kent, and L. T. Banaszak, J. Biol. Chem.

252, 7992 (1977).
[32] PHOSPHOPROTEINSTAINS 453

protein (chicken phosvitin or a-casein) as a positive control, and be

stained for phosphoprotein simultaneously with the unknown. Authentic-
ity of a stained band may be verified by the following procedure.l Two
sets of gels (a and b) are prepared, each containing the unknown and a
positive control in separate lanes. After appropriate fixation, as detailed
in step 1 above, treat the gel sets as follows:
AI. Fix set (a) in reagent A and set (b) in reagent B for 1 hr.
A2. Carry out step 3 above.
A3. Transfer both sets to a solution that will solubilize the protein
sample and permit it to diffuse from the gel [0.2 M NaCI, 4 M urea, or
0.1% sodium dodecyl sulfate (SDS) solution]. This may require a day and
several changes of the solution. Residual protein can be monitored in test
gels with Coomassie Blue stain. If either phosphate buffer or SDS is used
in this step it must subsequently be removed as in step 13 above.
A4. Use steps 4 through 9, as above.
By this strategem, phosphate liberated during hydrolysis is not
trapped in the gels of set (a) since CaC12 is absent from reagent A, but is
trapped at the phosphoprotein locus by the standard procedure to which
the gels of set (b) were subject. Step A3 results in diffusion from the gel of
all protein and nontrapped phosphate so that, following step A4, the
presence of a green band in the gels of set (b) [and the absence of such a
band from the gels of set (a)] is the consequence of a staining reaction with
trapped calcium phosphate and is not an artifact.
Entrapment of Enzymically Released Orthophosphate.4The ELP pro-
cedure is adaptable for the detection on polyacrylamide gels of enzymes
that release phosphate from their substrates. Pyruvate-uridine diphospho-
N-acetylglucosamine transferase has been located on gels after incubation
with its substrates, 1 m M phosphoenolpyruvate and 1 mM uridine diphos-
pho-N-acetylglucosamine, in the presence of 50 mM CaCI2. After steps 5
through 9, the enzyme was detected as a green band.

Method II. T h e U s e o f S t a i n s - A l l 4

The cationic carbocyanine dye, Stains-all {1-ethyl-2-(3-(1-ethyl-

naphtho[ 1,2 - d ]thiazolin - 2 - ylidene) - 2 - methylpropenyl)naphthol[ 1,2 - d]
thiazolium bromide}, stains many biopolymers varying shades of blue or
red. Unconjugated proteins are stained red, whereas phosphoproteins are

3 R. I. Zemell and R. A. Anwar, J. Biol. Chem. 250, 3185 (1975).

4 M. R. Green, J. V. Pastewka, and A. C. Peacock, Anal. Biochem. 56, 43 (1973).
454 RELATEDTECHNIQUES [32]
in shades of blue. However, glycoproteins are also stained blue, 5 as are
DNA and RNA. Although the staining procedure is easily performed, the
inherent lack of specificity should limit the use of this stain for monitoring
phosphoproteins, the phosphate content of which is verifiable by more
specific criteria. Sensitivity of the stain is at 0.3 nmol of phosphate.

Reagents
A. 25% (v/v) 2-propanol in deionized water
B. (Stock solution) 0.1% Stains-all in formamide
C. Working solution, prepared immediately before use: 10 ml of re-
agent B, 10 ml of formamide, 50 ml of 2-propanol, 1 ml of 3.0 M
Tris-HCl, pH 8.8. Prepare to 200 ml with deionized water
Procedure. Since direct light causes the gel background to become
opaque and stained bands to decolorize, both staining and destaining are
done in the dark.
1. Agitate alkaline gels or those containing SDS in reagent A for
15 min.
2. Stain gels overnight in reagent C.
3. Destain in deionized water.

Method III. Trivalent Metal Chelation 6

This method uses a trivalent metal ion as a mordant that will allow
staining of acidic phosphoproteins (phosvitins) with the protein dye
Coomassie Blue. Usually phosvitins do not stain permanently with this
dye. The method depends on differential staining: a blue band, corre-
sponding to phosvitin, is evident only when the trivalent metal ion AI3+ is
included in the staining reagent. Since AI3+ appears to permit staining by
reducing the net negative charge on the protein, 6 this method is of low
specificity. Sensitivity is at 0.13 nmol of phosphate.

Reagents
A. 0.05% Coomassie Brilliant Blue R250 in 25% (v/v) 2-propanol,
10% (v/v) acetic acid, 1% (v/v) Triton X-100, prepared in deion-
ized water
B. 0.1 M aluminum nitrate prepared in reagent A
C. 7% (v/v) aqueous acetic acid

5 L. E. King, Jr. and M. Morrison, Anal. Biochem. 71, 223 (1976).

6 j. Hagenauer, L. Ripley, and G. Nace, Anal. Biochem. 78, 308 (1977).
[33] WESTERNBLOTS 455

Procedure
1. Prepare two parallel-run gels (a and b), and stain gel (a) in reagent
A, and gel (b) in reagent B.
2. Destain both gels in reagent C.
A permanently stained blue band on gel (b) that is absent from gel (a)
suggests the presence of an acidic phosphoprotein.

[33] W e s t e r n Blots
By JAIME RENART and IGNACIO V. SANDOVAL

The blotting ~ of electrophoretically separated components onto solid

supports allows their identification, assists in the study of their binding to
other molecules, and sometimes leads to the isolation of such ligands on a
small scale. Because the proteins, extracted from gel, are bound to the
surface of a blotting support, blotting is particularly suitable when the
probes used for scrutinizing specific proteins either do not permeate the
matrix of the gel or do so only so slowly that their binding to protein
cannot be detected. In each instance, it is important to ensure that both
the gel electrophoresis system and the blotting support are adequate for
the type of experiment and for the protein probes that are used.
Technically, the simplest problem is the identification of proteins with
probes that recognize denatured proteins directly coupled to a blotting
support. For this situation, the following protocol can be used.
The protein mixture must be resolved into its components by electro-
phoresis in one or two dimensions. The sodium dodecyl sulfate (SDS)-
polyacrylamide gel systems of Laemmli2 or Neville3 (see also this volume
[12]) can be used to separate proteins in one dimension. The system of
Neville 3 has the advantage of using Tris-borate as running buffer, thereby
avoiding the lengthy washes required to remove glycine from gels in
which Tris-glycine is used. 2 However, complex mixtures of proteins are
incompletely resolved by this means, and separation in two dimensions
may be necessary. The O'Farrell system4 does so by using electrofocusing
in the first dimension and electrophoresis on SDS-polyacrylamide gels in
the second.
I j. Renart, J. Reiser, and G. R. Stark, Proc. Natl. Acad. Sci. U.S.A. 76, 3116 (1979).
2 U. K. L a e m m l i , Nature (London) 227, 680 (1970).
3 D. M. Neville, J. Biol. Chem. 256, 6328 (1971).
4 p. H. O'Farrell, J. Biol. Chem. 250, 4007 (1975).

Copyright © 1984by AcademicPress, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
[33] WESTERNBLOTS 455

Procedure
1. Prepare two parallel-run gels (a and b), and stain gel (a) in reagent
A, and gel (b) in reagent B.
2. Destain both gels in reagent C.
A permanently stained blue band on gel (b) that is absent from gel (a)
suggests the presence of an acidic phosphoprotein.

[33] W e s t e r n Blots
By JAIME RENART and IGNACIO V. SANDOVAL

The blotting ~ of electrophoretically separated components onto solid

supports allows their identification, assists in the study of their binding to
other molecules, and sometimes leads to the isolation of such ligands on a
small scale. Because the proteins, extracted from gel, are bound to the
surface of a blotting support, blotting is particularly suitable when the
probes used for scrutinizing specific proteins either do not permeate the
matrix of the gel or do so only so slowly that their binding to protein
cannot be detected. In each instance, it is important to ensure that both
the gel electrophoresis system and the blotting support are adequate for
the type of experiment and for the protein probes that are used.
Technically, the simplest problem is the identification of proteins with
probes that recognize denatured proteins directly coupled to a blotting
support. For this situation, the following protocol can be used.
The protein mixture must be resolved into its components by electro-
phoresis in one or two dimensions. The sodium dodecyl sulfate (SDS)-
polyacrylamide gel systems of Laemmli2 or Neville3 (see also this volume
[12]) can be used to separate proteins in one dimension. The system of
Neville 3 has the advantage of using Tris-borate as running buffer, thereby
avoiding the lengthy washes required to remove glycine from gels in
which Tris-glycine is used. 2 However, complex mixtures of proteins are
incompletely resolved by this means, and separation in two dimensions
may be necessary. The O'Farrell system4 does so by using electrofocusing
in the first dimension and electrophoresis on SDS-polyacrylamide gels in
the second.
I j. Renart, J. Reiser, and G. R. Stark, Proc. Natl. Acad. Sci. U.S.A. 76, 3116 (1979).
2 U. K. L a e m m l i , Nature (London) 227, 680 (1970).
3 D. M. Neville, J. Biol. Chem. 256, 6328 (1971).
4 p. H. O'Farrell, J. Biol. Chem. 250, 4007 (1975).

Copyright © 1984by AcademicPress, Inc.

METHODS IN ENZYMOLOGY, VOL. 104 All rights of reproduction in any form reserved.
ISBN 0-12-182004-1
456 RELATEDTECHNIQUES [33]

Proteins separated by electrophoresis remain embedded in acryl-

amide gel and must be blotted onto paper. For this purpose, either
diazophenylthio (DPT) paper 5 or diazobenzyloxymethyl (DBM) paper I is
available. These two papers are the best studied supports with respect to
efficiency of transfer, capacity for binding protein, reactivity conditions,
and stability. L4-6DPT paper is easier to prepare than DBM paper, 7 and the
following protocol, described by S e e d : is effective.
Materials
1,4-Butanediol diglycidyl ether (Aldrich 12,419-2)
2-Aminothiophenol (Aldrich 12,313-7)
Whatman 50 or 540 paper, or Schleicher & Schuell 589WH paper

Methods
All operations must be carried out in a fume hood. Gloves should be
used, since both 1,4-butanediol diglycidyl ether and 2-aminothiophenol
are toxic.
I. Cut the papei to the desired size (usually 16-18 cm) and place
about 20 g of paper in a heat-sealable polyester bag (Sears Seal-N-Save or
polyethylene bags). First add 70 ml of 0.5 M NaOH, which is to be
followed by 30 ml of butanediol diglycidyl ether. The bag is sealed, leaving
sufficient room for good mixing and the possibility for opening and reseal-
ing the bag. Rotate the bag end-over-end for 12-16 hr. After the coupling
period, remove reagents by pouring them into 1 M NH4OH and allow 24
hr for inactivation before disposal down the drain.
2. Mix l0 ml of 2-aminothiophenol with 40 ml of ethanol, and add this
mixture to the bag. Reseal and rotate for an additional l0 hr.
3. After step 2, papers are washed by sequential immersion and agita-
tion in ethanol, followed by washing in 0. I M HCI for a total of three
cycles of about 15 min each. Papers are rinsed in distilled water for 1-2
hr, immersed in ethanol, and finally dried in air. The paper should be
stored at - 2 0 ° in a desiccator.
4. Aminophenylthio (APT) paper is activated to form the diazo-
phenylthio (DPT) form by diazotization immediately prior to blotting.
Diazotization is performed by immersing the paper in a tray containing
100 ml of 1.2 M HCI at 4° to which 32 mg of NaNO2 dissolved in 1.2 M

5 B. Seed, Genet. Eng. 4, 91 (1983).

6 j. Reiser and J. Wardale, Fur. J. Biochem. 114, 569 (1981).
7 j. C. Alwine, D. J. Kemp, B. A. Parker, J. Reiser, J. Renart, G. R. Stark, and G. M.
Wahl, this series, Vol. 68, p. 220.
[33] WESTERNBLOTS 457

HC1 is added. Diazotization is allowed to occur for 30 min with occasional

shaking. The paper is briefly washed subsequently with several changes
of distilled water and immediately thereafter placed in contact with the
gel to be blotted.

Blotting of Proteins onto Paper

The transfer of proteins from gels onto paper, i.e., blotting, can be
performed either by passive diffusion I or by electrophoresis. 6 The rate
and yield of protein transfer are both related inversely to the size of the
protein and to the extent of cross-linking of the gel. Transfer of proteins
by electrophoresis is faster and more efficient and minimizes diffusion of
proteins over gel and paper surfaces that occurs during transfer by pas-
sive diffusion.
In both methods, the blotting paper is placed over a scouring foam pad
supported by a rigid plastic grid. The gel, containing the protein, is
washed three times for 10 min each in 250 ml of water and once for 10 min
in 50 mM sodium borate. The borate is at pH 8 for blotting by passive
diffusion and at pH 9.2 for electrophoresis. The gel is laid on the top of the
paper and successively covered with a foam pad and a rigid plastic grid.
Bubbles between gel and paper are carefully removed by rolling a pipette
over the gel surface. Finally, the gel is fixed in close contact with the
paper by fastening the plastic grids with rubber bands (Fig. 1). When
blotting is performed by passive diffusion, the paper should be beneath
the gel and on top of a dry foam pad so as to absorb the borate buffer
impregnating both the gel and the foam pad that lies on top of it. Most of
the protein diffusing from the gel within 4 hr reacts and becomes attached
to the paper. The rate of transfer may be increased by replacing the pad
underneath the blotting paper as many times as needed.
The blotting of protein by electrophoresis can be carried out by placing
the assembly of gel and paper, with the paper facing the anode, in an
electrophoresis tank filled with 10 mM sodium borate at pH 9.2. Transfer
is performed at 400 V (300-400 mA) to produce a voltage gradient of

PLASTIC GRID --C" I I I I I 1 1

FOAM PAD iiiiiiii~iii~i~iiiiiiii!iiii~i~i~iiiiiiii:i:i:i:i:ii~i~i~iiiiiiii:i~i~i:i:iiii~i~iiii:iii~i~i:i:i:i:iiiiii~iiii:i~i~i~i~i:i:i:~iiiiiiiiii:i~i~i~i~i:i:iiii~:i
I

POLYACRYLAMIDE GEL
BLOTTING PAPER ID

FOAM PAD -'- Ii!!!iii!~i!i!i!iii!i!i!!~i:i:i:!:!:!:!:i:i:i!i!i!ii~:!:~:~:~:~:!:i:i:i:i:i:i:!:~:~:~:~i!ii!:i:i:i:i:i:~ii!i!ii!i!i:i~i~iiii!ii:i:i~i~i~i~i~iiiii~i~i~i~i~i~i!~

PLASTIC GRID I I [--I I I

FIG. 1. Schematic of the protein-blotting sandwich used in electrotransfer experiments.
458 RELATEDTECHNIQUES [33]
32 V/cm for 1 hr at room temperature. 6 Lowering the pH of the buffer to
between 8 and 8.5 decreases the current and allows the voltage gradient to
be obtained with standard power supplies.

Other Transfer Media

In addition to DPT and DBM paper, cyanogen bromide-activated pa-
per 8 and cyanuric chloride paper 9 have been used to form stable covalent
bonds with the blotted proteins.
Whenever the formation of covalent bonds between protein and blot-
ting support is undesirable, nitrocellulose sheets, ~° DEAE paper, H or pa-
pers with attached affinity ligands for the protein under scrutiny may be
used. Although other buffers may be required for blotting onto these
supports, the method is essentially similar to that described for DPT and
DBM papers. Blotting of protein onto nitrocellulose can be performed
electrophoretically by immersing the assembly of gel and paper in 0.7%
acetic acid and using a voltage gradient of 6 V/cm for 1 hr at room temper-
ature. 10

Protein Identification
Finally, the specific protein under scrutiny must be singled out from
among all others that were blotted by using specific probes. For this
purpose, the blotting supports must first be treated in order to inactivate
remaining diazonium groups in DPT and DBM papers. The paper is
soaked for 15 min at room temperature with gentle rocking ~in 0.1 M Tris-
HC1 at pH 9.0 containing 10% ethanolamine and 0.25% gelatin (w/v).
Unreacted sites on nitrocellulose may be saturated with gelatin by soak-
ing the membrane in 10 mM Tris-HCl at pH 7.4, containing 0.9% NaC1
and 3% gelatin, for 1 hr at 40° with continuous rocking. 1° Bovine serum
albumin may be substituted for gelatin, but, owing to its frequent contami-
nation with IgG, its use should be avoided whenever it is anticipated that
protein A will be applied.
The most commonly used probes for protein are antibodies, since they
usually bind with high specificity and affinity to the denatured antigenic
protein covalently ligated to the blotting support. Whole sera, various
immunoglobulins fractionated with ammonium sulfate and further purified
by DEAE-chromatography, or monoclonal antibodies can be used as pro-

s R. A. Hitzeman, L. Clarke, and J. Carbon, J. Biol. Chem. 255, 12073 (1980).

9 H. D. Hunger, H. Grutzman, and C. Cotelle, Biochim. Biophys. Acta 653, 344 (1981).
l0 H. Towbin, T. Staehelin, and J. Gordon, Proc. Natl. Acad. Sci. U.S.A. 76, 4350 (1979).
11 T. McLellan and J. A. M. Ramshaw, Biochem. Genet. 19, 647 (1981).
[33] WESTERN BLOTS 459

tein probes. The amount of antibody used will vary with the titer. Anti-
body is added to the blotting support bathed in a saline buffer (10 mM
Tris-HCl, 0.9% NaCl, 0.1% gelatin, and 0.04% sodium azide at pH 7.4).
Routinely, a paper of the size of a standard slab gel is soaked for 2 to 24 hr
at room temperature in 30 ml of the saline buffer (50 to I00/zl buffer/cm 2)
containing either 30 to 100/zl of whole immune serum, 0.1-0.2 mg of
purified IgG, or 5 to 10/.tg of a high-affinity monoclonal antibody. Un-
bound antibody is washed thereafter with six changes of 200 ml of the
saline buffer for 1 hr each, at room temperature before labeling the anti-
body bound to the protein with a second labeled probe. The antibody is
most frequently treated with protein A from Staphylococcus a u r e u s A 1E
or with anti-antibodies beating iodine-125. ~°,~3The reaction between the
antibody and the labeled probe is performed by soaking the blot in 30 ml
of saline buffer containing 2 to 5 x 106 cpm 125I-labeled (100 Ci/mol) of
either protein A or anti-antibody for 4 hr at room temperature.
Anti-antibodies labeled with a photofluore such as rhodamine, ~° or
conjugated to horseradish peroxidase, ~° can be used with proteins blotted
onto nitrocellulose. After labeling, the unbound labeled probe is removed
by washing with the saline buffer (eight changes of 200 ml for 6 hr each)
and dried. Blots labeled with 125I a r e exposed to Kodak X-Omat film.
Nitrocellulose blots labeled with photofluores are inspected under
longwave UV light through a yellow filter.
Protein blotted onto DPT or DBM paper or onto nitrocellulose sheets
can be tested repeatedly with the same or new probes after removing the
old probe from the blotting support.l To remove either protein A or anti-
bodies bound to blotted proteins, the papers are incubated with l0 mM
Tris-HC1 at pH 7.5 containing 0.1 M 2-mercaptoethanol and 10 M urea,
for 30 min at 60°. ~ Alternatively, the paper is washed with 10 mM Tris-
HC1 at pH 7.5, containing 3 M potassium thiocyanate, for 30 min at 37°
and then with 10 mM HCI for an additional 30 min at room temperature. 6
An equally effective method of removing an old probe is by soaking the
paper in 0.25 M glycine-HC1 at pH 2.8 for 10 min at room temperature.
After removal of the probes by either method, papers are neutralized with
50 ml of 1 M Tris-HC1 at pH 7.5 for 30 min and washed with saline buffer
so that they may be probed again. Nitrocellulose blots can be cleaned
from old probes by incubation with 8 M urea, 0.1 M 2-mercaptoethanol,
and 5 mg bovine serum albumin per milliliter at 60° for 1 hr. Before
incubation with a new probe, the blots must always be reequilibrated with
the saline buffer.

12 S° W° Kessler, J. Immunol. 115, 1617 (1975).

is W. N. Burnette, Anal. Biochem. 112, 195 (1981).
460 RELATED TECHNIQUES [34]

The use of probes that bind only to native protein, rather than to a
denatured form as presented after SDS treatment, requires skill. Few
general rules can be given. The proteins must be resolved by electropho-
resis under nondenaturing conditions (i.e., omitting both SDS and urea)
or, if resolved under denaturing conditions, require renaturation. Partial
renaturation of some proteins subsequent to SDS-polyacrylamide gel
electrophoresis has been achieved by soaking the gels in buffers known to
stabilize them; generally 20% glycerol and 200 mM NaC1 are included. TM
The use of covalent attachment of proteins to DPT and DBM paper
may hinder the binding of the probe to protein but need not do so. In case
of interference, protein may be blotted onto nitrocellulose sheets, 1°
DEAE paper,11 or papers having attached affinity ligands for the protein. 15

Comments
The blotting of proteins resolved by gel electrophoresis onto paper has
been used for sorting antibodies interacting with them from whole se-
rum. 16This allows the immunization of animals with complex mixtures of
proteins containing one or more antigens of interest without lengthy purifi-
cation of antigen. Similarly, proteolytic digests of proteins resolved by gel
electrophoresis could be blotted onto solid supports and the antigenic
determinants of the protein characterized.

14 R. E. Manrow and R. P. Dottin, Anal. Biochem. 120, 181 (1982).

15 H. A. Ehrlich, S. N. Cohen, and H. O. McDevitt, Cell 13, 681 (1978).
t6 j. B. Olmstead, J. Biol. Chem. 256, 11955 (1981).

[34] F l u o r o g r a p h y for t h e D e t e c t i o n o f R a d i o a c t i v i t y in Gels

By WILLIAM M. BONNER

Fluorographic techniques can greatly increase the efficiency of radio-

active isotope detection. In these techniques, scintillants or fluors are
placed in the path of the radioactive disintegration. Light emitted by the
fluors in response to a radioactive disintegration can then be detected on
film. Because the wavelength of the emitted light depends on the fluor, not
on the energy of the radioactive disintegration, the same film can be used
to detect the small range of photoenergies from isotopes of widely varying

METHODS IN ENZYMOLOGY,VOL. 104 ISBN 0-12-182004-1

460 RELATED TECHNIQUES [34]

The use of probes that bind only to native protein, rather than to a
denatured form as presented after SDS treatment, requires skill. Few
general rules can be given. The proteins must be resolved by electropho-
resis under nondenaturing conditions (i.e., omitting both SDS and urea)
or, if resolved under denaturing conditions, require renaturation. Partial
renaturation of some proteins subsequent to SDS-polyacrylamide gel
electrophoresis has been achieved by soaking the gels in buffers known to
stabilize them; generally 20% glycerol and 200 mM NaC1 are included. TM
The use of covalent attachment of proteins to DPT and DBM paper
may hinder the binding of the probe to protein but need not do so. In case
of interference, protein may be blotted onto nitrocellulose sheets, 1°
DEAE paper,11 or papers having attached affinity ligands for the protein. 15

Comments
The blotting of proteins resolved by gel electrophoresis onto paper has
been used for sorting antibodies interacting with them from whole se-
rum. 16This allows the immunization of animals with complex mixtures of
proteins containing one or more antigens of interest without lengthy purifi-
cation of antigen. Similarly, proteolytic digests of proteins resolved by gel
electrophoresis could be blotted onto solid supports and the antigenic
determinants of the protein characterized.

14 R. E. Manrow and R. P. Dottin, Anal. Biochem. 120, 181 (1982).

15 H. A. Ehrlich, S. N. Cohen, and H. O. McDevitt, Cell 13, 681 (1978).
t6 j. B. Olmstead, J. Biol. Chem. 256, 11955 (1981).

[34] F l u o r o g r a p h y for t h e D e t e c t i o n o f R a d i o a c t i v i t y in Gels

By WILLIAM M. BONNER

Fluorographic techniques can greatly increase the efficiency of radio-

active isotope detection. In these techniques, scintillants or fluors are
placed in the path of the radioactive disintegration. Light emitted by the
fluors in response to a radioactive disintegration can then be detected on
film. Because the wavelength of the emitted light depends on the fluor, not
on the energy of the radioactive disintegration, the same film can be used
to detect the small range of photoenergies from isotopes of widely varying

METHODS IN ENZYMOLOGY,VOL. 104 ISBN 0-12-182004-1

[34] FLUOROGRAPHIC DETECTION OF RADIOACTIVITY 461

energies more efficiently than would be possible if the fiim had to interact
directly with those isotopes.

Weak fl-Emitters in Polyacrylamide Gels

For 3H, 14C, 33p, and 35S, the fluor must be placed inside the gel since
most of their radiation never emerges from the gel to expose the film.
However, most fluors are hydrophobic and not soluble in the hydrophilic
compounds used for keeping polyacrylamide gels solvated. Dimethyl
sulfoxide (DMSO) was found to have the necessary properties; 2,5-di-
phenyloxazole (PPO) dissolved in DMSO can be easily and efficiently
introduced into polyacrylamide gels. For 14C, 33p, and 35S, the increase in
sensitivity over autoradiography is approximately 15-fold. For 3H, the
enhancement in sensitivity is in excess of 1000-fold. The detection of 3H
in polyacrylamide gels is a major achievement of fluorography.
Original P P O - D M S O 1 and Related Procedures. Although DMSO and
other compounds used in fluorography have no known mutagenicity or
carcinogenicity, they are not without potential hazard, and appropriate
precautions should be used. DMSO readily penetrates the skin and may
facilitate the entry of dissolved substances such as PPO. Mixtures based
on glacial acetic acid must also be used with caution. The procedures
described here should be carried out in a hood and with gloves.
After electrophoresis, the gel may be either stained or processed di-
rectly for fluorography. Formaldehyde may be used to fix small proteins
and peptides in the gel. 2 All soaking steps should be performed with gentle
shaking in closed containers sufficiently large to contain the gel without
folding. Plastic freezer trays with tight-fitting lids can readily be obtained
for this purpose.
The following procedure is for 0.8-mm-thick gels (increase by a factor
of 3 for 1.5-ram-thick gelsl). The gels are soaked (1) twice for 10 rain each
in DMSO to remove water; (2) once in four volumes of a 22% (w/v)
solution of PPO in DMSO for 1 hr; (3) twice for 10 min each in water to
precipitate PPO within the gels. Gels are dried for about 1 hr under vac-
uum at 80° . Commercial vacuum gel dryers are available for this purpose.
For exposure overnight using unflashed film, sensitivities are approxi-
mately 20 dpm/mm 2 for 14C, 35S, and 33p and 300 dpm/mm2 for 3H.
Glacial acetic acid may be substituted for DMSO as the solvent; under
such conditions, the gels may be immersed directly in 22% PPO (w/v) in
glacial acetic acid, eliminating step 1. The staining pattern of the gel is

i W. M. Bonner and R. A. Laskey, Eur. J. Biochem. 46, 83 (1974).

2 G. Steck, P. Leuthard, and R. R. Biirk, Anal. Biochem. 107, 21 (1980).
462 RELATEDTECHNIQUES [34]

more faithfully retained when acetic acid is used instead of DMSO. PPO
can be recovered from either solvent for reuse by adding water and col-
lecting the precipitated PPO on a filter)
Some time after publication of the original procedure using PPO dis-
solved in DMSO, a commercial preparation, Enhance (New England Nu-
clear), became available. Since Enhance is based on acetic acid, gels can
be soaked directly in it for 1 hr, then in two 10-min washes of water.
Enhance may be considerably less expensive than PPO-containing solu-
tions if one does not recover the PPO for reuse.
Aqueous Procedures. Water-soluble fluorography preparations are be-
coming available as water-soluble fluors are found or synthesized. The
original water-soluble fluorographic procedure uses sodium salicylate. 4
This method is simpler than, and may be as sensitive as, the methods just
described, but its resolution is significantly less than that obtained with
PPO-DMSO. It is useful primarily when a few widely spaced radioactive
bands are to be detected.
The following procedure is used for a 1.5-mm-thick gel. Gels at neutral
or alkaline pH may be added directly to 1 M (16% w/v) sodium salicylate
for 30 min, but stained gels need to be presoaked for 30 min in water to
remove acetic acid. The gel is removed from the sodium salicylate and
dried at 80 ° directly on paper under an acetate sheet. The usual Mylar
sheets should not be used, since they become glued to the dried gel.
Commercial water-soluble fluorography solutions have appeared. The
procedures are basically identical to that used with sodium salicylate, but
the resolution is better and appears to be as good as that obtained with
PPO-DMSO. Two preparations presently available are Autofluor (Na-
tional Diagnostics) and Enlightning (New England Nuclear). None of the
above methods is obviously the best for every situation; investigators
must choose according to their requirements.

Weak fl-Emitters in Agarose Gels

After they have been dehydrated in several changes of absolute etha-
nol, agarose gels can be soaked in a 3% (w/v) solution of PPO in absolute
ethanol for 3 hr. 5 The PPO is precipitated thereafter by immersing the gel
in water. During drying, the gel should not be heated or heated only gently
so that it does not melt.

3 R. A. Laskey, A. D. Mills, and J. S. Knowland, Appendix in R. A. Laskey and A. D.

Mills, Eur. J. Biochem. 56, 335 (1975).
4 j. p. Chamberlain, Anal. Biochem. 98, 132 (1979).
R. A. Laskey, A. D. Mills, and N. R. Morris, Cell 10, 237 (1977).
[34] FLUOROGRAPHIC DETECTION OF RADIOACTIVITY 463

Weak r-Emitters on Solid Porous Supports

~Efficient fluorography of 3H and 14C on solid supports such as thin-
layer plates, paper chromatograms, and nitrocellulose filters can be ob-
tained by dipping the completely dried support into a slightly warm solu-
tion (about 37 °) of 0.4% PPO in 2-methylnaphthalene (Aldrich Chemical
Co.). 6 The plate or sheet is allowed to stand or hang vertically so that the
excess solution can run off. The solution solidifies as it cools to ambient
temperature. The support can then be exposed to flashed or unflashed film
at - 7 0 °. One caution is that plates should not be left uncovered over-
night, since the 2-methylnaphthalene is volatile. The levels of detection
are similar to those for polyacrylamide gels.

S t r o n g r - a n d T - E m i t t e r s in P o l y a c r y l a m i d e G e l s o r o n S o l i d S u p p o r t s

For strong r-emitters (32p) and T-emitters (125I), most of the radiation
passes through the film without exposing it. By means of an intensifying
screen on the other side of the film, some of this radiation is converted
into light, which then passes back into the film, exposing it. 7 Film cas-
settes can be purchased with intensifying screens permanently attached to
one or both faces. In general, calcium tungstate intensifying screens are
most suitable. Dried gels should be exposed at -70°; wet gels are better
exposed at 0 °, because freezing may crack them. In some cases, flashed
film and intensifying screens should not be combined, since some screens
contain enough endogenous radioactivity to blacken the film during a long
exposure. Approximately 1 dpm/mm 2 of 125I and 0.5 dpm/mm 2 of 32p can
be detected after exposure overnight.

Film Exposure Conditions

For fluorography the dried gel is placed in a cassette with the correct
film and exposed at -700. 8,9 The appropriate type of film is essential for
good results since the sensitivities of different types can vary by factors of
10.1 Kodak X-Omat AR or the equivalent from other manufacturers
should be used. Film images can be quantitated by microdensitometric
techniques, but, for small numbers of spots, another useful method is
merely to cut out the spots from the dried fluorographed gel. If the gel is
dried on thin paper and spotted around the edges with radioactive ink

6 W. M. Bonner and J. D. Stedman, Anal. Biochem. 89, 247 (1978).

7 R. A. Laskey and A. D. Mills, FEBS Lett. 82, 314 (1977).
s K. Randerath, Anal. Biochem. 34, 188 (1970).
9 U. Liithi and P. G. Waser, Nature (London) 205, 1190 (1965).
464 RELATEDTECHNIQUES [34]

(about 1/.~Ci of 14C per milliliter) the fluorograph can easily be placed in
register over the dried gel and the desired spots marked with pencil on the
paper backing while holding the film-gel sandwich in front of a strong
light. Spots are excised and placed in scintillation vials. Pieces of dried gel
infused with fluor will be recorded in a scintillation counter, but the geom-
etry must be kept constant for accurate quantitation. It is generally more
accurate to digest the pieces of dried gel overnight at 37° in 1 ml of a fresh
solution of 95 parts of 30% H202 and 5 parts of concentrated NH4OH in
capped scintillation vials. Plastic vials should be t~ed, since glass occa-
sionally explodes. The next morning, vials are cooled and a water-misci-
ble scintillation fluid is added. Since this is a homogeneous system, inter-
nal standards can be added to calculate counting efficiency and, hence,
the absolute amount of radioactivity in the area. This method also works
well for areas in which different isotopes are present, i.e., 3H and ~4C.
Flashing Film to Linearize Its Response.l° The response of film to
light is nonlinear. At low light intensities, the density of the resultant
image is approximately proportional to the square of the radioactive den-
sity. The response of the film can be made to approach linearity by expos-
ing it to a hypersensitizing light flash. A single flash of less than 1 msec
duration from an electronic flash unit is suitable, but it is usually neces-
sary to decrease its intensity with a Kodak Wratten 21 or 22 (orange) filter
and several layers of exposed film. In addition to the hypersensitized
grains, some of the grains in the film are also exposed by the flash, result-
ing in an increased background fogging. When the fogging density is about
0.15 ~, the film response is nearly linear. The distance between the film
and flash unit should be varied until that density is obtained. The density
can easily be measured by developing a piece of flashed film, cutting out a
small piece, and placing it in the cuvette holder of a spectrophotometer.
Since exposed film acts as a neutral density filter, any visible wavelength,
such as 650 nm, can be used. Once these conditions are ascertained, the
film is placed on a yellow background and flashed; its flashed face is then
placed against the gel.
Flashed film is useful for accurate microdensitometric quantitation of
spots as well as for two situations in which accurate quantitation may not
be required. When the areas of interest are near the minimum level of
detection, flashing the film results in a two- to threefold increase in rela-
tive density for them. 6a° This is because the nonlinear response of un-
flashed film underrepresents the faintest spots. When storage at - 7 0 ° is
unavailable, 3H and 14C fluorography is possible with flashed film but not
unflashed film at ambient temperature; however, the sensitivity is about
half that at - 7 0 ° .
lo R. A. Laskey and A. D. Mills, Eur. J. Biochem. 56, 335 (1975).
[34] FLUOROGRAPHIC DETECTION OF RADIOACTIVITY 465

Gel Preparation for Easier Fluorography

Fluorography is facilitatedif two factors are considcrcd cvcn before
acrylarnidc gels arc prcparcd. First, gels should be elastic for casc in
handling and resistance to cracking during drying. Blattlcr et al.l~ found
that, if the final acrylarnidc pcrccntagc (A) tirncs the final bisacrylarnidc
perccntagc (B) satisfiesthe cquation A B = 1.3, the gels will have satisfac-
tory elasticity throughout the range from 4 % acrylarnidc to 40%. This
laboratory has routinely used the formula A B = 1.5 for gels from 5 to
60% acrylarnidc and found these gels to have good elasticity and rcsis-
tancc to cracking. A n S D S gcl with 15% acrylarnidc and 0.1% bis-
acrylarnidc follows this formula and scparatcs proteins greater than
10,000 daltons. Stacking gels that arc rcrnovcd bcforc fluorography do not
nccd to follow this formula.
A second factor to bc considered is gel thickncss. Since the tirnc
rcquircd for a substance to diffuse a givcn distance is proportional to thc
square of that distancc (Fick's second law), thc use of thinncr gcls would
greatly decrease the time necessary for infusion with fluors and solvents.
In practicc, gels of 0.8 rnrn satisfyingthe formula A B = 1.5 can bc handlcd
easily and can bc infused with scintillantin less than I hr instead of thc
3 hr rcquircd for 1.5-rnrn gcls. Spaccrs and slot combs for such gels
can bc cut casily with scissors or a razor blade from skivcd Teflon tapc
(0.8 rnrn = 0.031 in.),availablc from most plastic supply houscs.

11 D. P. Blattler, F. Garner, K. Van Slyke, and A. Bradley, J. Chromatogr. 64, 147 (1972).
 SUBJECT INDEX 503

Subject Index

A Acid protease, affinity chromatography, 22

Aconitase, gel staining, 432
Aconitate hydrolase, s e e Aconitase
ACES, s e e N-2-Acetamido-2-aminoethane- Acrosin, gel staining, 429
sulfonic acid Actin, affinity chromatography, 50
N-2-Acetamido-2-aminoethanesulfonic Activation, 4-14, s e e a l s o specific com-
acid, pK, 405 pounds
N-2-Acetamidoiminodiacetic acid, pK, 405 Acyl carder protein, reversed-phase chro-
Acetate buffer, pK, 405 matography, 205
Acetic acid, membrane protein, solubiliza- ADA, s e e N-2-Acetamidoiminodiacetic
tion, 334 acid
Acetone, protein crystallization, 374 Adenine phosphoribosyltransferase
Acetylcholine, affinity chromatography, 50 affinity chromatography, 22
Acetylcholine receptor gel staining, 425
reconstitution, 342, 343 Adenosine deaminase
solubilization, 341,343 affinity chromatography, 47
Acetylcholinesterase, affinity chromatogra- gel staining, 430, 436
phy, 22 Adenosine kinase, affinity chromatogra-
Acetyl-CoA carboxylase, affinity chroma- phy, 22
tography, 22 Adenosine monophosphate deaminase, gel
N-Acetylgalactosamine transferase, affinity staining, 436
chromatography, 22 Adenosine (phosphate) deaminase, affinity
a-N-Acetylgalactosaminide-(2 --~ 6)-sialyl- chromatography, 22
transferase, affinity chromatography, Adenosine 5'-phosphosulfate sulfohydro-
22 lase, affinity chromatography, 22
Acetyl-fl-glucosaminidase, affinity chroma- Adenosinetriphosphatase,
tography, 47 affinity chromatography, 22
N-Acetylglucosaminidase, affinity chroma- DNA-dependent, affinity chromatogra-
tography, 22 phy, 24
fl-N-Acetyl-o-hexosaminidase, purifica- Adenylate cyclase
tion, 115, 116 affinity chromatography, 22
N-Acetyl-fl-hexosaminidase, affinity chro- gel staining, 426
matography, 22 Adenylosuccinate synthetase, dye-ligand
N-Acetyl-fl-hexosaminidase A, affinity chromatography, 109., 110
chromatography, 22 N2,N2'-Adipodihydrazidobis(N6-carbonyl-
Acetyltransferase, dye-ligand chromatogra- methyl-NAD), 364, 366
phy, 97 ADP-glucose pyrophosphorylase, s e e
Acid D-galactosidase, affinity chromatogra- Glucose-l-phosphate adenylyltrans-
phy, 25 ferase
Acid phosphatase ADP-ribosyltransferase dehydrase, affinity
affinity chromatography, 22 chromatography, 22
gel staining, 428 a2-Adrenergic receptor, affinity chroma-
reversed-phase chromatography, 209 tography, 50
504 SUBJECT INDEX

/3-Adrenergic receptor factors determining effect, 359, 360

affinity chromatography, 47, 50 optimization, 358, 359
reconstitution, 344-347 with polyethylene glycol, 356-364
solubilization, 341 polymer removal, 363
Adrenocorticotropin, reversed-phase purification, 356-364
chromatography, 208 strategy, 360-362
Adsorption, hydrophobic chromatography, two-phase system, 358
80-84 Affinity precipitation, 364-369
AES, s e e 2-Aminoethylsulfonic acid dehydrogenases, 364-369
Affinity chromatography, 3-233 in gels, 368, 369
antigens, 21, 47, 48, 54 pilot system, 366, 367
with dyes, 97-113 selectivity, 365
elution, 19, 20 Agarose
of enzymes, 21-46, s e e a l s o specific activated
enzyme coupling, 17, 18
HPLC, 212-223 quantitation, 14
adsorbents, 212-217 activation, with sulfonyl chlorides, 58,
affinity matrix preparation, 215 59
column packing, 218 dye coupling, 102-105
epoxy-silica, 213,215 fluorography, 462
lectin matrix, 217, 219 isoelectric focusing, 263
ligands, 219 ligand-substituted, 280
procedures, 218 preparation of tresylated support, 58,
silica-ligand, 214, 215 59
quantitation of bound ligand, 217 Alanine aminotransferase, affinity chroma-
use of lectins, 54, 55 tography, 22
of organic sulfonyl chlorides, 56-69 D-Alanine carboxypeptidase, affinity chro-
Affinity column, s e e a l s o specific type matography, 22
activation, 6-18 Alanine dehydrogenase, affinity chroma-
adsorption, 18, 19 tography, 22
elution, 19, 20 Albumin
Affinity electrophoresis, 275-281 affinity chromatography, HPLC, 219
agarose, ligand-substituted, 280 isotachophoresis, 294, 296
dextran, ligand.substituted, 280 reversed-phase chromatography, 205,
ligand-agarose, 277, 278 2O7
ligand preparation, 279-281 Alcohol dehydrogenase
limitations, 278, 279 affinity chromatography, 22, 67
polyacrylamide gel, 276, 277 HPLC, 219
procedure, 280, 281 reversed, 220-223
qualitative applications, 278, 279 affinity precipitation, 369
quantitative applications, 279 coupling, 64, 65
Affinity high-performance liquid chroma- to tosyl-agarose, 61, 62
tography, s e e Affinity chromatogra- HPLC, 216, 217
phy, HPLC Alcohol dehydrogenase-silica matrix, 220-
Affinity isoelectric focusing, s e e Isoelectric 223
focusing, affinity Aldehyde dehydrogenase, affinity chroma-
Affinity partitioning, 356-364 tography, 22
conditions, 358, 359 Aldehyde reductase, affinity chromatogra-
extractions phy, 22
preparative, 363 Aldehyde-silica, preparation, 215
repeated, 362, 363 Aldolase, gel staining, 432
 SUBJECT INDEX 505

Aldose reductase, affinity chromatography, a-Amylase

23 affinity chromatography, 23
Alkaline phosphatase gel staining, 428
affinity chromatography, 23, 47 Androgen, affinity chromatography, 50
dye-ligand chromatography, 97, 111 Angiotensin I converting enzyme, affinity
Alkylagarose chromatography, 47
capacity, 78, 79 Anthranilate synthase
derivatives, functional groups, 73 affinity chromatography, 23
flow rates, 78, 79 gel staining, 432
homologous series, 70-73, 90, 91 Anthraquinone dye, 99
mechanism, 92-96 Antibody, s e e a l s o specific type
preparation, 73-75 monocional
stability, 78, 79 affinity, 384, 385
structure, 74 affinity chromatography, 337
ultrastructural identity, 79, 80 chromatography, 387
Alkylsilica, reversed-phase chromatogra- crude antigens, 382
phy, 190-212 homogeneous antigen, purification,
Amide, coupling, 15 382, 383
Amine oxidase immobilization, 386, 387
affinity chromatography, 23, 27, 29, 47 vs. polyclonal, 385, 386
gel staining, 422 protein purification, 381-387
D-Amino-acid oxidase, gel staining, 422, specificity, 383
436 stability, 384
L-Aminoacyl-tRNA synthetase, dye-ligand protein purification, 381-387
chromatography, 97 Antibody-silica matrix, procedure, 220
ta-Aminoalkylagarose, 70 Aprotinin, reversed-phase chromatogra-
2-Aminoethylsulfonic acid, pK, 406 phy, 205
N*-(6-Aminohexyl)-5'-adenosinemonophos - APS, s e e 3-Aminopropanesulfonic acid
phate L-Arabinose kinase, affinity chromatogra-
affinity chromatography, 65, 66 phy, 23
coupling to tosyl agarose, 61 Arginase
to tresyl-agarose, 60 affinity chromatography, 23
Aminolevulinate dehydratase, s e e Porpho- antibody, 23
bilinogen synthase gel staining, 430, 436
Aminopeptidase, affinity chromatography, Arginine carboxypeptidase, affinity chro-
23, 47 matography, 23
3-Aminopropanesulfonic acid, pK, 406 L-Arginine deaminase, affinity chromatog-
2-Aminothiophenol, 456 raphy, 23
Aminotransferase, dye-ligand chromatogra- Argininosuccinase, s e e Argininosuccinate
phy, 97 lyase
Ammonia, determination, for resins, 9 Argininosuccinate, gel staining, 433
Ammonia buffer, pK, 406 Argininosuccinate lyase
Ammonium persulfate, substitutes, 240 gel staining, 436
Amphiphile, 329 simultaneous capture technique, 436
Ampholyte, 268 Aromatic-amino-acid decarboxylase, gel
commercial, 260 staining, 432
enzyme staining, 418 Arylsulfatase, gel staining, 428, 439
high molecular weight, 260, 261 Arylsulfatase A, affinity chromatography,
properties, 260 23
Amunine, reversed-phase chromatography, Asialoglycoprotein, affinity chromatogra-
209 phy, 50
506 SUBJECT INDEX

Asparaginase, affinity chromatography, 23 Borate buffer, pK, 406

Aspartase, s e e Aspartate ammonia-lyase Boronic acid-silica, 219
Aspartate aminotransferase, gel staining, Bovine serum albumin
425 affinity chromatography, 67
Aspartate ammonia-lyase, affinity chroma- retention mapping, 228
tography, 23 Brij 35, properties, 321
Aspartate carbamoyltransferase Buffer, 404-414, s e e a l s o specific type
affinity chromatography, 23 broad-range, 410, 411
gel staining, 425, 437,438 chelating effects, 409, 410
Aspartate-/3-decarboxylase, affinity chro- containers, 407
matography, 23 dependence on temperature, 405-407
D-Aspartate oxidase, gel staining, 422, 436 deuterated, 413-415
Aspartate transcarbamylase, s e e Aspartate effect of low temperature, 412
carbamoyltransferase of solvents, 411,412
Astroglial protein, affinity chromatogra- electrofocusing, 261
phy, 47 electrophoresis, 240, 242, 243
ATP-dependent enzyme, dye-ligand chro- evaluation, 408, 409
matography, 97 interactions, 406
Azo dye, 99 membrane protein solubilization, 310
metal ions, 409, 410
B NMR studies, 414
pK, 404-408
Bacteriorhodopsin, 329 selection, 404-408
crystallization, 377 n-Butanol, membrane protein solubiliza-
reversed-phase chromatography, 210 tion, 333, 334
Barbituric acid, for cyanate assay, 10 t e r t - B u t a n o l , protein crystallization,
Benzidine, 417 374
BES, s e e N , N-Bis(2-hydroxyethyl)-2-
aminoethanesulfonic acid C
Bicine, s e e N , N-Bis(2-hydroxyethyl)-
glycine Cacodylate, s e e Dimethylarsinic acid
BIGCHAP, properties, 309 Calmodulin
N, N'-Bis-3-(dihydroxyboronylbenzene)- affinity chromatography, 50
adipamide, 369 reversed-phase chromatography, 210
N, N-Bis(2-hydroxyethyl)-2-aminoethane- CAPS, s e e 3-(Cyclohexylamino)propane-
sulfonic acid, pK, 405 suifonic acid
3-[N-Bis(hydroxyethyl)amino]-2-hydroxy- Carbamoyl-phosphate synthetase, affinity
propanesulfonic acid, pK, 405 chromatography, 23
N,N-Bis(2-hydroxyethyl)glycine, pK, 406 Carbonate buffer, pK, 405, 406
[Bis(2-hydroxyethyl)imino]tris(hydroxy- Carbonate dehydratase
methyl)methane, pK, 405 affinity chromatography, 23
1,4-Bis(4-sulfobutyl)piperazine, pK, 406 retention mapping, 228
N, N '-Bis(3-sulfopropyl)ethylenediamine, reversed-phase chromatography, 210
pK, 405,406 Carbonic anhydrase, s e e Carbonate dehy-
1,4-Bis(3-sulfopropyl)piperazine, pK, 406 dratase
Bis-Tris, s e e [Bis(2-hydroxyethyl)imino]- 1,1 '-Carbonyldiimidazole, 12
tris(hydroxymethyl) methane activation, 12
1,3-Bis[tris(hydroxymethyl)methylamino]- coupling, 17
propane, pK, 405 l, 1 '-Carbodiimide
Bis-Tris propane, s e e 1,3-Bis[tris(hydroxy- activation, 12
methyl)methylamino]propane derivatization, 12
 sUBJECT INDEX 507

Carboxymethyldextran Chloroform-methanol, membrane protein

binding to DEAE, 121, 123 solubilization, 334
characterization, 122-124 Chlorotriazine dye, 99
degree of substitution, 122-124 3-(3 '-Cholamidopropyl)dimethylammonio-
as displacer, 118-133 1-propane sulfonate
principle, 118-122 properties, 309
preparation, 120-122 solubilization, 307, 312, 314, 332, 341
removal, from proteins, 131, 132 structure, 308
with serum proteins, 124, 125 Cholic acid, solubilization, 312, 332, 341
Carboxypeptidase, gel staining, 429 Choline acetyltransferase, affinity chroma-
Carboxypeptidase A, affinity chromatogra- tography, 23
phy, 23 Choline dehydrogenase, affinity chroma-
Carboxypeptidase B, affinity chromatogra- tography, 23
phy, 23 Cholinephosphate cytidylyltransferase,
Carboxypeptidase G2, dye-ligand chroma- affinity chromatography, 24
tography, 97, 111 Choriogonadotropin, affinity chromatogra-
Carboxypeptidase N, s e e Arginine car- phy, 47
boxypeptidase Chorismate mutase-prephenate-dehydro-
N-(3-Carboxypropionyl)aminodecyl- genase, affinity chromatography, 24
agarose, HPLC, 339 Chorismate synthase, affinity chromatogra-
Carcinoembryonic antigen, affinity chro- phy, 24
matography, 47 Chromatin, nonhistone, affinity chromatog-
Carcinogenic agent, avoidance, 417 raphy 51
Carcinoma antigen, isotachophoresis, 294 Chromatofocusing, 172
Cardiotoxin, affinity chromatography, 47 columns, 182
Catalase gradient, 182
affinity chromatography, 47 HPLC, 181-183
gel staining, 422 polybuffers, 183
Catechol O-methyltransferase, affinity polyethylenimine, 182, 183
chromatography, 23 protein fractionation, pl, 181-183
Cathepsin, affinity chromatography, 23 Chromatography, s e e a l s o specific type
Cathepsin B, gel staining, 429 HPLC, 133-233, s e e a l s o High-perfor-
C3b component, affinity chromatography, 50 mance liquid chromatography
Cellulase hydrophobic, 69-96
affinity chromatography, 23, 47 adsorption, 80-84
preparation of tresylated support, 58, 59 alkylagarose preparation, 73-75
Cetylammonium bromide, solubilization, capacity, 78, 79
332 column selection, 80-90
Cetylthiomethylammonium bromide effect of chain length, 70
gel, 254 of pH, 89
properties 321 of solvent, 86-90
Chaotropic reagent, 19 elution, 80-84
CHAPS, s e e 3-(3 '-Cholamidopropyl)di- techniques, 86-90
methylammonio- 1-propane sulfonate flow rates, 78, 79
CHES, s e e Cyclohexylaminoethanesul- homologous series, 70-73, 90, 91
fonic acid ligand density, 75-78
2-Chloroethanol, membrane protein solubi- matrix preparation, 73-75
lization, 334 mechanism, 92-96
Chloroformate principle, 72
activation, 12-14 purification, 71
coupling, 17 resolving power, 71
508 SUBJECT INDEX

stability, 78, 79 Corrinoid enzyme, affinity chromatogra-

use of chaotropic agents, 82 phy, 24
Chymotrypsin Corticosteroid-binding globulin, affinity
affinity chromatography, 24 chromatography, 50
coupling to tresyl-agarose, 64 Corticotropin, reversed-phase chromatog-
ionic strength effects, 354 raphy, 205
retention map, 174 Coupling, s e e specific type
Chymotrypsinogen C-reactive protein, affinity chromatogra-
ionic strength effects, 354 phy, 50
reversed-phase chromatography, 209 Creatine kinase
Cibacron Blue F3G-A affinity chromatography, 24
affinity chromatography, 97, 111,219 gel staining, 426
structure, 100 Crystallin, isotachophoresis, 294
Citrate buffer, pK, 405 Crystallization
Citrate synthase precipitation conditions, 370
affinity chromatography, 24 of proteins, s e e Protein crystallization
gel staining, 432 Cyanate ester, s e e a l s o specific substance
Clostripain, hydrophobic chromatography, analysis, 8-11
85, 87, 88 determination, for resins, 10
Coenzyme A-dependent enzyme, dye- molecular structure, 5
ligand chromatography, 97 l-Cyano-4-dimethylamino pyridinium
Collagen tetraltuoroborate, molecular structure,
PEG precipitation, 355 5
reversed-phase chromatography, 205, Cyanogen bromide
208 activation, 4-11, 70
Collagenase coupling capacity, 11
affinity chromatography, 24 with l-cyano-4-dimethyl aminopyri-
hydrophobic chromatography, 85, 87, dinium tetrafluoroborate, 7, 8
88 with N-cyanotriethylammonium, 7, 8
Collagen galactosyltransferase, affinity mechanism, 4, 6
chromatography, 24 with p-nitrophenyl cyanate, 7
Collagen glucosyltransferase, affinity with triethylamine, 6
chromatography, 24 formation of active groups, 4
Collagen glycosyltransferase, affinity ligand leakage, 11
chromatography, 24 Cyanogen brurnide-activated paper, 458
Column, s e e specific type, system Cyano-transfer reagent, 6
Complement component, affinity chroma- N-Cyanotriethylarnmonium, molecular
tography, 47 structure, 5
Complement factor, dye-ligand chromatog- Cyanuric chloride paper, 458
raphy, 97 Cyclic adenosine monophosphate, affinity
Conaibumin, retention mapping, 174, 228 chromatography, 50
Concanavalin A, 22-24 Cyclic nucleotide, affinity chromatography,
affinity chromatography, 65, 66 50
HPLC, 219 Cyclic nucleotide phosphodiesterase,
gel staining, 450 affinity chromatography, 24, 47
HPLC, 217 Cyclohexylaminoethanesulfonic acid, pK,
Concanavalin A-silica matrix, procedure, 406
220 3-(Cyclohexylamino)propanesulfonic acid,
Coomassie Brilliant Blue G250, 440, 441 pK, 406
Coomassie Brilliant Blue R250, 439 Cysteine S-conjugate N-acetyltransferase,
Coomassie protein stain, 439, 440 solubilization, 307
 SUlUECT INDEX 509

Cytidine deaminase, gel staining, 430, 436 Depth filtration, water purification, 393,
Cytochrome c 394
retention map, 174 Desalting, 203
reversed-phase chromatography, 205, Desorption, electrophoretic, 19
209 Detergent, s e e a l s o specific type
Cytochrome c oxidase anionic, HPLC, 138
affinity chromatography, 24 cation binding, 309
reversed-phase chromatography, 209 choice, 306-310
Cytochrome P-450, solubilization, 312, 313 critical micelle concentration, 309, 313-
Cytochrome reductase, affinity chromatog- 315, 321,322
raphy, 24 effect of lipids, 341
Cytokinin, affinity chromatography, 50 of salts, 341
Cytotoxic protease, affinity chromatogra- denaturation, 31 l
phy, 24 gels, 417
cationic, 253-255
D electrophoresis, Mr determination,
255
Dansyl derivative, 417 membrane protein solubilization, 305-
DEAE-BioGel, displacement chromatogra- 318
phy, 128 micelle
DEAE-cellulose, displacement chromatog- gel filtration, 326, 327
raphy, 128 Mr, 309, 321
DEAE-Sephacel, displacement chromatog- size, 320-323
raphy, 128, 129 molecular weight, 321
DEAE-Trisacryl M, displacement chroma- nonionic, 307, 333
tography, 125 properties, 309, 320-322
Decarboxylase, dye-ligand chromatogra- ionic, 321
phy, 97 nonionic, 321
Dehydrogenase protein crystallization, 377
affinity chromatography, HPLC, 218- protein-detergent ratio, 313-315
220 removal, 307, 318-328
affinity precipitation, 364-369 choice, 319, 320
dye-ligand chromatography, 97 detergent exchange, 325,326
Denaturation, with sodium dodecyl sulfate, gel filtration, 323-327
241 membrane protein, 318-328
Deoxy-BIGCHAP strategy, 323-325
solubilization, 307 ultrafiltration, 327, 328
structure, 308 separation, 322
Deoxycholic acid from protein, 318-328
removal, 319 solubilization, 311
solubilization, 332 sources, 308
Deoxynucleotidyltransferase Stokes radius, 320
affinity chromatography, 24 structures, 308
terminal, affinity chromatography, 32, 48 Detergent-protein micelle, 323
Deoxyribonuclease Dextran
affinity chromatography, 25 affinity chromatography, 50
gel staining, 428 ligand-substituted, 280
Deoxyribonucleic acid, affinity chromatog- Dextranase, gel staining, 428
raphy, 50 Dialysis, 372,375, 379
Deoxyribosylase, affinity chromatography, Diaminobenzidine, gel staining, 451
24 o-Dianisidine, 417
510 SUBJECT INDEX

Diazobenzyloxymethyl paper with DEAE-Sephacel, 128, 129

blotting, 456 with DEAE-Trisacryl M, 127
preparation, 456 elution, 115
Diazonium coupling, 15 high-performance liquid chromatog-
Diazophenylthio paper raphy, 132
blotting, 456 large-scale, 132, 133
preparation, 456 protein separation, 124-133
Dichlorotriazinyl dye, 99, 104 resolution, 128-131
Diethylene glycol monobutyl ether, mem- spacing, 125-128
brane protein solubilization, 335 spacing proteins, 120
Digitonin thin-layer, 133
properties, 309 Displacement electrophoresis, s e e Isota-
solubilization, 307, 341 chophoresis
structure, 308 Distillation, water purification, 398
Dihydrofolate reductase, s e e Tetrahydrofo- 5,5'-Dithiobis(2-nitrobenzoic acid), 17
late dehydrogenase DNA-dependent ATPase, s e e Adenosine-
Dihydrolipoamide reductase, affinity chro- triphosphatase, DNA-dependent
matography, 28 DNA-dependent RNA polymerase, s e e
Dihydropteridine reductase, affinity chro- RNA polymerase, DNA-dependent
matography, 24 DNA ligase, affinity chromatography, 24
1,2-Dihydroxyvitamin D, affinity chroma- DNA-nicking-closing enzyme, affinity
tography, 50 chromatography, 24
Dimethylarsinic acid, pK, 405 DNA nuclease, dye-ligand chromatogra-
N, N'-Dimethylbarbituric acid, for cyanate phy, 97
assay, 10 DNA polymerase
Dimethylformamide, membrane protein affinity chromatography, 24
solubilization, 335 dye-ligand chromatography, 97
Dimethylglutarate buffer, pK, 405 DNA polymerase fl, affinity chromatogra-
Dimethyl sulfoxide, protein crystallization, phy, 25
374 DNA-unwinding enzyme II, affinity chro-
Diol dehydrase, affinity chromatography, matography, 25
24 Dodecyl dimethylamine oxide, properties,
Diol-silica, preparation, 215 321
Dioxane, protein crystallization, 374 N,N-Dodecyldimethylamine N-oxide,
Dipeptidase, gel staining, 429 crystallization, 377
Dipeptidyl peptidase, affinity chromatogra- Dopamine fl-hydroxylase, s e e Dopamine
phy, 24 B-monooxygenase
DIPSO, s e e 3-[N-Bis(hydroxyethyl)- Dopamine B-monooxygenase, affinity
amino]-2-hydroxypropanesulfonic acid chromatography, 25
Disc gel electrophoresis, 237 Dye, s e e a l s o specific substance
Discontinuous electrophoresis, 244-255 chemistry, 98-101
buffers, 244, 245 chromatography, 102
cationic detergents, 253-255 coupling, 102-105
procedure, 244-253 to agarose, 102
Displacement chromatography, 113-133, to glass, 103
195, 196 to metal oxides, 102
ampholyte, 115-118 to polyacrylamide, 102
carboxymethyldextran spacers, 120, to silica, 103
125-128 immobilization, 97-113
with DEAE-BioGel, 128 nomenclature, 101
with DEAE-cellulose, 128, 129 parameters, 107
 SUBJECT INDEX 511

properties, 101, 102 gel, enzyme localization, 419, 420

purification, 101, 102 gradient, 199
sources, HPLC, 142
structure, 98-101 hydrophobic chromatography, 80-84
Dye affinity chromatography, HPLC, 219 ionic strength, 202-204
Dye-ligand chromatography, 97-I 13 isocratic, 198
applications, 109-113 isoelectric focusing, 274, 275
column properties, 107-109 nonvolatile solvents, 203
desorption, 108, 109 organic solvent composition, 200-202
dye concentration, 105 reversed-phase chromatography, 197-
dye selection, 105-107 199
effect of metal ions, 111 spacer arm destruction, 19
of zinc ions, Ill EMTA, s e e 3,6-Endomethylene-l,2,3,6-
elution, 108 tetrahydrophthalic acid
large scale, 109, 110 Emulphogen BC-720
operational parameters, 107-109 properties, 321
preliminary tests, 105 removal, 319
reactive dyes, 98-105 3,6-Endomethylene-l,2,3,6-tetrahy-
triazine dyes, 97 drophthalic acid, pK, 405
Endonuclease, affinity chromatography, 25
E Endopeptidase, affinity chromatography, 25
fl-Endorphin, reversed-phase chromatogra-
EDPS, s e e N,N'-Bis(3-sulfopropyl)- phy, 208
ethylenediamine Enolase
Elastase affinity partitioning, 361
affinity chromatography, 25 gel staining, 432
gel staining, 429 Enterokinase, s e e Enteropeptidase
Elastin, 25 Enteropeptidase, affinity chromatography,
Electrophoresis, 235-301, s e e a l s o specific 25
type Enterotoxin, C. p e r f r i n g e n s , iso-
buffers, 240, 242,243 tachophoresis, 294
elution, 20 Enzyme, s e e a l s o specific substance
nondenaturing activity, in gels, 417,418
gel composition, 240, 241 localization, 416-439
gel preparation, 240, 241 gel elution, recovery, 419
in sodium dodecyl sulfate, s e e Sodium staining, s e e Staining, enzyme
dodecyl sulfate-gel electrophoresis Epoxy-silica
Electrophoretic titration, retention map- coupling, HPLC, 216
ping, 228-231 preparation, 213,215
Ellman reagent, 17 EPPS, s e e N-2-Hydroxyethyipiperazine-
Elongation factor Ts, affinity chromatogra- N'-3-propanesulfonic acid
phy, 50 Erythropoietin, isotachophoresis, 294
Elongation factor Tu, affinity chromatogra- Esterase, gel staining, 428, 439
phy, 50 Estradiol, affinity chromatography, 50
Elution, 179-181, 185 Estrogen receptor, affinity chromatogra-
affinity chromatography, 19, 20 phy, 47
conditions, 19 Ethanol, protein crystallization, 374
nonspecific, 19 Ethanolamine buffer, pK, 406
specific, 19 Ether deoxylysolecithin, properties, 321
buffers, 202-204 Ethyl Cellosolve, membrane protein solu-
by electrophoresis, 20 bilization, 335
 512 SUBJECT INDEX

l-Ethyl-2-[3-(l-ethylnaphtho[ 1,2-d]- Formate buffer, pK, 405

thiazolin-2-ylidene)-2-methylpro penyl]- Formate dehydrogenase
naphthol[1,2-d]thiazolium bromide, affinity chromatography, 25
453,454 isotachophoresis, 294
Eugenol, 417 Formic acid, membrane protein solubiliza-
Exonuclease, affinity chromatography, 25 tion, 334
Formiminotransferase, affinity chromatog-
raphy, 25
F
Formycin B, 31
Factor V, affinity chromatography, 47 Formylmethionine aminopeptidase, affinity
Factor VIII, affinity chromatography, 47 chromatography, 25
Factor IX, affinity chromatography, 47 13-D-Fructofuranosidase, gel staining, 429
Fatty acid synthetase, affinity chromatog- Fructose-l,6-bisphosphatase, gel staining,
raphy, 25 428
Fc receptor, affinity chromatography, 50 Fructosyltransferase, gel staining, 425
Ferredoxin-NADP+ reductase, affinity L-Fucose dehydrogenase, affinity chroma-
chromatography, 25 tography, 25
Ferredoxin-nitrate reductase, affinity a-L-Fucosidase, affinity chromatography, 25
chromatography, 25 Fumarase, s e e Fumarate hydratase
Ferritin, reversed-phase chromatography, Fumarate hydratase
205,210 affinity chromatography, 25
Ferrochelatase, affinity chromatography, gel staining, 432
25
a -Fetoprotein
affinity chromatography, 47 G
purification, 116, 117
Fibrinogen, isotachophoresis, 294
Fibronectin /3-Galactofuranosidase, affinity chromatog-
affinity chromatography, 51 raphy, 25
PEG precipitation, 355 Galactosaminidase, affinity chromatogra-
Filtration, s e e specific type phy, 25
Flavokinase, s e e Riboflavin kinase D-Galactose kinase, affinity chromatogra-
Fluorescein isothiocyanate-antibody, gel phy, 26
staining, 450 a-Galactosidase, affinity chromatography,
Fluorography, 460-465 25
agarose gel, 462 13-Galactosidase, affinity chromatography,
/]-emitters, 461-463 25, 26
film exposure, conditions, 463, 464 /3-Galactosidase A2, affinity chromatogra-
flashing film, 464 phy, 25
gel preparation, 465 13-Galactosidase A3, affinity chromatogra-
polyacrylamide gels, 461,462 phy, 25
solid porous supports, 463 Galactosylhydroxylysylglucose transferase,
Fluorometer, flow, in size exclusion affinity chromatography, 26
HPLC, 163 Galactosyltransferase
FMN oxiaoreductase, affinity chromatog- affinity chromatograhy, 26
raphy, 25 gel staining, 425
Forskolin, 22 Gel, s e e a l s o specific type
Formaldehyde dehydrogenase, affinity composition, 240, 241
chromatography, 25 denaturing, 238
Formamide, membrane protein solubiliza- composition, 241
tion, 335 discontinuous, 255
 SUBJECT INDEX 5 13

preparation, 241-243 Glucocorticoid, affinity chromatography, 50

renaturation, 417 Glucocorticoid receptor, affinity chroma-
detergent, s e e Detergent gel; specific tography, 47
type Glucokinase, affinity chromatography, 26
enzyme activity, 417,418 Glucose, protein crystallization, 374
enzyme localization, 416-439 Glucoseamine-silica, 219
enzyme staining, 416-439 Glucose isomerase, affinity chromatogra-
granulated, 263-265 phy, 26
nondenaturing, 239-241 Glucose-6-phosphatase, affinity chromatog-
means of separation, 239 raphy, 26
reducing agents, 239 Glucose-l-phosphate adenylyltransferase,
rehydratable, 265, 272, 273 affinity chromatography, 22
stacking, function, 237 Glucose-6-phosphate dehydrogenase,
Gelatin, PEG precipitation, 355 affinity chromatography, 26
Gel electrophoresis Glucosephosphate isomerase, gel staining,
aggregates of proteins, 239 435
analytical, 237-255 Glucose- 1-phosphate uridylyltransferase,
apparatus, 246-249 gel staining, 427, 437
buffers, 240, 242, 243 a-Glucosidase
detergent gels, 253-255 autochromic method, 439
effect of temperature, 240 gel staining, 428, 439
gradient gels, 252, 253 sandwich-type incubation, 439
hexadecylpyridinium chloride, 239 a-D-Glucosidase, affinity chromatography,
molecular weight determination, 250- 26
252 fl-Glucosidase, gel staining, 429, 439
procedure, 249 Glucosyltransferase
protein stains, 249, 250 affinity chromatography, 26
sample loading, 248 gel staining, 426
sample preparation, 248, 249 fl-Glucuronidase, affinity chromatography,
SDS, 238 47
SDS-urea, 238 Glutamate decarboxylase, affinity chroma-
separating gel, 246-248 tography, 26
tetradecyltrimethylammonium bromide, Glutamate dehydrogenase
238 affinity chromatography, 26
two-dimensional, 239 affinity precipitation, 365, 367, 368
use of mercaptans, 240 gel staining, 422,436
Gel filtration, 154, s e e a l s o High-perfor- PEG precipitation, 354
mance liquid chromatography, size Glutamate synthase, affinity chromatogra-
exclusion phy, 26
separation of vesicles and micelles, Glutamic acid, protein crystallization, 374
326, 327 Glutaminase, gel staining, 429
Gel staining, 419, 420, s e e a l s o specific ~/-Glutamylcysteine synthase, affinity
enzyme chromatography, 26
copolymerization of substrate, 436, 437 7-Glutamyl hydrolase, affinity chromatog-
limitations, 420, 421 raphy, 26
postincubation capture, 436, 437 7-Glutamyltransferase, gel staining, 425
specific enzymes, 421-439 Glutaraldehyde, coupling, 15
Glass, preparation of tresylated support, Glutathione reductase, affinity chromatog-
58, 59 raphy, 26
a-1,4-Glucan branching enzyme, gel stain- Glutathione transferase, affinity chroma-
ing, 425 tography, 26
514 SUBJECT INDEX

Glutathione S-transferase GM-ganglioside-fl-galactosidase, affinity

gel staining, 425 chromatography, 26
isotachophoresis, 299 Gonadotropin, human chorionic, reversed-
Glyceraldehyde-3-phosphate dehydro- phase chromatography, 208
genase Gramicidin-S synthetase, affinity chroma-
affinity chromatography, 26 tography, 26
gel staining, 422 Growth factor, epidermal, affinity chroma-
hydrophobic chromatography, 71 tography, 50
purification, 71 Growth hormone
Glyceroaldehyde phosphate dehydro- affinity chromatography, 50
genase, affinity partitioning, 361 isotachophoresis, 294
Glycerol-3-phosphate dehydrogenase reversed-phase chromatography, 204,
affinity chromatography, 26 208
gel staining, 422,434, 436 GTP cyclohydrolase, affinity chromatog-
a-Glycerophosphate dehydrogenase, s e e raphy, 27
Glycerol-3-phosphate dehydrogenase Guanidine hydrochloride, 331
Glycerylpropyl-silica, preparation, 215 Guanine aminohydrolase, s e e Guanine
Glycinamide buffer, pK, 406 deaminase
Glycine, affinity chromatography, 50 Guanine deaminase, affinity chromatog-
Glycine buffer, pK, 406 raphy, 27
Glycineamide ribonucleotide transform- N2-Guanine RNA-methyltransferase,
ylase, affinity chromatography, 26 affinity chromatography, 27
Glycogen-agarose, 70 Guanosine, affinity chromatography, 51
Glycogen phosphorylase, affinity chroma- Guanylate cyclase, affinity chromatog-
tography, 26 raphy, 27
Glycogen phosphorylase b Guanyloribonuclease, s e e Ribonuclease Ti
adsorption-elution, 83 Guanylyltransferase, affinity chromatog-
elution, effect of pH, 88, 89 raphy, 27
hydrophobic chromatography, 71
purification, 71, 84
Glycogen synthase, hydrophobic chroma- H
tography, 85, 86
Glycolytic enzyme, dye-ligand chromatog- HCG-receptor complex, affinity chroma-
raphy, 97 tography, 47
Glycophorin, 329 Hemoglobin
Glycoprotein displacement chromatography, 133
antifreeze, 25 isotachophoresis, 294, 296
gel stains, 447-451 Hemopexin, affinity chromatography, 51
labeled-lectin stain, 449, 450 Hepatitis A virus, affinity chromatography,
membrane, affinity chromatography, 47 47
periodic acid-Schiff stain, 448, 449 HEPES, s e e N-2-Hydroxyethylpiperazine-
thymol-sulfuric acid stain, 447, 448 N'-2-ethanesulfonic acid
Glycoprotein/3-o-galactosyltransferase, HEPPS, s e e N-2-Hydroxyethylpiperazine-
affinity chromatography, 33 N'-3-propanesulfonic acid
Glycosidase Hexadecylpyridinium chloride gel, 239, 253
affinity chromatography, 26 Hexafluoroacetone, membrane protein
gel staining, 428 solubilization, 335
Glycylglycine buffer, pK, 406 Hexokinase
Glyoxalase affinity chromatography, 27
affinity chromatography, 26 coupling to tresyl-agarose, 61
gel staining, 433 dye-ligand chromatography, 110, 111
 SUBJECT INDEX 515

Hexosaminidase, affinity chromatography, detectors, 136, 137

27 gradient elution, 142
Hexosaminidase A, affinity chromatog- guard cartridges, disposable, 140, 141
raphy, 27 ion-exchange, 170-189
Hexosaminidase B, affinity chromatog- band broadening, 176
raphy, 27 characterization, 175
Hexosaminidase P, affinity chromatog- columns, 133-154
raphy, 27 capacity, 178, 179
Hibernation trigger, isotachophoresis, 294 characteristics, 176, 177
High-performance liquid chromatography, cleaning, 188, 189
133-233 comparisons, 176, 177
affinity matrix, 216, 217 preparative use, 179
chromatofocusing, s e e Chromato- selection, 173-177
focusing size, 178, 179
columns, 133-154, 161, 162 switching, 189
anion exchange, 170 wash solvents, 153
backflushing, 151 effect of anions, 186
bed irregularities, 151,152 of cations, 187
bore, 135 of charge, 171
cation exchange, 170 of urea, 188
cleaning, 188, 189 ehition, 179-181, 185
contamination, 153 guard columns, 181
effect of pH, 138 hydrophobicity, 173
failure, 137, 138 hydrophobic proteins, 187, 188
fittings, 134 isoelectric point, 173-175
frit cleaning, 150, 152 mechanism, 171,172
guard, 139-141 mobile phase selection, 180
microparticulate, 140 optimization, 183-185
high efficiency, 135 process, 170-173
installation, 134, 135 retention, 170-173
malfunctioning, 133 maps, 173, 174
correction, 144-149 sample preparation, 189
operation, 136-143 selectivity, 185-187
limits, 142 silica supports, 176, 177
packing ion-pair agents, 143
characteristics, 134 ligand coupling, 216, 217
peUicular, 140 mobile-phase additive, 138, 143
particulate matter, 138 problems, 143-150
pump, 134 pumps, 161
regeneration, 153 flow rate, 161
repair, 150-154 resolution, 139, 184, 185
of voids, 151, 152 decreasing, 146
sample load, 139 equation, 170, 184
saturator, 141, 142 reversed phase, s e e Reversed-phase
solvent, viscous, 153 chromatography
storage, 143 sample preparation, 136-139
switching, 189 sample volume, 139
testing, 135, 136 size exclusion, 154-169
troubleshooting, 143-150 applications, 169
washing, 152-154 calculations, 1 5 5 - 1 6 0
wash solvents, 153 columns, 133-154
516 SUBJECT INDEX

calibration, 155, 168 Histone, reversed-phase chromatography,

curve, 155, 156, 166, 167 209
comparison, 159, 160 Histoplasmin, isotachophoresis, 294
corrosion, 167 H-2K k antigen, affinity chromatography,
life, 167 47
maintenance, 163-167 HL-A antigen, affinity chromatography,
overload, 160 337
selection, 163-167 HLA-B, affinity chromatography, 47
wash solvents, 153 Horseradish peroxidase, gel staining,
desalting process, 160 451
elution, 157-159 Hyaluronidase, affinity chromatography,
equipment, 160-163 27
errors, 157 Hydrazide
guard column, 167 characterization, 16, 17
matrix, 162 quantitation, 16, 17
molecular weight standards, 166 Hydrazidoagarose, 14-18
Mr range, 156, 157 aldehyde-containing, 15, 16
operating conditions, 168, 169 amino-containing, 16
packing, 162, 164, 165 analysis, 16
pore size, 156 carboxyl-containing, 16
postcolumn reactions, 163 characterization, 16
protein standards, 155, 156 sulfhydryl-containing, 16
purification, 169 Hydrindantin, 9
resolution, 157-159, 169 Hydrolase
sample volume, 160 dye-ligand chromatography, 97
theoretical plates, 159 gel staining, 424, 428-430
theory, 155-160 3-Hydroxyacyl-CoA dehydrogenase, affin-
solvent impurities, 137, 138 ity chromatography, 27
troubleshooting N-2-Hydroxyethylpiperazine-N'-2-ethane-
abnormal operating pressure, 144 sulfonic acid, pK, 405
baseline noise, 149 N-2-Hydroxyethylpiperazine-N'-3-pro-
double peak, 148 panesulfonic acid, pK, 405
drifting baseline, 148, 149 to-Hydroxy-fatty-acid NADP-oxidoreduc-
fronting peaks, 148 tase, affinity chromatography, 27
ghost peaks, 143, 148 3-Hydroxybutyrate dehydrogenase, dye-
negative peaks, 148 ligand chromatography, 111, 112
peak tailing, 147, 149 Hydroxylase, dye-ligand chromatography,
retention time 97
decreasing, 144, 145 3-Hydroxy-3-methylglutaryl-CoA reduc-
increasing, 145, 146 tase, affinity chromatography, 27
split peak, 148-150 15-Hydroxyprostaglandin dehydrogenase,
truncated peaks, 148 affinity chromatography, 27
use of solvents, 137, 138 Hydroxysteroid dehydrogenase, gel stain-
Histaminase, s e e Amine oxidase ing, 422
L-Histidine aminotransferase, affinity 3a-Hydroxysteroid dehydrogenase, affinity
chromatography, 27 chromatography, 27
Histidinol dehydrogenase, affinity chroma- 20c~-Hydroxysteroid dehydrogenase, affin-
tography, 27 ity chromatography, 27
Histocompatibility antigen, isotachophore- 3fl-Hydroxysteroid dehydrogenase, affinity
sis, 294 chromatography, 27
 SUBJECT INDEX 517

20/3-Hydroxysteroid dehydrogenase, affin- Intrinsic factor, affinity chromatograhy, 50

ity chromatography, 27 Ion-exchange chromatography
3fl-Hydroxysteroid oxidase, affinity chro- electrophoretic titration curves, 225,
matography, 27 226
N-Hydroxysuccinimidochloroformate, 12 gel isoelectric focusing, 225, 226
Hypoxanthine-guanine phosphoribo- gradient elution, 142
syltransferase optimal pH, 223-233
affinity chromatography, 28 retention mapping, 227-229
gel staining, 425 Ion-exchange matrix, macroporous, 223,
225
Ion-exchange process, 170-173, s e e a l s o
I
High-performance liquid chromatogra-
Ia antigen, affinity chromatography, 337 phy, ion-exchange
L-Iditol dehydrogenase, affinity chromatog- charge distribution, 171
raphy, 32 Ionophoresis, s e e Isotachophoresis
Imidazole buffer, pK, 405 Isocitrate dehydrogenase
Imidocarbonate affinity chromatography, 28
analysis, 8-11 gel staining, 422
determination, for resins, 9, 10 Isocitrate lyase, gel staining, 432
formation, 5 Isoelectric focusing, 256-275
Immunoaffinity chromatography, 21, 47, affinity, 278
48, 54 ampholytes, 260, 261
Immunoglobulin, isotachophoresis, 294 apparatus, 258, 259
Immunoglobulin E, affinity chromatogra- buffers, 261,268
phy, 50 cascade, 268
Immunoglobulin E receptor commercial flatbeds, 259
reconstitution, 343,344 detection of proteins, 269-271
solubilization, 311,312, 341 electrode strips, 273
Immunogiobulin G, purification, 112 elution, 274, 275
Immunosorbent separation, 381-387 gels, 257, 258
Indolyl-3-alkane a-hydroxylase, affinity matrix, 262-265
chromatography, 28 preparation, 272, 273
Initiation factor elF-3, affinity chromatog- thickness, 258
raphy, 51 high-resolution, 256-275
Inosine 5'-monophosphate dehydrogenase, load capacity, 265, 266
dye-ligand chromatography, 109 pH gradient, 225, 259-262, 275
Inosinic acid dehydrogenase, affinity formation, 260
chromatography, 28 immobilized, 261,262, 269
Insulin power supplies, 258
affinity chromatography, 50 preparative, 256-275
crystallization, 377 printing, 274
reversed-phase chromatography, 205 recovery techniques, 271,272
Insulin receptor resolution factors, 266-269
affinity chromatography, 47 retention mapping, 229
reconstitution, 347 sample application, 273
Interferon separation distance, 268
affinity chromatography, 47 Isomerase, gel staining, 434, 435
dye-ligand chromatography, 97 Isopropanol, protein crystallization, 374
reversed-phase chromatography, 208, a-Isopropylmalate isomerase, affinity
209 chromatography, 28
518 SUBJECT INDEX

Isotachophoresis coupling, 18
acrylamide gel, 285-289 purification, 112
apparatus, 285-287 Legumin, affinity chromatography, 47
applications, 299-301 Leupeptin, 331
cooling, 297,298 Levatix dye, 99
electrolytes, 287, 288, 291,299, 300 LiChrosorb Diol column, 164
gel column, 285-289 LiChrospher column, 165
gel preparation, 287, 288 Ligand
photopolymerization, 287 coupling, 17, 18
preparative, 281-301 to agarose, 60
principles, 282-287 amino-containing, 18
procedures, 287-293 assay, 18
properties, 282, 283 HPLC, 216, 217
separation, 289 density
protein, 294, 295 charge density, determination, 77,
slab gel, 289-293 78
apparatus, 290 NMR, 76, 77
electrolytes, 291 quantitation, 75
printing, 292, 293 Lipase
procedures, 290-293 affinity chromatography, 28
spacer principle, 284, 285 gel staining, 428, 434, 435
zone-sharpening effect, 285 postincubation capture, 438,439
Isourea, bond instability. 11 sandwich-type indicator gel, 438, 439
simultaneous capture, 438,439
K Lipoamide dehydrogenase, s e e Dihydroli-
poamide reductase
Kallikrein, affinity chromatography, 28, Lipoprotein
47 high-density
2-Keto-3-deoxy-L-fuconate NAD oxidore- affinity chromatography, 51
ductase, affinity chromatography, 28 isotachophoresis, 295
Kininogen, isotachophoresis, 295 low-density, affinity chromatography,
Kohlrausch regulating function, 237 51
Kynurenine 3-monooxygenase, affinity Lipoprotein lipase, affinity chromatogra-
chromatography, 28 phy, 28
/3-Lipotropin, reversed-phase chromatogra-
L phy, 208
Lipoxygenase
a-Lactalbumin, retention mapping, 228 affinity chromatography, 28
Lactate dehydrogenase gel staining, 423
affinity chromatography, 28, 67 Lithium 3,5-diiodosalicylate, 331
affinity precipitation, 365-367 Lowry method, modifications for small
gel staining, 422 concentrations, 415, 416
Lactoglobulin, retention map, 174 Lubrol PX
/3-Lactoglobulin, separation, displacement properties, 309
chromatography, 124-127 solubilization, 307, 312
/3-Lactoglobulin A, ionic strength effects, structure, 308
354 Lucfferase, affinity chromatography, 28
Lactoyl-glutathione lyase, s e e Glyoxalase Lyase, gel staining, 432-434
Laemmli method, 455 Lymphocyte, affinity chromatography, 47
Lectin Lysine tRNA-synthetase, affinity chroma-
affinity chromatography, 54, 55 tography, 28
 SUBJECT INDEX 519

L-Lysine 6-aminotransferase, affinity HPLC, 338, 339

chromatography, 28 hydrophobic chromatography, 338,
Lysozyme 339
affinity chromatography, 28 integral, 329-339
retention map, 174 intrinsic, 305-318
Lysyl oxidase, affinity chromatography, 28 ion-exchange chromatography, 336
isotachophoresis, 294
phase separation, 335, 336
M
protein-detergent ratio, 313-315
Maclurin, reversed-phase chromatography, reconstitution, 340-347
210 reversed-phase chromatography, 201
aL-Macroglobulin, isotachophoresis, 294 solubility, 305
cx2-Macroglobulin, affinity chromatogra- criteria, 305, 306
phy, 47 solubilization, 305-318, 332-335
Malate buffer, pK, 405 approach, 305
Malate dehydrogenase buffers, 310
affinity chromatography, 28 comparisons, 317
dye-ligand chromatography, 111, 112 conditions, 310
gel staining, 422 strategy, 311,312
Malate synthase, gel staining, 432 stabilization, 312, 313
Malate thiokinase, affinity chromatogra- techniques, 303-347
phy, 28 Meromyosin, affinity chromatography, 51
Malic enzyme, affinity chromatography, 28 MES, s e e 2-(N-Morpholino)ethanesulfonic
Malonyl-CoA decarboxylase, affinity acid
chromatography, 28 Messenger ribonucleoprotein, affinity
Maltodextrin phosphorylase chromatography, 51
affinity chromatography, 28 Metal chelation, trivalent, protein staining,
purification, 84 454, 455
Mannosephosphate isomerase, gel staining, Metal ion
435 buffers, 409, 410
a-D-Mannosidase, affinity chromatogra- chelating agents, 409, 410
phy, 29 Methanol, protein crystallization, 374
Membrane Methionyl-tRNA synthetase, affinity chro-
isolation, 330-332 matography, 29
purification, 330-332 3-Methyladenine DNA-glycosylase, affinity
chaotropic ions, 331 chromatography, 29
chelating agents, 331 Methyl CeUosolve, membrane protein
solubilization, 331-335 solubilization, 335
Membrane filtration, microporous, water Methylmalonyl-CoA carboxyltransferase,
purification, 401 affinity chromatography, 33
Membrane protein Methylmalonyl-CoA mutase, affinity chro-
affinity chromatography, 337, 338 matography, 29
amphiphilic nature, 329 2-Methyl-2,4-pentanediol, protein crystalli-
characterization, 315-317 zation, 374
covalent chromatography, 338 N-Methylpyrrolidone, membrane protein
crystallization, 377, 378 solubilization, 335
detergents, s e e Detergent NS-Methyltetrahydrofolate-homocysteine
extraction with organic solvents, 333- methyltransferase, affinity chromatog-
335 raphy, 29
fractionation, 335-339 Methyltransferase mRNA, affinity chroma-
gel filtration, 336 tography, 29
520 SUBJECT INDEX

4-Methylumbelliferyl a-D-glucopyranoside, Neurophysin

gel staining, 439 affinity chromatography, 51
Micelle, s e e a l s o Detergent reversed-phase chromatography, 205,
detergent-protein, 323 208
/32-Microglobulin, reversed-phase chroma- Neurotoxin III, reversed-phase chromatog-
tography, 205, 208 raphy, 210
Microorganism, in water, destruction by Neutrophil migration inhibition factor,
UV, 399 affinity chromatography, 47
Migration inhibitory factor Neville method, 455
affinity chromatography, 47, 51 Nitrite reductase, affinity chromatography,
isotachophoresis, 294 29
Monoamine oxidase, s e e Amine oxidase Nitro blue tetrazolium, 434
Monobeads, 176, 177 Nitrocellulose, protein blotting, 458
Monochlorotriazinyl dye, 99, 104 Nitrogen, determination, for resins, 9
Monophenol monooxygenase, s e e Tyro- p-Nitrophenylchloroformate, 12, 14
sinase p-Nitrophenyl cyanate, molecular struc-
MOPS, s e e 3-(N-Morpholino)propanesul- ture, 5
fonic acid Nuclear magnetic resonance, buffers, 414
MOPSO, s e e 3-(N-Morpholino)-2-hydroxy- Nuclease, micrococcal, affinity chromatog-
propanesulfonic acid raphy, 29
Morpholine buffer, pK, 406 Nuclease T, affinity chromatography, 29
2-(N-Morpholino)ethanesulfonic acid, pK, Nucleic acid, coupling, 15
405 Nucleotidase, affinity chromatography, 29
3-(N-Morpholino)-2-hydroxypropanesul- 5'-Nucleotidase, solubilization, 307
fonie acid, pK, 405 Nucleotide, coupling, 18
3-(N-Morpholino)propanesulfonic acid, Nucleotide phosphodiesterase, affinity
pK, 405 chromatography, 29
mRNA cap, affinity chromatography, 51 Nucleotide phosphotransferase, affinity
Myoglobin, affinity chromatography, 47 chromatography, 29
Myosin, affinity chromatography, 47, 51 Nucleotide pyrophosphatase, affinity
Myosin kinase, affinity chromatography, chromatography, 29
29
O
N
Octylagarose, HPLC, 339
NAD kinase, affinity chromatography, 29 Octyl fl-glucopyranoside, crystallization,
NADPH-adrenodoxin reductase, affinity 377
chromatography, 29 Octyl glucoside
NADPH-cytochrome c reductase, affinity properties, 309, 321
chromatography, 29 solubifization, 307, 341
NADPH-ferredoxin reductase, affinity structure, 308
chromatography, 29 O'Farrell method, 455
NADPH-flavin oxidoreduetase, affinity Opiate, affinity chromatography, 50
chromatography, 29 Organic solvent, reversed-phase chroma-
NAD-protein ADP-ribosyltransferase, tography, 200-202
affinity chromatography, 29 Ornithine carbamoyltransferase
NAD(P) + transhydrogenase, affinity chro- affinity chromatography, 30
matography, 33 gel staining, 425
Neuraminidase, affinity chromatography, Oruithine decarboxylase, affinity chroma-
29, 32 tography, 30
 SUBJECT INDEX 521

Ornithine transcarbamylase, s e e Ornithine Phosphodiesterase

carbamoyltransferase affinity chromatography, 30
Orotate phosphoribosyltransferase, affinity dye-ligand chromatography, 97
chromatography, 30 gel staining, 428
Orotidine-5 '-phosphate decarboxylase, Phosphoenolpyruvate carboxykinase
affinity chromatography, 29 affinity chromatography, 30
Osmosis, reverse, water purification, 398, gel staining, 432
399 Phosphofructokinase
Ovalbumin affinity chromatography, 30
heterogeneity, 130 affinity partitioning, 361
retention map, 174 3-Phosphoglycerate kinase, affinity parti-
separation, displacement chromatogra- tioning, 361
phy, 124, 125, 130 Phosphoglycerate phosphomutase, gel
Oxidoreductase, gel staining, 421-424 staining, 435
6-Phosphogluconate dehydrogenase, affin-
P ity chromatography, 30
Phosphoglucose isomerase, affinity chro-
Papain, reversed-phase chromatography, matography, 30
210 Phosphoglycerate kinase, affinity chroma-
Paper, s e e specific type tography, 30
Parvalbumin, reversed-phase chromatogra- Phosphokinase, dye-ligand chromatogra-
phy, 209 phy, 97
Penicillin, affinity chromatography, 51 Phospholipase, affinity chromatography,
Pentachlorophenylchloroformate, 12 30
n-Pentanol, membrane protein solubiliza- Phosphohpase A2
tion, 334 affinity chromatography, 30
Pepsin, affinity chromatography, 30 gel staining, 428
Pepstatin, 22, 23 Phospholipase C
Peroxidase affinity chromatography, 30
affinity chromatography, 30 gel staining, 428
gel staining, 422 Phospholipid vesicle
Pharmacia Mono P column, 182 formation, 327
Pharrnacia Mono Q column, 177 gel filtration, 326, 327
Pharmacia Mono S column, 177 Phosphoprotein stain, 451-455
Pharmacia Polyanion Si column, 177 phosphate liberation, 451-453
Phenol, membrane protein solubilization, Stains-All, 453,454
334 trivaient metal chelation, 454, 455
Phenylalanine ammonia-lyase, affinity Phosphoprotein phosphatase, affinity
chromatography, 30 chromatography, 30, 31
Phenylalanine hydroxylase, affinity chro- Phosphoribosyltransferase, dye-ligand
matography, 30 chromatography, 97
Phenylaianine-tRNA synthetase, affinity Phosphotyrosine protein, affinity chroma-
chromatography, 30 tography, 47
Phenylmethylsulfonyl fluoride, 331 Phosphorylase a, affinity chromatography,
Phosphatase, gel staining, 430, 439 30
Phosphate buffer, pK, 405,406 Phosphorylase b
Phosphatidylcholine, affinity chromatogra- affinity chromatography, 30
phy, 51 reversed-phase chromatography, 209
Phosphatidylglycerophosphate synthetase, Phosphorylase kinase, affinity chromatog-
affinity chromatography, 30 raphy, 30
522 SUBJECT INDEX

Photopolymerization, isotachophoresis, stabilization, 351,352

287 Polyethylene glycol-triazine dye, 357
Phthalocyanine dye, 99 Polyethylenimine, chromatofocusing, 182,
Phytochrome, dye-ligand chromatography, 183
97 Polymer, removal, 363
PIPBS, s e e 1,4-Bis(4-sulfobutyl)piperazine Polynucleotide phosphorylase, s e e Polyri-
Piperazine-N, N'-bis(2-ethanesulfonic bonucleotide nucleotidyltransferase
acid), pK, 405 Polypyridine
Piperazine-N, N'-bis(2-hydroxypropanesul- activation, 11
fonic acid), pK, 405 with sulfonyl chloride, 11
Piperidine buffer, pK, 406 Polyribonucleotide nucleotidyltransferase,
PIPES, s e e Piperazine-N, N'-bis(2-ethane- affinity chromatography, 31
sulfonic acid) Polynucleotide 5'-triphosphatase, affinity
PIPPS, s e e 1,4-Bis(3-suifopropyl) pipera- chromatography, 31
zinc POPSO, s e e Piperazine-N,N'-bis(2-
Plasminogen, dye-ligand chromatography, hydroxypropanesulfonic acid)
97 Porphobilinogen synthase, affinity chroma-
Plasminogen activator, gel staining, 429 tography, 23
Poliovirus, affinity chromatography, 47 Post-proline cleaving enzyme, affinity
Polyacrylamide gel chromatography, 31
affinity electrophoresis, 276, 277 Precipitation
electrophoresis, 237-255 analytical, curve, 353, 354
fluorography, 461,462 fractional, protein, 351
isoelectric focusing, 262, 263 mechanism, 352
isotachophoresis, 285-289 protein, with polyethylene glycol, 351-
solid supports, 463 356
Polyacrylhydrazidoagarose, 14-18 Preelectrophoresis, 240
aldehyde-containing, 15, 16 Procion Blue H-B, structure, 100
amino-containing, 16 Procion Blue MX-4GD, affinity chromatog-
carboxyl-containing, 16 raphy, 112
derivatization, 15, 16 Procion Blue MX-R
preparation, 15 affinity chromatography, 219
sulfhydryl-containing, 16 structure, 100
Poly(adenosine diphosphate ribose) poly- Procion Brown MX-5BR, affinity chroma-
merase, affinity chromatography, 31 tography, 97, 219
Poly(A) polymerase, affinity chromatogra- Procion dye, 98-101, s e e a l s o specific dye
phy, 30 Procion Green H-4G, affinity chromatogra-
Polybuffer, chromatofocusing, 183 phy, 97, 110, 111
Polyethylene glycol Procion Green MX-5Br, affinity chroma-
affinity partitioning, 356-364 tography, 219
ionic strength effects, 354 Procion Red H-3B
precipitation affinity chromatography, 111, 112
curve, 353, 354 structure, 100
mechanism, 352 Procion Red H-8BN, affinity chromatogra-
preparation of tresylated support, 59 phy, 111,219
protein crystallization, 374 Procion Red H-E3B, affinity chromatogra-
protein-ligand interactions, 354, 355 phy, 97
protein-protein interactions, 354, 355 Procion Red H-4G, affinity chromatogra-
removal, 355, 356 phy, 97
role of buffer, 353 Procion Rubine MX-B, structure, 100
size availability, 352, 353 Procion Scarlet MX-G
 SUBJECT INDEX 523

affinity chromatography, 97 seeding, 380

structure, 100 solvent conditions, 373, 374
Procion Yellow H-A strategy, 380, 381
affinity chromatography, 111, 219 use of precipitants, 372
structure, 100 vapor diffusion, 375, 376
Procion Yellow MX-8G, affinity chroma- x-ray requirements, 370
tography, 97, 109 denaturation
Progesterone, affinity chromatography, 50 effect of detergents, 340-342
Proin, crystallization, 377 of lipids, 340-342
Prolactin displacement chromatography, 113-
affinity chromatography, 50 133
isotachophoresis, 294 heteroassociation, 355
Prolyl hydroxylase, affinity chromatogra- hydrophobic, 69-96, 319, s e e a l s o
phy, 31 Chromatography, hydrophobic
1,3-Propanediol, protein crystallization, HPLC, 187, 188
374 membrane, s e e Membrane protein
n-propanol, protein crystallization, 374 modifying reagents, 331
Properdin, affinity chromatography, 47 precipitation, polyethylene glycol, 351-
Propionyl-CoA carboxylase, affinity chro- 356
matography, 31 renaturation, 417
Prostaglandin cyclooxygenase, affinity retention map, 174
chromatography, 31 solubilization, 331-335
Protease stain, 439-441,447-455, s e e a l s o Stain-
affinity chromatography, 31 ing, protein
neutral, affinity chromatography, 29 gels, 447-451
postincubation capture, 438, 439 glycoproteins, 447-451
sandwich-type indicator gel, 438,439 silver, 441-447
simultaneous capture, 438, 439 colored images, 446
Protein destaining, 444, 445
assay, for dilute solutions, 415,416 image intensification, 444, 445
blot, s e e Western blot negative image, 442, 443
coupling, 18 positive image, 443,444
to agarose, 60 recycling, 444
crystallization, 370-381 sensitivity, 445
concentration method, 376, 377 standards, Mr, 250, 251
dialysis, 372, 375 Protein A
effect of concentration, 371 affinity chromatography, 66
of detergents, 377, 378 HPLC, 219
of solvents, 371 Proteinase, gel staining, 429
of temperature, 371,377 Protein kinase
factors affecting, 370-372 affinity chromatography, 31
hanging drop method, 376, 378 gel staining, 426
membrane proteins, 377, 378 Protein-ligand interaction, 354, 355
methodology, 372-378 Protein phosphatase, affinity chromatogra-
microdialysis, 379 phy, 31
optimization, 371 Protein-protein interaction, 354, 355
precipitating agents, 372-374 Psoriasis, protein, displacement chroma-
precipitating solvents, 374 tography, 129
precipitation conditions, 370 Pteroyl-ot-oligoglutamylendopeptidase,
sandwich box technique, 376, 378, affinity chromatography, 31
379 Pullulanase, affinity chromatography, 31
524 SUBJECT INDEX

Purine nucleoside phosphorylase, affinity wash solvents, 153

chromatography, 31 conditions, 199-211
Pyridine, membrane protein solubilization, counterions, 204
334 denaturation, 192
Pyridine buffer, pK, 405 desalting, 203, 204
Pyridine dinucleotide transhydrogenase, displacement model, 195
affinity chromatography, 31 effect of buffer, 202-204
Pyridine nucleotide-dependent oxidoreduc- of ionic strength, 202-204
tase, dye-ligand chromatography, 97 of pH, 204, 205
Pyridoxal kinase, affinity chromatography, eluants, 208-210
31 elution, 197-199
Pyridoxamine-5-phosphate oxidase, affinity gradient, 199
chromatography, 31 isocratic, 198
Pyrophosphatase organic solvents, 200-202
gel staining, 430 high-capacity
isotachophoresis, 294 flow rate, 207, 211
Pyrophosphate buffer, pK, 405 hydrophobicity, 207
Pyruvate carboxylase, gel staining, 435 ligands, 206
Pyruvate decarboxylase matrix, 206, 207
affinity chromatography, 31 pore size, 207
gel staining, 432 temperature, 207, 211
Pyruvate dehydrogenase, affinity chroma- ion pairing, 193,202-204
tography, 31 modulation, 191
Pyruvate kinase, gel staining, 426 membrane proteins, 201
Pyruvate-UDP-N-acetylglucosaminetrans- mobile phases, 208-210
ferase, affinity chromatography, 32 multisite interactions, 192, 193
protein, 190, 195
R elution, 204
purification, 208-212
Receptor, s e e a l s o specific type retention, 196
affinity chromatography, 50, 51 separation, 192-194
binding proteins, 50, 51, 55 retention
reconstitution, 342-347 behavior, 194-197
solubilization, 343 mechanism, 191, 192
Relaxin, reversed-phase chromatography, size-exclusion phenomenon, 193
205, 208 strategy, 193, 194
Remazol dye, 99 Rhodopsin, solubilization, 341
Renaturation, s e e Protein renaturation Riboflavin, affinity chromatography, 51
Renin, affinity chromatography, 32 Riboflavin kinase, affinity chromatography,
Resin, s e e specific type 25
Restriction endonuclease Ribonuclease
affinity chromatography, 32 affinity chromatography, 32
dye-ligand chromatography, 97 gel staining, 428
Retention map, 173, 174 retention map, 174
Retinal, affinity chromatography, 51 reversed-phase chromatography, 210
Reversed-phase chromatography, 190- Ribonuclease F~, affinity chromatography,
212 32
albumin, 195 Ribonuclease F2, affinity chromatography,
columns, 133-154, 208-211 32
efficiency, 207, 211 Ribonuclease H, affinity chromatography,
material, 205-207 32, 47
 SUBJECT INDEX 525

Ribonuclease T1, affinity chromatography, Shodex OH column, 165

27 Sialidase, s e e Neuraminidase
Ribonuclease "1"2,affinity chromatography, Sialyltransferase, affinity chromatography,
32 32
Ribonuclease U2, affinity chromatography, Silica
32 ligand substitution, 214, 215
Ribonucleic acid, coupling, 18 support, 176
Ribonucleotide reductase, affinity chroma- Silver stain, protein, 441-447, s e e a l s o
tography, 32 Protein stain, silver
Ribosomal protein, affinity chromatogra- Sm antigen, affinity chromatography,
phy, 51 48
R i c i n u s lectin, gel staining, 450 Sodium cholate
RNA ligase, affinity chromatography, 32 properties, 309, 321
RNA nuclease, dye-ligand chromatogra- structure, 308
phy, 97 Sodium deoxycholate, properties, 321
RNA polymerase Sodium dodecyl sulfate
affinity chromatography, 32 gel, 436, 455
basic protein, affinity chromatography, Laemmli formulation, 246
32 molecular weight determinations,
DNA-dependent, affinity chromatogra- 250-252
phy, 24, 32 Neville formulation, 247, 248
dye-ligand chromatography, 97 preparation, 245-248
gel staining, 426 protein standards, 250, 251
RNA polymerase III, affinity chromatogra- renaturation, 437
phy, 32 use of 2-propanol, 436, 437
RNP antigen, affinity chromatography, 48 properties, 321
removal, from gels, 418
S solubilization, 332
Sodium dodecyl sulfate-gel electrophore-
Salicylate hydroxylase, s e e Salicylate 1- sis, 238
monooxygenase gel preparation, 245-248
Salicylate 1-monooxygenase, affinity chro- gradients, Mr determination, 252, 253
matography, 32 protein interaction, 241,242
Secretin, reversed-phase chromatography, Solubilization, protein, with detergents,
209 305-318
Seeding, protein crystallization, 380 Somatostatin, affinity chromatography,
Separon HEMA column, 165 48
Sephacryl S-400, removal, 323 Sorbitol dehydrogenase, s e e L-Iditol dehy-
Serine acetyltransferase, affinity chroma- drogenase
tography, 47 Soybean trypsin inhibitor
Serine dehydratase, gel staining, 433 affinity chromatography, 65
Serotonin, affinity chromatography, 51 HPLC, 219
Serum albumin, dye-ligand chromatogra- coupling to tosyl-agarose, 61
phy, 97 to tresyl-agarose, 60, 61
Serum lipoprotein, dye-ligand chromatog- retention map, 174
raphy, 97 Spacing protein, carboxymethyldextran as
Serum protein spacers, 120
displacement chromatography, 124, 125 Spermine synthase, affinity chromatogra-
isotachophoresis, 294, 297 phy, 32
Sex steroid, affinity chromatography, 51 Sphingomyelinase, affinity chromatogra-
Shiga toxin, affinity chromatography, 48 phy, 32
526 SUBJECT INDEX

Staining T
enzyme
approaches, 419 TAPS, s e e 3-{[Tris(hydroxymethyl)
autochromic methods, 420 methyl]amino}propanesulfonic acid
copolymerization of substrate, 420 Taurine, pK, 406
effect of ampholytes, 418 Taurocholate, solubilization, 332
of SDS, 418 T cell growth factor, reversed-phase chro-
postincubation coupling, 420 matography, 210
precautions, 417,418 T4 DNA ligase, coupling to tresyl-agarose,
retention of activity, 417, 418 64
sandwich-type incubation, 420 TEA, s e e Triethanolamine
simultaneous capture, 419 TES, s e e 2-[Tris(hydroxymethyl)methyl-
gel, s e e Gel staining amino]ethanesulfonic acid
protein, 227, 249, 250, 292, 293, 419 Tetrabase, 417
Western blots, 458-460 Tetradecyltrimethylammonium bromide
Stains-All, s e e 1-Ethyl-2-[3-(l-ethylnaph- gel, 238, 254
tho[ 1,2-d]thiazolin-2-ylidene)-2- Tetrahydrofolate dehydrogenase, affinity
methylpropenyl] naphthol[ 1,2-d]thia- chromatography, 24
zolium bromide Tetrazolium salt, 434
Stokes radius, 320 Tetrodotoxin, affinity chromatography, 48
Subtilisin, gel staining, 429 Thermolysin, gel staining, 429
Succinate buffer, pK, 405 Thioester, coupling, 15
Succinate-semialdehyde dehydrogenase, Thioredoxin deaminase, affinity chroma-
affinity chromatography, 32 tography, 32
Succinate thiokinase, s e e Succinyl-CoA Threonine deaminase, s e e Threonine
synthetase dehydratase
Succinyl-CoA acetoacetyl-CoA trans- Threonine dehydratase, affinity chromatog-
ferase, affinity chromatography, raphy, 33
32 Thrombin, affinity chromatography, 33
Succinyl-CoA synthetase, affinity chroma- Thymidine kinase, affinity chromatogra-
tography, 32 phy, 33
Sulfobetaine, solubilization, 332, 337 Thymidylate synthase, affinity chromatog-
Sulfohydrolase, dye-ligand chromatogra- raphy, 33
phy, 97 Thymine-methacrylate, 219
Sulfonyl chloride Thymosin, reversed-phase chromatogra-
affinity chromatography, 56-69 phy, 209
coupling to agarose, 56--69 Thyroglobulin, reversed-phase chromatog-
immobilization, 56-69 raphy, 205,208
preparation of tosylated supports, 59 Thyroid peroxidase, affinity chromatogra-
of tresylated polyethylene glycol, phy, 33
59 Thyroid stimulating hormone, affinity
of tresylated supports, 58, 59 chromatography, 48, 50
Support, s e e a l s o specific type Thyrotropin
preparation of tosylated, 59 affinity chromatography, 48
of tresylated, 58, 59 reversed-phase chromatography, 208
Sweat protein, isotachophoresis, 298 p-Toluenesulfonyl chloride, 56, 57
Synchrom AX column, 182 Tosyl-agarose, coupling, 61, 62
Synchropak AX, 188 Tosyl chloride, s e e p-Toluenesulfonyl
Synchropak column, 165, 177 chloride
Synthetase, dye-ligand chromatography, N-Tosyl-L-phenylalanylchloromethyl
97 ketone, 331
 SUBJECT INDEX 527

Toyo Soda column, 177 Trinitrobenzenesulfonic acid, quantitation

Transamidase, gel staining, 425 of hydrazides, 15-17
Transcarboxylase, s e e Methylmalonyl-CoA Triosephosphate isomerase, gel staining,
carboxyltransferase 435
Transcobalamin II, affinity chromatogra- Tripeptide aminopeptidase, gel staining,
phy, 50 429
Transcriptase, reverse, affinity chromatog- Tris, s e e Tris(hydroxymethyl)-
raphy, 32 aminomethane
Transferase, gel staining, 424-427 Tris(hydroxymethyl)aminomethane, pK,
Transferrin 405
affinity chromatography, 50 2-[Tris(hydroxymethyl)methylamino]-
dye-ligand chromatography, 97 ethanesulfonic acid, pK, 405
isotachophoresis, 294, 296 3-{[Tris(hydroxymethyl) methyl]amino}pro-
retention mapping, 228 panesulfonic acid, pK, 406
Transferrin receptor, affinity chromatogra- N-[Tris(hydroxymethyl)methyl]glycine,
phy, 337 pK, 405
Transh~/drogenase, s e e NAD(P) + transhy- Triton N-101, solubilization, 312
drogenase Triton X-100
Transplantation antigen, affinity chroma- phase separation, 336
tography, 48 properties, 309, 321
Trehalose, gel staining, 429 removal, 319, 323-325
Tresyl-agarose solubilization, 307, 312
coupling, 60, 61 structure, 308
preparation, 62-65 Triton X-114, phase separation, 336
yield, 62-65 tRNA nucleotidyltransferase, affinity
Tresyl chloride, s e e 2,2,2-Trifluorethane- chromatography, 32
sulfonyl chloride Troponin C, 24
Tresyl-silica, preparation, 215,216 Trypsin
Triacylglycerol lipase, affinity chromatog- coupling to tresyl-agarose, 61
raphy, 28, 33 gel staining, 429
Triazine dye reversed-phase chromatography, 209
affinity chromatography, 97 Trypsin inhibitor, s e e a l s o specific type
elution, 108 affinity chromatography, 48
properties, 102, 103 retention mapping, 228
Triazine dye-polyethylene glycol, 357 reversed-phase chromatography, 205
Trichlorophenylchloroformate, 12 Tryptophanase, affinity chromatography,
Tricine, s e e N-[Tris(hydroxymethyl)- 33
methyl]glycine Tryptophan 5-monooxygenase, affinity
Triethanolamine, pK, 405 chromatography, 33
Triethylamine, for cyano transfer, 6 Tryptophan synthase, affinity chromatogra-
Trifluoroacetic acid, HPLC, 188 phy, 33
2,2,2-Trifluoroethanesulfonyl chloride, 56, TSK-gel PW column, 164
57 TSK-gel SW column, 164
activation, 62 Tubulin, affinity chromatography, 51
affinity chromatography, HPLC, 213, Tumor-specific surface antigen, isota-
216, 217 chophoresis, 294
coupling of proteins, 63 Tween 80, properties, 321
Triglyceride lipase, s e e Triacylglycerol Tyrosinase
lipase affinity chromatography, 33
Triiodothyronine, affinity chromatography, gel staining, 423
50 reversed-phase chromatography, 210
528 SUBJECT INDEX

Tyrosine aminotransferase, affinity chro- W

matography, 33
Tyrosine phenol-lyase, affinity chromatog- Water
raphy, 33 impurities, 391,392
metals, 396
U size, 395
for laboratory usage, 391-403
UDP-N-acetylenolpyruvylglucosamine microorganisms, 392
reductase, affinity chromatography, ultraviolet irradiation, 399
33 purification, 391-403
UDPgalactose 4-epimerase, affinity chro- activated carbon, 394, 395
matography, 33 apparatus, 401-403
UDPgalactose glycoprotein galacto- comparison, 400
syltransferase, s e e Glycoprotein/3-0- depth filtration, 393,394
galactosyltransferase distillation, 398
UDPglucose dehydrogenase, affinity chro- ion-exchange, 395-398
matography, 33 microporous membrane filtration,
UDPglucuronyltransferase, affinity chro- 401
matography, 33 process considerations, 401
UDPGpyrophosphorylase, simultaneous reverse osmosis, 398, 399
capture technique, 437 standards, 403
Ultrafiltration, water purification, 395 ultrafiltration, 395
Ultraviolet irradiation, water purification, ultraviolet irradiation, 399
399 treatment systems, 392-401
UmbeUiferyl derivative, 417 Western blot, 455-460
Urate oxidase, affinity chromatography, 33
Urea, 331
Urease, gel staining, 429, 436 Y
Uricase, s e e Urate oxidase
Urokinase Yeast hexokinase, dye-ligand chromatogra-
affinity chromatography, 33, 48 phy, 111
gel staining, 429
monoclonal antibody, 383 Z

V Z-DNA, affinity chromatography, 51

Zeppezauer procedure, 372, 375, 379
Valyl-tRNA synthetase, affinity chroma- Zwittergent, 3-14
tography, 33 properties, 309
Vapor diffusion, 375, 376 solubilization, 307
Vimentin, affinity chromatography, 51 structure, 308