APPLIED BIOCATALYSIS

APPLIED BIOCATALYSIS
Second Edition
Edited by
Adrie J.J.Straathof
Kluyver Laboratory for Biotechnology
Delft University of Technology
The Netherlands
and
Patrick Adlercreutz
Department of Biotechnology
Lund University
Sweden

harwood academic publishers

Australia • Canada • France • Germany • India • Japan • Luxembourg • Malaysia The Netherlands •
Russia • Singapore • Switzerland
 Copyright © 1994, 2000 OPA (Overseas Publishers Association) N.V. Published by
license under the Harwood Academic Publishers imprint, part of The Gordon and
Breach Publishing Group.

All rights reserved.

First Edition published 1994

This edition published in the Taylor & Francis e-Library, 2005.

“To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of
thousands of eBooks please go to www.eBookstore.tandf.co.uk.”

No part of this book may be reproduced or utilized in any form or by any means,
electronic or mechanical, including photocopying and recording, or by any
information storage or retrieval system, without permission in writing from the
publisher. Printed in Singapore.

Amsteldijk 166
1st Floor
1079 LH Amsterdam
The Netherlands

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library.

ISBN 0-203-30563-9 Master e-book ISBN

ISBN 0-203-34353-0 (Adobe eReader Format)

ISBN: 90-5823-024-4 (Print Edition)
 CONTENTS

PREFACE TO THE SECOND EDITION vi

PREFACE TO THE FIRST EDITION vii
CONTRIBUTORS ix

1. INTRODUCTION
1
Klaus B.Buchholz and Poul B.Poulsen
2. REACTIONS CATALYSED BY ENZYMES 18
Thorleif Anthonsen
3. ENZYMES AS PROCESSING AIDS AND FINAL PRODUCTS 62
Johannes Tramper and Poul B.Poulsen

4. CASE STUDIES IN THE APPLICATION OF BIOCATALYSTS FOR THE 103

PRODUCTION OF (BIO) CHEMICALS
Peter S.J.Cheetham
5. HOW TO GET THE BIOCATALYST 173
Marcel G.Wubbolts , Christopher Bucke and Stanislaw Bielecki
6. IMMOBILIZED BIOCATALYSTS 244
Sven Pedersen and Morten Würtz Christensen
7. PROTEIN ENGINEERING: DESIGN AND ENGINEERING ON THE 263
NANO SCALE
Steffen B.Petersen
8. BIOCATALYST PERFORMANCE 311
Antonio Balksteros and Laszlo Boross
9. BIOCATALYSTS IN NON-CONVENTIONAL MEDIA 339
Patrick Adlercreutz

10. PROCESS CONCEPTS FOR BIOCONVERSIONS 365

Adrie J.J.Straathof and Volker Kasche
11. BIOREACTOR DESIGN 392
Joaquim M.S.Cabral and Johannes Tramper
12. PATENT ASPECTS OF BIOCATALYSIS
441
Peter S.J.Cheetham and Philip J.D.Thomas
13. SOME COMMERCIAL AND FINANCIAL ASPECTS OF
BIOCATALYSIS RESEARCH AND DEVELOPMENT PROJECTS 464
Peter S.J.Cheetham

INDEX 506
 PREFACE TO THE SECOND EDITION

After the publication of the first edition of Applied Biocatalysis some five years ago, this
field has rapidly been developing. This is evident from the number and types of new
applications, but also from the state of the art for some of the important techniques, such
as protein engineering and the use of non-conventional media, for example.
Considering the continuing number of advanced courses in applied biocatalysis that are
being given by the Working Party, there was a need to revise the book. Thus numerous
items in this book have been corrected or updated. Moreover, several previous chapters
have been merged or split up. Additional improvements in the new edition are a two-level
Table of Contents, and a two-level numbering of the sections, figures, tables and
equations. All of this should facilitate the use of the book as a textbook by students in
courses such as the Advanced Course on Applied Biocatalysis.
We want to thank the previous editors for stimulating us to take up this job, and also all
the present and previous Working Party members for their contribution.
A.J.J.Straathof and P.Adlercreutz, Editors of the 2nd edition, 1999
 PREFACE TO THE FIRST EDITION

Applied biocatalysis is older than written history. Ancient records, picturing the
manufacture of foods and beverages, testify to the involvement of amylases and proteases
from microbial, plant or animal origin, without the knowledge of those using them. These
ancient applications can therefore be best described as an art and not as a technology or a
scientifically defined method.
The development of a scientifically based ‘applied biocatalysis’ began at the end of the
nineteenth century with the introduction of a standardised enzyme preparation, rennet, for
cheese making. Progress was slow, however, over the intervening years. The appearance
of detergent proteases and the introduction of genetic and protein engineering about
twenty years ago marked the beginning of the rapid increase in the discovery and
application of the scientific fundamentals of applied biocatalysis. Developments in this
young science are still rapid whilst training and education in the area are in their infancy.
The need for education and training was recognised by the Working Party on Applied
Biocatalysis of the European Federation of Biotechnology. One of the aims of the
Working Party has been to identify bottlenecks for the developments of the field and to
facilitate faster progress. To date this has been achieved through the organisation of
special, multi-disciplinary symposia. In 1988 the Working Party recognised that more
general progress in education and training was hampered by the lack of a good textbook
in this field and launched a project, based on the combined expertise of the members of
the Working Party from industry and academia, to remedy this in 1990.
To test the proposed format as an education tool an advanced course in applied
biocatalysis was organised in Murcia, Spain in November 1991 and given by members of
the Working Party. The lecture notes from this course were used as a first draft of the
book and subsequently revised, taking into consideration the many valuable comments
from course participants. The textbook, although authored by relatively few Working
Party members, has been refined by the concerted action of the entire group, through long
discussions in many meetings. Without these invaluable contributions the book would
have been quite different.
The book aims to provide teachers and students in biotechnology with a text for an
advanced course in applied biocatalysis whilst reviewing relevant basic principles. It will
therefore be useful to scientists and engineers entering the field and as a source of new
ideas and data for the specialist. The book itself will be used as the basis for a second
course in applied biocatalysis in Poland in 1994.
J.M.S.Cabrai, D.Best, L.Boross and J.Tramper, Editors of the 1st edition, 1994
 THE WORKING PARTY ON APPLIED
BIOCATALYSIS OF THE EUROPEAN
FEDERATION OF BIOTECHNOLOGY
(EFB)

A number of Working Parties of the EFB were set up to cover various scientific and
engineering areas of biotechnology. Among those Working Parties is the Working Party
on Applied Biocatalysis which has forty delegates from twenty-one countries. The terms
of reference of this Working Party are:

• To increase the understanding of biocatalysts and in particular their commercial and

other applications;
• To take initiatives in areas of growing scientific and industrial interest and importance
in the field of applied biocatalysis;
• To foster relationships between interested scientists and engineers in different European
countries by arranging meetings on topics relating to applied bio-catalysis;
• In particular the Working Party seeks to identify key topics which may be rate limiting
the development of European scientific and technological capabilities in applied
biocatalysis and to take appropriate steps to stimulate these areas, and/or to make
appropriate recommendations to the EFB;
• In addition the Working Party will increasingly be used as an expert source of technical
information and opinions by European Commission groups, for instance to propose
prospective research topics for EC research programmes.

The past and current activities of the Working Party on Applied Biocatalysis involve the
preparation of reports, the organisation of scientific meetings on topics of key scientific
interest and courses in the area of applied biocatalysis. The journal Biocatalysis and
Biotransformation is also published by Harwood Academic Publishers in association
with the EFB-Working Party on Applied Biocatalysis.
See also: http://www.dechema.de/englisch/europa/biotec/pages/biotec2a.htm
 CONTRIBUTORS

Patrick Adlercreutz
Department of Biotechnology
Center for Chemistry and Chemical Engineering
Lund University
P.O. Box 124
S-221 00 Lund
Sweden

Thorleif Anthonsen
Department of Chemistry
Norwegian University of Science and Technology
N-7491 Trondheim
Norway

Antonio Ballesteros
Departamento de Biocatalisis
CSIC Instituto de Catalisis
University of Autonoma
28049 Madrid
Spain

Stanislaw Bielecki
Institute of Technical Biochemistry
Lodz Technical University
Stefanowskiego str. 4/10
PL-90–924 Lodz
Poland

Laszlo Boross
Department of Chemistry and Biochemistry
University of Horticulture and Food Industry
Villány út 29–31
H-1114 Budapest
Hungary
Klaus Buchholz
Department of Carbohydrate Technology
Langer Kamp 5
D-38106 Braunschweig
Germany

Christopher Bucke
School of Biological and Health Sciences
University of Westminster
115 New Cavendish Street
London WIM 8JS
United Kingdom

Joaquim M.S.Cabral
Laboratório de Engenharia Bioquímica
Centro de Engenharia Biológica e Química
Instituto Superior Técnico
1000 Lisboa
Portugal

Peter S.J.Cheetham
Zylepsis Ltd.
6 Highpoint
Henwood Business Estate
Ashford, Kent
United Kingdom

Morten Würtz Christensen

Novo Nordisk A/S
Novo Allé, 8PS
DK-2880 Bagsvaerd
Denmark

Volker Kasche
Technische Universitāt Hamburg-Harburg
Denickestr. 15
D-21071 Hamburg
Germany

Sven Pedersen
Novo Nordisk A/S
Novo Allé, 8PS
DK-2880 Bagsvaerd
Denmark

Steffen B.Petersen
Biostructure and Protein Engineering Group
Aalborg University
Sohngaardsholmsvej 57
DK-9000 Aalborg
Denmark

Poul B.Poulsen
Novo Nordisk A/S
Novo Allé, 8PI, 03
DK-2880 Bagsvaerd
Denmark

Adrie J-J.Straathof
Kluyver Laboratory for Biotechnology
Delft University of Technology
Julianalaan 67
2628 BC Delft
The Netherlands

Philip J.D.Thomas
Eric Potter Clarkson
Park View House
58 The Ropewalk
Nottingham, United Kingdom

Johannes Tramper
Food and Bioprocess Engineering Group
Wageningen Agricultural University
P.O. Box 8129
6700 EV Wageningen
The Netherlands

Marcel G.Wubbolts
DSM Research
P.O. Box 18
6160 MD Geleen
The Netherlands
 1.
INTRODUCTION
KLAUS B.BUCHHOLZ1 and POUL B.POULSEN2
1 Department of Carbohydrate Technology, Langer Kamp 5, D-38106
Braunschweig, Germany. Tel: +49–531–380090; Fax: +49–531–3800988; E-
mail: [email protected]
2 Novo Nordisk A/S, DK-2880 Bagsvaerd, Denmark. Tel: +45–4442–3417; Fax:
+45–4498–0610; E-mail: [email protected]

ABSTRACT
In this book, applied biocatalysis is defined as the application of a biocatalyst to
achieve a desired conversion under controlled conditions in a bioreactor. This
chapter focuses on the history of this field.
The most ancient applications may best be described as an art and not as a
technology. The scientific developments in biocatalysis go back to the
beginning of the 19th century. They were very slow due to a lack of adequate
theories and methods. Only late in the 19th century did there occur a
breakthrough and an increase of research activities. A continuous technical
development can be traced to the beginning of the 20th century. Again, the
industrial development of enzymes was very slow initially. Only with the
appearance of the detergent proteases in the 1960’s, did the use of enzymes
increase. Subsequently, the first immobilized enzyme products were scaled up
to industrial application. Today, the number of successful applications of
biocatalysis is rapidly increasing, as will be shown in the subsequent chapters.

1.1 WHAT IS APPLIED BIOCATALYSIS?

Applied biocatalysis can be defined as the application of a biocatalyst to achieve a

desired conversion under controlled conditions in a bioreactor. A biocatalyst can either be
an enzyme, an enzyme complex, a cell organelle or a whole cell. The latter can be viable
growing or non-growing or non-viable. Furthermore, a biocatalyst can be free or
immobilized and this has far-reaching consequences, not only with respect to substrate
supply and mass transfer in general, but sometimes also, in case of viable cells, with
 Applied biocatalysis 2
respect the physiology. The source of biocatalysts can be of microbial, plant or animal
origin and examples of all three are found in this book.
In this first chapter the history of applied biocatalysis is described. At the end of the
chapter, an outline of the contents of the main body of the book will be given.

1.2 OVERVIEW OF HISTORY OF APPLIED BIOCATALYSIS

Applied biocatalysis has its roots in ancient China and Japan in the manufacture of food
and alcoholic drinks. Without knowing, man utilized microbial amylases and proteases,
in particular for the production of soy-derived foods. In Europe too, applied biocatalysis
has a long history. Cheese making has always involved the use of enzymes. As far back
as about 400 BC, Homer’s Iliad mentions the use of kid stomach for making cheese. It
was discovered that milk, which was stored in a bag made of a stomach of a recently
slaughtered calf, lamb or kid was converted into a semi-solid substance. Upon pressing of
this substance a drier material was obtained (namely cheese) which

(i) showed preferred properties compared to milk,

(ii) could easily be transported and
(iii) gained a flavor after some time.

These ancient applications may best be described as an art and not as a technology, a
scientifically based technique.
The scientific developments in biocatalysis go back to the beginning of the 19th
century. The first enzymatic action, notably starch hydrolysis by Diastase, was
acknowledged as a catalytic effect by Berzelius in 1835 and utilized industrially
following the findings and developments of Payen and Persoz on Diastase and its
application in the brewing industries about 10 years later. Continuous scientific
investigations, with emphasis on alcoholic fermentation, can be traced to the 1850’s with
investigations of Berthelot and Béchamp, the school of Pasteur and the controversy with
the chemical school and Liebig’s influence.
The first company based upon applied biocatalysis also dates back to the 19th century.
In 1874 Christian Hansen started a company in Copenhagen, Denmark. His company—
named Christian Hansen’s Laboratory to this day—was the first in the industrial market
with a standardized enzyme preparation, rennet, for cheese making. Rennet, a mixture of
chymosin (also called rennin) and pepsin, was and still is obtained by salt extraction of
the fourth stomach of suckling calves.
After this early period the research in the subsequent decades brought about significant
progress with the important findings of E.Fischer in 1894 on specificity and E.Buchner in
1897 on the pure chemical nature of alcohol formation, the key steps by Sumner and
Northrup crystallizing enzymes in 1926 and 1930, respectively, which represent the
turning points in identifying the nature of the catalytic agent.
A continuous technical development can be traced to the beginning of the 20th century,
 Introduction 3
where Röhm and the foundation of the Röhm and Haas company in 1907 may be seen as
a major technical and economic breakthrough.
Much material on the scientific and technological development has been summarized
in articles by Sumner and Myrbäck (1950), Sumner and Somers (1953), Ullmann (1914),
Tauber (1949), Neidleman (1991) and in the book of Roberts et al. (1995).

1.3 THE EARLY PERIOD UP TO 1880

1.3.1 Findings and Empirical Results

Stahl in his 1697 book Zymotechnika Fundamentalis explored the nature of fermentation
as an important industrial process, where zymotechnica should be the scientific basis
(Bud, 1992). The liquefaction of meat by gastric juice was noted by

Table 1.1 Ferments (enzyme activities) known till 1880.

Enzvme Ferment Catalyzed reaction Reference

source
Protease gastric juice meat liquefaction Spallanzani, 1783*
Cyanogen plant roots substrate: guaiacum Planche, 1810*
Glutin comp. wheat starch hydrolysis Kirchhoff, 1814*
Emulsin bitter almonds amygdalin hydrolysis Robiquet and Boutron, 1830*;
Chalard, Liebig and Wöhler,
1837*
Diastatic activity ptyalin starch hydrolysis Leuchs, 1831*
Amylase malt starch hydrolysis Payen and Persoz (1833)
Sinigrinase Faure, 1835*
Pepsin protein hydrolysis Schwann, 1836*
Trypsin protein hydrolysis Corvisart, 1856*
Saccharase yeast sucrose hydrolysis Berthelot, 1860*
Pectase plants pectin hydrolysis Payen (1874)
Pancreas lipolytic activity Bernard, 1856**
Pancreas extract fat and starch Dobell (1869)
hydrolysis
 Applied biocatalysis 4

Pancreatic pancreas fat hydrolysis Frankland (1885)

ferment
* in: Sumner and Somers (1953)
** in: Tauber (1949)

Spallanzani in 1783 (Sumner and Somers, 1953), the enzymatic hydrolysis of tannin was
described by Scheele in 1786 (Tauber, 1949), in 1814 Kirchhoff observed that a glutinous
component of wheat is capable of converting starch to sugar and dextrin, and Vogel
showed in 1817 that an infusion of oats would produce a fermentable sugar from milk
(Roberts et al., 1995).
Diastase was found to hydrolyse starch to dextrin and sugar (Payen and Persoz, 1833).
The isolation of diastase is described, following Payen and Persoz, by Knapp (1847): it is
precipitated from a malt extract and can be further purified by repeated dissolving in
water and precipitating by addition of alcohol. Diastase was the dominating object of
research throughout the century with about 10 to 20% of the publications dealing with it
during decades, due to its economic importance (see below).
“Pectase” was found in plants, both in soluble and non soluble form. It was able to
break down “pectose” into pectinic acid, and was attributed some similarity to diastase.
Neither materials could be crystallized (Payen, 1874). Claude Bernard was the first to
show lipolytic activity in pancreas in 1856 (Tauber, 1949), and Dobell (1869) found that
an extract from pancreas hydrolyzed both fat and starch; he gave a procedure for
extraction and stabilization and named this preparation “pancréatine”.
Further enzymatic conversions (ferment actions) observed in that early period,
summarized in Frankland’s list of soluble ferments (1885) and by Sumner and Somers
(1953) are summarized in Table 1.1.
Alcoholic fermentation was a dominating topic of the time in the field called
physiological chemistry which attracted a good deal of interest and activity of leading
scientists. Trying to identify the active principle, Berthelot showed, according to his own
interpretation, that a peculiar substance, formed by yeast, can transform sugar into
alcohol. This substance was nitrogen containing, could be precipitated by alcohol, and
was comparable to diastase. He stated, that this substance was different from yeast and
that there was no production of yeast cells (Anonymous, 1862; Berthelot, 1864). Thus he
claimed to have observed the transformation of sugar into alcohol without the
participation of any activity of living organisms. He was convinced that the formation of
alcohol from sugar was a genuinely chemical process. However, his experiments were
not performed under sterile conditions. Obviously unknown, in part anaerobic bacteria (in
tests where oxygen was excluded), must have produced small amounts of alcohol and
carbon dioxide, as well as hydrogen, acetic and butyric acid (Berthelot, 1857).
Pasteur presented a series of experimental results, showing unequivocally that
alcoholic fermentation proceeded only in the presence of living organisms. And he
showed, that in all cases where fermentation could be observed under exactly controlled
 Introduction 5
conditions, it was microorganisms (of different kinds, also present in the air), which must
be present in order to initiate fermentation. So he presented strong evidence against the
assumption of a “generatio spontanea” postulated by Gay-Lussac (Anonymous, 1862).
The overall research activity on soluble ferments, that is on enzymatic reactions, is
scarce in this period nevertheless seen as important (Berthelot, 1857, 1864). Thus in the
German “Journal für Praktische Chemie”, in the period from 1850 to 1860 no paper dealt
with soluble ferments (enzymatic activities) and 8 papers were published on fermentation
(Gährung, with the meaning of microbial activity); in the Bulletin de la Société Chimique
de Paris, one of the most important of the time for fermentation research, there was
published about one article per year in the 1860’s dealing with soluble ferments, which
signifies enzyme activity, and 3 to 4 dealing with fermentation (Table 1.2). It was only in
the 1880’s that research and publication activities rose significantly.

1.3.2 Technology
Work on technological issues was more important and even dominating over that on
basic aspects from the beginning of the 19th century: brewing (Roberts et al., 1995),
wine, bread and acetic acid production. These topics made up important parts of the
chemical technology of the time (Poppe, 1842; Knapp, 1847; Wagner, 1857; Payen,
1874). Knapp considered scientific and practical, industrial interest as equally important.
Diastase application was a major issue from the 1840’s and onwards. The treatment of
starch by acids or diastase yielded a gum syrup. Notably French products seemed to have
been produced enzymatically, as was obvious by their smell of malt (Knapp, 1847).
Sugar containing dextrin was produced, following a process of Payen and Heuzé by
treating 100 parts of starch with 5 to 8 parts of malt in water at 60 to 70°C (Wagner,
1857). This product was used mainly in France in bakeries, for the production of beer and
wines from fruits. The process is described in more detail, including applications, by
Payen (1874), also including a short calculation showing that the application of malt is
more economic than that of sulfuric acid. However, in a book on chemical technology by
Ost (1900) the process is not mentioned anymore.
Lab products were used to produce cheese (Knapp, 1847). Berzelius is cited with
details stating that 1 part of lab ferment preparation (essentially proteases) coagulates
1800 parts of milk, and that only 0.06 parts of the lab ferment is lost.

1.3.3 Theoretical Approaches

Some vital factor, “le principe vital”, was considered an important principle in the
chemical processes associated with the synthesis of materials isolated from living matter:
“All simple bodies in nature are subject to the action of two powers, of which one, that of
attraction, tends to unite the molecules of bodies one with another, while the other,
produced by caloric, forces them apart…A certain number of these simple bodies in
nature are subject to a third force, to that caused by the vital factor, which changes,
 Applied biocatalysis 6
modifies and surpasses the two others, and whose limits are not yet understood” (Beral,
1815, cited in Roberts et al., 1995).
In one of the books on technology in the period preceding Berzelius and Payen, Poppe
(1842) discussed mysterious ideas concerning fermentation: “Fermentation is seen as a—
at a time and under circumstances spontaneous—occurring mighty movement in a liquid
of different compounds…, which is due to the fact that several compounds act in
harmony with each other, others in opposition to each other, so that the first attract, the
latter reject each other”. “The sugar and a…material or a gum extract material act by
antagonist forces on each other so heavily, that they decompose and thus cause the
formation of alcohol (Weingeist)…” (Poppe, 1842).
Gay-Lussac had postulated a “generatio spontanea”, the hypothesis that a continuing—
rather mysterious—chain should be the cause of spontaneous generation (of organisms)
(summary by Anonymous, 1862). Pasteur clearly showed that this assumption was
without any empirical basis, and that fermentation was always due to inoculation by
organisms. Pasteur did change the disciplinary context of fermentation away from
chemistry redefining the subject to a new autonomous discipline, microbiology (Bud,
1992, 1993).
Berzelius, in 1839, had interpreted fermentation as caused by a catalytic force. He
postulated that a body by its mere presence could, by affinity to the fermentable
substance, cause its decomposition to the products.
About a decade later the book by Knapp (1847) develops a rather distinct picture of the
action of ferments (mainly referring to diastase and the work by Payen and Persoz)
speaking of “mighty chemical forces” and the ability, that one part (of the ferment) can
transform 2000 parts of starch into sugar (obviously a catalytic potential), and that “even
if one does not know anything of the chemical composition”. He speaks of a
“hypothetical body”, a “symbol”, and the most strange energy, by which the
transformation (of starch) occurs, and assumes, that the nitrogen containing parts (the
“Kleber”) of the seed are being transformed to a ferment, and that diastase is not a certain
substance, but rather a state or form (“Zustand”). Nevertheless, the process for the
isolation of diastase by extraction from malt and precipitation by alcohol is described, so
Knapp, based on Payen, approaches the phenomenon from a chemical and technological
point of view rather than from a mysterious one.
Again a decade later the distinction of organized and unorganized ferments was
developed. Wagner (1857) describes two types of ferments which can cause
fermentation; one is an organized (obviously living) body, like yeast, the other a protein
like a body which is in the state of decomposition. That segmentation is expressed more
precisely by Payen (1874). Fermentation is seen as a contact (catalytic) process of a
degradation (“Spaltungs-”) or addition process (with water). It can be performed by two
substances or bodies:

1. A nitrogen containing organic (non organized) substance, such as protein material

undergoing degradation.
2. An organized body, a lower class plant or an “infusorium”, such as with alcoholic
fermentation.
 Introduction 7
Probably the type of effect is the same insofar as the ferment of the second class produces
a body of the first class, possibly a big number of singular ferments (Payen, 1874).
In 1878 Kūhne named the latter class of substances: ENZYMES.
The scarce research activity (see above) and the slow progress in knowledge of soluble
ferments (enzymes) might be attributed to the lack of a leading paradigm. The debate on
mysterious processes, such as the “generatio spontanea” was not settled until Pasteur’s
investigations. Even then, soluble ferments had no chemical identity, the subject of
enzymatic catalysis remained obscure and bound to living processes.
The problems in understanding fermentation, and notably the economically most
important alcoholic fermentation, can be traced from a summary on the work of Berthelot
and Pasteur (Anonymus, 1862).

1.4 THE PERIOD FROM 1890 TO 1940

1.4.1 Growing Interest

A pronounced increase in papers on soluble ferments (Table 1.2) indicated the growing
interest, and several key findings initiated the enzymology of the 20th century.
From about 1894 onwards Emil Fischer investigated in a series of experiments the
action of different enzymes using several glycosides and oligosaccharides; the results
revealed specificity as one of the key characteristics of enzymes. In 1894 he compared
Invertin and Emulsin. He extracted Invertin from yeast, a usual procedure, and showed
that it hydrolyzed the -, but not the -methyl-D-glucoside. In contrast, Emulsin, a
commercial preparation from Merck, hydrolyzed the -, but not the -methyl-D-
glucoside. Furthermore, the L-sugar-derivative, methyl-L-glucoside was not a substrate.
Amongst a series of tests with different saccharides, he observed that invertin hydrolyzed
sucrose and maltose, but not lactose. An extract from a “lactose-yeast”, however, was
able to hydrolyse lactose, but not maltose. These observations are the most essential from
a range of which Fischer derived his famous theory on specificity (see below) (Fischer,
1909).

Table 1.2 Numbers of articles on soluble ferments in the mid-18th century.

Period Journal Number of articles on Number of articles on

soluble ferments fermentation**
1850 to J.Prakt. 0 1 per year
1861 Chem.
 Applied biocatalysis 8

1863 to Bulletin 1 per year (8 total) 3 to 4 per year

1871
Number of articles on diastase*
1885 Chem. Z. 14 12
Blatt
1886 Chem. Z. 9 5
Blatt
1887 Chem. Z. 18 11
Blatt
1889 Chem. Z. 21 2
Blatt
1890 Chem. Z. 29 1
Blatt
1891 Chem. Z. 20 4
Blatt
1895 Chem. Z. 35 18
Blatt
1900 Chem. Z. 38
Blatt
1905 Chem. Z. 73 12
Blatt
1910 Chem. Z. 94 24
Blatt
J. Prakt. Chem: Journal für Praktische Chemie
Bulletin: Bulletin de la Société Chimique de Paris
The journals include articles or reports on papers from others (notably French or German)
Chem. Z. Blatt: Chemisches Zentral-Blatt (abstract journal summarizing international
publications)
*Including similar activities, amylase, amyloglucosidase, not precisely differentiated
** Fermentation with the meaning of microbial activity

A few years after Fischer’s investigations Eduard Buchner published a series of papers
(1897, 1898) which signalized a breakthrough in fermentation and enzymology. The
experiments began in 1893 (Buchner, 1898b). In his first paper on alcoholic fermentation
without yeast cells (1897) he stated, in a remarkably short and precise manner, that “a
separation of the (alcoholic) fermentation from the living yeast cells was not successful
up to now”; in the following he described a process, which solves this task. He gives the
experimental details for the preparation of a cell free press juice from yeast cells, with
disruption, filtration under high pressure and further filtration. He then described the
 Introduction 9
formation of carbon dioxide from carbohydrates, sucrose, glucose, fructose and maltose;
no fermentation was observed with lactose and mannitol. No microscopic organisms were
present. Chloroform, an antiseptic inhibiting microbial growth, did not inhibit the
“fermentation” (which meant the enzymatic reaction). At elevated temperature protein
was precipitated, the activity reduced and finally destroyed. From these results, Buchner
derived essential new insights into the nature of alcoholic fermentation (see following
section).
In subsequent papers he communicated further important experiments, which also led
to immediate objections of other scientists active in the field. So Neumeister (1897)
reported in the same year that several tests to proof Buchner’s findings were negative; he
was skeptic towards the hypothesis to treat the agent (yeast press juice) as enzyme,
arguing with the complex function of the “zymase” and its poor stability as compared to
known enzymes. The argument, that proteases would be the reason for this, is disputed,
since yeast press juice would not show degradation by own enzymes. Neumeister
furthermore mentioned the analogy with a finding of Kühne, known for long time, that so
called plasma from muscles formed lactic acid. Buchner (1898a) responded to several
objections. To Neumeister’s remark on proteases he cited Hahn who had shown their
presence in yeast press juice in 1897. Stavenhagen had argued, that some bacteria and
some moulds were observed in Buchner’s press juice; Buchner calculated, that the
amount which was quantified, could not, by far, produce the amount of carbon dioxide
found by him. He furthermore gave new quantitative data on the amount of product
formed by a certain amount of press juice as a function of time under different conditions,
including an antiseptic agent. Schunk (1898) claimed that he had described a “crapp
ferment” earlier (1853/54), an unorganized ferment which showed alcoholic
fermentation, showing a slow formation of gas (carbon dioxide and hydrogen). His
argument could not hold since a quantitative basis excluding microbial activity as the
source of the fermentation was not shown. Further quantitative data were given by
Buchner and Rapp (1898c) on alcohol and carbon dioxide formation, also discussing the
sources of errors, the fermentation of different sugars with and without antiseptic agents
(Buchner and Rapp, 1898c) and the preparation of a dry product from yeast cell juice
active after resolving it in water (Buchner and Rapp, 1898d).
These many details deserve mention since they demarcate a breakthrough against an
established paradigm which taught that processes in living organisms—where alcoholic
fermentation was the most important example—were not of pure chemical nature, but
required a “vis vitalis” (a vital force). Now the chemical paradigm, which reduced all
reactions in physiological (or bio-) chemistry to chemistry without further hidden forces,
began to play the dominant role.
Further findings relevant for the establishment of the chemical nature of enzymatic
catalysis and technical application followed within rather short time. Croft-Hill
performed the first enzymatic synthesis, that of isomaltose, in 1898, allowing a yeast
extract ( -glycosidase) to act on 40% glucose solution (Sumner and Somers, 1953). In
1900 Kastle and Loevenhart found that the hydrolysis of fat and other esters by lipases is
a reversible reaction and that enzymatic synthesis can occur in a dilute mixture of alcohol
and acid (Sumner and Myrbäck 1950). This principle was utilized for the synthesis of
 Applied biocatalysis 10
numerous glycosides by Fischer and coworkers in 1902 and Bourquelot and coworkers in
1913 (Wallenfels and Diekmann, 1966).
Bertrand, in 1897, observed that certain enzymes required dialysable substances to
exert catalytic activity. He named these substances “coenzymes”. Sörensen pointed out
the dependence of enzyme activity on pH in 1909 (Sumner and Somers, 1953). An
important step entering physico-chemistry, and hence extending the theoretical basis of
enzymology, were the kinetic investigations and their interpretation by Michaelis and
Menten. They postulated that enzymatic action is due to the formation of an intermediate
compound between enzyme and substrate, and they presented a mathematical form still
used today (Sumner and Myrbäck, 1950).
The definite establishment of the chemical paradigm was the crystallization of urease
by Sumner in 1926, and further enzymes (trypsin etc.) by Northrup and Kunitz in
1930/31. In every known case the pure enzyme turned out to be a protein (Sumner and
Myrbäck, 1950).

1.4.2 Theoretical Developments

After a long initial period of about a hundred years, with mysterious theories and several
technical applications, where one (diastase) even achieved economic importance,
research on enzymes obtained a chemically scientific status.
E.Fischer (1909) in his work elaborated essential aspects of enzyme catalysis during
the 1890’s. The first aspect is specificity. The agents of the living cell (enzymes) are
optically active and therefore one might assume that the yeast cells with their asymmetric
agents can utilize only those sugars, the geometry of which is not too far from that of
glucose. These observations are the most essential from a range of others, and Fischer
therefrom deduced the famous picture of lock and key, which is a precondition for the
potential of an enzyme to have a chemical effect on the substrate. So he assumes that the
“geometrical form of the (enzyme) molecule concerning its asymmetry, corresponds to
that of the natural hexoses” (sugars).
This concept obviously is far from the former of unorganized ferments (the term for
soluble enzymes). Now an organized structure with high stereochemical precision is
considered as essential for the chemical potential to catalyze highly selective reactions.
The second aspect refers to the protein nature of enzymes. In 1894 Fischer (Fischer,
1909) stated that amongst the agents which serve the living cell the proteins are the most
important. He was convinced that enzymes are proteins. The role of this key problem
may be illustrated with a citation from Fruton (1979): “… the peptide theory was indeed
only a hypothesis fifty years after Franz Hofmeister and Emil Fischer advanced it…” (in
1902). The nature and structure of proteins remained unknown throughout the 19th
century; remarkably, technological applications were nevertheless put into practice since
the middle of the century (see above), based on their action, eventually recognized as
catalysis, only.
Proteins, such as albumin and casein, were included in the group of colloids, which
were attested a dynamic state of matter. “…The colloid possesses ENERGIA, …the
 Introduction 11
probably primary source of the force…of vitality”. Protoplasm was given mystical and
even magical properties, and was widely thought to “lose its virtue and disintegrate into
mundane proteins when extracted”. Willstätter, still in 1927, denied that enzymes were
proteins (Fruton, 1979).
Prior to the breakthrough by E.Buchner’s investigations, Berthelot had tried to
demonstrate cell-free (and thus chemical or enzymatic nature of the) conversion of sugar
to alcohol in the 1850’s, but indeed his approach could not be successful, as was obvious
from Pasteur’s clarification of the manifold sources of microbial infection, notably
inoculation from the air. It was these findings which were a condition for Buchner’s
concept to exclude microorganisms (at least to an essential extent, as he was able to
show) in his experiments. The progress in understanding and methods thus was necessary
for the chemical approach, even if Pasteur opposed it himself.
So the idea, the concept of enzyme action as a general principle in biochemical
reactions was that of Berthelot in the 1850’s, as to the priority. The experimental
verification however was Buchner’s work, and he earned the merit for it.
In order to illustrate the general scientific status of knowledge in microbiology and
(physiological) chemistry of the time, much less stringent as it might seem from the key
results of the leading scientists, we refer to a discussion of the Société Chimique de Paris
in 1897. There Maumené still argues with an “obscure” status of fermentations; he
attributes the action of solid ferments, or their membranes, to strong capillary forces
associated with their tiny forms; the soluble ferments were reduced to a singular agent,
water, which decomposes salts etc., arguments which he had put forward as early as
1858. To the contrary, Béchamp, also referring to his former work in 1853 to 1857, had
demonstrated that the moulds perform the inversion of sugar by a soluble ferment which
they form inside themselves and which they secrete. That was, as Béchamp argues, the
first demonstration that soluble ferments were not an alteration, or transition principle
(“principe d’altration”), but a basic, or original principle (“pure principe immidiat”)
(Bulletin, 1897).
Buchner initiated a new paradigm which, in strict contrast to that of Pasteur, stated that
enzyme catalysis, including complex phenomena like that of alcoholic fermentation, was
a chemical process not necessarily linked to the presence and action of living cells. In his
first paper he wrote that he presented the proof that (alcoholic) fermentation does not
require the presence of “such a complex apparatus as is the yeast cell”. The agent is a
soluble substance, without doubt a protein body, which he called “zymase” (Buchner,
1897). Referring to the heavy controversy on his findings and theory, where Stavenhagen
had pointed out the strict contradiction to Pasteur, Buchner stated that new experimental
findings cannot be disproved by older theories. Furthermore he remarked that the
contradiction to Pasteur was not as strict as it might be seen, since a modification of the
established paradigm would merge the two theories into one, saying that there is no
fermentation without zymase formed by organisms (Buchner and Rapp, 1898a).
Remarkably Fischer also concedes that much earlier other authors, Traube in 1858 and
Berthelot in 1860, had published the same theoretical concept, but that they could not
present the objective proof of it—so highlighting the central role of the experiment
 Applied biocatalysis 12
(Buchner and Rapp, 1898a). The same is true for Liebig’s position, which was
dominating at that time (Schunk, 1898). With Pasteur’s work the vitalistic theory in turn
gained the victory in the 1860’s (Buchner, 1898).
Buchner’s findings marked a new—chemical—paradigm leading research and theory
on enzymes. The activity in scientific research increased significantly due to the new
guidance (Table 1.2). Technical development also got a new, scientific basis on which to
proceed in a rational way.

1.4.3 Technological Development

With the exception of Christian Hansen the industrial development of enzymes was very
slow initially. J.Takamine began isolating bacterial amylases in the 1890’s in what later
became known as Miles Laboratories (today a part of Genencor).
Early applications and patents on enzyme application have been collected by
Neidleman (1991) (Table 1.3).
At the beginning of this century the production of plant lipases was performed by
mechanical disruption of the seed of ricinus after procedures of Nicloux and Hoyer.
These lipases were utilized for the production of fatty acids from oils and fats.

Table 1.3 Selected eaarly enzyme patents (adapted from Neidleman, 1991).

Inventors Year Enzyme Title

J.Takamine 1894 Amylases Process for making diastatic enzyme
J.Takamine 1906 Amylases Diastatic substance and production procedure
O.Röhm 1908 Trypsin, steapsin Preparation of hides for the manufacture of
leather
J.Takamine 1911 Amylases Enzyme
J.Takamine 1911 Amylases Amylolytic enzyme
L.Wallerstein 1911 Malt protease Beer and production procedure
L.Wallerstein 1911 Proteases Preparation of use in brewing
L.Wallerstein 1911 Pepsin Method of treating beer or ale
L.Wallerstein 1911 Papain Method of treating beer or ale
L.Wallerstein 1911 Bromelain Method of treating beer or ale
L.Wallerstein 1911 Yeast Method of treating beer or ale
O.Röhm 1915 Pancreatin Process for cleaning laundry of all types
S.Frankel 1915 Amylase Manufacture of diastase
 Introduction 13

I.Pollak 1915 Amylases Diastase preparations and production

procedure
I.Pollak 1915 Amylases Malt extract and production procedure
A.Boidin/J.Effront 1917 Amylases Process for treating amylaceous substances
A.Boidin/J.Effront 1917 Amylases Process of manufacturing diastases and toxins
by oxidizing ferments
V.G.Bloede 1918 Amylase Process of manufacturing vegetable glue
H.S.Paine/J.Hamilton 1922 Invertase Process for preparing fondant or chocolate soft
cream centers
J.Takamine 1923 Amylases, Enzymatic substance and production
protease, lipase procedure
A.Boidin/J.Effront 1924 Amylase, protease, Treatment of textile fabrics or fibers
lipase
Wallerstein Co. 1931 Amylases, Improvements in process of depilating hides
protease, lipase
M.Wallerstein 1932 Amylases or papain Method of making chocolate syrups
R.Douglas 1932 Amylases Process of preparing pectin
L.Wallerstein 1933 Invertase Invertase preparation and production
procedure
L.Wallerstein 1937 Proteases Process of chill-proofing and stabilizing beers
and ales
L.Wallerstein 1937 Proteases Rubber
L.Wallerstein 1938 Proteases Deproteinization of rubber latex

Details for a manufacture of 10 ton/week are given. It was pointed out that the reaction is
reversible and that an enzymatic synthesis of fat from glycerol and fatty acid was
described by Welter in 1911 (Ullmann, 1914). For the chill-proofing of beer proteolytic
enzymes have been used successfully since 1911 in the USA (Tauber, 1949). Lintner, as
early as 1890, observed that wheat diastase interacts in dough making. This effect was
extensively studied, the addition of malt extract came into practice, and American bakers
in 1922 used 30 million pounds of malt extract valued 2.5 million dollars (Tauber, 1949).
The production of pectinases began around 1930 for use in the fruit industry
(Schweizerische Ferment, now part of Novo Nordisk).
The use of enzymes for the manufacture of leather played a major role for the
industrial scale production of enzymes. For the preparation of hides and skins for tanning,
the early tanners kept the dehaired skins in a warm suspension of the dungs of dogs of
birds. Wood was the first in 1898 to show that the bating action of the unpleasant dungs
was caused by the enzymes (pepsin, trypsin, lipase) which they contained. In the context
 Applied biocatalysis 14
of Wood’s investigations the first commercial bate, called Erodin, was prepared from
cultures of Bacillus erodiens, based on a German patent granted to Popp and Becker in
1896. The bacterial cultures were adsorbed on wood meal and mixed with ammonium
chloride (Tauber, 1949).
In 1907 Röhm patented the application of a mixture of pancreatic extract and
ammonium salts as a bating agent (Tauber, 1949). His motivation as a chemist was to
find an alternative to the unpleasant bating practice. First he tried to apply an ammonia
containing aqueous solution, but long term tests were a failure. Since he knew Buchner’s
work on enzymes he came to assume that enzymes could be the active principle in dung,
and to look for sources which were technically feasible. Tests with pancreas were
successful, when he compared the results with those obtained with dung; whereas
amylase did not work. With this perspective he founded his company in 1907, which
successfully entered the market and had to move to Darmstadt after two years since the
space for expansion was not available in Stuttgart. In 1908 10 tonnes of the product with
the trade name Oropon were sold, followed by 53 tonnes and 150 tonnes in the
subsequent years. In 1913 the company worked with 22 chemists, 30 other employees
and 48 workers (Trommsdorf, 1976).
The history of the Röhm company makes obvious that the market for a new product
providing technical progress was an important factor, but that the background of
scientific knowledge on the principles of enzyme action was equally important as a
condition, leading experiments to a technically feasible solution. A systematic approach
towards the interrelation of scientific development and engineering aspects has been
published by Buchholz (1996).

1.5 DEVELOPMENTS SINCE 1940

Even the development of fermentation of citric acid (Pfizer), and penicillin (Beecham,
Glaxo, Merck, Pfizer, Squibb, Bristol Myers) in the 1920’s and 1940’s, respectively
(Turner, 1994) did not really trigger a scale-up of industrial applications of enzymes.
We must go forward to around 1955 before the development of enzyme production
was gaining speed by growing sales of bacterial amylase and protease. It began in a very
modest fashion. As an illustration, the turnover of the enzyme division of Novo Industri
(now Novo Nordisk), the leading enzyme manufacturer, did not exceed $1 million
annually until 1965. However, with the appearance of the detergent proteases, the use of
enzymes increased. Everybody wanted Biotex, the protease-containing detergent. At the
same time, an acid/enzyme process to produce dextrose using glucoamylase was used
increasingly in starch processing. By 1969, within only four years, Novo’s enzyme
turnover exceeded 50 million US$ annually, in 1997 Novo Nordisk’s enzyme division
had a turnover of approx. 650 million US$. The present global market is estimated to be
around 1.6 billion US$ (Stroh, 1998).
One question—which is as old as industrial enzymes—is: “Can enzymes be re-used?”
Some of the first attempts to reuse enzymes were described by Nelson at the beginning
 Introduction 15
of this century. But the enzymes absorbed to charcoal were very unstable. In the 1950’s,
Georg Manecke was the first really to succeed in making relatively stable systems;
however, he could not convince industry of the importance of further development of his
systems. It became the group of chemists working with Ephraim Katchalski-Katzir in
Israel who really opened the eyes of industry to the world of immobilized enzymes
(among Katzir’s co-workers were Klaus Mosbach, Daniel Thomas and Malcolm Lilly).
The first immobilized enzyme products to be scaled up to pilot plant and industrial
application (in 1969) were immobilized amino acid acylases (i.e. I. Chibata and
colleagues at Tanabe Seiyaku Co. in Japan) and penicillin G acylase
(M.D.Lilly/University College, London, Beecham Pharmaceuticals, England, and
G.Schmidt-Kastner, Bayer, Germany).
The largest immobilized enzyme product—even today—by volume is immobilized
glucose isomerase. The first commercial, enzymatic production of high fructose corn
syrups (HFCS) was in Japan in 1969 (Takasaki), utilizing heat treated Streptomyces cells
containing glucose isomerase in a batch reactor. In the USA Clinton Division of Standard
Brands (now ADM, USA) was the first company using “Takasaki immobilized glucose
isomerase” to make industrial quantities of HFCS around 1971. The sky-rocking sucrose
price, 1973–75, where the price of sucrose increased from 5–7 cents/lb to around 30
cents/lb speeded the interest for HFCS, and thereby immobilized glucose isomerase,
dramatically up. Companies like Novo Industri, and later Gist-brocades developed more
stable enzyme products, which were easy and cheaper to use. Resources were spent on
optimizing fermentation of glucose isomerase, immobilization processes, and the
application processes for the immobilized glucose isomerases. As a result productivity of
the commercial immobilized glucose isomerase products increased from approx. 500 kg
HFCS/kg immobilized enzyme product (1975) to approx. 15,000 kg/kg (1997).
Other immobilized enzyme product successes (where annual production of
immobilized enzymes has surpassed 1 ton/year) comprise: Aminoacylase (Amano),
hydantoinase (Smith Kline Beecham), lactase (Valio), lipase (Novo Nordisk), penicillin
G acylase (Gist-brocades, Smith Kline Beecham, Röhm) and penicillin V acylase (Novo
Nordisk/Gist-brocades) (Poulsen, 1984).

1.6 OUTLINE OF THIS BOOK

Originally almost all applications of biocatalysis involved hydrolytic reactions, except a

few, such as L-sorbose and ephedrine manufacture (Turner, 1994). Hydrolases still are
the main commercial enzyme class, but nowadays a much wider range of reactions is
being applied, either on a commercial scale or on a lab scale. The most important reaction
types are reviewed in chapter 2.
The range of commercial applications types of biocatalysis always has been very wide,
and is continuously expanding further. Therefore, the applications that are treated in this
book have been divided in two types:
 Applied biocatalysis 16
(i) Enzymes applied as processing aids and final products. These examples are reviewed
in chapter 3, and typically originate from the detergent, feed, textile and food industry.
In most of these examples the enzyme will be active in a complex matrix, usually
converting an undesired compound in this matrix and thereby improving its functional
properties. So the focus is on the substrate of the enzymatic reaction rather than on the
product.
(ii) Biocatalysts applied for the production of (bio) chemicals. These processes are
reviewed in chapter 4, and mostly originate from the fine-chemicals and food industry.
In these cases a (solution of a) relatively pure pre-formed precursor molecule is
converted into a (solution of a) target product. So, in contrast to category (i), the focus
is on the product of the reaction rather than on the substrate.

Conversions of primary feedstocks by fermentations (such as glucose to ethanol) are not

included in this book. However, fermentations are usually required to produce the
enzymes or cells in the first place, and therefore chapter 5 includes a review of this type
of fermentation. Chapter 5 also covers the other aspects of biocatalyst production, except
immobilization and protein and genetic engineering, which are treated in chapter 6 and 7,
respectively.
Chapters 8 to 11 deal with optimization of the reaction and process conditions of
biocatalytic processes. Finally, chapters 12 and 13 cover the patent aspects, and the
commercial and financial aspects, respectively.
Valuable comments and remarks by M.Turner and R.Bud are gratefully acknowledged.

1.7 REFERENCES

Anonymous (1862) Ueber die Gährung und die sogenannte generatio aequivoca.
(summary article) J. Prakt. Chem. , 85 , 465–472.
Beral, M.P.-J. (1815) Notes sur la fermentation. J. de Pharmacie et des Sciences
Accessoires , 1, 358–61.
Berthelot, M. (1857) Ueber die geistige Gährung. J. Praktische Chemie , 71 , 321–325.
Berthelot, M. (1864) Remarques sur la note de M.Béchamp relative à la fermentation
alcoolique. Bulletin Soc. Chim . 392–393.
Buchholz, K. (1996) Reflections on the history and scientific character of Biochemical
Engineering. Adv Molec Cell Biol , 15A, 117–134.
Buchner, E. (1897) Alkoholische Gährung ohne Hefezellen. Ber. D. Chem. Ges. , 30 ,
117–124.
Buchner, E. (1898) Ueber zellfreie Gährung. Ber. D. Chem. Ges. , 31, 568–574.
Buchner, E. and Rapp, R. (1898a) Alkoholische Gährung ohne Hefezellen. Ber. D. Chem.
Ges. , 31 , 209–217.
Buchner, E. and Rapp, R. (1898b) Alkoholische Gährung ohne Hefezellen (5.). Ber. D.
Chem. Ges. , 31, 1084–1090.
Buchner, E., and Rapp, R. (1898c) Alkoholische Gährung ohne Hefezellen (6.). Ber. D.
Chem. Ges. , 31, 1090–1094.
 Introduction 17
Buchner, E., and Rapp, R. (1898d) Alkoholische Gährung ohne Hefezellen (7.). Ber. D.
Chem. Ges. , 31, 1531–1533.
Bud, R. (1992) The zymotechnic roots of biotechnology. BJHS (The British Journal for
the History of Science) , 25 , 127–144.
Bud, R. (1993) The uses of life, A History of Biotechnology. Cambridge University
Press.
Bulletin (1897) Extrait des Procès—Verbaux des Séances. Bulletin Soc. Chim. , 17, 769.
Dobell, H. (1869) Bull. Soc. Chim. , 1, 506.
Fischer, E. (1909) Untersuchungen über Kohlenhydrate und Fermente . Berlin:
J.Springer.
Frankland, E. (1885) On chemical changes in their relation to microorganisms. J. Chem.
Soc. , 47, 159–183.
Fruton, J.S. (1979) In The Origins of Modern Biochemistry , edited by P.R.Srinivasan,
J.S.Fruton and J.T. Edsall, pp. 1–18. New York Academy of Sciences.
Knapp, F. (1847) Lehrbuch der chemischen Technologie , Vol. 2. Braunschweig:
F.Vieweg und Sohn.
Neidleman, S.L. (1991) Enzymes in the food industry: a backward glance. Food
Technology , 45(1), 88–91.
Neumeister, R. (1897) Bemerkungen zu Eduard Buchner’s Mittheilungen über “Zymase”.
Ber. D. Chem. Ges. , 30, 2963–2966.
Ost, H. (1900) Lehrbuch der Chemischen Technologie . Hannover: Verlag Gebr. Jänecke.
Payen, A. (1874) Handbuch der technischen Chemie . Nach A.Payens Chimie
industrielle, frei bearbeitet von F.Stohmann and C.Engler, Vol. II, p. 127. Stuttgart:
E.Schweizerbart’sche Verlagsbuchhandlung.
Payen, A. and Persoz, J.F. (1833) Mémoire sur la diastase, les principaux produits de ses
réactions, et leurs applications aux arts industriels. Annales de Chimie et de Physique ,
2me. Série 53 , 73–92.
Poppe, J.H.M. von (1842) Volks-Gewerbslehre oder allgemeine und besondere
Technologie . Stuttgart: Carl Hoffmann.
Poulsen, P.B. (1984) Biotechnology and Genetic Engineering Reviews , Vol.1. Newcastle
upon Tyne: Intercept Inc.
Roberts, S.M., Turner, N.J., Willets, A.J. and Turner, M.K. (1995) Biocatalysis , p. 1.
Cambridge: Cambridge University Press.
Schunk, E. (1898) Alkoholische Gährung ohne Hefezellen. Ber. D. Chem. Ges. , 31, 309.
Stroh, W.H. (1998) Industrial enzymes market: Growth experiences from new products
and movement into global market, Genetic Engineering News , March 1, pp. 11 and 38.
Sumner, J.B. and Myrbäck, K. (1950) In The Enzymes , Vol.1, Part 1, 1–27.
Sumner, J.B. and Somers, G.F. (1953) Chemistry and Methods of Enzymes , XIII–XVI.
New York: Academic Press.
Tauber, H. (1949) The Chemistry and Technology of Enzymes . New York: Wiley.
Trommsdorf, E. (1976) Dr. Otto Röhm—Chemiker und Unternehmer . Düsseldorf: Econ.
Turner, M. (1994) Biological catalysis and biotechnology. In The Chemical Industry ,
edited by A.Heaton, 2nd ed., pp. 306–372. Blackie Academic and Professional.
Ullmann, F. (1914) Enzyklopädie der technischen Chemie , Vol 5, p. 445. Berlin: Urban
und Schwarzenberg.
Wagner, R. (1857) Die chemische Technologie . Leipzig: O.Wiegand.
Wallenfels, K. and Diekmann, H. (1966) In Hoppe-Seyler , Vol. 6B, 1156–1210.
 2.
REACTIONS CATALYSED BY ENZYMES
THORLEIF ANTHONSEN
Department of Chemistry, Norwegian University of Science and Technology, N-
7491 Trondheim, Norway
Telephone: +47 73956206; Telefax: +47 73956255; Email:
[email protected] , [email protected]

ABSTRACT
The use of enzymes and whole cells as catalysts in organic chemistry is
described. Emphasis is put on the chemical reactions and the importance of
providing enantiopure synthons. In particular kinetics of resolution is in focus.
Among the topics covered are enzyme classification, structure and mechanism
of action of enzymes. Examples are given on the use of hydrolytic enzymes
such as esterases, proteases, lipases, epoxide hydrolases, acylases and amidases
both in aqueous and low-water media. Reductions and oxidations are treated
both using whole cells and pure enzymes. Moreover, use of enzymes in sugar
chemistry and to produce amino acids and peptides are discussed.

2.1 INTRODUCTION

2.1.1 A Brief Overview of Solved and Difficult Problems in Organic

Synthesis
Organic chemical synthesis is a mature and highly developed science. In particular for the
synthesis of pharmaceutical compounds, sophisticated tools for synthesis is more and
more needed. The key word for synthesis is selectivity which is necessary to obtain high
yield of a specific product. Synthesis is primarily concerned with the building up of
compounds and since organic chemistry is defined as “the chemistry of carbon
compounds” the formation of carbon-carbon bonds is central. A central role in such
reactions is played by the carbonyl group in ketones and aldehydes. Additions of
nucleophiles such as cyanide, enolates (aldol), alkylmagnesium halides (Grignard), ylides
(Wittig), halides (Reformatsky) etc. lead to formation of C-C bonds. Another important
 Reactions catalysed by enzymes 19
reaction for this purpose is combination of dienes with dienophiles (Diels-Alder). For use
in the laboratory there are numerous reduction and oxidation reactions whose specificity
are well known. Most of these reactions are based on the use of metals. To sum up there
are a large range of selective organic reactions available for most synthetic needs,
however, there is still one area where organic chemists are struggling, that is when
stereochemistry is involved, but considerable progress has been achieved in recent years.
Using chiral auxiliaries good enantiomeric excesses have been obtained in alkylations of
ketones and aldehydes (Evans, Masamune and others). Catalytic chiral epoxidations and
dihydroxylations have been performed using (−) or (+)-diethyl tartrate and cinchona
alkaloids as source of chirality, respectively (Sharpless). Chiral catalysts consisting of
binaphthol chelated metals have been successfully employed for reductions of ketones
(Noyori). Recently, using chiral binaphthol Mn (II) catalysts, racemic epoxides have been
resolved by hydrolysis (Jacobsen).

2.1.2 Why are Enzymes of Interest to an Organic Chemist?

Almost all chemical reactions need a catalyst to take place for instance by, acid, base,
metals etc. Why then bother to use enzyme catalysis in the laboratory or in process
industry? Well, there are some marvelous advantages that enzymes offer that are difficult
to obtain by conventional catalysis. First of all it is the selectivity and specificity that
enzymes show in their catalysis. No matter how simple or trivial the enzyme catalysed
chemical reaction is, this may be on three levels, chemo-, regio and stereo-selectivity or
specificity. It may be worth mentioning that the term selectivity is used when the starting
materials are prochiral and the products are stereoisomers produced in unequal amounts,
while specificity is used when the starting materials are stereoisomers which give
different products in the similar reaction. In particular the stereochemical properties of
enzymes are important, but also the fact that enzymes work under mild conditions is an
attractive property. The latter is becoming more and more important as greater demands
are made on chemical process industry concerning environmental aspects.

2.1.3 Classification of Enzymes

Enzymes, the hammer, the saw, the glue, briefly the tools of biocatalysis are by the
Enzyme Commission, International Union of Biochemistry and Molecular Biology,
which is a subdivision of the Federation of Biochemistry, divided into 6 classes according
to the chemical reactions they catalyse (IUBMB, 1992). The six classes are:

1. Oxidoreductases
2. Transferases
3. Hydrolases
4. Lyases
5. Isomerases
6. Ligases
 Applied biocatalysis 20

All of the complicated molecules of nature are made by enzymes, however, nature does
not use the same types of reactions to do the job. There are for instance at present no
enzymatic parallel to the Diels-Aider reaction although catalytic antibodies recently have
been produced for the purpose (Romesberg et al., 1998). On the other hand very recently
enzymes from spinach have been reported to break down TNT (Wilson, 1998)! The
classes that are currently most used by chemists are oxidoreductases, hydrolases and
aldolases, the latter belonging to the lyases. However, it still remains to be seen which
class of enzymes that will have most success in chemical and pharmaceutical industry.
There is for instance very promising results in synthesis of complicated oligosaccharides
using transferases. One may state that if the target is important enough, it will be possible
to get access to suitable enzymes for a process.

2.2 SYNTHESIS OF CHIRAL BUILDING BLOCKS

2.2.1 Importance of Enantiopure Compounds

No matter if a pair of enantiomers have exactly the same chemical and physical properties
such as melting point, boiling point and spectra and even show the same reactivity in an
achiral environment, they are in principle totally different compounds when they interact
with chiral molecules (Collins, Sheldrake and Crosby, 1992; Collins, Sheldrake and
Crosby, 1997; Sheldon, 1993). Such chiral molecules may be receptors or other proteins
of the body. It is well known that enantiomers may have different odor and taste. For
instance (S)-carvone tastes caraway while the (R)-enantiomer tastes of spearmint (Figure
2.1). A useful metaphor for interaction of receptors with the wrong enantiomer may be
trying to fit the left hand glove on the right hand. The effect of different enantiomers may
be particularly significant for drugs. Hence drugs that are chiral must be administered as
single enantiomers.

2.2.2 Chiral Drugs

Among the 1200 drugs presently (1998) under development, approximately half are
developed as single enantiomers, 32% are achiral and the remaining 17% are either
racemates or no decision has been made on the issue. In years to come it is expected that
these figures will be shifted even more towards enantiopure drugs. For some racemic
drugs already approved there are currently being developed processes for single
enantiomers.
It is common to call the most active enantiomer eutomer and the less active one
distomer. The ratio of the pharmacological activity of these are called the eudismic ratio
(ER) and Pfeiffer’s rule states that “the lower the effective dose of a drug the greater the
difference in the pharmacological effect of the optical isomers”. When two enantiometic
 Reactions catalysed by enzymes 21
drugs have different pharmacological effect it may be that the distomer has no effect at
all and just represents a ballast. -Blockers like propranolol (Figure 2.1) are examples of
this. In this case the (S)-enantiomer is 130 times as active as the (R)-form. In many cases
the distomer may have serious side effects like for the anesthetic ketamin (Figure 2.1).
The distomer, the (R)-enantiomer is a hallucinogen and since the drug is used as
racemate, hallucinations may be a postanesthesia sideeffect. The most well known drug
that belongs to this category is thalidomide (Figure 2.1). This drug was earlier used as a
sedative and the active drug is the (R)-form. The (S)-form is teratogenic and caused many
tragedies in the 60s. There are also examples of chiral drugs in which both enantiomers
have independent therapeutic value or they may even have positive effects in
combination.

Figure 2.1 Examples of chiral molecules where the enantiomers have different
biological activity. (S)-Carvone tastes of caraway while the (R)-
enantiomer tastes of spearmint. The (S)-form of asparagine tastes
 Applied biocatalysis 22

bitter while the (R)-enantiomer is sweet. The (S)-enantiomer of the -blocker

propranolol is 130 times as active as the (R)-form. (S)-Ketamin is an
anesthetic drug while the (R)-enantiomer is a halucinogen.
Thalidomide was earlier used as a sedative and the active drug is the
(R)-form. The (S)-form is teratogenic and caused many tragedies in
the 60s.

2.2.3 Chiral Building Blocks for Synthesis

A major reason why synthetic chemists have become interested in biological methods as
mentioned above, is that biocatalysis shows selectivity and specificity in catalysis. This
interest in turn is mainly due to the need to synthesise enantiopure compounds as chiral
building blocks for drugs and agrochemicals.
Chiral building blocks can be provided by three basically different methods;

i) by chemical transformation of enantiopure natural products

ii) by asymmetric synthesis
iii) by resolution of a racemic mixture.

Enzymes as chiral catalysts play a role in all three methods. In nature enzymes catalyse
all production of chiral compounds. In the laboratory enzymes can catalyse asymmetric
synthesis, as well as resolve racemates. Which of the three methods is chosen in different
cases depends on several factors, like price of starting materials, number of synthetic
steps, available production technology and know-how etc. There is at present a constant
ongoing development of synthetic methods and biotransformation is one field. Utilization
of method i) requires knowledge of classical organic synthesis, enzymes have already
played their role. Enzymes may play a part both in asymmetric synthesis and resolution.

2.3 HYDROLYTIC ENZYMES

As mentioned in part 2.1.3 hydrolytic enzymes are the most frequently used enzymes in
organic chemistry. There are several reasons for this. Firstly, they are easy to use because
they do not need cofactors like the oxidoreductases. Secondly, there are a large amount of
hydrolytic enzymes available because of their industrial interest. For instance detergent
enzymes comprise proteases, cellulases, amylases and lipases. Even if hydrolytic
enzymes catalyse a chemically simple reaction, many important features of catalysis are
still contained such as chemo-, regio- and stereoselectivity and specificity.
 Reactions catalysed by enzymes 23

2.3.1 Ester Hydrolysis and Synthesis

The natural task of hydrolytic enzymes is to hydrolyse ester, amide and glycoside bonds.
They have different substrate specificity, however, many hydrolases can accept a wide
range of substrates. Examples of ester hydrolysis are given in Figure 2.2a (asymmetric
synthesis from a dioic ester using an esterase), 2.2b (asymmetrization of a meso-diester
with a lipase) and 2.2c (resolution of a secondary ester catalysed by a lipase). Carboxyl
esterases and carboxyl lipases have in common that they act on carboxylic esters. They
differ in the structure of esters they hydrolyse. Since lipases in nature first of all
hydrolyse fats, i.e. triesters of glycerol with relatively uncomplicated (not branched)
alkanoic acids, they seem to prefer such substrates in general (Figure 2.3). Esterases on
the other hand are not that specific and may accept bulky acyl groups (R3). Moreover, the
structure of the alkoxy part is not so critical. A general observation, however, is that if the
stereocentre is far away from the centre of reaction (ester oxygen), the catalysis is not
stereospecific. With the currently available esterases and lipases it may be summarised
that:

• both esterases and lipases give mild hydrolysis

• lipases do not hydrolyse esters with “bulky” acyl groups
• hydrolysis of esters of secondary alcohols (R1, R2≠H) are more stereospecific

Of course the complete picture is more complicated.

Carboxylic esterases and carboxylic lipases have been tested with phosphate esters, but
without success. This is easy to understand on the basis of the mechanism which
 Applied biocatalysis 24

Figure 2.2 Production of enantiopure compounds using hydrolytic enzymes. In

(a) a prochiral diester is hydrolysed to yield predominance (in theory
100%) of one enantiomer. In the next example (b) a meso-diester is
hydrolysed to yield predominance (in theory 100%) of one
enantiomer of the monoester. If k1>k2 the (1S, 2R)-enantiomer is
formed to the greatest extent. Due to the preference of the enzyme
k4>k3 and the lower monoester (1R, 2S) will be removed fastest.
Hence both steps will lead to an increase of the upper enantiomer at
the monoester stage. If the reaction proceeds to completion, however,
the result will be another meso-compound, a diol. In example (c) a
racemic secondary ester is resolved by hydrolysis. One monoester is
hydrolysed faster than the other and this leads to kinetic resolution.
 Reactions catalysed by enzymes 25
has a tetrahedral transition state. Hydrolysis of phosphate esters requires a penta
coordinated transition state which has a completely different geometry. Classically, a
major difference between esterases and lipases is based on their kinetics. Using a slightly
soluble ester an esterase will show increasing activity as the substrate concentration is
increased. A typical lipase will not show increased activity until the substrate
concentration is beyond a point called the critical micellar concentration (CMC)
(Desnuelle, 1972). That is when the substrate concentration is so high that

Figure 2.3 Carboxyl esterases and carboxyl lipases both act on carboxylic
esters. They differ in the structure of esters they hydrolyse. Lipases
work best on substrates with relatively uncomplicated (not branched)
acyl parts. Esterases on the other hand are not that specific and may
accept bulky acyl groups (R3). Moreover, the structure of the alkoxy
part is not so critical. A general observation, however, is that if the
stereocenter is far away from the center of reaction (ester oxygen),
the catalysis is not stereospecific. Hydrolysis of esters of secondary
alcohols (R1, R2≠H) are more stereospecific.

the substrate is not soluble anymore and the enzyme will come in contact with droplets of
ester. This observation has been called interfacial activation. On a molecular basis lipases
are seen to undergo a rearrangement on an interface (Rubin, 1994). More specifically, the
active site is covered by a lid which opens in contact with a hydrophobic interface.
It is a well established technique to use enzymes in organic solvents or low-water
media (Koskinen and Klibanov, 1996). There are several reasons for choosing an organic
solvent as medium for catalysis by hydrolytic enzymes such as better solubility of
substrate and product, better stability of enzyme (most deactivating processes need water
to occur) and simpler removal of solvent. Moreover, the possibility to synthesize
carbohydrate derivatives, peptides and esters by reverse hydrolysis, are also interesting
aspects. Generally, enzyme catalysis depends on the medium. Properties of the medium,
which solvent, expressed by log P (Laane, 1987; Laane et al., 1987), co-solvents, the
amount of water in the system as expressed by the water activity (aw) etc. (Kvittingen,
1994) is treated more thoroughly in Chapter 9. Use of organic solvents in resolutions is
discussed in part 2.3.4.
It is possible to catalyse the formation of esters from acid and alcohol by a hydrolase.
However, the water formed in the reaction creates a problem for the equilibrium of the
 Applied biocatalysis 26
reaction, but also for the enzyme itself which tends to associate with the formed water
and become inactive (Kvittingen et al., 1992). Therefore transesterification is a much
more frequently used procedure for ester formation.

2.3.2 Amide Hydrolysis and Synthesis

Production of amino adds

Amino acids may be produced by biocatalysis either by asymmetric synthesis or

resolution (Figure 2.4). Addition of ammonia to the double bond of an , -unsaturated
carboxylic acid using a lyase such as fumarase is one example of the former. For instance
L-phenylalanine has been produced from cinnamic acid in

Figure 2.4 Biocatalytic synthesis of amino acids either by asymmetric

synthesis or resolution. Addition of ammonia to the double bond of
an , unsaturated carboxylic acid using a lyase such as fumarase
 Reactions catalysed by enzymes 27

or reductive amination by NH4 + of the carbonyl group of an -keto acid using

for instance alcohol dehydrogenase are examples of asymmetric
synthesis. The most frequently used method is by resolution of an -
amino acid obtained for instance by the Strecker synthesis, using
hydrolytic enzymes. All of the mentioned methods give
predominantly the natural L-form of the amino acid. In resolution
both the L- and the D-forms are obtainable. There are three
hydrolytic resolution processes, an ester of the carboxylic group may
be hydrolysed by an esterase or protease, an amide at the carboxylic
group may be hydrolysed by an amidase or an acyl group at the
amino group may be hydrolysed by an acylase.

Figure 2.5 Treating an aldehyde with ammonia and hydrogen cyanide produces
an -amino nitrile. By hydrolysis of the nitrile group an -amino
acid is produced. This synthesis is called the Strecker synthesis.

high yield (Hamilton et al., 1985). Another method is reductive animation by NH4 + of
the carbonyl group of an -keto acid using for instance alcohol dehydrogenase. The most
frequently used method, however, is resolution of an -amino acid obtained for instance
by the Strecker synthesis (Figure 2.5) using hydrolytic enzymes. All of the mentioned
methods give predominantly the natural L-form of the amino acid. By using resolution
both the L- and the D-forms are obtainable, however, since the L-form has higher
economic interests, processes are designed to give only this enantiomer. There are
basically three ways to use hydrolytic enzymes in resolution processes, an ester of the
carboxylic group may be hydrolysed by an esterase or protease, an amide at the
carboxylic group may be hydrolysed by an amidase or an acyl group at the amino group
may be hydrolysed by an acylase. The latter process is exploited commercially (Degussa
AG, Germany) (Leuchtenberger, Karrenbauer and Plöcker, 1984) in a membrane reactor
(hollow fiber) where the unreacted D-isomer is separated, racemised and recycled.
Another example is resolution of -amino acid esters under conditions of dynamic
resolution. Using catalytic amounts of pyridoxyl-5-phosphate, which forms a Schiff’s
base with the ester and not the acid, the unreacted D-ester was racemised in situ and for
instance L-tyrosine was obtained in 97% ee and 95% yield (Figure 2.6) (Chen, Huang
and Wang, 1994).
 Applied biocatalysis 28

Figure 2.6 By resolution of -amino acid esters under conditions of dynamic

resolution 100% of a single enantiomer may be produced. Using
catalytic amounts of pyridoxyl-5-phosphate, which forms a Schiff’s
base with the ester and not the acid, the unreacted D-ester may be
racemised in situ and for instance L-tyrosin has been obtained in 97%
ee and 95% yield.

Peptide synthesis

Formation of an amide bond (peptide bond) will take place if an amine and not an alcohol
attacks the acyl enzyme. If an amino acid (acid protected) is used, reactions can be
continued to form oligo peptides. If an ester is used the process will be a kinetically
controlled aminolysis. If an amino acid (amino protected) is used it will be reversed
hydrolysis and if it is a protected amide or peptide it will be transpeptidation. Both of the
latter methods are thermodynamically controlled. However, synthesis of peptides using
biocatalytic methods (esterase, lipase or protease) is only of limited importance for two
reasons. Synthesis by either of the above mentioned biocatalytic methods will take place
in low water media and low solubility of peptides with more than 2–3 amino acids limits
their value. Secondly, there are well developed non-biocatalytic methods for peptide
synthesis. For small quantities the automated Merrifield method works well.
 Reactions catalysed by enzymes 29
Nevertheless, one process for synthesis of the low calorie sweetener, Aspartame, which is
a methyl ester of a dipeptide, (Asp-Phe-OMe) involves a biocatalytic step. Aspartic acid
amino protected by benzyloxycarbonyl group, is reacted with two moles of phenylalanine
methylester catalysed by the protease thermolysin. The extra mole of ester makes the
dipeptide precipitate and it is later recycled. For details see section 4.6.

2.3.3 Resolution by Hydrolysis. Irreversible Reactions.

A racemate of a desired building block may often be produced easily by conventional
organic synthetic methods. For instance racemic amino acids can be obtained by the
Strecker synthesis from an aldehyde, ammonia and hydrogen cyanide (See 2.3.2).
Classic resolution has been performed by formation of diastereomeric salts which
could be separated. In a series of synthetic steps and when resolution is one step, it is of
utmost importance that the correct chirality is introduced at an early stage. When a
racemate is subject to enzyme catalysis, one enantiomer reacts faster than the other and
this leads to kinetic resolution (Figure 2.2c). Results of hydrolysis using lipase B from
Candida antarctica (CALB) and a range of C-3 secondary butanoates are shown in Table
2.1.
Since a similar reaction with two stereoisomers leads to different results (one
enantiomer reacts, the other does not) kinetic resolution should be called enantiospecific.
However, this does not seem to have become common terminology, so it may be better to
characterise the reaction by the enantiomeric ratio E. The enantiomeric ratio is the ratio
of the specificity constants of the enzyme for the two enantiomers

Table 2.1 Enantiomeric ratios (E) obtained in hydrolysis of butanoates 1a–15a and
transesterification of alcohols 1b–9b and 13b–15b using 2-chloroethyl
butanoate as acyl donor and lipase B from Candida antarctica (CALB) as
catalyst. For significance of R1 and R2 see Figure 2.9. E-values were
determined from chromatographically measured ee p and ee s at several degrees
of conversion and using the computer program E&K calculator version 2.03.
Note that the stereopreference of CALB changes from (R) to (S) when R2=F,
Cl and Br.
 Applied biocatalysis 30

Compound R2 R1 E Hydrolysis E Transesterific

1 H Ph 900 139
2 H CH2Ph 277 22
3 H CH2CH2Ph 600 319
4 F Ph 350 134
5 F CH2Ph 57 30
6 F CH2CH2Ph 710 193
7 Cl Ph 354 84
8 Cl CH2Ph 45 23
9 Cl CH2CH2Ph 170 42
10 OCH3 Ph 22
11 OCH3 CH2Ph 21
12 OCH3 CH2CH2Ph >100
13 Br Ph 999 35
14 Br CH2Ph 96 7
15 Br CH2CH2Ph 327 38

and it is a very important parameter for a resolution process. In resolution there are two
products called product and remaining substrate. Both products can reach very high
enantiomeric excess provided E is high. The enantiomeric excesses are termed ee p and ee
s respectively, and they depend on the degree of conversion c. Both ee p , ee s and c have
values between 0 and 1, but they are sometimes dealt with in %. For an irreversible
reaction such as hydrolysis the E-value may be calculated from either ee p or ee s and the
degree of conversion cat one single measurement according to (Chen et al., 1982):

Since the degree of conversion c under most circumstances (equal amounts of

enantiomers at the beginning of the reaction, no side reactions) is related to ee p and ee s
by:
 Reactions catalysed by enzymes 31

Figure 2.7 Energy profile of resolution or asymmetric synthesis. In both cases

the diasteromeric transition states leading to different enantiomers
have different energy and this difference G ‡ governs the
enantiomeric excess of products.

another expression may be used (Rakels, Straathof and Heijnen, 1993):

The advantage of this latter way to calculate E is that it does not involve c which may be
difficult to measure accurately. As opposed to ee p and ee s which are relative quantities
measured for inst in the same chromatogram, c is an absolute quantity. The most accurate
way, however, is to use a computer program to fit many measured data points at several
conversions to calculated curves for different E-values.
The E-value is related to the difference in free energy of activation for reaction with
the two enantiomers by G ‡ =−RT lnE. A reaction between the enzyme and one
enantiomer passes through a transition state of different energy from the transition state
resulting from reaction with the other (Figure 2.7) and E is constant throughout the
reaction. As mentioned above there are two products in a resolution process the
enantiomeric excesses of which depend on the degree of conversion. This feature is a
major difference from asymmetric synthesis. In asymmetric synthesis there is only one
product and the ee is independent of conversion.
In a resolution ee of the substrate fraction is zero when the reaction starts. Provided the
 Applied biocatalysis 32
enantiomeric ratio E is high, the product fraction will have high ee. For instance if E=19
(95:5), ee p will be 90% at the start of the reaction. As the reaction proceeds, ee p and ee s
will obviously change. The relationship between ee p , ee s and c

Figure 2.8 Enantiomeric excess of product (ee , full lines) and remaining
p of conversion calculated for
substrate (ee s , stippled lines) vs. degree
four different values of enantiomeric ratio E for an irreversible
reaction.

for four different values of E is shown in Figure 2.8 (E=3, 6, 20, and 100). Ideally, if E is
very high (E>100) both ee p and ee s will be close to 100% at 50% conversion and the
reaction virtually stops. Looking at the ee vs. conversion curves for the different E-values
it is inferred that ee p is at its maximum in the beginning of the reaction while ee s reaches
maximum at a later stage. This has an important consequence; even for a low
enantiomeric ratio, it is possible to obtain a very high ee s provided a reduced yield is
tolerable. From Figure 2.8 it is clear that even for a low E-value an ee s close to 100%
may be achieved if yield can be sacrificed. This difference between resolution and
asymmetric synthesis is very important. For this reason it may be easier to obtain the
remaining substrate with higher ee.
Lipase catalysed hydrolysis of racemic esters of the important chiron “solketal”, 1, 2-
O-isopropylidene glycerol, are not very stereospecific due to the fact that they are
primary esters. Secondary esters usually show much higher E-values. Table 2.1 shows E-
 Reactions catalysed by enzymes 33
values obtained for hydrolysis and transesterification of glycerol-like secondary
substrates and it is seen that they are generally very high. The results may be rationalised
on the basis of molecular modeling (Uppenberg et al., 1995). In Figure 2.9 it is shown
that the “small” group (−CH2−R2) at the stereocentre is located in the stereospecificity
pocket which is next to the site of reaction. With a primary ester as substrate the small
group would not reach this pocket.

Figure 2.9 Schematic model of the (R)-enantiomer of a secondary glycerol-like

substrate as its tetrahedral intermediate inside the active site cleft of
lipase B from Candida antarctica (CALB). The “large” group
(−CH2−O−R1) and the acyl part is inserted like a V-shape into the
cleft while the “small” group (−CH2−R2) at the stereocentre is
located in the stereospecificity pocket which is next to the site of
reaction.

2.3.4 Resolution in Organic Solvents. Reversible Reactions.

As mentioned above, in hydrolysis the ester may always be obtained with a high ee, but,
what if the alcohol is required? This problem may be circumvented if the reaction is
inverted. Instead of hydrolysis, a synthesis may be performed, an esterification or better a
transesterification in non-aqueous media (See Chapter 9). Since the enzyme shows the
same stereopreference no matter the direction of the reaction (hydrolysis or
transesterification), either the alcohol or the ester may be separated as the remaining
substrate. If the (S)-ester is the remaining substrate in hydrolysis the (S)-alcohol will be
the remaining substrate of transesterification (Figure 2.10).
For transesterification reactions a starting ester is needed. This is commonly called the
acyl donor and it acylates the enzyme to form the acyl enzyme. Enzyme catalysed
 Applied biocatalysis 34
transesterifications follow the ping-pong bi-bi mechanism in which AX (Substrate 1)
enters the enzyme, forms an enzyme-X (E-X) bond and expels A (Product 1) (Palmer,
1995). Then another substrate B (Substrate 2) reacts with E-X to liberate B-X (Product
2), leaving the enzyme in its original form. A detailed illustration is shown in Figure 2.11
exemplified by esterification of racemic 1-phenoxy-2-propanol (1 in Table 2.1) with
CALB using a butanoic acyl donor. The importance of the reaction is that preferably only
one of the enantiomers is esterified. In this case the (R)-ester was formed with very high
ee.
The mathematical apparatus presented in 2.3.3 is restricted to irreversible reactions.
Hydrolysis, due to the large excess of one reactant, water, is for practical purposes
irreversible. In a transesterification the concentration of the leaving alcohol (from

Figure 2.10 Hydrolysis of a racemic secondary ester or transesterification of

the corresponding secondary alcohol with CALB as catalyst both
yield the same enantiomer as product. The product of hydrolysis is
the (R)-alcohol while the product of transesterification is the (R)-
ester. The R,S-notation in this case is done on the assumption that R1
has higher priority than R2. This is not necessarily in the same order
as “large”, “small” in model considerations.
 Reactions catalysed by enzymes 35

Figure 2.11 Transesterification of a racemic mixture of a secondary alcohol (1-

phenoxy-2-propanol, 1 in Table 2.1) with a butanoic acyl donor
follows a ping-pong bi-bi mechanism in which Substrate 1 (acyl
donor) enters the enzyme, forms an acyl enzyme expelling Product 1
(the leaving alcohol from the acyl donor). Then another Substrate 2
(the enantiomers of the alcohol to be resolved) reacts with the acyl
enzyme to liberate Product 2 (the enantiomers of the produced
esters), leaving the enzyme in its original form. In a kinetic resolution
one of the enantiomeric alcohols reacts faster than the other to form
an excess of one enantiomer of the esters (ideally enantiopure, for 1
the (R)-ester was formed with very high ee). The success of the
resolution is expressed by the enantiomeric ratio E, which depends on
the difference in free energy of activation of the two diastereomeric
transition states. These are in turn related to the two tetrahedral
intermediates.

the acyl donor) will accumulate and eventually the reverse reaction will become
important. This will lead to reduced ee’s for the product and the substrate as the reaction
proceeds. For reversible reactions also the equilibrium constant, K eq has to be taken into
account and another set of formulas are used to determine E.

Computer programs for ping-pong bi-bi kinetics which use ee-values measured at several
degrees of conversion are available (Anthonsen, Hoff and Anthonsen, 1996; Rakels,
Straathof and Heijnen, 1993). If both enantiomers are available in pure forms, it is also
possible to determine E and K eq , from initial rate measurements.
 Applied biocatalysis 36

2.3.5 Problems with Reversibility

As mentioned in 2.2.2 a great problem with transesterification is that the reaction will
become reversible and the equilibrium constant will become important. The enantiomer
that reacts fastest in the forward direction will also react fastest in the

Figure 2.12 Calculated curves for ee (starting upper left corner) and ee
p a resolution with E=100 and three
(starting lower left corner) for s
different equilibrium constants, 10 000, 5 and 0.5. With a large K eq
the reaction is irreversible and the progress curves looks like the
examples of Figure 2.9. For reactions with smaller K eq-values a
dramatic effect is observed for ee s . The curve reaches a maximum,
as the reaction progresses further ee s is reduced and the curve never
reaches 100%. The effect of reversibility on ee p is not as dramatic.
The curve dips down at an earlier degree of conversion when K eq is
lowered.
 Reactions catalysed by enzymes 37

Figure 2.13 Effect on ee and ee -curves when changing the acyl

donor/substratepratio. In sthe example at far left equimolecular
amounts were used. The middle and right hand curves show the effect
of using 3 and 5 times excess of acyl donor respectively.

reverse direction. The effect of this is clearly inferred from Figure 2.12 in which the ee p
and ee s are calculated for a resolution with E=100 and for three different equilibrium
constants. With a large K eq the reaction is irreversible and the progress curve looks like
the examples of Figure 2.9. For reactions with smaller K eq-values a dramatic effect is
observed for ee s . The curve reaches a maximum, as the reaction progresses further ee s is
reduced and the curve never reaches 100% for ee s as it always does in the irreversible
case. The effect on the ee p curve is not as dramatic, it dips down at an earlier degree of
conversion. An obvious way to proceed is to push the reaction towards the product side
by increasing the concentration of the reactants. Figure 2.13 shows that the point where
decrease of ee s occurs is pushed towards higher conversion when the acyl
donor/substrate alcohol is changed through the series 1, 3 and 5, other factors being held
constant. Another way would be to change the nature of the alkoxy group of the acyl
donor, R1 in Figure 2.14. When R1 is changed so that the pKa-value of the corresponding
leaving alcohol is decreased, the reaction becomes less reversible. Completely
irreversible conditions are obtained when vinyl esters are used as acyl donors. After
reaction the expelled vinyl alcohol immediately tautomerises to the corresponding
aldehyde or ketone which provide the irreversibility. However, it has been shown that the
E-value also depends on the nature of the alkoxy group of the acyl donor even if the acyl
donors give the same acyl enzyme, f. inst. the butanoate of the enzyme (Hoff, Anthonsen
 Applied biocatalysis 38
and Anthonsen, 1996). Transesterification of 1-phenoxy-2-propanol shows that both the
E-value and the equilibrium constants vary a lot (Figure 2.15, Table 2.2). The acyl donor
giving irreversible conditions, vinyl butanoate, gives the lowest E-value.

Figure 2.14 Different acyl donors used for resolution of glycerol-like

compounds. Decreasing pKa-value of the corresponding leaving
alcohol (or acid when anhydrides are used) increases the acylating
power and the reaction becomes less reversible. (See also Table 2.2).
Completely irreversible conditions are obtained when vinyl esters are
used as acyl donors. After the reaction the expelled vinyl alcohol
immediately tautomerises to the corresponding aldehyde or ketone.
 Reactions catalysed by enzymes 39

Figure 2.15 Dependence on E-value and K eq of the nature of the alkoxy group
of the acyl donor in transesterification of 1-phenoxy-2-propanol,
circles: 2-chloroethyl butanoate, squares: 2,2,2-trichloroethyl
butanoate, triangles: butanoic anhydride, diamonds: vinyl butanoate,
filled symbols: product fraction, open symbols: substrate fraction.
The acyl donor giving irreversible conditions, vinyl butanoate, gives
the lowest E-value. (See also Table 2.2).

Table 2.2 Enantiomeric ratios E and equilibrium constants K eqobtained in

transesterification of 1-phenoxy-2-propanol using four different acyl donors.

Acyl donor E K eq
2-Chloroethyl butanoate 139 0.32
2,2,2-Trichloroethyl butanoate 26 =10
Butanoic anhydride 8 >20
Vinyl butanoate 4 >20
 Applied biocatalysis 40

2.3.6 How to Improve the Situation When the Enantiomeric Ratio is Low
A biocatalytic process consists of three participants, i) the catalyst (enzyme, whole cells),
ii) the substrate and iii) the medium. As a part of the medium may also be counted the
water activity of low-water media and the acyl donor for transesterifications. If it is
necessary to raise the E-value either of these four factors may be changed (Chen and Sih,
1989; Faber, Ottolina and Riva, 1993). A new catalyst may be found either among
available enzymes or by creation of new catalysts by mutating organisms or a crude
enzyme product may be purified (Quartey et al., 1996). In either case a simple screening
procedure is necessary. Recently a spectrophotometric method has been developed (Janes
and Kazlauskas, 1997). The more typical organic chemical ways of solving the problem
would be to change the substrate slightly, introduce for instance easily removable
protecting groups, or more easily to manipulate the medium. The lipase CALB (see 2.1.7)
was used to hydrolyse 2-butanoates of C-3 substrates and the substituents were varied
(R1=Ph, CH2Ph, CH2CH2Ph, R2=H, F, Cl, Br, OCH3 (see Figure 2.9) (Anthonsen and
Hoff, 1998). The results (Table 2.1) show that the enantiomeric ratio varied between 21
and close to 1000 which clearly illustrates the dramatic effect substrate variation may
have on the E-value. Inspection of the figures indicates that both size and electronic
effect may be significant. Similar effects were observed in transesterifications.
Results from the work mentioned above show that the very important chiron 8,
R1=CH2Ph, R2=Cl, gave only a modest E-value. When screening several solvent systems
it was found that this value increased considerably when the hydrolysis was performed in
30% acetone (Hansen et al., 1995). This effect was recently explained as being due to an
enantiospecific inhibition by the alcohol released upon hydrolysis without co-solvent
(Lundhaug et al., 1998). There are also many other examples of solvent effects in enzyme
catalysis, but it seems to be various origins of the effect (Anthonsen and Jongejan, 1997;
Kanerva et al., 1990; Zaks and Klibanov, 1988).
As discussed in part 2.3.4 choice of acyl donor in transesterifications may have
significant influence. The specificity of enzymes may also be influenced by the water
activity, aw. There are reports showing both positive and negative reports on E. Since
water is introduced in the medium other equilibria are established such as between the
acyl donor and water and between the produced ester and water, both leading to the
formation of acid in the system. How this influences the reaction is not clear.
 Reactions catalysed by enzymes 41

Figure 2.16 The mechanism of soluble epoxide hydrolase starts with a

nucleophilc attack by Asp333 on a carbon of the epoxide (usually the
least hindered one) to form a glycol monoester intermediate which is
stabilised by an oxyanion hole. A water molecule, activated by
His523 releases the 1,2-diol product. An aspartic residue, Asp495,
helps in this latter step. The overall reaction is a trans-anti periplanar
additon of water to the epoxide and thus it resembles acid catalysed
opening of epoxides. This mechanism implies that the configuration
is inverted provided the site of attack is a stereocentre.

2.3.7 Epoxide Hydrolases

Enantiopure epoxides and vicinal diols are important versatile chiral building blocks for
pharmaceuticals (Hanson, 1991). Their preparation has much in common and they may
also be converted into one another. These chirons may be obtained both by asymmetric
synthesis and resolution of racemic mixtures. When planning a synthetic strategy both
enzymic and non-enzymic methods have to be taken into account. In recent years there
has been considerable advance in non-enzymic methods as mentioned in part 2.1.1.
Formation of epoxides and vicinal diols from aromatics is important for the break down
of benzene compounds in nature (See part 2.6.5).
Enzyme catalysed hydrolysis of racemic epoxides is interesting from a practical point
of view. This reaction is catalysed by epoxide hydrolases (EHs, EC 3.3.2.3) (Archelas
and Furstoss, 1998). Mammalian EHs are the most widely studied and they are divided
into five groups among which the soluble (cytosolic) epoxide hydrolases (sEH) and
microsomal epoxide hydrolases (mEH) are best charactelised. The mechanism of sEH
from rat starts with a nucleophilic attack by Asp333 on a carbon of the epoxide (usually
the least hindered one) to form a glycol monoester intermediate which is stabilised by an
oxyanion hole. A water molecule activated by His523 releases the 1,2-diol product. An
 Applied biocatalysis 42
aspartic residue, Asp495, helps in this latter step (Figure 2.16). The overall reaction is a
trans anti periplanar addition of water to the epoxide and thus it resembles acid catalysed
opening of epoxides. This

Figure 2.17 Action of epoxide hydrolases on a mono-substituted epoxide (right

hand side of figure). Different epoxide hydrolases may prefer either
the primary or the secondary centre as site of attack and only reaction
at the secondary centre will result in inversion of stereochemistry. In
an EH catalysed hydrolysis proceeding with inversion, both the
unreacted epoxide and the product diol will have the same
configuration as defined by the R,S-convention. It may be worth
mentioning that this convention in some cases may be troublesome
for some epoxides such as glycidol derivatives. In such cases the
carbon atom at the stereocentre has to be introduced as a “phantom
atom” when deciding the priority of the ring derived substituent
relative to the exocyclic one. This implies, as shown on the left hand
side of the figure, that for glycidol itself (R=H) the ring derived
substituent will have priority while for all other substituents (R≠H) it
will not. The oxygen atom in the ring will of course always have
highest priority.

mechanism implies that the configuration is inverted provided the site of attack is a
stereocentre. The most reactive epoxides, and often the most interesting ones from a
synthetic point of view, are mono substituted. In this case only reaction at the secondary
centre will result in inversion of stereochemistry, reaction at the primary centre will not
(Figure 2.17). It should also be mentioned that different epoxide hydrolases may prefer
 Reactions catalysed by enzymes 43
either the primary or the secondary centre as site of attack.
In an EH catalysed hydrolysis proceeding with inversion, both the unreacted epoxide
and the product diol will have the same configuration as defined by the R,S-convention. It
may be worth mentioning that this convention in some cases may be troublesome for
some epoxides such as glycidol derivatives. In such cases the carbon atom at the
stereocentre has to be introduced as a “phantom atom” when deciding the priority of the
ring derived substituent relative to the exocyclic one. This implies, as shown in Figure
2.17, that for glycidol itself (R=H) the ring derived substituent will have priority while
for all other substituents (R≠H) it will not. The oxygen atom in the ring will of course
always have highest priority.

Figure 2.18 Resolutions of epoxides using microsomal epoxide hydrolases.

substituted styrene oxides gave very high E-values when R=Me or Et
(a) while cis-1,2-dialkyl substituted epoxides were attacked at the (S)-
carbon atom only, giving rise to one single product, the (R,R)-diol
(b). This is an example of what is known as enantioconvergence, ie.
two enantiomers reacting differently and thus giving rise to only one
product theoretically in 100% yield and ee.

Generally monosubstituted and cis-1,2-disubstituted epoxides are good substrates for EH

while tri-, tetra or trans-1,2-disubstituted ones are poor or non-substrates. Resolutions of
epoxides using microsomal epoxide hydrolases, mEHs show that cis-2-alkyl substituted
styrene oxides gave very high E-values when R=Me or Et (Figure 2.18a). A series of cis-
 Applied biocatalysis 44
1,2-dialkyl substituted epoxides were attacked at the (S)-carbon atom only, giving rise to
one single product, the (R,R)-diol (Figure 2.18b). This is an example of what is known as
enantioconvergence, i.e. two enantiomers reacting differently and thus giving rise to only
one product theoretically in 100% yield and ee. The results accord with the knowledge of
the active site of EH in which both of the substituents, R1 and R2 fit equally well when
the epoxide ring points in the same direction.
Microbial EHs have been used to resolve geminally disubstituted epoxides. In
particular species of Rhodococcus and Mycobacterium have given high E-values. A rule,
based on the size of the substituents, has been formulated in order to predict the
stereospecificity of these EHs. When using whole cells of Aspergillus niger and Bacillus
sulfurescens to hydrolyse styrene oxide, it was discovered that the former attacked the
stereocentre and produced the diol in (R)-configuration and left the (S)-epoxide
untouched. B. sulfurescens gave the same configuration of the diol, however, since the
(R)-epoxide was left unreacted the (R)-diol was a result of attack at the primary centre of
the epoxide. Logically, a mixture of the two organisms gave only one product, (R)-1-
phenyl-1,2-ethanediol in high yield and ee.

2.3.8 Derivatization of Sugars

The most characteristic feature of sugars chemically speaking, is all the more or less
chemically equivalent hydroxy groups. Performing selective synthetic transformations on
sugar molecules is therefore always a matter of protection and deprotection steps. The
need for selective methods is great and sugar chemists have for decades taken this
discipline to perfection. Nevertheless, tempted by the aspect of developing methods not
needing protection and deprotection steps, enzyme catalysis has been exploited.
Regioselective acylations have been the target, however, with modest success. Sugars
which are typically hydrophilic molecules, have been dissolved in solvents like pyridine
and DMF and acylated using hydrolytic enzymes. The selectivity obtained has been
mostly on the primary vs. secondary level, a kind of selectivity that is easily obtained by
classical methods. However, when more than one -CH2OH group is present, enzyme
catalysis may be of importance since acylation is mainly directed towards one specific
group. A drawback is the necessity to use high boiling solvents which are difficult to get
rid of. Due to low solubility of carbohydrates in organic solvents and lack of selectivity
of reactions, it may be concluded that this strategy for use of hydrolytic enzymes in the
carbohydrate field will only have limited value.

2.3.9 Glycosidases
Carbohydrates such as trioses, tetroses, pentoses, hexoses and polysaccharides are
extremely important molecules in nature. The biological significance of oligosaccharides
for cell-cell interaction is increasingly understood (Fukuda and Hindsgaul, 1994). The
simple carbohydrates are the building blocks of oligo- and polysaccharides. Biocatalysis
 Reactions catalysed by enzymes 45
is important for synthesis of simple carbohydrates (aldolases) as well as of oligo- and
poly-saccharides (transferases) (Augé and Crout, 1997; Drueckhammer et al., 1991;
Toone et al., 1989; Wong et al., 1995; Wong et al., 1995b).
The enzymes involved in break down and build up of oligo- and polysaccharides in
nature are either glycosidases or glycosyl transferases. The first type is a hydrolytic
enzyme and mainly catalyses break down of oligo- or polysaccharides. On a bulk scale
the most important ones are those acting on cellulose. The endoglucanases act primarily
on the amorphous disordered parts of cellulose while the cellobiohydrolases (EC 3.2.1.9
1), (exoglucanases) hydrolyse the crystalline part. They catalyse both the splitting of
cellulose crystals and the hydrolysis to yield cellobiose. Their extremely fascinating
mode of action is now revealed by high-resolution X-ray crystallography (Divne et al.,
1994; Divne et al., 1998).
The glycosidases act by two different mechanisms which is revealed by the
stereochemistry at the anomeric centre of the product (McCarter and Withers, 1994). In
one type of glycosidases the anomeric centre is directly attacked by a hydroxide to give a
product with inverted stereochemistry at the anomeric centre. In the other mechanism, the
anomeric centre is attacked by the carboxylate group of a glutamic acid residue to form
an intermediate in which the carbohydrate moiety is covalently bound to the enzyme
similar to in epoxide hydrolases (Figure 2.16) and serine hydrolases. Attack on this
intermediate by a nucleophile leads to the net result which is retention of the
stereochemistry at the anomeric centre.
There is a potential for use of glycosidase hydrolysis if carbohydrates like cellobiose is
a desired product. On the other hand use of glycosidases in synthesis may take place
either by reverse hydrolysis or by transglycosylation. A major problem encountered
compared with use of lipases in similar types of reaction, is that carbohydrates are
primarily soluble in water. Reverse hydrolysis is the term used when a sugar unit is
directly glycosylated by a nucleophile and water is the side product. There are two ways
available to influence the equilibrium in order to increase the yield of glycoside, either by
pushing the equilibrium by adding more sugar or by pulling the equilibrium by removing
water. In transglycosylation a starting glycoside is used. The aglycon part may be F, Ph,
p-Hiitrophenyl or a glycosyl unit. Present in the reaction mixture will be the starting
glycoside, the product glycoside and water. Success of transglycosylation depends on i)
that transglycosylation is faster than glycoside hydrolysis and ii) that rate of hydrolysis of
the produced glycoside is slower than rate of hydrolysis of starting glycoside. Although
much work has been done in trying to optimize yields by influencing the equilibrium or
using hydrophilic solvents such as DMSO or DMF, neither of the methods have been
proven to be generally useful. It is more likely that enzymatic synthesis of carbohydrates
will take place by using glycosyl transferases (see 2.4.1).
 Applied biocatalysis 46

2.4 TRANSFERASES

2.4.1 Glycosyltransferases
In recent years there has been tremendous progress in the use of glycosyl transferases.
This is due to a combination of challenging biological problems, production of relevant
enzymes by molecular biological techniques and skilled organic chemists.
Glycosyltransferases are divided into two groups according to which activated donors
they use for transfer of monosaccharides. The Leloir glycosyltransferases utilize eight
nucleoside mono- or diphosphate sugars, UDP-Glc, UDP-GlcNAc, UDP-Gal, UDP-
GalNAc, GDP-Man, GDP-Fuc, UDP-GlcUA and CMP-NeuAc. The Non-Leloir
glycosyltransferases utilize glycosylphosphates as activated donors.
Carbohydrate-mediated cell adhesion is an important event which can be initiated by
tissue injury or infection and is involved in metastasis. One such adhesion process is the
interaction between the glycoprotein E-selectin and oligosaccharides on the surface of
neutrophils (white blood cells). The ligand that E-selectin recognizes is the
tetrasaccharide sialyl Lewis X (SLex). Since SLex competes with white blood cells for
binding to E-selectin, thus inhibiting the adhesion process, it may useful as an anti-
inflammatory and anticancer agent.
 Reactions catalysed by enzymes 47

Figure 2.19 Sialylation of N-acetyl lactose by cytidyl monophosphate-N-

acetylneuraminic acid using 2,3-neuraminic acid transferase as
catalyst (upper box). Regeneration of the sugar nucleotide is shown
in the lower box. CMP is converted into CTP in two steps using two
different kinases. In the final step CMP-N-acetylneuraminic acid is
synthesised from CTP and N-acetylneuraminic acid (sialic acid) using
the appropriate synthetase. The formed pyrophosphate is converted
into inorganic phosphate. Altogether five different enzymes are
involved in the process.

Non-enzymic synthesis of SLex involves a large number of protection and deprotection

steps which are not suited for large scale production. However, enzymic processes using
transferases have been developed with great success. The crucial factor in order to
succeed is regeneration of the activated monosaccharides. Synthesis of SLex and related
oligosaccharides is performed on a large scale (kilograms) by Cytel Corporation, San
Diego, (Defrees et al., 1995; Ichikawa et al., 1992). Sialylation (N-acetylneuraminic acid)
of N-acetyl lactose by CMP-N-acetylneuraminic acid and 2,3-neuraminic acid
transferase is shown in Figure 2.19. The formed trisaccharide is fucosylated in the Glc 3-
position in the next step. Regeneration of CMP-N-acetylneuraminic acid is shown in the
 Applied biocatalysis 48
lower box. As inferred it involves a series of enzymes. If all of the enzymes are available,
the process is extremely simple to perform.

2.5 LYASES

The lyases comprise enzyme class 4. They are enzymes cleaving C-C, C-O, C-N and
other bonds by elimination, not by hydrolysis or oxidation. Lyases also catalyse addition
to double bonds. The types of reactions catalysed by lyases are decarboxylation
(decarboxylase), hydration/dehydration (hydratase/dehydratase), ammonia
addition/deamination (ammonia-lyase), cyanohydrin formation/cleavage (oxynitrilase),
aldol condensation/cleavage (aldolase), , or , -elimination (PLP dependent lyase)
and Claisen-type condensation (synthase, lyase) (van der Werf et al., 1994).

2.5.1 Aldolases
In several recent applications of enzyme catalysis, the substrates on which the enzymes
act are not the kind of substrates that are “natural” to the enzyme. However, enzyme
catalysed synthesis of hexoses in the laboratory depends solely on enzymes acting on
natural or near natural substrates. The relevant enzymes are the aldolases (EC 4.1.2
aldehyde-lyases) since they catalyse an aldol type of C-C bond forming aldol addition
reaction. The aldolases most commonly join two C-3 units, called donor and acceptor,
and two new stereocentra are formed with great stereoselectivity.
Aldolases may be divided into two groups according to their mechanism of action and
occurrence in nature. Aldolases from animals and higher plants (group I) use an amino
group in the enzyme to form a Schiff’s base intermediate to activate the aldol donors.
Those from lower organisms, bacteria and fungi, group II, use a metal ion, usually Zn2+
in the active site to form an enolate intermediate. The two mechanisms, exemplified by
fructose-1,6-diphosphate aldolase, a very important aldolase in synthesis and breakdown
of sugars, are shown in Figure 2.20. The most frequently used FDP-aldolase for synthesis
of specific carbohydrates is rabbit muscle aldolase (RAMA). It belongs to a group of
aldolases for which dihydroxyacetone phosphate (DHAP) is the sole acceptable donor
(Group A in Figure 2.21). However, several aldehydes and ketones may function as
acceptors. The different aldolases are commonly divided into four groups requiring in
addition to DHAP as donor, pyruvate (B), acetaldehyde (C) and glycine (D).
Since the range of substrates of aldolases are rather limited they are primarily useful
for synthesis of carbohydrate like compounds. Their utility cannot be compared to the
wide range of aldol reactions known in organic chemical synthesis. Thus it may be
concluded that aldolases may be used for synthesis of

rare carbohydrates
isotopically labeled carbohydrates
 Reactions catalysed by enzymes 49
carbohydrates with unusual heteroatoms

2.5.2 Ammonia Lyase, Hydratase

The addition of water or ammonia to a carbon, carbon double bond is catalysed by
hydratases and ammonia-lyases respectively. While more than a hundred hydratases

Figure 2.20 The two mechanisms of aldolases. Group I enzymes from animals
and higher plants use an amino group in the enzyme to form a
Schiff’s base intermediate to activate the aldol donors. Group II
enzymes from lower organisms, use a metal ion, usually Zn2+ in the
active site to form an enolate intermediate. The two mechanisms are
examplified by fructose-1, 6-diphosphate aldolase, a very important
aldolase in synthesis and breakdown of sugars.
 Applied biocatalysis 50
have been described only less than ten ammonia lyases are known. These enzymes can be
used to transform steroids, fatty acids, terpenoids, hydroxy acids and amino acids.
Several of these reactions have been commercialized.
The addition of water to fumaric acid catalysed by fumarase is a highly stereospecific
reaction and malic acid is formed as the sole product (Figure 2.22, X=H). The ammonia
lyase 3-methylaspartase catalyses the similar addition of ammonia to yield L-aspartic
acid. When unnatural substrates are used in these reactions (X ≠ H), less success is
experienced. An increasing X-group gives slow reaction rates.

2.6 OXIDOREDUCTASES

Enzymes belonging to class 1, the oxidoreductases, are responsible for reductions and
oxidations in nature. As opposed to hydrolases their action depend on cofactors,

Figure 2.21 The four groups (A, B, C, D) of aldolases according to their donor
 Reactions catalysed by enzymes 51
dependence.

mostly NAD+/NADH or NADP+/NADPH. When one mole of substrate is reduced, one

mole of cofactor is oxidised and vice versa. The enzyme remains unchanged. Since
cofactors are extremely expensive chemicals, they have to be regenerated in order to
assure an economically feasible process. This is discussed in Chapter 10. The alternative
is to use whole cells for reduction-oxidation processes. Then the cells will take care of the
regeneration process. For instance if growing cells of Baker’s yeast is reducing a ketone,
it is the sugar in the medium that is oxidized. The drawback of using whole cells is often
complicated work-up process.
The first sub-class of the oxido reductases is 1.1, and it comprises the dehydrogenases
which act on primary or secondary alcohols or hemiacetals. They are mostly used for
reduction of ketones and aldehydes. Two other categories are oxygenases and oxidases.
The latter is not much used in biocatalysis.

2.6.1 Reductions
Enantioselective reduction of unsymmetrically substituted ketones by dehydrogenases
yields secondary alcohols (Figure 2.23). This reaction is important since it is an
 Applied biocatalysis 52

Figure 2.22 Reactions catalysed by two lyases, a hydratase (upper) and an

ammonia-lyase (lower).

Figure 2.23 Enantioselective reduction of unsymmetrically substituted ketones

by dehydrogenases yields secondary alcohols. The reaction may
either follow Prelog’s rule (addition of hydride from re-side) or they
may not (anti-Prelog).
 Reactions catalysed by enzymes 53
asymmetric synthesis capable of giving 100% of product. There are numerous examples
of such reactions using the commercially available horse liver alcohol dehydrogenase
(HLADH) or yeast alcohol dehydrogenase (YADH). Both of which follow the so-called
Prelog’s rule for stereoselectivity which states that the hydride is delivered from the
reside of the keto group. Most commonly this leads to predominance of the (S)-alcohol.
Using whole cells for reductions eliminates the need for regeneration of NADH. Most
widely used is Baker’s yeast (Saccharomyces cerevisiae) which also in most cases
follows Prelog’s rule (Csuk and Glänzer, 1991). Baker’s yeast may also reduce
selectively carbon-carbon double bonds in certain cases when the double bond is
activated.

2.6.2 Oxidations
Oxidation of secondary or primary alcohols by dehydrogenases is usually not performed
biocatalytically. The reaction destroys a stereocentre, it is thermodynamically not
favoured and product inhibition is a problem. It is attractive only in cases where it is
necessary to discern between several hydroxy groups in a molecule. Microbial oxidation
of D-glucitol to yield L-sorbose is the key step in production of vitamin C (Reichstein
and Grüssner, 1934).

2.6.3 Hydroxylation of Carbon Centres, Mono-oxygenases

In addition to the dehydrogenases, oxygenases and oxidases are discerned as two groups
of oxidoreductases. From a synthetic viewpoint the latter is not a particularly useful
group. Oxygenases on the other hand, are very much involved in important reactions. The
mono-oxygenases insert one of the two oxygens into the substrate while di-oxygenases
insert both.
The mono-oxygenases which catalyse a series of oxidations such as hydroxylation,
epoxidation, heteroatom oxidation and Baeyer-Villiger oxidation (Figure 2.24), depend
on NADH or NADPH and cofactors usually Fe or Cu. A particularly important reaction
is the direct incorporation of molecular oxygen into non-activated carbon centres, such as
in synthesis of important steroidal drugs by microbial 11 -hydroxylation of
progesterone and 7 -hydroxylation of lithiocholic acid (Figure 2.25).

2.6.4 Oxidation of Alkenes

Enzymic asymmetric epoxidation of alkenes may be performed by pure mono-
oxygenases. However, due to practical problems such as need of cofactors, microbial
oxidation with whole cells has been more widely used for the purpose. One great
disadvantage however, is the toxicity of epoxides towards living cells.
 Applied biocatalysis 54

2.6.5 Oxidation of Aromatic Compounds

Both epoxide forming (mono-oxygenases) and epoxide hydrolyzing (EH) enzymes (See
part 2.3.7) are involved in the breakdown of benzene compounds in eukaryotes

Figure 2.24 Reactions catalysed by mono-oxygenases, hydroxylation of carbon

centres, aromatic hydroxylation, epoxidation of alkenes, heteroatom
oxidation and Baeyer-Villiger oxidation of a ketone.
 Reactions catalysed by enzymes 55

Figure 2.25 Examples of direct incorporation of molecular oxygen into non-

activated carbon centres. Catalysed by mono-oxygenase in Rhizopus
arrhizus or Aspergillus niger progesterone and lithiocholic acid have
been hydroxylated in the 11 - and 7 -positions respectively.

Figure 2.26 Oxidation (mono-oxygenases) of benzene compounds in

eucaryotes leads to epoxide in the first step. The epoxide is converted
into a trans-1,2-diol by epoxide hydrolyzing (EH) enzymes in a
second step. Procaryotes follow a different pathway using di-
oxygenases to form a dioxetane which is reduced to a cis-1,2-diol.
 Applied biocatalysis 56
.by forming epoxide in the first step and converting the epoxide into a trans-1,2-diol in
the next (Figure 2.26). The prokaryotes follow a different pathway using dioxygenases
and reductases to form a cis-1,2-diol. These enantiopure starting materials may be subject
to Diels-Alder reactions or Michael-type additions by their diene-system. They have been
widely used for synthesis of a number of bioactive compounds such as cyclohexanoids
(pinitol, myo-inositol), cyclopentanoids (prostaglandins) and rare carbohydrates.

2.7 ISOMERASES

The isomerases comprise enzyme class 5 and they are enzymes catalyzing changes within
one molecule. Important subclasses are racemaces and epimerases (EC 5.1) and
intramolecular oxidoreductases (EC 5.3).
In the production of the sweetener high fructose corn syrup (HFCS), glucose isomerase
is used to convert glucose into fructose (equilibrium mixture appr. 50:50) since fructose
is nearly three times as sweet as glucose. The enzyme belongs to the subclass
intramolecular oxidoreductases interconverting aldoses and ketoses (EC 5.3.1). About 8
million tonnes pr. year of glucose is treated in this way. For further details, see Chapter 4.
An emerging technology is connected to the versatile food additive alginate. It is
primarily used as a thickener in sauces, for ice cream etc. Alginate which is a copolymer
of D-mannuronic acid and L-guluronic acid is produced by sea weeds. The two sugar
units alternate in different patterns depending on organism, parts of organism and growth
conditions for production of the polymer. The ability to form gels is connected to
complexation with calcium ions and it is the L-guluronic acid sequences that are
responsible for this. The biosynthesis of alginate proceeds

Figure 2.27 Alginate which is a copolymer of D-mannuronic acid and L-

guluronic acid is produced by sea weeds. The biosynthesis of alginate
 Reactions catalysed by enzymes 57
proceeds via C-5 epimerization of a poly-D-mannuronan precursor. The
important ability of alginate to form gels is connected to
complexation with calcium ions. The polyguluronate sequences keep
the alginate chains together in an “egg-box”-model.

via a C-5 epimerization of a poly-D-mannuronan precursor (Figure 2.27). Cloning and

expression of the gene for the mannuronan C-5 epimerase enzyme (EC 5.1.3) opens for
the possibility to design alginates with specific chemo-physical properties (Ertesvåg et
al., 1994).

2.7.1 Racemisation
As discussed in part 2.2.3 biocatalysis may be used both in asymmetric synthesis and
resolution in order to obtain enantiopure compounds. A major difference between
asymmetric synthesis and resolution is that the former in theory may give 100% yield of
the wanted enantiomer. Resolution on the other hand can only give 50% yield since the
starting point is a mixture of 50% of each enantiomer. This is the classical disadvantage
of resolution.
In many cases the unwanted enantiomer may be racemised in situ so that finally only
one enantiomer results. This method is known as dynamic resolution or second order
asymmetric transformation and may be enzyme catalysed or non-enzyme catalysed (see
Figure 2.6). Provided the rate constants for these reactions are of ideal magnitude,
dynamic resolution may give high yield and high ee of one single enantiomer. The result
may then be comparable to asymmetric synthesis. The kinetics of racemisation and
description of various racemisation methods have been the subject of a recent review
(Ebbers et al., 1997). Depending on the actual rate constants for a system the resolution
will be more or less efficient. As for a regular resolution the success of the reaction will
depend on the enantiometic ratio E. However, ee of the product does not depend on the
conversion. The reason is simply that the enzyme continuously meets a racemic substrate
as opposed to in an ordinary resolution where the concentration of the fast reacting
enantiomer is decreasing with conversion. Hence the enantiomeric excess of the product
will be constant throughout the reaction and only depend on E in the following way: ee p
=(E−1)/ (E+1). If the E-value for an enzyme/substrate pair for instance is 10 ee p will be
82% while E=100 gives ee p =98%.
Enzyme catalysed racemisation is an attractive method. The enzymes are known as
racemases and they often need cofactors like pyrodoxyl phosphate (PLP) or bivalent
metal ions to function properly. The substrates used in racemisation reactions have two
features in common, i) the stereocentre carries a proton, ii) adjacent to the stereocentre is
a carbonyl group or another function that make the proton at the stereocentre acidic. Most
racemases work on amino acids and -hydroxy acids. The principle of those needing
PLP is formation of a Schiff’s base between the aldehyde of PLP and the amino group of
the amino acid (like in Figure 2.6).
 Applied biocatalysis 58

2.8 QUESTIONS

2.1 • Ways to obtain enantiopure chiral building blocks?

2.2 • What is a meso compound?
• What is a prochiral compound?
2.3 • How does reversibility influence ee s and eep in resolution?
• Explain general differences between resolution and asymmetric synthesis?
2.4 • What is dynamic resolution?
2.5 • Enzymic ways to produce amino acids?
2.6 • Mechanism for the epoxide hydrolases MEH?
• Consequence for stereochemical course of hydrolysis?
2.7 • What are oxygenases? Discuss the three different types.
• Substrates and products for mono oxygenases?
2.8 • Examples of reactions catalysed by dioxygenases.
2.9 • Mechanism for aldolases?
• How are aldolases divided according to donor type?
2.10 • Biosynthesis of alginate
• Significance of structure for gelation properties?

2.9 HINTS

2.1 • See part 2.2.3

2.2 • See Figure 2.2 for explanation.
• Does a meso-compound have stereocentra?
• What about a prochiral compound?
2.3 • See Figure 2.12
• Discuss yield and possibility to influence the ee.
2.4 • See Figure 2.6
2.5 • See Figure 2.4

2.6 • See Figures 2.16, 2.17 and 2.18.

 Reactions catalysed by enzymes 59

2.7 • See Figure 2.24

2.8 • See Figure 2.26
2.9 • See Figure 2.20 and 2.21
2.10 • See Figure 2.27

2.10 REFERENCES AND SUGGESTED FURTHER READING

Anthonsen, H.W., Hoff, B.H. and Anthonsen, T. (1996) Calculation of enantiomer ratio
and equilibrium constants in biocatalytic ping-pong bi-bi resolutions. Tetrahedron:
Asymmetry , 7, 2633–2638.
Anthonsen, T. and Hoff, B. (1998) Resolution of derivatives of 1,2-propanediol with
lipase B from C antarctica. Effect of substrate structure, medium, water activity and
acyl donor on enantiomeric ratio. Chem. Phys. Lipids , 93, 199–207.
Anthonsen, T. and Jongejan, J. (1997) Solvent effect in lipase catalysed racemate
resolution. Meth. Enzymol. , 286, 473–495.
Archelas, A. and Furstoss, R. (1998) Epoxide hydrolases: new tools for the synthesis of
fine organic chemicals. TIBTECH , 16, 108–116.
Augé, C. and Grout, D. (1997) Chemoenzymatic synthesis of carbohydrates. Carbohydr.
Res. , 305, 307–312.
Chen, C.-S. and Sih, C.J. (1989) General Aspects and Optimization of Enantioselective
Biocatalysis in Organic Solvents: The Use of Lipases. Angew. Chem. Int. Ed. Engl. ,
28, 695–707.
Chen, C.-S., Fujimoto, Y., Girdaukas, G. and Sih, C.J. (1982) Quantitative analyses of
biochemical kinetic resolutions of enantiomers. J. Am. Chem. Soc. , 104 , 7294–7299.
Chen, S.-T., Huang, W.-H. and Wang, K.-T. (1994) Resolution of amino acids in a
mixture of 2-Methyl-2-propanol/water (19:1) catalysed by alcalase via in situ
racemisation of one antipode mediated by pyridoxal 5-phosphate. J. Org. Chem. , 59,
7580–7581.
Collins, A.N., Sheldrake, G.N. and J.Crosby, E. (1992) Chirality in Industry , John Wiley
and Sons.
Collins, A.N., Sheldrake, G.N. and J.Crosby, E. (1997) Chirality in Industry II , John
Wiley and Sons.
Csuk, R. and Glänzer, B. (1991) Baker’s yeast mediated transformations in organic
chemistry. Chem. Rev. , 91, 49–97.
DeFrees, S., Kosch, W., Way, W., Paulson, J., Sabesan, S., Halcomb, R., et al. (1995)
Ligand recognition by E-selectin: Synthesis, inhibitory activity and conformational
analysis of bivalent sialyl Lewis x analogs. J. Am. Chem. Soc. , 117, 66–79.
Desnuelle, P. (1972) in The Enzymes , edited by P.Boyer, pp. 575–616. New York and
London: Academic Press.
Divne, C., Ståhlberg.J., Reinikainen, T., Ruohonen, L., Pettersson, G., KnowlesJ., et al.
(1994) The three-dimensional crystal structure of the catalytic core of
cellobiohydrolase I from Trichoderma reesei. Science , 265, 524–528.
 Applied biocatalysis 60
Divne, C., Ståhlberg, J., Teeri, T. and Jones, T. (1998) High-resolution crystal structures
reveal how a cellulose chain is bound in the 50 Angström long tunnel of
cellobiohydrolase I from Trichoderma reesei. J. Mol. Biol. , 275, 309–325.
Drueckhammer, D., Hennen, W., Pederson, R., III, C.B., Gautheron, C., Krach, T., and
Wong, C.-H. (1991) Enzyme catalysis in synthetic carbohydrate chemistry. Synthesis ,
499–525.
Ebbers, E., Ariaans, G., Houbiers, J., Bruggink, A. and Zwanenburg, B. (1997)
Controlled racemisation of optically active organic compounds: Prospects for
asymmetric transformation. Tetrahedron , 53, 9417–9476.
Ertesvåg, H., Doseth, B., Larsen, B., Skjåk-Bræk, G. and Valla, S. (1994) Cloning and
expression of an Azotobacter vinelandii mannuronan C-5 epimerase gene. J. Bacteriol.
, 176, 2846–2853.
Faber, K., Ottolina, G. and Riva, S. (1993) Selectivity-Enhancement of Hydrolase
Reactions. Biocatalysis , 8, 91–132.
Fukuda, M. and Hindsgaul, O. (1994) Molecular Glycobiology , Oxford: IRL Press.
Hamilton, B., Hsiao, H.-Y, Swann, W., Anderson, D. and Delente, J. (1985) Manufacture
of L-amino acids with bioreactors. TIBTECH , 3, 64–68.
Hansen, T.V., Waagen, V., Partali, V., Anthonsen, H.W. and Anthonsen, T. (1995) Co-
solvent enhancement of enantioselectivity in lipase-catalysed hydrolysis of racemic
esters. A process for production of homochiral C-3 building blocks using lipase B from
Candida antarctica. Tetrahedron: Asymmetry , 6, 499–504.
Hanson, R.M. (1991) The synthetic methodology of nonracemic glycidol and related 2,3-
epoxy alcohols. Chem. Rev. , 91, 437–475.
Hoff, B.H., Anthonsen, H.W. and Anthonsen, T. (1996) The enantiomer ratio strongly
depends on the alkyl part of the acyl donor in transesterification with lipase B from
Candida antarctica. Tetrahedron: Asymmetry , 7, 3187–3192.
Ichikawa, Y., Lin, Y-C., Dumas, D., Shen, G.-J., Garcia-Junceda, E., Williams, M., et al.
(1992) Chemical-enzymatic synthesis and conformational analysis of sialyl Lewis x
and derivatives. J. Am. Chem. Soc. , 114, 9283–9298.
IUBMB (1992) Enzyme Nomenclature 1992 . Academic Press, Inc.
Janes, L. and Kazlauskas, R. (1997) Quick E. A fast spectrophotometric method to
measure the enantioselectivity of hydrolases. J. Org. Chem. , 62, 4560–4561.
Kanerva, L.T., Vihanto, J., Halme, M.H., Loponen, J.M. and Euranto, E.K. (1990)
Solvent effects in lipase-catalysed transesterification reactions. Acta Chem. Scand. , 44,
1032–1035.
Koskinen, A.M.P. and Klibanov, A.M. Eds. (1996) Enzymatic reactions in organic
media . Blackie Academic and Professional, London.
Kvittingen, L. (1994) Some aspects of biocatalysis in organic solvents. Tetrahedron , 50,
8253–8274.
Kvittingen, L., Sjursnes, B., Anthonsen, T. and Halling, P. (1992) Use of salt hydrates to
buffer optimal water level during lipase catalysed synthesis in organic media: A
practical procedure for organic chemists. Tetrahedron , 48 , 2793–2802.
Laane, C. (1987) Medium-Engineering for Bio-Organic Synthesis. Biocatalysis , 1, 17–
22.
Laane, C., Boeren, S., Vos, K. and Veeger, C. (1987) Rules for Optimization of
Biocatalysis in Organic Solvents. Biotechnol. Bioeng. , 30, 81–87.
Leuchtenberger, W., Karrenbauer, M. and Plöcker, U. (1984) Scale-up of an enzyme
 Reactions catalysed by enzymes 61
membrane reactor. Process for the manufacture of L-enantiomeric compounds. Ann. N.Y.
Acad. Sci. , 434, 78–86.
Lundhaug, K., Overbeeke, P., Jongejan, J. and Anthonsen, T. (1998) Organic co-solvents
restore the inherently high enantiomeric ratio of lipase B from Candida antarctica in
hydrolytic resolution by relieving the enatiospecific inhibition of product alcohol.
Tetrahedron: Asymmetry , 9, 2851–2856.
McCarter, J. and Withers, S. (1994) Mechanisms of enzymatic glycoside hydrolysis.
Curr. Opinion Structural Biol. , 4, 885–892.
Palmer, T. (1995) Understanding Enzymes , 4 ed. Ellis Horwood.
Quartey, E.G.K., Hustad, J.A., Faber, K. and Anthonsen, T. (1996) Selectivity
enhancement of PPL-catalysed resolution by enzyme-fractionation and medium-
engineering: Synthesis of both enantiomers of tetrahydropyran-2-methanol. Enzyme.
Microb. Technol. , 19, 361–366.
Rakels, J.L.L., Straathof, A.J.J. and Heijnen,J.J. (1993) A Simple Method to Determine
the Enantiomeric Ratio in Enantioselective Biocatalysis. Enzyme Microb. Technol. ,
15, 1051–1056.
Reichstein, T. and Grassner, A. (1934) Eine ergiebige synthese der 1-ascorbinsäure (C-
Vitamin). Helv. Chim. Acta , 17, 311–328.
Romesberg, F., Spiller, B., Schultz, P. and Stevens, R. (1998) Immunological origins of
binding and catalysis in a Diels-Alderase antibody. Science , 279, 1929–1933.
Rubin, B. (1994) Grease pit chemistry exposed. Structural Biology , 1, 568–572.
Sheldon, R.A. (1993) Chirotechnology . Marcel Dekker, Inc.
Toone, E., Simon, E., Bednarski, M. and Whitesides, G. (1989) Enzyme-catalyzed
synthesis of carbohydrates. Tetrahedron , 45, 5365–5422.
Uppenberg, J., Øhrner, N., Norin, M., Hult, K., Patkar, S., Waagen, V., et al. (1995)
Crystallographic and molecular modeling studies of lipase B from Candida antarctica
reveal a stereospecificity pocket for secondary alcohols. Biochemistry , 34, 16838–
16851.
van der Werf, M., van den Tweel, W., Kamphuis, J., Hartmans, S. and de Bont, J. (1994)
The potential of lyases for the industrial production of optically active compounds.
TIBTECH , 12, 95–103.
Wilson, E. (1998) Spinach enzymes break down explosives. Chem. & Eng. News , 76 ,
40.
Wong, C.-H., Halcomb, R., Ichikawa, Y. and Kajimoto, T. (1995a) Enzymes in organic
synthesis: Application to the problems of carbohydrate recognition (Part 1). Angew.
Chem. Int. Ed. Engl. , 34 , 412–432.
Wong, C.-H., Halcomb, R., Ichikawa, Y. and Kajimoto, T. (1995b) Enzymes in organic
synthesis: Application to the problems of carbohydrate recognition (Part 2). Angew.
Chem. Int. Ed. Engl. , 34 , 521–546.
Zaks, A. and Klibanov, A. (1988) Enzyme catalysis in nonaqueous solvents. J. Biol.
Chem. , 263 , 3194–3201.
 3.
ENZYMES AS PROCESSING AIDS AND
FINAL PRODUCTS
JOHANNES TRAMPER AND POUL B.POULSEN*
Wageningen Agricultural University, Department of Food Technology and
Nutritional Sciences, Food and Bioprocess Engineering Group, Wageningen,
The Netherlands.
Telephone: +31 317 483204; Fax: +31 317 48 2237; E-mail:
[email protected]
*Novo Nordisk A/S, Novo Allé, 2880 Bagsvaerd, Denmark.
Telephone: +45 4442 3417; Fax: +45 4498 0610; E-mail: [email protected]

ABSTRACT
Many applications of enzymes exist today in many, very different industries. In
this chapter, first a short description is given of the various industries where
enzymes are used as processing aids or processed into final products. Further in
this chapter examples from the detergent, feed, textile and food industry are
worked out in detail, highlighting technical, commercial and social aspects to
reckon with when developing and applying enzymes for these purposes.

3.1 INTRODUCTION

Looking at the various enzymes provided by leading suppliers, no real major

breakthroughs among the major industrial enzymes have occurred since the introduction
of immobilized glucose isomerase in the seventies. Nevertheless, technological
innovations are driving a rapid growth of the industrial-enzyme market in existing and
new business areas (Wrotnowski, 1997). The current market is valued at about $1.6
billion (Stroh, 1998). The US, the largest area of global industrial consumption, is
estimated at 24.7% of dollar usage, followed by the European Union at 23%. It appears
that usage depends on gross domestic product per capita. Enzyme consumption becomes
affordable and increases rapidly as developing countries approach industrialization.
Projected sales of industrial enzymes in the US are given in Table 3.1.
Enzymes, in varying formulations, are used today in many different industries. Short
 Enzymes as processing AIDS and final products 63
descriptions of most applications are given in the following section (3.2), while some
major applications are worked out as cases in the later sections. For further, more detailed
information the reader is referred to e.g. Godfrey and West (1996). Magazines, in
particular ‘the Gist’ (Gist-brocades) and ‘BioTimes’ (Novo Nordisk) from companies
producing and marketing enzymes also contain much useful information and have been
used extensively in writing this chapter, in addition to the booklet ‘Enzymes at
Work’ (Novo Nordisk, 1992). To structure this chapter the industries are divided in three
categories, depending on the application of the enzyme preparation:

Table 3.1 Projected sales of industrial enzymes by application ($ US millions) for the
US. Source: Amvir Associates (adapted from Wrotnowski, 1997).

Year: 1996 1997 1997 1 (%) 2000 2006 AAGR 3

Application:
Food 169 173 7202 (45) 186 214 2.4
Detergents 160 176 550 (34.4) 234 414 10.0
Textiles 26.6 27.1 175 (11) 28.8 32.4 2.0
Leather 10.5 10.8 45 (2.8) 11.5 12.8 2.0
Paper & Pulp 1.0 1.2 20 (1.2) 1.7 4.0 15.0
Other 3.0 3.3 90 (5.6) 4.4 7.8 10.0
Totals 370 391 1600 466 685 6.4
1 Worldwide sales (Stroh, 1998)
2 Includes starch processing (11%)
3 Average Annual Growth Rate (%)

(1) Enzymes as final products

Detergent Industry
Cleansing-Agent and Personal-Care Industries
Pharmaceutical Industry
Animal-Feed Industry
Analytical Applications
(2) Enzymes as processing aids
Textile Industry
Leather Industry
Pulp and Paper Industry
Sugar Industry
Coffee Industry
(3) Enzymes in food and beverage production
 Applied biocatalysis 64
Dairy Industry
Beer Industry
Wine and Juice Industries
Alcohol Industry
Protein and Meat Industries
Baking Industry
Fat and Oil Industry

The category ‘Enzymes as industrial catalysts’, including the starch-processing, the

antibiotics and the fine-chemicals industry, is covered in chapter 4.

3.2 SHORT DESCRIPTION OF ENZYME APPLICATIONS

3.2.1 Enzymes as Final Products

Industries marketing enzyme preparations as (part of) the product to the final customer
are described in this category.

Detergent Industry

Since the mid-60s, the use of enzymes in detergents has been the largest of all enzyme
applications. Over half of all detergents presently available contain enzymes, in particular
proteases, amylases, lipases and cellulases. Besides improved washing efficiency, the use
of enzymes allows lower temperatures and shorter wash periods (of agitation) to be
employed, often after a preliminary period of soaking. Further in this chapter (section 3.3)
the detergent enzymes are worked out in more detail.

Cleansing-Agent and Personal-Care Industry

Enzymes are included in cleaning preparations for the dispersion of solids and films in
and on pipe-work, heat exchangers, tanks, etc. Cleaning solutions generally contain an
enzyme and a detergent mixture. The enzyme mixtures mainly contain amylases,
cellulases, lipases and proteases. They are reported to act even in systems containing high
contents of undissolved solids, such as in sewerage systems. The formulations depend on
the application in mind. For instance, microbial proteases, or trypsin and papain, might be
used for fouled dairy filters, alpha-amylases and beta-glucanases in yeast and cereal uses,
and cellulases, pectinases and glucanases for wines and fruit juices.
Enzymes are also used in cleansing agents for household and sanitation applications,
for instance for the removal of bathroom mold. Lipases are used in special cleansing
applications, such as contact-lens cleaners and windshield-washing fluids. Glucose
 Enzymes as processing AIDS and final products 65
oxidase/glucoamylase are used in toothpaste. In the last few years, particularly in Europe,
the use of enzymes in automatic dish-washing detergents has taken off as well.

Pharmaceutical Industries

In view of the essential part played by enzymes in the digestive and metabolic processes
in all living organisms, it is surprising that the development of medical applications for
enzymes has not been very extensive. One simple application, however, has been in use
since the nineteenth century; a crude mixture of porcine pancreatic enzymes called
‘Pancreatin’ is given orally as a digestive aid to people deficient in digestive enzymes as
a result of genetic disorders, surgical removal of the gall bladder or advancing age. A
selection of other enzymes which have become important therapeutic agents is shown in
Table 3.2.

Animal-Feed Industry

Animal feed is largely composed of plant and vegetable materials. The number of
materials suitable as feed components is often limited by the ability of animals to digest
them. Many cereals have a high energy content, but a proportion of it may be locked up
in the form of non-starch polysaccharides that cannot be metabolized by certain animals.
The feed meal from oil-seed processing is an example of a valuable feed component with
an underutilized carbohydrate content. The use of supplementary enzymes (Table 3.3) in
feed can be thought of as an extension of the animals own enzyme system to make certain
feed components more digestible.

Table 3.2 Some important therapeutic enzymes (adapted from Chaplin and Bucke,
1990).

Enzyme Reaction Use

Asparaginase L-asparagine+H2O→L-aspartate+NH3 Leukemia
Collagenase Collagen hydrolysis Skin ulcers
Factor VII Plasminogen→plasmin Blood clots
Glutaminase L-glutamine+H2O→L-glutamate+NH3 Leukemia
Hyaluronidase Hyaluronate hydrolysis Heart attack
Lysozyme Bacterial cell-wall hydrolysis Antibiotic
Rhodanase S2O3 2− + CN−→ SO3 2−+SCN− Cyanide poisoning
Ribonuclease RNA hydrolysis Anti-viral
 Applied biocatalysis 66

Streptokinase Plasminogen→plasmin Blood clots

Trypsin Protein hydrolysis Inflammation
Uricase Urate+O2→allantoin Gout
Urokinase Plasminogen→plasmin Blood clots

Table 3.3 Applications of enzymes in animal feeds (adapted from Cowan, 1992).

Enzyme Type Target raw material Expected benefit of use

-Glucanase -Glucans of barley & oats Reductions of sticky droppings

Improved feed utilization
Xylanase Arabinoxylans of wheat and rye Improved litter quality
Improved feed utilization
Protease Vegetable proteins Increased protein digestibility
Reduced nitrogen excretion
-Amylase Cereal starch Increased proportion of cereals possible in
piglet feeds
-Galactosidase Oligosaccharides of soy and Improved energy availability and reduced
other legumes scours in piglet feeds

It is particularly useful to be able to do this in the case of young animals whose digestive
systems are not fully mature. Later in this chapter (section 3.4) the animal-feed enzymes
are worked out further.

Analytical Applications

Enzymes are extremely valuable analytical tools, for both medical and non-medical
applications. Because of their high specificity they can be used to assay the amount of a
substance, even of another enzyme, in a complex mixture such as blood, urine or other
biological fluids. Many of the diagnostic tests carried out by physicians depend on
enzymes. Enzymes are routinely used to measure the concentrations of glucose, urea,
amino acids, ethanol and lactic acid in biological fluids and to identify proteins and
nucleic acids. Home diagnostic procedures also often depend on enzymes. For example,
diabetics must monitor the glucose content of their urine as an indicator of their need for
insulin. The glucose analysis is simply done by dipping an analytical-test stick in a urine
sample. The stick contains the enzymes glucose oxidase and peroxidase as well as a
reagent that registers the reaction with glucose by changing color.
 Enzymes as processing AIDS and final products 67
In the last decades new methods based on electrochemical and chemiluminescent
techniques have been developed, which make enzymatic methods very fast and extremely
sensitive. Enzyme electrodes are now used for continuous control of fermentation
processes. The combination of enzymes with semiconductor technology has resulted in
the development of a new generation of analytical devices, the biosensors. Enzymes are
also used as label in immunochemistry and as DNA-probes to change radio-active labels
(H3, I125). Heterogeneous enzyme immuno assays (ELISA’s) are used for detection of
thyroid hormones and infectious (AIDS) and non-infectious diseases. Enzymes, such as
horse radish peroxidase and alkaline phosphatase, are used for labelling of antigens and
antibodies. Many different compounds can be detected by this technique. A new area of
the application of ELISA’s is the detection of pesticides. Chemiluminescent substrates for
peroxidases and phosphatases make this method very sensitive and accurate. For the
detection of drugs homogeneous methods of analysis have been developed on the basis of
regulation of enzyme activity by an antibody system.

3.2.2 Enzymes as Processing Aids

The industries discussed under this heading use enzyme preparations to facilitate
processing, to prevent technical problems and to reduce wastes.

Textile Industry

Enzymes are being used increasingly in textile processing, mainly in the finishing of
fabrics and garments. The most important applications of enzymes are desizing,
biopolishing, jeans finishing and bleach clean-up. Desizing is an enzyme application
dating back to the beginning of the 20th century, but the others have all been developed in
the last decade. The applications of enzymes in the textile industry are described in more
detail in section 3.5.

Leather Industry

One of the oldest applications of industrially made enzymes is in the processing of hides
and skins for leather. In order to make the leather pliable, it is necessary to subject the
raw material to an enzyme treatment before tanning. This is called bating, whereby
certain protein components are dissolved and can be washed away. The degree to which
bating is applied is dependent on the desired character of the finished leather. Glove
leather, for example, should be very soft and pliable and is subjected to a strong bating,
whereas leather for the soles of shoes is only lightly bated. Leather for upper parts of
shoes falls between these two extremes with regard to bating. Historically, dog or pigeon
dung was used as a bating agent. Besides being a difficult process to control with
unpredictable results, the dung did not exactly contribute to the creation of a pleasant
working environment. The dung bates owed their softening effect to the action of
 Applied biocatalysis 68
proteases, and it was heralded as a great step forward in 1908 when the German chemist
Otto Röhm patented the first standardized bate. This was based on pancreatic enzymes
extracted from slaughtered animals and turned out to be a great success. Today, both
bacterial proteases and trypsin (the traditional pancreatic protease) are used for bating.
When preparing hides and skins for liming and unhairing, proper soaking of the raw
stock is a prerequisite for obtaining a good-quality leather. For some raw materials,
notably dried stock, satisfactory rehydration may be a difficult and time-consuming
process. By degrading inter-fibrillary protein using proteolytic enzymes, water absorption
is significantly facilitated and the soaking operation can be shortened.
The conventional way to remove hair from cowhides is to use harsh chemicals, namely
slaked lime and sodium sulfide. These chemicals completely dissolve the hair and open
up the fiber structure. Enzyme-assisted unhairing, with or without recovery of the hair, is
closely related to the conventional process. However, just by adding an enzyme, it is
possible to reduce the requirements for sodium sulfide and lime. This process gives a
very clean pelt, a high area yield and results in fewer chemicals in the waste water.
Degreasing with lipases is beginning to be used as an alternative to tensides and
solvents. Lipases hydrolyze not just the fat on the outside of the hides and skins, but also
the fat inside the skin structure. The advantage of using lipases is that they do not
interfere with the structure other than by hydrolyzing the fat. The lipase-based process is
also more environmentally acceptable than solvent- or tenside-based processes.

Paper and Pulp Industry

Until recently, the use of enzymes in the paper and pulp industry was not considered
technically or economically feasible. Quite simply, suitable enzymes were not readily
available, except for the limited use of enzymes to modify starch for paper coatings.
However, research by scientific institutions and enzyme producers has led to the
development of new enzymes that offer significant benefits for the industry, particularly
from the environmental point of view. Two examples of applications, in addition to the
starch modification, made possible by new enzyme developments are given here: bleach
boosting and pitch control.
In the manufacture of paper, starch-based adhesives are used either to strengthen the
paper base or for coating the surface of the paper. Raw starch is unsuitable for either
purpose. To achieve sufficient adhesive power with raw starch would require the
application of a solution that was far too thick for practical use. Instead, chemically
modified starch, with a much lower viscosity in solution, is used. As an economical
alternative to modifying the starch with aggressive oxidizing agents, the starch can be
treated with enzymes ( -amylases) to obtain the same thinning effect
The dominant chemical pulping process is the kraft process, which gives a very dark
brown pulp. The pulp must be bleached before it can be processed into fine paper grades.
Chlorine and derivatives of chlorine are by far the cheapest and most versatile bleaching
agents for chemical pulps. However, they have the disadvantage that chlorinated organic
substances (some of which are toxic) are formed during bleaching. The paper and pulp
 Enzymes as processing AIDS and final products 69
industry is under growing pressure from authorities, consumers and environmental
groups to reduce the use and the discharge of chlorinated organic compounds. By treating
the kraft pulp enzymatically prior to bleaching, it is possible to obtain a very selective
partial hydrolysis of the hemicellulose, which forms a precipitate on the fibers after the
kraft cooking process. The enzyme has two indirect effects—firstly, it is possible to wash
out more lignin from the pulp and, secondly, the pulp becomes more susceptible to the
bleaching chemicals. The technique is called ‘bleach boosting’ and gives a significant
reduction in the need for chemicals in the subsequent bleaching stage.
Pitch problems are common in paper mills. Pitch is a catch-all term for the mixture of
lipids, resins and other extraneous compounds found in wood. These compounds are still
present in the mechanical pulps as they enter the paper machines. Globules of pitch tend
to stick to processing equipment. The sticky patches can result in holes in the paper so it
has to be recycled or downgraded in quality. In the worst cases, the paper web can
rupture causing costly production stoppages. A commercial lipase has proved its abilities
to reduce pitch deposits significantly on rollers and other equipment. Lipase breaks down
triglycerides in the wood resin in the pulp in much the same way as fungal and bacterial
growth reduces the pitch content of wood during conventional seasoning. However,
unlike seasoning, where wood is stored for a long time, the enzyme acts immediately and
does not reduce brightness or yield.

Sugar Industry

Starch is a natural component of sugar cane. When the cane is crushed, some of the starch
is transferred into the cane juice where it remains throughout the subsequent processing
steps. Part of the starch is degraded by natural enzymes already present in the cane juice,
but if the concentration of starch is too high, starch may be present in the crystallized
sugar (raw sugar). If this is to be further processed to refined sugar, starch concentrations
beyond a certain level are unacceptable because the filtration of the sugar solution will be
too difficult and the crystallization of sucrose becomes problematic as starch interferes
with crystallization properties of sucrose (elongated crystals). In order to speed up the
degradation of starch, it is general practice to add concentrated -amylase, preferable
thermostable, during the evaporation of the cane juice.
Another polysaccharide, dextran, is not a natural component of sugar cane, but is
sometimes formed in the sugar cane by bacterial growth. This happens in particular when
the cane is transported or stored under adverse conditions (high temperatures and high
humidity). Dextran has several effects on sugar processing: clarification of the raw juice
becomes less efficient, filtration becomes difficult, heating surfaces become ‘gummed
up’, which affects heat transfer, and finally, crystallization is impeded resulting in lower
sugar yields. These problems may be overcome by adding a dextran-splitting enzyme at a
suitable stage of the process. It should be added that dextran problems may also be
encountered in the processing of sugar beets, although the cause of the dextran is
different. In this case, dextran is usually a problem when the beets have been damaged by
frost. The cure, however, is the same: treatment with a dextranase.
 Applied biocatalysis 70

Coffee Industry

Pectinases and galacto-mannanases are the main enzymes used in the coffee industry.
Pectinases increase the pulp removal efficiency and galacto-mannanases act as viscosity
reducers in production of instant coffee.

3.2.3 Enzymes in Food and Beverage Production

Without a fundamental understanding of the actual mechanism, people have been using
biological systems for many centuries for the production of food and beverages.
Although the understanding has dramatically improved, the biotechnological conversion
of agricultural products into food and beverages is still a complex process, usually
involving many enzymes of which a large number are produced by genetically modified
micro-organisms (Table 3.4).

Dairy Industry

During the last century an increase in cheese consumption, and at the same time a
decrease in the number of calves for slaughtering, has caused a world-wide shortage of
calf rennet. The fermentation industry has tried to compensate for this shortage by
developing alternatives, based on microorganisms. These microbial rennets are quite
satisfactory for the production of a number of cheeses and they have been accepted in
most countries as alternative coagulants. A development of the last decade is the
introduction of calf chymosin—the enzyme responsible for the characteristic curdling of
milk—produced by recombinant microorganisms like yeasts.
To hasten and improve cheese ripening and taste development, special mixtures of
lipases and proteases can be used. Enzyme-modified cheese is also a concentrated source
of the major flavor components found in mature cheese. As an ingredient, it allows the
food technologist to either increase a product’s flavor intensity, without increasing the
cheese solid content, or to reduce cheese solid content for production of ‘healthier’ (low-
fat) products, whilst maintaining the original flavor intensity of the product. Produced by
accelerating the natural biochemical changes that occur in traditionally ripened cheese,
enzyme-modified cheeses are manufactured using individual or multiple lipase and
protease enzyme preparations from cheap cheese off-cuts. Blends of microbial lipases
with a strong affinity for producing low-carbon-chain-length fatty acids are particularly
suitable. These volatile acids confer strong cheese flavor to the cheese slurry and allow
concentrates of up to thirty times the flavor intensity of normal cheese to be produced.
Similarly, protein modification aids flavor development of enzyme-modified cheeses.
The neutral fungal proteinase from Biocatalysts Ltd., has a high amino- and carboxy-
exopeptidase activity and is,
 Enzymes as processing AIDS and final products 71

Table 3.4 Food enzymes produced by genetically modified microorganisms (Source:

Association of Manufacturers of Fermentation Enzyme Products, Bruxelles, 22
April 1996).

Principal enzyme activity Application

-acetolactacte decarboxylase beverages
-amylase baking
beverages
cereal and starch
catalase egg
milk
chymosin cheese
cyclodextrin-glucosyl transferase (glucanotransferase) cereal and starch

-glucanase beverages
cereal and starch
glucose isomerase cereal and starch
glucose oxidase bakery
beverages
egg
salads
hemicellulase bakery
lipase bakery
fats and oils
maltogenic amylase bakery
beverages
cereal and starch
microbial rennet cheese
protease bakery
beverages
cereal and starch
cheese
fish
meat
salads
pullulanase bakery
beverages
cereal and starch
 Applied biocatalysis 72

xylanase bakery
beverages
cereal and starch

as such, suited for the generation of strong savory flavor without bitterness. Enzyme-
modified cheese is typically included in processed cheese, cheese spreads, cheese dips,
cheese snacks, biscuits and cheese cake, where there is a desire to reduce overall cheese
solid content, typically in low-fat, health-orientated products (source: Biocatalyst Ltd.
brochure).
Another established application in the dairy industry is the hydrolysis of lactose in milk
and whey by lactases. Diminished digestibility problems, increased sweetness and
prevention of lactose-crystal formation are the results. The lactose hydrolysis is worked
out as a case later in this chapter (section 3.6).
Hydrogen peroxide is used as a sterilant or preservative for milk and whey; it destroys
harmful micro-organisms. Any excess hydrogen peroxide left in the milk or whey after
treatment can be decomposed to water and oxygen using an enzyme (catalase).

Beer Industry

In the traditional brewing process, malt both acts as a raw material (starch and protein
source) and as an enzyme source. Improved process economics and reliability may be
obtained by replacing part of the malt with industrial enzymes and unmalted grains such
as barley.
Low-calorie beers can be produced by enzymatic hydrolysis of the unfermentable
polysaccharides, which make up a considerable amount of the caloric content of beer.
The wort contains more fermentable sugars after this treatment and is diluted after the
fermentation to reach the normal alcohol concentration.
Another typical enzyme application in the production of beer is the use of proteases,
such as papain or laccases in chill-proofing, which is the prevention of haze formation
that can occur at low temperatures during or after the maturation of beer. Finally, to
reduce filtration problems, -glucanases are used.
The enzyme, -acetolactate decarboxylase (ALDC), has been developed and
commercialized in the beginning of the nineties. ALDC catalyzes the decarboxylation of
-acetolactate to acetoin during the primary fermentation, thereby reducing the
production of diacetyl and, consequently, eliminating or greatly reducing the need for a
maturation period.

Wine and Juice Industry

A general objective of fruit processing is the maceration of whole tissue, degrading cell
 Enzymes as processing AIDS and final products 73
structure to as full an extent as possible, to release the maximum quantity of cell contents
into the process fluid stream. Conventional use of pectinase enzymes has partly addressed
this objective. Many commercial preparations vary distinctly in the balance of main
activities of polygalacturonase, pectin esterase and pectin lyase, which are critical to
optimal maceration. Application of these pectinases in the production of fruit juices
offers several advantages: a higher pressing yield, reduction of viscosity and improved
filtration and clarification. Apart from pectinases, (hemi)cellulases, amylases, glucanases
and proteases are also commonly used as processing aids. Depending on the type of fruit
(apple, orange, peach, grape, etc.), the manufacturing process and the end-product
quality, the enzyme supplier provides different enzyme ‘cocktails’ for optimal results.
Detailed information can be found in Voragen and van den Broek (1991).

Alcohol Industries

Starch derived from maize, potatoes, barley, cassava or other sources must be pretreated
with hydrolytic enzymes (amylases, amyloglucosidase, proteases), which carry out
liquefaction, saccharification and protein hydrolysis, respectively, before it can be
fermented by yeasts and other microorganisms into potable or non-potable alcohol.
Enzymes can be added in the form of malt (germinated barley) or koji (germinated rice),
but this is expensive. Therefore, industrial enzymes have nearly totally replaced malt and
koji as enzyme sources, thereby not only improving the economics but also the
predictability of the process.

Protein and Meat Industries

Enzymes, in particular proteases, are widely used to increase the value and availability of
proteins. Applications in the food industry enable intrinsic functional properties of
proteins, such as viscosity, whipping ability and emulsifying power, to be optimized by
controlled proteolysis. Examples include enzymatic modifications of soy and whey
proteins to make them functionally more suitable for food applications, hydrolysis of
wheat gluten, production of yeast extracts, production of gelatin from collagen and
preparation of peptones, which are hydrolyzed proteins used in microbiological growth
media. Proteases also are used advantageously in the recovery of protein from fish-
processing wastes, from the blood, offal and bones from slaughter-houses and in the
decolorization of hemoglobin to make this abundant source of protein visually more
acceptable. An example of an enzyme with high potential for these purposes is
transglutaminase (see section 3.7).
The tenderization of meat by natural maturation takes place in about 10 days at 2°C.
This slow maturation, brought about by the conversion of muscle into meat by
endogenous enzymes (neutral protease and collagenase), results in tender meat but has
the disadvantage of moisture loss and shrinkage of the tissues. Tenderization process
improvements using exogenous enzymes have been examined since 1940. On a
 Applied biocatalysis 74
commercial scale, plant proteases, such as papain (from papaya) and bromelain (from
pineapple), are used in the tenderization of meat. These proteases are capable of digesting
connective tissue and muscle protein. A practical problem is how to achieve an even
distribution of the enzymes in the tissue. If preparations are sprinkled on the surface of
the meat, the interior remains tough. Repeated injection under pressure is an alternative.

Baking Industry

The raw material of bread is flour, a mixture of starch, protein (gluten), lipids, glucan and
some wheat enzymes. Upon mixing flour, yeast and water, many complex biochemical
and biophysical reactions take place in the dough phase, catalyzed by the wheat enzymes
and by the CO2-producing yeast. This continues in the baking phase, finally resulting in
bread. The addition of extra enzymes to the dough enables the baker improved control of
the baking process, the advantages being:

– introduction of different baking processes

– shorter process time
– slowing-down of staling
– independence of flour variability
– replacement of unnatural processing acids, i.e. chemicals, with natural enzymes

The application of enzymes has a long history in the flour, milling and baking industries
(Haarasilta and Pullinen, 1992). Cereal-based enzymes have been used for decades in the
form of different malt products, such as malt flour, malt extract and malt-based
improvers. The use of fungal -amylases became widespread during the 1960’s. -
Amylases produce dextrins, which are further broken down to sugars by naturally
occurring -amylases, thus improving yeast fermentation, bread volume and crust color.
-Amylases (now mainly from bacterial sources) are best, but also -amylases reduce
rate of staling, thereby prolonging the freshness of baked products.
Proteases from various sources are commonly used to optimize the rheological
properties of bread doughs made from strong wheat or, in biscuit and wafer production,
to reduce gluten elasticity. Hemicellulases (or cellulases and pentosanases) have been
used to improve the baking properties of stiff rye flours. Hemicellulases can also be used
to improve dough properties and bread quality in the production of wheat breads.
References also exist describing the role of oxidoreductases (lipoxygenase, glucose
oxidase, catalase, peroxidase, ascorbic-acid oxidase, sulphydryl oxidase, glutathione
dehydrogenase) and lipase. Combined addition of oxidative and hemicellulolytic
enzymes improves dough and bread properties and is a good alternative to the chemical
additives, such as potassium bromate, whose use is being increasingly restricted by
legislation (Haarasilta and Pullinen, 1992).
 Enzymes as processing AIDS and final products 75
Fat and Oil Industries

Natural fats and oils can be used directly in products, either individually or as mixtures.
In many cases, however, it is necessary to modify their properties, particularly their
melting characteristics, to make them suitable for particular applications. Therefore, the
oils and fats industry has developed several modification processes using enzyme
technology. In particular, lipases (and lately cutinases), phospholipases and pectinases
can be used for interesterification processes, ester syntheses and in olive-oil extraction.
The main current potential application of lipase-catalyzed fat-modification processes is
in the production of valuable confectionery fats; for instance, alternative methods of
obtaining cocoa-butter equivalents by converting cheap palm-oil fats and stearic acid to
cocoa-butter-like fats. The reaction is executed in a water-poor medium, such as hexane,
to prevent hydrolysis. At least one commercial application exists. Loders Croklaan
(Unilever) has an enzymatic interesterification plant in Wormerveer, the Netherlands.
Many other new potential applications of lipases have been proposed of which some will
certainly be economically feasible. Examples and details can be found in chapter 9 of this
book.
Pancreatin, a mixture of enzymes from the pig pancreas, can improve the emulsifying
properties of lecithin. By the action of the enzyme phospholipase A-2, lecithin is
converted to lyso-lecithin, which has better emulsifying properties. An example is the
improvement of the stability of mayonnaise, margarines and cosmetics. In general
enzymes made by (genetically modified) microorganisms are largely replacing those
from animal sources.
As in fruit processing, pectolytic and cellulolytic blends are used to enhance the
maceration of olive-oil pulp, thereby improving olive-oil extraction yield. These enzyme
blends can also be used for avocado-oil extraction.

3.3 DETERGENT INDUSTRY

3.3.1 Introductory intermezzo: Whiter than white

The thought ‘the hotter, the better’ is not valid anymore for washing of textiles. Modern
detergents have been improved with enzymes which degrade dirt most efficiently at
temperatures of 30 to 40°C, thus saving energy and environment. Even before the first
world war enzymes were added to detergents. These enzymes came from the pancreas of
cows and had a limited stability in detergent products. Up to the sixties these enzymes
were used in pre-soaking of laundry. Then the detergent manufacturers changed to
bacterial enzymes with the result that the recommended washing temperatures
permanently decreased.
In the last decade use of modern biotechnology has caused a revolution in the detergent
industry. Newly developed enzymes with higher activities and stabilities have become
available. In addition to improving the properties of enzymes, biotechnologists are also
 Applied biocatalysis 76
continuously in search of novel enzymes. The genetic information of a novel, suitable
enzyme occurring in a troublesome organism is transferred to bacterial or fungal cells
which are easy to cultivate, thus making available large quantities of the desired
enzymes. In case of detergents, enzymes only reach the consumer in relatively small
quantities, and are not literally consumed. As a result detergent enzymes were quickly,
that is to say in the beginning of the nineties, produced by means of genetically modified
micro-organisms without giving much publicity to it. The goals aimed at, more stable
enzymes working at temperatures which reduce the consumption of energy at an
economic price, both benefited producer and consumer.

3.3.2 Introduction (Thole and Velink, 1996)

Starch, fats and proteins cause ugly stains in textile. However, under the influence of the
right enzymes these disappear as snow for the sun. Already for decades detergent
manufacturers tried to further refine the concept of enzymatic stain removal. Currently
enzyme engineers and biotechnologists work together to develop more ideal detergent
enzymes.
The first patent on the use enzymes in detergents originates from 1913. In that year
Otto Röhm patented his idea to add trypsin to the laundry. He combined the well-known
properties of the tryptic enzymes, i.e. degradation of fat and protein residues, with the
assumption that many stains are caused by fat and protein residues. This concept
appeared immediately on the market under the name of Burnus and was sold throughout
Europe the then coming 50 years. During that time there were few new developments. In
1945 the Swiss company Gebr. Schnyder marketed the product named BIO-38, which
contained a purified pancreatic trypsin. A real improvement followed in 1957 when BIO-
40 was introduced; instead of the pancreatic enzymes, this product contained a bacterial
protease. Nevertheless, both types were still most active at neutral pH, while a high
activity was desired at a high pH as most detergents in that time consisted to a large
extent of soda. The solution for this problem came from Novo Industri in Denmark. This
company developed a microbial protease, Alcalase™, with a high activity and stability at
pH 8–10; this product was marketed in 1958. A decade later more than 80% of the
detergents in Germany contained enzymes. However, in the beginning of the seventies
the use lowered. A public debate started as result of reports saying that enzymes can
cause allergic reactions. It appeared that this only occurred with people working in the
production process. Adaptation of the formulation process for making the enzymes
solved this problem (see following intermezzo).

3.3.3 Intermezzo: Enzyme Formulations and Safety Precautions

Industrial enzymes must be tested and formulated for the maximum degree of safety in
handling. As enzymes are proteins they may cause allergic reactions in susceptible
people. For instance, early attempts to use proteases in detergents foundered because
 Enzymes as processing AIDS and final products 77
workers in detergent plants developed hypersensitivity. This was combatted by
developing dust-free enzyme granulates, in which the enzyme is incorporated into an
inner core, containing inorganic salts and sugars as preservative, and bound with
reinforcing fibers of carboxymethyl cellulose or a similar protective colloid (Figure 3.1).
This core is coated with inert waxy materials made from paraffin oil or polyethylene
glycol plus various hydrophilic binders, dispersed in the wax. This combination of
materials both prevents enzyme dust formation and protects the enzymes against damage
by other detergent components during storage.
Until the end of the sixties enzyme products such as the detergent proteases were just
powder products. Today very few powdered-enzyme products remain. All detergent
enzyme products from the larger enzyme suppliers are either liquid formulations or
granulated and further protected by coatings. Today formulation techniques really have
become a science with MAC-values in production facilities of 10–100 nanogram/m3 air.
It is further recommended that the use of such safe enzyme products shall be planned
such that the liquid enzyme product is not spilled and allowed to dry and aerosol
formation shall be prevented. With these simple rules in mind, industrial enzymes are
very safe.

Figure 3.1 Cross-section of dust-free enzyme granule (adapted from Riisgaard,

1990).

3.3.4 Recent developments (Thole and Velink, 1996)

In the mid eighties a totally new and revolutionary enzyme product was marketed by
Novo Nordisk (at that time Novo Industri); this time a cellulase named as Celluzyme™.
This enzyme possesses not only the capacity to remove stains, it also has laundry-
 Applied biocatalysis 78
softening properties (Figure 3.2). Another breakthrough followed a couple of years later
in 1988 when again Novo launched a new enzyme, i.e. their lipase Lipolase™. This was
the first industrial enzyme product produced by means of recombinant-DNA technology.
Today detergent enzymes are classified in four categories:

1. proteases for removal of protein residues,

2. amylases for starch degradation,
3. lipases for fat hydrolysis, and
4. cellulases for laundry-softening and color-regaining properties.

For use in a detergent an enzyme must suffice the following criteria. In the first place it
must have an adequate activity and stability at alkaline pH. It must also remain active and
stable in a broad temperature interval (e.g. 10–60°C). For domestic use the trend is to low
temperatures (see 3.3.7 In-depth intermezzo). Additionally an enzyme should be resistant
to hydrolysis and oxidation by other substances in the detergent, e.g. proteases. For
enzymes with laundry-softening properties two additional criteria apply, i.e. the enzyme
must show a high fiber separation and at the same time not impair the fibers.
Developments aim to further satisfy these criteria.
 Enzymes as processing AIDS and final products 79

Figure 3.2 How Celluzyme™ acts on the fabric surface (BioTimes 3, Sept.
1996; reprinted with permission).
 Applied biocatalysis 80

3.3.5 Intermezzo: Old clothes look like new (BioTimes, September 1996 and
March 1998)
Washing and wear soon cause colored cotton fabrics to lose their brightness and become
fuzzy and look old. The illustrations show why (Figure 3.2). As the garment is used,
fibers on the fabric surface begin to break and form split ends, in a similar way to hair.
These protruding micro-fibrils become entangled forming small pills that easily trap
particles of dirt. As a result, the fabric starts to look faded and worn, even though it may
be quite new. Celluzyme is an enzyme complex that removes protruding micro-fibrils and
small pills (see also 3.5.1) on the surface, as well as the dirt trapped inside them. With the
removal of the micro-fibrils, the fuzzy appearance of the surface disappears.
Celluzyme has a number of positive effects, but the color clarification effect is of
particular interest. The perceived color clarity of a piece of new cotton declines after only
a few washes. However, a remarkable rejuvenation occurs after about 10 washes with a
detergent containing Celluzyme. This is when the fibers have become sufficiently frayed
for the enzymes to gain access to the micro-fibrils. The wetting and drying associated
with washing opens up the microstructure of the fibers. Further washes in Celluzyme
restore the color of the cotton almost to its original appearance. In the end, the cotton
looks almost as good as new. Without Celluzyme present, the sharpness of the color
continues to decline after every wash. However, Celluzyme can also progressively restore
the color of a worn fabric. The aging process is put into reverse and the fabric starts to
look better and better. The color clarification effect is a unique feature of Celluzyme. It is
worth noting that unlike proteases or amylases that give an immediate effect after one
wash cycle, the color clarification effect of Celluzyme is cumulative. Repeated washing
with Celluzyme gives the greatest benefits.
Now an improved cellulase product called Carezyme™ is also available. Carezyme is a
mono-component enzyme preparation, i.e. it contains a single cellulase activity, whereas
Celluzyme is a multi-component enzyme preparation. Carezyme has been chosen on the
basis of a high activity on micro-fibrils but not on the intact fiber. In more scientific
terms, Carezyme has a high activity on amorphous cellulose, which is the major
constituent of the micro-fibrils, and a low activity on crystalline cellulose, which is the
major constituent of the intact fiber. This means that effective color care can be obtained
without significant fiber damage or loss of strength. In addition, doses needed for color
care are too low to give a significant effect on the intact fibers. The exact dose required
varies from detergent to detergent depending on its formulation. Carezyme works well at
the relevant pH and temperatures for washing household laundry.
 Enzymes as processing AIDS and final products 81
3.3.6 Mixing Enzymes Creates Improved Performance (Chemisch Weekblad
31/32, August 1996; BioTimes, March 1996)
The development of mono-component enzyme preparation using gene technology,
introduces the possibility to produce optimal blends for specific applications. Humicola
EG5 cellulase, for instance, is known to have significant anti-pilling and depilling activity
which is useful in removing the gray fuzz from faded colored cellulose-containing fabrics
such as cotton but may be less useful in anti-greying effect. Other types of cellulases, for
example bacterially derived cellulases including Bacillus sp. have a more pronounced
anti-greying effect with lesser depilling and anti-pilling. Accordingly, it would be useful
to combine a Bacillus sp. derived cellulase with a Humicola EG5 cellulase in the same
mixture to take advantage of their relative strengths. The advantage of this would be
particularly important given a cellulase which, in addition to its valuable characteristics,
also had a serious drawback such as strength loss of treated fabric. Thus, an aggressive
cellulase can be used in a lesser quantity to achieve the beneficial effect of that cellulase,
e.g. anti-pilling and depilling, and supplemented with a less aggressive cellulase to
achieve an incremental benefit in terms of another effect, e.g. anti-greying, with the end
result being less total strength loss with full realization of cellulase benefits.
Appropriate combination of proteases and amylases can result even in synergy effects
in removing food stains. Stains of cocoa, potatoes, pasta sauce, fruit, gravy, etc., are
mainly a combination of protein and starch. The likely explanation for the synergy is that
the amylase breaks down starch on the surface of the stain allowing the protease to attack
the protein in a more effective way.

3.3.7 In-depth Intermezzo: Towards Even Lower Wash Temperatures (van

Ee, 1994)
In response to both environmental and economic pressures wash temperatures have gone
down steadily. They are now reaching levels that hamper the effective performance of the
proteases in detergents. At low temperatures, the interaction between protease and protein
stain becomes less efficient. This interaction takes place at the active site of the enzyme
(Figure 3.3). Specific subsites S1, S2, S3, S4 towards the N terminus, and S1’, S2’, S3’,
S4’ towards the C terminus of the protease maintain electrostatic interactions with the
corresponding sites (P1-P4 and P1’-P4’) in the substrate. Apparently, at very low
temperatures, the stronger interactions, i.e. between S1 and P1 and S4 and P4, are out of
balance, resulting in lower performance. Gist-brocades has investigated the effects of
amino-acid substitutions at these specific locations of their commercial enzyme, using
site-directed mutagenesis, aiming at the development of superior enzymes for washing at
lower temperatures. The results obtained with one mutant enzyme are distinctly better
than the ‘natural’ protease, i.e. less is needed of the new enzyme to get the same
perceived wash performance especially at low temperatures. Apparently, amino-acid
substitutions within or between binding pockets allow for the generation of new proteases
specifically suited for cold water washing at low enzyme dosages.
 Applied biocatalysis 82

3.4 ENZYMES AS FEED ADDITIVES

3.4.1 Introduction (van Dijck and Geerse, 1993)

Commercial compound feeds for livestock are formulated to provide starch, protein and
lipid components which are digested with high efficiency. In poultry diets, the

Figure 3.3 Schematic representation of the binding subsite of Maxacal®,

presently Purafect (Genencor), showing some key amino acids which
play an important role in substrate recognition. The arrow shows the
position of the scissile bond (the Gist, 57 (1994), 9; with permission
of Gist-brocades and Genencor; the latter has bought this enzyme
from Gb).

major source of energy is cereals, usually maize. If, however, for economic reasons,
instead of maize, other cereal grains (such as barley, rye, or wheat) are used, productive
values are much lower than would be predicted from their nutrient content. Furthermore,
feeding such grains is often associated with sanitary problems, sticky droppings and poor
litter quality. The magnitude of the problem depends on the type of cereal, the specific
cultivar and on climatic conditions, in particular at the ripening phase of the grain.
Physiological research has linked the nutritional problems to the presence of soluble
 Enzymes as processing AIDS and final products 83
non-starch polysaccharides (NSP) in these cereals. Cereals contain, besides starch and
protein, varying amounts of NSP derived from the plant cell wall matrices. It is mainly
composed of cellulose, -glucans, and pentosans (arabinoxylans). Both the soluble -
glucans as well as the pentosans form viscous solutions in water and from several feeding
studies this was pin-pointed as a limiting nutritional factor in non-maize cereal-based diet
formulations.

3.4.2 Upgrading Barley (Stobart, 1993)

Barley is generally not included in poultry and pig rations due to its low energy content.
This results primarily from the high dietary fiber content of barley. More specifically, one
particular fiber component, the mixed-linked -glucan, has the major impact. This
particular carbohydrate is the primary component of the barley endosperm cell walls,
which surround and protect the enclosed nutrients such as starch and protein. Further,
mixed-linked -glucan can cause viscous conditions in the intestinal lumen, thereby
reducing feed intake and nutrient digestion and absorption. This can in turn lead to wet
and sticky excreta and digestive disorders.
Research with poultry established that the addition of microbial -glucanases could at
least partially alleviate these problems. The first widespread application of this
knowledge was in Finland in 1984. Since then, it has been established that -glucanase-
based multi-enzyme products are more consistent and cost effective. Such products are
now utilized by the animal feed industry in many countries at times when barley is
attractively priced relative to other grains.
The efficacy of such products in poultry is amply demonstrated by data collected from
many studies. In addition to the potential reduction in feeding costs, when including
cheap barley supplemented with enzymes, broiler producers have also reported savings in
litter costs and fewer carcass downgrades due to hock and breast burns. Similar benefits
have been reported for other types of poultry, including laying hens and turkeys.
This success with poultry quickly stimulated interest in applying the same principals to
pig production. Although initial trials were disappointing, further work led to the
development of -glucanase-based multi-enzyme products that proved effective in
barley-based pig diets. As for poultry, barley-based diets supplemented with enzymes can
give the same piglet performance as wheat, indicating an improvement in the digestible
energy of about 8%.
Barley is often considered to be somewhat prophylactic with regards to digestive
disorder in piglets. This seems to be accentuated by including enzymes. This is
presumably due to the enzymes stimulating digestion in the upper sections of the
intestinal tract, which in turn reduces readily available substrate for bacterial proliferation
in the lower tract.
 Applied biocatalysis 84
3.4.3 Upgrading Wheat (Stobart, 1993)
Initial attempts to apply the enzymes used for upgrading barley in wheat-based diets were
encouraging, although generally not economically feasible. Further research established
that pentosanase-based products were more cost effective. This quickly led to the
development of pentosanase-based multi-enzyme products. As for barley, commercial
application indicated that such products could significantly improve the nutritive value of
wheat. Chicken production data demonstrated that addition of the enzyme improved live-
weight by 3% and feed conversion by 6%. The conclusion from this is that wheat
supplemented with enzymes can have a metabolizable energy similar to that of maize. In
addition to elevating the productive value, enzyme supplementation of wheat-based
poultry diets can reduce litter problems, leading to fewer carcass blemishes and dirty
eggs. Other trials and commercial experiences suggest that the addition of enzymes can
maintain performance in diets low in protein and essential amino acids.
The application of pentosanase-based enzyme products to wheat-based piglet diets is
also cost-effective, with improvements in live-weight gain and feed conversion of about
5% commonly reported. Further, lower incidences of digestive disorders

Figure 3.4 The reaction catalyzed by phytase. Free inorganic phosphate is

liberated from the substrate phytic acid (Myo-inositol-hexaphophate;
adapted from Walsh et al., 1993).

and less between animal variation in live-weight have also been observed on such diets.

3.4.4 Milk Replacers

It is common practice in Western Europe and North America to separate the domestic-
animal ‘production mother’ from her young as soon as possible. The earlier a piglet or
calf can be weaned off its mothers milk, the better from the breeders point of view
because there is a greater chance of the mother becoming pregnant again. The suckling of
the young acts as a crude form of birth control. Feeds based on cereals and vegetable
proteins are not normally fed to piglets as their own digestive system is adapted to milk.
 Enzymes as processing AIDS and final products 85
But the addition of industrial enzymes allows these proteins to be digested so that the pig
producer can use a more economical blend. In some cases, it has been possible to
completely replace the skimmed milk normally used in early grower feeds. Furthermore,
there is no loss of performance.

3.4.5 Enzymes and the Environmental Impact of Intensive Agriculture

(Walsh, Power & Headon, 1993; and references cited therein)
Supplementation of the diet with selected enzyme activities may promote a decrease in
the overall pollutive effect of animal excreta. This is particularly true in the case of
dietary phosphorus, a large proportion of which remains unmetabolized by mono-
gastrics.
In the region of 60–65% of the phosphorous present in cereal grains exists as phytic
acid (myo-inositol- hexaphosphate) which, accordingly, represents the major storage form
of phosphate in plants. However, in this form, the phosphate remains largely unavailable
to mono-gastrics as these species are devoid of sufficient, suitable, endogenous
phosphatase activity that is capable of liberating the phosphate groups from the phytate
core structure (Figure 3.4). The animals’ inability to degrade phytic acid has a number of
important nutritional and environmental consequences. Phytic acid is considered anti-
nutritional in that it chemically complexes a number of important minerals such as iron
and zinc, preventing their assimilation by the animal. The lack of available phosphorus
also forces feed compounders to include a source of inorganic phosphate (such as
dicalcium phosphate) in the feed, with the result that a large proportion of total phosphate
is excreted. It has been estimated that in the USA alone, 100 million tons of animal
manure is produced annually, representing the liberation of somewhere in the region of 1
million tons of phosphorus into the environment each year. The potential pollutive effect
of this in areas of intensive pig production is obvious. Many countries are enacting tough,
new anti-pollution laws in an attempt to combat the adverse effect of animal waste on the
environment.
Several microbial species (in particular fungi) produce phytases (EC 3.1.3.8). The
incorporation of suitable, microbially derived phytases in the diet can confer the ability to
digest phytic acid on the recipient animals. This would have a threefold beneficial effect:
the anti-nutritional properties of phytic acid would be destroyed; a lesser requirement for
feed supplementation with inorganic phosphorous would exist; and reduced phosphate
levels would be present in the faeces. Several trials have confirmed that the inclusion of
phytase in animal feed promotes at least some of these effects. However, the enzyme is
not yet used in many countries. This may be explained, in part, by the fact that most
microbial species only produce low levels of phytase activity which, obviously, has an
effect on the cost of the finished product. It seems likely that widespread utilization of
phytase within the industry will only be made possible by the production of this enzyme
from recombinant sources, and at least two major enzyme companies are marketing such
an enzyme for a number of years now.
Within foreseeable time plants will also produce enzymes on a commercial scale. The
 Applied biocatalysis 86
Dutch company Plantzyme succeeded in producing the feed enzyme phytase in rape seed.
Using modern biotechnology the phytase gene has been introduced into the genetic
material of rape seed. In the rape-seed plant the enzyme is produced in the seeds. The
advantage of that is that it can be easily harvested, namely in the same manner as regular
rape seed. Besides the enzyme is packed in the seed, such that it can be stored for several
years. The seeds can be added directly to the feed, without having to isolate the enzyme
first. Feed for chickens and pigs mainly consists of wheat and barley and it already also
contains a small amount of rape seed. The research is in the final phase before the
product goes into the admittance and registration procedure.

3.4.6 Phytase Stabilization (the Gist, 1/72 (1998) 10–11)

As said above, the enzyme phytase, produced by recombinant micro-organism, has been
used in a small number of countries for a couple of years now. However there was one
drawback. Phytases are not very stable at the higher operating and processing
temperatures applied by certain compound feed manufacturers. The enzyme is actually
inactivated by a combination of water and higher temperatures. Many compound feed
manufacturers pellet their feed mixes and each has a different preferred operating
temperature dependent on the type of feed and pelleting process. The higher temperatures
(typically above 80°C) offer the benefit of low dusting pellets and reduced bacterial
activity (especially Salmonella). Some countries even insist on higher pelleting
temperatures to minimize bacterial activity in the feed. But the activity of the phytase
added suffers and feed conversion figures likewise.
Gist-brocades recently developed a more temperature-stable, low-dusting, high-
performance phytase additive for compound feeds. To reduce the effect of water and
temperature a new type of starch carrier is introduced that better stabilizes the active
ingredient and that has a larger particle size, actually a granulate, thus reducing the effect
of water and higher temperature on the product during feed processing. Furthermore, the
storage life is extended as well. In hotter countries the effect will be even more
pronounced than in cooler northern climates as the net benefit of good stability during
storage of the feed will be felt even more.

3.4.7 Conclusions (Stobart, 1993)

The addition of microbial enzymes to animal feed, almost unheard of 15 years ago, is
now common in many countries. These enzymes have the advantages that they are
entirely natural, leave no residues, and work only on feed components. Enzymes increase
the degree to which feed components are broken down, giving the animal producer two
options when using feed enzymes. The first is to add the enzymes to normal feeds to give
a cost-effective improvement in production. The second is to use enzymes to upgrade
lower digestibility feedstuffs, allowing the inclusion of lower-cost raw materials in feeds
without affecting performance.
 Enzymes as processing AIDS and final products 87
There is also evidence now that feed enzymes can play an important role in reducing
pollution from agriculture. By improving nutrient digestion and retention, enzymes,
including phytase can reduce total manure output and the content of nitrogen and
phosphorous in manure. Thus feed enzymes can help agriculture respond to the
increasing environmental demands of the general public.
As commercial interest grows in this area and more cost-effective microbial enzymes
become available, it is inevitable that the application of enzymes in feed will further
expand. In areas where feed enzymes are already widely applied, they have been
acclaimed as the most important development in mono-gastric feeding this decade.

3.5 TEXTILE INDUSTRY

3.5.1 Introduction
In the case of fabrics made from cotton or blends of cotton and synthetic fibers, the warp
(longitudinal) threads are coated with an adhesive substance known as a ‘size’. This is to
prevent the threads breaking during weaving. The most common size is starch or starch
derivatives. After weaving, the size must be removed again in order to prepare the fabric
for finishing (bleaching, dyeing, printing, water- or crease-proofing, etc.). This process
(desizing) may be carried out by treating the fabric with strong chemicals such as acids,
bases or oxidizing agents. However, starch-splitting enzymes (amylases) have been
preferred for many years due to their high efficiency and specific action. Amylases bring
about complete removal of the size without any harmful effects on the fabric. Another
advantage of enzymes compared to the strong chemicals mentioned above, is that the
enzymes are harmless to the environment, so the waste water from the process is more
acceptable from the environmental point of view. Further details in paragraph 3.5.2.
Cotton and other natural fibers based on cellulose can be improved by an enzymatic
treatment known as bio-polishing. As the name suggests, the treatment gives the fabric a
smoother and glossier appearance. The treatment is used to remove ‘fuzz’—the tiny
strands of fiber that protrude from the surface of yarn. A ball of fuzz is called a ‘pill’ in
the textile trade. These pills can present a serious quality problem since they result in an
unattractive knotty fabric appearance. After bio-polishing, the fabric shows a much lower
pilling tendency. The other benefits of removing fuzz are a softer and smoother handling,
and superior color brightness. In paragraph 3.5.3 an example is worked out. In 1996 a
new enzyme has been launched and is the answer to another request from the market—
the need for a bio-polishing enzyme to remove fuzz without affecting the strength of the
fabric (BioTimes, September 1996). The new enzyme is a protein-engineered, mono-
component acid cellulase which has a gentler action and therefore does not result in as
much weight loss as conventional acid cellulases. This enzyme product is targeted
specifically at linen and ramie or blends containing these natural fibers. Linen is a time-
honored fiber that is back in fashion in Europe. Ramie is a common fiber for textiles in
 Applied biocatalysis 88
South East Asia and Japan.
Many ‘casual’ garments are subjected to a wash treatment to give them a slightly worn
look. A prime example is the stonewashing of denim jeans (see intermezzo 3.5.4). In the
traditional stonewashing process, the blue denim is faded by the abrasive action of
pumice stones on the garment surface. Nowadays, denim finishers are using a special
cellulase product to accelerate the abrasion. The cellulase works by loosening the indigo
dye on the denim in a process known as ‘bio-stoning’. A small dose of enzyme can
replace several kilograms of stones. The use of fewer stones results in less damage to the
garments, less wear on machines and less pumice dust in the laundry environment and of
course less stones have to be used. Bio-stoning has opened up new possibilities in denim
finishing by increasing the variety of finishes. For example, it is now possible to fade
denim to a greater degree without running the risk of damaging the garment. Productivity
can also be increased because laundry machines contains fewer stones and more
garments. More details in paragraph 3.5.5.
Bleaching with sodium hypochlorite is another tool in finishing blue jeans (see 3.5.6).
Natural fabrics such as cotton are normally bleached with hydrogen peroxide before
dyeing. Bleaches are highly reactive chemicals and any peroxide left on the fabric can
interfere with the dyeing process. That is why a thorough ‘bleach clean-up’ is necessary.
The traditional method is to neutralize the bleach with a reducing agent, but the dose has
to be controlled precisely. Enzymes present a more convenient alternative because they
are easier and quicker to use. A small dose of catalase is capable of decomposing
hydrogen peroxide to water and oxygen. Compared with the traditional clean-up
methods, the enzymatic process results in less-polluted waste water or reduced water
consumption.

3.5.2 Desizing (BioTimes, June 1998)

Before weaving the warp yarn is coated in liquefied starch. It is the commonest
constituent of the sizes used to treat cotton yarns and it is present in small quantities on
unwashed denim and other woven cotton fabrics. In the early 1980s, enzymes began to be
used for the desizing of denim because they offered some advantages over traditional
alkaline processes. In all these years, enzymatic desizing has been limited to the use of
amylases. However, in December 1997, Novo Nordisk introduced an improved desizing
product called DeniPrime™ offering a unique combination of lipases and amylases.
DeniPrime also includes a wetting agent and dispersion agent so it is fully formulated and
ready to use.
The lipases in DeniPrime are for the removal of tallow—a waxy white fat composed
mainly of triglycerides. For denim finishers, it is often impossible to know what type of
size has been used on the denim. If tallow has been used and it is not fully removed at the
start of the processing, streaks and crack marks may result from uneven abrasion. As
denim garments are too expensive to reject, the streaks and crack marks are often filled in
by hand with dye. All denim finishers are aware of these quality problems and therefore
DeniPrime can be seen as a form of insurance. By using DeniPrime, finishers can be sure
 Enzymes as processing AIDS and final products 89
time after time that they have removed all the size if it is composed of starch, animal fat
or vegetable oils. DeniPrime can also be used on ecru (undyed) fabric, which is dyed in
garment form after desizing.

3.5.3 Enzymatic Polishing of Jute/Cotton-blended Fabrics (Sreenath et al.,

1996)
Jute is the second most common natural fiber produced in the world. It is a strong and
stiff fiber. Jute (Corchorus capsularis) is grown extensively in Bangladesh, India and
China. Raw jute is processed into coarse fibers and used for the production of cords and
textiles. With the rise in sales of synthetic fibers, traditional jute markets have been lost,
and researchers have sought to develop new products. One application is manufacture of
strong, durable fabrics made from 20 to 30% jute and 70 to 80% cotton blends. However,
fabrics made from these blends have a distinct prickly sensation when in contact with the
skin. This is due to rigid jute fibers protruding from the surface. Unless the quality of the
fabric is improved, garments made of these blends are not smooth and soft and will not
perform well.
The properties of jute fiber can be improved through biochemical retting. By removing
the pectin sheath, the jute fiber is softened. Enzyme treatments can be carried out either
before or after weaving. In either case, the jute fiber is smoothed through bio-polishing.
The concept of bio-polishing was first developed in Japan. The objectives were to create
a smooth fabric and soften the fibers without the use of traditional, topically applied
chemicals. In cotton fabrics, the protruding fibers are removed by bio-polishing the fabric
surface using cellulases. Jute fiber consists not only of cellulose but also of
hemicellulose, pectin and lignin. Microbial pectinases and xylanases allow selective
removal of pectin and xylan thereby making the jute softer. Further reactions with
cellulase could result in selective removal of protruding fibers. In their study, Shreenath
et al. show that mixtures of microbial enzymes indeed remove protruding jute fibers from
the surface of jute/cotton fabrics, improving the quality of the jute-blended fabrics.

3.5.4 Intermezzo: Stonewashing (Pols, 1997)

Blue jeans have, since Jacob W.Davis and Levi Strauss in 1873 applied for a patent,
always kept their original form, a model pant that was popular in and around Genoa,
according to legend; thus the name jeans. Nevertheless, fashion also influences blue
jeans. The generation which grew up in the sixties invented all kinds of ways, e.g.
swimming in the sea with it, to give their blue jeans a worn and bleached appearance.
Industry went along with this new idea in fashion and introduced the stone-washed blue
jeans. The new blue jeans were literally stone-washed in large drums with pumice, using
about 1 kilogram of stone for 1 kilogram of jeans. During 1 hour of washing about half of
the pumice is pulverized, which implies a heavy burden on the environment. A
manufacturer which so washes 100,000 jeans, has at one side of the factory 18 m3 of wet
 Applied biocatalysis 90
pulverized pumice and at the other side an enormous storehouse with stones.
Additionally, such a stone wash means a heavy attack on the duration of life of the jeans
and of the washing machines. Besides, filling a machine half with stones is not very
efficient.

3.5.5 Enzymes Replacing Stones (BioTimes, June 98)

Cellulases, as the name suggests, degrade cellulose, which is the basic building block
found throughout the plant kingdom. As denim is made of cotton, it too consists of
cellulose. Cellulases can be used to give denim a worn, abraded look. This is a major
application for industrial enzymes in the textile industry today. The enzyme weakens the
fiber ends (fibrils) protruding from the fabric surface, a mechanism we have seen before.
Then mechanical action breaks off the weakened fiber ends from the body of the fabric.
Indigo dye is removed at the same time giving denim jeans a faded, worn look. There are
many variations on the faded look and it can be achieved not just with enzymes but with
pumice stones and other abrasives as well as bleach.
The cellulases for denim finishing introduced by Novo Nordisk since the mid 1980s
are what are known as multi-component enzymes. They contain a range of different
cellulases including endocellulases and exocellulases, according to which parts of the
cellulose molecule they attack. There also cellobiohydrolases and cellobiases. Some of
these cellulases are more suited for denim finishing than others. Modern techniques for
developing enzymes allow a selection to be made from this array of different cellulases.
Then a selected cellulase (typically an endoglucanase) can be produced on a large scale
without side activities. These are known as mono-component enzymes and they are more
refined products and much more precise in their effect. The different mono-component
products can be combined in order to obtain garments with a desired finish. Other
advantages of mono-component cellulases are better strength retention and reduced
weight loss.

3.5.6 Enzymes for Bleaching (BioTimes, March 97)

Jeans have gone from work wear to fashion wear. As a result, denim has undergone more
change and development in the last 10 years than in the preceding 100 years. Bleaching is
one of the tools of fashion designers and denim finishers for creating new looks. About
one third of all denim garments are bleached. Sodium hypochlorite is the normal
bleaching agent for jeans but it has been the focus of criticism. Chlorinated organic
compounds are a by-product in the waste water and they are known to pose a threat to the
environment. Strict regulations for discharges exist in some countries and an intensive
search is on for alternatives. Novo Nordisk has introduced such an alternative in the form
of the enzyme laccase. Laccase is an oxidoreductase catalyzing oxidation-reduction
reactions, while almost all other commercial enzymes are hydrolases. The laccase doesn’t
work on its own. If you put the laccase into a wash liquor with denim jeans, nothing
 Enzymes as processing AIDS and final products 91
happens because the indigo is totally insoluble. A mediator is needed—a kind of
chemical middleman between the enzyme and the indigo. The commercial preparation is
supplied as a light brown slurry containing both the laccase and the mediator. In the
presence of water these two substances would react together instantly but they are unable
to do so in the commercial formulation because the slurry is non-aqueous. As soon as the
slurry is poured into the wash liquor, the reaction starts because water is present and the
wash liquor turns a shade of red! The red color disappears when the bleaching reaction is
over. The reddish color comes from the ‘mediator radical’—an oxidizing agent which is
used up in the bleaching process. It oxidizes indigo to soluble degradation products. In
this way, some of the indigo can be removed. Though the chemistry is complex, in
practice the process is very simple to control with no risk of over bleaching once you
have established the right dose. The reaction stops by itself so there is no need to
inactivate the laccase.
Classic blue denim consists of an indigo warp (yarn running vertically through the
weave) and a white weft (yarn running horizontally). In conventional bleaching with
sodium hypochlorite, the white yarn becomes whiter and the blue yarn becomes lighter.
The result is low contrast between the white and the blue. When using the new laccase-
based technology, the bleaching action is specific to indigo. The natural off-white colors
of the white yarns are retained giving a darker look that is impossible to achieve with
hypochlorite. Another application is the bleaching of sulfur bottom denim where two
types of dyestuff are used—indigo and sulfur dye (black in color). The enzymatic
bleaching system is so specific that only the indigo is oxidized, not the sulfur dye. This
dye shows through more prominently giving the denim a gray ‘cast’. The new enzymatic
bleaching system is not likely to replace hypochlorite completely. For one, hypochlorite
is extremely cheap and, for another, the enzymatic system can not achieve the ‘sky’ blue
color characteristics of hypochlorite bleaching. However, other fashionable looks can be
achieved and there are many process advantages compared with hypochlorite. For
example, enzymatic bleaching is much simpler to control and works in mild conditions.
The laccase formulation is also suitable for the bleaching of delicate fabrics such as
stretch denim, which contains elastane. Hypochlorite, on the other hand, attacks the
elastane so that it loses some elasticity. Any fabric dyed with indigo can be bleached,
including lyocell and chambray (lightweight denim).

3.5.7 Putting it All Together (BioTimes, March 97)

Novo Nordisk has also introduced another application of the laccase formulation, namely
for stonewashing in combination with cellulase. It enhances the abrasion effect. In this
case, a lower dose of the enzyme/mediator slurry is sufficient. Low doses have a mild
bleaching effect though the end-result doesn’t look like bleaching. Instead the denim
looks as if it has been subject to more abrasion. Cellulases can give a stonewashed look
to jeans but they have certain limitations when used on their own. If denim finishers want
a highly abraded effect, they usually prefer a combination of cellulases and stones to
obtain the desired level of abrasion. With the laccase slurry stones are no longer needed.
 Applied biocatalysis 92
By treating the denim with a low dose of the laccase formulation after a cellulase
treatment, a comparable look is obtained. The price of enzymes is roughly the same as
the price of the stones they replace but when other savings are taken into account, the
stone-free process works out cheaper. Without stones, there is more room for garments.
Therefore a laundry’s throughput of garments increases. Pumice is an abrasive material
that can cause damage to the denim. In addition, it wears out the drums of washing
machines necessitating costly repairs. After stonewashing the garments must be destoned
and this is labor-intensive work.
One of the biggest arguments against stones is the solid waste they leave behind. The
pumice grit clogs up drains so it should be removed from waste water. Huge amounts of
pumice sludge are thus produced. Substituting pumice stones with enzymes removes this
environmental headache. With all these disadvantages, it may only be a matter of time
before stones and stonewashing are consigned to the history books on denim. Now denim
finishers can substitute stones completely with enzymes—either cellulases on their own
for light abrasion or in combination with the laccase preparation for heavy abrasion.
Advances in biotechnology have moved denim finishing out of the Stone Age!

3.6 LACTOSE HYDROLYSIS (Cheetham, 1994)

3.6.1 Introduction
Over two-thirds of the world’s adult, particularly people in developing countries, cannot
drink very much ordinary milk. The reason is that they have a deficiency

Figure 3.5 Lactose hydrolysis (Cheetham, 1994).

in lactase, the enzyme essential for digesting lactose. Lactose is a carbohydrate unique to
milk and it is also referred to as milk sugar. Lactose is also the major sugar present in
milk. Its hydrolysis (Figure 3.5) into its constituent sugars, glucose and galactose, is thus
essential for making milk (products) enjoyable for all adults. Also the resulting mixture
of galactose and glucose is sweeter, is less prone to crystallization and is more soluble
 Enzymes as processing AIDS and final products 93
than the original lactose. Large quantities of whey are produced as a by-product of cheese
making and contains up to 5% lactose. A little whey is used in animal feeds and as a
fermentation feedstock, but most is waste. Acid hydrolysis of lactose is possible, but
color and bitterness are rapidly produced, requiring expensive purification so that
commercialization has not taken place. Hydrolysis with galactosidase (lactase) is an
attractive alternative.

3.6.2 Applications
By pre-treating milk with lactase, all adults can enjoy milk and a whole range of other
lactose-free dairy products can be made such as ice cream and yogurt. In the production
of ice cream, lactose hydrolysis can also be used to improve certain properties such as the
texture, sweetness and tendency to crystallize. The increased sweetness is also
advantageous in the manufacture of flavored milk products because less sugar needs to be
added.
The hydrolysis of lactose conveys a number of desirable properties to whey—increased
sweetness for one. Lactase-treated whey can be used as a sweetener in various foods such
as ice cream, bakery products, beverages and confectionery. This treatment can save
large quantities of whey from going to waste and presenting a potential pollution hazard.

3.6.3 Technical Factors

Lactases are produced by many yeasts such as Klyveromyces lactis. Lactases produced by
molds such as Aspergillus niger are especially useful as they have acid pH optima
suitable for processing dairy materials such as whey, whereas the yeast enzymes are
usually used for treating milk since they have neutral pH optima. Lactases are still
relatively expensive enzymes when compared with other industrial hydrolases such as -
amylases and proteases. Immobilization can reduce enzyme treatment costs, but fouling
by undissolved solids such as milk proteins reduces efficiency. Procedures have,
however, been developed that are not prone to fouling. Whey is easily contaminated with
micro-organisms, but effective cleaning and sanitation procedures were developed,
including the use of anti-microbial additives that did not harm the enzyme and that were
consistent with the safety requirements of a food processing operation.

3.6.4 Commercial Factors

The following favorable aspects have determined the commercial success of lactose
hydrolysis using lactases, despite of a number of disadvantages:

• Enzymes were already used in the dairy industry, e.g. rennet for clotting of milk.
• Lactases can be used in the production of milk for lactose-intolerant people,
concentrated milk products, fermented milk products, animal feed ingredients and as a
 Applied biocatalysis 94
food ingredient.
• Very large amounts of cheap whey are produced.
• Food and beverage applications for lactase-treated whey syrup have been developed.
• Use of whey hydrolysate as fermentation feedstock has been developed.
• The technology developed was suitable for commercialization via joint ventures
between the inventor company and ‘user’ companies.

As disadvantages can be listed:

1. The real need for lactose-free milk products is limited as most lactase-deficient adults
can tolerate larger amounts than they are likely to consume.
2. Milk treatment tends to be confined to batch processes since continuous processing
with immobilized enzyme results in microbial contamination, which is encouraged
because mild conditions of temperature and pH have to be used.
3. Enzyme treatment and processing costs are relatively high, for instance lactase costs
are estimated at 200 $/ton of product, as compared with immobilized glucose
isomerase and invertase treatments that cost 2–3 $/ton product and 15–30 $/ton of
product, respectively.
4. Lactose crystallizes easily, so that concentration prior to transportation of whey for
treatment can be difficult. Whey is very prone to microbial contamination so that
maintenance of hygienic operation conditions is a major processing problem,
especially as the lactases are not thermostable and so contamination cannot be
prevented by operating at elevated temperatures.
5. Despite the huge quantities of whey produced, production is geographically very
dispersed. Therefore there are few plants with a big enough out-put to allow cost-
effective treatment, especially when taking into account transportation costs.
6. Whey-based products are in competition with starch-based sweeteners.
7. Most lactose-intolerant people are too poor to buy lactase-treated milk.
8. Some whey has found alternative applications, such as fermentation to produce
ethanol.

Despite of this long list of disadvantages commercial processes have been developed.

3.6.5 Commercial Processes

The first of these commercial processes to be reported was in Milan, based on Snam
Progetti immobilized-enzyme technology, and was said to have a capacity to treat 100 m3
milk/day. The Finnish company Valio Oy has, since 1975, operated a commercial scale
process that is reported to use two 0.6 m3 columns of immobilized enzyme with a
capacity of 80 m3/day, operating using whole whey as well as demineralized and ultra-
filtered whey. Other commercial ventures have originated from the immobilized-lactase
technology developed by Corning (USA) and exploited via joint ventures with the
following companies. The Milk Marketing Board and Specialist Dairy Ingredients (UK)
operate a plant processing 20 m3 of sweet whey/day with a reactor productivity of 1
 Enzymes as processing AIDS and final products 95

m3/40 kg of immobilized enzyme/h. Coivre and Union Laitere Normande (France)

operate a 10 m3/day semi-industrial plant. Most interesting is the Nutrisearch plant
(Kentucky, USA) with a capacity to process 380 m3 of cottage-cheese whey/day,
producing a protein-rich retentate from fermentation, with the permeate treated with 1500
kg scale lactase columns, and with the resulting hydrolysate used to produce bakers yeast.

3.6.6 Conclusions
As in the analogous hydrolysis of sucrose by invertase, the hydrolysis of lactose is easy.
The real challenge lies in developing a cost-effective process for the production of a
relatively low-value-added product; in particular overcoming processing problems, such
as microbial contamination, that make efficient use of the enzyme very difficult. Because
of the relatively low-added-value margins the range of technical options available are
limited. Nevertheless some commercial processes exist. However despite those successes
widespread use of lactase technology has not occurred. Certainly companies such as Tate
and Lyle, who sought to develop immobilized-lactase systems for sale to milk-processing
companies and the Damrace Corp. who developed ultra-filtration reactors for lactase
hydrolysis have been unsuccessful.

3.7 MICROBIAL TRANSGLUTAMINASE IN FOOD PROCESSING (Zhu et

al., 1995; and references cited therein)

3.7.1 Introduction
Transglutaminase (protein-glutamine -glutamyltransferase, EC 2. 3. 2. 13) is an enzyme
capable of catalysing acyl-transfer reactions introducing covalent cross-links between
proteins as well as peptides and various primary amines. When the -amino groups of
lysine residues in proteins act as an acyl acceptor, -( -Glu)-Lys bonds are formed both
intra- and inter-molecularly. Without primary amines in the reaction system, water
becomes the acyl acceptor and the -carboxyamide groups of glutamine residues are
deamidated, becoming glutamic acid residues. The transglutaminase-catalysed reactions
can be written as:

These transglutaminase-catalysed reactions can be used to modify the functional

properties of food proteins. Transglutaminase has been used to catalyze the cross-linking
of a number of proteins, such as whey proteins, soy proteins, gluten, myosin and
 Applied biocatalysis 96
actomyosin. The modification of food proteins by transglutaminase may lead to textured
products, help to protect lysine in food proteins from various chemical reactions,
encapsulate lipids and/or lipid-soluble materials, form heat- and water-resistant films,
avoid heat treatment for gelation, improve elasticity and water-holding capacity, modify
solubility and functional properties, and produce food proteins of higher nutritive value
through cross-linking of different proteins containing complementary limiting essential
amino acids.
Transglutaminase has been found in animal and plant tissues and microorganisms.
Since the 1960s, the purification, characterization and application of Ca2+-dependent
transglutaminase of animal origin, mainly guinea-pig liver, have been intensively studied.
A process chart of transglutaminase production from different sources is shown in Figure
3.6.
Guinea-pig liver has been the sole source of commercial transglutaminase for decades.
The scarce source and the complicated separation and purification procedure for
obtaining tissue transglutaminase have resulted in an extremely high price of the enzyme,
about USD 80 for one unit. It is thus not possible to apply such tissue transglutaminase in
food processing on an industrial scale. Separation and purification of transglutaminase
from fish tissue and plant tissue are still in their infancy. Recently efforts have been made
to search for transglutaminase derived from microorganisms. Transglutaminases were
found in cultures of Streptoverticillium sp. and Streptomyces sp. Microbial fermentation
makes it possible to achieve mass production of transglutaminase from cheap substrate. A
number of examples of the application of microbial transglutaminase in food processing
have been announced and are discussed below.

3.7.2 Applications
The production of transglutaminase by microorganisms makes it possible to apply this
enzyme in a variety of food processes. An overview of the application possibilities for
microbial transglutaminase in food processing is given in Table 3.5. A few of these
examples will be described in some detail below, in order to show the simplicity of the
treatment with microbial transglutaminase and the positive effects that can be obtained.
 Enzymes as processing AIDS and final products 97

Figure 3.6 Process chart of transglutaminase production from different sources

(Zhu et al., 1995).

In meat processing it is of great interest to maximize the yield of marketable products.

This includes development of methods for re-structuring low-value cuts and trimmings to
 Applied biocatalysis 98
improve their appearance, flavor and texture and to enhance market value. Re-structuring
treatment usually involves size reduction, reforming and binding. In such a treatment,
transglutaminase can have a very important

Table 3.5 Overview of applications of microbial transglutaminase in food processing

Source Product Effect

Meat Hamburger, meatballs, stuffed Improved elasticity, texture, taste and
dumplings, shao-mai flavor
Canned meat Good texture and appearance
Frozen meat Improved texture and reduced cost
Molded meat Restructuring of meat
Fish Fish-meat paste Improved texture and appearance
Krill Krill-meat paste Improved texture
Collagens Shark-fin imitation Imitation of delicious food
Wheat Baked foods Improved texture and high volume
Soya bean Mapuo-Doufu Improved shelf-life
Fried Tofu (Aburaage) Improved texture
Tofu Improved shelf-life
Vegetables and Celery Food preservation
fruits
Casein Mineral absorption promoters Improved mineral absorption in intestine
Cross-linked protein Allergenicity reduction
Gelatin Sweet foods Low-calorie foods with good texture,
firmness and elasticity
Fat, oil and Solid fats Pork-fat substitute with good taste,
proteins texture and flavor
Plant proteins Protein powders Gel formation with good texture and taste
Seasonings Seasonings Improve taste and flavor

function. Minced meat and other food ingredients can be mixed with transglutaminase,
shaped, packed in pressure-resistant containers and retorted to manufacture meat products
such as hamburgers, meatballs, stuffed dumplings and shao-mai (a typical Chinese food).
Such foods show improved elasticity, texture, taste and flavor. Minced beef and pork,
flour, onion, skim-milk powder and condiments can also be mixed with water and
 Enzymes as processing AIDS and final products 99
microbial transglutaminase, packed with sauce in bags and retorted to make raw
hamburgers.
Methods have been reported for manufacturing fish-meat paste containing
transglutaminase to manufacture kamaboko (Japanese fish-meat paste) with good texture
and whiteness. Another processing method is the manufacture of shark-fin imitation food
with transglutaminase. Shark-fin is considered as a delicious and healthy (functional)
food in Southeast Asia. An imitation of shark-fin is prepared by cross-linking gelatins,
collagens or a mixture thereof with transglutaminase and making a gel from the product.
The collagen ingredient may be collagen fibers, collagen fibrils, collagen solutions or
mixtures thereof. A shark-fin imitation food can be prepared for instance by treating
gelatin in water at pH 7 with transglutaminase, extruding the solution through holes, thus
forming fibrous gel and, finally, drying the product.
A method has also been developed for manufacturing storage-stable retort mapuodoufu
(doufu is tofu in Chinese). Mapuo-doufu, braised tofu with minced beef and chili pepper,
is one of the most typical hot-spiced dishes in Sichuan Province, China. In this method,
retort mapuo-doufu that can be preserved at room temperature for a long time, is
manufactured by treating soy-bean-milk solutions with coagulating agents and
transglutaminase at temperatures up to 80°C to manufacture tofu (soybean curd),
optionally cutting the tofu into pieces, putting it in heat-resistant containers with minced
beef and seasonings, and retort sterilization.
Freshness of vegetables and fruits can be maintained by coating with a membrane
containing transglutaminase and proteins. Cut celery, treated with an aqueous solution
containing transglutaminase, proteins, gelatins and Partner-S (natural bacteri-cide from
spices) and then heated at 50°C for 5 min to form coating membranes, was kept at 20°C
for 3 days. It showed only up to 300 bacterial cells/g, compared to 2×106 without
treatment.
A method using transglutaminase for reducing the allergenicity of some food proteins
and/or peptides has been developed as well. s1-Casein (23 kDa) was treated with
transglutaminase at 25°C for 20 h in water to manufacture cross-linked casein (approx.
90 kDa), which was less allergenic.
A promoting material for absorption of minerals in the human body can be prepared by
deaminating of casein through treatment with transglutaminase. The resulting material
promotes absorption of minerals in intestine and can be used in the food industry and for
medicines, for instance mineral supplement formulations for adults, children and infants.
The casein is soluble in neutral and slightly acid conditions and can keep minerals
solubilized in the intestine.

3.7.3 Perspectives
It is of paramount interest to search world-wide for new protein sources and to broaden
the application potentials of existing proteins for human consumption. In developing
countries, many people are still suffering from starvation and efforts are being focused on
producing acceptable protein foods from non-animal proteins, to solve the problem of
 Applied biocatalysis 100
protein deficiencies. On the other hand, in addition to the awareness of health problems
caused by obesity, people in developed countries are increasingly aware of the
environmental burden caused by the surplus of livestock. Facing a novel food product,
consumers are very sensitive to the properties such as flavour, nutritional value,
appearance, shelf life and palatability. In this respect, protein modification by enzymes,
especially by microbial transglutaminase whose mass production can be achieved by
fermentation from cheap substrates, is one of the most promising alternatives in
developing novel protein foods.

3.8 CONCLUDING REMARKS (Stroh, 1998)

The industrial enzyme business is undergoing an unprecedented wave of corporate

consolidation as well as growing cooperation with start-up companies to obtain cutting-
edge technologies against competition. Enzymes provide biological solutions to industrial
problems, using the catalysts of living cells in large-scale processes and are expanding
globally in acceptance and use. To a large degree, industrial enzyme development has
been the province of European companies, dominated by the Danish leader, Novo
Nordisk, Gist-brocades (the Netherlands) and Genencor International (USA). A decade of
other major suppliers of commercial enzymes come from Japan, UK, Germany and
France. The wild card lies in Asia. In just a few years, China has emerged as a visible
factor in this global business. According to statistics compiled by the Chinese Enzyme
Preparation Industry Association, over 200,000 metric tons of bulk enzyme preparations
are produced annually, increasing historically at a rate of 10% per year.
A dramatic industry shakeup occurred over the past three years with companies
merging or being acquired with consequent rationalization of both production facilities
and product ranges. Gist-brocades and Solvay (Brussels, Belgium) sold their technical
enzymes business (primarily those for detergents, textiles and large-scale starch
conversion) to Genencor International, which is now the second largest producer of
industrial enzymes after Novo Nordisk.
There are two main reasons for the consolidation of the major market players,
especially within the important detergents segment: prices of the new commodities, with
the exception of the latest-launched oil- and fat-fighting lipase enzymes, have become
extremely competitive. Only companies with large-scale production have been able to
survive the consequent rigorous cost-cutting waves. It is unlikely that competitors that do
not achieve economies of scale or devote significant resources to R&D would have the
staying power.
The engines of growth for the industrial enzymes business may lie in start-up
companies that interest and challenge major supplies with their novel enzyme products
and techniques. New technologies, such as DNA shuffling of natural genes, pathway
discovery technology and novel host systems, are expanding the tools available for
enzyme R&D. For example, using standard culturing methods, researchers could clone
three or four enzymes per year. In contrast, building libraries from nucleic acids nets a
 Enzymes as processing AIDS and final products 101
thousand novel enzymes annually. The latest technologies skip the traditional culturing
steps and immediately extract the DNA from the uncultured microbes. The nucleic acids
are converted to cloned libraries or gene banks, which are screened for enzymatic
activity. All in all, more than a decade of enzyme-technology startups have become
players on the industrial-enzymes market.

3.9 REFERENCES AND FURTHER READING

BioTimes, a quarterly magazine available from Novo Nordisk, Enzyme Promotion Dept.,
Novo Allé, DK-2880 Bagsvaerd/DK.
Chaplin, M.F. and Bucke, C. (1990) Enzyme technology . Cambridge: Cambridge
University Press.
Cheetham, P.S.J. (1994) Case studies in applied biocatalysis—from ideas to products. In
Applied biocatalysis , edited by J.M.S.Cabral, D.Best, L.Boross and J.Tramper, pp. 47–
108. Singapore: Harwood Academic Publishers.
Cowan, D. (1992) Advances in feed enzyme technology. Agro-Food-Industry Hi-Tech ,
May/June, 2, 9–11.
Dijck, van P.W.M. and Geerse, C. (1993) More problems solved, the Gist , 55, 6–7.
Ee, van J.H. (1994) Towards even lower wash temperatures, the Gist , 57, 8–10. the Gist,
a quarterly magazine available from Gist-brocades, Delft, the Netherlands
(http://www.gist-brocades.com).
Godfrey, T. and West, S., eds. (1996) Industrial Enzymology (2nd edition). London:
MacMillan Press.
Haarasilta, S. and Pullinen, T. (1992) Novel enzyme combinations: A new tool to
improve baking results. Agro-Food-Industry Hi-Tech , May/June, 12–13.
Pols B. (1997) Biotechnologie voor spijkerbroeken, NRC Handelsblad , 30 May.
Riisgaard, S. (1990) The enzyme industry and modern biotechnology. In Proceedings 5th
European Congress on Biotechnology , Vol. 1 , 31–40. Munksgarrd.
Sreenath, H.K., Shah, A.B., Yang, V.W., Ghana, M.M. and Jeffries, T.W. (1996)
Enzymatic polishing of jute/cotton blended fabrics. J. Ferm. Bioeng. , 81, 18–20.
Stobart, H.G.A. (1993) Using enzymes to improve the productive value of poultry and
pig feedstuffs. Agro-Food-Industry Hi-Tech , January/February, 27–29.
Stroh, W.H. (1998) Industrial enzymes market: Growth experienced from new products
and movement into global market. Genetic Engineering News , March 1, pp. 11 and 38.
Thole, E. and Velink, M.-P. (1996) Wasmiddelen verrijkt met enzymen. Chemisch
Magazine , March, 109.
Voragen, A.G.J. and van den Broek, L.A.M. (1991) Fruit juices. In Biotechnological
innovations in food processing , 187–210. Oxford: Butterworth-Heinemann.
Walsh, G.A., Power, R.F. and Headon, D.R. (1993) Enzymes in the animal-feed industry.
TIBTECH , 11 , 424–430.
Wrotnowski, C. (1997) Unexpected niche applications for industrial enzymes drives
market growth. Genetic Engineering News , February 1.
 Applied biocatalysis 102

Zhu, Y., Rinzema, A., Tramper, J. and Bol, J. (1995) Microbial transglutaminase—a
review of its production and application in food processing . Appl. Microbiol.
Biotechnol. , 44 , 277–282.
 4.
CASE STUDIES IN THE APPLICATION OF
BIOCATALYSTS FOR THE PRODUCTION
OF (BIO)CHEMICALS
PETER S.J.CHEETHAM
Zylepsis Ltd., 6 Highpoint, Henwood Business Estate, Ashford, Kent, UK
Tel: +44–1233–660555; Fax +44–1233–660777
Updated and adapted with the aid of Adrie J.J.Straathof

ABSTRACT
The list of commercial enzymatic and microbial processes for the conversion of
a precursor into a desired product is rapidly expanding. In this chapter,
seventeen entries from this list are treated as case studies. The examples
included span a wide range of reaction type, process configuration, substrate
type, and product application area.
The case study examples show the biocatalysis is proving successful in a
growing number of processes in a range of industries. Biocatalysis technologies
are now established as a standard industrial approaches not just for specialty
materials, but also for commodities such as 6-aminopenicillanic acid in the
pharmaceutical industry, high fructose corn syrup for beverages and foods, and
acrylamide in the chemicals industry. The examples given in this chapter are
merely a selection, but appear to show that there are no obvious common factors
making for success or failure. This may be because of the multiplicity of
technical factors involved and their complex inter-relationships, but most
especially because of the importance of commercial factors that must also be
satisfied, such as described in Chapter 13.
Nevertheless some general common themes do appear such as:
• A good IPR position (See Chapter 12).
• A low cost manufacturing process.
• Easy scaling up to a commercial scale.
• A satisfactory safety and regulatory situation.
All making for an acceptable time and cost to reach the market and resulting
in a big return on investment given a sufficiently big market in which good
 Case studies in the application of biocatalysts 104
profit margins can be earned.

4.1 INTRODUCTION

The applications of biocatalysis that are treated in this chapter are bioconversions that
fulfil a number of criteria. These are that they:

• describe a reaction or a set of consecutive reactions in which a pre-formed precursor

molecule is converted into a target product; rather than fermentation processes in
which the desired product is produced de novo from a general feed stock such as
glucose via the primary metabolism.
• involve the use of either free enzymes, immobilised enzymes, or combinations of
enzymes still associated with their parent cells (although of course fermentation is
usually required to produce the enzyme or cells in the first place). The distinction
between a fermentation and a bioconversion process, or a fermentation stage and a
bioconversion stage of a process, however, is not clear-cut. In some fermentations
precursors are fed, for instance phenylacetic acid in the penicillin G fermentation, and
during some bioconversions with whole cells the cells may still be metabolically fully
active.
• deal with the production of both speciality chemicals, that are sold principally on the
basis of their functional properties, and also commodity chemicals usually sold in
larger quantities and lower prices on the basis of price, supply and quality factors etc.
Therefore the products of the bioconversion reaction will usually be isolated as pure
chemicals, in contrast to the situation for the processes treated in Chapter 3. Some
processes are on the borderline. The case study of lactose hydrolysis is treated in
Chapter 3. Lactose hydrolysis has been carried out to remove lactose from milk, but
also to produce a fermentable sugar solution from whey.
• have been reported to be operated on a commercial scale, or have been successfully
scaled-up and announced to be commercialised at a scale of usually> 100 kg/a,
although in some cases of very expensive products this figure may be lower.

Table 4.1 shows an extensive list of applications that fulfil these criteria. Nevertheless,
such a list cannot be complete, because not all information is in the public domain.
Moreover, the number of applications, especially in the field of chiral intermediates,
seems to be growing more rapidly than ever (see Stinson, 1996, e.g.).
Clearly, most of the products in Table 4.1 are chiral compounds. None of the products
is racemic, and only a few are achiral. The biocatalysts are (combinations) of enzymes or
cells. If the key enzyme has been indicated it may be used pure, partly purified in a cell-
free extract, or in a whole cell. For each option, the biocatalyst may be used free or
immobilized. If the name of a microorganism has been indicated, usually several of its
enzymes are active in the catalysis. The entries that are displayed in bold are treated in the
case studies further on in this chapter, in the same order as in Table 4.1.
 Applied biocatalysis 105
These case studies are an attempt to demonstrate the complexity of the
commercialisation of biocatalytic (biotransformations/bioprocessing) processes, and to
compare the technical and business factors that all contribute to the success or failure of
particular biocatalysis-based industrial projects. This chapter is definitely not a
straightforward review of the science involved, such information is mostly readily
available from other published sources. Instead, in each example some background
information has been given, and then important technical and commercial factors that
have influenced the success or future of the product in the market place are highlighted.
Examples are included that:

• span a wide range of the commercial applications of biocatalysis in pharmaceuticals,

agrochemicals, food etc.

Table 4.1 (Bio) chemicals commercially produced by biocatalysis. Entries in bold face
are treated as case studies in this chapter.

a Carbohydrates
Product Key substrate Biocatalyst
N-acetyl-D-neuraminic acid GlcNAc+pyruvate Neu5Ac-aldolase
butyl -glucoside maltose -glucosyltransferase
-cyclodextrin amylose cyclodextrin glucosytransferase

-cyclodextrin amylose cyclodextrin glucosytransferase

-cyclodextrin amylose cyclodextrin glucosytransferase

dextran sucrose dextransucrase
frucose syrup (HFCS) glucose glucose isomerase
fructooligosaccharides inulin inulinase
fructooligosaccharides sucrose fructosyl transferase
galactomannan guar gum -galactosidase
galactooligosaccharides lactose -galactosidase
glucodextrins amylose -amylase
gluconic acid glucose Aspergillus niger
glucooligosaccharides sucrose+maltose -glucosyltransferase
glucose-6-phosphate glucose+acetylphosphate hexokinase+acetate kinase
glucose syrup glucodextrins glucoamylase+pullulanase
2-keto-L-gulonic acid L-sorbose G. oxydans+Bacillus sp.
 Case studies in the application of biocatalysts 106

leucrose sucrose dextransucrase

maltose amylose -amylase
palatinose (isomaltulose) sucrose isomaltulose synthase
L-sorbose (Vit. C precursor) sorbitol Acetobacter sp.
trehalose amylose Rhizobium+Arthrobacter sp.

b Fatty acids and derivatives

Product Key substrate Biocatalyst
1-palmitoyl-2-oleyl-3-stearoyl- vegetable oil+fatty acid Candida antarctica lipase B
glycerol
-decalactone ricinoleic acid yeast strains
-decalactone 11-hydroxypalmitic yeast strains
acid
decyl oleate oleic acid Candida antarctica lipase B
ethyl glucoside esters fatty acid Candida antarctica lipase B
ethyl heptanoate/2-methylbutanoate acid+alcohol esterases
fatty acids triglycerides Rhizopus niveus lipase
isopropyl myristate myristic acid Candida antarctica lipase B
isopropyl palmitate palmitic acid Candida antarctica lipase B
cis-3-hexenol linoleic acid lipoxygenase+hydroperoxide
lyase
methyl ketones fatty acids Penicillium roqueforti
octyl palmitate palmitic acid Candida antarctica lipase B
PEG400 monostearate stearic acid Candida antarctica lipase B
unsaturated fatty acids triglycerides lipase

c Steroids produced by biocatalysis

Product Key substrate Biocatalyst
androsta(di)enedione sitosterol or cholesterol Mycobacterium sp.
cortisol 17 -acetoxy-11-deoxycortisol Curvularia lunata
11 -hydroxy-progesterone progesterone Rhizopus arrhizus

prednisolone hydrocortisone Arthrobacter simplex

 Applied biocatalysis 107

prednisone cortisone Arthrobacter simplex

triamcinolone 9 -fluoro-16 -hydroxycortisol Streptomyces roseochromogenes
trimegestone triketone precursor S. cerevisiae

d Peptides and -lactams produced by biocatalysis

Product Key substrate Biocatalyst
7-ACA cephalosporin C D-amino acid oxidase +
glutaryl acylase
7-ADCA cephalosporin G penicillin acylase
6-APA penicillin G penicillin G acylase
6-APA penicillin V penicillin V acylase
aspartame precursor Z-L-Asp+DL-Phe-OMe thermolysin
cefalexin 7-ADCA+D-PhGly deriv. penicillin acylase
cefuroxime cephalothin Rhodospanidium toruloides
insulin precursor Thr-O-tBu-B30-insulin trypsin
L-malyl-Tyr-O-Et L-Tyr-OEt aminopeptidase
oligolysine L-Lys-OEt trypsin
oligoLys/Arg L-Lys-OEt+L-Arg-OEt trypsin

e Amino acids and derivatives produced by biocatalysis

Product Key substrate Biocatalyst
L-alanine DL-aspartic acid L-aspartate decarboxylase
L-alanine DL-acetyl-alanine aminoacylase
D-aspartic acid DL-aspartic acid L-aspartate decarboxylase
L-aspartic acid fumaric acid aspartase
citrulline L-arginine arginine deimidase
L-cysteine thiazolinone derivative Ps. thiazolinophilum
L-DOPA L-tyrosine tyrosine phenol lyase
L-DOPA DL-acetyl-DOPA aminoacylase
L-homophenylalanine DL-homo-Phe-amide aminopeptidase
D-homophenylalanine-amide DL-homo-Phe-amide aminopeptidase
D- p -hydroxyphenylglycine DL-hydantoin hydantoinase+carbamoylase
L-tert-leucine trimethylpyruvate leucine dehydrogenase +
formate dehydroenase
 Case studies in the application of biocatalysts 108

L-lysine DL-2-aminocaprolactam racemase+aminolactamase

L-methionine DL-acetyl-Met aminoacylase
L-ornithine L-arginine arginase
L-phenylalanine DL-acetyl-Phe aminoacylase
L-phenylalanine DL-Phe-OEt chymotrypsin
L-phenylalanine phenylpyruvic acid transaminase
L-phenylalanine trans-cinnamic acid phenylalanine ammonia lyase
D-phenylglycine DL-hydantoin hydantoinase+carbamoylase
L-tryptophan L-Ser+indole E. coli
L-tryptophan DL-acetyl-Try aminoacylase
D-valine DL-hydantoin hydantoinase+carbamoylase
L-valine DL-valineamide aminopeptidase
L-valine DL-acetylvaline aminoacylase
D-valineamide DL-valineamide aminopeptidase

f Nucleoside derivatives and precursors of analogues produced by biocatalys

Product Key substrate Biocatalyst
carbovir precursor (RS)-2-azabicyclo [2.2.1] hept-5-en-3-one lactamase
epivir precursor (+/−)-BCH1891 cytidine deamin
5′-guanosine monophosphate RNA ribonuclease
5′-inosine monophosphate 5′-guanosine monophosphate deamidase

g Other chiral products produced by biocatalysis

Product Key substrate Biocatalyst
acetyl-(+)-trans-2- (+/−)-trans-2- Candida antarctica
methoxycyclohexanol methoxycyclohexanol lipase B
(R)-allethrolone (RS)-allethrolone lipase
L-carnitine -butyrobetaine Agrobacterium
(S) -2-chloropropionic acid (RS) -2-chloropropionic acid dehalogenase
(cis, 1S)-1,2-dihydrocatechols substituted benzenes benzene dioxygenase
diltiazem precursor (RS)-MPGM2 a.o. S. marcercans lip
 Applied biocatalysis 109

(R)-1,2-epoxyalkanes 1-alkenes Nocardia corallina

(R) -glycidol (RS) -glycidyl butanoate pig pancreas lipase
(R) -glycidyl butanoate (RS) -glycidyl butanoate pig pancreas lipase
HPOPS3 POPS3 Beauveria bassiana

(S)- -hydroxyisobutyric acid isobutyric acid Candida rugosa

L-malic acid fumaric acid fumarase

(R)-methoxyacetyl derivative (RS)-1-phenylethylamine Pseudomonas lipase
(S)-naproxen (RS)-ethyl naproxen esterase
oxamniquine de-hydroxy-oxamiquine Aspergillus sclerotior
D-panthotenic acid DL-pantoyl lactone lactonase
L-phenylacetylcarbinol benzaldehyde+pyruvate pyruvate decarboxy
pravastatin mevastatin Streptomyces
carbophilus
(R)-styrene oxide styrene Nocardia corallina
Trusopt precursor4 oxosul phone Neurospora crassa

h Other non-chiral products produced by biocatalysis

Product Key substrate Biocatalyst
acrylamide acrylonitrile nitrile hydratase
acrylic acid acrylamide amidase
6-hydroxynicotinic acid nicotinic acid Achromobacter xylosoxydan
5-methylpyrazine-2-carboxylic acid 2,5-dimethylpyrazine Pseudomonas putida
nicotinamide 3-cyanopyridine nitrile hydratase
1 BCH189=4-amino-1-(2-hydroxymethyl-[1,3]oxathiolan-5-yl)-1H-pyrimidin-2-one (Strohl,
1997)
2 MPGM=p-methoxyphenyl-glycidic acid methyl ester
3 HPOPS=(R)-2-(4-hydroxyphenoxy)propionic acid; POPS=(R)-2-phenoxypropionic acid
(Dingler et al., 1996)
4 Trusopt=(4S,6S)-hydoxysulphone heterocyclic compound (A.J.Blacker and R.A.Holt, in:
Collins et al., 1997)

• carry out a variety of different reactions such as isomerisation, resolution of racemates,

peptide bond synthesis etc.
• involve the use of either free or immobilised biocatalysts
• involve activity towards both naturally produced and synthetic organic chemicals
• have proved to be big commercial successes, as well as some examples that have been
only partially successful, since much can be learnt from such failures.
 Case studies in the application of biocatalysts 110
In every case the information provided has been obtained by collating public domain
sources of information, but unfortunately very often little data is available, particularly on
commercial aspects, even for products that have proved to be big successes. Thus
microbial biotransformations for steroid modification, particularly stereoselective
hydroxylations, such as the use of Rhizopus arrhizus to convert progesterone into anti-
inflammatory and other drugs via 11- -hydroxyprogestrone, have proved to be very
successful. However, comparatively little useful information exists from public domain
sources, despite (or perhaps because) a market of hundreds of millions $/a exists for such
microbially transformed steroids (cortisone, aldosterone, prednisolone and prednisone
etc.) produced by microbial hydroxylation and dehydrogenation reactions coupled with
complimentary chemical steps.
In some cases particular processes are especially important because of the value of the
products produced and also the amounts of enzyme used, i.e. from both the enzyme
producer and user standpoints. Therefore the examples on fructose syrups and 6-
aminopenicillinic acid have been expanded to include much information on the business
and technical strategies employed, and detailed process economic aspects respectively.
Many other examples of biocatalysis-based commercial processes and products also exist
with many others undergoing development. These other products also exhibit the same
important technical and commercial features that are identified for the case studies in this
chapter.

4.2 HIGH FRUCTOSE SYRUPS

4.2.1 Background
Glucose syrups have been used in the food industry for a long time. Fructose is
significantly sweeter than glucose. No effective chemical isomerisation methods are
possible, and other sources of fructose, for instance by the hydrolysis of inulin, are not
yet performed on large scale. Therefore an enzyme isomerisation technology has been
developed (Jensen and Rugh, 1987; White, 1992; Pedersen, 1993).

4.2.2 Technical Features

• The discovery of a non-cofactor dependent enzyme (glucose isomerase) that would

isomerise glucose into fructose (see Figure 4.1).
• The development of cost-effective immobilised cell isomerisation processes.
• The development of a chromatographic enrichment process to produce HFCS (high
fructose corn syrup) containing 55% fructose for use in soft drinks especially Pepsi
and Coca Colas.
 Applied biocatalysis 111

Figure 4.1 Glucose isomerisation.

• HFCS processes could not have been developed if efficient enzyme-based production
methods for converting starch into glucose had not already been developed.

4.2.3 Commercial Features

• Glucose isomerisation was the only cost-effective method of producing HFCS;

recently, inulin hydrolysis has become a minor competitor.
• Technology is now relatively mature.
• The first glucose isomerase to be used industrially was operated by the Clinton Corn
Co. (Standard Brands, USA) from 1967–1970 on a small scale. However, the glucose
isomerase process was only really introduced in the early 1970s, when sucrose prices
were at a historic high, enabling HFCS to gain market share rapidly in products that
had been traditionally sweetened with sucrose or glucose.
• By-products of starch refining and HFCS production are significant and reduce HFCS
production costs by 30–45% (Table 4.2). This is because the corn is 70% starch on a
dry weight basis, it also contains 10% protein, 4.5% fat and 2.7% crude fibre. In
addition much of the glucose syrup produced is fermented to produce ethanol for fuel
use.
• HFCS was rapidly adopted by big soft-drink manufacturers.
• Two main HFCS products are made, containing 42% fructose, and 55% fructose for
soft drink applications.
 Case studies in the application of biocatalysts 112
• Economies of scale are important, especially as the big corn processing companies such
as A.D.M.Cargill and Staley are in close competition.
• Starch costs are the main element in the production costs for HFCSs. Because starch is
significantly cheaper than sucrose, HFCS production using glucose isomerase is more
economical than by the hydrolysis of sucrose by invertase,

Table 4.2 Major components of US high fructose syrup’s variable production costs.

Component Production costs (1982) (%)

Corn 50
Energy 20
Labour 10
Chemicals 10
Enzymes 5
Miscellaneous 5
Total 100

especially as isomerase enzyme costs per tonne of product manufactured are about
one third those of the comparable invertase costs. Corn (maize) is the preferred raw
material because it can be grown in huge monocultures and is easily dried, making it
easy to store.

4.2.4 Disadvantages

• Glucose isomerase can only produce HFCS containing 42% fructose under normal
operating conditions. More thermostable glucose isomerases would allow operation at
higher temperatures thus shifting the equilibrium of the reaction towards the formation
of higher fructose concentrations.
• Glucose isomerase has a higher pH optimum than is required in the preceding starch
liquification and saccharification steps so that pH adjustment is necessary. Also the -
amylase used to carry out saccharification requires calcium ions for full activity, but
calcium inhibits glucose isomerisation, necessitating its removal by ion-exchange
treatment prior to isomerisation.
• Glucose isomerase is an intracellular enzyme with relatively poor stability, making
purification and immobilisation difficult.
• HFCS is in competition from other sweeteners such as sucrose and aspartame.
• Cost-effective HFCS production is critically dependent on the profit contribution of
 Applied biocatalysis 113
products from starch refining such as gluten, starch and glucose syrups.
• The EC have imposed an embargo on large-scale HFCS manufacture in Europe so as to
protect European sugar beet farmers. This embargo has been extended to HFCS
manufacture by inulin hydrolysis.

4.2.5 Technical Details

Major producers of glucose isomerase are Novo-Nordisk, Genencor and Cultor (now
Danisco). They all use different microbial sources of glucose isomerase. Novo and
Genencor produce immobilised cell products. For instance Novo sell immobilised cells of
Streptomyces murinus with a productivity of 15 tonne dry weight syrup/kg immobilised
cells and an operational life-time of 100–500 days.

Table 4.3 Estimated high fructose syrup production capacities, 1988.

Country/continent, company 42% 55% Total Share (%) Plants

US, Archer Daniels Midland 500 1475 1975 24 4
US, A.E.Staley 392 1190 1582 19 4
US, Cargill 285 681 966 12 3
US, CPC International 188 324 512 6 3
US, American Fructose 116 277 393 5 2
US, Hubinger 113 154 267 3 1
US, Coors Biotech Products 23 70 93 1 1
US Total 1617 4171 5788 69 18
Canada 50 50 0.6 3
South America 53 43 96 1 6
Europe 420 420 5 18
Asia (mostly Japan) ≥1800 ≥22 40
Australia 4 4 0.05 1
Total 8330 100 86
Capacities are expressed in 1000 ton equivalents dry solids (adapted from White, 1992).

Because a relatively low-value product is produced, the cost-effectiveness of the enzyme

is crucial to success. Therefore Cultor have produced a regenerable immobilised enzyme.
Using this product the process is started using columns of support that have been only
 Case studies in the application of biocatalysts 114
partially saturated with enzyme. It is subsequently operated by the in-line addition of
fresh glucose isomerase soluble enzyme, which is absorbed by the remaining spare-
capacity of the enzyme columns. Thus a constant reactor productivity of HFCS can be
maintained for very long periods, despite the continuous loss of activity during use.
An alternative approach has been taken by Genencor, who have genetically engineered
their glucose isomerase so as to improve stability and thus productivity, but the attitude
of legislators and the public to genetically engineered products is an uncertain factor in
this.

4.2.6 Conclusions
A very big HFCS industry has developed despite competition from aspartame. Glucose
isomerase is easily the biggest selling immobilized enzyme in the world and so is a
significant biotech product in its own right. Large corn starch processing companies such
as Archer Daniel Midland and Staley together produce almost 107 ton/a of HFCS, mostly
in the USA (Table 4.3). HFCS production appears to have only become successful in
areas with available local supplies of starch, that are nett sugar importers, that have well
developed processed food and beverage industries as customers, and that offer fiscal
support to the industry. Figure 4.2 illustrates the dramatic rise in world HFCS production
since 1975.

Figure 4.2 The increase in world HFCS production 1975–1995 (est.).

Reprinted from Pedersen (1993) by courtesy of Marcel Dekker Inc.
 Applied biocatalysis 115
As a result of the success of HFCS the types of caloric sweeteners consumed has changed
considerably (Figure 4.3). If the impact of non-caloric high intensity sweeteners,
principally aspartame, is taken into account the importance of sucrose is diminished more
significantly.
The only other sugar isomerase to be used on a commercial scale is the isomaltulose
synthase used to isomerase sucrose into isomerase sucrose into isomaltulose (=
palatinose), which is finding use either as a non-cariogenic sugar (Mitsui Sugar Co.
Japan) or as a precursor of the bulking agent isomalt (Palatinit Co. Germany, subsidiary
of Südzucker Co.). Some other sugar isomerases show some potential, for instance for the
conversion of D-galactose into D-tagatose.
Alternative methods of producing fructose include the hydrolysis of inulin from
chicory using the enzyme inulinase. This is carried out on a modest scale by several sugar
companies in Western Europe. A second very inventive method was developed by the
Cetus Corp., but failed to become successful. This involved an ingenious biocatalytic
process for the manufacture of epoxides such as propylene oxide, via the intermediate
propylene chlorohydrin. D-Fructose was a side-product of this process, and was produced
as follows. Hydrogen peroxide was required for making the epoxide. This could be
generated in situ from glucose using glucose-2-oxidase, with D-arabino-2-hexosulose
produced as a side-product, which can be hydrogenated directly into D-fructose.

4.3 GLUCOSE PRODUCTION

The corn starch processing industry provides a very good example of how biocatalysts
can be successfully employed in bioprocessing on a very large commercial scale. Just
 Case studies in the application of biocatalysts 116

Figure 4.3 Per capita consumption of caloric sweeteners in the United States.
(From Hacking, 1986. Economic Aspects of Biotechnology.
Reproduced with kind permission from Cambridge University Press,
UK).

a glance at a flow-diagram of the various steps involved in maize starch processing

(Figure 4.4) indicates that a refinery-type operation is involved, in which all of the
components and side-products of the starch: oil, wheatgerm, protein (gluten), husk (bran)
and corn-steep liquor, as well as the starch, is extracted and used. Indeed the commercial
 Applied biocatalysis 117
viability of the entire process is dependent an obtaining income from such side-products,
as well as the starch and glucose and fructose syrups. Also whereas many of the products
have been produced for many years, the impact of new biotechnology has been two-fold.
Firstly to improve the production of such traditional products as glucose syrups by the
use of enzymes such as glucoamylase and pullulanase; and secondly to enable the

Figure 4.4 Corn wet milling.

 Case studies in the application of biocatalysts 118
production of new products such as high fructose syrups using glucose isomerase and
other new products such as cyclodextrins using cyclodextrin glucosyl transferase (Schenk
and Hebeda, 1992).
One of the factors that has allowed the rapid post-WW2 expansion of the corn starch
processing industry has been rising agricultural productivity resulting in a steady increase
in per capita cereal production, despite the rising world population (Figure 4.5). By the
year 2000, starch production estimates are for about 900×106 ton/a, about 75% from corn,
and with a (1990) value of 80 $/ton. Also over the

Figure 4.5 Average supply of cereal grain in kilogram per world inhabitant
over the period 1961–1979. From J.Mackay (1981). Cereal
production. In Cereals: A Renewable Resources. Theory and
Practice, pp. 5–23. Eds: Y.Pomeranz and L.Munck. St Paul, Minn,
USA: American Association of Cereal Chemists.

period that HFCSs have rapidly gained market share the price of corn has decreased, for
instance from 3.16 $/bushel in 1981 to 1.95 $/bushel in 1986, resulting in a starch costs
from 1.74 $/bushel to 0.79 $/bushel. This is especially significant as starch is the major
cost element in HFCS production (Table 4.2). Starch processing and therefore HFCS
production is concentrated in a relatively small number of companies in N. America,
 Applied biocatalysis 119

some of which operate several very large plants each producing over 105 ton/a HFCS
(Table 4.3), so that the proportion of corn produced that is used for HFCS production is
now very appreciable (Table 4.4). Analysis of production economics shows that the corn
raw material is the largest single cost with energy, labour, chemicals and enzyme costs
also significant. It is interesting that although the process is completely dependent on the
use of enzymes they only comprise ca. 5% of the total production costs (Table 4.2).
The role of the enzymes is three-fold. Firstly there is the use of very thermostable -
amylases to pre-thin the gelatinised starch, reducing its viscosity so that it can be easily
handled and further processed. This process is conceptually very similar to many other
commercial uses of hydrolases, especially proteases and glycosidases. Pre-thinning takes
place at 105°C and the thermostable -amylase from B. licheniformis actually has a
temperature optimum of almost 100°C.

Table 4.4 Corn utilisation for sweetener production in the U.S.a (in million bushels).

Products 1980/81 1989/90 1992/93 b

HFCS 160 368 404
Glucose and dextrose 182 193 216
Total sweetener 342 561 620
Total U.S. corn crop 7525 8770
Sweetener share of total crop (%) 7.5 7.1
a From: Sugar and Sweetener Outlook and Situation. U.S. Department of Agriculture, Economic
Research Service, June 1983, September 1992.
b Forecast

Secondly the pre-thinned starch is hydrolysed (saccharified) to glucose syrup using

glucoamylase (amyloglucosidase). In the last decade glucoamylase has been
supplemented with pullulanase in order to more completely debranch the -1–6 side-
chains of the starch since the glucoamylase cannot act on either these -1–6 branches or
the -1–4 bonds adjacent to the branches. Despite numerous attempts immobilised
glucoamylase has not been employed for saccharification of the mainstream pre-thinned
starch because despite their much greater activities the immobilised enzymes fail to
produce syrups of sufficiently high glucose content (≥96 DX is required, i.e. ≥96% of the
glucose units is at the reducing chain end).
Thirdly there is the conversion of the glucose syrups into HFCS using immobilised
glucose isomerase. Use of soluble enzyme is not possible because of its high cost, and
because it is an intracellular enzyme and is only stable when used still associated with its
parent cell. The activities and costs of these enzymes are given in Table 4.5. These starch
enzymes rank with some of the largest enzymes in world market.
 Case studies in the application of biocatalysts 120
An additional feature is that the glucose syrups can also be fermented to produce
ethanol for use in fuels. Obviously HFCS is a more valuable product, but power ethanol
production does generate significant revenues.

Table 4.5 Estimated activities and costs of enzymes involved in HFCS production.

-amylase glucoamylase glucose isomerase

Typical activity (U/mg protein) 3000 70 5
Turnovers required (mol/kg HFCS) 0.6 6 3
Typical used life time (min) 90 3000 200,000
Productivity1 (ton HFCS/kg protein) 450 35 330
Cost contribution to HFCS production2 1.7 3.5 3.8
(US$/ton HFCS)
Enzyme costs1 (US$/kg protein) 750 125 1250
1 Calculated from preceding rows.
2 H.S.Olsen, in: Rehm and Reed (1998), vol. 9, pp. 663–736. Total costs are 238 US$/ton HFCS.

4.4 -DECALACTONE

4.4.1 Background

• (R)- -Decalactone contributes much of the characteristic taste and aroma of peach and
many other flavours. Chemically synthesised -decalactone has been cheaply
available for a long time, but the consumer demand for naturally flavoured food and
beverages that arose in the mid 1980s created a strong demand for the (R)- -
decalactone isomer as a natural food flavour molecule. This definition of natural grade
required its production by entirely enzyme-based steps. In turn this led to the
development of a number of biotransformation processes to make natural -
decalactone.

4.4.2 Technical Features

• Manufacturing processes for (R)- -decalactone have been developed by a number of

flavour companies using yeast such as Yarrowia lipolytica and S. cerevisiae selected
for their ability to partially -oxidise the hydroxy mono-unsaturated fatty acid
 Applied biocatalysis 121
(ricinoleic acid) that occurs naturally as a major constituent of castor oil. The lead for
these processes was given by original Japanese research published in the early 1980s.
• -Decalactone production takes place under low pO2 conditions using pre-grown cells
to which castor oil or its hydrolysate is added. 4-Hydroxydecanoic acid is the main
microbial product, with -decalactone accumulating up to a concentration of 5–10%.
However de novo production from glucose is poor, only trace levels of -decalactone
being produced that are far from economic.
• Extensive downstream processing of the fermentation broth is carried out in order to
obtain -decalactone that is sufficiently pure for use as a flavour free of any off-
flavours, but development of a manufacturing process was relatively easy because
existing fermenters could be used together with an appropriate combination of standard
downstream unit operation.
• In some cases, alkylricinoleate is used as the precursor to reduce foaming problems.
• Yields of product are enhanced by acidification of the broth prior to downstream
processing by solvent extraction and distillation; so as to cause lactonisation of the 4-
hydroxydecanoic acid into -decalactone. Yields are also improved by minimisation
of by-product formation, especially of 3-hydroxy- -decalactone which has no value as
a flavour, but which is formed by lactonisation of the 3,4-dihydroxy decanoic acid that
accumulates once the fermentation becomes aerobic again after the -decalactone
production period has ceased. Attempts have also been made to produce some
additional -decalactone by converting this byproduct into 3,4-unsaturated -
decalactone, and then stereoselective reduction by S. cerevisiae into -decalactone.

4.4.3 Commercial Features

• Because -decalactone is a flavour chemical its quality is assessed by its taste quality
rather than by its chemical purity, and is used in only very low concentrations to
achieve the desired effects.
• Market acceptance was eased by the previous consumption of -decalactone from fruit
sources, and as a chemically synthesised flavour chemical.
• The value of -decalactone is enhanced by its use as a species characteristic flavour for
some flavours such as peach, and as an important flavour contributor to a variety of
other products.
• Initial sales prices were high, in the order of 4,−8,000 $/kg, but as the range of uses, and
the volume of sales of -decalactone increased, sales prices fell to ca 500$/kg. Thus,
in just over 10 years, -decalactone has moved from being very much a specialty
chemical to much more of a commodity status with several producers and much bigger
volumes of sales. However, even these lower prices are still far higher than for the
chemically synthesised -decalactone. (Gatfield, 1997; Maume and Cheetham, 1991).

4.4.4 Conclusions
 Case studies in the application of biocatalysts 122
Strain selection and bioprocess development of a process to meet the big and sustained
market for natural -decalactone has been created. By extension of this approach other
flavour lactones are also being manufactured, including -decalactone and others from
appropriate precursors such as either 11-hydroxypalmitic acid for the -decalactone, or
by a two step process from linoleic acid via its 13-hydroperoxide derivative.

4.5 6-AMINOPENICILLANIC ACID (SEMI-SYNTHETIC PENICILLINS)

4.5.1 Background
It is well known that penicillin G was first discovered by Fleming in 1932. Penicillins are
excellent active and broad spectrum therapeutic agents. However, the penicillins that can
be produced in high yields by fermentation (penicillins G and V) are relatively
ineffective, and also because many microorganisms have natural resistance or have
acquired tolerance by mutation. Also, penicillin G is not stable in the stomach and so has
to be administered by injection. Therefore a range of semi-synthetic penicillins have been
developed in which the phenylacetyl and phenoxyacetyl side-chains of penicillin G and V
are removed, and then replaced, e.g. by D-phenylglycine forming ampicillin, which was
the first semi-synthetic antibiotic introduced by Beechams and Bayer in ca. 1961; or by
p-D-hydroxyphenylglycine to form amoxicillin etc. This process involves hydrolysis of
penicillin G or V to form 6-aminopenicillinic acid (6-APA), followed by resynthesis
using a different side-chain. Chemical deacetylation to produce 6-APA is possible, and
was used originally, but the -lactam ring is labile and the process requires the use of
low temperatures (about −80°C), absolutely anhydrous conditions and the use of organic
solvents, making the process difficult and expensive.
 Applied biocatalysis 123

Figure 4.6 -Decalactone formation.

 Case studies in the application of biocatalysts 124

Figure 4.7 Conversion of penicillin into 6-APA.

4.5.2 Technical Advantages

• This single enzyme reaction replaces several chemical steps and the use of organic
solvent in the pre-existing process. A reaction temperature close to ambient can be
used.
• The microbial sources of penicillin amidases/acylases required for side-chain removal
were found and were quickly commercialised as whole-cell biocatalysts.
• The enzyme preparations are free of undesirable -lactamase which degrades the
penicillin.
• However, the retention of activity upon reuse was poor and undesirable side-activities
were present. Thus the step forward was to extract the enzyme and purify it free of
penicillin lactamase which destroys antibiotic activity, and to then immobilise the
enzyme in an active and stable form so as to allow efficient reuse (ca. 1000 cycles),
thereby making the catalyst cost contribution to overall process costs quite small.
• Phenylacetic acid produced by hydrolysis of the penicillin G can be recycled and fed
back as a precursor into the Penicillium fermentation.
• Genetic engineering techniques to improve penicillin amidase yields during
fermentation are now employed thereby reducing biocatalyst process costs.
• Yields of penicillin G and V precursors from fermentation have been increased ca. 104
fold over a 25 year period making available much cheaper precursors (see Figures
13.14 and 13.15).

4.5.3 Technical Disadvantages

• At pH<7.5 the equilibrium conversion is too low whereas at pH > 8 the substrate and
 Applied biocatalysis 125
product degrade too fast (see Chapter 10).
• The penicillin amidase reaction generates an acid product and so the pH control of
reactors is very important requiring the use of stirred tank reactors in series, rather than
packed-bed reactors which generally have a higher volumetric productivity. More
recently, recycle reactors have been developed.
• Penicillin acylase is an intracellular enzyme so that its isolation is relatively difficult
and expensive.
• Different enzymes, and therefore different source microorganisms are required for the
efficient hydrolysis of penicillin V and G.

4.5.4 Commercial Factors

• Production of semi-synthetic antibiotics is now a widely adopted and mature

technology.
• Up to 9×103 ton/a is produced world-wide by a number of companies, such as
Beechams, Toyo Jozo, Pfizer, Gist-brocades (now DSM) etc. The majority is produced
using penicillin G acylase, although some companies prefer to use the V acylase. This
large commercial potential was obviously a very big incentive to companies to develop
this technology.
• The market for the penicillin acylase is still quite moderate at ca. 8–10×106 $ in 1988.
The annual use for the immobilized enzyme was estimated in 1993 at 1000 kg/a.

4.5.5 Technical Details

Processes operate using immobilised penicillin V or G acylases derived from fungi such
as Bovista plumbea, and such as E. coli respectively. Following the discovery of
penicillin acylases commercial processes were developed very rapidly. Productivities of
up to 2,000 kg of 6-APA/kg immobilized enzyme are obtained with operating lifetimes in
excess of 103 h.

4.5.6 Conclusions

These processes have operated successfully for 25 years and now result in over 10×106 $
sales of penicillin derivatives world-wide. In many cases penicillin acylases have been
developed in-house by the user companies such as SmithKline Beecham.
For the analogous side-chain removal reaction required for semi-synthetic
cephalosporin manufacture more complicated processes have been developed. The side-
chain of cephalosporin C can be split off enzymatically, but only after its amino group has
been removed by a combined enzymatic and spontaneous reaction sequence (Matsumoto,
1993). Amongst other companies, Hoechst has replaced their chemical process by a two-
enzyme process, thus reducing the amount of waste from 31 ton to 0.3 ton per ton of 7-
 Case studies in the application of biocatalysts 126
ACA. A single-enzyme process might simplify this process in the near future.
The production of 7-ADCA, which also traditionally was carried out chemically, in
1989 has been replaced by Gist-brocades (now DSM) by an enzymatic process so as to
avoid the use of halogenated solvents.

Table 4.6 Fixed capital investment, 6-APA production facility.

Plant section Cost Fraction

(× 1000 $) (%)
Fermentation 7170 34.2
Clarification 1123 5.3
Extraction 895 4.3
Pen. G recovery 366 1.8
6-APA production 1045 5.0
Solvent recovery 2340 11.2
Offsites 5184 24.7
Buildings and general services 2833 13.5
Subtotal 20956 100.0
Contingency (20% of subtotal) 4191
Total fixed capital investment 25147
Reproduced from Harrison and Gibson (1984), with kind permission from Elsevier; copyright
Applied Science Publishers Ltd., UK,

Despite this considerable success, biocatalysts have not proved quite so successful in
other steps of the semi-synthetic antibiotic process. The coupling of side-chains to 6-
APA/7-ADCA using enzymes is technically possible, but most commercial processes still
carry out this step by chemistry because the enzymes that couple the side methyl esters or
amides also hydrolyse these activated side-chains as well as the desired product.
Depending on the selectivity of the enzyme, large excesses of the activated side-chains
are required to obtain sufficient yields. However, DSM-Deretil is now implementing a
process for cephalexin production using an enzymatic coupling of 7-ADCA to D-
phenylglycine, which is activated as an amide or ester.
Amidases have also proved useful in other processes, such as in the production of p-
hydroxyphenylglycine (see elsewhere in this chapter) and in the selective hydrolysis of
-lactams so as to produce enantiomerically pure intermediates for anti-AIDS
carbocyclic nucleoside drugs.
 Applied biocatalysis 127

4.5.7 Detailed Process Economics

The fermentation step to produce penicillin G/V is the major cost element in the overall
process to produce 6-APA. This is substantially due to the high cost of sterile engineering
(Table 4.6 and 4.7). Clarification, extraction and solvent recovery steps are also
significant, a reflection of the dilute and impure composition of fermentation broths. The
concentration of 6-APA in the final broth has a big effect on total process costs. Thus
increasing final 6-APA concentrations from 1.2–6.0% have been calculated to reduce
production costs by over 50% (Table 4.8). By contrast the 6-APA production step cost is
quite small, and is less that half the cost of the solvent recovery process (Table 4.6). The
costs of the immobilized enzyme is not insignificant; in a recent calculation it was
estimated at 2.5 $/kg 6-APA (Rasor and Tischer, 1998).

Table 4.7 Estimated operating costs for 6-APA production. March 1983 cost basis US $.

Annual Cost Product Cost

($) ($/kg 6-APA)
I. Variable Costs
A. Raw Materials
I. FERMENTATION MEDIA
Molasses 1474000 5.36
Cornsteep liquor 27500 0.10
Potassium phenylacetate 937750 3.41
Potassium monohydrogen phosphate 195250 0.71
Potassium dihydrogen phosphate 233750 0.85
Other media components 9625 0.04
2. PEN G RECOVERY AND CONVERSION TO 6APA
Potassium acetate 1509750 5.49
MIBK 332750 1.21
Penicillin acylase 973500 3.54
Filter aid 324500 1.18
Other raw materials 63250 0.23
B. Utilities
1. STEAM 924000 3.36
 Case studies in the application of biocatalysts 128

2. ELECTRICITY 1215500 4.42

3. COOLING WATER 101750 0.37
Total Variable Cost 8322875 30.27
II. Fixed Costs
A. Labour
1. OPERATING LABOUR 1727000 6.28
2. MAINIENANCE LABOUR 404250 1.47
3. OVERHEAD AND SUPERVISION 1067000 3.88
B. Maintenance Supplies 753500 2.74
C. Taxes and Insurance 503250 1.83
D. Depreciation 1677500 6.10
Total Fixed Cost 6132500 22.30
Total Operating Cost 14455375 52.57
Reproduced from Harrison and Gibson (1984) with kind permission from Elsevier. Copyright
Applied Science Publishers Ltd, UK.

4.6 ASPARTAME

4.6.1 Background
Aspartame is a high intensity dipeptide sweetener, ca. 200 times as sweet as sucrose. It
was originally developed by G.D.Searle & Co. prior to their acquisition by Monsanto.
Chemically synthesised aspartame has rapidly acquired a major share of the world high
intensity sweetener market, particularly in soft drinks. Until recently it has all been
supplied by a monopoly supplier, the Nutrasweet Corp (a Monsanto-Ajinomoto joint
venture) protected by product patents. Recently biocatalytic methods

Table 4.8 Process economics—conventional versus novel production of antibiotics.

Production costs ($ millions)

Conventional: Novel:
 Applied biocatalysis 129

1.2% final concentration 6.0% final concentration

Raw materials 30.2 14.6
Labour 7.3 5.6
Utilities 16.1 4.2
Equipment 19.0 7.8
Buildings 1.6 0.7
Direct costs 74.2 32.9
Overhead (60% of labour) 4.4 3.3
Total costs 78.6 36.2
Annual production (kg) 2,700,000 2,700,000
Unit costs ($/kg) 29.40 13.55
Data originally from: J.Leslie Glick, Genex Corporation

of production have been established, but Nutrasweet still retain 75% of the 1×109 $ high
intensity sweetener market.

4.6.2 Technical Factors (Oyama, 1992)

• An enzyme has been discovered that is extremely suitable for synthesising

benzyloxycarbonyl-aspartame, a precursor of aspartame (Asp-Phe-methyl ester).
• It is enantioselective and forms the peptide bond only with L-Phe-OMe and not with D-
Phe-OMe. This enables the use of cheaper racemic precursor (also because D-Phe-
OMe can be racemised, see Figure 4.8), whereas the competing chemical synthesis
process of Nutrasweet has to use more expensive L-Phe-OMe.
• It is regioselective and does not react with the -carboxylate of the aspartic acid and so
no bitter tasting -aspartame is formed.
• It is substrate selective as it lacks esterase activity and so does not hydrolyse the methyl
group, which is essential for sweetness, from the Phe-OMe.
• The aspartame precursor precipitates during the reaction, which pulls the reaction
equilibrium to completion.

4.6.3 Commercial Factors

• The profit potential of aspartame was a major factor in the takeover of Searle by
Monsanto. (Current aspartame sales are ca. 104 ton/a, equivalent to ca. 850 × 106 $/a).
• The success of the Nutrasweet Co. in establishing aspartame in the market. This has
 Case studies in the application of biocatalysts 130
occurred especially because it is very suitable for use in diet (low calorie) drinks and by
using novel marketing strategies, for instance by advertising

aspartame, an ingredient, to consumers in the hope that they will continue to buy only
aspartame containing products even after Nutrasweet patents expire and competing
sources of this sweetener appear.
• A joint venture was established between Tosoh, the inventors of the enzyme synthesis,
and DSM to commercialise the process in the Netherlands via the Holland Sweetener
Company joint venture company. This process now makes 2000 ton/a of aspartame
using the enzyme process.
• The expiry of Nutrasweet patents, first in Europe, then in the USA, has opened up the
market to new manufacturers. Nutrasweet were selling their aspartame at high profit
margins, exploiting their position as monopoly supplier, they were able to cut their
prices very substantially so as to compete with the Holland Sweetener Co. product.
• Record high EC tariffs have been imposed on imported (chemically synthesised)
aspartame.
• Bioeurope developed a Micrococcus caseolyticus strain that contains an enzyme that is
also specific for the amine group of the phenylalanine, thus eliminating the need to
protect the amine group of aspartic acid. This process was developed for Hoechst, who
have not commercialised it, perhaps because they have their own proprietary high
intensity sweetener (Acesulfame) which they have developed in-house. However,
Bioeurope have used this technology to produce N-L-malyl-L-tyrosine which is being
sold as a tan-accelerator for use in cosmetics.

4.6.4 Disadvantages

• The enzyme process is not significantly cheaper than the chemical method so that the
aspartame made by both methods are very competitive.
• Regulatory approval for aspartame produced by Nutrasweet was very long and
expensive, although this was no longer such a big problem for Holland Sweetener Co.

4.6.5 Technical Details

• The metalloprotease thermolysin, obtained from Bacillus thermoproteolyticus, a strain

of B. stearothermophilus, is used as a crude preparation in an aqueous medium. The
enzyme is recovered from the reaction mixture by ultrafiltration with a yield of >95%.
• An alternative process using immobilised thermolysin in column reactors with
substrates supplied in ethyl acetate—water as the solvent seems to be feasible.

4.6.6 Conclusions
Aspartame is manufactured using an enzyme-based route by the Holland Sweetener Co.
 Applied biocatalysis 131
This product is estimated to hold 30% of the European aspartame market. A new 2,000
ton/a plant was built by the end of 1993, which should provide sufficient capacity to
allow supply to Pepsi and Coca Cola who together represent ca. 70% of the total market
for aspartame.
In the future aspartame can expect to encounter competition from new high intensity
sweeteners such as sucralose which is produced by Tate & Lyle/Johnson & Johnson and
alitame (Pfizer), which have advantages such as even higher sweetness and, in the case of
sucralose, heat stability. In response Nutrasweet are busy developing a new very high
intensity sweetener (Sweetener 2000), which is reputed to be 10,000 times as sweet as
sucrose.
There are several other examples of enzyme peptide synthesis. The conversion of
porcine insulin into a human insulin precursor, by replacing the B chain C-terminal with
threonine, is still an important alternative to microbially produced human insulin.
Peptides that are produced from ethyl esters of L-amino acids by BioEurope are used as
ingredients of cosmetics.

4.7 CEFUROXIME

4.7.1 Background

Cephalosporins are -lactam antibiotics that block microbial cell wall synthesis. The
original cephalosporin, Cephalosporin C, has only weak antibiotic activity. Therefore
much more powerful second generation cephalosporins were developed by side-chain
modification. Modifications at C7 are most effective but modifications at position 3 are
also important so as to increase in vivo activity. Synthesis of the second generation
cephalosporin cefuroxime requires the replacement of the C3 acetoxy side-chain of the
precursor with a carbamate group. Chemical methods proceed via a hydroxylated
intermediate which causes problems due to a tendency to lactonise at low pHs. Therefore
development of a biocatalysis step was initiated in order to achieve selective reaction
under mild conditions.

4.7.2 Technical Factors

• A highly active structure was predicted based on a structure-function analysis of

existing molecules.
• A microorganism with the required activity was found as Glaxo were already
developing esterases to convert cephalosporin C into desacetyl cephalosporin C and
some of these proved able to produce cefuroxime and to be cost effective on a process
scale (Figure 4.9).
• Modification of the C3 side-chain can be carried out together with addition of the C7
 Case studies in the application of biocatalysts 132
furan side-chain, i.e. without isolation of the intermediate, thus reducing the number of
steps in the process.
• Cefuroxime could not be administered orally because it is highly ionised at
physiological pHs and so has only poor lipid solubility.

4.7.3 Commercial Factors

• The excellent biological activity of the cefuroxime, including a broad spectrum of

effective antibacterial activities and good -lactamase stability, has created a
Applied biocatalysis 133
 Case studies in the application of biocatalysts 134

Figure 4.8 Aspartame synthesis.

Figure 4.9 Conversion of cephalothin into cefuroxime.

substantial and continuing market for this antibiotic. For instance it is generally
more active than cephalothin against gram negative bacteria, including
enterobactericeae such as E. coli.
• Cefuroxime is only poorly absorbed from the gastrointestinal tract and so was
originally only an injectable product, but a commercially successful new orally-
administered cerufoxime ester product has been developed which has created a new
surge in sales. This acetoxyethyl pro-drug form of cefuroxime is rapidly hydrolysed in
the brush-border mucosal epithelial cells of the duodenum and small intestine, but
requires special formulation so as to ensure efficient dissolution following
investigation.

4.7.4 Technical Details

The process is carried out using whole cells of Rhodosponidium toruloides cultured in a
yeast-like form to replace the C3 acetoxy group with carbamate. Public domain sources
of information do not indicate how many, and which enzymes are involved; for instance
whether the same enzyme is involved in the reactions at the C3 and C7 positions.
 Applied biocatalysis 135

4.7.5 Conclusion
This process has been operated successfully by Glaxo on a commercial scale for over 17
years. An important advance has been the development of an acid stable form of
cefuroxime by esterification of the C4 ester group that has allowed the production of
products suitable for oral administration. This new product Cefuroxime axetil (trade
name Zinnat) has rapidly achieved market sales of ca. 300 $ in 1990–91

Figure 4.10 L-Aspartic acid synthesis.

compared with declining sales of ca. 50×106 $ for the original Cefuroxime, which now
has considerable completion, having stimulated the development of competing drugs.

4.8 L-ASPARTIC ACID

4.8.1 Background
Aspartic acid has been used in pharmaceuticals and foods etc. for some time, for instance
as an acidulant. More recently, demand has been stimulated because it is a component of
the dipeptide high intensity sweetener aspartame.

4.8.2 Technical Factors

• The enzyme acts stereoselectively to produce only the required L-isomer (Figure 4.10).
• Originally a fermentation process for the production of L-aspartic acid was established.
This was modified into an immobilised enzyme process, but since the extracted
enzyme is not very stable, an efficient continuous process was not possible. Therefore
an immobilised cell system was developed with a very long operational lifetime.
Another raw material for L-aspartic acid is maleic anhydride, which is first converted
 Case studies in the application of biocatalysts 136
into fumaric acid and then into the L-aspartic acid.
• Recovery and isolation of the L-aspartate is easy by crystallisation.

4.8.3 Commercial Factors

• The demand is for the natural L-isomer, rather that racemic DL-aspartate produced by
chemical methods.
• The chief advantage of the immobilized cell process is a significant reduction (60%) in
the cell production (biocatalyst) costs due to the improved efficacy of use. Costs of
labour, fuel, raw materials etc. were very similar to those of the comparable batchwise
process.
• Significant market demand for the product existed.

4.8.4 Technical Details

A single enzyme, L-aspartate ammonia lyase obtained from E. coli is used acting on
ammonium fumarate substrate. Little cell activity was lost upon immobilisation. Initially
polyacrylamide was used as the immobilisation medium, and later cross-linked K-
carrageenan was used, as higher operational life-times for the biocatalyst were obtained.
The immobilized cell activity is very stable with a half-life of 120 days, while achieving
95% conversion of substrate into product.

4.8.5 Conclusions
This process has been operated successfully by the Tanabe Seiyaku Co. in Japan since
1973. Similar processes have since been commercialised by other companies, such as the
Kyowa Hakko Co., often using different immobilisation methods such as polyurethane.
The same immobilized cell approach has also been used by Tanabe since 1974 in their
commercial process for the production of L-malic acid from fumarate using the hydratase
activity of Brevibacterium ammoniagenes cells.
Also Tanabe have extended the use of that aspartic acid producing process by using the
L-aspartic acid as the substrate for L-alanine production using P. dacunae cells with L-
aspartate decarboxylase activity. This process has been operating since 1982 using
sequential columns of immobilised E. coli and P. dacunae cells (Chibata, Tosa and
Takamatsu, 1987). Also, DL-aspartic acid can be used as the feed in this process. Then,
D-aspartic acid is obtained as an additional product, for which there is a modest demand.

4.9 D-P-HYDROXYPHENYLGLYCINE
 Applied biocatalysis 137

4.9.1 Background
D-p-Hydroxyphenylglycine is an important component of certain semi-synthetic
antibiotics such as the semi-synthetic cephalosporins cefadroxil and cefatrizine and the
semi-synthetic penicillin amoxicillin, with a combined world market in excess of 3×109
$/a. Synthesis was possible from DL-5-monosubstituted hydantoins (cyclic ureides of -
amino acids) provided that a selective D-hydantoinase could be found, which would be
competitive with chemical methods.

4.9.2 Technical Factors

• Enzyme activity on a D (non-natural) configuration, non-protein cyclic amino acid

derivative appears unlikely. However the D-hydantoinase reaction is very similar to
the dihydroxypyrimidase present in pyrimidine metabolism. The original hydantoinase
used was obtained from calf liver, but subsequently many active microorganisms were
found, particularly a strain of B. brevis. The resulting D-N-carbamoyl amino acid can
then be converted into product by treatment with nitrous acid (Figure 4.11).
 Case studies in the application of biocatalysts 138

Figure 4.11 Formation of D-p-hydroxyphenylglycine (D-HPG).

• Only the D-p-hydroxyphenylhydantoin is attacked by the enzyme. Unreacted L-

hydantoin racemises readily under mildly alkaline process conditions thus allowing
yields of product close to 100% to be obtained without extra recycling steps. In recent
years, enzymatic racemisation has been achieved, so that racemases can be added if
the racemisation becomes rate-limiting.
• Re-use of cells by using immobilised cell catalyst is possible.
• Agrobacterium radiobacter was found by the De Bi-sclavo Co (Siena, Italy) to process
both the D-hydantoinase and carbamoylase. Thus direct formation of product is
possible without the nitrous acid treatment required after the D-hydantoinase reaction.
The carbamoylase is especially high yielding because CO2 is formed as a product, thus
pulling the reaction towards completion.
• This process also depends on the efficient chemical synthesis of the DL-hydantoin
precursor.
 Applied biocatalysis 139
• Other biocatalytic methods of producing D-p-hydroxyphenylglycine have not proved
competitive, for instance transaminase based processes require glutamate to be
supplied. Others include the hydrolysis of N-acyl derivatives by acylase and amides by
aminopeptidase (DSM), the use of L-specific hydantoinases and immobilised subtilisin
for the resolution of D,L-2-acetamido-p-hydroxyphenylacetic acid methyl ester
(Bayer).

4.9.3 Commercial Factors

• A large commercial demand for D-p-hydroxyphenylglycine both for penicillin and

cephalosporin antibiotics had been established and provided the market-pull for the
development of a new process. Kanegafuchi were reported to be producing 300–700
ton of D-p-hydroxyphenylglycine by this method in the early 1980s. The present
market is estimated at 1000 ton/a.
• Despite the development of this process the market for hydantoinase is still quite small,
estimated to be ca. 2×106 $ in 1988, representing ca 14% of the products sales price.
• Several other companies have developed patented processes to D-
hydroxyphenylglycine. These include Bayer, DSM, SNAM-Progetti and Ajinomoto.
DSM has got access to several technologies and is now carrying out the hydantoinase
process at DSM Deretil in Spain.

4.9.4 Conclusions
The Kanegafuchi and Recordati companies have operated this process commercially since
the mid 1980s (Kanegafuchi began the process in 1983) using B. brevis and A.
radiobacter respectively, probably using immobilized cells. Ajinomoto are reported to
have a D-hydantoinase/D-carbamoylase process using a Pseudomonas strain. The
analogous process to produce phenylglycine required as a precursor of antibiotics such as
ampicillin has yet to be commercialised as this process is still too complex and expensive
when compared with the existing chemical process, which is well optimised and uses
depreciated equipment, to allow pharmaceutical manufacturers to switch to the biological
process.
Some other companies still use chemical resolution methods to obtain the D-p-
hydroxyphenylglycine, as well as the D-phenylglycine, so that competition between
microbially and chemically produced material takes place. Hydantoinases also have
considerable potential in other areas, for instance Ajinomoto have developed a method for
producing L-phenylalanine using a newly isolated strain of Pseudomonas. Continued
screening of microorganisms has found strains with hydantoinase suitable for the
production of L-glutamate, L-lysine, D or L-valine, L-leucine, L-tryptophan and L-
phenylalanine from various precursors, and so wider commercial applications for
hydantoinase technology may be expected in the future. D-Valine production for use as a
precursor of the insecticide fluvalinate is of particular interest. Kanegafuki have also used
 Case studies in the application of biocatalysts 140
hydantoinases ‘in-reverse’ to carry out stereoselective enzymic cyclisation to form a D-
hydantoin from the corresponding N-carbamoyl-α-alkylated amino acids.
Although fermentation processes to unnatural amino acids, such as D-amino acids as in
this example, are not generally possible, interesting exceptions exist. For instance the
Tanabe Seiyaku Co. have developed a fermentation process for D-alanine. This depends
on the formation of L-alanine from pyruvate, followed by

Figure 4.12 L-tert-Leucine synthesis.

D-alanine production by alanine racemase, and then preferential excretion of the D-

alanine across the cell membrane into the medium.

4.10 L-TERTIARY LEUCINE

4.10.1 Background
This unnatural acid is used as a chiral intermediate for the synthesis of a number of
products. Chemical asymmetric synthesis was very difficult and so the stereoselective
synthetic properties of enzymes were exploited to carry out a selective reduction reaction.
The stereoselective hydrolysis of protein amino acid esters had already been
commercialised by Tanabe in Japan using immobilised aminoacylase, and selective
reduction reactions using whole yeast cells are already used in a number of processes,
such as the selective reduction of the anti-cancer drug Coriolin.

4.10.2 Technical Features

• Methods of synthesising amino acids, and especially enantiomerically pure non-natural

amino acids, were developed using cofactor requiring redox enzymes (Figure 4.12).
• Good reactor productivities and cofactor recycling efficiencies with reuse of the
 Applied biocatalysis 141
enzyme and also the cofactor, were achieved by coupling the cofactor to polyethylene
glycol (PEG) to increase its size, so as to facilitate retention in ultrafiltration reactors.

4.10.3 Commercial Features

• The complicated technology used makes this approach economically uncompetitive

with fermentation processes for the production of naturally occurring amino acid, such
as L-phenylalanine, L-glutamic acid or L-lysine, for which large markets exist, and for
which significant economies of scale in production costs can thus be gained.
• Specialised and expensive 2-oxo acid precursors are often required.
• The market demand for the product is only low in volume (kg/a).

4.10.4 Technical Details (Bommarius, Schwarm and Drauz, 1998)

• PEG-derivatised cofactor together with leucine, dehydrogenase is used in combination

with the formate dehydrogenase required to regenerate the cofactor.
• L-tert-leucine is formed from trimethyl pyruvate, a 2-oxo acid.
• Reactor productivities of 640 g.l−1.day−1 and a cofactor recycling efficiency of 130,000
(mol product formed/mol cofactor used) are achieved making cofactor costs very low.
• complete conversion of the oxo acid avoids the requirement of ion exchange
chromatography in the downstream processing. A semi-continuous batch process has
been developed, which can be used campaign-wise.

4.10.5 Conclusions
Small-scale kilogram commercial production of L-tert-leucine is carried out by Degussa.
This approach is also used by some other companies to produce speciality amino acids
such as L-phenylalanine, L-methionine or L-valine. Degussa are also reported to use
ultrafiltration reactors containing aminoacylase for producing enantiopure amino acids by
resolution of the racemic N-acetyl derivative on a 240 ton/a scale. This process, although
successful, does not provide a markedly superior approach to the use of whole cells
supplied with glucose or lactate as a source of reducing equivalent for cofactor
regeneration. Other examples exist of the use of biocatalysts in the production of
unnatural amino acids useful as intermediates in chemical syntheses. These include the
DSM process for the resolution of amino acid amides using the aminopeptidase of
Mycobacterium neoaurum, for instance to make intermediates for D- -alkyl acid drugs.
Sophisticated membrane-based bioreactors have also been developed by Sepracor and
Bend Technology for the production of chiral pharmaceutical intermediates.
Recently, a lipase-catalysed kinetic resolution has been developed to produce L-tert-
leucine via a lactone of its N-benzoyl derivative. There is also reputed to be another
bioroute to L-tert-leucine using transaminase technology.
 Case studies in the application of biocatalysts 142

4.11 L-LYSINE

4.11.1 Background
Only a few amino acids are produced on a large scale. L-Glutamate (800,000 ton/ a) is
used as a flavour. DL-Methionine is used as a feed additive (300,000 ton/a).
L-Lysine is an essential amino acid and is used in very large quantities to supplement
human foods and animal feeds so as to improve their nutritional quality. Efficient
fermentation for its production have been developed in Japan. An alternative production
process method involves first the chemical synthesis of DL- -amino- -

Figure 4.13 L-Lysine synthesis.

caprolactam and then the selective hydrolysis of the L- -amino-6-caprolactam to give L-

lysine (Figure 4.13). This contrasts with the fermentation methods in which the last steps
in the pathways leading to the production of lysine involve either the decarboxylation of
meso , -diaminopimelic acid, or proceed via -amino adipic acid (Nakayama, 1985).

4.11.2 Technical Factors

• Microorganisms with both the required activities were discovered that surprisingly
acted with 100% specificity on the isomer of an unnatural substrate, and that did not
further metabolise the lysine product.
• Fortuitously simultaneous selective reaction and racemisation reactions can be achieved
by using two different microorganisms with lactamase and lactam racemase activities
respectively, since both have the same pH optimum (pH 8–9). Thus quantitative yields
of L-lysine are obtained, rather than the 50% yields usual in resolution processes. In
 Applied biocatalysis 143
addition extra process steps are eliminated and increased production costs avoided.
• DL-Amino acid resolution processes for various amino acids had already been
successfully pioneered by the Tanabe Seiyabu Co. using immobilised A. oryzae
aminoacylase acting on DL-N-acetyl acids. This step was first carried out as a
continuous immobilized enzyme process in 1969.

4.11.3 Commercial Factors

• Economies of scale of production are a significant advantage, ca. 300,000 ton/ a of

lysine are produced with demand growing at 7% per year. In 1982 this was only 40,000
ton/a, and Toray (Japan) was reported to produce 10% of this via DL- -amino- -
caprolactam. This compound is easily available to Toray.
• The competition is with fermentation processes which rely on the over-production of L-
lysine from cheap glucose sources by Corynebacterium glutamicum strains used on a
500 m3 scale and producing ca. 120 g/l lysine. In some cases the lysine is used in the
form of a spray-dried fermentation broth containing lysine sulphate.
• Animal feeds are a major market, especially for monogastric animals. This is because
their nutritional requirements are high and the lysine content of traditional feed such as
soy or maize is low. A supplementation of feeds with individual essential amino acids
is necessary because a high protein contents lead to excessive manure production,
especially by pigs.

4.11.4 Technical Details

This process uses cell suspensions of Cryptococcus laurentii which possess high L- -
amino- -caprolactamase activity, together with Achromobacter oboe cells with high -
amino- -caprolactam activity.

4.11.5 Conclusions
Although this enzymatic process fills only a niche in the L-lysine market, it is a
successful example of a general method for amino acid resolution. It has some superior
features compared to the Tanabe L-aminoacylase approach. The L-lysine can be extended
to non-protein amino acids such as the use of P. putida aminopeptidase to resolve DL-
homophenylalanine to produce precursors for the anti-hypertensive drug Enalapril. A
similar approach has also been used for the production of L-cysteine from DL-2-amino-
2-thiazoline-4-carboxylate using Sarcina lucea, which is remarkable in that both
isomers form L-cysteine.
Recently, Tanabe has developed a process for the production of D-lysine from L-lysine
by successive chemical racemisation and microbial degradation of the remaining L-lysine.
 Case studies in the application of biocatalysts 144

4.12 L-PHENYLALANINE

4.12.1 Background
Traditional commercial requirements for L-phenylalanine have been small (less than 50
ton/a) and had been satisfied by the use of aminoacylase to resolve chemically
synthesised DL-N-acetylphenylalanine. However with the advent of aspartame as a high
intensity sweetener a very big derived demand for L-phenylalanine was generated. As a
result a number of companies began to develop bioconversion and fermentation processes
to produce L-phenylalanine.

4.12.2 Technical Factors

• A range of suitable precursors were available; such as trans-cinnamic acid,

phenylpyruvic acid, -acetamidocinnamic acid and benzylhydantoin (Figure 4.14).
• Many microbial sources of enzymes suitable for converting them into L-phenylalanine
were discovered.

Figure 4.14 Enzymatic L-phenylalanine synthesis routes.

 Applied biocatalysis 145

4.12.3 Commercial Factors

• Due to the great success of aspartame the demand for L-phenylalanine is very big
(about 6,000 ton/a currently) and prices have fallen considerably with increased
volumes of production.
• Production of L-phenylalanine proved to be an attractive target for new ‘startup’
biotechnology companies as well as established companies. However, this opportunity
disappeared once the aspartame manufacturers’ own L-phenylalanine production plant
came on stream.
• The Nutrasweet© Corp. are now selling L-phenylalanine, rather than just using it in-
house. They quote its potential as a chiral intermediate for the production of rennin
inhibitors, HIV protease inhibitors, taxol and other pharmaceuticals.

4.12.4 Disadvantages

• Precursor costs are high by comparison with simple cheap fermentation nutrients, and
are potentially subject to problems in supply.
• Well developed and large scale fermentation processes to produce several amino acids
such as lysine, had already been developed. This know-how was then used as a basis
to develop very cost-effective L-phenylalanine fermentation processes.
• Demand for L-phenylalanine was exclusively from the Monsanto-Ajinomoto joint-
venture Nutrasweet© Company (the first manufacturer of aspartame).
• The aspartame process of the Holland Sweetener Co. can use DL-phenylalanine,
because their enzyme is L-specific and the D-phenylalanine can be racemised in a
recycle.

4.12.5 Technical Details

Several biotransformation processes were developed. As an example Allelix used
screening to isolate a Corynebacterium capable of converting -acetamidocinnamic acid
into L-phenylalanine. Lactate was added to the media to stimulate NADH regeneration.
The key enzymes involved were cloned to increase copy numbers, the cells were used in
immobilized form and phenylalanine recovery was facilitated by precipitation techniques.
This approach was also explored by the Tanabe Co. in Japan. By contrast Genex used the
phenylalanine ammonia lyase activity of Rhodococcus rubra to convert trans-cinnamic
acid.

4.12.6 Conclusions

Several bioconversion processes were developed to a production scale. These included

the use of the precursors trans-cinnamic acid by Genex, -acetamidocinnamic acid by
 Case studies in the application of biocatalysts 146

Allelix and Tanabe, and phenylpyruvic acid by Purification Engineering. This process,
for instance, was operated successfully on a 600 ton/a scale and the Genex process was
optimised to a stage at which the biocatalyst represented only 6% of product costs.
However, none of these processes were finally adopted. The fermentation approach to
producing L-phenylalanine proved ultimately successful. This was because a high
yielding fermentation process was developed using an over-producing strain giving ca 25
g.l−1 of phenylalanine, and also because the Nutrasweet company preferred to backwards
integrate into the production of the raw materials for aspartame and so obviously
preferred their own in house fermentation process to third party processes. Also, the
Holland Sweetener enzyme process for aspartame manufacture selectively uses the
phenylalanine methyl ester in racemic (DL) mixtures, which is cheaper than the L-
phenylalanine ester that has to be used by the non-selective chemical synthesis process.
Although fermentation has proved to be the preferred method for producing most L-
amino acids, biotransformation is preferred for unnatural amino acids, for instance L-tert-
leucine (see elsewhere in this chapter) and also for the production of D-homophenyl
alanine as a precursor of the ACE inhibitor drugs such as Enalapril (Merck, Sharp &
Dohme). In the future a substantial demand for D-alanine as a precursor for Alitame,
Pfizer’s potential new high intensity sweetener, may emerge. Similarly D-valine may
become useful as a precursor of the pyrethroid insecticide Fluvalinate.

Figure 4.15 Kinetic resolution of (RS)-2-chloropropanoic acid.

4.13 (S)-2-CHLOROPROPANOIC ACID

4.13.1 Background (see Taylor, 1998)

Chlorinated phenoxypropanoic acid herbicides have been very widely used following the
introduction of products such as Dalapon in the 1950s. More recently Mecoprop and
 Applied biocatalysis 147

related molecules have been used for post-emergent control of broad leafed weeds and
Fluazifop and Fenoxaprop for the post-emergent control of grass weeds at low dose
levels. Original products were all racemic (RS) mixtures, but only the (R)-isomer has
herbicide activity. Therefore an enantioselective manufacturing process for these
herbicides is highly desirable on environmental grounds, thus creating a demand for (S)-
2-chloropropanoic acid, the key intermediate in their synthesis (Figure 4.15). A single
isomer intermediate would also reduce the cost of synthesis and allow greater flexibility
in the formulation of the end product herbicide.

4.13.2 Technical Factors

• The ICI (now Zeneca) team had a good background in benzene-cis-glycol

biotransformations.
• Bacterial strains were isolated that could grow on R-CPA [(R)-2-chloropropanoic acid],
but not on S-CPA, because although both isomers were dehalogenated by different R
and S specific dehalogenase isozymes, the (S)-lactic acid formed from the S-CPA
could not be metabolised.
• A chiral chloroacetic acid was found to be a good inducer of the required R-CPA
dehalogenase and good growth of cells was obtained in minimal media containing
cheap carbon sources.
• The R-CPA dehalogenase proved active on substrates containing up to 1 M CPA and
with complete dehalogenation of the R substrate to produce S-CPA with an
enantiomeric excess of over 98%.
• Chemical mutagenesis produced a strain in which the S-CPA dehalogenase had been
completely eliminated so that cheaper whole P. putida cells could be used for the
bioconversion, rather than more expensive extracted enzyme.
• Care had to be taken that free chloride ions didn’t corrode the stainless steel of the
fermenters.
• Cells with elevated levels of R-CPA dehalogenase were produced by cloning the P.
putida gene for the enzyme and expressing it in E. coli. The recombinant strain can be
grown in continuous culture and achieved dramatic improvements in productivity.
• Stereoselectivity is high because the enzyme acts directly at the chiral centre of the
molecule, making dechlorination superior to other methods of resolving the racemate,
such as by selective esterification of the R-isomer followed by separation of the
required (S)-isomer, and then hydrolysis and racemisation of the R-isomer; or by the
selective hydrolysis of octyl (R)-2-chloropropionate, and then separation and
hydrolysis of the required (S)-isomer.
• Spray-dried cells have a stable enzyme activity enabling easy storage and transport of
the dehalogenase biocatalyst.
• Isolation of product from the lactic acid occurs by solvent extraction.
 Case studies in the application of biocatalysts 148

4.13.3 Commercial Factors

• A wide range of herbicide products exist, all of which require the same (S)-2-
chloropropanoic acid precursor, especially the Zeneca Fusilade product.
• Sales of stereochemically pure herbicides were rising indicating that they give
competitive market advantage over competitors still selling racemic products.
• (S)-Lactic acid is produced as a side-product.
• The technology developed was sufficiently novel to allow strong patents to be obtained.
• Other processes to (S)-2-chloropropanoic acid have been developed based on lipase
technology, and on fermentation of glucose to (R)-lactic acid and subsequent chemical
conversion.
• A collaboration with A.H.Marks Ltd. helped to solve substrate quality issues.

4.13.4 Technical Details

Whole cells are used in stirred tanks with pH control, producing (S)-2-chloropropanoic
acid in 50% yield from the racemate (0.3 M) with an enantiomeric excess of over 95%.
This approach was selected in preference to other methods of resolution such as acylation
of the racemate and then stereoselective hydrolysis. The dehalogenase enzyme is specific
for substrates with a carboxyl group and a 2-chloro or bromo substituent. No cofactor or
metal ion is required and reaction involves an inversion of configuration.

4.13.5 Conclusions
The first few tonnes of the (S)-2-chloropropanoic acid were made in 1985–86, and the
first production begun in 1987 of 500–1000 ton/a quantities for customers in

Figure 4.16 Kinetic resolution of (RS)-glycidyl butyrate.

the USA and Japan as well Europe. The first production using the genetically engineered
biocatalyst began in 1991.
A full-scale production plant has been built by Zeneca (formerly ICI) at Huddersfield,
UK with a capacity of 2,000 ton/a. The material produced is probably all intended for
 Applied biocatalysis 149
internal use to make herbicides, thus increasing the profit margins possible, by selling the
final products rather than just the intermediate. Zeneca also plans to extend this
technology to the production of other stereochemically pure chemicals such as (S)-2,3
dibromopropionate and 2-hydroxybutyrate. Other applications of these microorganisms
are being explored for use in waste-treatment.

4.14 GLYCIDYL BUTANOATE

4.14.1 Background
C-3 synthons are important in the manufacture of a number of products. These
intermediates include (R) and (S)-glycidyl derivatives. Small scale chemical synthesis
from D or L-serines and other raw materials has been carried out, but all these routes
have problems, such as the use of expensive substrates or production of polluting wastes.
Therefore various biocatalytic methods for producing enantiopure glycidyl butanoate
have been attempted (Elferink et al., 1991).

4.14.2 Technical Factors

• This is one process in which the initial research was carried out in an academic lab.
Whitesides and co-workers showed that resolution by stereoselective enzyme
hydrolysis of racemic (+/−) glycidyl butanoate was possible (Figure 4.16).
• The process allows both (+) and (−) material to be obtained in high yields.
• The process was successfully optimised to minimise substrate and enzyme costs and to
allow operation at high substrate concentrations.

4.14.3 Commercial Factors

• Glycidyl butanoate is used as a precursor of a range of different products such as (+)-

brefelden A, (S)- -blocker drugs such as (S)-timolol (Merck), antiviral compounds
and other products.
• There is an increasing trend towards the production of enantiomerically pure products
so as to minimise side-effects of drugs.
• Both (R)-(−) glycidyl butanoate and (R)-(+)-glycidol are required for commercial
products.
• A competing technology is the Sharpless asymmetric epoxidation, which uses chiral
titanium complexes as the catalyst. Arco uses this for production of (R)-glycidol and
other epoxy alcohols in commercial quantities.
 Case studies in the application of biocatalysts 150

4.14.4 Technical Details

Several enzymatic processes have been developed. The DSM-Andeno method uses
porcine pancreatic lipase, which selectively hydrolyses the (R)-(+) glycidol with purities
of at least 97% enantiomeric excess, suitable for subsequent chemical processing to
produce enantiomerically pure products.

4.14.5 Conclusions
A method based on the use of PPL has been commercialised by DSM-Andeno and also
other companies. Another manufacturing process involving an epoxide intermediate is
the lipase-catalysed resolution of the methyl ester of p-methoxyphenylglycidate, which is
an intermediate in the process of Diltiazem, a cardiovascular drug.
Other similar lipase/esterase resolution processes have been developed such as the use
of Bacillus thai esterase to produce the substituted propanoic acids that are precursors of
non-steroidal anti-inflammatory drugs, such as naproxen and ibuprofen etc., and the
formation of chiral amines by Celgene. Other methods start from prochiral precursors and
have the advantage that enantioselective synthesis allows the production of particular
isomers in yields approaching 100%, rather than the 50% yields characteristic of
resolution processes. For instance Hoechst have patented the production of enantiomers
using Pseudomonas fluorescens lipase to either acylate diols or hydrolyse diacetate
esters.
Lipase-catalysed esterification instead of ester hydrolysis is also being used, a.o. by
Glaxo Wellcome for a kinetic resolution that yields (1S, 2S)-trans-2-methoxycyclo-
hexanol, a key intermediate in the synthesis of tricyclic -lactam antibiotics.
One of the best examples of commercial use of lipases is for the production of esters
such as isopropylmyristate for cosmetic use, which is currently carried out by Unichema
on a scale of several thousands ton/a.

4.15 OXAMNIQUINE

4.15.1 Background
Oxamniquine is a very effective anti-schiztosomidal agent used in the treatment of
bilharzia and related conditions. Chemical synthesis was possible, except for the final
step which required the hydroxylation of a methyl group attached to an aromatic ring.
Therefore biological approaches were examined (Figure 4.17).
 Applied biocatalysis 151

Figure 4.17 Oxamniquine synthesis.

4.15.2 Technical Factors

• A microorganism was found by Pfizer with the required activity, that was not inhibited
by substrate or product, and that gave a high yield of easily extracted product.
• The microorganism used was an Aspergillus strain, and Pfizer already had con-
siderable experience using Aspergillus for citric acid production and other uses.

4.15.3 Commercial Factors

• Structure-function analysis of other drugs performances indicated the structure of the

target molecule (oxamniquine).
• At the time of introduction (late 1960s and early 1970s) alternative drugs had been
found to have toxic side-effects.
• Oxamniquine had good anti-parasite activity, especially versus South American strains,
and that was relatively free of side effects.

4.15.4 Technical Details

No efficient chemical procedure for the hydroxylation could be developed. The process
was operated using Aspergillus sclerotiorum used as whole cells with controlled addition
of precursor. Use of immobilised cells or cell extracts was tested but never developed
(Hinge, 1990).

4.15.5 Conclusions
Oxamniquine was produced by Pfizer on a commercial scale through the 1970s in their
Brazilian plant close to the markets for this drug. It was successful, but did not make ‘big
money’ and is now only produced occasionally as it has become superseded by more
effective drugs.
Several hydroxylation biotransformations have been commercialised, such as steroid
hydroxylation, e.g. the 11- -hydroxylation of progesterone by Rhizopus arrhizus, 6-
 Case studies in the application of biocatalysts 152
hydroxylation of nicotinic acid by Pseudomonas acidovorans and the conversion of -
butyrobetaine into L-carnitine by an Agrobacterium strain. The hydroxylation of a methyl
group, analogous to that carried out in the oxamniquine process, is carried out in the
synthesis of the useful intermediate 3-hydroxyisobutyric acid by Candida rugosa. Also,
the production of dihydrocatechols, HPOPS and pravastatin involve hydroxylations (see
Table 4.1).

Figure 4.18 L-Ephedrine synthesis via phenylacetylcarbinol.

4.16 L-PHENYLACETYL CARBINOL (L-EPHEDRINE PRECURSOR)

4.16.1 Background
Ephedrine was originally isolated as the active agent present in plant extracts used in
ancient Chinese medicine for respiratory ailments. As long ago as 1921 the formation of
optically active phenylacetyl carbinol (PAC) from benzaldehyde and pyruvate by brewers
yeast and cell-free yeast extracts was reported. The PAC can then be reductively
animated to produce optically active L-ephedrine (Figure 4.18). L-Ephedrine is widely
used in the treatment of asthma and hay fever as a bronchodilating agent and
decongestant.

4.16.2 Technical Factors (Rogers, Shin and Wang, 1997)

• In the plants of origin ephedrine is formed from L-phenylalanine by a pathway

involving side-chain shortening, addition of a C2N unit and then methylation by S-
adenosyl methionine.
• The reaction has been shown to be carried out by a pyruvate decarboxylase and
involves thiamine pyrophosphate in the formation of ‘activated’ acetaldehyde from
pyruvate, which then condenses with benzaldehyde. Evidently, pyruvate
decarboxylase, a crucial enzyme for ethanol biosynthesis, is used in an unnatural way
 Applied biocatalysis 153
for acyloin condensation.
• The reaction is high yielding, can be carried out using a cheap and readily available
biocatalyst, and did not require sterile conditions.
• Reaction also takes place with substituted aromatic aldehydes, in particular the product
of the reaction with 3-hydroxybenzaldehyde was converted into an ephedrine analogue.
• Benzaldehyde and acetaldehyde react to form benzoic acid side-product, therefore
benzaldehyde addition to the reaction is controlled so as to minimise benzyl alcohol
production and maximise PAC yields.
• Benzaldehyde resistant Saccharomyces mutants were developed.

4.16.3 Commercial Factors

• L-Ephedrine is a well accepted and widely used drug so that markets are significant.
• PAC production form benzaldehyde by yeast is still operated on a large scale by Knoll
(BASF, Germany) and Malladi Drugs (India). Knoll supply about one third of the
world market of ephedrine in competition with Ganes Chemicals Inc. who process
ephedrine from natural Chinese sources, which also covers about one third of the world
market. Now Roussel Corp. have entered the market with ephedrine synthesised by
Malladi Drugs & Pharmaceuticals (Bombay, India) and converted by Wyckoff
Chemicals (Michigan, USA). Nowadays about 1000 ton/ a R-PAC is being produced.
• PAC production was reported to be carried out on an 10 m3 reactor scale using 200 kg
of baker’s yeast.

4.16.4 Conclusions
More efficient chemical procedures were developed and as a result the yeast-based
process is now no longer in such extensive commercial operation.

4.17 ACRYLAMIDE

4.17.1 Background
Acrylamide is required in very large quantities as the pre-polymer of the polyacrylamide
that is very widely used in polymer and flocculant applications. The chemical
manufacture of acrylamide has been established for a long time. The original process
involved treatment of acrylonitrile with sulphuric acid at 90°C. More recently processes
have been introduced that require the use of copper catalysts and high temperatures (80–
140°C), but result in the formation of large quantities of toxic waste, including HCN. The
expensive copper catalyst used is difficult to regenerate. In addition the chemical process
produces acrylamide that requires considerable purification, for instance because the
 Case studies in the application of biocatalysts 154
acrylamide tends to polymerise under the high temperatures used, thereby increasing
production costs.

4.17.2 Technical Factors

• Nitto Chemicals had already developed a microbial process to eliminate acrylonitrile

from the factory waste streams.
• A microorganism was discovered that converts acrylonitrile into acrylamide at ‘low’
temperatures, thus reducing energy costs (Figure 4.19).
• The microorganism used has a high endogenous nitrile hydratase ratio when urea was
used as an inducer in the presence of cobalt ions. (The nitrilase is undesirable as it
converts the acrylamide further into acrylic acid).

Figure 4.19 Synthesis of acrylamide.

• The amount of nitrile hydratase in the cell-free extracts accounts for more that 50% of
the total soluble protein.
• Quantitative conversion of acrylonitrile is achieved so that separation of acrylamide
from unreacted acrylonitrile is no longer necessary. A key advantage of this process
appears to be that because less impurities are produced by the microbial process,
polyacrylamide of higher molecular weight can be manufactured so that lower in-use
concentrations can be used, creating considerable savings for the end-user.
• The new biocatalytic process was quite easy to integrate into existing manufacturing
process.
• Despite the high productivity of their original commercial strains Nitto have continued
to search for improved strains, for instance see US 5 153 858.

4.17.3 Commercial Factors

• A very large world market of ca. 200,000 ton/a, at a current market price of about 2
$.kg−1, exists, sufficient to provide a substantial market for a new process and to
enable significant economies of scale in production costs to be achieved.
• Acrylamide is a freely traded bulk chemical for which even small reductions in
 Applied biocatalysis 155
production costs can give significant market advantage. Major producers include Dow,
Nalco, American Cyanamid, Mitsubishi and Mitsui Toatsu.
• The bioprocess is much less polluting, making the process cheaper to operate in areas
where pollution control measures are rigorous and therefore expensive to implement.
• The process is novel and can therefore be protected by patenting.

4.17.4 Technical Details

Pseudomonas chloraphis cells were used first, and more recently Rhodococcus
rhodochrus J1. Cells are immobilised in polyacrylamide particles and used in column
reactors operated at below 10°C. The acrylamide is produced in 100% yield, and is so
pure that polymerisation inhibitors have to be added to prevent spontaneous polymerisa-
tion. Both acrylonitrile and acrylamide inhibit the nitrile hydratase; the nitrile hydratase is
extremely stable. Therefore acrylonitrile is fed to maintain a level of 6% resulting in the
accumulation of acrylamide of 66% (w/v), after which is it simply decolourised and
concentrated (Yamada and Kobayashi, 1996).

4.17.5 Conclusions
This is one of the very few examples of a successful biotransformation process for a
commodity chemical. Nitto Chemicals (Japan) are now reported to operate this process on
a 30,000 ton/a scale using immobilised cells produced in-house. Related enzymes have
been proposed for a variety of applications, but the only one to be attempted on a large
scale is the ICI process for the detoxification of cyanide containing waste. Because of
their broad substrate specificities nitrile converting enzymes are also being extensively
evaluated for a variety of bioconversions, such as to produce the precursors of
caprolactam used to make nylon, and for production of nicotinic acid and nicotinamide
(vitamin B6). The latter compound is now being produced commercially using a nitrile
hydratase. The first enantioselective nitrile degrading enzyme was found in a strain of
Acinetobacter by workers at Asahi Co., Japan) and this S-selective nitrilase has been used
to form the analgesic drug (S)-(+)-ibuprofen from its nitrile precursor.
A new development is a bioprocess for the hydrolysis of acrylonitrile to ammonium
acrylate, which is a key component of polymers used in products as diverse as paints,
dyes, cosmetics, plastics, papers and even disposable nappies. The big advantage of the
bioprocess is that acrylonitrile hydrolysis would otherwise be very energy intensive.
Another notable new advance is the development of hydroxynitrile lyases for the
synthesis of enantiomerically active aromatic and aliphatic cyanohydrins. For instance, an
S-specific hydroxynitrile lyase has been obtained from Hevea brasiliensis and the
resulting (S)-cyanohydrin can be used to obtain both hydroxy acids and aminoalcohols.
 Case studies in the application of biocatalysts 156

4.18 6-HYDROXYNICOTINIC ACID

4.18.1 Background
Substituted nicotinic acid derivatives are useful in the synthesis of pesticides and
pharmaceuticals as specific inhibitors of NAD and/or NADP dependent enzymes. 6-
Hydroxynicotinic acid is a very useful intermediate in the synthesis of such inhibitors.

4.18.2 Technical Factors

• The chemical synthesis of 6-substituted nicotinic acids is difficult and together with the
separation of by-products makes this an expensive operation.
• Selective and high yielding microbial hydroxylation is possible (Figure 4.20).

Figure 4.20 6-Hydroxynicotinic acid synthesis.

4.18.3 Commercial Factors

• A range of different products is possible, all from the key 6-hydroxynicotinic acid
precursor; for instance via chlorination of the hydroxyl or carboxyl groups of the 6-
hydroxynicotinic acid.

4.18.4 Technical Details

Achromobacter xylosoxydans has been used to carry out the selective hydroxylation in
high yield, using the enzyme which catalyses the first step in nicotinic acid degradation.
The whole-cell biotransformation process has been scaled-up to 12 m3, which is
sufficient to produce high purity 6-hydroxynicotinic acid for the subsequent chemical
reactions. The hydroxylation is oxygen requiring, so that oxygen transfer rate-limits the
reaction.

4.18.5 Conclusions
Lonza (Switzerland) have successfully developed this process. They also run several
other microbial oxidation processes on a commercial scale. One uses a bacterial strain in
 Applied biocatalysis 157

a>100 ton/a, 50 m3 fed batch process to convert -butyrobetaine into L-carnitine, which
is used in pharmaceuticals as a thyroid inhibitor, as a slimming aid and in sports foods. Its
physiological role is to facilitate the transfer of fatty acids across cell membranes.
Another bioprocess of Lonza is the oxidation of dimethyl-pyrazine into 5-methyl-2-
pyrazinecarboxylic acid.

4.19 GENERAL CONCLUSIONS

Ideas for new R&D projects arise from market demands for new and/or improved
products and also to exploit newly developed technologies. The former effect, (market
pull) is generally essential for success.
Biocatalysts can be used to make new products, improved processes for pre-existing
products, and to make new and improved intermediates for use in other processes. The
general objective is to produce the best quality product at a minimum production cost by
developing processes in which the yield and purity of the product and its functional
characteristics are maximised. Therefore complex problems involving not just scientific
questions but also legal and commercial issues must be solved if success is to be
achieved. These goals are achieved both by adapting existing technical knowledge and by
generating new know-how by research. A key objective is to optimise the overall process
and not just the scientifically most interesting biotransformation step. This is because the
success or failure of R&D projects is ultimately judged not so much by how scientifically
creative they are, but rather in terms of their ‘bottom line’: the sales and profits they earn
for the companies operating them.
Despite the large number of biocatalysis projects, both successful and unsuccessful,
carried out world-wide, an important conclusion would appear to be that there are few
simple and generally applicable rules for success. Each of the case-studies presented in
this chapter represent individual solutions to particular problems and little widely
applicable technology would appear to have been developed, for instance even now,
despite a lot of work over 30 years very few generally applicable and easily scaled-up
immobilisation techniques have been developed.
However, some important technical and commercial lessons can be learnt, particularly
by reference to Tables 4.9 and 4.10 in which important scientific and commercial features
of the case-studies have been collated and limiting factors identified. The following
commercial and technical topics are suggested as being important.

4.19.1 Commercial (see Table 4.11)

• In order to achieve commercial success a variety of technical and commercial ‘hurdles’

must be passed. Failure in one important respect probably results in the failure of the
whole enterprise, irrespective of how well the product or process shapes-up in other
respects.
 Case studies in the application of biocatalysts 158
• A new technology, such as the use of biocatalysts, provides new entry opportunities for
companies wishing to move into new product areas.
• Biocatalysts tend to be used for the production of speciality (performance) chemical
products, rather than for the manufacture of bulk (commodity) chemicals. Good
examples of commodity chemicals produced by biocatalysts do exist, such as
acrylamide and glucose syrups. Other products from bioreactions, such as HFCS’s and
6-APA, have, due to their success, also acquired the status of commodity chemicals.
• Many applications of biocatalysts are now mature and so growth of only a few percent
per year can be expected. Substantial growth can still be expected in new areas such as
the use of biocatalysts in bioorganic syntheses or in paper and pulp treatment.
• Three main challenges can be recognised.

1) Development of a new process that will compete successfully with either an already
well-established chemical process, or with
2) A well-established first-generation fermentation or enzyme process.
3) The development of new processes for new functional molecules or materials, in
which case a major problem is the costs and time required to achieve regulatory
approval for the product.

Table 4.9 A comparison of technical and commercial features of the case-study

processes.

Product Fructose - 6-APA Aspartame Cefuroxime

syrup Decalactone precursor
Cells (C) or C&E C E E C
Enzyme (E)
used
Free (F) or I F I F F
immobilised
(I) biocatalyst
used
Number of 1 several 1 1 a
enzymes
involved
Type of Isomerase -Oxidase Acylase Peptidase Transferase
enzyme (s)
Batch (B) or C B B a B
continuous (C)
process
Source of Streptomyces yeast strains Bovista B. Rhodosporidium
 Applied biocatalysis 159

Enzyme (s) murinus and (e.g. Yarrowia plumbea, stearothermophilis toruloides

others lipolytica, S. E.coli, a.o.
cerevisiae)
Company ADM, Staley IFF, Quest, etc. Gist- Holland Glaxo
etc. brocades Sweetener Co. Wellcome
(DSM),
etc.
Company USA etc. USA, Ireland, Many Netherlands UK
etc.
Application of Sweetener Flavour Antibiotic High intensity Antibiotic
product component sweetener
New product Y N Y N Y
(Yes/No; Y/N)
Substantially Y N N Y Y
new process
(Yes/No)
Product used E I I E I
predominantly
internally (I)
or is sold on
(E)
Commercial S S S S S
success (S) or
failure (F)
Date process 1967 late 1980s 1961 mid 1980s ca. 1978
started
(a) has not been made public

Product L- D-p- L-tert L-Lysine (S)-2-

Aspartic Hydroxyphenylglycine Leucine Chloro
acid acid
Cells (C) or C C E E C
Enzyme (E)
used
Free (F) or I a Ib F F
immobilised
(I) biocatalyst
used
 Case studies in the application of biocatalysts 160

Number of 1 1 or 2 2 2 1
enzymes
involved
Type of Ammonia Hydantoinase Dehydrogenases Lactamase and Dehalog
enzyme (s) lyase (+ carbamoylase) racemase
Batch (B) or C C B/C a B
continuous
(C) process
Source of E. coli B. brevis and B. cereus, Cryptococcus P. putida
Enzyme (s) Agrobacterium Candida bodinii laurentii and
radiobacter Achromobacter
obae
Company Tanabe Kanegafuki, Recordati, Degussa Toray Zeneca
Seiyaku DSM Industries
a.o.
Company Japan a.o. Singapore, Italy, Spain Germany Japan UK
Application of Acidulant Antibiotic component Chiral Food/feed Herbicid
product and intermediate additive
component
of
aspartame
New product N N Y N N
(Yes/No;
Y/N)
Substantially Y Y Y N Y
new process
(Yes/No)
Product used I/E I I – I
predominantly
internally (I)
or is sold on
(E)
Commercial S S c S S
success (S) or
failure (F)
Date process 1973 1983 late 1980s 1980s late 1980
started
(b) Enzymes are retained by used of an ultrafiltration reactor
(c) The success or failure of recent processes cannot be stated yet
 Applied biocatalysis 161

Product (R)- Oxamniquine L-PAC Acrylamide 6-

Glycidyl Hydroxynicotinic
butanoate acid
Cells (C) or E C C C C
Enzyme (E)
used
Free (F) or F F I I F
immobilised
(I) biocatalyst
used
Number of 1 1 1 I 1
enzymes
involved
Type of Lipase Hydroxylase Decarboxylase Nitrile Hydroxylase
enzyme (s) hydratase
Batch (B) or B B B C B
continuous
(C) process
Source of Pig A. sclerotiorum S. cerevisiae Rhodococcus Achromobacter
Enzyme (s) pancreas rhodochrous xylosooxydans
Company DSM Pfizer Knoll Nitto Lonza
Andeno
Company Netherlands Brazil USA Japan Switzerland
Application of Chiral Anti- Anti-asthma Prepolymer Chiral intermediate
product intermediate schiztosomiadal drug
drug
New product N Y Y N N
(Yes/No;
Y/N)
Substantially N Y Y Y Y
new process
(Yes/No)
Product used – I I E –
predominantly
internally (I)
or is sold on
(E)
Commercial S d S S c
success (S) or
 Case studies in the application of biocatalysts 162

failure (F)
Date process late 1980s late 1960s 1920s mid 1980s late 1980s
started
(d) These processes operated successfully for some time, but are no longer used now

Table 4.10 Factors limiting bioprocess development.

• Synthetic, rather than degradative, reactions are rarely possible.

• The required biocatalysts, for instance an enzyme with the right stereospecificity, is rarely
available ‘off-the-shelf’ and/ or at reasonable cost.
• Biocatalysts are often insufficiently stable to meet process and economic targets.
• Reactants are often not permeable to cells, and so react only slowly, or are unstable under the
conditions of operation.
• General scientific know-how about the biochemistry, microbiology etc. of the system being
studied is very often insufficient such that new basic research has to be done before proper
targeted development work can begin.
• Process intensity achieved is too low, and therefore processes are too expensive and logistically
difficult to operate.
• Product isolation and purification capabilities (downstream processing) are usually ineffective
and/or expensive so that the yield, quality and cost of product is poor.
• Scale-up of technology is difficult and/or expensive, for instance too many unit operations are
involved.
• The development of new technology often required the recruitment of people with the new
skills required for its research and development.
• Raw materials are not readily and cheaply available in the quantities and purities required for
industrial production.
• High capital costs are frequently involved in starting up the processes, especially when new
sophisticated technology has to be used. Therefore, existing equipment and sites should be
utilised as much as possible.
• Formulation of the active ingredients, so as to maximise its activity and stability etc., frequently
requires considerable effort.
• Regulatory and safety testing costs and delays can be excessive.
• Patent issues including infringements and the expiry of patent protection before R&D costs are
recouped, can create problems.
• When long R&D time-scales are necessary the resulting products can be superseded by
 Applied biocatalysis 163

different technologies producing different products but with similar functionalities.

• Market size and demand for the product can be too small to justify R&D costs.
• Profit margins achieved can be too small to justify R&D costs.
• Time-scales to commercial production can be too long.

Table 4.11 Some important commercial factors.

1. Significant market demand.

2. Substantial profit margins.
3. Economies of scale of production.
4. Availability of low cost raw materials.
5. Patentable (proprietary) technology.
6. Weak competition (including competition from other different products with the same or
overlapping functionalities.)
7. Ability to produce a range of products with different market uses from a single process.
8. Regulatory approval (for pharmaceutical and food products).

• Biotransformations expertise is spread throughout the industrialised countries and so

competition may arise from a wide range of countries.
• The high capital costs and lengthy construction times associated with the establishment
of sophisticated biotech processes can be a considerable entry barrier to companies
wishing to sell products produced using biocatalysts.
• When success is achieved, it has to be maintained in a commercial environment. Hence
even very successful products can become easily and rapidly superseded. Even when
success has been sustained over a long period of time adaptation and improvement has
been necessary in order to maintain competitive advantage. For instance, the glucose
isomerase and penicillin acylases used now are very much improved over those
originally commercialised 20 years ago. This is partially due to technical advances and
partially due to continuously increasing consumer and customer demands and
standards.
• A continuous search goes on for better and/or cheaper precursors and catalysts for large
turnover/highy profitable products. These may be derived from chemistry or
biotechnology. What is certain is that a new product is not required in order to
commercialise new biocatalysis technology: indeed the converse may actually be a
more attractive and lower risk business strategy. For instance both ascorbic acid and
ephedrine were introduced as commercial products produced using biotech stages
 Case studies in the application of biocatalysts 164
before WW2, but development of these manufacturing processes still goes on. Similarly
steroid biotransformations have been the basis of several commercially successful
processes but new approaches are still being explored, notably the pioneering work by
the Sumitomo Chemical Co. who have developed recombinant yeast strains with
improved C17- and C21 hydroxylase activities due to the expression of fused pairs of
cytochrome P450 C17- monooxygenase and reductase and cyt P450 C21
monooxygenase and reductase.
• In the ‘supply-chain’ from raw materials, through intermediates to final products (see
Figure 4.21), higher profit margins are often obtained by selling the final product to the
ultimate consumer, which may be householders, farmers, physicians, a health service,
or to a agrochemical, food or pharmaceutical products manufacturer. Thus it is often
preferable, especially for large companies, to concentrate on applying biotechnology to
the improved production of products for their own consumer products, rather than
selling intermediates to end-product manufacturers. A typical supply chain is
fermentation raw materials, to 6-aminopenicillanic acid or D-p-hydroxyphenylglycine,
to the antibiotic molecule amoxicillin, to the final product Clamoxil. Note that very
often a key intermediate is used as a ‘captive’, rather that being sold on a mercantile
basis, so as to maximise commercial benefits.
• Some of the products formed are functional ingredients in their own right, such as high
fructose syrups, cefuroxime etc. Others are intermediates that can be either freely
traded, such as acrylamide as a pre-polymer and 6-APA and p-hydroxyphenylglycine
as precursors of semi-synthetic antibiotics. Alternatively some intermediates tend to be
used in-house by the company producing them such as (S)-2-chloropropanoate and L-
tert-leucine.
 Applied biocatalysis 165

Figure 4.21 Examples of processes using different raw materials processing

technologies and producing products with different end uses.

• In some cases the development of a new product or ingredient generates a derived

demand for a precursor or a biocatalyst necessary to make it, such as the demand for
L-phenylalanine generated by aspartame, and the demand for glucose isomerase used
to make high fructose syrups. In many cases these derived demands are sufficiently
big to make materials such as L-phenylalanine and glucose isomerase important
commercial targets in their own right.
• Irrespective of the route of commercialisation, in most cases the development of
technology that is protectable by patenting is very important so that a proprietary
position can be established.
• Very many technical and commercial factors are important such as the utility of the
product (cost-benefit relationships), ease of scale-up, the productivity of the process
etc. Process design, for instance, involves a series of choices, such as the use of
isolated enzyme or intact microorganism, use of free or immobilised cell or enzyme,
use of mutant or genetically modified cell, or batch or continuous processing etc. Such
choices depend on other factors such as the availability and cost of precursors, product
purity required, intended scale of operation and existing skills and equipment available
within the Organisation.
• The size of the company involved is also very important. Obviously a project with only
relatively modest market and profit potential may be of considerable importance to a
small company, but would probably not prove attractive to a large multinational
company with much larger sales and profit criteria.
• Biotechnology is intrinsically interdisciplinary in nature. This makes the technology
very interesting. However, inter-disciplinary work tends to be expensive, and its
management is very challenging if successful R&D and commercial exploitation are to
be achieved.

4.19.2 Technical Points (see Table 4.12)

• The use of biocatalysts is most often successful when the target materials cannot be
easily produced by more established methods, such as chemical synthesis: especially
when regioselective and/or enantioselective reactions are necessary. Thus, although in
many processes, such as fructose and acrylamide synthesis, the biocatalyst carries out
the crucially important reaction in the process; biocatalysts are particularly effective
when integrated with other chemical steps, such as in the synthesis of aspartame etc.
so as to achieve easy selective reaction. This big advantage of biocatalysts is likely to
become increasingly important as the traditional sources of chiral molecules are
limited and already well exploited, whereas the supply of new naturally occurring
biocatalysts by selective isolation from nature is effectively unlimited.
• Higher value applications allow much more sophisticated technology to be used; the
 Case studies in the application of biocatalysts 166
technical options for use in the production of low margin products are quite limited.
• The effectiveness of a biotransformation process does not depend on whether the
biocatalyst acts on its normal physiological substrate, or on another non-naturally
occurring chemically synthesised molecule or material.

Table 4.12 Some important technical factors.

Factor Examples
1. Reactions possible that are not possible using HFCS, cefuroxime, L-tert-leucine, (S)-2-
chemistry. chloropropanoic acid, -decalactone
2. Specificity of reaction, including
a. substrate specificity aspartame
b. positional specificity oxamniquine, -decalactone
c. stereospecificity L-lysine, (S)-2-chloropropanoic acid
3. Allows milder process conditions e.g. 6-APA, acrylamide, L-PAC
temperature, sterility etc.
4. Reduces number of process steps required. p-hydroxyphenylglycine (esp. Recordati
process), L-lysine
5. Existing in-house know-how, esp. ease of oxamniquine, cefuroxime, L-aspartic acid,
integration into existing production facilities. acrylamide
6. Eliminates the need to use organic solvent in 6-APA
processing.
7. Readily available biocatalyst. L-PAC, glycidyl butanoate
8. Immobilisation of biocatalyst to allow its reuse L-aspartic acid, HFCS, L-tert-leucine
or continuous use.
9. Genetic engineering to improve HFCS, (S)-2-chloropropanoic acid*
biocatalyst/process.
10. Use of biocatalysts in combination with other aspartame, (S)-2-chloropropanoic acid, L-
separate chemical steps. PAC
* Most commodity biochemicals are now made using enzymes produced by gene cloning.

• Science of an appropriate level of sophistication should be employed that is consistent

with, for instance, convenient scale-up and easy marketing of the product.
• Virtually all of the successful developed processes utilise microbial enzymes or cells.
Plant and animal cells are slower growing, susceptible to contamination, have more
complex nutritional requirements, and are altogether more difficult to work with
 Applied biocatalysis 167
despite considerable progress has been made with some systems, such as the production
of -methyldigoxin by Digitalis lanata cells.
• A comparison of the various case-studies shows that there is often no pattern apparent
in the use of extracted enzyme rather intact cells, or the use of immobilised biocatalyst
rather than free biocatalyst. Chapter 5 gives a more detailed discussion on this point.
• In order to achieve the desired reactions a large variety of different microorganisms and
enzymes have been used; bacterial, yeast and fungal strains, and dehydrogenase,
isomerase, decarboxylase, acylase, protease, dehalogenase, and glycosidase enzymes
etc. Thus each process tends to use a different microorganism, either as the biocatalyst
or as a source of enzyme. Each strain requires individual research into media
development and fermentation optimisation etc. which is expensive.
• In most reactions only a single enzyme is involved. Some exceptions occur, either when
a microorganism possesses both the required activities, as is the case in the Recordati
process for D-p-hydroxyphenylglycine; or when two enzymes can be easily used under
the same conditions, such as in the L-lysine process from caprolactam. In this respect
the promise that genetic engineering can provide microbial strains with additional
cloned enzyme activities has been disappointing so far, as in the case of ascorbic acid.
However, the use of producer strains containing multiple gene copies or cloned genes
from another organism has become standard in the production of bulk industrial
enzymes such as acylases, proteases and lipases, in order to obtain more efficient
producer organisms.
• Down-stream processing, those steps in the process required for the isolation,
purification and concentration of the product, are absolutely critical if pure chemicals
are to be produced cost-effectively. Very often the DSP steps represent over 50% of
the total manufacturing costs.
• Usually the size of individual markets for speciality chemicals or the number of user
companies are too small to justify the development of new biocatalysts. Hence,
research often tends to be confined to new uses for existing biocatalysts, especially
those with a sufficiently broad substrate specificity.
• Development of new or adapted chemical reactions is often required so as to
successfully integrate chemical and biological steps into a process.
• Use of new biocatalysts sometimes allows previously unexploited raw materials to be
used.
• Most of the processes involve using a blocatalyst to carry out a selective degradation or
modification of the precursor(s). The industrial enzyme world market is dominated by
sales of cheap hydrolytic enzymes such as proteases and acylases. Only in rare cases,
such as aspartame syntheses, is the biocatalyst used to form a new bond and thereby
make bigger and more complex molecules. This contrasts with the power and
flexibility of synthetic organic chemistry. Similarly only one of the processes
described, L-tert-leucine, involves cofactor regeneration without the need to use whole
cells, despite the fact that most of the enzymes known to science require cofactors for
activity.
• The enzymes normally involved in synthetic reactions, such as carbon-carbon bond
 Case studies in the application of biocatalysts 168
formation are not very suitable for easy use in biotransformation reactions because i) they
are rarely secreted as extracellular enzymes, ii) their synthesis is tightly controlled and
linked to the primary metabolic needs of the cell, making over-production difficult, and
iii) they usually require activated coenzyme-linked substrates. Therefore fermentation
is still usually the preferred method of carrying out multi-step and synthetic reactions.
• Whereas soluble enzymes can often be used by technically relatively unsophisticated
users, immobilised biocatalysts require users with a much greater degree of skill and
higher grade facilities.
• The range of commercial applications that biocatalysts based processes can be used for
is currently limited, because despite the very large number of different enzymes that
are known to science, only a small proportion can be used easily and cheaply. That is
the range of possible applications is technically limited by the availability of suitable
catalysts. Hence the importance of enzymes with high stereo-selectivity, but with a
broad substrate specificity, and that are easy to use on an industrial scale so that the
same biocatalyst can be used in a number of different processes.
• The functional characteristics of a biocatalyst that are most important vary from
application to application. Thus in some cases long operational stabilities are most
important, whereas for other uses reaction selectivity or resistance to organic solvents
are most important.
• Improving process intensity to reduce costs and increase logistic convenience is vital.
Thus the use of high substrate concentrations and achieving high degrees of conversion
to product with little side-product formation is very important. The combining of
successive process steps, such as the use of extractive bioconversion approaches, is
promising; but so far the only good case-study of a commercial process using this
approach, even on a small-scale, is the L-tert-leucine process.
• A significant number of the processes employ continuous processing, usually by using
immobilised biocatalysts. This is in great contrast to the fermentation industry where
very few continuous commercial processes operate. Processes can either be dedicated,
in which the whole process is optimised towards production of a single valuable
product. Alternatively the process can take the form of a ‘refinery’-style process, such
as HFCS production, in which a complex raw material is processed to produce a range
of co-products all of which make an important contribution to the commercial success
of the overall process.
• Simply developing a highly productive biocatalyst is not sufficient. It is the productivity
of the overall process that determines success or failure.
• In most of the case studies quite well-defined and pure raw materials are used. This
prevents clogging of immobilised enzyme columns and microbial contamination, and
also simplifies the downstream processing.
• In most of the processes examined academic research has had comparatively little direct
influence, in-house company research having solved most of the problems. Notable
exceptions include the work of Yamada et al. on acrylamide, Lilly and coworkers on
semi-synthetic penicillins, Whitesides and his group on glycidyl butyrate, Slater on
chloropropanoic acid, and Wandrey and Kula’s groups on L-tert-leucine. Often there
 Applied biocatalysis 169
has to be a considerable emphasis on the use or adaptation of already established science,
rather than the carrying-out of new basic research.
• Also in most of the case studies the biocatalyst required is produced by the company
wishing to use it or by a company with which it has a close customer-supplier
relationship; rather than be supplied as a ‘commodity’ catalyst by a specialist supplier
company. This situation may of course change now that some new enzymes such as a
range of lipases and esterases are commercially available in bulk quantities and at
moderate prices from a variety of suppliers.
• Microbial cells with useful enzyme activities are not yet available commercially in the
same scale and variety that enzyme preparations are sold.
• In order to succeed a multi-disciplinary approach has to be taken involving scientists,
engineers and commercial experts with different and complimentary skills. In
particular an effective marketing input into the choice of R&D targets is essential.
• Usually a biocatalyst-based process is in competition with other approaches, for
instance enzyme resolution processes to produce single isomer products are in
competition with their isolation from natural sources and chemical asymmetric
synthesis processes.
In conclusion biocatalysts have acquired many useful applications in a variety of
different industries. These genuine commercial achievements are particularly
impressive when judged in comparison with a realistic view of technical progress in
other closely related fields. Thus one very large pharmaceutical company has
introduced only three new fermentations in the last 25 years. In the future big
advances are necessary in order to make substantial progress, for instance by
learning how to carry out synthetic rather than just degradative reactions, so that
effective competition with established chemical syntheses can be established.

4.20 ACKNOWLEDGEMENT

We would like to thank all those members of the Working Party and others, who have
helped to compile this chapter by contributing information. In particular Will v.d. Tweel
(DSM) who has been particularly helpful.

4.21 QUESTIONS

1) Compare and contrast the important features of a pharmaceutical and food orientated
process. Do the very different markets for the products made affect the type of
technology used?
2) Give examples and explanations of commercial biocatalysis projects in which (i)
positional specificity, & (ii) stereospecificity are very important.
3) Compare and contrast processes that (i) use whole cells and free enzyme, or (ii)
 Case studies in the application of biocatalysts 170
immobilised biocatalysts and free (non-immobilised) biocatalysts.
4) Can you identify differences in the technology employed in biocatalysis processes
depending on whether the process is designed to product a new product, or an
improved version of a pre-existing product?
5) Explain why such a diverse range of microorganisms have been used as the source
biocatalysts in these processes.
6) What do you consider to be the most important general advantages and disadvantages
of enzyme and cell biocatalysts when used in large-scale processes, as compared with
other production technologies such as chemistry?
7) Can you identify some of the chief problems that may prevent biocatalytic processes
from achieving commercial success, and even cause established processes to become
superseded?
8) Explain the importance of the following

a) economies of scale of production,

b) raw materials costs,
c) multiple uses for the product,
d) purity of product formed,
e) joint-venture agreements,
f) utility of side-products,
g) value of patents; on the success of products produced by biocatalysts-based
processes.

Illustrate your answer with examples from real-life commercial processes.

4.22 REFERENCES AND FURTHER READING

Further details and references to details of the processes mentioned in this chapter can be
found in the following:
Birkett, J.A. and Robinson C. (1990) Biotransformation of cephalosporins: an adjunct to
chemistry. In Real products from Industrial Biocatalysis , Royal Society of Chemistry.
Bommarius, A.S., Schwarm, M. and Drauz, K. (1998) J. Molecular Catalysis B,
Enzymatic , 5 , 1–11.
Cheetham, P.S.J. (1995) Principles of industrial biocatalysis and bioprocessing. The
applications of enzyme in industry . In Handbook of Enzyme Biotechnology , 3rd
edition, edited by A.Wiseman, pp. 83–234 and 419–552. London: Ellis Horwood.
Chibata, I., Tosa, T. and Takamatsu, S. (1987) Continuous L-alanine production using
two different immobilized cell preparations on an industrial scale. Methods in
Enzymology , 136, 472–478.
Collins, A.N., Sheldrake, G.N. and Crosby, J. (eds.) (1992) Chirality in Industry .
Chichester: John Wiley and Sons.
Collins, A.N., Sheldrake, G.N. and Crosby, J. (eds.) (1997) Chirality in Industry II .
Chichester: John Wiley and Sons.
 Applied biocatalysis 171
Copping, L.G., Martin, R.E., Pickett, J.A., Bucke, C. and Bunch, A.W. (eds.) (1990)
Opportunities in Biotransformations , SCI/Elsevier Applied Science.
Dingler, C., Ladner W., Krei, G.A., Cooper, B. and Hauer, B. (1996) Preparation of (R)-
2-(4-hydroxyphenoxy) propionic acid by biotransformation. Pesticide Science , 46, 33–
35.
Elferink, V.H.M., Breitgoff, D., Kloosterman, M., Kamphuis, J., van der Tweel, W.J.J.
and Meijer, E.M. (1991) Industrial developments and biocatalysis. Recueil des Travaux
Chimiques des Pays Bas , 110, 63–74.
Gatfield, I.L. (1997) Biotechnological production of flavour-active lactones. Advances in
Biochemical Engineering , 55, 221–238.
Glöckner, R. and Roduit, J.P. (1996) Industrial bioprocesses for the production of
substituted aromatic heterocycles. Chimia , 50, 413–415.
Gotfredsen, S.E., Ingvorsen, K., Yale, B. and Andersen, O. (1985) The scope of
biocatalysts in organic chemical processing. In Biocatalysts in Organic Synthesis ,
edited by J.Tramper, H.C.van der Plas and P.Linko, pp. 3–18. Amsterdam: Elsevier.
Hacking, A.J. (1986) In Economic Aspects of Biotechnology , edited by J.Baddiley et al.
UK: Cambridge University Press.
Harrison, F.G. and Gibson, E.D. (1984) Approaches for reducing the manufacturing costs
of 6-APA. Process Biochemistry , 19, 33–36.
Hinge, R.D. (1990) Large-scale manufacture of oxamniquine. In Real products from
Industrial Biocatalysis , Royal Society of Chemistry.
Jensen, V.J. and Rugh, S. (1987) Industrial scale production and application of
immobilised glucose isomerase. Methods in Enzymology , 136, 356–370.
Kamphuis, J., Boesten, W.H.J., Broxterman, P.B., Hermes, H.F.M., van Balken, J.A.M.,
Meijer, E.M. and Schoemaker, H.E. (1990) New developments in the chemo-
enzymatic production of amino acids. Advances in Biochemical Engineering , 42 ,
134–186.
Keller, R., Schlingmann, M. and Woernle, R. (1987) Production of 5’-ribonucleotides
using immobilized 5’-phosphodiesterase. Methods in Enzymology , 136, 517–522.
MacKay, J. (1981) Cereal Production in Cereals: A renewable resource. Theory and
Practice , edited by Y. Pomeranz and L.Munck, pp. 5–23. St Paul, Minn, USA:
American Association of Cereal Chemists.
Maume, K.A. and Cheetham, P.S.J. (1991) The production of -decalactone by
fermentation of castor oil. Biocatalysis , 5, 79–97.
Matsumoto, K. (1993) Production of 6-APA, 7-ACA and 7-ADCA by immobilized
penicillin and cephalosporin amidases. In: Tanaka et al., pp. 67–88.
Meyer, H.-P. (1991) Microbiology in the contemporary chemical industry. Biotech.
Forum Eur. , 8 , 602–606.
Morgan, B.P. and Moynihan, M.S. (1997) Steroids. In Kirk-Othmer encyclopedia of
Chemical Technology , edited by J.I.Kroschwitz and M.Howe-Grant (eds.), 4th ed., pp.
851–921. New York: John Wiley & Sons.
Nakayama, K. (1985) Lysine. In Comprehensive Biotechnology , edited by M.Moo-
Young, volume 3, pp. 607–620. Oxford: Pergamon Press.
Neidleman, S. (1991) Industrial Chemicals. In Biotechnology—The Science and the
Business , edited by V. Moses and R.E.Cape, pp. 17, 297–310. Harwood Academic.
Oyama, K. (1992) The industrial production of aspartame. In: Collins et al., pp. 238–247.
Pedersen, S. (1993) Industrial aspects of immobilised glucose isomerase. In: Tanaka et
 Case studies in the application of biocatalysts 172
al., pp. 185–208.
Polastro, E.T., Walker, A. and Teeuwen, H.W.A. (1989) Enzymes in the fine chemicals
industry: dreams and realities. Biotechnology , 7, 1238–1241.
Poulsen, P.B. (1984) Current applications of immobilised enzymes for manufacturing
purposes. In Biotech & Genetic Eng. Reviews , 1, 121–140.
Rasor, P. and Tischer, W. (1998) Process benefit of enzyme immobilisation. Paper
presented at “Advances in industrial biocatalysis” , InBio Europe ’98. Amsterdam:
Spring Innovations Ltd.
Rehm, H.J. and Reed, G. (eds.) (1998) Biotechnology , 2nd edition, vol. 8a
Biotransformations I. Weinheim: Wiley-VCH.
Rogers, P.L., Shin, H.S. and Wang, B. (1997) Biotransformation for L-ephedrine
production. Advances in Biochemical Engineering , 56, 33–59.
Schenck, F.W. and Hebeda, R.E. (editors) (1992) Starch hydrolysis products . New York:
VCH.
Sheldon, R.A. (1993) Chirotechnology: Industrial synthesis of optically active
compounds . New York: Marcel Dekker.
Stinson, S.C. (1992) Chiral Drugs. Chemical and Engineering News , Sept. issue, 46–77.
Stinson, S.C. (1996) Technological innovation thrives in fine chemicals industry.
Chemical and Engineering. News , July issue, 35–61.
Straathof, A.J.J. (1994) Industrially applied bioconversions of carbohydrates.
Carbohydrates in Europe , 10, May issue, 5–8.
Strohl, W.R. (editor) (1997) Biotechnology of antibiotics , 2nd ed. New York: Marcel
Dekker.
Syldatk, C., Laufer, A., Muller, R. and Hobe, H. (1990) Production of optically pure D
and L- -amino acids by bioconversion of D,L-5-monosubstituted hydantoin
derivatives. Advances in Biochemical Engineering , 41, 28–75.
Tanaka, A., Tosa, T. and Kobayashi, T. (eds.) (1993) Industrial application of
immobilized biocatalysts . New York: Marcel Dekker Inc.
Taylor, S.C. (1998) Singled out for success. Chemistry in Britain , May , 23–25.
Tramper, J. (1985) Immobilising biocatalyst for use in syntheses. Trends in Biotech. , 3,
45–50.
White, J.S. (1992) Fructose syrup: Production, properties, and applications. In Starch
hydrolysis products , edited by F.W.Schenck and R.E.Hebeda, pp. 177–199. New
York: VCH.
Yamada, H. and Kobayashi, M. (1996) Nitrile hydratase and its applications to industrial
production of acrylamide. Bioscience Biotechnology and Biochemistry , 60, 1391–
1400.
Yamada, H. and Shimizu, S. (1985) Microbial Enzymes as catalysts for the synthesis of
biologically useful compounds. In Biocatalysts in Organic Synthesis , edited by
J.Tramper, H.C.van der Plas and P.Linko, pp. 19–40. Amsterdam: Elsevier.
 5.
HOW TO GET THE BIOCATALYST
MARCEL G.WUBBOLTS1, CHRISTOPHER BUCKE3 and STANISLAW
BIELECKI2
1 DSM Research, P.O. Box 18, NL-6160 MD Geleen, The Netherlands. e-mail:
[email protected]
2 Institute of Technical Biochemistry, Lodz Technical University,
Stefanowskiego str. 4/10, PL-90–924 Lodz, Poland, e-mail: stanb@ck-
sg.p.lodz.pl
3 School of Biological and Health Sciences, University of Westminster, 115 New
Cavendish Street, London WIM 8JS, United Kingdom. e-mail:
[email protected]

ABSTRACT
Finding a catalyst, be it an isolated enzyme or a whole cell system, for a
biocatalyst application is frequently not an easy task. With a desired product in
mind, one has to consider variously suited starting compounds, which are
preferably commercially available or cheaply synthesized, and the enzymatic
steps leading to the products of interest. Furthermore, the chemo-, regio-and
stereoselectivity of the biocatalyst towards functional groups of the starting
compounds should be taken into account.
The quest for biocatalysts usually starts by screening among existing
enzymes, which are either commercially available or described in the open or
patent literature, for the application one has in mind. If suitable biocatalysts are
not available, screening for novel activities is required. This entails the search
for the desired activity using samples from nature that are likely to contain it,
using techniques of various degrees of sophistication that range from simple
growth tests to elaborate high-throughput screening programs with for instance
chromogenic substrates. As a result of screening for novel activities, the
biocatalyst has to be prepared from strains containing the desired enzyme (s)
and, depending on the application, purification of enzymes may be required to
actually apply the biocatalyst.
This chapter describes some of the common methods that are used to obtain a
biocatalyst, illustrated by examples, and includes the description of means to
produce enough of the active material in a form suited for application in an
actual process.
 Applied biocatalysis 174

5.1 BIOCATALYST SELECTION

5.1.1 Introduction
Any organization which does not attempt to innovate is moribund and consequently,
industry is a constantly seeking new or improved biocatalysts. Improvements will be
sought in response to commercial needs, some fundamental, such as the requirement for a
totally new catalyst, some more detailed, such as a desire to change specific parameters in
an existing process.
In the food and feed industries, the production of glucose isomerase provides an
excellent example of the location and development of a totally new catalyst, whereas the
search for variants of a-amylases which do no longer require calcium ions for stability is
an example of a desire for detailed change (to minimize the need for deionisation of
glucose syrups which are to be isomerized using glucose isomerase). The nature of the
process in which the new biocatalyst is to be used must be considered at the very outset of
research so that time is not wasted in exploring blind alleys. For instance, a sialidase from
Vibrio cholerae can be considered seriously for use in the synthesis of novel
oligosaccharides as potential therapeutic agents (Theim and Sauerbrei, 1991) but enzymes
from the same organism would never be acceptable for use in the food industry for safety
reasons.
In the fine chemicals industry, new developments frequently deal with novel
compounds, requiring totally new routes, or with more efficient routes to existing
compounds, requiring more subtle changes in an existing procedure. Investigation of
chemical as well as biological alternatives (‘route scouting’) is required to thoroughly
evaluate production alternatives ranging from complete chemical synthesis to biocatalytic
methods. Biocatalysis procedures can include biotransformations using an enzymatic step,
bioconversions with (living) cells that may catalyze multiple steps starting from a
precursor molecule or complete fermentative routes using simple sugars resulting in the
formation of a desired product. Both in the food and feed as well as in the fine chemicals
industries and in any other industrial area where enzymes can be applied, biocatalytic
routes will only be selected for development to a commercial process, when they are most
appropriate based on economical, environmental and regulatory aspects.
Similar to standard organic synthetic methods, various biocatalytic processes can lead
to the product of interest as illustrated by the various routes to D- and L-amino acids
illustrated in Figure 5.1. At the ‘route scouting’ stage, all alternatives have to be
thoroughly evaluated. A crucial factor that influences the selection of a specific
biocatalytic route and thus the nature the biocatalyst we are looking for is process
economics. For instance, the availability and cost price of the starting compound (s) can
force such a decision. In addition, when only 50% of the starting compound is converted
 How to get the biocatalyst 175
to the desired product (resolution), the importance of the cost price of this starting
compound is even more obvious. In the illustrated example (Figure 5.1), routes for amino
acid production that suffer from the latter disadvantage are the aminopeptidase, the
acylase and the esterase routes. Racemization steps and recycle loops have to be
introduced to attain proper product yields. Biocatalytic resolutions with spontaneous in
situ racemization (hydantoinase/carbamoylase) as well as routes that start from prochiral
compounds such as -keto acids (amino transferase and amino acid dehydrogenase
routes) are more advantageous in this respect.
Crucial to the development of a biocatalytic route is furthermore the selection of a
readily available biocatalyst with sufficient activity, selectivity and stability, which
preferably should be commercially available. Other factors that play an important role are
the scale at which the process should be run and the time it takes to develop such a
biocatalytical process.
 Applied biocatalysis 176

Figure 5.1 Various routes that can be applied to manufacture chiral D- and L-
amino acids.
 How to get the biocatalyst 177

Table 5.1 Biocatalyst classification and use1.

E.C. Name Total Available Utility for Processing Utility for Synthetic
class 2 Applications 3 Application 3
1 Oxidoreductases 861 90 ++ +++
2 Transferases 970 90 ± +
3 Hydrolases 898 125 +++ +++
4 Lyases 316 35 ± ++
5 Isomerases 140 6 +++ ±
6 Ligases 118 5 + +
1 Adapted from Faber (1997)
2 Number of uniques E.C. entries in LIGAND, April 1998 (source:
http://www.genome.adjp/kegg/docs/enzyme_stat.html)
3 Ranging from very usefull (+++) to of little use (±)

5.1.2 Availability of Enzymes or Cells for Biocatalysis

The commercial availability of enzymes or whole cell biocatalysts for a desired
biotransformation is frequently a limiting factor for commercial application of
biocatalysts. Enzymes that are cheaply available are typically used in detergents,
processing of food, feed and textiles, as well as in waste management applications. Most
of these are hydrolytic enzymes, but also isomerases (e.g. glucose isomerase) and
oxidoreductases are used on industrial scale (Table 5.1).
Enzymes that are suited for application in biocatalysis are mostly hydrolases, but also
oxidoreductases, lyases and, to a lesser extent, transferases are useful. Obviously, the
focus of bulk enzyme producers is different from the main interests of those who want to
apply enzymes in biocatalytic applications. Fortunately, a growing number of companies
has become active in the field of enzyme production for biotransformations and by now a
large number of enzymes suited for biotransformations has become commercially
available (Table 5.1).
Consideration of the sources of the enzymes used commercially on a significant scale
indicates that they come from a relatively small number of organisms, with fungi
predominating over bacteria, and rather few enzymes being isolated from plant and
animal sources (Table 5.2). In more detail, it is clear that industry favors Aspergillus and
 Applied biocatalysis 178
Bacillus species as producers of enzymes but that for some purposes, such as the
isomerization of glucose, actinomycete-derived enzymes are preferred because they have
proved to be particularly effective. The reasons for these preferences are simple: many of
the large scale uses of enzymes are connected with the food industry and derive directly
from traditional food fermentation processes in which Aspergillus and/or Bacillus species
predominate. (It is, perhaps, surprising that very few commercial enzymes are obtained
from the lactic-acid bacteria, which play such an important part in traditional food
preservation processes). Those organisms are Generally Recognized As Safe (GRAS) and
enzyme-producing companies are now very familiar with the growth characteristics of
the organisms and the isolation and purification (if necessary) of biocatalysts from them.
Inevitably, therefore, there will

Table 5.2 Organisms most frequently used for industrial enzyme production1.

Species Number 2 Species Number 2

Fungi Bacteria
Aspergillus niger 37 Actinoplanes missouriensis 1
Aspergillus oryzae 8 Bacillus licheniformis 8
recombinant Aspergillus oryzae 5 Bacillus subtilis 20
Aspergillus spp. 13 Bacillus spp. 11
Chaetomium sp. 1 recombinant Bacillus subtilis 5
Humicola insolens 4 Clostridium sp. 1
Mucor spp. (inc. Rhizomucor) 6 Klebsiella spp. 2
Penicillium spp 8 Lactobacillus tormentum 1
Rhizopus spp 9 Microbacterium arborescens 1
Trichoderma longibracheatum 10 Micrococcus lysodeikticus 1
Trichoderma spp 5 Pseudomonas fluoroscens 1
Yeasts 4 Streptomyces spp 3
1 Data from AMANO, Gist-Brocades, GENENCOR, NAGASE, NOVO-Nordisk, PRIMALCO,
RHONE POULENC and SOLVAY as described in Godfrey & West 1996.
2 Corrected for multiple uses of the same enzyme, i.e. in formulations for different purposes.

be pressure from biocatalyst production organizations to select familiar microbes as

sources of new enzymes. Already there are examples of the use of the familiar microbes
as genetically manipulated hosts for the production of enzymes such as chymosin. It does
not follow that those seeking new biocatalysts should ignore the potential of previously
 How to get the biocatalyst 179
unexplored organisms as sources of new enzymes. As applied biocatalysis advances and
industry becomes increasing familiar, and comfortable, with the use of enzymes,
increasingly adventurous processes are being considered which require more specialized
biocatalysts. There is rapid growth of the number of chemo-enzymatic syntheses being
developed, requiring biocatalysts, which are robust and yet have great specificity. For the
robustness, which is required if a biocatalyst is to be used in polar organic solvents, the
structurally-rigid enzymes from thermophiles are attractive: for synthetic specificity,
enzymes catalyzing secondary product synthesis in plants may be required. Both types of
enzyme may be expensive to isolate from their natural sources but now the potential to
clone the genes encoding such enzymes into a suitable hosts, such as Aspergillus and
Bacillus species or Escherichia coli means that any enzyme may be produced relatively
inexpensively. Nevertheless, these enzymes remain more costly than detergent and
processing enzymes, due to the scale at which they are produced. Consequently, the
contribution of the biocatalyst costs to the production costs can be substantial.
Conversely the value of the product (and potential profits) allow more advanced
techniques to be used in the development of new biocatalysts.
Whole cell biocatalysts are more difficult to obtain and apply than enzymes. Although
numerous strain collections exist that can supply strains with known biotransformation
activities (see paragraph 5.2.1), one has to be able to cultivate the micro-organisms or
perform expensive toll-fermentations elsewhere to obtain enough

Table 5.3 Biocatalyst types and their potential for use in industrial applications.

Type of Biocatalyst Utility for Industrial Application 1 Commercially available

Enzymes
Pure enzymes +++ Yes
Crude enzymes ++++ Yes
Whole cells
Wild type strains ++ Yes
Recombinant strains +++ No
Abzymes ± No
Ribozymes – No
1Ranging from very useful (++++) to of no current use (−)

biocatalyst for an industrial process. This is especially true for recombinant whole cell
biocatalysts, which can be genetically engineered to have increased enzyme levels and
produce less by-products, but which are not easily acquired.
Novel biocatalysts, such as catalytic antibodies or ribozymes, that have recently been
 Applied biocatalysis 180
developed are for as yet far away from industrial application (see paragraph 5.4.2 and
5.4.3), mostly because these catalysts are not available in amounts or at costs that are
commercially interesting (see Table 5.3) and because they do not have obvious
advantages over conventional biocatalysts.

5.1.3 Catalysis by (Isolated) Enzymes or Whole Cells

Once the appropriate starting compound has been selected and the number of routes to
the product has been reduced, questions concerning the biocatalyst arise. It largely
depends on the reaction that has to be catalyzed whether it is possible to use cell-free
extracts, or whether it is necessary to use purified enzyme preparations (see paragraph
5.6) or even growing or resting whole cells. Some of the criteria that play a role in
deciding what catalyst to choose have been listed in Table 5.4.
Biocatalysts based on hydrolases (E.C. class 3, Table 5.2) are mostly used as (purified)
enzymes since they are cofactor independent, since these preparations are commercially
available and because a number of hydrolases can be applied in organic solvents.
Oxidoreductases (E.C. class 1) however, are relatively complex enzymes, which require
cofactors and frequently consist of more than one protein component. Thus, despite the
fact that efficient cofactor regeneration systems for NADH based on formate
dehydrogenase (FDH) have been developed (Bradshaw et al., 1992; Chenault &
Whitesides, 1987; Wandrey & Bossow, 1986, chapter 10) and that also an NADPH
dependent FDH has been isolated (Klyushnichenko, Tishkov & Kula, 1997), these
enzymes are still mostly used as whole-cell biocatalysts.
Enzymes that catalyze addition-elimination reactions, lyases (E.C. class 4) are cofactor
independent and suited for cell-free applications. Nevertheless, it has proven more
economical to use resting cells in large scale industrial applications such as the
production of acrylamide (Yamada & Tani, 1983).

Table 5.4 Comparison of the use of enzymes in cell free or intact cell
biotransformations.

Cell Free Purified Growing Resting

Extract Enzyme Cells Cells
Activity moderate high moderate1 high
Specificity moderate high moderate1 moderate
Cofactor recycling required required not required not required
Substrate transport not required not required required required
By-products few very few many2 few-many2
 How to get the biocatalyst 181

Cost of biocatalyst moderate high low-moderate low-moderate

Re-use possible no yes no yes
Solvent tolerance low-moderate low-moderate low-moderate low
Concentration high high low high
tolerance
Product high high low-moderate high
concentrations
Product recovery easy-complex easy complex easy-complex
Equipment cheap cheap expensive cheap
Waste little little cell mass cell mass
1, 2 Can be increased (1) or reduced (2) by using recombinant strains.

Biocatalysts that catalyze group transfer reactions (transferases, E.C. class 2),
isomerizations (isomerases, E.C. class 5) as well as ligase-based biotransformation
reactions (E.C. class 6) typically are used as cell free or purified enzymes. Some of these
enzymes, such as glucose isomerase, are extensively used in the food industries on a large
scale (see Table 5.1). However, relatively few industrial biocatalytic applications exist,
especially since enzymes of these classes generally are highly substrate-specific, making
them less suited for wide range synthetic applications.

5.1.4 Process Parameters

When looking for a suitable biocatalyst, one has also to consider the (operational) activity
that is required for commercial application and the operational conditions that will be
used in the process (e.g. temperature, salt concentration, pH, organic solvents, substrate
and product concentration) will have to be addressed as well. If the reaction is optimally
performed at for instance high temperatures, thermophilic organisms are more likely to
provide the desired enzymes than mesophilic strains (see paragraph 5.4.1). And vice
versa, psychrophiles operate well at lower temperatures and, since they do not require
excessive heat treatment to be inactivated, are easily killed following the process.

5.1.5 Selection of a Suitable Biocatalyst

Once the process criteria have been identified various strategies can be followed to obtain
the biocatalyst for the desired biotransformation. Most commonly, first a literature, patent
and electronic media search is performed (‘database mining’) in order to find established
biocatalysts that are known to catalyze the desired reaction or that catalyze a reaction that
is similar to it (see 5.3.2). Databases that are becoming
 Applied biocatalysis 182

Table 5.5 Useful CD-ROM databases and Internet addresses to obtain biocatalysis
information.

Name CD-ROM or URL Description

Knight-Ridder Knight-Ridder Information OnDisc Chemical Engineering
and Biotechnology
Abstracts
Biotransformations Chapman & Hall Biotransformations Kieslich/Warwick
Biotransformations Club
DBGET http://www.genome.ad.jp/dbget/dbget.links.html Integrated multi-
database network
EXPASY http://expasy.hcuge.ch/sprot/enzyme.html Integrated multi-
database network
ENTREZ http://www3.ncbi.nlm.nih.gov:80/htbin- Integrated multi-
post/Entrez/ database network
SRS http://www.embl-heidelberg.de/srs/srsc Integrated multi-
database network
Biocatalysis http://dragon.labmed.umn.edu/~Iynda/index.html Biodegradation and
Biocatalysis Database
MDB http://cgsc.biology.yale.edu/metab.html Metabolic Data Base
Metabolic http://www.biobase.com/emphome.html Metabolic pathways
pathways
EMP http://biobase.com/emphome.html/homepage.html/ Database of Enzymes
and Metabolic Pathways
NCBI http://www.ncbi.nlm.nih.gov/ National Center for
Biotechnology
Information
PDB http://www.pdb.bnl.gov/ Protein Data Bank
SWISS-PROT http://expasy.hcuge.ch/sprot/sprot-top.html Protein sequences
GENBANK http://ncbi.nlm.nih.gov:2555/r_genbank2.html Nucleotide and protein
sequences

increasingly useful to obtain information on biocatalysis are either commercially

available on CD-ROM or can be freely accessed via the internet (Table 5.5). The
integrated use of multiple databases can thus lead one to a suitable source to develop a
biocatalyst from. These (commercially) available biocatalysts may be directly capable of
 How to get the biocatalyst 183
performing the desired reaction or may be forced to do so by using unnatural substrates,
by using inhibitors that block alternative reactions or by applying different reaction
conditions.
If no suitable biocatalysts can be found among those identified previously, a choice has
to be made between isolation of a novel biocatalyst from nature’s biodiversity (see 5.3.1)
and genetic modification of existing biocatalysts (see 5.3.5).

Table 5.6 Frequently used culture collection information centers and national culture
collections.

Culture Address
Collections
ATCC American Type Culture Collection, 12301 Parklawn Drive, Rockville, MD-
20852, USA
NCIMB National Collections of Industrial and Marine Bacteria, 23 St. Machar Drive,
Aberdeen AB2 I RY, UK
NRRL Northern Regional Research Laboratories, Agricultural Research Service,
USDA, 1815 N. University Street, Peoria, IL 61604, USA
CBS Centraal Bureau voor de Schimmelcultures, Julianalaan 67, 2628 BC, The
Netherlands
DSMZ Deutsche Sammlung für Mikroorganismen und Zellkulturen, Mascheroder
Weg 1b, 3300 Braunschweig, Germany
IFO Institute for Fermentation Osaka, 17–85, Juso-Honmachi 2-chome, Yodogawa-
ku, Osaka, Japan
IAM Institute of Applied Microbiology, University of Tokyo, Tokyo, Japan
CMI Commonwealth Mycological Institute, Ferry Lane, Kew Gardens, Surrey TW9
3AF, UK
Information
Centers
WDC Worldwide Directory of Culture Collections, World Data Center for
Collections of Cultures of Microorganisms, RIKEN, 2–1 Hirosawa, Wako,
Saitama 351–01, Japan
MSDN Microbial Strain Data Network, Institute of Biotechnology, Cambridge
University, 307 Huntingdon Road, Cambridge CB3 OJX, UK
ICECC Information Center for European Culture Collections, Mascheroder Weg 1,
4400– Braunschweig Germany
 Applied biocatalysis 184

5.2 APPLICATION OF EXISTING BIOCATALYSTS

5.2.1 Whole Cell Biocatalysts

A vast amount of knowledge in the area of bioconversions is available through literature
and many of the strains described in publications and patents have been submitted to
culture collections, from which they can be retrieved. A number of frequently used
culture collections and information centers on culture collections is listed in Table 5.6.
From these collections, strains that have been carefully typed and characterized can be
obtained. In many cases, information on biotechnological applications of the strains is
included in the collection and suitable growth media are provided.
A well-known way to accomplish a desired bio transformation is the use of unnatural
substrates for already (commercially) available biocatalysts.

Figure 5.2 Production of 5-methylpyranozoic acid (5-MPA) by P. putida

ATCC 33015. The upper part shows the degradation route for p-
xylene, the lower part the production of 5-MPA, which is not
attacked by benzoic acid dioxygenase.

An example of a process, developed by Lonza, that was based on a strain that was simply
ordered from the ATCC strain collection is the production of 5-methylpyranozoic acid
from 2,5-dimethylpyrazine byPseudomonas putida ATCC 33015 (Kiener, 1992). This
strain is capable of degrading toluene and xylenes via benzoic acid derivatives. It was
serendipitously discovered, that 2,5-dimethylpyrazine was regioselectively hydroxylated
to 5-methylpyranozoic acid but that this product remained untouched since toluic acid
dioxygenase does not accept this compound as a substrate (Figure 5.2). 5-
Methylpyranozoic acid is of use for the synthesis of glipizide, a blood glucose lowering
drug.
 How to get the biocatalyst 185
Another example is the commercialized formation of R-phenylacetylcarbinol (R-PAC)
from benzaldehyde and pyruvate by fermenting brewer’s yeast (see chapter 4.16;
Neuberg & Hirsch, 1921). Pyruvate decarboxylase is responsible for the formation of an
activated acetaldehyde from the decarboxylation of pyruvate and the condensation of this
activated acetaldehyde with benzaldehyde. Nowadays, about 1000 tons per annum (tpa)
R-PAC is produced this way.
Physiological optimization of enzyme synthesis by variation of the culture parameters
is usually required to enhance the catalytic activity of whole-cell biocatalysts to such a
level that it can be applied in a biocatalytic process. In addition, physiological conditions
can influence the selectivity of the reaction, since enzymes with opposite selectivities can
be differentially expressed. In some cases, genetic engineering is required to obtain
biocatalysts with a desired selectivity that does not consume the product of choice (see
5.3.5). Alternatively, one may choose to isolate the desired activity from the culture in
order to use the biocatalyst in an enzyme reactor.

5.2.2 Existing Enzyme-Biocatalysts

The number of enzymes that are of use in biocatalysis and that have become
commercially available has increased considerably over the last few years. This is

Table 5.7 A number of commercial suppliers of enzymes for biocatalysis.

Company Country Company Country

Amano Int. Japan Miles Laboratories USA
Biocatalysts UK Novo Nordisk Denmark
Biozyme UK Oriental Yeast Japan
Boehringer Germany Recordati Italy
Calbiochem USA Röhm & Haas Germany
Diversa USA Rhone-Poulenc France
Fluka Switzerland Sigma Chemicals USA
Genzyme UK Toyo Yozo Japan
Genencor USA Worthington USA
Meito Sangyo Japan

mostly true for hydrolases and particularly for extracellular enzymes such as lipases. A
number of commercial suppliers of enzymes is represented in Table 5.7. In order to
facilitate finding the proper enzyme for a given application, a number of enzyme
 Applied biocatalysis 186
suppliers such as Boehringer, Fluka and most notably Diversa (see also 5.3.5) offer
‘screening kits’ containing limited amounts of a diversity of enzymes that can rapidly be
tested for biotransformation of a given substrate. Once an enzyme catalyzing the desired
reaction has been identified, it can be ordered in larger quantities to develop a process.
From the ‘existing’ enzymes, lipases and proteases are, from the biocatalytic point of
view, the most studied enzymes and their substrate specificities are well understood.
Lipases, their genetics, biochemistry, structures and their application in biocatalysis are
well documented (reviewed in Gilbert, 1993; Vaeger & Wohlfarth, 1993; Nagao & Kito,
1990; Theil, 1995). These enzymes are particularly useful since they display a
considerable resistance to solvent denaturation and can be used in (practically) anhydrous
media and, on top of that, can be cheaply acquired.
It is obvious that the hydrolytic capacity of lipases is not restricted to triacylglycerols,
but that for instance condensation reactions can be carried out and that unnatural esters
can be synthesized by means of commercially available lipases, provided that the proper
reaction conditions are chosen. Klibanov and coworkers have shown that the
enantioselectivity and stability of various enzyme catalyzed reactions are affected
significantly by organic solvents. By choosing the right reaction conditions, the activity
but also the selectivity of existing biocatalysts can be varied to better suit process
requirements: the enantioselectivity E of a transesterification reaction was shown to be
markedly reduced in the presence of anhydrous organic solvents and was dependent on
the solvent dipole moment. Another parameter is temperature as illustrated by a
secondary alcohol dehydrogenase from the thermophilic Thermoanaerobacter
ethanolicus. It was shown that the stereochemical properties of the enzyme are also
strongly dependent on temperature (Pham, Philips & Ljungdahl, 1989), which can be
explained when one considers the relation between the enantioselectivity E and the
differences in Gibbs free energy between the R- and S-transition states G‡:– RTlnE=
‡
G (Faber, 1995).

Table 5.8 Biological diversity of microorganisms, estimated from the rate of discovery
of new species1.

Microorganisms Known Estimated Percentage known

bacteria 4,760 800,000–3,000,000 0.2–0.6%
fungi 69,000 1,500,000 5%
algae 40,000 190,000–1,800,000 0.4–24%
1A.Bull, Genetics of Industrial Microorganisms ’94

A commercialized example illustrating the potential of proteases under non-natural

conditions is the formation of aspartame from N-protected-L-aspartate and L-
 How to get the biocatalyst 187
phenylalanine methyl ester catalyzed by thermolysin (see chapter 4.6; Isowa et al., 1979).
A DSM-Tosoh joint venture has commercialized this process for large-scale production
of this low caloric sweetener. A big advantage of this process is the stability of the
thermolysin, a metallo-endoproteinase from the thermophilic bacterium Bacillus
thermoproteolyticus, in the presence of organic solvents (Oyama, 1986).

5.3 SELECTION OF NOVEL BIOCATALYSTS

If no suitable biocatalyst for the desired biotransformation is available or when the

activity or selectivity of known biocatalysts is insufficient to develop a commercial
process, a search for a novel biocatalyst is required.

5.3.1 Sampling Sites

Based on the rate at which novel microorganisms are being discovered, one can make an
estimation of the amount of different organisms that have been identified up to now
relative those that remain uncharacterized (Table 5.8). The overwhelming biochemical
diversity present in nature makes that the isolation of microorganisms with novel
enzymatic activities is worthwhile (reviewed in Bull, Goodfellow & Howard Slater,
1992). Preferred sites to find novel enzymes include so-called mega-diversity countries
such as Mexico, Colombia, Brazil or India and biodiversity hot spots like rain forest and
deep seas.
In order to be successful in finding a biocatalyst that is adapted to the required process
conditions, one could sample at a site where similar conditions apply. As such sampling
sites with low biodiversity, but high selection pressure such as hypersaline ponds, acidic
or alkaline lakes, deserts, hot springs or sites polluted with chemicals are particularly
useful.

5.3.2 Screening Procedures

Cheetham (1987) has presented a possible flow diagram for the procedures involved in a
classical screening program (Figure 5.3). To perform a primary screen for the
 Applied biocatalysis 188

Figure 5.3 Flow diagram of screening operations.

desired activity using large numbers of organisms cheap, simple, rapid and selective
detection methods should be available. This is especially true for high throughput
screening programs that make use of robots and preferably simple analytical techniques
 How to get the biocatalyst 189
like colorimetry or fluorescence to screen thousands of colonies per day.
Once the reaction type and possible intermediates involved in the desired reaction are
known, a primary screen is developed, which is aimed at identifying organisms

Figure 5.4 Production of cis, cis-muconic acid from benzoic acid by the
consecutive action of benzoate-1,2-dioxygenase, dihydroxy-
dihydrocyclohexa-2, 4-diene-l-carboxylic acid dehydrogenase and
catechol-1,2-dioxygenase.

that can transform a certain compound. Afterwards, a secondary screen is performed on

isolates that gave a positive signal in the primary screen and at this stage product
characterization, reaction rates and selectivities are evaluated. When finally a suitable
biocatalyst has been identified, this strain is optimized for production.
Designing primary screening procedures can be difficult. The procedure should
discriminate against false positives but at the same time allow substrate analogues or
chromogenic substrates to be accepted. Also, false negatives (e.g. glycosidases are known
that do not hydrolyze chromogenic substrates) pose a problem, since the desired activity
may not be found at all.

Growth

One of the most powerful primary screening tools remains growth of an organism on the
starting compound or on analogues thereof as a sole carbon, nitrogen or for instance
phosphorous source. Care has to be taken however, not to miss out on positives that do
perform the desired step, but cannot metabolize the reaction product(s) and therefore do
not demonstrate growth. Mizumo et al. (1988) have screened thousands of benzoic acid
assimilating strains to obtain an efficient producer of cis, cis-muconic acid (Figure 5.4).
In order to circumvent benzoate-inhibition problems in the ultimate process,
microorganisms were isolated at elevated benzoic acid concentrations. Several promising
cis, cis-muconic acid producers were obtained by subjecting these strains to mutation in
order to eliminate the cis, cis-muconic acid degrading enzyme. Indeed, a mutant was
isolated which was capable of producing cis, cis-muconic acid with a quantitative yield of
44.1 g/1 (48 h) by successive feeding of benzoic acid.
Nitrile-hydrolyzing enzymes, such as nitrile hydratase, nitrilase and amidase, have
great potential as catalysts for producing high-value amides and acids from the
corresponding nitriles and can be obtained by screening for microorganisms that can
utilize nitriles as nitrogen source for growth. A wide variety of microorganism able to
 Applied biocatalysis 190
degrade nitriles and distinct hydration pathways has thus been found and described
(Figure 5.5). Nitto Chemical Industry has commercialized the production of acrylamide
from acrylonitrile using Rhodococcus rhodochrous J1 (30,000 tons per annum), which
was found after a screening program (see section 4.17). It has now replaced the initial
production strain Pseudomonas chlarapsis B23. The latter has a nitrile hydratase activity
towards acrylonitrile that is at least 3000 times higher than

Figure 5.5 Degradation routes for nitriles. The first route is a two-step reaction
involving a nitrile hydratase, which converts the nitrile to the amide,
and an amidase, which converts the amide to the corresponding acid.
The second pathway involves direct hydrolysis of the nitrile to the
carboxylic acid and ammonia by a nitrilase.

the amidase activity, and as a consequence more than 99% of the nitrile is converted into
the acrylamide without the formation of acrylic acid as a by-product (Nagasawa and
Yamada, 1989). To increase the nitrile hydratase activity of P. chlororaphis B23, the
culture conditions were optimized. The addition of ferric ions, essential for the nitrilase,
and methacrylamide as an inducer significantly enhanced the formation of nitrile
hydratase (Yamada et al., 1986). The activity has been increased even further by
mutagenesis (Ryuno, Nagasawa & Yamada, 1988). As a result of such optimization of
culture conditions the amount of nitrile hydratase enzyme in the current production strain,
Rhodococcus rhodochrous J1 has increased up to almost 50% of the total soluble protein
in the cell (Nagasawa, Shimizu & Yamada, 1993; Yamada, 1998; Nagasawa et al., 1991).
In spite of the enormous amount of work performed on nitrile degrading enzymes, for
quite some time no enantioselective nitrile hydratase had been described. Based on a
laborious screening program, examining about 800 isolated strains and 500 culture
strains, researchers at Asahi (Japan) have isolated an Acinetobacter species which
contains a highly specific S-nitrilase, converting for instance 2-(4′-isobutylphenyl)
propionitrile to S-(+)-ibuprofen (Figure 5.6; Yamamoto et al., 1990), a well-known
analgesic.
The Ajinomoto company has developed a commercial route to L-cysteine, which is
currently applied at 1000 tpa scale, by screening for cultures that produced Lcysteine
from DL-2-amino- 2-thiazoline-4-carboxylic acid, which can be synthesized from 2-
 How to get the biocatalyst 191
chloroprop-2-enoic acid methyl ester (Sano et al., 1977; Yokozeki et al., 1987) (Figure
5.7). Enantioselective hydrolysis of L- and spontaneous racemization of D-2-amino- 2-
thiazoline-4-carboxylic acid constitutes a process with complete conversion of the
starting compound.

Clearing and precipitation zones

A technique that has proven successful for compounds that are poorly soluble in aqueous
solution is the formation of clearing zones (‘halo formation’) around active

Figure 5.6 Production of S(+)-ibuprofen from rac. 2-(4’-isobutylphenyt)-

propionitrile using an S-specific nitrilase from Acinetobacter
(Yamamoto et al., 1990).

colonies that are plated on agar plates containing the substrate at an supersaturated
concentration. By transformation of the insoluble compound to a soluble product, a
translucent halo is formed around a colony that produces active enzyme. The use of agar
plates containing casein to isolate protease excreting microorganisms is an illustration of
this technique.
The inverse technique, which makes use of complexation of compounds to the product
such that it becomes insoluble (e.g. fatty acids with barium cations), is also applied.
These methods are also applicable for strains that do not grow on the screening
compound due to lack of a complete metabolic pathway for the starting material. A
variation on this theme is the trapping of radioactive tracers, such as 14C
 Applied biocatalysis 192

Figure 5.7 Screening of microorganisms capable of hydrolyzing DL-amino-2-

thiazoline-4-carboxylic acid has led to a process developed by the
Ajinomoto company (currently applied at a scale of 1000 tpa) for the
production of L-cysteine (Sano et al., 1977; Yokozeki et al., 1987).

Figure 5.8 Fluorogenic substrates to screen for lipases and esterases (upper
molecule) and esterases (lower molecule). A pyrene group functions
as the fluorophore and fluorescence is quenched by a
trinitrophenylamino moiety attached to the same molecule (Hennetter
et al., 1998).

labeled CO2 that is generated by cells growing on a suitable, radioactive carbon source or
 How to get the biocatalyst 193
liberated from a radioactive substrate by decarboxylases. Radioactive CO2 can be bound
to filter paper that has been soaked in Ba(OH)2 and is detected by autoradiography. In
such a way, microtiter plates have been used to screen large numbers of cultures for
decarboxylase activity.

Chromogenic substrate analogues

Chromogenic substrates that are highly similar to the desired starting compound are of
great use to screen for organisms that carry out the desired biotransformation reaction.
Chromogenic substrates that release chromophors, such as the yellow compound ortho-
or para-nitrophenol, or fluorescent groups such as fluorescein, umbelliferone, or
rhodamine can rapidly be analyzed and are well suited for automation. For hydrolases in
particular, numerous chromogenic substrates are available (reviewed in Wolley and
Petersen, 1994). One has to realize however, that the chromogenic group may cause steric
interference with the desired biotransformation, which can lead to false negative results.
Recently, a series of self-quenching fluorogenic substrates for lipases and esterases
were reported (Hennetter et al., 1998), which carry a fluorophore (pyrene) and a
fluorescence quencher (trinitrophenylaminoacyl) group. Cleavage of the quencher chain
from the triglyceride substrate leads to an increase of the fluorescence, which can easily
be detected and is suitable for automation (Figure 5.8).
 Applied biocatalysis 194

Figure 5.9 Examples of derivatization reagents for the colourimetric detection

of primary and secondary amines and ketones.

Complexation of the product

Derivatization or complexation reagents that react or form a coloured complex with

functional groups of the product formed provide another technique to screen for novel
biocatalysts (see Figure 5.9). One should be conscious however, that some reagents
cross-react with cell constituents and that this should be avoided by analyzing cell-free
culture fluids thus reducing the isolation of false positives in a screening program. An
elegant solution to this problem is the use of so-called agar overlay screens, which are
rapid and have successfully been used (Morin, Hummel & Kula, 1986) to screen for D-
hydantoinases using PDMB, which reacts with primary amines to a coloured Schiff base
(Figure 5.10). First, cells were grown on agar plates and the colonies that appeared were
covered with a layer of agar containing the substrate dihydrouracil. The formation of the
hydrolysis product N-carbamoyl-3-aminopropanoic acid, detected by PDMB, indicated
the presence of hydantoinase activity with only few false positive signals, when
comparing the data with nucleic acid methods (see section 5.3.3) and activity assays
(Figure 5.10) (Morin & Leblanc, 1998). Another example of this complexation method is
the screening for strains producing L-3,4-dihydroxy-phenylalanine (L-DOPA), an
effective pharmacological agent in the treatment of Parkinson’s disease, on plates.
Microorganisms that were able to hydroxylate L-tyrosine to L-DOPA formed colonies
that turned violet-black,
 How to get the biocatalyst 195

Figure 5.10 Agar overlay screening procedure to screen microbial populations

for hydantoinase activity (Morin, Hummel and Kula, 1986). A screen
for dihydropyridinase activity based on Schiff base formation with
PDMB (upper panel) led to the identification of numerous strains
with D-hydantoinase activity, which in combination with a D-
carbamoylase is employed to produce D-amino acids.

as a result of the reaction of the accumulating L-DOPA with ferrous ions added to the
agar plates (Tanaka, Yoshida & Nokayama, 1974).

Indicator strains

Another strategy that has been succesfully pursued is the application of indicator
microorganisms, which display a physiological reaction, such as growth or non-growth,
towards a desired product. Amino acid deficient (auxotrophic) strains or strains that are
strongly inhibited by the desired product are applied in a layer on top of the population of
strains that is investigated. This is illustrated by the work of Wagner and coworkers
(Gross, Syldatk & Wagner, 1987), in which microorganisms were screened for their
capacity to produce L-amino acids from the corresponding D,L-hydantoins. The use of a
tryptophan-auxotrophic yeast in an overlay assay enabled the detection of L-tryptophane
producing microorganisms from D,L-indolylmethylhydantoin by monitoring growth of
the yeast strain. Similarly, benzyl-penicillin amidase producing strains were found by
overlaying plates with a layer of Serratia marcescens culture. S. marcescens is coloured
red and resistant to benzylpenicillin amide, but not to the product 6-aminopenicillanic
acid (6-APA), and colonies where growth of the coloured indicator bacteria was lacking
were found to contain the desired enzymatic activity.

Table 5.9 Completed genome sequencing projects1.

Species Complete genome size (nt) Year

Archeae Archaeoglobus fulgidus 2,178,400 1997
Methanobacterium thermoautotrophicum 1,715, 377 1997
Methanococcus jannashii 1,664,977 1996
Pyrococcus horikoshii 1,738,505 1998
Eubacteria Aquifex aeolicus 1,551,335 1998
Bacillus subtilis 4,214,807 1997
Borrelia burgdorferi 1,230,663 1997
 Applied biocatalysis 196

Chlamidia trachomatis 1,042,519 1998

Escherichia coli 4,639,221 1997
Haemophilus influenzae 1,830,135 1995
Heliobacter pylori 1,667,867 1997
Mycobacterium tuberculosis 4,411,529 1998
Mycoplasma genitalium 580,073 1995
Mycoplasma pneumoniae 816,394 1996
Rikettsia prowazekii 1,111,523 1998
Synechocystis sp. 3,573,470 1996
Treponema pallidum 1,138,011 1998
Fungi Saccharomyces cerevisiae 12,069,313 1997
Animals Caenorhabditis elegans 97,000,000 (appr.) 1998
1 Status: January 1999

Use of antibodies

Antibodies raised against highly purified enzymes can serve to screen biological samples
for variants that catalyze the same reaction, but that are for instance better suited to
perform under process conditions or have a more desirable substrate spectrum. Screening
using polyclonal antibodies against the L-hydantoinase, hydantoin-racemase and L-N-
carbamoylase of Arthrobacter aurescens DSM 3747 has been successful late strains with
higher productivities for the synthesis of L-amino acids (Siemann, Syldatk & Wagner,
1993).

5.3.3 Screening by Using Molecular Biology Techniques

A number of scientific consortia have taken up the formidable endeavor to sequence the
genomes of Homo sapiens and a number of medically and commercially important
microorganisms. Some of the sequence projects have been completed (Table 5.9) and this
provides a wealth of information which can be of use for biocatalytic applications.

Database mining

Based on homologies between the genes from sources for which a clear function of the
encoded enzyme has been defined on basis of biochemical evidence, one can predict the
function of a large number of sequence homologues within these sequenced organisms.
 How to get the biocatalyst 197
Having the complete sequence of an organism available makes it easy to isolate genes
which on basis of computer prediction may encode a desired enzymatic activity, In this
light, it is interesting to note that the genomes of four Archeae, which frequently are
extremophiles of considerable interest for biocatalysis, have been completed. Also,
knowing the genome sequence of host strains that are suited for whole cell biocatalyst
development, such as E. coli, B. subtilis and S. cerevisiae (Table 5.9) opens ways to
construct specialized recombinant strains that are devoid of unwanted side reactions by
making specific deletions on the chromosome. The information obtained from genome
analysis (genomics) is of particular interest for construction of strains designed to
produce a desired product from a cheap precursor molecule such as glucose. Assembly of
pathways towards a certain product by introducing the proper genetic information
(metabolic pathway engineering) can theoretically provide the most cost-effective route
and this technology holds great promise for the decades to come.

Screening by polymerase chain reaction

When such detailed information as a chromosomal sequence is not available for a strain
with an established biocatalytic potential, the use of the polymerase chain reaction (PCR)
technique can help out. By using the amino acid sequence of enzymes with a similar
activity as the enzyme to be cloned, one can design degenerate oligonucleotides (PCR
primers) complementary to this amino acid sequence. With these primers, one can
attempt to amplify the desired gene and clone it into an expression vector. Similarly, the
technique of DNA to DNA hybridization (‘Southern’ and colony blotting) can be used to
identify clones that contain genes which are homologous to those encoding known
enzymes. Here again, the desired gene is brought to expression in the proper host, using a
suitable vector. The constructs resulting from both the PCR and ‘Southern’ approach can
be employed to produce the enzyme for either whole-cell or cell-free biocatalytic
applications. Using a 122bp probe of a D-hydantoinase gene from Pseudomonas putida
Morin and coworkers were able to identify strains carrying hydantoinase genes in a
selective screen (Morin & Leblanc, 1998). The advantage of this hybridization procedure
is that genes can be identified even when the conditions of the selection do not lead to
induction of the enzyme. A ‘classical’ screening protocol based on enzyme activity
would lead to more false negative results.

Expression libraries

One of the major problems in exploiting biodiversity is that many of the microorganisms
that occur in nature cannot be cultured under laboratory conditions. Nevertheless, these
cells may contain activities that are relevant for development of a biocatalytic route. To
get access to this wealth of catalytic possibilities, one has to make use of molecular
biology techniques. In effect, expression gene libraries are constructed from biotopes
rather than from defined microorganisms (reviewed in Dalbøge and Lange, 1998). The
 Applied biocatalysis 198
DNA from an environmental sample is cloned in an expression vector and genes that lie
on the cloned DNA are brought to expression (a socalled ‘expression library’). By
performing screens for a desired enzymatic activity, specific clones corresponding to that
activity are isolated and the genes encoding the enzymes can be further sub-cloned. In
such a way, recombinants selected for a specific process or specific process conditions
can be constructed. Depending on the application and process requirements, one may
decide to purify the enzyme (CloneZyme™) from such a recombinant. This concept was
introduced by the company Recombinant Biocatalysis Inc. (now Diversa Corp.) for
exploration of microrganisms from hot springs, which are frequently not culturable in the
laboratory. Novo Nordisk has also used the concept of expression cloning using a cDNA
expression library to obtain over 200 different enzymes of industrial relevance, such as
arabinanases, endo-glucanases, galactanases, mannanases, polygalacturonases, pectin
lyases, pectin methylesterases, proteases, rhamnogalacturonases, lipases and xylanases as
well as exo-acting enzymes (Dalbøge, 1998).

5.3.4 Screening by Novel Analytical Techniques

One of the main prerequisites that is required for rapid analysis of screening samples
(intermediate to high throughput screening) is a sufficiently discriminative analytical
technique, which can differentiates between a positive and negative screening result.
Analytical techniques, based on HPLC, LC-MS, GC, GC-MS etc. are insufficiently fast
and require too much handling to be used for a primary screening round, especially when
the amount of strains to be analyzed is high. A novel analytical method is based on
Fourier Transform Infrared Spectroscopy (FT-IR), which contains much more spectral
information than for instance UV spectroscopy. Previously, FT-IR has proven very
powerful to discriminate between strains of Streptomyces (Naumann et al., 1991). The
novel approach with the acronym DRASTIC (Diffuse Reflectance Absorbance
Spectroscopy Taking In Chemometrics), combines FT-IR with powerful computational
techniques and is a promising candidate for screening application, since it can
discriminate between structurally related metabolites in dried samples from liquid culture
(Winson et al., 1997). Using DRASTIC, different ampicillin concentrations in mixtures
with washed E. coli HB101 cells, grown in complex media, could be discriminated
(Figure 5.11).

5.3.5 Genetic Modification of Existing Biocatalysts

Natural and accelerated evolution

Since the middle of the nineteenth century we have introduced increasing quantities of
complex, man-made (anthropogenic) chemical compounds into the natural environment.
Many of these substances are recycled by microbial action, whilst others are poorly
 How to get the biocatalyst 199
biodegradable. This challenge from the chemical industry has resulted in the evolution of
new and unexpected enzyme activities and is a striking example of in vivo genetic
engineering operating on a large scale. Mutational events leading to the acquisition of
novel enzyme activities include transfer of genes, gene duplication, gene fusion,
recombination between genes, deletion or insertion of

Figure 5.11 Diffuse Reflectance Absorbance Spectroscopy Taking In

Chemometrics (DRASTIC), a FT-IR-based method for rapid
screening for metabolites. Different concentrations of ampicillin
(ranging from 0 to 5 mg/mL) were mixed with a constant amount E.
coli cells, dried and analyzed by FT-IR (Winson et al., 1997).

gene segments and one or more single site mutations, or combinations of these. Despite
the fact that nature thus provides us with new ‘variations on a theme’, such novel enzyme
activities have to be found and isolated. Furthermore, enzymes that have evolved fairly
recently are not likely to make good catalysts in terms of affinity and turn over numbers.
A well known example is the haloalkane dehalogenase from Xanthobacter autotrophicus
GJ10, an enzyme thought to be of recent origin (Pries et al., 1994), which is an inefficient
dehalogenating enzyme with a high KM for its natural substrate. It has to be produced at
considerable levels compared to total cell protein in order to supply the cell with
sufficient carbon for growth (reviewed in Janssen, Pries & van der Ploeg, 1994).
Another example that illustrates the plasticity of genes in the environment is the
construction of a 3-chlorobenzoate, 4-chlorobenzoate and 3,5-dichlorobenzoate
degrading Pseudomonas species from two distinct Pseudomonas strains (Reineke &
Knackmuss, 1980). Pseudomonas sp. B13 is able to grow on 3-chlorobenzoate, but
cannot grow on 4-chlorobenzoate and 3,5-dichlorobenzoate. P.putida mt-2, however
 Applied biocatalysis 200
contains a toluate-1,2-dioxygenase (see Figure 5.2), encoded by the xylXYZ genes of the
TOL plasmid, that readily oxidizes 3-chlorobenzoate and 3,5-dichlorobenzoate. By
selective enrichment in a chemostat, exconjugants of these two Pseudomonas strains have
been isolated which can degrade both 4-chlorobenzoate and 3,5-dichlorobenzoate as well
as 3-chlorobenzoate. In a similar way, running a chemostat under strong selective
pressure can alter the substrate specificity of biocatalysts, their affinity towards
substrates, their regulation, product inhibition, temperature stability etc.

Figure 5.12 The ICI (Zeneca) process for polyphenylene production. In

addition to P. putida NCIMB 11767, recombinants of E. coli were
used with “improved cost effectiveness”.

Genetic engineering of biocatalysts

Genetic engineering nowadays allows us to control gene expression independent of the

natural inducers of enzymatic activity and independent of the natural host that once
harboured the enzyme. This is especially relevant since natural isolates are specialized in
degradation of the educt, not synthesis of a product. Frequently, efficient degradation of
the desired product is observed, which can be prevented by using a host different from
the natural isolate.
In living cells, regulation of enzyme activity can occur at various stages during a
production process. If such control mechanisms pose a problem to the efficient
application of whole cell biocatalysts, regulatory mutants or specific recombinants must
be created. If a strain does not perform adequately, strain improvement strategies can be
used that are based on mutagenesis followed by selection of a desired phenotype
(‘classical’ strain improvement) or by genetic engineering.
 How to get the biocatalyst 201
An example of a significant improvement of a biocatalyst by genetic engineering is the
polyphenylene process of Zeneca (Ballard et al., 1994). For quite some time, it had been
impossible to produce high molecular weight polyphenylene, a polymeric material with
interesting material properties. A biological route to polyphenylene was developed by ICI
based on cis-1,2-dihydroxycyclohexa-3,5-dieneyclohexadiene (cis-diol) produced by the
benzene dioxygenase from P. putida NCIMB 11767 (Figure 5.12). In order to do so,
chemical mutagenesis rounds to obtain: i.) a cis-diol accumulating strain that didn’t
degrade the desired compound; ii.) obtain a mutant with constitutive expression of the
dioxygenase and iii.) obtain a mutant that is not subject to catabolite repression by
glucose, were required. The resulting strain, P. putida UV4, was used until Zeneca had
cloned the benzene dioxygenase genes for use in recombinant strains with “improved cost
effectiveness” (Ballard et al., 1994).
 Applied biocatalysis 202

Figure 5.13 Production of cis-diols using P. putida F39/D and some synthetic
applications.

A closely related case is that of the use of cis-diols for synthetic applications as elegantly
explored by Hudlicky and coworkers, who have used P. putida F1 or rather P. putida
F39/D, a cis-diol dehydrogenase negative mutant, to produce a variety of cis-diols
(Figure 5.13). These molecules are difficult to construct by organic chemical synthesis
and serve as very powerful synthons for chiral synthesis (Butora et al., 1996; Faber,
1995; Hudlicky, Endoma & Butora, 1996a; Hudlicky et al, 1996; Hudlicky et al., 1996c).
 How to get the biocatalyst 203
Here again, E. coli; recombinants expressing the todC1C2BA genes, have proven more
successful than P. putida F39/D, especially since induction by toluene can be omitted
(Hudlicky et al., 1996c). A particularly elegant example of the use of a rare enzyme
produced by a genetically-modified organism is found in the production of guluronate-
enriched alginate for use in the encapsula-tion of Islets of Langerhans for the in vivo
production of insulin in diabetics. Encapsulation in alginate protects the islet cells from
the host’s immune system but high mannuronate alginate is itself antigenic and capsules
become coated with leukocytes. High guluronate alginates are not antigenic but are not
readily available. The alginates used commercially are extracted from brown algae, but
very similar polyuronates are produced by bacteria such as Azotobacter vinelandii and
Pseudomonas aeruginosa. The first polymeric product in the synthesis of alginate in
Azotobacter vinelandii (as in algae) is polymannuronate: guluronate is produced by
epimeri-sation of mannuronate residues within the polymannuronate (Pindar and Bucke,
1973). This reaction, is catalyzed by epimerases. Brown algae are inconvenient sources
of the epimerase and A. vinelandii has several different epimerases and, more
significantly, several alginate lyases which depolymerize the alginate. Hence the bacteria
are not satisfactory practical sources of the epimerase either. The problem has been
solved by cloning the gene for one of the epimerases into Escherichia coli and the
recombinant organism has been used as a source of sufficient epimerase to produce
several kilograms of high-guluronate alginate. (Ertesvåg and Skjåk-Braek, in press).
Characterization of enzymes that catalyze the individual reactions is an ever
developing field and new enzymatic activities are constantly discovered and
characterized at the molecular and genetic level. As a result, a vast arsenal of cloned
enzymes is now available and many industrial enzymes are currently produced from
genetically engineered enzyme producers. Relatively few industrial processes however,
have been realized that utilize recombinant whole-cell biocatalysts. Genetically
engineered whole-cell biocatalysts can be beneficial when one wants to prevent an
undesired side reaction catalyzed by an isolated organism and space-time yields can be
improved by elevating enzyme dosages using molecular biology techniques.
Furthermore, regulation of enzyme systems can be controlled and multistep reactions can
be assembled using genetic information from various sources. In short, genetic
engineering enables the construction of biocatalysts, specifically tailored to a process.

Metabolic pathway engineering

Despite the successes of strain improvement by mutagenesis and selection, especially in

amino acids and penicillin production, we will focus more on examples that make use of
genetic engineering techniques to improve the production of desired molecules in a
fermentation process. Metabolic pathway engineering or MPE, which can be defined as
the purposeful modification of networks of metabolic reaction in microorganisms, whole
plants or animals, is especially suited to produce molecules that are common to nature by
fermentation. The technology, which has a long development time, is especially suited
for generic products and is most commonly applied for the synthesis of proteogenic
 Applied biocatalysis 204
amino acids from glucose (see also Figure 5.1).
One of the most well known cases is that of indigo. The serendipitous discovery that
cloning of the naphthalene dioxygenase genes in E, coli leads to the formation

Figure 5.14 The Genencor process for indigo production. See text for details.

of the blue dye indigo by has instituted an ambitious research project, first at Amgen and
later at Genencor to develop a green route for indigo biosynthesis. In the final E. coli
 How to get the biocatalyst 205
strain, 18kb of DNA comprising 15 open reading frames was introduced and stabily
maintained (Figure 5.14). The host strain contained an aroG* mutation, which enhances
the carbon flux through chorismate pathway, an enhanced trp operon, a mutation in trpB
(tryptophane synthase, B subunit) and the naphthalene dioxygenase genes. The indigo
that was produced however, was slightly coloured red, due to the formation of indirubin
(Figure 5.14). Addition of an isatin hydrolase encoding gene from P. pulida, resolved this
problem. The production of ‘bio-indigo’ from glucose was performed up to 300,000L
scale and dye to stain 200,000 linear yard of denim was produced at a US$0.90/lb, which
is equal to the market price of chemical indigo (Bialy, 1997). Despite the success, this
example has not been commercialized, probably since the cost price of chemical indigo is
substantially lower than the market price.

Protein engineering

Protein engineering is another important technology to try and elucidate the structure-
function relationship of proteins and to improve enzymes to function better under process
conditions. Protein engineering and rational protein design have not always been
successfully used to improve existing biocatalysts, since due to the complexity of
proteins our knowledge of structure-function relationship is good enough to explain, but
insufficient to predict (see chapter 7). Nevertheless numerous examples exist where
rational changes have led to an improved enzyme. In the production of starch-derived
syrups, site-directed mutagenesis is now being applied to obtain bacterial a-amylases with
a reduced need for calcium ions to stabilize the enzyme. As a consequence, the costs may
be lowered of deionising glucose syrup prior to isomerization by glucose isomerase,
which is inhibited by calcium ions.

Directed evolution and gene shuffling

To go one level further in customizing enzymes towards a process, one can apply
directed enzyme evolution, a technique whereby a gene is subjected to random
mutagenesis followed by strong selection for a desired phenotype. This technology does
not require prior knowledge on the three dimensional structure of a given biocatalytic
enzyme. DirectEvolution™ is offered on a commercial basis by Diversa Corp., offering
customers enzymes that can be made more active under a specific set of conditions of pH
and temperature.
Directed evolution involves multiple rounds of random mutation and selection
combined with gene shuffling to evolve enzymes towards desired properties (reviewed in
Arnold and Moore, 1997; Kuchner and Arnold, 1997). The group of Arnold has
succeeded in ‘evolving’ a dimethylformamide (DMF)-sensitive esterase for the cleavage
of the loracarbef-p-nitrobenzyl ester into an esterase that remains active in 15% DW
(Moore et al., 1997). Most of the mutations that had been found in the solvent-resistant
mutants could not have been predicted using current computational methods.
 Applied biocatalysis 206
DNA or gene shuffling is a related technique (patented by Maxygen) whereby, within a
family of related genes, domains are interchanged randomly. It is a powerful technique
since it combines useful mutations from individual genes. Libraries can be generated by
random fragmentation of related genes, followed by reassembly of the fragments in a
polymerase reaction. By screening a large number of such ‘shuffled’ genes, variants with
novel, desired characteristics can be obtained (Zhang, Dawes & Stemmet, 1997; Crameri
et al, 1998).

Table 5.10 Temperature classification of microorganisms.

T range (°C) T optimum (°C)

Psychrophiles 0–25 10–15
Mesophile 10–42 20–37
Thermophile 30–70 45–60
Extreme thermophile 45–80 60–75
Hyperthermophile 85–110 85–105

5.4 NOVEL BIOCATALYSTS

5.4.1 Thermostable Enzymes

Thermophilic and especially hyperthermophilic microorganisms, which are able to
survive at temperatures above 100°C (Table 5.10), contain enzymes that are stable at
elevated temperatures. These thermostable enzymes are of special interest for application
in biocatalysis, since most processes are optimally performed at elevated temperatures
(Arrhenius law), provided that reaction components are stable. In addition to this increase
in reaction rates, thermostable enzymes are particularly interesting since they have a
higher denaturation temperature and rigidity than mesophilic enzymes, which can be
employed to rapidly purify thermophilic enzymes from recombinant mesophiles such as
E. coli by boiling. Other advantage is the storage stability of these enzymes at room
temperature, which leads to lower costs. Thermophilic enzymes have therefore found
applications in production of specialty chemicals, processing of animal feeds, starch
processing, baking and food processing, brewing and fermentation, pulp and paper
industry, in detergents and in waste treatment systems.
To increase the thermostability of existing enzymes by genetic engineering techniques
is difficult, since molecular determinants that increase thermostability are hard to predict.
Comparisons of heat labile and heat stable structures do however indicate that
 How to get the biocatalyst 207
thermostable enzymes contain fewer amino acids that are vulnerable to heat such as Asn,
Gln (deamination), Cys (disulfide bond destruction) or Asp (peptide bond hydrolysis). In
addition, thermostable enzymes are more compact by strong internal hydrophobic
interactions, absence of deep spaces, more helix-forming amino acids, more ion pairs on
the surface, internal salt bridge networks, restriction of C- and N-terminal mobility,
stabilization of -helix dipoles, smaller and fewer surface loops, and more Pro residues
in these loops. A new round of screening to isolate thermostable enzymes or a ‘directed
evolution’ programme to obtain thermostable variants of an industrially relevant enzyme
are therefore more likely to succeed than ‘rational’ protein engineering.

5.4.2 Catalytic Antibodies (Abzymes)

The immune system that is present in higher eukaryotes is a highly diverse molecular
recognition system which on average consists of 108 to 1010 different antigen binding

Figure 5.15 Diels Alder reaction of N-ethylmaleimide and

tetrachlorothiophene dioxide, catalyzed by monoclonal antibody 1E9
raised against a hapten that mimics the transition state (Hilvert et al.,
1989).

molecules, the antibodies. As early as 1969, Jencks postulated that the mechanism of
recognition by antibodies could be of use for catalysis:

“If complementarity between the active site and the transition state contributes
significantly to enzyme catalysis, it should be possible to synthesize an enzyme
by constructing such an active site. One way to do this is to prepare an antibody
to a haptenic group which resembles the transition state of a given reaction. The
combining sites of such antibodies should be complementary to the transition
state and should cause an acceleration by forcing bound substrates to resemble
 Applied biocatalysis 208
the transition state”.

Finding such catalytically active antibodies at that time was looking for a needle in a
haystack. The number of antibodies from an immunized animal that bind the transition
state analogue hapten is low and the amount of antibodies that is catalytically active on
the actual substrate is only a fraction of that. It took until 1986 with the development of
the hybridoma technology for producing monoclonal antibodies before this concept could
be verified experimentally. This technology allows the screening of thousands of
monoclonal antibodies for binding of the analogues and for catalytic activity on the actual
substrate. Furthermore, the amount of the catalytically active antibodies with the desired
activity is very much higher. Most importantly, the hybridoma cell lines are immortal and
can be thus be used to produce homogenous preparations of specific antibodies without
batch-to-batch variations. Catalytic antibodies or abzymes have been employed to
catalyze a diverse group of reactions such as hydrolysis and synthesis of esters and
amides, -elimination reactions, Claisen rearrangements, Diels-Alder reactions (see
Figure 5.15), cis-trans isomerization reactions, cyclization reactions, transesterifications,
photo-induced dimerization and cleavage reactions, redox reactions, decarboxylation
reactions and even aromatic substitutions (reviewed in Benkovic, 1992; Lerner, Benkovic
& Schultz, 1991).
Despite the apparent diversity of the catalytic antibodies and the fact that they can
catalyze reactions that are uncommon to enzymes (e.g. Diels-Alder and Claisen
rearrangement reactions), industrial applications of these catalysts have not yet been
realized. This is mostly due to their cost of production, which is very high compared to
that of enzymes, and due to the relatively low specific activity of abzymes, which is
usually orders of magnitude lower than that of enzymes. Other technical difficulties for
production of catalytically active antibodies are that they are prone to product inhibition,
that it is difficult to find and synthesize stable transition state analogues and that these are
frequently non-immunogenic. Novel developments however, such as phage display
technology and especially the production of single chain antibodies in recombinant
Escherichia coli strains (Brooks et al. 1996), making antibody production independent
from immunogenicity of a hapten and facilitating antibody (fragment) production in high
amounts, could bring this technology into a stage that is more competitive with enzymes
in the years to come.

5.4.3 Ribozymes
The notion that RNA existed prior to enzymes and that RNA molecules can be
catalytically active is by now well accepted. Indeed, RNA molecules have been observed
to catalyze phophodiester bond breaking and synthesis (as occurs during replication and
splicing) but also reactions more distant to its structure, such as amide bond formation
(Wiegand, 1997). RNA thus provided the means to assemble peptides which may have
led to the protein world of today (Zhang and Cech, 1998). Reactions catalyzed by RNA
molecules have thus far not been employed in biocatalysis and it is unlikely that
 How to get the biocatalyst 209
ribozymes ever will find a way into in industrial applications due to the limited diversity
of reactions that can be catalyzed and their relatively low reaction rates.

5.5 FERMENTATION OF BIOCATALYSTS

Whether a biocatalyst consists either of an isolate from a natural environment, of a

classically improved strain or of a strain improved by various types of genetic
engineering, the biocatalyst that is obtained after a screening effort will eventually be
applied in a process. In order to obtain sufficient amounts of active biocatalyst, the cells
have to be produced by fermentation.

5.5.1 General Prerequisites for the Biocatalyst

Bacteria, yeast and filamentous fungi are mostly applied as biocatalysts, since these cells
grow at a relatively high rate and since established growth conditions can be applied
using inexpensive media. Whether living or resting whole cells (bioconversion) or
enzymes in isolated or crude form are applied (biotransformation) determines how a
biocatalyst is to be prepared. General prerequisites for production of microorganisms for
biocatalytic purposes are:

• The biocatalyst must be available as pure culture

• The biocatalyst must be genetically stable (in particular recombinant systems)
• Storage of stocks of the biocatalyst without loss of quality should be possible
• Growth on inexpensive media is preferable and growth should be rapid
• Production of the desired metabolite or enzyme has to take place (may require
induction)
• High productivities are desirable
• Production of a biocatalyst must be possible on a large scale
• The biocatalyst should not be pathogenic to animals or plants

The use of a recombinant strains as biocatalysts has the advantage that once the
fermentation conditions for the host microorganisms have been found, the conditions to
be used for the recombinant strain are mostly close to these. Furthermore, the need for
induction can be removed, undesired side reactions can be avoided and enzymes from
pathogenic strains can safely be produced in recombinants of GRAS organisms.
Therefore, biocatalysts that are amenable to genetic engineering are frequently preferred
over those that are not.

5.5.2 Storage of Biocatalysts

In order to develop a reproducible fermentation process the preservation and long term
 Applied biocatalysis 210
maintenance of microorganisms, especially when these have been genetically optimized,
is essential. An number of methods are available that are all aimed at preservation of the
cells for extended periods in a state that is suited for their application in biocatalysis.
Established methods include:

• Storage at low temperatures: cells, treated with cryoprotective agents such as glycerol
(10% v/v) or DMSO (5% v/v), are stored in liquid nitrogen (appr.—200°C) or in a
freezer maintained at −60 to −80°C. Biocatalysts stored in this way can be maintained
for a period of up to 5 years.
• Freeze drying: freeze drying or lyophilisation is one of the most practiced means of
biocatalyst preservation, since it is easily performed, inexpensive and can be used to
preserve strains for periods up to 10 years. Freeze dried stocks are mostly stored at
room temperature, which does not require expensive cryo-storage equipment.
Commonly, 20% w/v skimmed milk powder or 12% w/v sucrose is used to prevent
detrimental effects of the freeze drying process to the cells.
• Spore forming microorganisms can be preserved by isolation of spores and storing
dried spores in sterile soil or on carrier beads.

An inoculum for an industrial scale fermentation is prepared from such preserved strains
by inoculating the stock in a small volume (e.g. 5mL) and performing stepwise increases
of the culture volume in order to reach the proper inoculation volume (typically 1–10%)
of the production fermenter.
 How to get the biocatalyst 211

Figure 5.16 Typical laboratory scale fermentor for batch or continuous culture
studies (Courtesy of L.H. Fermentation Ltd).

5.5.3 Process Development

To develop a process for production of the biocatalyst, small scale cultures (5mL to 5L)
are used. Initially, tubes or shake flasks are commonly applied, especially when a large
number of values for a given variable such as media composition, temperature and pH are
evaluated. As development progresses and quantitative data are required, more controlled
conditions than those provided by Erlemneyer flasks are necessary. Commonly,
 Applied biocatalysis 212
continuously stirred tank reactors (CSTR, Figure 5.16) are used, which provide a more
controlled environment regarding for instance the dissolved oxygen tension (DOT) and
pH.

Figure 5.17 Growth of a propene utilising Mycobacterium sp in batch culture

exhibiting typical growth kinetics: 1) lag phase; 2) acceleration; 3)
exponential phase; 4) deceleration; 5) stationary phase; 6) decline.

Growth Characteristics of Microorganisms

In a well mixed bioreactor a homogeneous suspension exists and typical growth kinetics
an be observed as illustrated in Figure 5.17. Six phases can be distinguished: the lag
phase; acceleration phase; the exponential growth phase; the deceleration phase; the
stationary phase and the phase at which death/decline occurs.

Lag Phase

A microorganism that is inoculated into a medium with a different composition than the
previous culture medium or is transfered from one physicochemical environment (e.g.
temperature, pH) to another does not grow immediately. Usually, a certain time is
required to adjust to the new environment before growth can continue. This is called the
 How to get the biocatalyst 213
lag phase.

Exponential Phase

Once the cells have adjusted to growth in the new environment the specific growth rate
(h−1) of the organism will gradually increase up to a maximum value max (when is
constant such as at max, then dx/dt= x where x is the cell or biomass concentration).
The actual value of max depends on the organism, available nutrients and
physicochemical conditions. In any growth sufficient medium will vary with pH and
temperature. Most unicellular bacteria exhibit broad pH optima for growth in the range
pH 6–8.5. Yeasts and filamentous fungi tend to exhibit lower pH optima (e.g. pH 4–6)
than bacteria. Growth of these organisms at low pH minimises the chance of bacterial
contamination.
Under defined physicochemical conditions the growth rate of any microorganism also
depends on its nutritional status and availability of nutrients. A single microorganism can
be cultured with a very wide range of growth rates and cells harvested after growth at
different rates can exhibit significantly different biochemical characteristics. It is often
the case that the highest specific enzyme activities ( mole product/min.mg cells) are not
found at the highest growth rates. Hence medium design and control of nutrient supply
are critical in optimising activity. Nutritional control of growth rate can be achieved in
two ways. Microorganisms will grow faster in media that contains a rapidly
metabolisable carbon source (often a simple carbohydrate) and that supplies components
such as amino acids and vitamins which would otherwise have to be synthesized de novo,
than in media without these components or with a complex carbon source. Alternatively
growth rate may be controlled by controlling the supply of a nutrient at limiting levels.
This can be seen by reference to the Monod equation for cell growth.

where S is the concentration of the growth limiting substrate and Ks is the saturation
constant, a measure of the affinity of the cell for the substrate (Ks is inversely
proportional to the affinity). In principle growth may be limited by the supply of any
essential nutrient but the choice of limiting nutrient can have a dramatic effect on cell
physiology.

Deceleration and Stationary Phase

In a closed system the growth rate of a microorganism will ultimately decline to zero. In
the absence of other effects this is bound to occur because of nutrient depletion and the
 Applied biocatalysis 214
deceleration should follow that predicted by the Monod equation. However, particularly
in shake flask cultures, the lack of control of pH can lead to cessation of growth before
nutrient depletion. Cessation of growth does not imply loss of viability and many
microorganisms can remain viable for a considerable time in this non-replicating state.
The deceleration and early part of stationary phase are characterised by a redirection of
the metabolic activity of the cell towards a period of dormancy which can include the
induction of new enzymes and decline of others directly associated with growth. With
some organisms, particularly the filamentous fungi and bacteria, depletion of a nutrient
other than the primary carbon source can lead to the induction of new and complex
pathways for the synthesis of secondary metabolites. These include many valuable
antibiotics and pharmacologically active agents. The secondary metabolite production
phase is often referred to as Idiophase as distinct from the growth or Trophophase.

Figure 5.18 Schematic of a large scale (pilot/production) fermenter with three

impellers. (From Stanbury and Whitaker 1984). Typical geometrical
ratios are given in Table 5.11.

Death/Decline

Cells in stationary phase maintain viability by a gradual utilisation of endogenous energy

 How to get the biocatalyst 215
reserves to supply vital functions. Eventually these will decline and in many cases this
leads to loss of viability. Some organisms, however, have the capacity to differentiate to
form resting stages such as spores or cysts with extended survival capacity and the
potential to regerminate in a nutritionally richer environment.

5.5.4 Bioreactor Design and Operation

Stirred tank reactors tend to dominate existing large-scale fermentation processes. The
basic design of a stirred tank reactor is illustrated in Figure 5.18, its typical geometric
ratios in Table 5.11. These can range from the 1–10 L laboratory reactor, through pilot-
scale (50–5000 L) to production scale (50,000–500,000 L). The essential elements are: a
vessel capable of being sterilised, operated and sampled aseptically for a number of days;
a system of aeration and agitation sufficient for meeting the oxygen demands of the
fermentation without damaging the organism; pH and temperature control.
Much information on microbial physiology and strain characteristics can be gleaned
from studies on a laboratory scale. When these have been adequately

Table 5.11 Typical dimensional ratios for bioreactors with 3 multibladed impellers.

Ratio Value
L/D 1–3.5
P/D 0.25–0.5
Z/P 0.8–1.0
Y/P 1.0–1.3
W/P 1.0–1.3
V/P 1.0–1.4
Baffle width/D 0.08–0.1

defined to allow prediction of the productivity and therefore scale of an industrial

process, a programme of scale-up can begin. Scale-up is a complex process requiring an
interdisciplinary combination of chemical and mechanical engineering and microbial
physiology. While the general engineering principles of scale-up have largely been
established, the behaviour of a microorganism cultured on a larger scale is more difficult
to predict (Trilli, 1986). Although it is not the intention to reiterate the principles of scale-
up here it is pertinent to make a few comments on the types of problems encountered.
Probably the major problem in the scale-up of fermentation processes lies in
maintaining aeration efficiency to aerobic cultures. Aeration is normally achieved by
sparging with air and the rate determining step in a non-viscous fermentation is the
 Applied biocatalysis 216
transfer of oxygen across the gas-liquid interface, The oxygen transfer rate (OTR) in a
fermenter is given by:

where CL, is the concentration of dissolved oxygen in the fermentation broth, C* is the
saturated dissolved oxygen concentration in the fermentation broth (under the operating
conditions), ‘a’ is the gas/liquid interface area per liquid volume and KL is the mass
transfer coefficient which can be considered as the reciprocal of the resistance to transfer
of oxygen from gas to liquid. It is difficult to measure KL and ‘a’ separately so the two
terms are usually combined to give the volumetric transfer coefficient KLa which is a
measure of the aeration capacity of the fermenter. With reference to this equation,
agitation of a fermentation broth can be seen to improve KLa by decreasing bubble size,
so maximising ‘a’ for a fixed volume of air, and increasing gas ‘hold up’ time (i.e. the
time that a bubble is in contact with the liquid). Although a target of scale up might be to
maintain the same KLa, measurement of KLa becomes difficult or non representative with
increasing bioreactor size (Trilli, 1986; Stanbury and Whittaker, 1984).

5.6 ISOLATION OF BIOCATALYSTS: DOWNSTREAM PROCESSING

5.6.1 Introduction
When appropriate, biocatalysts on an industrial scale are applied as in tact whole cells or
as cell lysates. If biocatalyst purification is not required on the basis of productivity,
specificity or product quality demands, it is usually omitted. In those cases, cells are
harvested by filtration or centrifugation and used directly or a lysate is prepared and the
biocatalyst is used without further purification (‘crude lysate’).
Frequently however, enzymes found in the fermentation broth at the time of harvesting
are quite different from the state in which they can be used as industrial biocatalysts,
reagents in clinical chemistry or as therapeutic agents, The recovery of an enzyme from
microbial culture, its concentration and purification will require careful and effective
sequential operations, which are called downstream processing (Wheelwright, 1989).

5.6.2 Downstream Processing (DSP)

The nature of DSP operations will be influenced by the type of fermentation, physical
and chemical properties of the enzyme and unwanted by-products or contaminants,
enzyme concentration and location (intracellular, extracellular), scale of operation, waste
treatment implications, enzyme stability and desired specifications and the marketable
 How to get the biocatalyst 217
price of the enzyme.
The various stages of enzyme downstream processing generally fall into the categories
depicted in Figure 5.19; i.e. recovery, low resolution purification, and high resolution
purification. The recovery stage starts with the spent medium and ends with a well
clarified supernatant, containing the enzyme protein which is suitable for the resolution
stages. If the enzyme is secreted as a soluble protein the recovery step comprises only a
concentration step to remove excess liquid. However, many enzymes are located inside
the cell mass and, therefore, a recovery procedure is required to break the cells, separate
out the debris and recover the enzyme protein, either from the liquid inside the cell or
from the cellular debris by extraction. Once a soluble concentrate is obtained, purification
consists of steps which remove impurities and increase the relative concentration or
purity of the enzyme. The methods applied in particular stages of enzyme downstream
processing are shown in Table 5.12.

5.6.3 Downstream Processing Facilitated by rDNA Technology

The content of desired enzymes at the end of fermentation processes should be as high as
possible in order to ease the downstream processing. This may be achieved by a proper
choice of microorganisms, optimising the fermentation conditions (e.g. culture media,
inducer, morphology, growth rate) or by RDNA technology.
Applied biocatalysis 218
 How to get the biocatalyst 219

Figure 5.19 General downstream process flow chart (adapted from

Wheelwright, 1989).

Table 5.12 Various stages and methods involved in enzyme downstream processing.

Stages Steps Typical methods

Recovery Pretreatment of Adjustment of temperature, pH, ionic strength, addition
fermentation broth of flocculation agents
Extracellular enzyme: Centrifugation
Cell removal (solid Filtration
liquid separation)
Intracellular enzyme: Centrifugation
a) cell recovery Filtration
(solid—liquid
separation)
b) cell desintegration Mechanical: high pressure homogenization
Non-mechanical: osmotic shock, organic solvent, enzyme
lysis
c) debt is removal Centrifugation
(solid-liquid separation) Filtration
Aqueous two-phase separation
Purification Intra- and extracellular
enzyme:
Concentration
a) precipitation Inorganic salts, organic solvents, charged polymers,
b) ultrafiltration affinity precipitation
(reverse osmosis)
c) adsorption anionic and cationic matrices alumina, celite, calcium,
phosphate
High d) drying evaporation, spray drying, freeze drying Ion-exchange
resolution chromatography,
Hydrophobic interaction chromatography, Affinity
chromatography, Gel filtration, Aqueous two-phase
separation

Overexpression
 Applied biocatalysis 220
Molecular cloning is now a standard procedure to overproduce specific enzymes of use in
biocatalysis in a ‘host’ microorganism that is suited for the process. The host strain
should fulfil the objectives of downstream processing, which are high recovery, high
purity, reproducibility and low cost scale-up.
E. coli is a popular host for expression of foreign genes and numerous gene expression
systems for in this microorganism have been developed. Alternative expression systems
are yeast, Aspergillus sp., Kluyveromyces lactis, Bacillus subtilis, Pseudomonas sp., and
Streptomyces sp. One of the problems with overexpressing foreign genes in E. coli
recombinants in particular is that overproduced proteins can be present in an inactive,
non-native form called inclusion bodies that are of limited use in biocatalysis (Kane and
Hartley, 1988). The benefit of inclusion formation for downstream processing is that the
proteins of interest are generally between 20–80% pure within the inclusion bodies. For
high value therapeutic proteins that can be refolded into a an active structure by using
solubilization agents such as guanidinium chloride, urea, detergents, alkaline pH or
organic solvents (Marston, 1986), this is beneficial. For biocatalytic purposes, the costs of
refolding a inclusion body protein to an active enzyme are mostly too high to consider
production of biocatalyst in such a manner. In such a case it is better to prevent the
formation of inclusion bodies by changing the culture conditions (e.g. growth at 28°C) or
use a different host, which is less prone to inclusion body formation or which excretes the
protein of interest.

Heterologous production of an enzyme

The overproduction of an enzyme that is common to that host itself is an example of

homologous overexpression of a gene. In contrast, heterologous expression is meant
when a gene normally not present in a host microorganism has been introduced. For
instance a gene from eukaryotic origin expressed in a prokaryote such as E. coli. The
heterologous overproduction of polypeptides in E. coli may be unsuccessful because the
protein is unproperly folded or recognized as foreign and degraded. Proteolysis in E. coli
can be prevented or significantly reduced by changing the E. coli host strain (e.g.
protease deficient mutants), by altering the cellular location of the foreign protein or by
making hybrids of the heterologous gene with a bacterial gene, thus producing a fused
protein product. The problem of proteolysis can also be tackled by chemical means by
using serine protease inhibitor such as phenyl methyl sulfonyl fluoride (PMSF), since
among the eight proteases found in E. coli six are of the serine type (two are
metalloenzymes).
The protection of over-produced foreign protein in E. coli has for example been
accomplished by making hybrids of the foreign and the galactosidase gene. After
recovery of such fusion proteins, an additional step, is necessary to cleave the fused
polypeptide off in order to isolate the desired polypeptide. This is most conveniently done
by specific proteases (Table 5.13). Convenient cleavage sites with a unique recognition
site can be placed between the C-terminus of the prokaryotic sequence and the N-
terminus of the foreign sequence. Some linker amino acid sequences and the means to
 How to get the biocatalyst 221
cleave at the respective positions are listed in Table 5.13.
Alternatively, cleavage can be done chemically, but here specificity of cleavage at a
single defined position is complicated, since most chemical cleavage methods react with
single amino acids (Table 5.13). For instance, chemical cleavage at the C-terminal side of
methionine residue can be effected using CNBr, but the use of this method is limited,
since most proteins are likely to contain internal methionine residues.
The enzymatic methods are most promising, but on the other hand chemical methods
are relatively independent of tertiary structure of the cleavage site and have technical
scale-up advantages. Chemical methods are therefore very well suited for production of
bioactive peptides cleaved off from a fusion protein.

Affinity tagging

The fusion method is very advantageous for downstream processing, especially when the
fused protein is produced in soluble form, thus circumventing the need for renaturation.
In addition, it allows the insertion of affinity tails or tags by fusion of

Table 5.13 Specific cleavage methods for proteins. The peptide bond susceptible to the
cleavage is marked, (adapted from Nilsson et al., 1988)

CLEAVAGE SITE METHOD

CHEMICAL METHODS:
Asn-Gly Hydroxylamine at pH 9
Asp-Pro Weak Acid
Cys- 2-nitro-5-thiocyanobenzoate
His- N-bromosuccinimide
Met- Cyanogen bromide
Phe- Li/Methaneamine/N-bromosuccinimide
Trp- N-bromosuccinimide
Tyr- Bromide/HCl
ENZYMATIC METHODS:
HisXxx- Ala64-Subtilisin
Tyr- Chymotrypsin
ProXxx-GlyPro Collagenase
Gly(Asp)3–5Lys- Enteropeptidase
IleGluGlyArg- Factor Xa
 Applied biocatalysis 222

-IleValGlyGlvThrVal Factor XIIa

GluAlaGluAla- Yeast -mating factor
ProPheArg- Plasma kallekrein
Glu- Staphylococcal V8 protease
GlyProArg- Thrombin

the gene of interest to a gene encoding a protein or peptide with strong affinity to a
ligand. The expressed fusion protein can then be purified easily by binding to the ligand,
followed by site specific cleavage at the junction between the two proteins. The ideal
affinity tail or tag should be:

• Strong, but weak enough to be disrupted without denaturing the proteins

• Preferentially small and monomeric
• Not be interfering with secretion
• Stable to proteolysis

The junction between the affinity tail and the product protein should be in a flexible
region to enable independent folding of two polypeptides and to facilitate the cleavage
(Nilsson and Abrahamsen, 1990). Some affinity tails and tags are shown in Table 5.14
and described below.
Staphylococcal protein A (SPA), with its immunoglobulin binding ability, has been
used as an affinity tail for the purification of a human insulin growth factor-I (IGF-1)
fusion. The insertion of an acid-labile Asp-Pro cleavage site at the fusion point allowed
the separation of the protein A moiety from IGF-1 (Nilsson et al., 1985).
An export-affinity fusion vector containing the gene cgt encoding cyclomaltodextrin
glucanotransferase (CGTase) from Bacillus circulans var. alkalophilus has been
described

Table 5.14 Affinity tails useed for purification of gene products by gene fusion.

Affinity tails Ligand Secretion References

-galactosidase TPEG No Ullman, 1984

Protein A (SPA) IgG Yes Moks et al., 1987

Poly Arg ion exchanger No Sassenfeld and Brewer, 1984
Smith et al., 1984
Protein A, Protein G IgG No Strandberg et al., 1991
 How to get the biocatalyst 223

Synthetic domain of SPA IgG Yes Stahl et al., 1991

Synthetic domain of SPG IgG, HSA Yes Stahl et al., 1991
Cyclomaltodextrin
glucanotransferase -cyclodextrin Yes Hellman and Mantsala, 1992
Maltose binding protein crosslinked amylose Yes Bedouelle and Duplay, 1988
Endoglucanase cellulose Yes Grenwood et al., 1984
AspTyrLysAspAspAspAspLys Monoclonal Antibody Yes Hopp et al., 1988
Chloramphenicol
acyltransferase Chloramphenicol No Nilsson et al., 1988

by Hellman & Mantsala (1992). CGTase binds to various sugars polymers, enabeling
purification to near homogeneity in a single step using -cyclodextrin ( -CD) coupled
agarose. The hybrid proteins can be eluted from the column with a -CD solution under
mild conditions at pH 7.5. Collagenase, factor Xa and thrombin are potential proteases
for site specific cleavage of the CGTase hybrids to release the product from the CGTase
affinity tail, since there are no recognition sites for these proteases in CGTase.
Systems to generate fusions to the gene encoding the maltose binding protein (malE) of
E. coli, which is exported to periplasmic space, have been developed. MalE-fusions can
be purified by the affinity of the MalE portion to crosslinked amylose (Bedouelle and
Duplay, 1988). Similarly, the cellulolytic enzyme endoglucanase (encoded by the cenA
gene) allowed, via a gene fusion, the binding of alkaline phosphatase to cellulose.
The currently most used tag to facilitate protein detection and purification is the
hexahistidine tag, which makes use of the affinity of stretches of histidines to divalent
metal ions in combination with immobilized metal affinity chromatography (IMAC) over
a Ni++ nitrilotriacetic acid agarose (Bush et al., 1991; Janknecht et al, 1991). Specific
monoclonal antibodies directed against the His tag can be used to monitor gene
expression. Alternatively, fusions to polyarginine tags facilitate purification due to the
unusual basicity of the fused protein and its strong binding to cation-exchange matrices
(Sassenfeld and Brewer, 1984; Smith et al., 1984).
Another system for purification of recombinant proteins expressed in yeast and E. coli
is based on the production of fusion proteins that bear an 8 amino acid N-terminal
extension (‘flag’) which can be bound by a monoclonal antibody in a calcium dependent
manner. The Lys-Arg pair proceeding the flag segment is recognized by the KEX2
processing protease, which removes the leader from the desired protein. The flag fusion
product is secreted into the culture medium, where it can be identified by its biological
activity or by anti-flag antibody. The proteins
 Applied biocatalysis 224

Table 5.15 Lytic systems available for use in E. coli to facilitate downstream processing.
(Adapted from Osborn and Cooney, 1990; Osborn, Eberiel and Cooney, 1989).

Cloned system
Source of gene promotor Triggering Mode of action Remarks
lyric system agents
Autolytic Osmotic shock, Degradation of Lysis in exponentially
enzyme of E. cephaloridine, cell wall growing cells,
coli mocenomycine, peptidoglycans to inhibition of lysis at
infection with different extent 10 mM Mg2+
phage phiX174,
induction of the
cloned phiX174
lytic genes
Colicin lytic Not fully, High lethality and
enzymes elucidated. instability of the
Control of action Colicin lytic genes
difficult
Colicin E 1 kil lac Temperature shift Alteration of Inhibition of lysis at,
above 30°C membrane 10–20 mM Mg2+
permeability
Colicin E 2 celB lac Temperature shift Degradation of Slow release of -
to 42°C bacterial DNA galactosidase, higher
release, alkaline
phosphatase
inhibition of lysis at
10–20 mM Mg2+
Colicin E 3 hic/cel own Mitomycin C Inhibition of Non-specifying
C promotor protein synthesis, release of cellular
gene product proteins
associate with
membrane
Cloacin DF 13 gene Mitomycin C Inhibition of Gene H necessary for
H protein synthesis, both the lytic and
gene product killing effect upon
associate with induction
membrane
Bacteriophage genes lac Isopropyl-p-D- S gene product S and R gene
Lambda S.R.R thiogalactoside alters cell products are essential
z (IPTG) membrane and for lysis, Rz gene
transports the R products are
 How to get the biocatalyst 225

and R, gene necessary for lysis in

products to the medium containing
periplasm to high Mg2+
allow cell wall concentration.
degradation by Addition of cyanide
them (10 mM),
chloramphenicol (40
g/ml) or chloroform
after induction of
cloned genes caused
almost immediate
lysis. Intracellular
proteins are
efficiently released
from both logarithmic
and stationary phase
Bacteriophage gene e lac II IPTG Muramidase is Gene t is required for
T4 coded. lysis and seems to
Degradation of have the same
cell wall function to the
lambda S gene
Bacteriophage gene lac IPTG Gene E product Gene E product has
phiX174 E triggers the E. no lysozyme like
coli autolytic activity inhibition of
system lysis at 50 mM Mg2+
or at 6>PH>8
gene pl Temperature shift 97% of -
E from 30°C to 42° galactosidase was
C released after 60 sec.
sonication

can be separated from the flag by proteolysis with enterokinase to yield the flag-free
product in the active form.

Induced cell lysis

Another possible application of recombinant DNA in downstream processing of

intracellular proteins is facilitation of cell lysis by using cloned lysis genes or by
activating natural autolytic enzymes in the producing strains. The lysis of the cells would
occur in a controlled manner after being triggered by external factors such as temperature
or chemical induction. The lytic system may offer an alternative to existing methods of
 Applied biocatalysis 226
cells disruption or may be used in conjunction with them. In Table 5.15 some of the lytic
systems available for use in E. coli are presented.

5.6.4 Recovery of Enzymes

Pretreatment of Fermentation Broth

When after a fermentation, the cell suspension leaves the fermenter, it is cooled down to
maintain enzyme stability and activity and to prevent introduction of contaminating
microorganisms. One may be tempted to change the pH to achieve better separation of
solids, but care should be taken that changes in pH value do not lead to reduced enzyme
stability. The use of chemical preservatives is frequently not acceptable at this stage, due
to cost arguments and because most preservatives are undesirable or prohibited in the
final enzyme preparation. The ionic strength of the medium should be kept low when
adsorption to various matrices are considered and medium constituents or pH control
agents should be selected with this point in mind. The high ionic strength of the broth
may require dilution with demineralized water, which is disadvantageous for processing,
since larger volumes cause increased handling problems, require larger equipment sizes
and lead to an increased overall processing time.

Solid—Liquid Separation

The separation of solids from liquids is a central operation in enzyme downstream

processing and is one of the, most difficult aspects of large scale processes. It includes
the separation of biomass from fermentation broth, the removal of cell debris and the
collection of enzyme precipitates.
The solids of the fermentation broth are often colloidal in nature and are difficult to
remove directly without coagulating or flocculating agents. Cell flocculation can be
improved by neutralization of the charges on the surfaces of microbial cells. This
includes changes in the pH and the addition of a range of compounds which alter the
ionic environment. Flocculating agents include inorganic salts, mineral hydrocolloids and
organic polyelectrolytes. In some cases it may also be necessary to add filter aid, in the
form of diatomaceous earth for example, before filtration.

Table 5.16 Some characteristic of centrifuges.

Type of Advantages Disadvantages

Centrifuge
 How to get the biocatalyst 227

Perforated Good dewatering, easy to clean, Limited solids capacity, solids

basket washing of cake possible recovery laborious, low centrifugal
force, discontinuous
Decanter Suitable for slurry with high solids Low centrifugal force
centrifuge concentration, high input of slurry,
continuous operation
Disc bowl Partly continuous operation, high Poor dewatering—works only with
centrifuge centrifugal force, liquid discharge under low solids content
pressure, laborious cleaning
Nozzle bowl Continuous operation, high centrifugal Poor dewatering
centrifuge force
Multichamber Increase in solids capacity, no loss of Solid recovery laborious,
bowl efficiency up to complete filling of discontinuous operation
chambers
Tubular bowl Good dewatering, high centrifugal Limited solids capacity, solids
force, easy to clean, bowl easy to recovery laborious, discontinuous
remove operation

Centrifugation

Centrifuges that operate in batch or continuous mode, which vary in their manner of
liquid and solid discharge, their unloading speed and their relative maximum capacities
are available. When choosing a centrifuge for a specific process it is important to ensure
that the centrifuge will perform the separation at the planned production rate and will
operate reliably with minimum manpower (Wheelwright, 1989; Wang et al., 1979;
Stanbury and Whitaker, 1984). Some of the critical characteristics of different centrifuges
are summarized in Table 5.16. Different types of centrifuges, such as the perforated bowl
basket centrifuges, the solid bowl scroll centrifuge, the disc-bowl centrifuge, the multi-
chamber centrifuge, and the tubular bowl centrifuge are shown in Figure 5.19.

Filtration

Filtration separates components according to their size. Efficiency depends on the shape
and compressibility of the particles, the viscosity of the liquid phase and the driving
force, which is the pressure created by overpressure or by vacuum. Filtration can be
performed either as dead-end filtration, where the feed stream flows perpendicular to the
filter surface (Lee, 1989) or as tangential flow filtration, where the feed stream flows
parallel to the filter and the filtrate diffuses across it. Examples of the former are the
continuous rotary vacuum drum filter, where a rotary vacuum filter has a filter medium
covering the surface of a rotating drum and the filtrate is drawn through the drum by an
 Applied biocatalysis 228
internal vacuum and the plate and frame filter press, a batch type filter.
Tangential flow filters make use of the fact that a tangential flow or cross flow
generates high shear across the membrane surface which prevents filter clogging by
suspended solids. Cross flow filtration can be applied to recover the cell suspension or
the cell-free filtrate as a product. Designs of filter housing include flat sheet (plate and
frame or spiral configuration), hollow-fiber, tubular or rotating filter units and the
membranes are constructed from polymer materials, such as cellulose ester,
polypropylene, polyvinyl chloride, polysulphone, polyacrylic or polycarbonate.
Problems encountered with filtration are that membrane fouling can occur, which
causes a decline in flux with time under constant operating conditions. Furthermore,
concentration polarization, the effect that the increased concentration of components on
the membrane surface reduces the flux due to the additional hydrodynamic resistance, is
observed. This effect can be minimized in cross-flow filtration, by applying high flux
rates across the membrane surface (Wang et al., 1979; Lee, 1989).

Cell Disintegration

Various enzymes are produced intracellularly. Hence, following cell harvesting, an

efficient disruption process to disintegrate the cell to release the intracellular proteins is
needed. Some types of cells are broken readily by gentle treatment, while others are very
resistant to breakage. A number of cell disruption methods have been developed:

Physical-Mechanical Methods

Liquid shear. The basis of the technique is that a suspension of cells (about 12% w/ v) is
forced through a small orifice under very high pressure (up to 140 MPa). Cell disruption
in a Manton Gaulin homogenizer appears to be caused by shear forces due to passage
through a small channel combined with a disruptive effect due to the sudden drop in
pressure as the suspension passes into the outlet. The disruption is proportional to the
number of passes (N) and the operating pressure (P) and can be described by the
following equation:

where R is the fraction of the cells disrupted, K is a dimensional (MPa−a) disruption

constant and a is a measure of cell resistance to disruption (e.g., for S. cerevisiae, C. utilis
and E. coli the value of a equals 2.9, 1.77, 2.2 respectively).
Solid shear. In this technique, a frozen cell paste at −20°C is forced through a narrow
orifice under very high pressure. The shear forces exerted by the passage of the extruded
paste is aided by ice crystal formation in the frozen paste. On the laboratory scale the
Hughes press or X-press are used.
 How to get the biocatalyst 229
Agitation with abrasives. The rapid agitation of a microbial cell suspension with glass
beads or similar abrasives forms the basis of these techniques. A typical apparatus is the
Dyno-Mill. The rate of cell breakage is dependent on the size and concentration of the
beads, the type, concentration and age of cells, the agitator speed, the flow rate through
the chamber, the temperature and the arrangement of the agitator discs.
Freezing and thawing of a microbial cell paste causes ice crystal formation and
subsequent thawing results in some disruption of cells. Although the technique has the
advantages of apparent simplicity and low-temperature operation it is not extensively
used.

Chemical Methods

Detergents. Under appropriate conditions of pH, ionic strength and temperature,

detergents (ionic: sodium lauryl sulphate, sodium deoxycholate, sodium cholate and
cetyldiethyl-ammonium bromide, or nonionic: Tweens and Tritons), can be used to lyse
cells. Detergents may however cause enzyme inactivation and may need to be removed
before purification.
Osmotic shocki. This technique involves washing of the cells in buffer solution to
remove growth medium and resuspension in 20% buffered sucrose. After equilibration,
the cells lose some internal water. The cell paste obtained after centrifugation is rapidly
dispersed in water at approximately 4°C. A sudden decrease in sucrose concentration will
cause disruption of cells. The same effect can be achieved with decrease of salt
concentration.
Alkali treatment. Alkali treatment may be used for hydrolysis of cells provided that the
desired enzymes will tolerate a pH of 11.5 to 12.5 for 20 to 30 minutes.
Enzyme Treatment. There are a number of enzymes which hydrolyze the microbial cell
wall constituents. Enzymes exhibiting these activities include lysozyme, enzyme from
snail extract and lytic enzyme systems of microbial origin composed of proteases,
mannases, 1,3- -glucanases and chitinase. Although the method is probably the most
gentle available it is not an economical method for large scale application due to the cost
of the lytic enzymes (Kula and Shutte, 1987).

Extraction

In many cases, enzymes of interest are associated with insoluble parts of the cell, such as
mitochondria, endoplasmic reticulum or membranes, and an extraction procedure is
necessary. Not all of these enzymes can survive solubilization in the absence of their
normal cellular milieu. The extraction buffers should not only have the correct pH and
ionic strength but also the appropriate stabilizers should be added (Table 5.17).
 Applied biocatalysis 230

Cell debris and nucleic acid removal

Removal of the debris produced by mechanical disruption can be improved if the particle
diameter is increased. This can be done either by flocculation, coagulation or by binding
to the various chemical bioprocessing aids such as polyvinylpyrrlidone, Whatman CDR
(cell debris remover), chitosan or Biocryl bioprocessing aids (BPAS). The adsorbed
materials can be removed by filtration or centrifugation.
Nucleic acids can be removed from a lysate by precipitation with a variety of
compounds, such as cetyltrimethylammonium bromide (CTMAB), streptomycin
sulphate, prolamine sulphate, polyethylenimine, polylysine or manganese chloride.

Table 5.17 Some of the reagents used in extraction buffers.

Compounds Uses
EDTA Chelators of cations, particularly heavy metals
EGTA Specific for Ca++
Dithiothreitol Protection of active sites
Mercaptoethanol sulphydryl groups from oxidation
Tween 20 Solubilization of membrane bound
Triton X-100 proteins or disruption of vesicles
Substrates, Help to stabilize the enzyme
Competitive inhibitors against heat inactivation or extremes of pH
PMSF, EDTA, Inhibitors of degradative enzymes
Alkylating reagents such as proteinases and certain glycosidases

Precipitation of protein may also occur in some cases, resulting in loss of enzyme
activity. Nuclease treatment is a very effective method of nucleic acid removal (Melling
and Phillips, 1975), but of limited use for enzymes which are intended for applications in
molecular biology.

5.6.5 Isolation and Purification

Concentration

The amount of extracellular enzyme in culture filtrate or intracellular enzyme in cell-free

extract is usually low. Moreover, the enzyme to be purified represents a fraction of total
protein. A variety of techniques has been devised for water removal and the majority of
these effect partial purification of the enzyme in question (Melling and Phillips, 1975;
 How to get the biocatalyst 231
Bell, Hoare & Dunnill, 1983).

Precipitation

Salts. Salts trap water and unveil the hydrophobic areas on the surface of enzyme
molecules resulting in aggregation their precipitation. Salting out depends on
hydrophobic interaction, alteration of pH or ionic strength, polarity and temperature. The
most commonly used reagents are ammonium and sodium sulphate.
Organic solvents. Addition of organic solvents decreases the solubility of proteins by
reducing the dielectric constant of the medium. For the precipitation of enzymes,
methanol, ethanol or propanol are mostly used, but acetone and diethyl ether can also be
employed. The principal disadvantage of organic solvents is their tendency to cause
structural damage of enzyme molecule.
High molecular weight polymers. Other organic compounds which can be used for the
precipitation of enzymes are water-soluble, neutral polymers such as polyethylene glycol
(PEG). These have the advantages of being non-toxic, non-flammable and non-
denaturing to proteins.
Affinity precipitation. In this technique, specific interactions between enzymes and
their complementary ligands are exploited. The ligand is attached to a soluble carrier and
this forms complexes with the desired protein, which precipitates out of the solution. The
precipitate is washed with buffer, to remove entrapped contaminants and the purified
protein is dissociated. The precipitation depends upon the nature of the macroligands,
which are either homo-bifunctional ligands, containing two identical ligand molecules on
a spacer molecule (e.g. N2N2′− adipodilhydrazido-bis-(N6-carbonylmethyl-NAD), or bis-
Cibacron Blue F3GA) or hetero-bifunctional ligands. The latter ligands have two
functions, one for binding the target proteins and the other for causing precipitation. The
macroligands are designed in such a way that the precipitation is reversible and induced
by changing the solution temperature or pH. Chitosan and polyacrylamide with benzoic
groups are examples of the carriers for preparing hetero-bifunctional ligands.

Adsorption

Some of the high cost enzymes are concentrated by adsorption chromatography. After
adsorption, the adsorbent with bound enzyme is centrifuged and washed in order to
remove unbound protein and solutions. Elution is performed by manipulation of pH
and/or ionic strength.

Ultrafltration

Ultrafiltration separates macromolecules in the molecular weight range from 1000 to

about 100,000. The membranes used are characterized by a nominal molecular weight
 Applied biocatalysis 232
cut-off, and may be isotropic, with the pore openings equally small on both sides of the
membrane, or anisotropic, with the conically shaped pores characterized by smaller
openings at the process side and larger outlet at the permeate side. With the latter
membranes the faster flow rates and more precise control of the process can be achieved.
Membranes are made from different materials. Cellulose nitrate ultrafiltration
membranes are of limited chemical and thermal compatibility and with imprecise cut-off.
The membranes from polyvinylidene fluoride, polyacrylonitrile or polysulphone possess
good chemical compatibility are more stable over a very wide pH range and are easy to
clean. Problems with membrane fouling can usually be overcome by treatment of the
membranes with detergents, proteases or with acid or alkaline solutions.
Ultrafiltration results in little loss of enzyme activity and can be used for concentration
and fractionation on basis of molecular weight, for removal of salts and low molecular
weight species, as well as for changing salt composition by diafiltration.

Affinity crossflow filtration

The separation of biomolecules with a 10-fold difference in molecular weight is about the
best that can be achieved using ultrafiltration. A highly purified protein can be obtained
by combining affinity interaction and membrane separation: affinity cross-flow filtration.
A target protein binds to a specific ligand immobilized on a high molecular weight carrier
which can be retained by the membrane. The unbound contaminants pass through the
membrane and afterwards the target protein is desorbed. The ligand used in affinity cross-
flow filtration can be bound to water-insoluble particles, such as starch granules, agarose
beads, liposome, silica nanoparticles, sepharose beads or dead microbial cells, or to
water-soluble high molecular weight polymers, such as dextran or polyacrylamide.

Drying/Evaporation

Drying or evaporation is usually applied to sufficiently thermostable enzymes whose

activity is not affected by increasing concentrations of other solutes produced by these
techniques from crude culture filtrates. Freeze drying is slow and expensive, but may be
used to obtain stable enzyme preparations, which can be stored easily for further
processing. Spray drying is used in many industrial enzyme production processes to
obtain dry, powdered final product. Rotary, falling or climbing film evaporators may be
used for concentration of enzymes and since these operate under vacuum at temperatures
below 40°C, heat denaturation is avoided. The tendency to foam causes problems during
evaporation and enzyme solutions should therefore be degassed.

5.6.6 Chromatography
The basis of chromatography is the differential distribution of solutes in a column packed
with a support (the stationary phase) through which an eluent solvent or buffer (the
 How to get the biocatalyst 233
mobile phase) is flowing. Successful application of chromatographic techniques relies on
the resolving power. Each of the various modes of chromatographic separation have
unique advantages that dictate where and when in the purification process these
techniques should be used.

Ion Exchange Chromatography

Ion exchange chromatography (IEC) is the most versatile and commonly applied
chromatographic procedure in the initial purification steps due to its ability to handle
large volumes of feed in a short time with a relatively small amount of a support. In this
technique the biomolecules are separated according to differences in their charge. At a
given pH (one unit below or above the pI value of the enzyme of interest) a mixture of
enzymes carrying different net charges, in a buffer solution of low ionic strength, is
applied to the column packed with a gel bearing either negative (cation exchanger) or
positive charges (anion exchanger) (Table 5.18). Molecules with a net charge opposite to
that of support are bound reversibly to the support by electrostatic forces. The bound
molecules are eluted from the column by an increasing salt or pH gradient. IEC has a
great resolving power and can be applied to purify a number of proteins under a variety
of conditions. Optimization of the separation is performed by changing pH, ionic strength
and elution modes (stepwise or linear gradient). Effectively designed IEC procedures can
give tremendous product enrichment from just one chromatographic step and are

Table 5.18 Ion exchange substituents.

Cation exchange
Carboxymethyl (CM)

Sulfoethyl (SE)

Sulfopropyl (SP)

Phosphate (P)

Sulfonate (S)

Anion exchange
Diethyloaminoethyl (DEAE)

Quaternary aminoethyl (QAE)

Quaternary amine (Q)

 Applied biocatalysis 234

relatively inexpensive since simple salts are commonly used for product elution. In
addition, since base-stable IEC gels are commercially available, cleaning and
depyrogenation are effected with inexpensive agents, such as sodium hydroxide or
ethanol.
After ion exchange chromatography the enzyme preparation is in a more workable
volume and is then ready for further purification. Depending on the requirement for the
next step, the enzyme solution can be desalted by gel filtration or ultrafiltration or the
buffer can be changed by diafiltration.

Hydrophobic Interaction Chromatography

Hydrophobic interaction chromatography (HIC) separates biomolecules according to

differences in their hydrophobicity. A crude enzyme solution with high salt concentration
is applied to a column with hydrophobic ligands. Hydrophobic regions of the enzyme
molecules bind to the hydrophobic ligands (e.g. octyl- or phenylagarose). Hydrophilic
molecules are washed from the column. The enzymes adsorbed to the hydrophobic
matrix are eluted using a gradient of decreasing salt concentrations, by varying the pH or
by decreasing of polarity of elution buffer by addition of ethylene glycol.
Hydrophobic interaction chromatography also has very high resolving power and can
be used as a concentration step during purification. The necessity of high ionic strength to
promote protein adsorption makes HIC a good technique to follow a salt precipitation
step. This technique can be used in batch mode but, generally, it is not the easiest or
cheapest technique to use at process scale.

Affinity Chromatography

Affinity chromatography is a very powerful technique which separates biomolecules

according to differences in their biological function or chemical structure. The stationary
phase for affinity chromatography consists of a matrix to which ligands are covalently
attached.
The ligands used for enzyme purification can be specific to the desired enzyme
(substrate, substrate analogue, enzyme inhibitor, antibody), specific for different classes
of enzyme (AMP, NAD, PLP) or of limited predefined specificity (dye affinity
chromatography with Procion, Cibacron dyes).
 How to get the biocatalyst 235

Table 5.19 The media used for separation in gel filtration.

Material Example
Cross-linked dextran Sephadex
Cross-linked polyacrylamide Biogel-P
Cross-linked agarose Superose
Sepharose CL
Cross-linked beaded cellulose Cellulofine
Agarose Sepharose
Bio-Gel A
Composite of polyacrylamide and dextran Sephacryl
Composite of polyacrylamide and agarose Ultrogel AcA
Hydroxylated acrylic polymer Trisacryl
Ethylene glycol—methacrylate copolymer Fractogel
Porous glass CPG
Bio-Glass

Affinity chromatography is very attractive because of the simplicity of its operation and
very large volumes can be processed on a relatively small column in a short time,
yielding small volumes of a concentrated enzyme solution with a high purity. The major
drawbacks of affinity chromatography are the high cost of affinity supports and the
lability of some affinity ligands (e.g. proteins, antibodies). This chromatographic
technique can be applied at any stage of purification, but is not normally recommended at
an early stage.
Gel Filtration

In gel filtration chromatography the enzymes are separated according to their size. The
separation depends on the different abilities of the particular protein molecules to enter
the pores within the gel beads. Large molecules, which cannot enter even the largest
pores, pass through the column faster. Molecules are eluted in order of decreasing size,
which for globular components, corresponds well to their molecular weight. Common gel
filtration materials are shown in Table 5.19.
Of all the techniques presented in this chapter, gel filtration offers the lowest resolution
and causes dilution of the enzyme. To achieve significant separation of the enzymes in
solution, the sample volume should be no greater than five percent of the column bed
volume. Of the purification procedures, gel filtration is usually selected as the last
resolving step, when the sample volume is smaller and the fractionated mixture is less
complex. Concentration of the enzyme is necessary before further processing. Gel
 Applied biocatalysis 236
filtration is suitable for fast buffer exchange, is easy to use, and the results are easy to
interpret.

Control and Documentation of Downstream Processing

The analysis of enzyme purity and activity after an individual purification step is vital to
assess the effectiveness of this separation technique. An increase in specific activity of
purified enzyme after each purification step is the indicator of efficiency. The purity of
enzymes can be assessed by a variety of techniques, including polyacrylamide gel
electrophoresis (PAGE), SDS-PAGE, isoelectrofocusing, gradient gel electrophoresis,
immunoelectrophoresis, blotting or HPLC. A material balance should be constructed for
the whole downstream processing operation, illustrating the purity and recovery yield of
the enzyme. It must be fully documented in such a way that anybody can obtain the same
product by following the same procedures.

5.7 ACKNOWLEDGEMENTS

The authors wish to acknowledge Will van den Tweel and David Leak for their
contributions to the first edition of ‘Applied Biocatalysis’, which in part have been
integrated in this chapter.

5.8 QUESTIONS

5.1. When screening for a novel activity, one preferably starts with for instance soil
samples with a high degree of diversity. Preferably, soil from biodiversity hot spots is
used. What are these and why is the chance of finding a desired activity higher in such
areas? Could you think of a sampling site with lower diversity that nonetheless can be
suited for input in a screening exercise?
Non-culturable microorganisms, which cannot be cultured in the laboratory but which
constitute the vast majority of microbial life, can be a valuable source of biocatalysts.
How would one proceed to get access to enzymes from such microorganisms?

5.2. Multi-enzyme systems can be used for ‘one pot’ synthesis of desired compounds,
which is especially advantageous when nature does not provide the right pathway.
Enzymes that are not found together in any living system can be combined to catalyze the
desired series of reactions. What problems may occur when enzymes obtained by
screening from different sources are used in a ‘one pot’ synthesis?
 How to get the biocatalyst 237

5.3. Catalytic antibodies or ‘abzymes’ are antibodies that have been developed for
biocatalytic reactions. How are ‘abzymes’ produced? Why is it essential to know the
detailed mechanism of a reaction for development of a catalytic antibody to catalyze it?
What are the main limitations of ‘abzymes’ that have restricted the industrial use of such
catalysts?

5.4. Aspartame is an alternative sweetener that is produced via a biocatalytic route.

Aspartame is composed of two L-amino acids, which? Why can’t aspartame be made out
of racemic amino acids? How could one find an enzyme to couple the two amino acids
(think of hydrolases, even though these normally cleave peptide bonds)? What criteria
should one have in mind when doing so? How could one reverse the peptide cleavage
reaction? Would you use the purified enzyme to produce aspartame?

5.5. The potential of lyases for the synthesis of optically active compounds are of
commercial interest, because these enzymes are stereospecific and do not require
complicated cofactor recycling procedures. What types of reactions are catalyzed by
lyases? Lyases typically catalyze reversible reactions. How can you push the equilibrium
in the desired direction?

5.6. Mention some of the advantages and disadvantages of using isolated enzymes
rather than whole cells in biocatalysis. Is it always needed to purify enzymes for a
biocatalytic application? To what degree should one purify?

5.7. The purification of proteins is typically performed by chromatographic techniques

such as ion exchange chromatography, hydrophobic interaction chromatography, metal
chelate chromatography, covalent chromatography, affinity chromatography or gel
filtration (permeation) chromatography. The specific enzyme activity and the degree of
purification are two quantities that describe the purification process. What are the
physicochemical characteristics of proteins that are exploited for each of the
chromatography types described above. How are the ‘specific enzyme activity’ and the
‘degree of purification’ defined? Lysozyme is a small protein (129 amino acids, 14,600
Da) with a high isoelectric point (11.0). Design a strategy to purify this enzyme based on
these two characteristics: which of the chromatographic techniques described above
would you use and in which sequence?

5.9. HINTS AND ANSWERS

5.1. See sections 5.3.1 and 5.3.3 (molecular biology techniques).

 Applied biocatalysis 238

5.2. Consider reaction conditions as well as kinetics.

5.3. See section 5.4.2.

5.4. Some hints can be found in chapter 9.

5.5. Some hints can be found in chapter 2.

5.6. Think in terms of specificity and productivity.

5.7. See section 5.6.

5.10 LITERATURE

5.10.1 Suggested Reading

Atkinson, T., Scawen, M.D. and Hammond, P.M. (1987) Large scale Industrial
Techniques of enzyme recovery. In Biotechnology 7a , edited by H.Rhem and G.Reed,
p. 279. Weinheim: VCH Verlag.
Bergquist, P.L. and Morgan, H.W. (1992) In Molecular Biology and Biotechnology of
Extremophiles , edited by R.A.Herbert and R.J. Shar, pp. 44–75. Glasgow and London:
Blackle.
Bull, A.T., Goodfellow, M. and Howard Slater, J. (1992) Biodiversity as a source of
innovation in biotechnology. Annu. Rev. Microbiol. , 46 , 219–252.
Faber, K. (1995) Biotransformations in organic chemistry, 2nd Edition. Berlin: Springer-
Verlag.
Herbert, R.A. and Sharp, R.J. (eds.) (1992) ‘Molecular Biology and Biotechnology of
Extremophiles’. Glasgow and London: Blackle.
Higgins, J.J., Lewis, D.J., Daly, W.H., Mosqueira, F.G., Dinnill, P. and Lilly, M.D.
(1978) Investigation of the unit operations involved in the continuous flow isolation of
-galactosidase from Escherichia coli. Biotechnol Bioeng. , 20, 159–182.
Kieslich, K., van der Beek, C.J., de Bont, J. and van den Tweel, W.J.J. (eds.) (1998) New
frontiers in screening for microbial catalysis. Amsterdam: Elseviers Science B.V.
Lerner, R.A., Benkovic, S.J. and Schultz, P.G. (1991) At the crossroads of chemistry and
immunology: catalytic antibodies. Science , 252, 659–667.
Melling, J. and Phillips, B.W. (1975) Large scale extraction and purification of enzymes.
In Handbook of Enzyme Biotechnology , edited by A.Wieseman. Sussex: Ellis
Horwood Ltd.
Naumann, D., Helm, D., Labischinski, H. and Giesbrecht, P. (1991) In Modern
techniques for rapid microbiological analysis , edited by W.H.Nelson, pp. 43–96. New
 How to get the biocatalyst 239
York: VCH Publishers.
Ransac, S., Rogalska, E., Gargouri, Y., Deveer, A.M.T.J., Paltauf, F., de Haas, G.H. et al.
(1990) Stereoselectivity of Lipases. J. Biol. Chem. , 265, 20263–20270.
Stanbury, P.F. and Whitaker, A. (1984) Principles of Fermentation Technology.
Pergamon Press.
Trilli, A. (1986) In Industrial Microbiology and Biotechnology , edited by A.L.Demain
and N.Solomon, pp. 277–307. American Society for Microbiology.
Wolley, P. and S.B.Petersen. (1994) Lipases. New York: Cambridge University Press.

5.10.2 Cited References

Arnold, F.H. and J.C.Moore. (1997) Optimizing industrial enzymes by directed
evolution. Adv Biochem Eng Biotechnol. , 58, 1–14.
Atkinson, T., Barstow, D.A., Jones, S.A., Minton, N.P. and Sherwood, R.F. (1984)
Cloning enzyme genes for overproduction. Biochem Society Transaction , 1, 215–218.
Ballard, D., Blacker, A.J., Woodley, J.M. and Taylor, S.C. (1994) Polyphenylenes from
Biosynthetic cis-Dihydroxycyclohexadiene. In Plastics from microbes , edited by
D.P.Mobley, pp. 139–168. Munchen: Carl Hanser Verlag.
Bedouelle, H. and Duplay, P. (1988) Production in Escherichia coli and one-step
purification of bifunctional hybrid proteins which bind maltose. Eur. J. Biochem. , 171,
541–549.
Bell, D.J., Hoare, M. and Dunnill, P. (1983) The formation of Protein Precipitates and
Their Centrifugal Recovery. Downstream Processing. In Adv. Biochemical
Engineering , edited by A.Fiechter, pp. 1–72. Berlin: Springer-Verlag.
Benkovic, S.J. (1992) Catalytic antibodies. Annu. Rev. Biochem. , 61, 29–54.
Bialy, H. (1997) Indigo gives the blues to chemists. Nature Biotechnology , 15, 110.
Bradshaw, C.W., Fu, H., Shen, G.J. and Wong, C.H. (1992) A Pseudomonas sp alcohol-
dehydrogenase with broad substrate-specificity and unusual stereospecificity for
organic-synthesis. J. Org. Chem. , 57, 1526–1532.
Brooks, L., Suckling, C.J. and Stimson, W.H. (1996) Construction of a single-chain Fv
from an antibody which catalyses a Diels Alder cycloaddition. Biochem. Soc. Trans. ,
24, 313S.
Bush, G.L., A.M.Tassin, H.Friden and D.L.Meyer. (1991) Secretion in yeast. Purification
and in vitro translocation of chemical amounts of prepro-alpha-factor. J Biol Chem. ,
266, 13811–13814.
Butora, G., Hudlicky, T., Fearnley, S.P., Gum, A.G., Stabile, M.R. and Abboud, K.
(1996) Chemoenzymatic synthesis of the morphine skeleton via radical cyclization and
a C10–C11 closure. Tetrahedron Lett. , 37, 8155–8158.
Chenault, H.K. and Whitesides, G.M. (1987) Regeneration of nicotinamide cofactors for
use in organic synthesis. Appl Biochem Biotechnol , 14, 147–97.
Cheetam, P.S. (1987) Screening for novel biocatalysts. Enzyme Microb. Technol. , 9,
194–213.
Crameri, A., S.A.Raillard, E.Bermudez and W.P.Stemmer. (1998) Molecular evolution of
an arsenate detoxification pathway by DNA shuffling. Nature , 391, 288–291.
Dalbøge, H. (1998) Screening and expression cloning of fungal enzyme genes of
industrial relevance. In New frontiers in screening for microbial catalysis , edited by
 Applied biocatalysis 240
K.Kieslich, C.J.van der Beek, J.de Bont and W.J.J.van den Tweel, pp. 29–36.
Amsterdam: Elseviers Science B.V.
Dalbøge, H. and Lange, L. (1998) Using molecular techniques to identify new microbial
biocatalysts. TIBTECH , 16, 265–272.
Ertesvåg, H. and Skjåk-Bræk, G. (1999) In Carbohydrate Biotechnology Protocols ,
edited by C.Bucke, in press. Totowa, N.J.: Humana Press.
Gilbert, E.J. (1993) Pseudomonas lipases: biochemical properties and molecular cloning.
Enz. Microb. Technol. , 15, 634–645.
Gross, C., Syldatk, C. and Wagner, F. (1987) Screening method for micro-organisms
producing L-amino acids from D,L-5-monosubstituted hydantoins. Biotechn.
Techniques , I, 85–90.
Hellman, J. and Mantsala, P. (1992) Construction of an Escherichia coli export—affinity
rector for expression and purification of foreign proteins by fusion to
cyclomaltodoxtrin glycotransferase I. Bio/Technol. , 23, 19–34.
Hermetter, H., Paltauf, F., Zenzmaier, E., Schlacher, A., Schwab, H. and Faber, K. (1998)
Fluorescense screening for lipolytic enzymes. In New frontiers in screening for
microbial catalysis , edited by K.Kieslich,
C.J.van der Beek, J.de Bont and W.J.J.van den Tweel, pp. 53–60. Amsterdam: Elseviers
Science B.V.
Hilvert, D., Hill, K.W., Nared, K.D. and Auditor, M.T.M. (1989) Antibody catalysis of a
Diels-Alder reaction. J. Am. Chem. Soc. , 111, 9261–62.
Hudlicky, T., Endoma, M.A.A. and Butora, G. (1996a) New chiral synthons from the
microbial oxidation of bromonaphthalenes. Tetrahedron Asymmetry , 7, 61–68.
Hudlicky, T., Pitzer, K.K., Stabile, M.R., Thorpe, A.J. and Whited, G.M. (1996b)
Biocatalytic syntheses of protected D-mannose-d5, D-mannose-d7, D-mannitol-
2,3,4,5,6-d5 and D-mannitol-1,1,2,3,4,5,6,6–d8. J. Org. Chem. , 61, 4151–4153.
Hudlicky, T., Tian, X., Konigsberger, K., Maurya, R., Rouden, J. and Fan, B. (1996c)
Toluene-dioxygenase-mediated as-dihydroxylation of aromatics in enantioselective
synthesis. Asymmetric total syntheses of pancratistatin and 7-deoxypancratistatin,
promising antitumor agents. J. Am. Chem. Soc. , 118, 10752–10765.
Inoue, A. and Horikoshi, K. (1989) A Pseudomonas thrives in high concentrations of
toluene. Nature , 366, 264–266.
Isowa, Y., Ohmori, M., Ichikawa, T., Mori, K., Nonaka, Y. and Kihara, K. (1979) The
thermolysin-catalyzed condensation reactions of N-substituted aspartic and glutamic
acids with phenylalanine alkyl esters. Tetrahedron Lett. , 28, 2611–2612.
Jaeger, K.E. and S.Wohlfarth. (1993) Bacterial lipases: Biochemistry, molecular genetics
and applications in biotechnology . Bioengineering , 9, 39–46.
Janknecht, R., G.de Martynoff, J.Lou, R.A.Hipskind, A.Norandeim and
H.G.Stunnenberg. (1991) Rapid and efficient purification of native histidine-tagged
protein expressed by recombinant vaccinia virus. Proc Natl Acad Sci USA , 88, 8972–
8976.
Janssen, D.B., F.Pries and J.R.van der Ploeg. (1994) Genetics and biochemistry of
dehalogenating enzymes. Annu. Rev. Microbiol. , 48, 163–191.
Kane, J.F. and Hartley, D.L. (1988) Formation of recombinant protein inclusion bodies in
Escherichia coli. TIBTECH , 6, 95–100.
Kiener, A. (1992) Enzymic oxidation of methyl groups on heteroarenes. A versatile
method for the preparation of heteroaromatic carboxylic acids. Angew. Chemie
 How to get the biocatalyst 241
(Weinheim) , 104, 748–749.
Klyushnichenko, V., Tishkov, V. and Kula, M.R. (1997) Rapid SDS-Gel capillary
electrophoresis for the analysis of recombinant NADP(+)-dependent formate
dehydrogenase during expression in Escherichia coli cells and its purification. J
Biotechnol. , 58, 187–95.
Kuchner, O. and F.H.Arnold. (1997) Directed evolution of enzyme catalysts. TIBTECH ,
15, 523–530.
Kula, M.-R. and Shutte, H. (1987) Purification of proteins and the disruption of microbial
cells. Biotechnol. Prog. , 3, 31–42.
Lee, L.-S. (1989) The primary stages of protein recovery . J Biotechnol. , 11 , 103–118.
Marston, F.A.O. (1986) The purification of eukaryotic polypeptides synthesized in
Escherichia coli. Biochem. J. , 240, 1–12.
Mizumo, S., Yoshikawa, N., Seki, M., Mikawa, T. and Imada, Y. (1988) Microbial
production of cis, cis-muconic acid from benzoic acid. Appl. Microbiol. Biotechnol. ,
28, 20–15.
Moore, J.C., Jin, H.M., Kuchner, O. and Arnold, F.H. (1997) Strategies for the in vitro
evolution of protein function: enzyme evolution by random recombination of improved
sequences. J Mol Biol , 272, 336–47.
Morin, A. and Leblanc, D. (1998) Passive and active screening of D-hydantoinase-
producing microorganisms. In New frontiers in screening for microbial catalysis ,
edited by K.Kieslich, C.J.van der Beek, J.de Bont and W.J.J. van den Tweel, pp. 133–
142. Amsterdam: Elseviers Science B.V.
Morin, A., Hummel, W. and Kula, M.R. (1986) Rapid detection of microbial
hydantoinase on solid medium. Biotechnol. Letters , 8, 571–576.
Nagao, A. and Kito, M. (1990) Lipase-catalyzed synthesis of fatty acid esters useful in
the food industry. Biocatalysis , 3, 295–305.
Nagasawa, T., Shimizu, H. and Yamada, H. (1993) The superiority of the third-
generation catalyst, Rhodococcus rhodochrous J1 nitrile hydratase, for industrial
production of acrylamide. Appl. Microbiology Biotechnol. , 40, 189–195.
Nagasawa, T., Takeuchi, K., Nardi-Dei, V., Mihara, Y. and Yamada, H. (1991) Optimum
culture conditions for the production of cobalt-containing nitrile hydratase by
Rhodococcus rhodochrous 1. Appl. Microbiol. Biotechnol. , 34, 783–788.
Nagasawa, T. and Yamada, H. (1989) Microbial transformation of nitriles. TIBTECH , 7,
153–158.
Naumann, D., Helm, D., Labischinski, H. and Giesbrecht, P. (1991) The characterization
of microorganisms by Fourier-transform infrared spectroscopy (FT-IR) In Modern
techniques for rapid microbiological analysis , edited by W.H.Nelson, pp. 43–96. New
York: VCH Publishers.
Neuberg, C. and Hirsch, J. (1921) Uber ein kohlenstoffkettenknipfendes Ferment
(Carboligase). Biochem. J. , 115, 282–310.
Nilssen, B., Hoimgren, E., Josephson, J., Gatenbeck, S., Philipson, L. and Fuhlen, M.
(1985) Efficient secretion and purification of human insulin-like growth factor I with a
gene fusion vector in staphylococci. Nucleic Adds Res. , 13, 1151–1162.
Nilssen, B. and Abrahmsen, L. (1990) Fusions to staphylococcal protein A. Methods
inEnzymol. , 185, 144–161.
Norris, P.R. and Ingledew, W.J. (1992) In Molecular Biology and Biotechnology
ofExtremophiles , edited by R.A. Herbert and R.J.Sharp, pp. 115–142. Glasgow and
 Applied biocatalysis 242
London: Blackle.
Oyama, K. (1986) Enzymatic synthesis of Aspartame in organic solvents. In Biocatalysis
in organic media , edited by C.Laane, J.Tramper, M.D.Lilly, pp. 209–224.
Tanaka, A., Tosa, T. and Kobayashi, T. (1993) Industrial application of immobilized
biocatalysts. New York: Marcel Dekker, Inc.
Pham, V.T., Phillips, R.S. and Ljungdahl, L.G. (1989) Temperature-dependent
enantiospecificity of secondary alcohol dehydrogenase from Thermoanaerobacter
ethanolicus. J. Am. Chem. Soc. , 111, 1935–1936.
Pindar, D.F and Bucke, C. (1973) The biosynthesis of alginic acid by Azotobacter
vinelandii. Biochem. J. , 152, 617–622.
Pries, F., Van der Ploeg, J.R., Dolfing, J. and Janssen, D.B. (1994) Degradation of
halogenated aliphatic compounds: The role of adaptation. Microbiol. Rev , 15, 279–
295.
Reineke, W. and Knackmuss, H.J. (1980) Construction of haloaromatics utilizing
bacteria. Nature , 277, 385–386.
Ryuno, K., Nagasawa, T. and Yamada, H. (1988) Isolation of advantageous mutants of
Pseudomonas chloraphis B23 for the enzymic production of acrylamide. Agric. Biol.
Chem. , 52, 1813–1816.
Sano, K., Yokozeki, K., Tamura, F., Yasuda, N., Noda, I. and Mitsugi, K. (1977)
Microbial conversion of DL-2-amino 2-thiazoline-4-carboxylic acid to L-cysteine
and L-cystine: screening of microorganisms and identification of products. Appl
Environ Microbiol , 34, 806–810.
Sassenfeld, H.M. and Brewer, S.J. (1984) Apolypeptide fusion designed for the
purification of recombinant proteins. Bio/Technology , 2, 76–81.
Siemann, M., Syldatk, C. and Wagner, F. (1993) Detection and comparison of strains
with selective L-hydantoin cleaving activity using polyclonal antibodies. Biotechnol.
Tech. , 7, 361–366.
Smith, J.C., Derbyshire, R.B., Cook, E., Viney, J., Brewer, S.J., Sassenfeld, H.M. and
Bell, L.D. (1984) Chemical synthesis and cloning of a poly-arg gene fragment
designed to aid polypeptide purification. Gene , 32. 321–327.
Tanaka, Y., Yoshida, H. and Nokayama, K. (1974) Production of L-DOPA by mutants of
Pseudomonas melanogenum. Agric. Biol. Chem. , 38, 633–539.
Theil, F. (1995) Lipase-supported synthesis of biologically-active compounds. Chem.
Reviews. , 95, 2203–2227.
Thiem,J. and Sauerbrei, B. (1991) Chemoenzymic synthesis of sialyloligosaccharides
with immobilized sialidase. Angew. Chem. Int. Ed. Engl. , 30, 1521–1523.
Wandrey, C. and Bossow, B. (1986) Continuous cofactor regeneration—utilization of
polymer bound NAD(H) for the production of optically active acids. Biotechnol
Bioeng. , 3, 8–13.
Wang, D.I.C., Cooney, C.L., Demain, A.L., Dunnill, P., Humphrey, A.E. and Lilly, M.D.
(1979) Fermentation and enzyme technology. John Wiley and Sons.
Wheelwright, S.M. (1989) The design of downstream processes for large scale protein
purification. J. Biotechnology , 2, 89.
Wiegand, T.W., Janssen, R.C. and Eaton, B.E. (1997) Selection of RNA amide synthases.
Chem Biol ., 4, 675–683.
Winson, M.K., Goodacre, R., Woodward, A.M., Timmins, É.M.Jones, A., Alsberg, B.K.
et al. (1997) Diffuse reflectance absorbance spectroscopy taking in chemometrics
 How to get the biocatalyst 243
(DRASTIC) A hyperspectral FT-IR-based approach to rapid screening for metabolite
overproduction. Analytica Chimica Acta , 348, 273–282.
Yamada, H. and Tani, Y. (1983) Microbial production of amides from nitriles, Patent
WO 8301784.
Yamada, H., Ryuno, K., Nagasawa, T., Enomoto, K. and Watanabe, I. (1986) Optimum
culture conditions for production by Pseudomonas chloraphis B23 of nitrile hydratase.
Agric. Biol. Chem. , 50, 2854–2865.
Yamada, H. (1998) Screening of novel enzymes for the production of novel compounds.
In New frontiers in screening for microbial catalysis , edited by K.Kieslich, C.J.van
der Beek, J.de Bont and W.J.J.van den Tweel, pp. 13–17. Amsterdam: Elseviers
Science B.V.
Yamamoto, K., Ueno, Y., Otsubo, K., Kawakami, K. and Komatsu, K. (1990) Production
of S-(+)-ibuprofen from a nitrile compound by Acinetobacter sp. strain AK226. Appl.
Environ. Microbiol. , 56, 3125–3129.
Yokozeki, K., Majima, E., Izawa, K. and Kubota, K. (1987) Screening of microorganisms
producing D-p-hydroxyphenylglycine from DL-5-(p-hydroxyphenyl) hydantoin. Agric.
Biol. Chem. , 51, 963–964.
Zhang, B. and T.R.Cech. (1998) Peptidyl-transferase ribozymes: trans reactions,
structural characterization and ribosomal RNA-like features. Chem Biol. , 5, 539–53.
Zhang, J.H., G.Dawes and W.P.Stemmer. (1997) Directed evolution of a fucosidase from
a galactosidase by DNA shuffling and screening. Proc Natl Acad Sci USA , 94, 4504–
4509.
Zillig, W., Holz, I., Janekovic, D., Klenk, H.-P, Imsel, E. and Trent, J. (1990)
Hyperthermus butylicus, a hyperthermophilic sulfur-reducing archaebacterium that
ferments peptides. J. Bacteriol. , 172, 3959–3965.
 6.
IMMOBILIZED BIOCATALYSTS
SVEN PEDERSEN and MORTEN WÜRTZ CHRISTENSEN
Novo Nordisk A/S, Novo Allé, 8PS, DK-2880, Bagsvaerd, Denmark
Telephone: +45 44422239; Fax: +45 44421237; Email: [email protected]

ABSTRACT
Immobilized biocatalysts are enzymes, cells or organelles (or combinations of
these) which are in a state that permits their reuse (The Working Party of
Immobilized Biocatalysts, 1983). Examples are insoluble enzymes, e.g. used in
a fixed bed reactor or soluble enzymes, e.g. used in a semipermeable membrane
reactor. This chapter will describe methods of industrial interest for making
biocatalysts insoluble.

6.1 WHY IMMOBILIZE?

In general the use of immobilized biocatalysts makes continuous processes possible. This
facilitates process control, which optimizes product yield and quality. Other advantages
of immobilized biocatalysts are: 1) that they do not become mixed with the product,
which makes product recovery easier and 2) that the biocatalyst, in the case of enzymes,
usually becomes more stable.
Immobilized glucose isomerase represents by far the largest application of
immobilized biocatalysts (see section 4.2). There are two main reasons for immobilizing
glucose isomerase: 1) glucose isomerase is an expensive enzyme, because of low
fermentation yield and low catalytic activity. The low fermentation yield is because it is
an intracellular enzyme and the low catalytic activity is caused by glucose not being the
natural substrate for the enzyme (xylose is). Reuse of the enzyme is therefore necessary
in order to make an industrial process economical. 2) fructose and glucose are not stable
at the conditions where industrial isomerization is carried out (60°C, pH 7.5). Typical
degradation products are organic acids and carbonyl compounds. The reaction time
should therefore be minimized and this is done by using a fixed bed reactor with
immobilized enzyme. It will require excessive amounts of liquid enzyme to obtain
reaction times of comparable size.
 Immobilized biocatalysts 245
Examples of immobilization methods for commercially available glucose isomerase
are given in Table 6.1 (Pedersen, 1993). The activity units are those used by the
manufacturers.
Two other immobilized enzymes have reached large scale industrial application:
penicillin amidase and lipase.
Penicillin amidase is used industrially to produce 6-aminopenicillanic acid (6-APA)
from penicillin G or V (see section 4.5). Acid is produced during the process and this will
inactivate the enzyme. One way of overcoming this problem is by using a fixed bed
reactor with immobilized enzyme. The substrate is pumped very rapidly

Table 6.1 Examples of commercial immobilized glucose isomerases.

Manufacturer Trade Enzyme Immobilization Typical

Name Source method values of
initial space
velocities at
60°C (bed
volumes per
hour)
CPC (Enzyme G-zyme G S. Purified GI adsorbed on an 6
Bio-Systems) 994 olivochromogenes anion-exchange resin.
Walon and Stouffs, 1980.
Genencor Spezyme S. rubiginosus Purified GI adsorbed on an 3.9
International (600 anion exchange resin
IGIU/g) consisting of DEAE-
cellulose agglomerated
with polystyrene and TiO2.
Antrim and Auterinen,
1986
Nagase Sweetase S. Binding of heat-treated 1.4
phaechromogenes cells to an anion-exchange
resin, granulated.
Ishimatsu, 1973.
Novo Nordisk Sweetzyme S. murinus Crosslinking of cell 2.1
A/S T (300 material with
IGIU/g) glutaraldehyde, extruded.
UOP Ketomax S. PEI-treated ceramic 5.2
100 olivochromogenes alumina with
glutaraldehyde crosslinked,
purified G.I.Rohrbach,
1981.
 Applied biocatalysis 246

through the enzyme bed, so that only weak acidic conditions are formed. The acid is then
neutralized and the stream recycled to the column for repeated 6-APA production. An
additional benefit of using immobilised enzyme in this process is that no rigorous
purification method, which might cause degradation of the fragile 6-APA molecule is
required.
Lipases are extraordinary enzymes in the sense that the lipase is functioning on an
interface between two phases (oil/water). This interface and the low amount of water are
obtained by using immobilized lipase. The lipase is adsorbed on carriers such as anion
exchange resins or adsorbers and covered with a thin film of water.

Figure 6.1 Crosslinking with glutaraldehyde

6.2 CROSS-LINKING WITH GLUTARALDEHYDE

Both cells and enzymes can be crosslinked with glutaraldehyde. Inert protein or
polyamines may be added to the reaction mixture.
It is probably the -amino groups of the proteins lysine residues, that react with
glutaraldehyde forming a simple Schiffs base as suggested by Branner-Jørgensen (1978).

6.2.1 Cross-linking Whole Cells

 Immobilized biocatalysts 247
In practice the process is carried out like this: (Sweetzyme Q, previously manufactured
by Novo Nordisk A/S).
A cell concentrate (Bacillus coagulans glucose isomerase) is produced by
centrifugation of the culture broth. The cells are then disrupted by pumping the cell
concentrate through a Manton Gaulin homogenizer cell with a single stage homogenizing
valve assembly. The cells are then crosslinked with glutaraldehyde, diluted and
flocculated with a cationic flocculant to give a clear water phase. The mixture is filtered,
and the moist crosslinked aggregate is extruded by means of an axial extruder. Finally,
the particles are dried in a fluid bed dryer and sieved. The particle fraction 300–1000 m
is used for fixed bed operation.

6.2.2 Crosslinking of Enzymes

Crosslinking of enzyme crystals with e.g. glutaraldehyde results in the so-called CLECs
produced by Altus Biologics inc.
CLECs are insoluble and have a high operational stability in aqueous systems as well
as in organic solvents (Wang et al., 1997).
A high enantioselectivity has been reported with lipase-CLECs (Persichetti et al.,
1996).

6.3 ADSORPTION ON CARRIERS

Enzymes can be adsorbed on various types of materials e.g. silica gel, metal oxides, glass
and organic polymers. Depending on the nature of the carrier material, adsorption can be
accomplished by hydrogen bonding; hydrophobic interaction and ionic forces.
Several factors are important for the choice of a carrier:

• Mechanical strength
• High enzyme capacity and retained activity
• Chemical inertness
• Physical stability
• Low cost, possibility for regeneration
• Food grade (for the use in food applications)

The adsorption should be strong enough to insure no enzyme leakage under process
operation. In the characterization of a carrier material, adsorption isotherms of enzyme
adsorbed to the carrier can illustrate the enzymes affinity towards the carrier material and
the maximum amount of enzyme, which can be adsorbed on the material (Rexová-
Benková, 1992 and Gitlesen, Bauer & Adlercreutz, 1997).
Using the assumption of monolayer adsorption on the carrier, Langmuir type of
adsorption isotherms can be constructed. Curves can be fitted to equation (Eq. (6.1)),
 Applied biocatalysis 248
where [A] is the amount of adsorbed enzyme, K L is the concentration of free enzyme at
half-saturation of carrier and [A] Max is the asymptotic maximum amount of enzyme
adsorbed.

Eq. (6.1)

Applications that proceed in aqueous media are even more depending on a very strong
binding of enzyme to the carrier compared to applications in organic media. This is due
to the difference in protein solubility in the two media. In the case of aqueous media, the
adsorbed enzyme can be cross-linked with an bifunctional reagent (typically
glutaraldehyde) to stabilize and minimize any enzyme leakage. However, using cross-
linking eliminates the possibility for re-using the carrier and is therefore only used for
applications, where cost of enzyme and carrier is low.
The carrier materials used for enzyme adsorption must have a highly porous character
and a pore size distribution, which should facilitate a free diffusion of enzyme into the
carrier. Furthermore, substrate and product should also be able to diffuse freely. This is
especially important in the case of large protein substrates, where diffusion into the pore
can be a problem (e.g. whey protein and casein).
Even though, the immobilization procedure should be engineered to maximum retained
activity of immobilized enzyme, it is difficult to measure the amount of active enzyme on
the carrier without an active site titration. However, this has been done in the case of
immobilized trypsin, although only covalent immobilized. (Daly and Shih, 1982)
Techniques have been developed to follow the enzyme adsorption. An immmunogold
staining technique for localisation of lipase on ion-exchange resins has been successfully
used for this purpose (Ison et al., 1990). A more direct method, using 14C-labelled lipase
and micro audioradiography, was developed by Larsen et al. (1995). Recently, an elegant
method, using confocal microscopy and flourescein isothiocyanate labelled lipoxygenase,
visualized the enzyme distribution in the porous support (polyacrylamide beads) without
any cutting of the beads (Pinto and Macáis, 1995). This technique has also been used for
the distribution analysis of prestained yeast cells within polyamide microcapsule (Green
et al., 1995).
The choice of particle size depends more on the type of reactor used in the process. In
batch-operation, the particle size distribution can be broader than in the case of packed
bed reactors.

6.3.1 Hydrophobic Support Materials

A wide range of different supports for hydrophobic adsorption of enzymes is
commercially available. Hydrophobic polymers are the most frequently used and they are
typically based on acrylic, divinylbenzene-styrene copolymers or propylene polymers,
 Immobilized biocatalysts 249
but also hydrophobic CPG (controlled pore glass), made by silanization, has been used
(Bosley and Clayton, 1993).
Below, the construction of an adsorption isotherm is described.
The Accurel EP100 carrier, macro porous polypropylene granulates, was obtained
from Akzo Nobel. Particle size was in the range of 200–1000 m. 1 gram of carrier was
washed with 96% ethanol and water prior to the immobilization. Lipase solution (purified
and concentrated solution of Humicola lanuginosa lipase (103 KLU/ml), pH 7.5) was
added to the wet carrier. The lipase loadings were in the range of 125–500 KLU/g carrier.
The amount of adsorbed lipase activity was calculated from the difference in lipase
activity in the supernatant before subtracted the lipase activity after adsorption.
The lipase activity unit (LU) is described in publication AF 95.1/2GB obtainable from
Novo Nordisk A/S. The experimental data was fit to Eq. 6.1 with the non-linear
regression program GraphPad Prism.
The [A] Max and K L were estimated to 546 KLU/g and 9.5 LU/ml, respectively. KLU
is an abbreviation for 1000LU.

6.3.2 Adsorption on Silica

To illustrate adsorption on silica immobilization of lipases on macroporous silica will be
described.
The diameter of lipase globules is generally around 50Å. Pedersen and Hansen (1994)
discovered that lipases adsorb well to silica, if the adsorption is carried out

Figure 6.2 Adsorption isotherm of Humicola lanuginosa lipase.

 Applied biocatalysis 250
around pH 4.5, and the expressed lipase activity is high provided that the pore diameter
of the macroporous silica material is from 12 to 45 times the diameter of the lipase
molecule.
In practice the immobilization is carried out like this:
The silica carrier is a product from Grace described in Biocatalyst Supports SG BC
1E/June 87.
1 g of carrier (Grace 6, pore diameter 1500A, particle size 0.5–1 mm) is washed in
0.2M acetate buffer of pH 4.5 for half an hour and filtered. 186.000 LU of Humicola
lanuginosa lipase is dissolved in 5 ml of deionized water and added to the carrier. pH is
adjusted to 4.5 and the carrier and the lipase solution is slowly agitated by rotation for
two hours followed by vacuum filtration. The filtrate is analyzed for hydrolytic activity
(LU/ml) in order to determine the amount of adsorbed (loaded) lipase (98%). The
immobilised lipase is air dried, the moisture content is adjusted to 10% by weight, and
the sample is analyzed for BIU (68 BIU).
The lipase activity unit (BIU, batch interesterification units) is described in publication
AF 206–2 obtainable from Novo Nordisk A/S.
A system consisting of a precolumn containing ion exchange resin saturated with water
and an enzyme column (containing 4.5 g immobilized lipase prepared as described
above) in series was set up. The function of the precolumn is saturation of the substrate
with water. An oil mixture consisting of 28.6% (w/w) lauric acid and 71.4% (w/w) soy
bean oil was pumped through the columns. The temperature in the columns was kept at
60°C. The flow rate was adjusted in order to keep a constant conversion of 14%
incorporated lauric acid in the soy bean oil. Samples were taken 3–5 times per week and
analyzed by removing the free fatty acids and partial glycerides by column
chromatography and methylating the triglycerides followed by capillary gas
chromatography of the methyl esters. The flow rate, which is proportional to the activity
was plotted against time when the conversion was 14±1% incorporated lauric acid. A half
life of 850 hours was obtained.

6.3.3 Ion Exchange Resins

Adsorption of enzymes on ion exchange resins is commonly used in the glucose
isomerase process. An example is Spezyme, the product from Genencor Int., which uses
purified glucose isomerase from Streptomyces rubiginosus adsorbed onto an anion-
exchange resin consisting of DEAE-cellulose agglomerated with polystyrene and TiO2
(Antrim and Auterinen, 1986).
Ion exchange resins are normally reused. The spent enzyme on the carrier particles is
washed away with a salt or sodium hydroxide solution. After washing with water, fresh
enzyme can be adsorbed onto the particles. Regeneration can be carried out in the column
or in a separate tank.
Immobilized enzymes based on the adsorption may obtain high activities. Another
interesting aspect of the technology is that the so-called on column loading (OCL)
process is possible. At the beginning the ion exchange resin is only partly loaded with
 Immobilized biocatalysts 251
enzyme. During fixed bed operation fresh enzyme can be added to the incoming substrate
solution and will adsorb to the carrier during passage through the column. The advantage
is that the activity of the immobilized enzyme in the column can be kept constant until
the capacity of the carrier is used (Antrim, Leoyd and Auterinen, 1989).

6.4 COVALENT ATTACHMENT

The covalent linking of enzymes to supports is mainly of use in cases where stabile and
strong enzyme attachment is important. However, coupling to amino acids residues
responsible for the catalytic activity should be avoided to preserve the active site. The
carrier can be activated using a number of chemical approaches depending on the nature
of the carrier e.g. Sepharose® or CPG and the enzyme of interest. This field has been
investigated thoroughly, for reviews see references.
Common for the entire activated carriers is the use of expensive activation chemistry
that, in turn, restricts the industrial use into areas with high margin e.g. pharmaceutical
applications.
The number of commercially available activated carriers for covalent attachment is
therefore small in comparison to commercially available enzyme adsorbent materials.
Eupergit C®, Röhm GmbH, is an epoxy activated co-polymer of methacrylamide,
N,N′-methylene-bis(methacrylate) and glycidyl methacrylate. This support is commonly
used when the coupling chemistry should be epoxy based. Deloxan®, from Degussa, is
polysiloxane-based material activated with amino-groups. Cyanogen bromide activated
materials have been widely used. In this case, CNBr-activated

Figure 6.3 Coupling to CNBr activated polysaccharide.

Figure 6.4 Coupling chemistry of Eupergit®C.

 Applied biocatalysis 252

Figure 6.5 Principle of enzyme coupling to Deloxan®DAP with

glutaraldehyde.

cross-linked agarose is commercially available (Sephacryl ®CNBr, from Amersham

Pharmacia Biotech).
Figures 6.3–6.5 illustrate the coupling chemistry of the above mentioned activated
carriers. For additional coupling chemistry, see Hermanson et al., 1992.
The obtained Shiff base is in many cases not stable enough and can therefore be
reduced (with cyanoborohydride) to make a more stable bond. The coupling of enzyme to
Deloxan® can also be performed with carbodiimides.
The amino acids used for the covalent coupling are types with functional side chains
e.g. Cys, Lys, Tyr and Asp, comprising thiol, amino, phenolic and carboxylic groups. The
reactivity of these residues will be determined by the chosen coupling chemistry and its
conditions, where Lys and the N-terminal amino group are the most frequently used
amino acids for covalent attachment.
In some cases where enzymes have a poor degree of surface amino groups, the
covalent linkage can be established through the glycosidic chains by periodate oxidation
of enzyme with subsequent coupling to activated nylon (López, Braun & Klein, 1996).
Immobilisation buffers have to be optimised with respect to pH, salt concentration etc.
for each couplings chemistry (Hermansson et al., 1992). Furthermore, the degree of
attachment can be important for the stability and retained activity of the immobilised
enzymes. Blanco, Calvete and Guisan (1989) have investigated the effect of multipoint
attachment vs. singlepoint attachment of immobilised chymotrypsin. A high degree of
attachment resulted in a more rigid immobilised enzyme comprising higher activity and
stability. Further studies on the stabilization by covalent multipoint attachment have been
 Immobilized biocatalysts 253
reported by Rosell, Fernandez-Lafuente & Guisan, 1995 and Fernandez-Lafuente et al.,
1995. See also Chapter 8.
Numerous applications with covalently bound proteases or galactosidases, for the use
within the food area, have been reported (Pitcher, 1980). Only immobilized -
galactosidase, by covalent attachment to porous silica (Corning method), for the
hydrolysis of acid whey has been used in semi-industrial scale (Gekas and López-Leiva,
1985). The use of immobilized trypsin in the production of predigest protein for infant
formula is a potent application (Ge and Zhang, 1993). In many cases, the conversion of
protein substrates can be inhibited due to diffusions problems and fouling of the support
material.

6.5 GRANULATION

In section 6.3.1 it was described, that macroporous silica particles are very effective
carriers for lipase immobilization, provided that the pore diameter is larger than about
500Å. For industrial use the particle size cannot be less than 200–300 m.
Commercial materials with these specifications are relatively expensive and because
cheap particulate silica materials with the right pore diameter are available a granulation
process has been developed by Pedersen, Larsen, and Aasmul (1995).
The granulated particles exhibit technical properties equal to or almost equal to the
materials described in section 6.3.1.
In practice the granulation process is carried out like this:
250 g of Celkate T-21 (a synthetic magnesium silicate from Manville) was washed
with 3 volumes of 0.1M acetate buffer, pH 4.5 for 30 minutes followed by filtration.
Humicola lanuginosa lipase concentrate in an amount corresponding to 500.000 LU/ g of
Celkate T-21 was added together with 3 volumes of 0.1 M acetate buffer, pH 4.5, and
stirred for two hours at room temperature. After vacuum filtration the immobilized lipase
was dried for 24 hours at room temperature, the moisture
 Applied biocatalysis 254

Figure 6.6 Granulation of silica.

content adjusted to 10% and analyzed to 14.3 BAUN/g (BAUN, Batch Acidolysis Units
Novo, measures the initial rate of incorporation of decanoic acid into high oleate
sunflower oil (10% water, 70°C)). The filtrate contained 27565 kLU, corresponding to an
adsorption of 78%.
65 g of the thus dried immobilized lipase on Celkate T-21 powder was introduced into
a high speed mixer with an impeller, which can be operated with a speed of 900 rpm. 55g
of a 5% (w/w) gelatin solution was atomized onto the powder with running impeller.
Hereafter 0.1g of Aerosil 200 silicium dioxide (Degussa) was added. The formed
granulate was dried at room temperature and sieved (300–700 m). The moisture content
was adjusted to 10% and analyzed to 5.9 BAUN/g.

6.6 GEL ENTRAPMENT

By this technique a 3-dimensional network is formed around the biocatalyst. Depending

on the method the biocatalyst will in some cases be chemically bound to the network.
The method is widely used for immobilization of viable cells, where the open gel
makes transport of nutrients/metabolites to and from the cell possible, but it has not
gained any industrial importance in immobilizing enzymes. The reasons for this are the
 Immobilized biocatalysts 255
toxicity of the monomers, the difficulties in scale-up and the high pressure drops that will
develop during fixed bed operation due to the elasticity of the gels.

6.6.1 Water Soluble Monomers

Polyacrylamide

Wet cells (15g) in saline (80 ml) are mixed with an equal volume of a 10% w/v monomer
solution (85% acrylamide/15% NN′ methylene bis-acrylamide). Initiator is added (1 ml
5% NNN′N′-tetramethyldiethylamine and 4 ml 5% ammonium persulphate). Gelation
occurs within 20 minutes at 5°C.

Alginate

Wet cells (5g) are suspended in 15 ml of a 1% sodium alginate solution. The mixture is
then forced through a syringe needle into a 0.05 M CaCl2-solution containing the
necessary nutrients for the cell. Gelation occurs almost immediately.

Immobilization in foams/sponge

Several papers have described the immobilization of enzymes by entrapment in foams.

Especially polyurethane foam (PUF) has been widely used (Hu, Korus & Stormo, 1993;
Ferreriradias and Dafonseca, 1995; LeJeune et al., 1997). Hu et al. found that the
immobilization of -galactosidase using PUF was predominantly accomplished by
covalent coupling of the enzymes primary amino groups and isothiocyanate groups of the
prepolymer. The further entrapment in the micropores will facilitate the immobilization
of enzymes. The Hypol 3000 prepolymer (Hampshire Chemical Company) is a very
commonly used polyurethane prepolymer and together with surfactans (e.g. Pluronic
from BASF) as foam booster, very porous materials can be obtained (LeJeune et al.,
1997).

6.6.2 Sol-gel Matrices

Immobilization of lipases in silica gels by gel entrapment has recently received interest
(Sata et al., 1994; Reetz, 1997; Reetz et al, 1996). This so-called sol-gel process involves
hydrolysis of Si(OR)4 in the presence of enzyme. The formed Si-monomers crosslink in
the presence of an acid or base catalyst into an amorphous SiO2-network entrapping the
enzyme. The use of alkylsilanes RSi(OCH3)has been reported to give lipase preparations
with unusually high catalytic activities (Reetz, 1997).
 Applied biocatalysis 256

6.6.3 Biocatalytic Plastics

The problem of high pressure drops with gel entrapped materials has been overcome by
entrapping the enzymes in plastic materials such as polystyrene and
polymethylmethacrylate (Wang et al., 1997). The method involves chemical
acryloylation of the enzyme to provide a polymerizable functionality, formation of non-
covalent ion pairs between the enzymes and a surfactant, solution of these ion pairs in an
organic solvent followed by addition of vinyl monomers, a crosslinker, and an initiator to
give the desired vinyl polymer with the entrapped enzymes.

6.7 MEMBRANES

In the past many attempts have been done to implement membranes into immobilized
enzyme reactors, especially within immobilized enzymes for food processing (Pitcher,
1980 and Bakken, Hill & Amundson, 1992). None of the attempts (e.g. immobilized -
galactosidase to produce low-lactose milk) have materialized into industrial scale yet.
Immobilization of lipases on membranes have also been described and several
bioreactors were developed (see review, Balcao, Paiva & Malcata, 1996). The
immobilization can be done by simple physical adsorption of the lipase on hydrophobic
hollow fibers or flat sheets where polypropylene types are the preferred (e.g. Accurel or
Celgard) (Bouwer, Cuperrus & Derksen, 1997).
Covalent immobilization of lipase on nylon fibers has been done, using the enzymes
carbohydrate groups as chemical link. Oxidation of the lipases carbohydrates with
periodate provides aldehyde groups for the binding to hydrazide activated nylon (López,
Braun & Klein, 1996).
So far the main application for membrane-immobilized lipases is the hydrolysis of
triglycerides, but also synthesis of simple carboxylic esters, peroxycarboxylic acids and
diltiazem intermediates have been described (Malcata and Hill Jr., 1995; Balcao, Paiva &
Malcata, 1996; Bouwer, Cuperrus & Derksen, 1997; Matsumae et al., 1994 and López,
Braun & Klein, 1996). In case of diltiazem, a potent Ca channel blocker, the synthesis of
an intermediate for this pharmaceutical is currently being produced by Tanabe Seiyaku
Co Ltd. on large scale by a Serratia marcescens lipase immobilized on a hollow-fiber
ultrafiltration membrane (Matsumae et al., 1994; Tosa and Shibatani, 1995). The lipase-
catalyzed hydrolysis of the diltiazam intermediate (I) was first developed by DSM-
Andeno using Mucor miehei lipase (Sheldon, 1996).

6.8 IMMOBILIZED VIABLE CELLS

 Immobilized biocatalysts 257
The most important field of application of immobilized cells is the use in environmental
technology (Buchholz and Kasche (1997)).
In other applications there is now an emerging use of immobilized viable cells in the
wine and beer industry.

Figure 6.7 Lipase catalyzed hydrolysis of diltiazam intermediate.

Wine

The secondary fermentation of Champagne production is now carried out by immobilzed

cell technology. Moet & Chandon utilize a plug of immobilized yeast cells for bottle
carbonation (Mensour et al., 1997) .

Beer

The use of immobilized cell technology in beer production has recently been reviewed
(Mensour et al., 1997). Industrial processes for maturation of beer and production of
alcohol free beer are now in operation.
Cultor Ltd. (Finland) and Tuchenhagen(Germany) have developed a process, where
yeast cells are adsorbed on the surface of the carrier developed for glucose isomerase
(Spezyme, Table 6.1). The high volumetric productivity of the immobilized yeast cells
make a conversion of -acetolactate to acetoin possible with only a few hours residence
time in the packed bed columns.
The Belgian company Alfa Laval in association with Schott Engineering (Germany)
has developed processes similar to Cultor using the porous glass bead carrier Siran®
(made by Schott Engineering).

Biotransformations in organic media

There is a very high amount of literature published on this topic, but industrial
applications are scarce. Biorefining in petrochemistry (Setti, Lanzarini & Pifferi, 1997)
appears to be an interesting new development, but still with a lot of practical problems to
be solved. The immobilization method in biotransformations in organic media is mostly
 Applied biocatalysis 258
entrapment of whole cells.

6.9 CONCLUDING REMARKS

When immobilized glucose isomerase was introduced in the early seventies, it was
believed, that other industrial applications of immobilized enzymes would soon be found,
but this turned out not to be true. The main limitation to the use of immobilized enzymes
is that the substrate has to be soluble and highly purified in order to avoid clogging of the
enzyme bed. In the case of glucose isomerase the cost of purifying the syrup (carbon
treatment, ion exchange) before the fixed bed isomerase reactors is substantially higher
than the enzyme costs.
Only in recent years with the development of lipases new industrial applications of
immobilised enzymes have been found. The main commercially interesting processes are
modifications of fats and oils (Haumann, 1997) and simple ester synthesis. Furthermore,
the use of immobilized lipases in the pharmaceutical and agro-chemical industry has
increased since some lipases are able to catalyze a number of enantioselective reactions
(Patel, Nagarajan & Kilara, 1996; Morgan et al., 1997 and Balkenhohl et al., 1997).
Large-scale lipase-catalyzed conversions in organic media has been reviewed by
Sheldon, 1996. A number of very different immobilisation methods have been developed
over the years. For large scale industrial applications, the price parameter is very
important, and glutaraldehyde crosslinking or adsorption are the methods of choice.

6.10 QUESTIONS

6.1. Compare immobilisation by crosslinking and covalent attachment. What are the
strong and weak points of each method?

6.2. PEI (polyethylene imine) is often used as an additive when crosslinking enzymes
with glutaraldehyde. Explain why.

6.3. Explain why immobilised enzymes are inefficient catalysts, when the substrate is
macromolecular. Which type of immobilisation would be best to apply?

6.4. How do you measure the surface area of carrier particles

6.11 HINTS AND ANSWERS

6.1. Compare the two immobilisation methods with respect to the parameters listed in
section 6.3.
 Immobilized biocatalysts 259

6.2. Write the chemical formula for PEI. How many primary amine groups does it
contain?

6.3. Hints to be found in sections 6.3 and 6.4

6.4. N -adsorption and Hg-porosimetry. An example can be found in Barros et al.

(1998). 2

6.12 REFERENCES AND SUGGESTED FURTHER READING

The following books are suggested for further reading about immobilized biocatalysts.
Rosevear, A., Kennedy, J.F. and Cabral, J.M.S. (1987) Immobilized Enzymes and cells .
Bristol, England: IOP Publishing Ltd.
Woodward, J. (1985) Immobilised cells and enzymes—a practical approach . Oxford:
IRL Press.
Koskinen, A.M.P. and Klibanov, A.M. (1996) Enzymatic reactions in organic media .
London: Chapmann & Hall.
Gemeiner, P. (1992) Enzyme Engineering—Immobilized Biosystems . London: Ellis
Horwood Ltd.
Story, B.K. and Schafhauser-Smith, D.Y. (1994) Immobilization of polysaccharide-
degrading enzymes . Biotechnology & Genetic Engineering Reviews, Vol 12., edited
by M.P.Tombs. Andover: Intercept Ltd.

FURTHER REFERENCES

Antrim, R.L. and Auterinen, A.-L. (1986) Die Stärke , 38, 132.
Antrim, R.L., Leoyd, N.E. and Auterinen, A.L. (1989) New isomerisation technology for
high fructose syrup production. Die Stärke , 41, 155.
Bakken, A.P., Hill, C.G. and Amundson, C.H. (1992) Hydrolysis of Lactose in Skim
Milk by Immobilized Beta Galactosidase Bacillus-Circulans. Biotechnology and
Bioengineering , 39, 408–417.
Balcao, V.M., Paiva, A.L. and Malcata, F.X. (1996) Bioreactors with immobilized
lipases: State of the art. Enzyme and Microbial Technology , 18, 392–416.
Balkenhohl, F., Ditrich, K., Hauer, B. and Ladner, W. (1997) Optically-active amines via
lipase-catalyzed methoxyacetylation. Journal für Praktische Chemie-Chemiker-
Zeitung , 339, 381–384.
 Applied biocatalysis 260

Barros, R.J., Wehtje, E., Garcia, F.A.P. and Adlercreutz, P (1998) Physical
characterization of porous materials and correlation with the activity of immobilized
enzyme in organic medium . Biocatalysis and Biotransformation , 16, 67–85.
Bosley, J.A. and Clayton, J.C. (1993) Blueprint for a lipase support: Use of hydrophobic
controlled-pore glasses as model systems. Biotechnology and Bioengineering , 43,
934–938.
Bouwer, S.T., Cuperrus, F.P. and Derksen, J.T.P. (1997) The performance of enzyme-
membrane reactors with immobilized lipase. Enzyme and Microbial Technology , 21,
291–296.
Branner-Jørgensen, S. (1978) On the mechanism of protein crosslinking with
glutaraldehyde. 4 th Enzyme Engineering Conference , M4.
Buchholz, K. and Kasche, V. (1997) Biokatalysatoren und Enzymtechnologie . Weinhein,
Germany: VCH.
Daly, W.H. and Shih, F. (1982) Covalent binding of trypsin to hydrogels. American
Chemical Society Symposiums Series , 186, 133–148.
Fernandez-Lafuente, R., Rosell, C.M., Rodriguez, V. and Guisan, J.M. (1995) Strategies
for enzyme stabilisation by intramolecular crosslinking with bifunctional reagents.
Enzyme and Microbial Technology , 17, 517–523.
Ferreiradias, S. and Dafonseca, M.M.R. (1995) The effect of substrate hydrophobicity on
the kinetic-behavior of immobilized Candida rugosa lipase. Biocatalysis and
Biotransformation , 13, 99–110.
Ge, S.-J. and Zhang, L.-X. (1993) Predigestion of soybean proteins with immobilized
trypsin for infant formula. Applied Biochemistry and Biotechnology , 43, 199–209.
Gekas, V. and López-Leiva, M. (1985) Hydrolysis of Lactose: A Literature Review.
Process Biochemistry , 20, 2–12.
Green, K.D., Gill, I.S., Khan, J.A. and Vulfson, E.N. (1995) Microencapsulation of yeast
cells and their use as a biocatalyst in organic solvents. Biotechnology and
Bioengineering , 49, 535–543.
Gitlesen, T., Bauer, M. and Adlercreutz, P. (1997) Adsorption of lipase on polypropylene
powder. Biochemica et Biophysica Acta-Lipids and Lipid Metabolisme , 1345, 188–
196.
Haumann, B.F. (1997) Structured lipids. INFORM , 8, 1004–1016.
Hermansson, G.T., Mallia, A.K. and Smith, P.K. (1992) Immobilized Affinity Ligand
Techniques . California, US: Academic Press Inc.
Hu, Z.C., Korus, R.A. and Stormo, K.E. (1993) Characterization of immobilised enzymes
in polyurethane foams in a dynamic bed reactor. App. Micro. Biotechnol. , 39, 289–
295.
Ishimatsu, Y. (1973) US Patent 3915797.
Ison, A.P., Dunnill, P., Lilly, M.D., Macrae, A.R. and Smith, C.G. (1990) Enzymatic
interesterification of fats: Immobilization and immunogold localization of lipase on
ion-exchange resins. Biocatalysis , 3, 329–342.
Larsen, A.M., Barfoed, M. and Ursin, C. (1995) Localisation and concentration profiling
of immobilized 14C-labelled lipase on ion-exchange resin by micro autoradiography.
European Conference of Biotechnology , MAP-120.
LeJeune, K.E., Mesiano, A.J., Bower, S.B., Grismley, J.K., Wild, J.R. and Russell, A.J.
(1997) Dramatically stabilized phosphotriesterase-polymers for nerve agent
degradation. Biotechnology and Bioengineering , 54(2), 105–114.
 Immobilized biocatalysts 261

López, J.L., Braun, B. and Klein, E. (1996) Immobilization of Candida rugosa lipase to
nylon fibers using its carbohydrate groups as the chemical link. Biotechnology and
Bioengineering , 51, 327–341.
Malcata, F.X. and Hill Jr., C.G. (1995) Industrial utilization of a hollow-fiber membrane
reactor for the controlled lipolysis of butterfat. Enzyme Engineering XII , edited by M.-
D.Legoy and D.N.Thomas. Annals of the New York Academy of Science, Vol. 750,
401–407.
Matsumae, H., Furui, M., Shibatani, T. and Tosa. T. (1994) Production of optically-active
3-phenylglycidyl acid ester by the lipase from Serratia marcescens on a hollow-fiber
membrane reactor. Journal of Fermentation and Bioengineering , 78(1), 59–64.
Mensour, N.A., Margaritis, A., Briens, C.L., Pilkington, H. and Russell, I. (1997) New
developments in the brewing industry using immobilised yeast cell bioreactor systems.
J. Inst. Brew. , 103, 363–370.
Morgan, B., Stockwell, B.R., Dodds, D.R. andrews, D.R., Sudhakar, A.R., Nielsen, C.M.,
Mergelsberg, I. and Zumbach, A. (1997) Chemoenzymatic approaches to sch-56592, a
new azole antifungal. JAOCS, 74. 1361–1370.
Patel, M.T., Nagarajan, R. and Kilara, A. (1996) Lipase-catalyzed biochemical reactions
in novel media—a review. Chemical Engineering Communications , 153, 365–404.
Pedersen, S. (1993) Industrial aspects of immobilized glucose isomerase. Industrial
Application of immobilized biocatalysts , edited by A.Tanaka, T.Tosa and
T.Kobayashi. New York: Marcel Dekker, Inc.
Pedersen, S. and Hansen, T.T. (1994) Immobilization of thermostable microbial lipase by
adsorption to macroporous inorganic carrier particles. US Patent 5342768 .
Pedersen, S., Larsen, A.M. and Aasmul, P. (1995) Method for production of an
immobilized enzyme preparation and use of the immobilized enzyme Preparation.
Patent Application PCT/DK/00076.
Persichetti, R-A., Lalonde, J.J., Govardhan, C.P., Khalaf, N.K. and Margolin, A.L. (1996)
Candida-rugosa lipase—enantioselectivity enhancements in organic solvents.
Tetrahedron Letters , 37 (36), 6507–6510.
Pinto, M.C. and Maciás, P. (1995) Determination of intraparticle immobilized enzyme
distribution in porous support by confocal scanning microscopy. Biotechnology
Techniques , 9(7), 481–486.
Pitcher, W.H. (1980) Immobilized Enzymes for Food Processing . Boca Raton, FL: CRC
Press.
Reetz, M.T., Zonta, A., SimpelkampJ., Rufinska, A. and Tesche, B. (1996)
Characterization of hydrophobic sol-gel materials containing entrapped lipases.
Journal of Sol-Gel Science and Technology , 7, 35–43.
Reetz, M.T. (1997) Entrapment of biocatalysts in hydrophobic sol-gel materials for use in
organic chemistry. Adv. Mater. , 9(12), 943–954.
Rexová-Benková, L. (1992) Affinity chromatography. Enzyme Engineering—
immobilized biosystems , edited by P.Gemeiner. Chicester: Ellis Horwood Ltd.
Rohrbach, R.P. (1981) US Patent 4268419.
Rosell, C., Fernandez-Lafuente, R. and Guisan, J.M. (1995) Modification of enzyme
properties by the use of inhibitors during their stabilisation by multipoint covalent
attachment. Biocatalysis and Biotransformation , 12, 67–76.
Sata, S., Murakata, T., Ochifuji, M. and Suzuki, T. (1994) Development of immobilized
enzyme entrapped within inorganic matrix and its catalytic activity in organic medium.
 Applied biocatalysis 262
Journal of Chemical Engineering of Japan , 27(6), 732–736.
Setti, L., Lanzarini, G. and Pifferi, P.G. (1997) Whole cell biocatalysis for an oil
desulfurization process. Fuel Processing Technology , 52, 145–153.
Sheldon, R.A. (1996) Large-scale enzymatic conversions in non-aqueous media.
Enzymatic Reactions in Organic Media , A.M.P.Koskinen and A.M.Klibanov. London:
Chapman & Hall.
The Working Party on immobilized Biocatalysts (1983) Enzyme Microb. Technol. , 5,
304.
Tosa, T. and Shibatani, T. (1995) Industrial Applications of Immobilized Biocatalysts in
Japan. Enzyme Engineering XII , edited by M.-D.Legoy and D.N.Thomas. Annals of
the New York Academy of Science, Vol. 750, 364–375.
Walon, R.G.P. and Stouffs, R.H.M. (1980) European patent application 0036662 A2.
Wang, P., Sergeeva, M.V., Lim, L. and Dordick, J.S. (1997) Biocatalytic plastics as
active and stable materials for biotransformations. Nature Biotechnology , 15, 789–
793.
Wang, Y.F., Yakovlevsky, K., Zhang, B.L. and Margolin, A.L. (1997) Cross-linked
crystals of subtilisin—versatile catalyst for organic synthesis. Journal of Organic
Chemistry , 62(11), 3488–3495.
 7.
PROTEIN ENGINEERING: DESIGN AND
ENGINEERING ON THE NANO SCALE
STEFFEN B.PETERSEN
Biostructure and Protein Engineering Group, Biotechnology Laboratory,
Aalborg University, Sohngaardsholmsvej 57, DK 9000 Aalborg, Denmark
Telephone: +45 96 358469; Telefax: +45 98 142555; Email:
[email protected]

ABSTRACT
Protein Engineering has become a tool in academia and industry alike for the
production of proteins with new or optimized properties. In essense protein
engineering uses genetic engineering techniques to alter the gene coding for the
protein. Whereas Nature as a “blind watchmaker” has done protein engineering
throughout the 3.8 billion year history of life, modern protein engineering
attempts to do rational changes to the gene in question. As such protein
engineering is a science in dramatic development. The present chapter addresses
the processes in Nature that facilitated the evolution of life, the importance of
the physical chemical properties of water is stressed. The essential physical
chemical techniques for studying the consequences of protein engineering is
briefly outlined. Finally several protein engineering case stories are presented,
highlighting the potentials and limitations of the technique.

7.1 INTRODUCTION

Protein Engineering is the art of modifying an existing protein or creating, de novo, a

protein of pre-specified properties. Protein Engineering has excited the whole
biotechnological and pharmaceutical community, because of its potential for optimising
the function of industrially important enzymes, or protein-based pharmaceuticals towards
existing or novel applications. Protein Engineering holds the promise of providing
environmentally more friendly alternatives to existing industrial processes that today are
based on non-enzymatic processes. The virtual explosion in the number of protein
sequences that has been deduced from genetic sequences has grown exponentially over
 Applied biocatalysis 264
the last decade, with the consequence, that our knowledge about natures successful
protein engineering projects has improved dramatically as well. If we identify a protein in
e.g. E. coli and find homologues in a range of other organisms we may compare such
sequences and extract important information about which parts of the amino acid
sequence that are essential for its stability and function. Often it is found that a particular
protein occur in both bacteria and “higher” organisms such as Homo sapiens. This is e.g.
true for subtilisin that, as the name indicates, originally was found in Bacillus subtilis—
true homologues are now known to be expressed in man, Homo sapiens. The
philosophical implications of this realisation is simple but intimidating. Man must share a
common ancestor with the lowly micro-organism, Bacillus subtilis. This is most likely
contrary to what is believed by the general public. A comforting fact is that this common
ancestor must have lived about 1 billion years ago. This is surely a long time ago,
considering that Homo sapiens appeared on the evolutionary scene only 1–2 million years
ago. It is relevant to observe that both the Bacillus subtilis and Homo sapiens from an
evolutionary perspective have managed to survive and evolve from this common root in
the shadows of biological evolution. From this vantage point the terminology high and
low organisms appear meaningless—both organisms and their ancestors have proven able
to survive until today. Whether we, the representatives of the species Homo sapiens, will
prove able to sustain our own living conditions in the future appear less than certain to
the author.
Over the last century we have recorded an amazing technological trend towards
miniaturisation of complex gadgets, such like electronics and various sensors. A
computer, that 40 years ago impressed its users through its blinking lights and ability to
compute, today will appear like a hopeless and primitive invention. The size needed for
one bit to be processed has shrunk several orders of magnitude, and at the same time the
speed at which the process can occur has shown a similar impressive trend. It is always
tempting to believe that the frontline of technological achievements represents the
ultimate, but history has shown us that this only seldom is true. Whereas the current
status represents a size-reduction of pre-existing macroscopic entities such as electric
capacitors, coils and connectors onto a semi-conductor based matrix, serious research is
now in progress aiming at producing molecular based electronical components. Purposely
engineered proteins will most definitely find their way into this realm as information
storage and computation units. This is particularly so, since several proteins have the
intrinsic ability to populate more than one discrete conformational or energetic state. E.g.
bacteriorhodopsin can adopt two states in the presence of retinal, one in which the retinal
is an all-trans conformation, and one in which the conformation around one double bond
has changed into a as conformation. Rhodopsin is currently being investigated for its
potential for information storage, since the above mentioned trans-to cis-conversion is
reversible and controllable.
In the present small review we will interest ourselves primarily for the art of
engineering proteins. In this process we will find ourselves recovering important
information from seemingly distant areas of science, such as physical chemistry and
molecular evolution. Since the literature is full of a myriad of physical chemical methods
 Design and engineering on the nano scale 265
we will first dwell on methods used to test the native or engineered protein for stability
and function. In this context it is necessary to define some necessary fundamental aspects
of physics and chemistry. It is the authors hope that the present text will succeed in
providing the reader with the necessary vocabulary as well as a useful overview over this
complex but exciting topic.

7.2 THE 3.8 BILLION YEAR EXPERIMENT

Currently we believe that earth is about 4.5 billion years old, and that all elements stem
from inter-galactic dust and remnants from violent catastrophic events, super novas,
where a star enters the end stage of its life cycle and finally explodes. Atoms heavier than
iron, e.g. chromium and uranium, probably stems directly from such explosions, whereas
most lighter elements, such like carbon, nitrogen and oxygen, are from more “normal”
fusion processes in the sun or other stars. Hydrogen is omnipresent. We may therefore
assume that all atomic elements necessary for life exist in the immediate environment
around stars. As planets condensed from the interstellar dust gravitational energy was
converted to thermal energy. For the first hundred million years Earth’s thermal energy
was very high and more volatile compounds such like CO2 and H2O would move at
speeds, that would allow them to escape the Earth gravitational field. Earth is still too
small and too hot to retain some of the lightest elements such as helium. Gradually the
temperature dropped to levels where compounds such as H2O would largely remain
attached to the surface of Earth. Therefore, presumably all H2O stems from collisions
with comets, that are known to contain large amounts of water, although it is not known
why. Curiously the deuterium/hydrogen ratio of three comets, that has been studied with
spectroscopy so far is considerably higher that what is found on Earth. Comet impacts
were very common in the early stages of Earth geological history, and have gradually
ceased to low levels. Thus H2O, which is often heralded as the molecule without which
life could not exist, actually stem from comet impacts that most certainly have destroyed
most if not all of life that may have existed at the time of impact. In various layers of rock
(strata) we have found remnants of cellular life dating back 3.8 billion years. It is truly
humbling to compare this number with the 2 million years, that Homo sapiens have
walked on the planets surface. In this general context it should also be mentioned that
amble evidence exist that life has been threatened repeatedly and in one case almost
completely destroyed due to collisions with major celestial objects. The latest such event
took place 265 million years ago, and it appear as if more than 95% of the life forms on
Earth vanished shortly thereafter. Some paleo-biologists have concluded that Earth
appear to suffer such impacts with an apparent periodicity of 65 million years. It is
interesting that one of the largest challenges to life on Earth appear to have been
molecular oxygen, which initially was toxic or maybe even deadly to most living
organisms. Living cells at that time had developed and optimised themselves toward a
world largely depleted of oxygen. A plausible model for the survival of cellular life
appear to be the symbiotic invasion of one prokaryote into the intra cellular space of
 Applied biocatalysis 266
another. Over aeons of time the invader transformed into what we today call
mitochondria, the organelle that deals with energy production in the living cell. This
appears to have been a successful solution to a large group of living cells, that we today
call the eukaryotes, the cells that have a defined cell nucleus. Energy production today is
intimately linked with molecular oxygen—as a matter of fact many organisms cannot live
without it. Interestingly the mitochondria (as well as the chloroplasts) have their own
genome, which surprisingly exhibit a somewhat different genetic code.

7.2.1 Water—So Simple and Yet So Complex

During their education students are presented for water as a simple, tri-atomic molecule,
which due to its hydrogen bond capacity is excellent as a solvent. It is off course true that
water only consists of two hydrogen atoms and one oxygen atom, and that it displays an
exquisite ability to serve as an acceptor and donator of hydrogen bonds. But water is in
many aspects highly complex despite its deceptive simplicity. We are tempted to interpret
the chemical world in terms of its covalent composition—but as we will find later, non-
covalent bonds, such as hydrogen bonds may be an essential key to the deeper
understanding of e.g. proteins. When we raise the question: “What 3D structure does the
molecule(s) have?” the question may have no proper answer unless we state during which
time window we are looking, i.e. if we ask for the 3D structure of a protein during 1 pico-
second (10−12 sec) we may find that the core of the protein is fairly well defined, whereas
there appears to be significant motion of the amino acid side chain atoms on the protein
exterior. Conversely if we study water, and ask for the water structure over a period of
seconds, we can interpret the data as monomeric water molecules, that on the average are
involved in hydrogen bond formation to its immediate neighbours. However, if we ask
about the 3D structure of water over a period of 1 pico-second, water is not behaving as a
monomeric entity—it is an any instant linked with a large, local network of other water
molecules. The water molecules that are bound to the protein surface, in many aspects are
integral parts of the protein structure—Nuclear Magnetic Resonance data show that they
display life-times in the bound state in the low milli-second range, thus exceeding the
life-time of a typical substrate molecule in the active site. From the substrate molecule’s
point of view, the water that is bound to the protein surface is part of the protein. Such
life-time considerations may be of importance if water molecules have to be forced out of
the active site cleft in order to facilitate substrate binding. Complete removal of such a
water-based protein solvent shell is exceedingly difficult, and a normal “dry” protein
preparation must be assumed to contain about 10% w/w water.
The interaction of water with electromagnetic radiation (photons) is fascinating. As
seen in Figure 7.1 (adapted from Jackson, 1975) water attenuates electromagnetic
radiation very efficiently—except in a very narrow frequency window, which exactly
coincides with what we call “visible light”. When moving away from the visual range,
the absorption coefficient increases rapidly with over 6 to 8 orders of magnitude, thus
rendering the water non-transparent. The functional properties of the (human) eye is
deeply rooted in the physical properties of water. Photosynthesis is also based on
 Design and engineering on the nano scale 267
absorption of photons in the visible and near UV range. It is very tempting to consider
that early life developed in the earth primordial oceans, and in order to succeed in
harvesting energy from the little light that reached into the ocean, any efficient
mechanism had to operate on visual light. The evolution of the eye (which has repeated
itself at least twice throughout evolution) had to converge on a sensory device optimised
for the “visual” range of exactly the same reasons. Why should nature develop a photon-
sensor (an eye), for photons which were not present in the environment? If we let our
imagination soar and consider the (very faint) possibility of life developing in the
methane and ammonia rich environment on e.g. Jupiter, any light harvesting mechanism
developed will be very different, since the radiation window is bound to be very different.
Similarly Superman’s super-vision indicates that he stems from a quite different world,
albeit not Jupiter.

Figure 7.1 The absorption coefficient of water as a function of the frequency of

the electromagnetic radiation. Adapted from Jackson, 1975. (See
Colour Plate I)

7.2.2 Our Genes, Our Heritage

Life on our beautiful planet has thus repeatedly risen like the bird “Phoenix” from the
ashes and despite the onslaught of cosmic rays and celestial missiles have become what it
is today. Exposed to toxic oxygen its living organisms managed to convert into new
forms of life that could thrive on oxygen. This is a strong manifest of the adaptability of
life. Every time life was threatened with extinction somehow it recovered and a burst of
new life forms emerged. It is evident that Nature has experimented extensively with all
 Applied biocatalysis 268
living organisms, including man. Most of these experimental life forms are extinct today.
Some progeny survived, probably because they through mutational events acquired
unknown but necessary properties that allow them to proliferate. Today Earth is teeming
with life, and it may appear strange to look for similarities between Homo sapiens and
Bacillus subtilis as we commented briefly about in the introduction. However such
comparisons can be extremely useful. Some are highly surprising as the one below (see
Figure 7.2) where we have compared the amino acid sequence from human cytochrome C
with cytochrome C’s from maize. Note that the similarity is extensive—only a few amino
acids have

Figure 7.2 Human cytochrome C compared with cytochrome C from maize.

All identical positions are hatched. 72 positions out of 105 are totally
conserved. (See Colour Plate II)

been substituted even when we compare what we normally would regard as distant
organisms. We are thus dealing with an enzyme, that through several hundreds of million
years have been kept virtually unchanged—even if new organisms have developed from
existing ones.
Who should have believed that a protein without which a human cannot live, is ~70%
identical to the cytochrome C from the “lowly” plant maize? We are truly linked
evolutionarily with all other living organisms whether we like it or not. The superiority of
Homo sapiens is certainly not verified in what we know about our genetic heritage.
However, the extent of identity between homologous proteins can also be small—this is
typically true if the proteins appear late in evolution such as hemoglobin. It is proposed
that such proteins are far from optimised, and are thus displaying a higher rate of amino
acid substitutions. It appear reasonable to assume that cytochrome C has had a similar
molecular evolutionary history with high initial rates of amino acid changes. The close
similarity between the two cytochrome C’s in Figure 7.2 thus hint at Homo sapiens and
maize stemming from an ancestral organism, where the optimisation of cytochrome C
was essentially complete.

7.3 PROTEIN SEQUENCE ANALYSIS

At the onset of a protein engineering project often only the amino acid sequence of the
protein, or fragments thereof, are known. Without access to a three-dimensional structure
one has to explore whatever information is available at the sequence level. It is tempting
 Design and engineering on the nano scale 269
to believe that, knowing the amino acid sequence and the force fields that are appropriate
for describing molecular motion, one can calculate the 3-D structure. However, it is
plainly impossible to do such an exhaustive conformational search. If we assume that a
10° accuracy around the torsion angles is sufficient for our purpose, then we need to
calculate 14400 conformations for each residue. If we have a peptide of length N, we will
thus have to calculate the conformational energies for (36*36)N different conformations.
This task grows exponentially with N and is simply impossible to do, even at relatively
short chain lengths.
Based on the discussion above it is clear that we need a different approach, that
somehow can take advantage of available knowledge. We will here address some
relevant cases.

7.3.1 The Protein Sequence Shares Significant Homology with Other

Proteins
In this case we have information about what most likely is a class of proteins. If the 3-D
structure of one of these proteins is known, we can immediately start predicting the 3-D
structure of the protein of interest. If the homology is more than 40% residue identity,
this approach is fairly accurate in the cases reported so far and is fully acceptable as a
basis for protein engineering activity on the protein in question (see e.g. Pastora and
Lesk, 1991). In practical terms, one assumes that the 3-D structure that is known for the
homologous protein essentially share its folding and other global features with the new
protein. Present day computer graphics and molecular modelling software allows for
rapid generation of models of mutated structures—and a rational scheme of mutations
will therefore lead to a plausible 3-D structure for the new protein. Problems may arise in
modelling loop structures at the surface of the protein, because it is often at such
locations in the sequence that the largest differences are found between two homologous
proteins. However experience in handling this type of modelling is growing rapidly. In
particular, the availability of structural data bases is crucial for the success of this part of
the modelling endeavour. We will revert to this later in this chapter.

7.3.2 Only Part(s) of the Protein Sequence Displays Homology to Other

Protein Sequences
This situation is much more difficult to handle and any structural predictions are, at best,
risky. Often one finds that only a few short stretches of the amino acid sequence can be
matched to other protein sequences. It is, of course, possible to align the remaining parts
of two such sequences using the Needleman-Wunsch algorithm, while still maintaining
the local matches. However, any alignment that is based on a purely mathematical string
description may fail, especially if the sequences are very dissimilar. In addition, one has
to consider the limitations of automatic sequence alignments—most methods have no
way of including structural information. If a residue is positioned at the surface of the
 Applied biocatalysis 270
protein, a hydrophobic residue could be replaced with a hydrophilic, maybe even a
charged, residue without great difficulties, since both types of residues are found on the
surface of proteins. However, the same argument is not true for hydrophobic residues
residing in the hydrophobic core of the protein, where hydrophilic residues are rare and
charged residues are almost totally absent. Standard automatic sequence alignment
assumes that a certain score can be assigned to a given substitution of one amino acid for
another (e.g. LYS for TRP), independent of whether the substitution takes place on the
surface or in the hydrophobic core of the protein. The score matrices that are applied are
built upon a careful analysis of substitution patterns found in families of selected related
proteins. By studying the substitution patterns in protein families, substitution scores
have been derived for all possible pairs of amino acids, based on the average likelihood
that such a substitution takes place. The sequence alignment program applies a selected
score matrix, to evaluate whether it can achieve a better match, by shifting one sequence
with respect to the other and by introducing gaps and deletions in the two sequences.
When a maximum has been found within the boundary conditions set by the program or
the user, the procedure has run to an end. The result is often good, but it may also fail
because of the reasons stated above. In Figure 7.3 we illustrate this feature with a local
alignment of several structures of trypsin but, in addition to the standard alignment, a
structural alignment based on the 3D structures and also the water accessibility for the
residues in this selected segment of the trypsin sequence are shown. The alignments are
very similar but not identical—close to both gap regions the sequences are aligned
somewhat differently, leading to a misalignment of the segment GYHF in the automatic
sequence alignment. Likewise, the fragment CYKSR close to the second gap region
displays a similar problem. The conclusion from such observations is that, short of 3-D
structural

Figure 7.3 Sequence and structural alignments of the N-terminal segments of

two trypsins: 1trm (trypsin) and 2gch (gamma chymotrypsin). The
notation 1trm and 2gch are the one used in the Brookhaven Protein
Data Bank (PDB). aa-1trm-1D and aa-2gch-1D refer to the output
from a sequence alignment including 7 trypsin-like sequences (PDB
codes: 1trm, 2gch, 1ton, 3est, 1hne, 3rp2, 1sgt). aa-1trm-3D and aa-
 Design and engineering on the nano scale 271
2gch-3D are the sequence alignments that result from a 3D-structural
alignment of the two protein structures. The traces ss-1trm-3D and
ss-2gch-3D are the associated secondary structural assignments of the
two fragments. The wa-1trm-3D and wa-2gch-3D are the calculated
side chain solvent accessibility for the fragments.

information, sequence alignment is most often a good tool for making a crude alignment
but, even in high homology cases as illustrated here, errors are likely to occur,
particularly in the loop regions of the protein. This type of error originates in the fact that
the score matrix cannot distinguish between surface residues and buried residues and,
since the number of surface residues is much smaller than the number of buried residues,
the score matrix will have a heavy bias towards substitution patterns that are correct for
the protein interior. We have illustrated graphically the best 3-D match for the two
structures for 1trm and 2gch. The core of the proteins that is widely conserved is shown
in red and, in blue and green, the regions (predominantly loops) where the structures did
not overlap.

7.3.3 Search for Motifs with Known Function

A collection of known patterns of amino acids has been assembled and is available from
the author or from the EMBL, Data Library (Bairoch 1991, EMBL 1991). The data base
is called PROSITE and contains in the excess of 500 different motifs that one can search
for in a sequence of unknown function.

7.3.4 Predict the Secondary Structure

Several methods exist for predicting the secondary structure of proteins. These methods
are often applied in unison with hydrophobicity plots with the purpose of identifying
structural features of importance for the interpretation and subsequent modification of the
protein structure. Unfortunately, none of them are very accurate, but certain important
features can often be deducted from such predictions. In
 Applied biocatalysis 272

Figure 7.4 A subset consisting of a total of 75000 amino acids in available 3D

structures in the Brookhaven Protein Data Base has been screened for
secondary structural preferences. Three categories has been
classified: , and turn. In (A) the total count has been given. Note the
different total abundance of the amino acids. In (B) the percentages
are given (number of a selected amino acid in a given secondary
structural category divided with the total count of that amino acid).
The third column in each entry gives the % abundance of an amino
 Design and engineering on the nano scale 273
acid type. The program PDBASE (Petersen, 1991) was used for the analyses.
(See Colour Plate III)

Figure 7.4 we have illustrated the overall occurrence of the various residues in different
secondary structural classes in approximately 500 structures in the PDB. It is interesting
to observe the relatively uneven preferences for the different amino acids, GLY and PRO
have a clear bias towards the coil structure, whereas ALA prefers the -helix over the -
sheet environment.

Table 7.1 Conformational parameters for -helical, -sheet and -turn residues in 29
proteins. The P alfa , P beta and P turn are the conformational parameters for the
different secondary structures, the fi’s are the frequencies for occurrence of a
given residue in the b-turn. (Adapted from Prevelige and Fasman, 1989).

P alfa P beta P turn f1 f2 f3 f4

Glu 1.51 Val 1.70 Asn 1.56 Asn 0.16 Pro 0.30 Asn 0.19 Trp 0.17
Met 1.45 Ile 1.60 Gly 1.56 Cys 0.15 Ser 0.14 Gly 0.19 Gly 0.15
Ala 1.42 Tyr 1.47 Pro 1.52 Asp 0.15 Lys 0.12 Asp 0.18 Cys 0.13
Leu 1.21 Phe 1.38 Asp 1.46 His 0.14 Asp 0.11 Ser 0.12 Tyr 0.13
Lys 1.16 Trp 1.37 Ser 1.43 Ser 0.12 Thr 0.11 Cys 0.12 Ser 0.11
Phe 1.13 Leu 1.30 Cys 1.19 Pro 0.10 Arg 0.11 Tyr 0.11 Gln 0.10
Gln 1.11 Cys 1.19 Tyr 1.14 Gly 0.10 Gln 0.10 Arg 0.10 Lys 0.10
Trp 1.10 Thr 1.20 Lys 1.01 Thr 0.09 Gly 0.09 His 0.09 Asn 0.09
Ile 1.08 Gln 1.10 Gln 0.98 Tyr 0.08 Asn 0.08 Glu 0.08 Arg 0.09
Val 1.06 Met 1.05 Thr 0.96 Trp 0.08 Met 0.08 Lys 0.07 Asp 0.08
Asp 1.01 Arg 0.93 Trp 0.96 Gln 0.07 Ala 0.08 Thr 0.07 Thr 0.08
His 1.00 Asn 0.89 Arg 0.95 Arg 0.07 Tyr 0.07 Phe 0.07 Leu 0.07
Arg 0.98 His 0.87 His 0.95 Met 0.07 Glu 0.06 Trp 0.06 Pro 0.07
Thr 0.83 Ala 0.83 Glu 0.74 Val 0.06 Cys 0.05 Gln 0.04 Phe 0.07
Ser 0.77 Ser 0.75 Ala 0.66 Leu 0.06 Val 0.05 Leu 0.04 Glu 0.06
Cys 0.70 Gly 0.75 Met 0.60 Ala 0.06 His 0.05 Ala 0.04 Ala 0.06
Tyr 0.69 Lys 0.74 Phe 0.60 Phe 0.06 Phe 0.04 Pro 0.03 Ile 0.06
Asn 0.67 Pro 0.55 Leu 0.59 Glu 0.06 Ile 0.03 Val 0.03 Met 0.06
Pro 0.57 Asp 0.54 Val 0.50 Lys 0.06 Leu 0.03 Met 0.01 His 0.05
 Applied biocatalysis 274

Gly 0.57 Glu 0.37 Ile 0.47 Ile 0.04 Trp 0.01 Ile 0.01 Val 0.05

7.3.5 The Chou-Fasman Method

The most widely used prediction scheme for protein secondary structure has been,
beyond doubt, the Chou-Fasman method (Prevelige and Fasman, 1989; Fasman, 1989;
1985; Chou and Fasman, 1974a; 1974b; 1977; 1978; 1979). The Chou-Fasman method
relies on the calculation of conformational parameters P , P and Pi, which are the ratios
between the actual occurrence of a given residue in a secondary structure ( -helix, -
strand and turn) and the number one should expect if the residue occurred without any
structural preference (see Table 7.1). Based on such parameters Chou and Fasman
formulated rules that resulted in a prediction of secondary structure for a given amino
acid sequence. Chou and Fasman were capable of obtaining 70–80% accuracy on novel
sequences using the predictive scheme (Prevelige and Fasman, 1989). Since the Chou-
Fasman method orginally was not defined in the form of a rigorous computer algorithm,
the applications by other researchers have been somewhat subjective in their nature,
leading to various claims on the accuracy of the method. Several reports seem to agree
that 55–60% is the average performance. The predictive scheme of Chou and Fasman has
now been formulated in a program by the Fasman group (Prevelige and Fasman, 1989),
which should result in more reproducible results.

7.3.6 The GOR Method

Whereas the output space in the Chou-Fasman method is limited to 3 states: , and
coil the GOR method utilizes up to 81 states (Garnier, Osguthorpe and Robson, 1978;
Garnier and Robson, 1989; Gibrat, Garnier and Robson, 1987). The approach uses
information theory as its basis and represents protein structures by residue torsion angels
and . The 81 different states then subdivide the torsion angle space in to small
segments. These segments are not equally probable—certain torsion angles are much
more likely to occur than others—90,7% of all torsion angles in a 68 protein set
containing 11,237 residues were close to -helical and -strand torsion angles,
respectively and 3,3% were found in a region where GLY and ASN are normally found
(Garnier and Robson, 1989). By grouping the 81 different states, a reduced number of
states could be defined. The GOR method predicts correctly 53,2% of a test set into 3
states and 41,1% in 4 states.
The GOR method has been further developed to include also pair correlation’s. In
helices such pair correlation’s are found to reflect our basic perception of the helix
geometry. E.g. LEU and VAL are favorable for helix structures if they are separated by 3
or 4 residues but unfavorable if separated by ±2. Likewise -strands and turns have also
 Design and engineering on the nano scale 275
been studied for pair correlation’s.
When including pair correlation 61,0% of the residues were correctly predicted in a 3
state model and 51,9% in a 4 state model. The database encompassed 74 proteins from
the Brookhaven Protein Data Bank.
It is difficult to compare quantitatively the GOR method with the Chou-Fasman
method since it is implicit that a method operating with a much larger number of output
categories will exhibit a lower success rate. The authors of the present paper rather
recommend use of more than one scheme for predicting protein secondary structure.

7.3.7 Neural Network Based Predictions

Lately a new computer based protein structural prediction method has appeared (Qian
and Sejnowski, 1988; Bohr et al., 1988; Holley and Karplus, 1989; Bohr et al, 1990;
Petersen et al., 1990; McGregor, Flores and Sternberg, 1989). Neural networks can be
trained on pairs of matching questions and answers to a given problem. For example, if
one trains a neural network on matched pairs of protein sequences and secondary
structures, the neural network will learn how to map the sequence onto the secondary
structure. Given that the set of questions/answers is sufficiently large, the neural network
can be expected to answer with reasonable accuracy a question to which one has no
answer at the moment. In order to obtain a high quality answer, the problem (question)
that is posed to the neural network has to be within the general concept area that the
network has been trained on. Thus, after training the network on the known set of
sequences/secondary structure pairs, one can pose a question in the form of a protein
sequence not included in the training set and the neural network will now answer in the
form of a secondary structural prediction for that particular sequence. The accuracy of
neural network predictions are similar to
 Applied biocatalysis 276

Figure 7.5 A neural network trained to predict both secondary structural

assignment as well as distance constraints. The amino acid sequence
for a protein is fed into the input layer of the network. The network
selects a local context around a central residue, here “R”. The trained
network then passes on the information to the next so-called “hidden”
layer, which then passes the once processed input information on to
the final ouput layer, that presents the predictions the network
proposes for “R” in terms of a secondary structural assignment as
well as the distance constraints. Note that the local context is
influencing the networks prediction. (See Colour Plate IV)

or slightly better than the Chou-Fasman method. The training is computational-intensive

but certainly feasible on present day computers. When training is done, a neural network
can easily run on a standard PC.
It has also been demonstrated that the neural network approach can be utilised to
predict the 3-D backbone folding of a protein related to the proteins the neural network
has been trained upon (Bohr et al., 1990). A neural network is trained upon matching
pairs of protein sequences and secondary structure as well as C distance constraints
(i.e. the distance between 2 residue C -atoms larger or smaller than a selected threshold
 Design and engineering on the nano scale 277
distance, in this particular case 8 Å). After training on a set of 13 proteases of 4 different
structural classes, the network was presented with the sequence of a protease not included
in the training set but showing high homology to one of the proteins in the training set.
The network predicted the secondary structure (see Figure 7.5) as well as the C distance
constraints, which were then used to generate a folded structure using novel minimization
techniques especially designed for this type of problem. However, since this is a case
study, 3-D prediction based on neural networks will need further clarification before the
technique is of general applicability.

7.3.8 Identify Residues at the Surface of the Protein in Question

It is known that in 95% of all charged residues are located at the protein surface. The
residual 5% are found inside the protein, but typically paired in a salt-bridge with another
charged residue of opposite charge. Likewise, prolines shows a marked preference for
coil-like structures, which most often are also surface positioned. Figure 7.3 shows how
the different amino acid residues are distributed in the various types of secondary
structures of known 3-D protein structure. Assuming that we do not know the 3-D
structure, we may still have reasons to believe that we can engineer the protein’s
characteristics in a more general sense. We may want to shift the iso-electric point of the
protein: we can do that by introducing charged residues in the sequence at points where
we have reasons to believe that we are modifying a surface location of the protein. In the
simplest case we substitute a LYS or ARG with an ASP or GLU, or vice versa, leading to
a net change in charge of 2 at neutral pH. By substituting a charged residue for another
charged residue we ensure with some certainty that we maintain most of the hydration at
that point of the protein sequence, which may have important implications for protein
stability. There are, of course, pitfalls in this approach: if we assume that all charged
residues are surface positioned we may, be change, substitute one of the residues in a
buried salt bridge in the protein. The protein will probably not survive as a functional
protein—we have introduced a major electrostatic repulsion in the core of the protein
where previously we had a stabilizing electrostatic interaction. Likewise, if a
hydrophobic residue is correctly assigned to being a surface residue—and is located in a
surface loop, it is tempting to assume that one could substitute it with a hydrophilic,
maybe even charged residue in most cases. However, there are examples where residues
residing in loops at the surface have their side chain almost totally buried, by pointing the
side chain into the loop gap instead of exposing its main chain atoms to the solvent. If we
introduce, at such a location, a residue which cannot be buried, without great penalties in
terms of destabilizing the protein, the loop environment may attempt to compensate, by
allowing the hydrophilic side chain to obtain access to the solvent surface and, as a result,
change some of the other torsion angles in the loop. If this compensation is not possible
we should anticipate that the protein will display marked conformational changes or will
be dramatically destabilised. A detailed analyses of a very large set of proteins for which
either the 3D structure was determined experimentally, or the 3D structure could be
assumed to be identical due to high levels of sequence identity to a protein for which the
 Applied biocatalysis 278
3D structure was known has been performed in my laboratory (Petersen et al., 1998). In
this study we differentiated between different discrete levels of solvent exposure, i.e. a
residue can have 10, 20, 30 or 100% of its side chain solvent exposed. The different
amino acids distributed themselves into 3 fairly distinct classes of highly polar or charged
residues (H,R,K,Q,E,N,D), weakly polar (G,A,P,S,T) and hydrophobic residues
(V,L,I,M,F,W,Y,C). Histidine resembled the weakly polar class more than the highly
polar or charged, probably reflecting the fact that most protein structures have been
determined at or close to neutral pH, where histidine can be expected to be electrically
neutral. It is interesting to note that the Q and N, which carry no side-chain charge,
populate the protein surface in a manner very similar to the classical titratable residues
(R,K,E,D,Y).

7.3.9 Identify Active Site Residues

Based on experimental observations, many functional features are often known about the
protein in question, aside from its amino acid sequence. Let us assume that one has
information that identifies the protein as belonging to a specific functional class of
enzymes, such as the trypsins, but a sequence alignment to any of the other members of
the family, for which the 3D structure is known, comes out negative. Despite a lack of
such information it is still possible to consider some types of protein engineering
activities. We must be prepared to err more often than in the optimal case where the
structure is known but, if the chances for a successful outcome are not zero, this approach
may still be worthwhile pursuing in an industrial setting, where the potential benefits of a
successful engineering process may be extremely attractive. If one can identify where the
active site residues are located one can engineer the function of the enzyme, by
substituting one or more of the residues close to the active site residues. Sometimes the
identity of the active site residues are known because of obvious sequence homology. At
other times it is more obscure and only more advanced methods can help identifying such
residues. One such method is MULTIM, developed in our laboratory. The method allows
for a semi-automatic sequence alignment of several sequences and shows graphically
those segments of the sequence that can be aligned. Here we have illustrated this in
Figure 7.6 by aligning a set of trypsin-like proteases for which the 3-D structures are
known. It is interesting that MULTIM finds 2 out of three of the active site residues, as
well as a third domain located at the surface of the protein but diametrically opposite to
the active site. We have no interpretation of this finding yet.

7.4 PHYSICAL CHEMICAL METHODS

7.4.1 Testing the (Engineered) Protein

In order to engineer a protein, we need a solid understanding of the native protein. If we
 Design and engineering on the nano scale 279
do not know our proteins initial properties, the engineering endeavour becomes futile.
Although 3D structural determinations have received major attention in this context, and
will be given considerable space in the present text as well, we will also highlight various
methodologies and concepts that provide very useful information about our protein.
Many of those methodologies are spectroscopic in nature. We will highlight the common
underlying principles sometimes at the

Figure 7.6 A “MULTIM” plot of 6 trypsin like sequences. Each of the

horizontal bars represents a protein sequence. Every 50 amino acids a
vertical line is drawn inside the horizontal bar. The bars are organised
pair-wise and if a common motif is found in all the analysed
sequences a connecting line is drawn between the motif locations in
all the sequence pairs. In this particular case the motif is defined as 3
 Applied biocatalysis 280
residues out of 6 has to be identical, before the program regards it as a
conserved motif. A colour code is used to indicate the statistical
significance of the motif. (See Colour Plate V)

expense of adapting a specific terminology’s that only apply to one of them. Throughout
this part we will use frequency or energy as the common descriptor of spectral position.
In Figure 7.7 an overview is given between the various energies, wavelengths,
frequencies and the commonly used names for the various spectral ranges. Molecular
modelling and physical chemical theory may predict that a certain set of mutations lead to
improvement in the stability or function of the protein. However, only direct testing of
the relevant protein properties will reveal whether the protein modification was
successful. In practice, this process is an

Figure 7.7 The relationship between photon energy in kcal/mole, wavelength

in meter and frequency in Hz is shown schematically. The commonly
used acronyms for each major frequency range is indicated at the
bottom. (See Colour Plate VI)

iteration towards a target, where each mutation reveals to which extent the adapted
strategy was appropriate.
A wide range of techniques exist, which each of them provide their own advantages. It
is relevant to be observant towards how much protein is needed for the investigation and
which intrinsic or extrinsic probe that is being monitored.
 Design and engineering on the nano scale 281

7.4.2 Protein Molecular Energy Levels

A protein is in principle just another type of molecule constructed from amino acids, each
of which are composed from a set of atoms. We typically encounter the atoms C,H,O,N
and S in the normal amino acids. In more rare circumstances we may find selenium in
“selenium cystein” residues, where selenium substitutes for sulfur, likewise amino acids
may be covalently modified with phosphate, thus occasionally phosphor, P, may be
present. Whatever the exact composition, the atoms are linked with covalent bonds,
which are rather stable with dissociation energies of 50–150 kcal/ mole (See Table 7.2).
Please note that this energy level is reached by even visual light. The fact that
photosynthesis is capable of operating with visual light might already have warned us
that single and double bond energies were similar to the photon energies in question,
since the process (although exceedingly complex in all its details) essentially is capable
of rearranging single and double bonds in CO2 and sugars.

Table 7.2 Standard bond lengths and dissociation energies for typical chemical bonds. ‡)
peptide bond length—the peptide bond has partial double bond character.

Bond Length (Å) Bond Energy (kcal/mole)

C–C 1.54 88
C=C 1.34 167
C=C 1.20 229
C–H 1.09 103
C–N 1.47 (1.34‡) 70
C=O 1.24 173
S–H 1.34 81
N2 1.10 226

At higher energy level ionisation of the molecule may result, i.e. one or more electrons
may separate from the molecule—for the simplest atom we have, the neutral hydrogen,
this event occur at 314 kcal/mole, for atomic carbon this value is somewhat lower, 260
kcal/mole. It is considerable more difficult to obtain double or triple ionisation, due to the
positive charge on the nucleus, e.g. C+ ionisation takes place at 562 kcal/mole and C++
at 1036 kcal/mole. Similar values can be expected for protein ionisation.
If the molecule does not disintegrate by bond breaking or ionisation after taking up the
energy provided by the photon, then what happens? Several other possibilities exist. The
protein may dissipate the energy in the form of low-energy vibrations or rotations or it
 Applied biocatalysis 282
may populate excited electronic states. Combinations are typically observed as we will
see later. At this point in time one may wonder how the protein molecule survives the
onslaught of photons in the visible and UV-range as well as the occasional cosmic
bombardments that takes place. Why does the protein not break apart spontaneously
when illuminated?
The optical properties of various objects can be characterised by their ability to absorb
or transmit electromagnetic radiation (photons) of a given wavelength. If an object is
transparent, the photons pass through the object without depositing the energy they are
carrying, conversely if the object absorbs the photons, it also absorb the energy they
carry. The earth atmosphere is absorbing strongly in the UV range, and is thus effectively
protecting us from the majority of the potentially damaging UV photons from the Sun.
The human body tissues absorb all light in the visible range, but most tissues except bone
structures are fairly transparent to X-rays. This is quite interesting since it tells us an
important message—somehow the energy of the incident photons have to match some
intrinsic property of the object, or they will pass unhindered, although scattering may
take place. In the case of the X-rays, they interact with the Calcium atoms of the bone
structure. Therefore an X-ray photon may not damage our protein at all, if its energy does
not match any possible energy transitions, although it may be diffracted from the protein.
The diffraction process is in nature quite similar to a collision of two solid balls, as we
see it in the billiard game. Of the two objects, one is stationary (but it does not have to be

Figure 7.8 A differential scanning calorimeter investigation of the triglyceride

lipase cutinase from Fusarium solani pisii at pH 8.5. The buffer 20
mM glycine (flat trace) and the protein solution in the glycine buffer
have been studied separately. A distinct signal at about 55 degree C
defines the thermal denaturation of the enzyme. These scans were
 Design and engineering on the nano scale 283
performed using a scan rate of 90 degree/hour.

stationary) and the other is moving, and if its trajectory crosses the location of the other
object, the energy of the moving object is distributed between the two. If the collision is
perfectly elastic the energy remains with the moving ball—if it is inelastic some level of
energy transfer will take place. We observe very similar type of events at the molecular
level when photons collide with atoms or molecules.
At this point in time we have only addressed the covalent bonds of the protein—but the
protein contains many other bonds, most of them so-called hydrogen bonds. These bonds
are intrinsically weak, each stabilises the protein with energies of the order of 0.5–4
kcal/mole. Similarly it is known that the activation energy for hydrogen-deuterium
exchange of exchangeable hydrogens is about 1.85 kcal/mole. They are thus not
absorbing photons in the visible range (see Figure 7.1)—but they will respond to infrared
and microwave radiation of a proper wavelength. It is therefore interesting that when we
heat a protein, most proteins tend to loose their 3D structure at 40–100 degree C. If an
ideal “black body” is heated in this temperature range, it will radiate photons at a
frequency in excess of 2·1013 Hz, which corresponds to an energy of at least 1 kcal/mole,
thus coinciding or overlapping with the energy range needed to break hydrogen bonds.

7.4.3 Differential Scanning Calorimetry

The thermal stability of a protein can be assessed using Differential Scanning
Calorimetry (DSC) (see Figure 7.8). The techniques relies on a carefully controlled
constant rate heating of two small ml sized samples, one of which contains the protein
dissolved in a buffer, and a reference sample, typically containing the buffer alone. The
heating rate may range from 1 degree per hour to 90 degree per hour. The temperature
range that may be covered range from a few degree C to 110 degree C. The upper limit is
feasible since the sample cells are mildly pressurised. Most proteins will exhibit a thermal
transition somewhere in the range 30–80°C. Aside from obtaining the Tm (the thermal
transition temperature) the enthalpy for the transition appear as the integral of the
transition peak. A detailed line shape analysis may provide information about the
mechanism of unfolding and in favourable cases the enthalpy for unfolding.

7.4.4 Spectroscopy Based Techniques

In a wider sense, all types of Spectroscopy expose the protein sample to a pulse or a
continuous stream of photons. The wavelength or frequency of these photons may be held
constant or it may be scanned through a preset range. Some sort of appropriate detector is
then positioned close to the sample. The relationship between what the detector records
and what we knew we exposed the sample to initially can be unravelled to various
relevant pieces of information such as how much of our protein that still possess its
native structure. Photons may behave either as electromagnetic waves or as particles.
 Applied biocatalysis 284
This will explain many of the features that we will address later in this section. Whether
or not a photon will interact in some useful manner with our protein or not, is solely
determined by its energy at the time when it collides with the protein molecule. If the
energy is not matching an intrinsic property of the molecule, the photon may even pass
straight through the protein sample—i.e. the protein sample is transparent for the
incoming photon beam. At other energy levels we may observe the energy initially is
absorbed and after some delay time is re-emitted, possibly with some apparent loss of
energy. A fraction of the energy absorbed may be converted into heat in the sample.
Although the basic spectroscopic principle appear very simple, the practical
implementations of the spectroscopic method at various photon energy levels appear
dramatically different due to the radically different technologies needed for the photon
generation as well as for detection of the emitted photon. The wavelength range covered
is very large, as is seen in Figure 7.7. In this context we will dwell only with techniques
of relevance for protein studies. We will point out which types of effects the photon is
believed to introduce.

7.4.5 UV Spectroscopy—CD and Fluorescence

The ultraviolet range of the spectrum covers from about 10 to 380 nm, corresponding to
from 3000 to 75 kcal/mole. The range normally in use for protein studies is from 170–
380 nm (168–75 kcal/mole). The spectral range is actually split in two, one, the Far-UV
is from 10 to 200 nm and the Near-UV is from 200 to 380 nm. In the near UV the
aromatic residues TRP, TYR and PHE all absorb, whereas in the Far-UV the carbonyl
groups of the protein backbone gives rise to strong bands.
 Design and engineering on the nano scale 285

Figure 7.9 A schematic representation of a fluorescence spectrometer

consisting of a lamp, an excitation monochromator with associated
slits, a sample chamber, and in a 90 degree angle to the excitation
beam an emission monochromator with associated slits and finally a
detector. (See Colour Plate VII)

If the incoming light is polarised, any chiral features of the molecule that absorbs will be
highlighted—in particular changes in such chiral features, e.g. when the protein is
unfolding, these features will stand out. The secondary structural elements of proteins are
chiral in their nature. This phenomenon can be followed with the so-called circular
dichroism (CD) .
A significant technical feature is that oxygen gas has a strong electronic absorption in
the Far-UV range, of what reason the spectrometer is often flushed with nitrogen gas
 Applied biocatalysis 286
during an experiment.
Another technique that often utilises the UV spectral range is Fluorescence
Spectroscopy. It also relies on a UV excitation, and subsequent emission perpendicular to
the incident beam (see Figure 7.9). The emission can either take place with the same
frequency (resonance fluorescence) or at a lower frequency (stimulated fluorescence).
The latter phenomenon is rooted in the ability of the UV excited state to interact with the
local environment, typically through the excitation of vibrational states of the
surrounding part of the protein molecule or of the solvent molecules.
E.g. tryptophane residues of proteins excite at 290–295 nm but they emit photons
somewhere between 310 and 350 nm. The missing energy is deposited in the tryptophane
molecular environment in the form of vibrational states. While the excitation process is
complete in pico-seconds, the relaxation back to the initial state may take nano-seconds.
While this period may appear very short, it is actually an extremely relevant time scale
for proteins. Due to the inherent thermal energy, proteins move in their (aqueous)
solution, they display both translational and rotational diffusion, and for both of these the
characteristic time scale is nano-seconds for “normal” proteins. Thus we may excite the
protein at time 0 and recollect some photons some nano seconds later. With the invention
of lasers, as well as of very fast detectors, it is completely feasible to follow the protein
relax back to its ground state with sub-nano second resolution. The relaxation process
may be a simple exponential decay, although tryptophane of reasons we will not dwell on
here display a multi-exponential decay.
In the case of fluorescence we may benefit from using polarised light. If the excitation
light is vertically polarised, a comparison of the emitted light intensities at vertical and
horizontal polarisation will reveal a difference in intensity. This difference is caused by
the molecular motion that has taken place in the time between the initial light pulse hit
the protein and the time at which the protein reemits the residual energy in the form of a
photon of similar or lower energy. We thus have a unique possibility to follow the motion
of the protein at the nanosecond range. We are limited by the natural relaxation time of
the tryptophane residue, but more long-lived probes may be chemically attached to the
protein surface.
The fluorescence response of tryptophanes and other aromatic residues may be used to
monitor the protein folding state as well. If we at time zero expose our protein to a high
concentration of denaturant, we can follow the time evolution of this unfolding event
using the fluorescence emitted from the endogenous tryptophanes. Conversely we can
take the unfolded protein at time zero and transfer it to an environment which allow
protein to refold. By extrapolation to zero denaturant concentration we can determine the
intrinsic folding and unfolding rates at a given temperature.
Despite the high energy of the excitation, fluorescence emission may be quenched by a
range of quenchers, such as iodide (I− and Cu2+). The net effect of a quencher is that the
observed fluorescence emission reduces as the quencher concentration increases
(typically milli-molar concentrations are employed). This is due to an energy transfer
between the excited fluorophore and the quencher. The quencher thus provides an
alternative relaxation pathway for the excited molecule. Not surprisingly the effect of the
 Design and engineering on the nano scale 287
quencher is correlated with its ability to reach the excited residue. If the residue is buried
in the protein interior the quencher will not be very efficient, whereas it will be highly
effective for a surface residue. Please note that since some quenchers are electrostatically
charged, they will be attracted to protein surface patches that carry the opposite polarity.
This effect is off course pH dependent.

7.4.6 Infrared Spectroscopy

With vibrational or infrared spectroscopy one can study the interaction between photons
with energy 0.1–10 kcal/mole and the protein molecule. As we alluded to above any
photon in this energy range cannot lead to molecular disintegration. Instead they will
excite various types of bond vibrations or deformations. Thus the two atoms in a bond
may display an oscillating bond length. A criteria for an infrared excitation is that the
bond must possess a permanent dipole moment. This will always be the case if the two
atoms are of different types and are therefore fulfilled for many highly interesting
components of a protein system, e.g. the peptide backbone carbonyl group. It is actually
such that this particular excitation is sensitive towards the secondary structural element in
the protein that it is a part of. A high resolution IR spectrum of a simple gas e.g. carbon
monoxide reveals a wealth of fine structure, due to the presence of many closely spaced
vibrational and rotational energy levels, from which important information about
molecular conformation and force constants for the vibrational modes can be extracted.
In solution, and in particular in protein solutions, such fine structure is not seen,
presumably due to the inherent complexity of the protein system itself, as well as
interaction(s) with the solvent. We are therefore having relatively simple broad line
spectra from which we may extract information. IR spectroscopy can also be done in a
time resolved mode, with time resolution of the order of 10th of nano-seconds. This
feature has allowed for nano-second time-resolved studies of the kinetics of CO binding
to myoglobin, and to studies of retinal binding to bacteriorhodopsin as well. A time
resolved IR study need to be triggered by an external event, such as a laser flash. In some
cases substrate molecules have been produced with a light sensitive protecting group
attached, that would detach when exposed to a laser flash. These compounds are called
“caged” molecules.

7.4.7 Nuclear Magnetic Resonance Spectroscopy

Nuclear Magnetic Resonance (NMR) spectroscopic technique utilises photons to interact
with the nuclei that display weak magnetic properties. In order to observe this effect, the
sample is immersed in a strong magnetic field, 1000 times the earth magnetic field or
even stronger. Under these circumstances certain nuclei, such like 1H, 13C, 15N, 31P and
17O are absorbing photons in the radio frequency range of the electromagnetic spectrum.
Each nucleus absorbs in a frequency range that is proportional to the field strength
applied and the type of the nucleus. After a certain period of time the nuclei will re-emit
 Applied biocatalysis 288
the radiation and relax to their ground state. The time for this relaxation is in the range of
milliseconds to seconds. The strongest signal is observed from the hydrogen-isotope, 1H.
The exact frequency where a given 1H nucleus will absorb is influenced by the local
electronic environment of the nucleus. In a classical (and probably only partially correct)
picture the local electronic environment is providing a magnetic shield for the nucleus.
This effect results in different absorption frequencies for protons in different chemical
environments such like CH3, the methyl group and C6H6, benzene. The difference is not
large, a few parts per million (ppm) but with modern NMR techniques this is easily
observable. If one wants to observe such chemical differences it is clear that the magnetic
field must not vary across the sample space beyond the ppm resolution that one wants to
achieve.

7.4.8 The Magnetic Field—Super Conducting Magnets

Strong magnetic and highly homogeneous fields is now produced routinely with super
conducting magnets. In these magnets a strong electric current is passed through a coil
made from a special metallic alloy, that looses its electrical resistance at extremely low
temperatures. Thus the coil is immersed in liquid helium, which guarantees a temperature
of about −269°C, or 4K, the boiling point for liquid helium. A strong current is started in
the super cooled coil, and the current generator is disconnected—the current in the coil
will run for years without any discernible reduction. This single coil cannot provide the
necessary homogeneity of the magnetic field—and a number of correctional secondary
coils, both super conducting and room temperature coils, are strategically placed. When
small adjustable currents are passed through these coils a highly homogeneous magnetic
field can be created.

7.4.9 Nuclear Magnetic Relaxation

Albert Einstein realised that relaxation from an excited state can occur due to
spontaneous emission or stimulated emission. The spontaneous emission from the excited
state occur without any interference from the environment. The stimulated emission is
induced by collisions with photons with energy very similar to the energy difference
between the excited state and the ground state. The ratio between the probability for
spontaneous emission and a stimulated emission is given by

Where v is the frequency of the photon, h is Plancks constant, k is Boltzmans constant

and T is the temperature in Kelvin. Thus when h.v << k.T stimulated emission dominates.
Conversely, at high frequencies h.v >> k.T the spontaneous emission is most likely. A
more detailed picture of the frequency and temperature dependence is presented in Figure
 Design and engineering on the nano scale 289
7.10. Note that in the case of fluorescence, where the energies involved indicates that
spontaneous emission in the form of a simple single transition should dominate.
Nevertheless the typical pathways back to the ground state appear to involve multiple
transitions where the excited state interchanges low energy photons with the
environment. We thus have a case where the dynamics of the environment may facilitate
a more efficient but complex multistep pathway back the ground state than spontaneous
emission provides.
Neutral hydrogen in the magnetic fields in interstellar space may have excited state life
times that is measured in years—in contrast hydrogen in liquid water at room temperature
display excited state lifetimes of 3–4 seconds. Why is this? How come that the liquid
water can lead to a more efficient relaxation of the proton

Figure 7.10 The probability ratio for spontaneous versus stimulated emission
as a function of photon frequency. The extreme left curve
corresponds to a 100 K case, the extreme right to a 1000 K case,
where the two middle traces represent “biological” range. (See
Colour Plate VIII)

excited state? A simple picture provides us with a reasonable model. The thermal energy
of water is equivalent to a black body irradiating photons with a specific energy
distribution. Some of the photons will have energies sufficiently close the emission
energy, and stimulated emission will occur. It is interesting to note that a specific
 Applied biocatalysis 290
temperature may be optimal for relaxation. Thus an increase in temperature may lead to
an increase or decrease in lifetime of the excited state. We can also deduce that a change
in magnetic field strength also will change the efficiency of the relaxation process, since
the excitation energy needed is proportional to the field strength.
From Figure 7.10 it is seen that spontaneous emission according to the Planck theory
of Black body radiation as well as Einstein’s work starts to dominate above 1013 Hz at
300K, this corresponds to the infrared range of the electromagnetic spectrum. Note, that
if the temperature increases the zero crossing point moves into the visual and UV range.

7.4.10 The NOESY Spectrum

In NMR the resonance frequency is in the range of 106–109 Hz. Thus stimulated emission
is expected to dominate the relaxation process in NMR. The stimulation of emission from
a particular nucleus may occur through interaction with the solvent or through interaction
with another nucleus in the same molecule. Two protons in the same molecule is sensing
each other through space, they will display different lifetimes of the excited state in the
presence and absence of an excited state of the other proton. It has been found
empirically that such protons display significantly different relaxation if the distance is
less than approximately 5Å. This cut-off imply that the typical environment will provide
a more efficient relaxation than the other proton, if the distance exceeds 5Å. Conversely,
we may extract distance information from the observed lifetimes if the distance to one or
more other protons is less than 5Å. This is the underlying principle of the so-called
nuclear Overhauser effect (nOe), from which the spectroscopist can deduct molecular
conformation in molecules ranging from small organic molecules to 25 KD protein
molecules.

7.4.11 Spin-Spin Coupling and Correlation Spectroscopy

Aside from the relaxation behaviour the proton is also revealing the presence of other
magnetic nuclei in the molecular environment. This type of interaction is conducted
through the bonds connecting the nuclei and is limited to 3–4 bonds. This interaction is
therefore indicative of the electronic nature of the covalent bonds. If one proton is having
a single neighbouring proton located on a vicinal carbon atom, then a highly useful effect
occur. Whereas a single isolated proton display a single excited state, and therefore emits
radiation with a single frequency, the two neighbouring vicinal protons will each display
two excited states of different energy, and therefore emit at two frequencies. The two
nuclei are said to be spin-spin coupled through the chemical bonds. This is due to the fact
that the neighbouring proton may have the magnetic moment (spin) of its nucleus aligned
in parallel or anti-parallel with the external magnetic field. More complex coupling
patterns may result from systems with a larger number of protons.
The coupling through chemical bonds can be utilised to unravel the connectivity of
more complex systems, such as proteins. In proteins we have a continous backbone of
 Design and engineering on the nano scale 291
connected peptide bonds, each of which have their own side-chain ranging from the
single proton of glycine to the much more complex aromatic side-chain of the amino acid
tryptophane.
In a protein we may have a large number of protons, typically of the order of
thousands. Each of these will give rise to one or more emissions, pending on the number
of protons to which they are coupled. The protein is composed of a combination of the 20
different naturally occurring amino acids. Thus if the protein contains 200 amino acids,
many amino acids occur more than once in the protein. Often these identical amino acids
occur at slightly different locations in the NMR spectrum. This is due to spatial effects
often caused by local electrostatic interaction between dipol moments in the individual
residues and electric charges located on the surface. Another significant effect is the
shielding effect caused by the pi-electron systems of the aromatic side-chains of
phenylalanine, tyrosin and tryptophane. These effects fortunately allow for a complete
sequential assignment of proteins of a size up to about 25 kD. Some ambiguities may
persist, but they are relatively few.

7.4.12 NMR Based 3D Structure Determination

The 3D structure of a small to medium sized protein can be obtained through NMR
experiments. Here we will give a highly simplified presentation, that still may help
grasping the underlying principles. Initially the correlation between different resonance’s
are unravelled using 2D correlation spectroscopy, a technique which results in a 2D
spectrum, where cross-peaks occur between nuclei that are less than 4 bonds separated
form each other. A careful inspection of such cross peaks will make a complete
sequential assignment of the various peaks in spectrum to a single proton in a particular
amino acid possible.
The second step is to acquire information about spatial constraints. This is done
through the recording of one or more 2D NOESY spectra, where the all cross-peaks
indicate that the two nuclei in question is less than 5 A from each other. Some of the
cross peaks will stem from 2 protons located on the same residue and seeing each other
through space. The other non-trivial NOESY peaks are identified and assigned to protons
on specific residues. A list of such distance constraints is compiled, where each entry
indicates that the two atoms in question are less than 5 Å apart. Given that enough such
constraints can be identified it is possible to construct a 3D model for the protein. The 3D
model represents one possible 3D structure that satisfies most or all of the constraints.
Several such 3D models may be generated, and each of these may represent a true picture
of the protein. These solutions typically differ in the exact conformation of solvent
exposed loops of the peptide chain, and as such they may actually be interpreted as a
plausible dynamic picture of the protein in solution.

7.4.13 Isotope Enrichment and Hydrogen-deuterium Exchange

 Applied biocatalysis 292
In some cases it is advantageous to enrich the protein with stable, NMR active isotopes
such like 13C and 15N. This is costly in isotopes but it may save weeks of NMR
experimental time or interpretation time. The advantage of isotope enrichment is caused
by the additional spectral dispersion it will introduce in the NMR spectra. An additional
dimension will be available—the 13C and/or the 15N spectral dimensions. Likewise
important information may be gained from studying how the exchangeable protons
(typically the peptide NH protons) may be exchanged with deuterium. This type of
experiment is in principle very simple—the protein is simply immersed in D2O and the
time development of the NMR signals is recorded. Any proton signal, that disappear from
the spectrum, is an exchangeable proton, and the speed at which this happens will contain
information about whether it is part of a stable secondary structure or not.

7.5 X-RAY CRYSTALLOGRAPHY

X-ray crystallographic studies of proteins have been of central importance for the
development of a detailed understanding of structure-function relationships in proteins
and, thereby, for the successful development of protein engineering as a science. In this
context it is important to note that although we today know, or can deduce from genetic
sequencing, about 150.000 protein sequences, only approximately 1200 have been
resolved structurally. Thus, our 3-Dimensional structural understanding of proteins is
very limited compared to our understanding of protein sequences. Here we will focus
upon particular issues of importance when using X-ray diffraction techniques on proteins,
rather than addressing the topic more widely.

7.5.1 The Crystal

X-ray diffraction analysis cannot be performed without a well-behaved crystal of the
protein. Several crystal forms of the protein may exist, which can present a distinct
advantage in certain contexts and a serious difficulty in others. In some cases the protein
may not appear to be crystallisable at all, which obviously prevents X-ray structure
determination. This situation is typical for membrane proteins, although a few have been
crystallised. In order to circumvent these problems, robots have been designed that allow
for a systematic screening for proper crystallisation conditions in terms of solvent
composition: salt, organic solvent, pH, temp etc. Although this is a definite advantage,
one should keep in mind that the number of possible conditions is, for all practical
purposes, infinite, since the different solvent components may be increased and decreased
independently of each other. Thus there is still no substitute for knowledge about the
proper composition solvent. The dream about the robot that solves all our problems is
still absent from present day reality.
 Design and engineering on the nano scale 293

7.5.2 Data Collection

The experimental data created in an X-ray experiment is a diffraction pattern, resulting
from the interaction between the X-rays diffracted from the various atoms in the protein
molecule that are larger than the wave length of the X-ray beam. Since the typical wave
length applied is 1,8 to 2 Angstrøm, the X-ray data do not contain information about the
hydrogens in the protein, whereas carbon, oxygen and nitrogen contribute to the date set.
Data collection can either take place using a socalled rotating anode or an area detector.
The latter invention allows the experimenter to collect a very large number of data points
simultaneously and thus reduces the time required to obtain a full data set, suitable for the
reconstruction of the protein structure. This, in combination with modern computer
technology and recent software developments, has increased markedly the pace at which
new protein structures have been solved. An additional factor has undoubtedly been the
current success of modern biotechnology, where the end product is often a protein.

7.5.3 Heavy Atom Derivative

The diffraction pattern is only detected as a pattern of intensities at given locations
around the crystal—there is inherently no phase information available in a single data set.
In order to overcome this problem, crystals are either soaked with heavy atoms (e.g.
Hg2+) or entirely new crystals are grown from a medium containing the appropriate
heavy metal ion. Alternatively, one can modify the protein chemically and, for example,
incorporate iodines at well defined locations. The presence of one or more large atoms in
the structure causes additional diffraction of the incoming X-ray beam. By comparing a
diffraction data sets in the presence and absence of heavy atoms, the X-ray
crystallographer can obtain the missing phase information and, subsequently, he can
attempt to solve the structure. The latter involves a transformation of the phase and
intensity information to cartesian space co-ordinates (x, y, z) for the atoms that gave rise
to the diffraction pattern. This transformation is only possible if the phase information is
available.

7.5.4 Refining the Structure

The accumulated knowledge about protein as well as other bio-molecular structures is
growing constantly, permitting increased understanding of the motion and stability of
these molecules. Software capable of utilising such knowledge is now available and
proper usage of such programs can result in an additional refinement of the model beyond
that possible using only the experimental data. This approach is in common use and a few
words of caution may be appropriate: inherently the refining process is based on existing
knowledge but, by assuming that this knowledge is complete, one does run the risk of
excluding solutions that satisfy all experimental date.
 Applied biocatalysis 294

7.5.5 The Brookhaven Protein Data Bank (PDB)

After refining the protein structure the X-ray crystallographer (or NMR spectroscopist)
submits his refined structure to the Brookhaven Protein Data Bank. However, structures
that have been solved as part of an industrial co-operation may be kept secret by an
industry which has financed the study wholly or in part. Currently approximately 700
structures are available, in the data bank but some of these have only been deposited as
their sequence, whilst the data are being worked up.

7.6 PROTEIN ELECTROSTATICS

The importance of electrostatic effects in protein structure and function has long been
recognized (Tanford and Kirkwood, 1957; Perutz, 1978; Warshel and Russell, 1984;
Rogers, 1989). It is generally believed that electrostatic field effects are among the first
that come into play when two protein molecules approach each other. The reason for this
is that electrostatic fields reach much further out than any other force fields that we
associate with protein-protein interaction. It is reasonable to expect that motion in the
combined electrostatic field of both molecules helps to orient the molecules for
subsequent docking.
When we have access to an interesting protein structure and we would like to engineer
the molecule, it is not enough to limit one’s preparations to a computer graphics study of
the local 3-D environment around the residue(s) in question. It
 Design and engineering on the nano scale 295

Figure 7.11 A: The electrostatic surface potential of trypsin (left) and gamma
 Applied biocatalysis 296
chymotrypsin (right) displayed at the pH values 7 (upper trace) and 9 (lower
trace). The color scale used ranged from −4*kT to +4*kT—blue
denotes negative potential and red positive potential. Note the
considerable differences in the protein electrostatic appearance as a
function of pH and enzyme. B: A superposition of the protein
backbones of trypsin (blue) and gamma chymotrypsin (cyan). The
active site residues are high-lighted in green (histidine), red Aspartate
(Asn in the case of 2gch) and yellow (serine). (See Colour Plate IX)

may be of paramount importance to understand if or how one perturbs the electrostatic

field of the protein by introducing new residues in the protein sequence. In order to
illustrate this important fact we have calculated the electrostatic fields of the two
enzymes 1trm and 2gch, which we aligned in Figures 7.3 and 7.6. The result is shown in
Figure 7.11a. It is clear that the electrostatic fields have changed significantly, although
Figure 7.11b shows that the backbones of the two enzymes are almost super-imposable.
However, we are still not fully capable of describing the charge-charge interaction in
an entirely satisfactory way. Part of the reason for this difficulty is due to the complex
dielectric properties of the system: the protein environment, which is believed generally
to be of low dielectric constant (2–4), and the solvent water, which can be described by a
much higher dielectric constant (80 or less). The presence of salt, counter ions,
membranes or protein complexes may modify the latter in particular. Interactions, which
need to include the boundary between the two dielectric media, are complex and several
methods have been developed (Warshel and Russell, 1984; Rogers, 1989). Some methods
are very rigorous, but so computer-intensive that they cannot be used to describe
molecules as large as a normal sized protein. Various approaches have been developed,
involving the introduction of cut-off distances (typically 10 Å), but the inherent long-
range nature of electrostatic interactions makes this approach less than satisfactory.
Macroscopic models have been developed that describe the protein and the water as
macroscopic dielectric materials. In their simplest form these models use a distance-
independent dielectric function, i.e. a simple Coulomb interaction. Others may apply a
distance-dependent dielectric function. The more detailed implementations include a
descriptions of the protein-solvent boundary in terms of solvent accessibility and ionic-
strength effects (Gilson and Honig, 1988).
Microscopic models aim at describing the charge-charge and charge-solvent interaction
in full molecular detail, thus circumventing the use of a dielectric constant (Warshel and
Russell, 1984; Russel and Warshel, 1985; van Belle et al., 1987). Although this may
seem like the proper approach, practical limitations on computer capacity and speed
make approximations unavoidable, especially in the solvent description.

7.7 PROTEIN ENGINEERING

In Figure 7.12 a schematic representation of the Protein Engineering work-cycle is

 Design and engineering on the nano scale 297
presented. A protein (centre) is identified, and is produced either by purification of a
natural source for the protein or by genetic engineering (lower cycle) using an
appropriate host organism. The 3D structure of the protein is solved using e.g. X-ray, or
if feasible using NMR. If none of these two approaches are possible, the 3D structure
may be predictable, if the protein shares sufficient homology with a protein for which the
3D structure is already known. After the structure has been produced, molecular
modelling and computational approaches are used in order to investigate which amino
acid changes one could consider experimentally. In

Figure 7.12 The protein engineering cycle. The lower circle represents the
genetic engineering activities—the upper the structural
determination, modelling and protein characterisation. (See Colour
Plate X)

parallel, activity measurements and physical chemical investigations of the native protein
are performed. After consulting with the genetic team, the agreed upon mutants are
produced and purified. The resulting mutant protein(s) are tested for activity and physical
 Applied biocatalysis 298
chemical properties. If fully successful the engineering process may terminate here, most
likely one or more iterative cycles may be needed before a novel optimised protein have
been produced.

7.8 CASE STORIES

7.8.1 The Engineering of a Molecular Switch in T4-Lysozyme

Brian Mathews’ group in Oregon has worked extensively with lysozyme and, in 1989,
they reported a remarkable result: by substituting both THR21 and THR142 with CYS they
could reversibly block enzyme function by establishing, under oxidising conditions, a
disulphide bridge between positions 21 and 142. Upon exposure to reducing agents the
disulphide bridge disappears and enzymatic function is fully restored. Position 21 and
142 are located at the surface but at opposite sites of the active cleft of the enzyme. The
engineered disulphide bridge is thus physically blocking the active site from interacting
with substrates.

7.8.2 Enhancing the Thermostability of a Protein by Decreasing the

Entropy of Unfolding
In surface positioned loops a high occurrence of GLY, AT A and PRO is found. Upon
unfolding the PRO residues will still restrict the motions of the backbone of the protein
since the torsion angles can only alter by altering the pucker of the ring system of PRO
which is locked into the backbone. In contrast, a GLY in the same position will give rise
to a maximum degree of freedom around the torsion angles since GLY does not have any
side chain and hence do not restrict intramolecular reorientation to any significant extent
compared to PRO. This difference can be viewed as a dramatic difference in the gain of
entropy following unfolding. Two such mutations were studied in the case of lysozyme:
GLY77→ALA and ALA82→PRO. At neutral pH a small but significant increase in the
denaturation temperature was reported. The largest effect was found for the ALA82→PRO
where a 2.1 ±0.5°C increase was measured.

7.8.3 Introducing Salt Bridges to Enhance Thermostability

In an investigation (Barlow and Thornton, 1983) of the known protein structures, it has
been found that charged residues seem to prefer an environment containing oppositely
charged side groups. About one third of the charged groups participate in salt bridge
formation and it is concluded that the charge contributed to tertiary structure stabilization
and not to the stability of secondary structure. It has been proposed (Rogers 1989) that
entropy is the driving force for surface-exposed groups whereas enthalpy is dominant for
 Design and engineering on the nano scale 299

buried charge pairs in proteins. Introduction of a salt bridge (ASP70-HIS31) into T4

Lysozyme stabilized the folded state by 3–5 kcal/ mol (Anderson, Becktel and Dahlquist,
1990).
A structural feature in proteins that deserves attention in this context is the macro
dipole of -helices (Hol, 1985a,b; Hol, van Duijnen and Berendsen, 1978; Hol, Halie and
Sander, 1982). In an a-helix the local dipole moments of the carbonyl groups as well as
the amide N-H are aligned and result in a macro dipole with the positive pole near the N-
terminal and the negative pole at the C-terminal. This dipole may be further stabilized by
oppositely charged residues in the immediate vicinity of the terminal. This has indeed
been observed and, in this context, it is interesting to refer the reader to Figures 7.4
addressing the location of PRO’s in -helices and the likelihood of PRO sitting next
neighbor to another amino acid, which is very high for GLU. Since PRO is almost
exclusively found in the first turn of -helices the PRO-GLU pair therefore cluster close to
the positive pole of the helix dipole thus lending support to the postulate cited above.
Similarly we note that the ARG and LYS are somewhat under represented as we should
expect based on the same argument.

7.8.4 Barnase—a Model System

The small 110 a.a. ribonuclease from B. amyloliquefaciens, barnase, has been studied
extensively in A.R.Fersht’s laboratory (Sali, Bycroft and Fersht, 1988; Matouschek et al.,
1990; Bycroft et al., 1990; Serrano and Fersht, 1989) using NMR as an essential
experimental technique. The -helices of barnase have been studied with respect to
stability and it was found that mutation of the THR6 and THR26 residues located at the N-
terminal of the helix (Serrano and Fersht, 1989) could destabilize the protein with up to
2.5 kcal mol−1. The THR and SER residues are capable of facilitating the formation of an
additional hydrogen bond in the first turn of the -helix. If THR is substituted with ASP
or GLU no marked change in stability was observed, probably due to the known charge-
dipole interaction between the negative charge of the ASP or GLU and the positive end
of the helix macro-dipole.
NMR has been used to study the dynamics of the folding process for barnase
(Matouschek et al., 1990; Bycroft et al., 1990). By observing the time course of
hydrogen-deuterium exchange, in particular, of the peptide back-bone of the protein, it
has been demonstrated that the folding process can be unraveled in at least two stages.
The first stage is very rapid, with a rate constant of 250–500 s−1, whereas the second is
one order of magnitude slower, 12–28 s−1. It is proposed that the second stage rate
constant is to a large extent dependent on the slow isomerisation of PRO’s 21, 47 and 64.

7.8.5 Subtilisin Proteases

With subtilisin nature has provided us with a model system for protein engineering
studies (Wells and Estell, 1988). Being a small single domain serine protease
 Applied biocatalysis 300
(MW~27,500) with no cofactors or metal ions requirement for its function, it displays
Michaelis-Menten kinetics and it is secreted in large amounts by a wide variety of
Bacillus species. Subtilisin is also among the most important industrial enzymes due to its
use in laundry detergents. Protein engineering strategies for subtilisin have focused on a
number of aspects, namely catalysis, substrate specificity, thermal and oxidative stability
and pH profile. We will describe briefly each of these aspects.

7.8.6 Catalysis
Early mutational studies on subtilisin were aimed at probing the importance of the active
site residues for catalysis. Carter and Wells (1988) made mutants of subtilisin BPN (from
Bacillus amyloliquefaciens) with each of the three active residues (ASP32, HIS64 and
SER221) replaced by ALA. In every case there was a ~105 fold decrease in kcat. Mutating
all three residues at one time produced a similar decrease in kcat. The residual catalytic
activity (103-fold higher than non-enzymatic catalysis) was presumed to result from
favorable interactions in the transition state complex. ASN155 can form a hydrogen bond
with the scissile peptide bond CO and mutation of this residue does indeed reduced kcat
by a factor of 102–103. Interestingly enough, the ASN155→GLY/SER221→WALA has its
kcat increased fivefold, which is explained by an alternate hydrolytic mechanism (Storer,
1991), where water would interact directly with the scissile bond. Takagi et al. (1990)
found that the mutation LEU31→ILE produces a 2–6 fold increase in kcat for peptide
substrates in subtilisin E. ILE is indeed present in some natural forms of subtilisin
displaying higher activity than subtilisin E. ILE32 is presumed to improve orientation of
active site ASP32.

7.8.7 Substrate Specificity

Subtilisin has a broad substrate specificity in the P1 position (residue contributing the CO
of the scissile bond). Efforts have been made to engineer it into a more specific protease.
Replacement of GLY66 (at the bottom of the specificity pocket) by bulkier residues
increases preference towards small uncharged substrates (Estell and Wells, 1988).
Charged substitutions or residues 156 and 166 can form salt bridges with P1 side chains,
thus causing catalytic preference to residues of the opposite charge. Specificity to GLU at
P1 increased 1900-fold over the wild type in the mutant GLU156→GLN/GLY166→LYS.
Substrate-assisted catalysis occurs in the HIS64→ALA, where HIS at P2 position can
replace the missing HIS64. This mutant form shows a 800-fold relative preference for
substrates containing His in the P2 position (Carter et al., 1989; 1991). Recent efforts
have been made to produce mutant forms having preference for ester substrates and
higher aminolysis rates, for peptide synthesis applications (Bonneau et al., 1990;
Abrahmsen et al., 1991).
 Design and engineering on the nano scale 301

7.8.8 pH Profile
There has been a number of efforts to engineer a broader or shifted pH profile into
subtilisin. The acidic leg of the pH profile is controlled by HIS64 protonation and a
downward shift in the pKa of HIS64 was expected to extend the acidic leg, since loss of
activity is probably due to HIS protonation. Engineering of surfaces charges did cause a
downward pKa shift in the HIS (Russel, Thomas and Fersht, 1987), with a concomitant
broadening in the acidic leg of the pH profile. In the most extreme case the mutant
ASP99→LYS/GLU156→LYS had its pKa shifted down one unit. The observed shifts could
be predicted theoretically using a linearised Poisson-Boltzmann model to describe the
charge-charge interactions (Sternberg et al., 1987). Loss of activity in the alkaline region
seems to be due to Tyrosine ionisation, as an apparent pKa of 10.5 indicates (Estell and
Wells, 1988). To test this hypothesis the mutant TYR104→PHE was prepared and
extension of the alkaline leg was indeed observed (Estell and Wells, 1988). These results
demonstrate the possibility of engineering the pH profile of subtilisin. However,
stabilising mutations have to be found if wild type catalytic rates are to be attained.

7.8.9 Stability
Improved stability has been another major goal of subtilisin engineering projects.
Stability to oxidising agents has been an industrial concern. Oxidation of MET222 rapidly
inactivates subtilisin and therefore replacements at this position have been tried. A
systematic analysis of the 19 possible 222mutants showed activities ranging from 130
down to 0.3 percent of wild type activity. The mutants MET222→ALA and MET222→SER
were stable to H2O2 oxidation over long periods of time (Wells et al., 1987), but with a
decrease in specific activity. Groen et al. (1990) used a combination of mutagenesis and
chemical modification to produce an oxidation-resistant mutant of Bacillus lentus
subtilisin with an activity close to wild type. The mutation MET222→CYS is
complemented with chemical modification of the cysteinyl group, rendering it resistant to
oxidation.
Diverse attempts have been made to improve stability to irreversible denaturation,
which can be caused by autolysis, high temperature, extreme pH, denaturants, etc.
Engineering disulfide bonds had a mixed success: the first 7 mutants produced did not
show a significantly higher autolytic stability. Moreover, the strength of the bond did not
correlate with the observed increase in stability (Mitchinson and Wells, 1989). In one
case a disulfide bridge was introduced based on homology with a subtilisin variant of
known structure, protease K. The mutant showed a decreased autolytic stability,
indicating the need to import other features in the protease K structure. Takagi et al.
(1990) reported the inclusion of a disulfide bridge in subtilisin E, based on homology
with aqualysin I. The GLY61→CYS/SER98→CYS showed increased thermal stability
while keeping a wild type catalytic efficiency.
Stabilization through inclusion of salt bridges has also been tried. Erwin et al. (1990)
 Applied biocatalysis 302
used the PROTEUS software to predict salt bridge formation in structures containing
charged mutations. Of the secreted mutated forms there was only one, GLN19→GLU,
which displayed significantly higher autolytic stability. Evidence for salt bridge
formation was corroborated by X-ray analysis of the mutated structure. There was no
evidence for salt bridge formation in the remaining structure, indicating the poor
predicting ability of PROTEUS. This may be due to the over simplistic model used to
describe the charge-charge interaction in the salt bridge.
Narhi et al. (1991) recently reported an enhancement in the thermal stability of aprA-
subtilisin by three point mutations. The mutations were ASN109→SER and ASN218→SER
to prevent cyclisation with the adjacent glycines and ASN76→ASP in the Ca2+binding
loop. The mutant form also exhibits improved stability to detergent denaturation with
little dependence on calcium concentration. Subtilisin 8350 (derived from subtilisin BPN’
via six site-specific mutations) was found to be 100 times more stable than the wild type
enzyme in aqueous solution and 50 times more stable than the wild type in anhydrous
dimethylformamide (Wong et al., 1990)
The oxidative stability of subtilisin has been extensively studied and improved stability
has been engineered. In subtilisin BPN’ two methionines, MET124 and MET222 are
especially susceptible to oxidation. To prevent the negative influence caused by the
formation of methionine sulfoxide the MET can be substituted with ALA, SER or LEU,
without loosing more than 12–53% of the activity. One such mutant MET222→ALA is
currently in use as a commercial detergent enzyme: ‘Durazyme’ (Riisgard, 1990).

7.8.10 Subtilases
A comprehensive review has been published on protein engineering strategy for
subtilases (Siezen, 1991). The review describes more than 50 subtilases known and their
homology. Two main classes could be identified by studying the N-terminal domains.
191 structurally conserved core residues were identified and 18 of these are highly
conserved, 9 of which are GLY. Whereas the core regions show considerable homology,
the loop regions are much more variable.

7.8.11 A Metal Ion Switch in Trypsin

Metal ions play an important role in both protein structure and function (Boel, 1990;
Tainer, Roberts & Getzoff, 1991). By introducing a HIS instead of an ARG in the
immediate environment of the active site HIS57 in trypsin a Cu2+binding site was created
(Higaki, 1990). With Cu2 +bound the enzyme was inactive—upon removal it becomes
active.

7.8.12 Insulin
Native human insulin forms associates under physiologically relevant conditions.
 Design and engineering on the nano scale 303
Monomers build asymmetric dimers, which in turn form hexamers (of 3 dimers). Higher
associates also are formed. It was of great interest to construct an engineered modified
form of human insulin where the monomer was prevented from forming a dimer, and
thereby also hexamers and higher associates. At the same time the biological potency, of
course, needed to be maintained, so residues involved in or providing for receptor
interactions could not be mutated. Since a wealth of information is available on insulins
of various origins, and since the 3-D structure of insulin was available in PDB, Brange et
al. (1988) could design mutations aimed at producing a true monomeric insulin. Two
routes were investigated: at the dimeric interface a -sheet-like hydrogen bond pattern is
present between two strands, one from each monomer. Since the interface, in addition,
represents a very close spatial fit, mutations introducing larger residues, which did not
perturb the local hydrophobicity, were made and they did indeed obstruct the formation
of dimeric insulin. In addition, mutations which modified the electrostatic interactions
between the two monomers were introduced at the edge of the interface. The
PROB28→ASP mutant was found to prevent the formation of dimeric insulin very
efficiently. The reason for this is that B28 in one monomer is located close to two
negatively charged residues GLUB21 and GLUA4. The interpretation of this particular
result is that electrostatic repulsion involving ASPB28 GLUB21 and GLUA4 destabilizes
the dimer.

7.8.13 De novo Design of Proteins

Many groups are currently working towards designing proteins de novo (Sander, 1991).
The first aim in such a design process is to engineer successfully a protein structure of a
predefined fold. Obviously and structures and mixtures thereof can and have been
considered. A four helix bundle ‘Felix’ was created by Richardson’s group (Hecht,
1990). As judged from CD the protein is 50–65% helical in solution. In the case of Felix,
the protein was synthesized as a single peptide strand. Another approach, promoted by
Mutter et al. (Mutter and Vuileumier, 1989), utilizes a Template-Assembled Synthetic
Protein strategy, where an organic chemical molecule serves as the template onto which
short peptide segments can be covalently attached. By proper choice of the peptide
segments, one can promote various types of secondary structures in the peptide domain.
Both -, -and - structures have been synthesized.
De novo design of proteins with pre-specified enzymatic activity is one of the most
complex goals of protein engineering. So far only a few results have been reported, but
they are extremely interesting because they do prove that we are at the level where our
insight into protein structure is sufficiently advanced to allow for de novo enzyme design.
Helichrome, an artificial hemeprotein that can hydrolyze aniline, has been reported
(Sasaki and Kaiser, 1989). The enzyme is active but significantly weaker than the natural
enzyme.
A four helix fold has also been used for de novo design of chymotrypsin-like enzyme
(Hahn, Wieskaw and Stewart, 1990). The trypsin catalytic triad SER, HIS and ASP has
 Applied biocatalysis 304
been incorporated at the ends of three of the four helices. A weak esterase activity was
indeed found but a 10-fold reduction in activity compared with the natural enzyme
indicates that the steric environment is far from optimal for the active site analog in this
design.

7.9 CONCLUSION

Protein Engineering has already contributed dramatically to our understanding of protein

structure and function. The possibility of doing protein engineering has in turn
encouraged a large number of X-ray Crystallographic groups to solve interesting protein
structures, often as part of a publicly funded protein engineering effort. Many European,
North American and Japanese funding agencies have uniquely identified protein
engineering as an enabling technology of great importance for their industry, in particular
for the biotechnological industry. More visionary individuals foresee bold applications of
engineered proteins outside the biotechnology and medical area of science in the future,
an expectation the author fully share. Whatever the framework, the de novo design
strategy is extremely interesting. At this point in time it is far from being a real industrial
alternative to modifying, and thereby optimising, existing protein structures of interest.
However, the wealth of structural and functional information it potentially can create
makes it a key activity for the future success of Protein Engineering.

7.10 ACKNOWLEDGMENTS

This manuscript is an extension and a major revision of the manuscript for the first
edition of this book that I wrote together with Dr. Paulo Martel, now at ITQB, Oireas,
Portugal. His contributions directly and indirectly are much appreciated. The author is
grateful for the generous support he has received from Nordjysk Energifond, Obelsk
Familiefond as well as Mål-2.

7.11 EXERCISES

7.1. Name the major classes of macromolecules in biology. Outline the molecular
differences between them. Is enzyme like function limited to proteins?

7.2. How many different ways can a penta-peptide be composed from the 20 natural
occurring amino acids. Answer the same question for a dodeca-peptide.
 Design and engineering on the nano scale 305

7.3. Where is it most likely to find a hydrophobic amino acid—on the surface of a
protein or buried in the interior?—Same question for a hydrophilic residue.

7.4. What happens to the fraction of hydrophobic residues if you compare a 100 AA
protein with 2000 amino acid protein. Assume that both proteins are globular.

7.5. A protein has a pI of 6.2. What will you expect of its solubility as a function of
pH?

7.6. A patient is displaying signs of a dysfunctional protein. The protein is identified

and investigated. Compared to the normal human protein the only change detected was
that an alanine had been substituted with a tryptophane residue. Fluorescence based
thermal stability investigations of the mutant and normal protein revealed that the mutant
protein displayed a thermal transition 10 degree below the normal protein, which
unfolded at 60 degree C . What can you deduct from these observations?

7.7. A Differential Scanning Calorimetry experiment gave the results that Protein X
entered a thermal transition at 65 degree C. No other transition was observed in the
temperature range 10–110 degree C. The fluorescence response from the single
endogenous tryptophane in protein X was monitored as a function of temperature as well.
A distinct change was observed at 45 degree C, but nothing above this temperature.
Which of the two temperatures will most likely represent the collapse of the 3D structure.
Explain these observations.

7.8. A protein is composed of two distinct domains linked together with a short peptide
fragment. The protein is poorly soluble and no crystallization attempts have succeeded.
The protein was studied with circular dichroism. The band at 223nm was followed as a
function of temperature as well as of pH for both the intact protein as well as the isolated
domains. For the intact protein a clear thermal transition occur at 55 degree C at pH 5.5
and the pH profile show a inverted bell shaped curve with a minimum at pH 5.5. The
isolated domains show transition temperatures of 48 and 30 degree C, respectively. The
intact protein was then studied with Differential Scanning Calorimetry as a function of
pH and again a bell shaped curve was obtained with a thermal stability maximum around
pH 5.5. What can you deduct from these observation with respect to the protein?

7.12 HINTS AND ANSWERS

7.2. A pentapeptide can be composed in 205=20*20*20*20*20=3.200.000 different

ways. For the dodeca peptide the answer is 2012 different ways.
 Applied biocatalysis 306

7.3 A hydrophobic amino acid will most likely be found in the hydrophobic core—a
hydrophilic residue will (at least for water soluble proteins) most likely be located on the
surface.

7.4 The larger the protein, the more hydrophobic residues it will contain in its core.
This problem is essentially a problem of evaluating the ratio between the surface area and
the volume as a function of radius. The surface area of a sphere is 2*pi*r2—the volume
of a sphere is 4/3 * pi*r3—thus the relative growth in surface area is slower than the
growth in volume for a given change in radius.

7.5 The solubility at the isolelectric point is lower than the solubility at pH values away
from pI. One way of explaining this is that the electrostatic repulsion between different
protein molecules is at a minimum at pI.

7.6 The A-> W mutation in all likelyhood took place in the core of the protein. Since
W is much larger than A the hydrophobic core of the protein cannot pack as efficient as
before and the melting point is as a consequence much lower.

7.7 A DSC investigation describe the global changes of the protein as a function of
tempearture. A fluorescence experiment which report the response of a single
tryptophane residue is essentially a local measurement of the environment around the
tryptophane.

7.8 The fact that the individual domains each exhibit a lower transition temperature
than the intact two-domain protein and the observation that the transition temperature is
higher than for the individual domains indicate that the two domains are making multiple
non-covalent bonds (Hydrogenbonds, hydrophobic contacts, salt bridges etc) to each
other. The single thermal transition of the intact two domain protein indicate that both
domain go through a single cooperative transition.

7.13 REFERENCES

Abrahmsen, L., Tom, J., Burnier, J., Butcher, K.A., Kossiakoff, A. and Wells, J.A. (1991)
Engineering subtilisin and its substrates for efficient ligation of peptide bonds in
aqueous solution. Biochemistry , 30, 4151–4159.
Anderson, D.K., Becktel,W.J. and Dahlquist, F.W. (1990) pH-Induced Denaturation of
Proteins: A single Salt Bridge Contributes 3–5Kcal/mol to the Free Energy of Folding
of T4 Lysozyme. Biochemistry , 29 , 2403–2408.
Aqvist, J., Luecke, H., Quiocho, F.A. and Warshel, A. (1991) Dipoles localized at helix
termini of proteins stabilize charges. Proc. Natl. Acad. Sci. USA , 88 , 2026–2050.
 Design and engineering on the nano scale 307
Bairoch, A. Swiss-Prot , Department de Biochemie Medicale, Centre Medical
Universitaire, 1211 Geneva, Switzerland.
Barlow, D. and Thornton, J.M. (1983) Ion Pairs in Proteins. J. Mol. Biol. , 168 , 867–885.
van Belle, D., Couplet, C., Prevost, M. and Wodak, S.J. (1987) Calculations of
Electrostatic Properties in Proteins. J. Mol. Biol. , 198, 721–735.
Boel, E., Brady, L., Brzozowski, A.M., Derewenda, Z., Dodosn, G.G., Jensen, V.J. et al.
(1990) Calcium Binding in -Amylases: An X-ray diffraction Study at 2.1 A Resolution
of two enzymes from Aspergillus. Biochemistry , 29, 6244–6249.
Bohr, H., Bohr, J., Brunak, S., Cotterill, R.M.J., Lautrup, B., Nerskov, L. et al. (1988)
Protein Secondary Structure and Homology by Neural Networks: The alpha helices of
Rhodopsin. FEBS Letters , 241, 223–228.
Bohr, H., Bohr, J., Brunak, S., Cotterill, R.M.J., Fredholm, H., Lautrup, B. and Petersen,
S.B. et al. (1990) A novel approach to prediction of the 3-Dimensional structures of
protein backbones by neural networks. FEBS Letters , 261, 43–46.
Bonneau, P.R., Graycar, T.P., Estell, D.A. and Jones, J.B. (1990) Alteration of the
specificity of subtilisin BPN by site-directed mutagenesis in its Sl and Sl’ binding sites.
J. Am. Chem. Soc. , 113, 1026–1050.
Branden, C. and Tooze, J. (1991) Introduction to Protein Structure. New York and
London: Garland Publishing.
Brange, J., Ribel, U., Hansen, J.F., Dodson, G., Hansen, M.T., Havelund, S. et al. (1988)
Monomeric insulins obtained by protein engineering and their medical implications.
Nature , 383, 679–682.
Bycroft, M., Matouschek, A., Kellis Jr, J.T., Serrano, L. and Fersht, A.R. (1990)
Detection and characterization of a folding intermediate in barnase by NMR. Nature ,
346, 488–490.
Carter, P., Nilsson, B., Burnier, J.P., Burdick, D. and Wells, J.A. (1989) Engineering
subtilisin BPN for site-specific proteolysis. Proteins , 6, 240–248.
Carter, P., Abrahmsen, L. and Wells, J.A. (1991) Probing the mechanism and improving
the rate of substrate-assisted catalysis in subtilisin BPN. Biochemistry , 30, 6142–6148.
Carter, P. and Wells, J.A. (1988) Dissecting the catalytic triad of a serine protease.
Nature 7, 332(6164), 564–568.
Chou, P.Y. and Fasman, G.D. (1974a) Conformational Parameters for amino acids in
helical, -sheet and random coil regions calculated from proteins. Biochemistry , 13 ,
211–222.
Chou, P.Y. and Fasman, G.D. (1974b) Prediction of Protein Conformation.
Biochemistry , 13, 222–245.
Chou, P.Y. and Fasman, G.D. (1977) -Turns in proteins. J. Mol. Biol. , 115, 135–175.
Chou, P.Y. and Fasman, G.D. (1978) Prediction of Secondary Structure of Proteins from
their amino acid sequence. Adv. Enzymol. , 47, 45–148.
Chou, P.Y. and Fasman, G.D. (1979) Prediction of -Turns. Biophys. J., 26, 367–384.
Crippen, G.M. and Havel, T.F. (1988) Distance Geometry and Molecular Conformation.
Taunton, Somerset, UK: Research Studies Press Ltd and New York: John Wiley &
Sons Inc.
Crippen, G.M. (1989) Distance Geometry. In Computer Aided Design , edited by
G.Richards, 61–65. IBC Technical Series.
EMBL Data Library, Postfach 10.2209, D-6900. Germany: Heidelberg.
 Applied biocatalysis 308
Erwin, C.R., Barnett, B.L., Oliver, J.D. and Sullivan, J.F. (1990) Effects of engineered
salt bridges on the stability of subtilisin BPN. Protein Engineering , 4(1), 87–97.
Fasman, G.D. (1989) Prediction of Protein Structure and the Principles of Protein
Conformation . New York: Plenum Press.
Fasman, G.D. (1989) The Development of the Prediction of Protein Structure. In
Prediction of Protein Structure and the Principles of Protein Conformation , edited by
G.D.Fasman. New York: Plenum Press.
Fasman, G.D. (1985) A critique of the utility of the prediction of protein secondary
structure. J. Biosci. , 8, 15–23.
Freedman, R. and Wetzel, R. (1991) Protein Engineering. Current Opinion in
Biotechnology , 2.
Garnier, J. and Robson, B. (1989) The GOR Method for predicting Secondary Structures
in Proteins. In Prediction of Protein Structure and the Principles of Protein
Conformation , edited by G.D.Fasman. New York: Plenum Press.
Garnier, J., Osguthorpe, D.J. and Robson, B. (1978) Analysis of the accuracy and
implications of simple method for predicting the secondary structure of globular
proteins. J. Mol Biol. , 120, 97–120.
J.Gibrat, Garnier, J. and Robson, B. (1987) Further developments of protein secondary
structure prediction using information theory. New parameters and consideration of
residue pairs. J. Mol. Biol. , 198. 425–443.
Gierasch, L.M. and King, J. (eds.) (1990) Protein Folding , American Association for the
Advancement of Science.
Gilson, M. and Honig, B. (1988) Energetics of Charge-Charge Interactions in Proteins.
Proteins , 3, 52–52.
Gilson, M.K. and Honig, B.H. (1987) Calculation of the electrostatic potentials in an
enzyme active site. Nature , 330, 84–86.
Groen, H., Bech, L.M., Branner, S. and Breddam, K. (1990) A highly active and
oxidation-resistant mutant subtilisin-like enzyme produced by a combination of site-
directed mutagenesis and chemical modification. Eur. J. Biochem. , 194, 897–901.
Hahn, K.W., Wieskaw, A.K. and Stewart, J.M. (1990) Design and Synthesis of a peptide
having Chymotrypsin-Like Esterase Activity. Science , 248, 1544–1547.
Hecht, M.H., Richardson, J.S., Richardson, D.C. and Ogden, R.C. (1990) De novo
design, Expression and Characterization of Felix: A four helix bundle Protein of
Native-Line Sequence. Science , 249, 884–891.
Higaki, J.N., Haymore, B.L., Chen, S., Fletterick, R.J. and Craik, C.S. (1990) Regulation
of Serine Protease Activity by an engineered Metal switch. Biochemistry , 29, 8582–
8586.
Hol, W.G.J. (1985a) The role of the n-helix dipole in protein function and structure.
Prog. Biophys. Mol. Biol. , 45, 149–195.
Hol, W.G.J. (1985b) Effects of the n-helix dipole upon the functioning and structure of
proteins and peptides. Adv. Biophys. , 19, 133–165.
Hol, W.G.J., van Duijnen, P.T and Berendsen, H.J.C. (1978) The n-helix dipole and the
properties of proteins. Nature , 273, 443–446.
Hol, W.G.J., Halie, L.M. and Sander, C. (1982) Dipoles of -helix and -sheet: Their
role in protein folding. Nature , 294, 532–536.
Holley, L.H. and Karplus, M. (1989) Protein secondary structure prediction with a neural
network. Proc. Natl. Acad. Sci. USA , 86, 152–156.
 Design and engineering on the nano scale 309
Jackson, J.D. (1975) Classical Electrodynamics. Wiley.
Kabsch, W. and Sander, C. (1983) Dictionary of protein secondary structure: pattern
recognition of hydrogen bonded and geometrical features. Biopolymers , 22, 2577–
2637.
Loomis, W.F. (ed.) (1988) Four Billion Years . Sunderland, Massachusetts: Sinauer
Associates.
Matsumura, M. and Matthews, B.W. (1989) Control of Enzyme Activity by an
Engineered Disulphide Bond. Science , 243, 792–794.
Mathews, B.W., Nicholson, H. and Becktel, W.J. (1987) Enhanced protein
thermostability from site-directed mutations that decrease the entropy of unfolding.
Proc. Natl. Acad. Sci. USA , 84, 6663–6667.
Matouschek, A., Kellis Jr, J.T., Serrano, L., Bycroft, M. and Fersht, A.R. (1990)
Transient folding intermediates characterized by protein engineering. Nature , 346,
440–445.
McGregor, M.J., Flores, T.P. and Sternberg, M.J.E. (1989) Prediction of -turns in
proteins using a neural network. Protein Engineering , 2, 521–526.
Mitchinson, C. and Wells, J.A. (1989) Protein engineering of disulfide bonds in subtilisin
BPN. Biochemistry , 28, 4807–4815.
Mutter, M. and Vuileumier, S. (1989) A Chemical Approach to protein Design-Template-
Assembled Synthetic Proteins (TASP). Angewandte Chemie , 28, 535–676.
Narhi, L.O., Stabinsky, Y., Levitt, M., Miller, L., Sachdev, R., Finley, S. et al. (1991)
Enhanced stability of subtilisin by three point mutations. Biotechnol. Appl. Biochem. ,
13, 12–24.
Nei, M. and Koehn, R.K. (1983) Evolution of Genes and Proteins. Sunderland,
Massachusetts: Sinauer Associates.
Oxender, D.L. and Fox, C.F. (eds.) (1988) Protein Engineering . New York: Alan R.Liss,
Inc.
Pastora, A. and Lesk, A.M. (1991) Brave new Proteins: What evolution reveals about
proteins. Current Opinion in Biotechnology , 2, 592–599.
Perutz, M.F. (1978) Electrostatic effects in Proteins. Science , 201, 1187–1191.
Petersen, S.B., Bohr, H., BohrJ., Brunak, S., Cotterill, R.M.J., Fredholm, H. et al. (1990)
Training Neural Networks to analyze biological sequences. Trends in Biotechnology ,
8, 304–308.
Petersen, S.B. (1991) PDBASE: Poor Man’s Structural Protein Database. In Site-Directed
Mutagenesis and Protein Engineering , edited by M.R. El-Gewely. Amsterdam:
Elsevier.
Petersen, S.B., Jonson, P.H., Fojan, P., Petersen, E.I., Petersen, M.T.N., Hansen, S. et al.
(1998) Protein Engineering the surfaces of proteins. J. Biotech. , 66, 11–26.
Prevelige, P. and Fasman, G.D. (1989) Chou-Fasman Prediction: of the Secondary
Structure of Proteins. In Prediction of Protein Structure and the Principles of Protein
Conformation , edited by G.D.Fasman. New York: Plenum Press.
Qian, N. and Sejnowski, T.J. (1988) Predicting the secondary structure of globular
proteins using neural network models. J. Mol. Biol. , 202, 865–884.
Regan, L. (1991) Protein Design. Current Opinion in Biotechnology , 2, 544–550.
Riisgard, S. (1990) in Proc. 5th Eur. Congr Biotech. , edited by C.Christiansen, L.Munck
and J.Villadsen. Copenhagen: Munksgard.
Rogers, N.K. (1989) The Role of Electrostatic interactions in the Structure of Glubular
 Applied biocatalysis 310
Proteins in Prediction of Protein Structure and the Principles of Protein Conformation ,
edited by G.D.Fasman. New York : Plenum Press.
Rogers, N.K. (1986) The Modeling of Electrostatic Interactions in the Function of
Globular Proteins. Prog. Biophys. Mol Biol , 48, 37–66.
Russell, A.J., Thomas, P.G. and Fersht, A.R. (1987) Electrostatic effects on modification
of charged groups in the active site cleft of subtilisin by protein engineering. J. Mol
Biol. , 193, 803–813.
Russel, S.T. and Warshel, A. (1985) Calculations of Electrostatic Energies in Proteins. J.
Mol. Biol. , 185, 389–404.
Sali, D., Bycroft, M. and Fersht, A.R. (1988) Stabilization of protein structure by of
interaction of -helix dipole with a charged side chain. Nature , 335, 740–743.
Sander, C. (1991) De novo Design of Proteins. Current Opinion in Structural Biology ,
1 , 630–637.
Sasaki, T. and Kaiser, E.T. (1989) Helichrome: Synthesis and enzymatic activity of a
designed Hemeprotein. J. Am. Chem. Soc. , 111, 380–381.
Schultz, G.E. and Shirmer, R.H. (1978) Principles of Protein Structure. New York,
Berlin, Heidelberg, Tokyo: Springer Verlag.
Serrano, L. and Fersht, A.R. (1989) Capping and -helix stability. Nature , 342, 296–
299.
Siezen, R.J., de Vos, W.M., Leunissen, J.A.M. and Dijkstra, B.W. (1991) Homology
modelling and Protein Engineering strategy of subtilases, the family of subtilisinlike
serine proteases. Protein Engineering , 4, 719–737.
Sternberg, M.J.E., Hayes, F.R.F., Russell, A.J., Thomas, P.G. and Fersht, A.R. (1987)
Prediction of electrostic effects of engineering of protein charges. Nature , 330, 86–88.
Storer, A.C. (1991) Engineering of proteases and protease inhibition. Current Opinion in
Biotechnology , 2, 606–613.
Takagi, H., Morinaga, Y, Ikemura, H. and Inouye, M. (1988) Mutant subtilisin E with
enhanced protease activity obtained by site-directed mutagenesis. J. Biol. Chem. , 265
(36), 19592–19596.
Takagi, H., Takahashi, T., Momose, H., Inouye, M., Maeda, Y. and Matsuzawa, H.
(1990) Enhancement of the thermostability of subtilisin E by introduction of a disulfide
bond engineered on the basis of structural comparison with a thermophilic serine
protease. J. Biol. Chem. , 256(12), 6874–6878.
Tainer, J.A., Roberts, V.A. and Getzoff, E.D. (1991) Metal binding sites in Proteins.
Current Opinion in Biotechnology , 2, 582–591.
Tanford, C. and Kirkwood, J.G. (1957) Theory of Protein Titration Curves. J. Am. Chem.
Soc. , 79, 5333–5347.
Warshel, A. and Russell, S.T. (1984) Calculations of Electrostatic Interactions in
Biological Systems and in Solutions. Q. Rev. Biophys. , 17, 283–422.
Wells, J.A. and Estell, D.A. (1988) Subtilisin: an enzyme designed to be engineered.
TIBS , 13, 291–297.
Wells, J.A., Powers, D.B., Bott, R.R., Katz, B.A., Ultsch, M.H., Kossiakof, A.A. et al.
(1987) Protein Engineering of Subtilisin. In Protein Engineering , edited by Oxender
and Fox. Alan R.Liss.
Wong, C.-H., Chen, S.-T., Hennen, W.J., Hennen, W.J., Bibbs, J.A., Yang, Y.-F. et al.
(1990) Enzymes in organic synthesis: use of subtilisin and a highly stable mutant
derived from multiple site-specific mutations. J. Am. Chem. Soc , 112, 945–953.
Wutrich, K. (1986) NMR of Proteins and Nucleic Adds . New York: Wiley.
 8.
BIOCATALYST PERFORMANCE
ANTONIO BALLESTEROS1 AND LASZLO BOROSS2
1 Departamento de Biocatalisis, Instituto de Catalisis, CSIC, 28049 Madrid,
Spain
Email: [email protected]
2 Department of Chemistry and Biochemistry, University for Horticulture and
Food Industry, Budapest, Hungary
Email: [email protected]

ABSTRACT
In the case of most enzymic transformations the reaction rate can be described
as a hyperbolic function of the concentration of substrate; the characteristic
parameters of these hyperboles are the Vmax and the KM values, which can be
determined easily by different linearized plots. Different factors such as
temperature, pH, chemical modification of the functional groups in the side
chains of the protein, reversible inhibitors, activators, allosteric effectors,
influence the catalytic activity of the enzymes.
Since the protein scaffold is commonly not very stable, many methods have
been used for stabilization: presence of additives, immobilization by multiple-
point attachment, stabilization by chemical or biochemical modification and by
protein engineering, and several others.

8.1 ENZYME KINETICS AND MECHANISMS

8.1.1 Introduction
For application of a biocatalyst we must know its basic properties, the substrate
specificity and the kinetic characteristics. The substrate specificity is a relatively
uncomplicated topic, it can be determined with simple experiments, and for the most
important enzymes many data are available. Determination of the kinetic properties of an
enzyme is a more complex problem. A detailed description of an enzymic catalysis
requires extensive data about the structure of the whole protein molecule, the structure of
 Applied biocatalysis 312
its active centre, the mechanism of the reaction, the rate constants of the individual steps
of the catalytic process, the stability of the active conformation, the action of stabilizers,
activators, inhibitors etc. There are many excellent books available about enzymes, their
structure, mechanism, kinetics, stability etc. However, for practical application of an
enzyme much less information is needed.
The knowledge of fundamental enzymology is always advantageous for everyday’s
work with enzymes, and especially useful for development of new enzymic technologies.
From this point of view the most important data about an enzyme are the basic kinetic
characteristics and knowledge about the factors which influence these values. We want to
summarize the general features of these characteristics in this chapter.

Table 8.1 Meaning of the KM constant in the Michaelis-Menten equation

for some simple enzyme mechanism.

Reaction mechanism Meaning of K M Remarks

(Formation of kinetically stable EP complex)

8.1.2 Characterization of the Catalytic Properties of Enzymes for Their Use

in Industrial and Analytical Processes
Two characteristics, the Michaelis constant KM and the maximal velocity Vmax are the
most important numeric data. The well-known Michaelis-Menten equation describes the
relationship between the initial reaction rate and the substrate concentration with these
two constants. The actual form of the rate equation of an enzymic process depends on the
chemical mechanism of the enzymic transformation of the substrate to product (Table
8.1).
In all cases an enzymic process is composed of several consecutive reaction steps.
Even the simplest Michaelis-Menten type “rapid equilibrium mechanism” involves two
steps, the binding of the substrate, S, to a specific site in the active centre, and the
chemical transformation of the bound S to product P, during which the enzyme becomes
free again. The Michaelis constant characterizes the affinity of the enzyme to its
 Biocatalyst performance 313
substrate, i.e. the equilibrium (dissociation constant) of the reversible ES complex
formation from the free enzyme (E) and substrate (S). The rate of the chemical
transformation of the substrate to the product under optimal conditions is characterized
by the value of the Vmax. When S is present in high concentration the enzyme is saturated
with S and the enzyme exists in ES complex form. The first order rate constant, kcat,
characterizes the rate of the formation of P from ES in this monomolecular reaction. In
the rapid equilibrium mechanism k2=kcat, and this step is the rate limiting step in the
overall enzymic process. When the concentration of the substrate is low, the enzyme is
not saturated and the overall rate of the enzymic reaction, the rate of the formation of P
from S, could be described by the equation:

(8.1)

where kcat/KM corresponds to an apparent second order rate constant of the simplified
bimolecular reaction:

kcat/KM is also called the catalytic efficiency of the enzyme. In the rapid equilibrium
mechanism k2 is much less than k−1, and KM is practically equal to the dissociation
constant Ks of the ES complex. Therefore k2/KM=k2 k1/k−1=k2/Ks. This is the lower limit
of the efficiency. When k2 is much higher than k−1 k2/KM= k2k1/k2=k1, which is the
upper limit of the efficiency, and in optimal cases it approximates the rate of a diffusion
controlled reaction, 108−109 M−1s−1.
In the formation of the ES complex noncovalent bonds (H-bonds, ionic bonds, van der
Waals bonds, hydrophobic interactions) bind the substrate to the enzyme. Experimentally
determined rate constants of the formation of various ES complexes are less than the
diffusion rate 107–109 M−1s−1 in most cases (Hammes, 1992; Fersht, 1985). This
indicates that in addition to the diffusion some other chemical changes occur, requiring
little energy, such as desolvation or some minor alteration of the conformation of the
enzyme. The change in conformation of the enzyme upon the ES complex formation has
been shown for many enzymes (induced fit theory). These reversible conformational
changes could happen in small consecutive steps, each of them requiring little activation
energy. The enzyme protein molecule undergoes continual fast, small changes of
conformation (“dynamics of protein structure” or “conformational motility” (Somogyi,
Welch and Damjanovich, 1984; Rosenberg and Somogyi, 1986). Certain changes are
more frequent than others. In some cases the apparent rate constant of the conformational
changes can be characterized (Vas and Boross, 1974). From the point of view of catalysis
only some conformations of the enzyme molecules are active (“conformational fit”,
Straub & Szabolcsi, 1964).
 Applied biocatalysis 314
The formation and the dissociation of EP complex into the free product and free
enzyme molecules, as E+S ES EP E+P, are similar to those of the ES complex.
For many enzymes of the living cells, the forward and the backward reactions are equally
important.
In enzymic reactions the central ES EP transformation is very fast, and the value of
kcat is very high. In addition to correctly oriented binding of the substrate at the active
center of the enzyme, an effective decrease in activation energy of this reaction step
might also be provided by stabilization of the transition state of the substrate molecule in
the ES complex.
Some energy diagram models of simple enzymic reactions are shown in Figure 8.1. A
schematic model for an advantageous binding of the substrate on the enzyme active
center is illustrated in Figure 8.2.
From a thermodynamic point of view the most important reaction characteristic for
practical application is its free enthaply change G°. According to the fundamental
equation G°=−RTlnK, the equilibrium constant of the reaction is determined by G°. A
high negative value (−20 kJ/mol or even less) usually implies that the reaction results in
high yield and quantitative transformation of substrate to product.
 Biocatalyst performance 315

Figure 8.1 Model energy diagrams for non-enzymic reactions (A), enzymic
reaction following the rapid equilibrium mechanism (see Table 8.1)
(B) and enzymic reaction following Briggs-Haldane kinetics (C). “E”
represents the activation energy of transition and the positive and
 Applied biocatalysis 316

negative indices refer to forward and reverse reactions. The ‡ indicates the
transition states. The standard free enthalpy change of the whole
reaction is independent from the reaction mechanism.

Figure 8.2 A simple schematic model illustrating binding of a hypothetic

substrate in stretched conformation and two different conformations
of the active site of the free enzyme.

However, when the value of G° is close to zero, or even positive, the reaction reaches a
nonfavorable equilibrium and the reverse reaction is favoured.
We should note that in enzymology the G free enthalpy change is used, because the
change in volume during the reactions is negligible. Therefore U is equal with H (
H= U+P V, and G= H−T S). Some authors use the “free energy” change F =
U−T S expression.
Many enzymic reactions have high negative G° values, for example glycosyl or
peptide bond hydrolysis reactions in aqueous media, oxidations with oxygen as substrate
etc. Some thermodynamic data of industrially applied enzymic reactions are described by
Bruns and Schulze (1962), Tewari (1990) and Biselli, Kragl and Wandrey (1995).
For a better understanding of the enzyme catalysis in nature, experimental and
theoretical studies characterize the free energy profiles and catalytic efficiencies of
enzymes under different conditions, which may define the performance of an enzyme in
maintaining a constant flux or a constant pool concentration of the product, working
under irreversible or reversible conditions etc. (Albery and Knowles, 1976; Stackhouse et
al., 1985; Pettersson, 1992; Somogyi, Welch and Damjanovich, 1984). Only a few
enzyme reactions have been analyzed in detail and further experimental investigations are
necessary to characterize the enzymes, to draw general conclusions, and to deduce how
much their evolution approximated the requirements for “optimal catalysis”.
Since the free enthalpy change of a reaction does not depend on a specific catalyst, it is
 Biocatalyst performance 317
not possible to change its value by the use of another enzyme. Although the rate of the
product formation could be faster if a “better” enzyme is used, the rate of the reverse
reaction should be accelerated to the same extent. However, alteration of the reaction
environment could change the value of the free enthalpy change. Such factors are the
temperature, pH, ionic strength and, in some cases, the presence and the concentration of
various ions.
Many enzymes, which transform two different substrates to one or two product(s),
could be characterized using equation (8.1), if the concentration of one substrate is high
enough to saturate the enzyme. If the two substrate molecules bind to the enzyme
independently from each other, the calculated KM values will reflect the affinity of the
substrate to the complex of the other substrate molecule and the enzyme. Further, the
Vmax will characterize the rate of the reaction at the excess concentrations of both
substrates (the enzyme is saturated by both substrates). However, this could be just a
coarse approximation, and there are kinetic analytical methods for a more exact
characterization of such two-substrate enzymic reactions, which could run on different
ways e.g. random Bi-Bi, ping-pong Bi Bi mechanisms (Keleti, 1986; Fersht, 1985; Segel,
1975; Cornish-Bowden, 1995).
Theoretically, all chemical reactions are reversible and, in the presence of enzymes, a
dynamic equilibrium state will be reached rather than a total transformation of substrate
molecules to product molecules. The dynamic equilibrium could be characterized by the
Haldane equation:

(8.2)

in which Vf and Vp are the maximum velocities in the forward and reverse directions,
respectively, and Kp and Ks the dissociation constants of the EP and ES complexes
(Keleti, 1986; Fersht, 1985).
If the concentration of one substrate is too small, the equilibrium will be non-
favourable, the reaction stops after formation of few product molecules. This is
advantageous for the reverse reactions. In this way, synthesis is possible with hydrolytic
enzymes in organic solvents (see Chapter 9).

8.1.3 Factors Affecting Enzyme Activity

a) Effect of pH

In general, enzyme active sites contain various acidic or basic amino acid residues.
Interaction between them and their simultaneous interactions on the substrate influence
the catalytic process (Segel, 1975; Fersht, 1985; Keleti, 1986).
 Applied biocatalysis 318
A good example of this interaction in catalysis is the hydrolysis of the bacterial cell
wall polysaccharide by lysozyme. This enzyme contains two carboxylic groups at its
active site and, in active enzyme one must be in dissociated—COO−, the other in the
undissociated—COOH form. Therefore, the pK’s of the two carboxylic groups are
different. This difference in dissociation constant is a consequence of the neighbouring
amino acid residues and of the interactions between the functional groups in the
microenvironment.
A similar situation exists in the acidic, aspartic-type proteases, where two acidic amino
acid residues must interact to split the peptide bond in the substrate. These carboxylic
groups also must dissociate differently, and therefore their pK values must be different.
The study of the pH dependence of the enzyme activity, and, in particular the pH
dependence of the Vmax (kcat) and KM values of the enzyme provides important
information about the structure—function relationship of the enzyme. The most
important features of the pH dependences of the enzymic reactions are summarized
below.
Plots of the activity of various enzymes as a function of pH give different curves, in
some cases bell-shaped. The pH dependence of the activity could be the result of three
fundamental effects of H+ ion concentration:

(i) The overall conformation of a protein molecule is affected by the acidity or alkalinity
of the solution, because the interaction of the conformation-determining groups
depends upon their ionization state. This pH dependence is very complex and there is
no general equation for its description. Fortunately, large, inactivating conformational
changes occur usually in pH regions far from the pH optimum of the activity.
Therefore, enzyme inactivation through pH-mediated conformational changes usually
does not cause difficulties in the industrial application of enzymes.
(ii) In many cases the substrate binding site contains basic or acidic groups. The charge
of such groups can be essential for ES complex formation. Consider a general case in
which the substrate binding site contains two acidic groups: deprotonation of one site
is a prerequisite for substrate binding, deprotonation of the other group results in
abolition or reduction of the complex forming ability. Similarly, substrate molecules
can also have basic, and/or acidic groups, whose ionic dissociation can influence
binding to the enzyme. As a generalisation four dissociating groups (two on the
enzyme, two on the substrate molecule) might be found which determine
fundamentally the pH dependence of ES complex formation. (The acidic dissociation
of these groups in the ES complex form are characterized with pKES1 and pKES2).
(iii) pH can affect the Vmax value of the enzyme: dissociation of some groups could be a
prerequisite for the action of the catalytic site, and the dissociation of other groups
could abolish or decrease this activity. In a general treatment of this pH dependence
one can simplify the problem to the dissociation of a dibasic acid again, in which the
deprotonation of the first group increases the activity and that of the second decreases
the activity.

Analysis of the pH dependencies of KM and Vmax give important information about those
 Biocatalyst performance 319
groups which are in the active center and which participate in substrate binding or in
catalytic action.
It must be emphasized that the effect of pH refers to the effect of the hydrogen ion
concentration around the enzyme molecule, i.e. in the microenvironment. In the case of
immobilized enzymes, the pH in the bulk solution and in the microenvironment of the
enzyme, could be different: the “partitioning effect” and “diffusion limitation effects”
promoted by the solid support could change the H+ concentration around the
insolubilized enzyme molecule. Further, the covalent chemical modification of acidic or
basic functional groups in the enzyme molecule outside of its active center could also
change the pH dependence (by affecting the H+ concentration in the close vicinity of the
active center). The covalent immobilization of enzymes is performed frequently by
binding through lysine amino groups, thus changing their pK values and therefore,
possibly altering the pH dependence of the activity of such enzymes, where no diffusion
limitation effects or partitioning of H+ ions occur. An excellent summary of various
effects of immobilization on the enzymes was written by Trevan (1980).
The microenvironment inside the protein molecule, created by neighbouring groups at
the active center, can influence the pK value of another group. For example, in pepsin
Asp 32 is hydrogen-bonded to Ser 35, and Asp 215 is H-bonded to Thr 218. This
suggested that COOH…. OH hydrogen bond formation over the three residues in the
primary sequence might be advantageous. On this basis Ido et al. (1991) performed
interesting experiments to change the pH dependence of the activity of HIV-1 (Human
immunodeficiency virus type 1) protease. This is an aspartic type protease but it has an
unusual pH dependence: its activity range is pH 4–6, instead of pH 2–6, which is normal
for aspartic type proteases. The authors changed the Ala 28 residue to Ser 28, so
introducing an OH group close to the Asp 25 in the active site. This A28S mutant
protease exhibited a decreased pKE2 as compared with that of the wild type enzyme, and
further, both pKES1, and pKES2 decreased significantly. This indicates that the
replacement of Ala 28 by Ser 28 resulted in a new H-bond formation with Asp 25,
increasing its acidity.
A desirable change in the pH dependent activity can also be demonstrated in simple
experiments. Penicillin acylase of E. coli catalyses the synthesis of ampicillin and
cephalexin by condensation of -aminophenylacetic acid with 6-aminopenicillanic acid
and 7-aminodeacetoxycephalosporanic acid respectively, although more slowly in the
later case. Forney and Wong (1989) deduced that the protonated -amino group in the
aminophenylacetyl moieties inhibits binding to the enzyme. They wanted to determine
whether it was possible to alter the substrate specificity of penicillin acylase and to select
enzymes that efficiently hydrolyse substrates with -aminophenylacetyl moiety at low
pH. The authors used -aminophenylacetyl-leucine (APAL) as a substrate analog of
ampicillin and cephalexin. They cloned the gene for acylase of E. coli ATCC 11105,
transfered it to a leucine-auxotroph E. coli and selected mutants which could cleave
APAL, thus providing leucine for growth in a low pH medium. The purified mutant
enzyme exhibited 10-fold greater second order rate constant kcat/KM for APAL
hydrolysis. This experiment suggests that proper selection might result in advantageous
change of the original pH dependence of an enzyme and in isolation of appropriate
 Applied biocatalysis 320
enzymes for desired reactions.
An interesting observation is that an enzyme may exhibit different pH activity profiles
for various neutral substrates. The explanation of this is that the enzyme binds or
transforms such various substrates differently. For example, Taka amylase has different
pH optima for long chain amyloses and for low molecular mass substrates. Some specific
chemical modifications of the side chains of the enzyme may also alter the pH activity
profiles. Kobayashi, Miura and Ichisima (1992) modified the lysine amino groups using
bifunctional reagent o-phtalaldehyde, and observed a pronounced shift in the pH-
dependence of oligomaltoside hydrolysis.

b) Effect of temperature

The rate of a chemical reaction depends on temperature, and this dependence can be
described by the empirical Arrhenius equation:

(8.3)

where k is the rate constant, Ea is the activation energy of the reaction and A is the
“action constant”.
In the case of enzymes, the rate of the catalysed reaction increases regularly with
increasing temperature. However, the probability of the unfolding of the
threedimensional conformation of the protein molecule also increases, as there is more
energy available to split the non-covalent interactions between side chains. In some cases
it has been demonstrated that such noncovalent interactions play a dominant role in the
stability of the native conformation. For example, Brosnan, Kelly and Fogarty (1992)
demonstrated that the irreversible thermoinactivation of -amylase of Bacillus
stearothermophilus at 90°C is related to changes of the hydrophobic interactions in the
molecule.
Disruption of side chain interactions results in denaturation of the protein, and the rate
of inactivation follows first order kinetics in the simplest cases. The plot of the logarithm
of the remaining activity (In at) versus time gives a straight line, the slope of which is the
negative value of the inactivation rate constant.
The heat inactivation of many enzymes, and in particular of those which are composed
of subunits, is not a simple process, and the In at versus t plots are more complex.
Usually the oligomeric forms of enzymes are slower inactivating than their monomeric
forms (Szajani, Ivory and Boross, 1980). Similarly, specific interactions between
different protein molecules could result in higher heat resistance.
At low temperature (T1 and T2 in Figure 8.3) the rate of heat inactivation is slow as
compared to the rate of the catalysed reaction. At elevated temperature (T3, T4) the
increased heat-inactivation rate results in a faster decrease in the number of the active
catalyst molecules. As a consequence of this the rate of the enzymic reaction becomes
 Biocatalyst performance 321
slower and slower with time. The mean value of [P] / t, i.e. the apparent rate of the
reaction, is not constant but becomes a decreasing function of t. Further increase in
temperature (T5 , T6) results in even faster catalysis at the beginning, but also faster
inactivation, terminating the reaction sooner. The mean value of product formation rate
decreases rapidly (Figure 8.3).
The sum of the two effects of the increase in temperature, i.e. the combination of the
increase of the catalysis rate and the increase of the rate of inactivation, gives

Figure 8.3 The effect of temperature of an enzyme reaction and the effect of
the time-period of the activity measurements on the apparent
temperature optimum (after Wiseman, 1975). The index numbers
indicate the increase of temperature. It is important to note that in all
cases the decrease of the rate of product formation is the consequence
of partial inactivations only, i.e. the concentration of substrate must
be enough to saturate the enzyme even at time t2.

a curve which passes through an optimum when the measured rate of the enzymic
reaction is plotted as a function of the temperature. It is easy to understand from the
explanation above, that the shape of the curve depends on the length of the time of the
measurement. The shorter the periods of measurements the higher the measured mean
activities and the higher therefore the apparent temperature optimum. Consequently, the
“temperature optimum” is not an exact characteristic property of an enzyme. However,
from the point of view of the practical application of the enzyme, the reaction time
chosen for the enzymic treatment determines an optimum temperature, at which the yield
of the product is maximum. Therefore, the analysis of the temperature dependence of the
enzymic reaction gives very practical information to the engineer in the choice of the best
 Applied biocatalysis 322
conditions for the biocatalytic operation.

c) Covalent chemical modifications

The functional groups in a protein have chemical reactivities somewhat different to those
of the same groups in a small peptide. The reasons for this are fundamentally the
differences in the various side-chain interactions and steric hindrances. Some reagents
react specifically with one type of functional group such as—SH reagents, lysine amino
or arginine guanidino group modifying chemicals (“specific chemical modification”). In
other cases the chemical reactivities of the same type of groups (e.g.—OH or—SH
groups) in the protein molecule are very different and only the extremely reactive group
will be modified (“selective chemical modification”). Examples for such cases are the
“—OH proteases” and other “serine enzymes”, in which specific sidechain-interactions
result in a specially high nucleophilic reactivity of the—OH group in the active centre.
The other—OH groups in the same enzyme react much slower with the selective
reagents, like diisopropylfluorophosphate (DIFP), phenylmethylsulfonylfluoride (PMSF),
etc. Selective reagents like PMSF could be used for inactivation of serine proteases in
extracts to protect other enzymes against proteolytic damages.
The chemical reactivity of the functional groups in a protein can be characterized by
measuring their rate of modification, identifying groups of high reactivity, surface
exposed groups of regular reactivity, temporarily surface exposed groups (Vas and
Boross, 1974), sterically hindered surface exposed groups of decreased reactivity, or
buried functional groups. Such experiments help to better understand the chemical nature
of some parts of the protein molecule.
Specific chemical modifications produce changes in the properties of the enzyme.
When single modification of one group results in inactivation but does not change the
conformation of the enzyme molecule, this indicates that the group is essential for the
catalytic activity. In other cases, modification of the side chains in the enzymes may
cause specific, and sometimes drastic changes in the catalytic properties and in thermal
stability.
Many recent experiments demonstrate such effects. Morand and Biellmann (1991)
modified the -amylase of B. licheniformis with a polyaldehyde derived from -
cyclodextrin, and reduced the Schiff bases formed on amino groups with a suitable
reagent. The number of the reactive lysine amino groups decreased from 8 to 3.5 per mol
enzyme. The catalytic activity of amylase decreased with increasing number of the
modified amino groups to about 75% of the untreated enzyme. Thermal stability of the
enzyme increased, the half life values at 80°C being 4.7, 5.6, 6.2 and 7.0 minutes for the
native enzyme and for enzymes containing 1, 2 and 4.5 modified lysine groups,
respectively.
Selective chemical change of the serine—OH group to cysteine—SH in enzymes can
be performed with extremely reactive serine residues in the active sites by the use of
phenylmethylsulfonyl fluoride and, subsequently, thioacetic acid (Polgar and Bender,
1966). This selective chemical modification demonstrates the essential role of an—OH
 Biocatalyst performance 323
group in the catalytic process. Recently Slade et al. (1991) showed that in E. coli
penicillin acylase, Ser 290 could be converted to Cys, resulting in complete enzyme
inactivation, suggesting the fundamental, nucleophilic attacking role of this group
towards the substrate, as in other serine type peptide hydrolysing enzymes. Similar
experiments were also performed with penicillin acylase of K. citrophila (Martin et al.,
1991).
The essential—SH group in D-glyceraldehyde-3-phosphate dehydrogenase and the
imidazol residues of the ribonuclease are also more reactive because of side-chain
interactions in the active center. Such functional groups may have such extremely high
reactivity that an equivalent amount of the reagent causes full inactivation of the enzyme.
Applied biocatalysis 324
 Biocatalyst performance 325

Figure 8.4 The Lineweaver-Burk plot (A) and the Hanes plot (B) of typical
enzyme kinetics in presence of a competitve (a) noncompetive (b),
mixed type (c) and uncompetitive (d) inhibitor.

8.1.4 Effects of Reversible Binding Inhibitors

Inhibitors are molecules which decrease the activity of the enzymes and can be divided in
two classes: reversible binding inhibitors and irreversible binding inhibitors.
The reversible binding inhibitors form enzyme-inhibitor complexes in a reversible
manner. Competitive inhibitors bind to the substrate binding site of the enzyme,
competing with the substrate molecule for the same binding site of the protein. The
presence of high substrate concentrations supresses the binding of the inhibitors and vice
versa. Therefore the Vmax value of the enzyme remains unchanged but the KM value
increases. The Lineweaver-Burk plot in such cases gives straight lines which intersect at
the same point on the ordinate axis and the slopes of which are greater at higher
concentration of the inhibitor (Figure 8.4), according to the function:

(8.4)

where Ks and Ki are the dissociation constants of the enzyme-substrate and of the
enzyme-inhibitor complexes, respectively.
Non-competitive inhibitors do not disturb the binding of the substrate but upon binding
they inactivate the catalytic site. The Lineweaver-Burk plots for this type of inhibition
give straight lines, crossing each other at one point on the abscissa, and the slope of
which is higher with increasing concentration of the inhibitor, according to the function:

(8.5)

Accordingly, non-competitive inhibitors do not influence the Ks (or KM) value of the
enzyme but decrease the Vmax.
The mixed-type inhibitors combine the effects of the competitive and noncompetitive
inhibitors: binding at the active center decreases the affinity of the enzyme towards the
substrate molecule and also decreases the rate of transformation of the bound substrate. In
their presence, the straight line plots intersect in the fourth quarter of the Lineweaver-
Burk plot, according to equation:
 Applied biocatalysis 326

(8.6)

where Ki and are the binding constants of the inhibitor to the free enzyme and to the
ES complex, respectively.
Uncompetitive inhibitors can bind to the enzyme-substrate complex only, but not to the
free enzyme molecule. The Lineweaver-Burk plots in such cases give parallel straight
lines for activity-substrate concentration profiles, measured at different concentrations of
the inhibitor (Figure 8.4), according to equation:

(8.7)

Some reversible inhibitors bind to the enzyme at a site different from the active site but
cause apparently competitive, noncompetitive, mixed-type or incompetitive inhibitions.

8.1.5 Activators and Regulators

The activity of key-enzymes of metabolism are necessarily regulated. One type of
regulation is carried out by reversibly binding allosteric regulators (activators or
inhibitors), molecules which bind reversibly to a “regulatory binding site” outside of the
active center. Their binding changes the conformation of the protein, which could result
in increased or decreased catalytic activity. Allosteric enzymes are always composed of
several subunits. In the case of “concerted mechanisms” the enzyme molecules in their
free state can exist in two conformations: one (R, “relaxed”) is catalytically active and
can bind the substrate and also the activator; the other, (T, “tensed”) is inactive and can
bind only the inhibitor (Monod, Wyman and Changeaux, 1965). The conformation of all
subunits of the enzyme is the same, either R or T. The binding of an allosteric regulator
on one subunit freezes the conformation not only of this subunit but also that of the others
in the same enzyme molecule. In allosteric enzymes, the substrates themselves are
allosteric regulators and their regulatory effects are called homotropic interactions. When
the dependence of the catalytic activity of these enzymes as a function of the substrate
concentration is described by a sigmoidal curve, it suggests the existence of a positive
cooperativity. Negative cooperativity causes a deformation of the hyperbolic activity-
substrate concentration profile.
In some cases simple inorganic ions can also affect the conformation of the enzyme
molecules, thus functioning as regulators.
 Biocatalyst performance 327

8.2 STABILITY AND STABILIZATION OF BIOCATALYSTS

Industrial exploitation of biocatalysts requires that they be rendered more stable to

withstand harsh operating conditions (high temperatures and other conditions that favour
the loss of catalytic activity), hence increasing biocatalyst life. The following sections
discuss the mechanisms of enzyme stability and stabilization.

8.2.1 Inactivation of Enzymes. Stability of Enzymes

The structure and function of enzymes is determined by both the amino acid sequence
and the surrounding solvent. The overall stability of proteins is characterized by a subtle
balance of intra- and inter-molecular interactions. The basic principle of the structure
(and of the stability) of the proteins is related to the nature of its normal environment: for
(water) soluble globular proteins this is the minimization of the hydrophobic surface area,
whereas the contrary is the case for membrane proteins (Jaenicke, 1991).
Stability of an enzyme is defined as its ability to retain catalytic activity under
specified conditions (Martinek and Berezin, 1978). The quantitative criterion for stability
is the value of the first-order rate constant of monomolecular inactivation, king kin.

(8.8)

The decrease in kin is an indication of stabilization.

According to Tanford (1968), protein denaturation is a process involving a major or
minor change of its three-dimensional native structure, without altering the amino acid
sequence. A change (or unfolding) of the structure of an enzyme impairs the correct
arrangement of the active site and, hence, results in enzyme inactivation. Therefore,
stabilization means preventing this change and preserving the native structure of the
protein (Gianfreda and Scarfi, 1991).
When an enzyme looses activity it is due to partial unfolding of the polypeptide chain.
The inactivating agent disrupts the delicate balance of noncovalent bonds. The difference
between the free energies of the folded and unfolded states is small, e.g. 5–20 kcal/mol.
Initially this unfolding is reversible, as long as no covalent changes in the primary
structure occur. However, if the agent maintains its effect, irreversible covalent changes
take place. This can be described by the well known scheme of Lumry and Eyring (1954)

(8.9)
 Applied biocatalysis 328
where N is the native, catalytically active enzyme; U, the unfolded, denatured protein;
and I, the irreversibly inactivated enzyme. Enhanced stability will mean a shift in the
conformational equilibrium toward the native conformation. Such a shift will increase the
lifetime of the native conformation by decreasing the rate of denaturation, increasing the
rate of renaturation, or both (Stellwagen, 1984).
When comparing the stability of mesophilic and thermophilic enzymes, many factors
that enhance stability have been discussed (Mozhaev and Martinek, 1984): Electrostatic
interactions, hydrogen bonds, intramolecular disulfide bonds, hydrophobic interactions,
protein-protein contacts etc. As a generalization, enhanced stability is the result of many
concomitant stabilizing factors and stabilization is achieved through an increase in
internal and a decrease in external hydrophobicity. Tomazic and Klibanov (1988),
studying several amylases, found that deamidation of Asn and/or Gln residues and Cys
oxidation were the major processes taking place in irreversible thermal inactivation.
In the majority of papers, the kin of enzyme at one temperature is assumed to be a
measure of thermostability. Moreover, the dependence of log kin on 1/T (Arrhenius plot)
is only valid over a narrow temperature range. In general, when temperature dependence
changes, one can observe (Ugarova, Rozhkova and Berezin, 1978):

i) an increase in Ea (the activation energy of thermoinactivation) with temperature.

Stabilization will increase with a decrease in temperature (“low-temperature
stabilization”). The modified enzyme will be more stable than the native one at
temperatures lower than the intercept point of both Arrhenius lines.
ii) A decrease in Ea will lead to a “high-temperature stabilization”.

Sadana and collaborators (Sadana, Raju and Shahin, 1989) have proposed an empirical
stability index (SI) for enzyme deactivation, which makes more quantitative the effect of
different variables on enzyme stability.
For single-step inactivations,

(8.10)

where a1 is the ratio of the specific activities of the inactivated state to the native one.

8.2.2 Stabilization of Enzymes

The ideal approach to the stabilization of an enzyme would be to identify the mechanism
of its inactivation and then design a procedure which would prevent the said mechanism
(Klibanov, 1983). Many different methods have been proposed for rendering the protein
stable under the conditions of the catalytic process and these are reviewed below.
 Biocatalyst performance 329

a) Stabilization by additives

Additives are soluble species that do not interact covalently with the protein.
In general, additives may be classified (Gray, 1988) into the following groups:

i) Substrates and other similar ligands. If the species binds to the native, active form of
the protein, then the equilibrium folded-unfolded will be affected, resulting in
preservation of enzyme activity. Other ligands (products, effectors, inhibitors) can also
stabilize in the same way.
ii) Organic molecules of low molecular weight (cosolvents). Two classes of cosolvents
may be distinguished. Polyols of three or more carbons increase the stability of the
protein, owing to the preferential exclusion of the additive from the protein vicinity.
On the other hand, more hydrophobic cosolvents usually decrease the stability of the
threedimensional form of the protein; the less hydrophilic the solvent the greater the
effect on Tm (temperature at which half the protein is unfolded).
iii) Ionic species, a) Metal ions. In many thermostable enzymes from thermophiles, the
binding of Ca+2 is an important stabilizing factor, due to the formation of a multipoint
binding to amino acid residues far apart, which results in freezing the protein structure
in the most stable active form, b) Non-specific additives. These act due to preferential
exclusion of the additive from the immediate environment of the protein.
iv) Polymers. Stabilization of proteins by polysaccharides, proteins and synthetic
molecules has been reviewed (Schmid, 1979). The stabilizing properties of
polyethylene glycol increase with concentration and chain length (Monsan and
Combes, 1984). Since the polymers are very hydrophilic it is reasonable to assume that
they act by preferential hydration of the protein.

b) Stabilization by immobilization

Melrose (1971) reviewed the instances in which direct comparisons of the stabilities of
native and immobilized enzyme pairs were reported: In 30 cases the insoluble enzyme
was more stable, in 8 the soluble, whilst there was little difference in the remaining 12
cases. Immobilization does not lead necessarily to stabilization; ho ever, this is mainly
true in the immobilization of self-destructing enzymes (proteinases). A comprehensive
review on enzyme stabilization by immobilization has been written byKlibanov (1979).

i) Binding to a support

If unfolding is regarded as a compulsory step in enzyme inactivation, then if an enzyme

globule is made rigid, it will be more difficult to disrupt its active site. (Of course, the
stiffness must not increase too much to yield a “frozen” active center). A good rigidity is
achieved when the protein molecule is linked to a solid support by several strong
 Applied biocatalysis 330
chemical bonds. Gabel (1973) showed that the greater the number of linkages in a
trypsin-Sephadex complex, the more stable the protein against inactivation.
Presently, one of the general principles of enzyme stabilization is that of multipoint
covalent attachment of the molecule to a support. In practice, difficulties arise because it
is not easy to find congruency between support and enzyme surfaces. Martinek et al.
(1977) devised a strategy of binding the enzyme to a complementary surface in the
support, so that the multipoint linkage affects not only a small portion of the protein
surface: This was achieved by modifying the enzyme with an analogue of a monomer
(acrylic) and then copolymerizing the resulting preparation with the monomer. An
Arrhenius plot of the rate of inactivation of -chymotrypsin chemically entrapped in
acrylic gels is presented in Figure 8.5. By extrapolating at 60°C, chymotrypsin,
immobilized in polymetacrylate and in polyacrylamide, is 1000 and

Figure 8.5 Temperature dependence of the first-order rate constant (min−1) for
mono-molecular thermoinactivation of free and immobilized -
 Biocatalyst performance 331
chymotrypsin. Conditions: 5 mM tris-HCl, pH 8.0, 3M KC1. , native
enzyme; , chymotrypsin modified with acryloyl chloride; acryloyl
chymotrypsin chemically entrapped in polymethacrylate gel ( ,
water-soluble enzyme; , water-insoluble enzyme) ; , acryloyl
chymotrypsin chemically entrapped in polyacrylamide gel. (Adapted
from Martinek et al.,).

200 times as stable as the free native enzyme, respectively, whereas the acryloylated
chymotrypsin is destabilized when compared with the native protein. Similarly, further
extrapolation shows that at 100°C the immobilized enzyme would be 108 times more
stable than the native one.
In the process of contacting enzyme and support, many variables can be changed in
order to promote immobilization by multiple bond formation. Supports with a great
density of activated sites (for subsequent linkage to the ubiquitous amino groups in
proteins) have been obtained. The insolubilization of micrococcal endonuclease on
CNBr-activated agarose yielded derivatives 300 to 700 times more stable than the native
enzyme at 40–50°C (Guisan and Ballesteros, 1979; Guisan, Serrano and Ballesteros,
1993). Insolubilization of other enzymes on agarose activated with 2,3-epoxy-1-propanol
and then oxidized with periodic acid-, yielded very stable insoluble derivatives at 50°C.
Trypsin immobilized on agarose was 25000 times more stable than the native (Guisan
and Blanco, 1987), whereas penicillin acylase was stabilized 50000-fold (Alvaro et al.,
1990). An unstable lipase from Candida rugosa was made 150-fold more stable (Otero,
Ballesteros and Guisan, 1988).
Multipoint attachment to a support protects the enzyme from inactivation by organic
solvents. Mozhaev et al. (1990) have recently demonstrated that covalent linkage to
polyacrylamide gel stabilizes -chymotrypsin from denaturation by alcohols, the
stabilizing effect increasing with the number of bonds between the protein and the
support.
Although less well studied, stabilization of enzymes by multipoint fixation can also be
obtained by non-covalent bonding to matrices or by covalent bonding to soluble
polymers, like dextrans (Schmid, 1979).

c) Stabilization by chemical and/or biochemical modification

Chemical modification of proteins has been extensively studied over the years to identify
which amino acids are involved in catalysis. Much less work has been carried out on its
influence on enzyme stability. Chemical modification of proteins may yield stabilization,
destabilization or no effect at all. Martinek and Berezin (1978) reported the dependence
of the thermostability of chymotrypsin on the degree of alkylation of its amino groups: up
to 30% alkylation the stability rose slightly; at 90% substitution stability increased
markedly, with a maximum (110-fold) at 95%; stability fell to nearly initial values when
100% amino groups were modified. (With these modifications, the optimum pH of the
enzyme can change and one must therefore be cautious in comparing two different
 Applied biocatalysis 332
species, native and modified). Covalent binding of polyethylene glycol to enzymes and
proteins (see e.g., Inada et al., 1988) has recently been found to be a good method for
stability in apolar solvents.
Intramolecular crosslinking with bifunctional reagents is an important method of
protein stabilization. These bonds can be covalent or noncovalent, and their presence
hinders the unfolding of the macromolecule. Indeed, nature utilises these linkages to
stabilize proteins (Torchilin and Martinek, 1979). The success of in vitro intramolecular
crosslink formation in a protein will rely on the good match between the length of the
bifunctional reagent and the distance of the groups to be linked. Torchilin et al. (1978)
have suggested a rigidification of enzyme molecules with the corresponding functional
groups. For a number of reasons, the majority of studies have been carried out with
glutaraldehyde. In some cases the stabilization is a result of inter-subunit crosslinking. If
the crosslinking reaction is carried out with a large concentration of enzyme, there is a
fair possibility of obtaining intermolecular crosslinking; in applied enzymology there is a
wealth of data on the production of stable polymeric aggregates of enzymes.
The most important biochemical modification of enzymes is glycosylation, the extent
of which can control the bioactivity of the polypeptide (Rademacher, Parekh and Dweek,
1988). A constant feature of glycosylation is microheterogeneity: at any glycosylation
site the oligosaccharide antenna can be different or the site may be unoccupied. The
presence of sugar in a protein usually makes crystallization much more difficult. In
biotechnology, it is possible to manipulate the glycosylation of a polypeptide by selecting
the type of cell used for gene expression, engineering of the production process, etc.
Although it was accepted that only eukaryotes can glycosylate, it is now known that
many glycoproteins are also found in bacteria. Presently, the exact role of the sugar
moieties in enzymes or proteins has not been elucidated. It is common knowledge that, in
higher animals the presence of the carbohydrates imparts resistance to proteolysis and
consequently, longer half-life to the protein in the circulation system. Concerning the role
of the oligosaccharides in stabilization of proteins, not many studies have specifically
addressed this point (Rua et al., 1993). The isozymes of invertase became denatured at
rates that decreased as their carbohydrate content increased (Arnold, 1969).
Glucoamylase possesses many short oligosaccharide chains linked along its polypeptide
backbone. Glycoenzymes with this type of sugar structure are extremely stable to
denaturation (Pazur and Aronson, 1972).
Other post-translational modifications of proteins (e.g., phosphorylation) are extremely
important mechanisms of regulation of enzyme activity. Very little work has been done
on the effects of such modifications on enzyme stabilization.

d) Stabilization by protein engineering

It is now possible, by site-directed mutagenesis of the gene, to change any amino acid
residue in a protein. In enzymology, this technique is leading to new insights into the
process of catalysis (Knowles, 1987). The ultimate goal is to design tailor-made enzymes
with specific catalytic properties and stability. One answer is the engineering of
 Biocatalyst performance 333
additional sulphydryl groups to achieve a more stable, crosslinked protein. In
glycoenzymes, altering the process of glycosylation, (i.e., replacing an Asn residue) is an
alternative approach. Yutani et al. (1987) have studied the role of individual residues in
stabilizing the conformation of the -subunit of tryptophan synthase. They replaced Glu
49, which is buried in the interior of the protein, by any of the other amino acids, and
found that the stability of the resulting enzyme tended to increase linearly with increasing
hydrophobicity of the substituted residue.
A traditional in vivo approach has been to select stable enzymes from organisms that
grow in extreme conditions—engineering by nature. This highlights the importance of
extensive screening for enzymes (Cheetham, 1987), and has provided insight into
mechanisms of stabilization.
A comprehensive description demonstrating the usefulness of protein design to
enhance the stability of enzymes can be found (Protein Engineering) in Chapter 7.

e) Miscellaneous

Zaks and Klibanov (1984) discovered a new way of enzyme stabilization by enzyme
dehydration. Dry pancreatic lipase, when placed in organic medium containing only
0.015% water, becomes much more stable than the normal enzyme preparation in an
aqueous medium (see Chapter 9).
Martinek et al. (1981) also studied stabilization in systems of low water content.
Several enzymes have been microencapsulated into reversed micelles formed by
surfactants in apolar organic solvents (see Chapter 9). The enzymes retained their
catalytic activity and substrate specificity.
Oligomers or enzyme aggregates are often more stable than the constituent monomers.
Based on this idea of protein-protein interactions, Shami, Rothstein and Ramjeesingh
(1989) have proposed a new approach to stabilization using antibodies specific for the
enzyme.

8.2.3 Other Biocatalysts

In general enzymes are more stable in their natural microenvironment (i.e. in cells). This
increased stability can be due to different effectors such as membranes or ligands.
Intracellular enzymes can be further stabilized if the whole cells are immobilized, for
example, entrapped in a polymeric gel.
If the gene coding for the enzyme of interest is extrachromosomal, then immobilization
of cells is a good way to ensure retention of the plasmid during growth.

8.3 QUESTIONS

8.1. Why are the pKa values of different aspartic acid side chains in a protein not
 Applied biocatalysis 334
identical?

8.2. By which mechanisms does the pH value influence the catalytic performance of an
enzyme?

8.3. Is it of importance to know the kinetic parameters kc and KM if you are going to
at not?
use an enzyme as catalyst in an industrial process? Why or why

8.4. Calculate the ratio v /V

max under inhibition
a) Ki=5 mM, [I]=5 mM, icompetitive
the following conditions: Ks=2.5 mM, [S]=0.5 M
b) Ki=5 mM, [I]=5 mM, uncompetitive inhibition
c) Ki=5 mM, [I]=5 mM, noncompetitive inhibition
d) Ki= =5 mM, [I]=5 mM, mixed inhibition

8.5. In one text you find that an enzyme expresses the highest catalytic activity at 35°C
and in another text it is claimed that the same enzyme is maximally active at 48°C.
Suggest a possible explanation!

8.6. Does immobilisation stabilize enzymes?

8.4 HINTS AND ANSWERS

8.1 and 8.2. See “Effect of pH” in paragraph 8.1.3.

8.3. The k value gives information on how fast the enzyme can work when saturated
cat The higher the kcat value is, the lower amount of enzyme can be used.
with substrate.
The KM value should be compared with the substrate concentration that should be used.
If the KM value is very low compared to the substrate concentration in the reactor, then
the enzyme will always work near the maximal rate and the enzyme will thus be
efficiently used. If the KM value is high compared to the substrate concentrations to be
used, the rate will vary rather much with the substrate concentration. In the latter case it
will be important to choose a reactor type which operates at relatively high substrate
concentrations.

8.4. The ratios are:

a) 1/1.01
b) 1/2.005
c) 1/2.01
d) 1/2.01
 Biocatalyst performance 335

8.5. The “optimal temperature” is not a welldefined characteristic. Most probably the
two measurements have been carried out using different reaction times and enzyme
inactivation causes a reduction in activity if the reaction time is long. See paragraph
8.1.3. “Effect of temperature”.

8.6. Sometimes, but not always. See “Stabilization by immobilization” in paragraph

8.2.2.

8.5 REFERENCES AND SUGGESTED FURTHER READING

The following texts are suggested for those who want a more extensive overview of
different topics treated in this chapter:
Damjanovich, S., Keleti, T. and Tron, L. eds. (1986) Dynamics of Biochemical Systems .
Amsterdam: Elsevier Science Publishers.
Fersht, A. (1985) Enzyme Structure and Mechanism , W.H.Freeman, Reading.
Hammes, G.G. (1992) Enzyme Catalysis and Regulation . New York: Academic Press.
Keleti, T. (1986) Basic Enzyme Kinetics . Budapest: The Publising House of the
Hungarian Academy of Sciences.
Krautz, K. and Waldmann, H. (1995) Enzyme Catalysis in Organic Synthesis . New
York: VCH Publishers Inc.
Mozhaev, V.V. (1993) Mechanism-based strategies for protein thermostabilization.
Trends Biotechnol , 11 , 88–95.
Ó Fágáin, C. (1997) Stabilizing Protein Function . Berlin: Springer.

More detailed literature references:

Albery, W.J. and Knowles, J.R. (1976) Evolution of enzyme function and the
development of catalytic efficiency. Biochem. , 15, 5631–5640.
Alvaro, G., Fernandez-Lafuente, R., Blanco, R.M. and Guisan, J.M. (1990)
Immobilization—stabilization of penicillin G acylase from Escherichia coli. Appl.
Biochem. Biotechnol. , 26, 181–196.
Arnold, W.N. (1969) Heat inactivation kinetics of yeast -fructofuranosidase. A
polydisperse system. Biochim. Biophys. Acta , 178, 347–353.
Biselli, M., Kragl, U. and Wandrey, C. (1995) Reaction engineering for enzyme-
catalyzed biotransformations. In Enzyme Catalysis in Organic Synthesis , edited by
K.Krautz and H.Waldmann, pp. 125–130. New York: VCH Publishers. Inc.
Brosnan, M.P., Kelly, C.T. and Fogarty, W.M. (1992) Investigation of the mechanisms of
irreversible thermoinactivation of Bacillus stearothermophilus -amylase. Eur. J.
Biochem. , 203, 225–231.
Bruns, F.H. and Schulze, C. (1962) Acylase 1. Reindarstellung, physikalisch-chemische
Eigenschaften und Identität mit Hippernikase. Biochem. Zeitsch. , 336, 162–181.
 Applied biocatalysis 336
Cheetham, P.S.J. (1987) Screening for novel biocatalysts. Enzyme Microb. Technol. , 9,
194–213.
Chibata, I., (1978) Immobilized Enzymes . Tokyo: Kadansha.
Cornish-Bowden, A. (1995) Fundamentals of Enzyme Kinetics . London: Portland Press.
Fersht, A. (1985) Enzyme Structure and Mechanism. Reading: W.H.Freeman.
Forney, L.J. and Wong, B.C. (1989) Alteration of the catalytic efficiency of penicillin
amidase from Escherichia coli. Appl Environ. Microb. , 55, 2556–2560.
Gabel, D. (1973) The denaturation by urea and guanidinum chloride of trypsin and N-
acetylated-trypsin derivatives bound to sephadex and agarose . Eur. J. Biochem. , 33,
348–356.
Gianfreda, L. and Scarfi, M.R. (1991) Enzyme stabilization: state of the art. Molec. Cell.
Biochem. , 100, 97–128.
Gray, C.J. (1988) Additives and enzyme stability. Biocatalysis , 1, 187–196.
Guisan, J.M. and Ballesteros, A. (1979) Preparation of immobilized sepharose-
micrococcal nuclease derivatives: activity and stability. J. Solid-Phase Biochem. , 4,
245–252.
Guisan J.M. and Blanco, R.M. (1987) Stabilization of trypsin by multi-point attachment
to aldehyde-agarose gels. Ann. NY Acad. Sci. , 501, 67–72.
Guisan, J.M., Serrano, J. and Ballesteros, A. (1993) Stabilization of micrococcal
endonuclease on agarose gels highly activated with CNBr. Biocatalysis , 8, 81–89.
Hammes, G.G. (1992) Enzyme Catalysis and Regulation . New York: Academic Press.
Inada, Y., Takahashi, K., Yoshimoto, T., Ajima, A., Matsushima, A. and Saito, Y. (1988)
Application of PEG-enzyme and magnetite-PEG-enzyme conjugates for
biotechnological processes. Trends Biotechnol. , 6, 131–134.
Ido, E., Kezdy, F.J., Han, H., and Tang, J. (1991) Kinetic studies of human
immunodeficiency virus type I protease and its active-site hydrogen bond mutant
A28S. J. Biol. Chem. , 226, 24359–24366.
Jaenicke, R. (1991) Protein stability and molecular adaptation to extreme conditions. Eur.
J. Biochem. , 202, 715–728.
Keleti, T. (1986) Basic Enzyme Kinetics , Akademic Kiado, Budapest.
Klibanov, A.M. (1979) Enzyme stabilization by immobilization. Anal. Biochem. , 93, 1–
25.
Klibanov, A.M. (1983) Stabilization of enzymes against thermal inactivation. Adv. Appl.
Microbiol. , 29, 1–28.
Kobayashi, M., Miura, M. and Ichishima, E. (1992) Modification of subsite Lys residue
induced large increase in maltosidase activity of Taka-amylase A. Biochem. Biophys.
Res. Comm. , 183, 321–326.
Knowles, J.R. (1987) Tinkering with enzymes: what are we learning?. Science , 236,
1252–1258.
Lehninger, A. (1970) Biochemistry . New York: Worth Publishers Inc.
Lumry, R. and Eyring, H. (1954) Conformation changes of proteins. J. Phys. Chem. , 58 ,
110–120.
Martin, J., Slade, A., Aitken, A., Anche, R. and Virden, R. (1991) Chemical modification
of serine at the active site of penicillin acylase from Kluyvera citrophila. Biochem. J. ,
280, 659–662.
Martinek, K. and Berezin, I.V. (1978) General principles of enzyme stabilization. J.
Solid-Phase Biochem. , 2, 343–385.
 Biocatalyst performance 337
Martinek, K., Levashov, A.V., Klyachko, N.L., Pantin, V.I. and Berezin, I.V. (1981) The
principles of enzyme stabilization. VI. Catalysis by water-soluble enzymes entrapped
into reversed micelles of surfactants in organic solvents. Biochim. Biophys. Acta , 657,
277–294.
Martinek, K., Klibanov, A.M. Goldmacher, V.S. and Berezin, I.V. (1977) The principles
of enzyme stabilization. I. Increase in thermostability of enzymes covalently bound to
a complementary surface of a polymer support in a multipoint fashion . Biochim.
Biophys. Acta , 485, 1–12.
Melrose, G.J.H. (1971) Insolubilized Enzymes—Biochemical applications of synthetic
polymers. Rev. Pure Appl. Chem. , 21, 83–119.
Monod, J. Wyman, J. and Changeaux, J.P. (1965) On the nature of allosteric transitions: a
plausible model. J. Mol. Biol. , 12, 88–118.
Monsan, P. and Combes, D. (1984) Effect of water activity on enzyme action and
stability. Ann. NY Acad. Sci. , 434, 48–63.
Morand, P. and Biellmann, J.F. (1991) Modification of -amylase from Bacillus
licheniformis by the polyaldehyde derived from -cyclodextrin and -amylase
thermostability. FEBS Lett. , 289, 148–150.
Mozhaev, V.V. and Martinek, K. (1984) Structure-stability relationships in proteins: new
approaches to stabilizing enzymes. Enzyme Microb. Technol. , 6, 50–59.
Mozhaev, V.V., Sergeeva, M.V. Belora, A.B. and Khmelnitsky, Y.L. (1990) Multipoint
attachment to a support protects enzyme from inactivation by organic solvents: -
chymotrypsin in aqueous solutiuons of alcohols and diols. Biotechnol. Bioeng. , 35,
653–659.
Otero, C., Ballesteros, A. and Guisan, J.M. (1988) Immobilization/Stabilization of lipase
from Candida rugosa. Appl. Biochem. Biotechnol. , 19, 163–175.
Pazur, J.H. and Aronson, N.N.Jr. (1972) Glycoenzymes: enzymes of glycoprotein
structure. Adv. Carbohyd. Chem. Biochem. , 27, 301–341.
Pettersson, G. (1992) Evolutionary optimization of the catalytic efficiency of enzymes.
Eur. J. Biochem. , 206, 289–295.
Polgar L. and Bender, M.L. (1966) A new enzyme containing a synthetically formed
active site. Thiol-subtilisin. J. Am. Chem. Soc. , 88, 3153–3154.
Rademacher, T.W., Parekh, R.B. and Dweek, R.A. (1988) Glybiology. Ann. Rev.
Biochen. , 57, 785–838.
Rosenberg, A. and Somogyi, B. (1986) Conformational fluctuations, thermal stability and
hydration of proteins, studies by hydrogen exchange kinetics. In Dynamic of
Biochemical systems , edited by S. Damjanovich, T.Keleti and L.Tron, pp. 101–112.
Amsterdam: Elsevier.
Rua, M.L., Diaz-Maurino, T. Fernandez, V.M., Otero, C. and Ballesteros, A. (1993)
Purification and partial characterization of two distinct lipases from Candida
cylindracea. Biochim. Biophys. Acta. , 1156, 181–189.
Sadana, A., Raju, R.R. and Shahin, E. (1989) A stability index for enzyme deactivation.
J. Biotechnol. , 12 , 135–152.
Schmid, R.D. (1979) Stabilized soluble enzymes. Adv. Biochem. Eng. , 12, 41–118.
Segel I.H. (1975) Enzyme Kinetics . New York: Wiley Interscience Publ.
Shami, E.Y., Rothstein, A. and Ramjeesingh, M. (1989) Stabilization of biologically
active proteins. Trends in Biotechnology , 7, 186–190.
Slade, A., Horrocks, A.J., Lindsay, C.D., Dernbar, B. and Virden, R. (1991) Site-directed
 Applied biocatalysis 338
chemical conversion of serine to cysteine in penicillin acylase from Escherichia coli
ATCC 11105. Effect on conformation and catalytic activity. Eur. J. Biochem. , 197,
75–80.
Somogyi, B., Welch, G.R. and Damjanovich, S. (1984) The dynamic basis of energy
transduction in enzyme. Biochim. Biophys. Acta , 768, 81–112.
Stackhouse, J., Namblar, K.P., Benbaum, J.J., Stauffer, D.M. and Benner, S.A. (1985)
Dynamic transduction of energy and internal equilibria in enzymes: a reexamination of
pyruvate kinase. J. Am. Chem. Soc. , 107, 2757–2763.
Stellwagen, E (1984) Strategies for increasing the stability of enzymes. Annals N.Y.
Acad. Sci. , 434, 1–6.
Straub, F.B., and Szabolcsi, G. (1964) Dynamic aspects of enzyme structure. Mol. Biol.
Izdat Nauka (Moscow) , 182–187.
Szajani, B.Ivory, K. and Boross L. (1980) Preparation and practical utilization of a highly
active immobilized form of porcine kidney amino-acylase. Acta. Biochem. Biophys.
Acad. Sci. Hung. , 15, 295–303.
Tanford, C. (1968) Protein denaturation. Adv. Prot. Chem. , 23, 121–282.
Tewari, Y.B. (1990) Thermodynamics of industrially-important, enzyme catalysed
reactions. Appl. Biochem. Biotechnol. , 23, 187–203.
Tomazic, S.J. and Klibanov, A.M. (1988) Mechanisms of irreversible thermal
inactivation of Bacillus -amylases. J. Biol. Chem. , 263, 3086–3091.
Torchilin, V.P., Maksimenko, A.V., Smirnov, V.N., Berezin, I.V., Klibanov, A.M. and
Martinek, K. (1978) The principles of enzyme stabilization. III. The effect of the length
of intra-molecular cross-linkages on thermostability of enzymes. Biochim. Biophys.
Acta , 522, 277–283.
Torchilin, V.P. and Martinek, K. (1979) Enzyme stabilization without carriers. Enzyme
Microb. Technol. , 1, 74–82.
Trevan, M.D. (1980) Immobilized enzymes . Chichester: J.Wiley & Sons Ltd.
Ugarova, N.N., Rozhkova, G.D. and Berezin, I.V. (1978) A physico chemical analysis of
soluble and immobilized enzyme stabilization. J. Solid-Phase Biochem. , 3, 119–131.
Vas, M. and Boross, L. (1974) An approach for the determination of equilibrium constant
of structural motility. Eur. J. Biochem. , 43, 237–244.
Wiseman, A. (1975) Handbook of Enzyme Biotechnology . Chichester: Ellis Horwood.
Yutani, K., Ogasahara, K., Tsujita, T. and Sugino, Y. (1987) Dependence of
conformations stability on hydrophobicity of the amino acid residue in a series of
variant proteins substituted at a unique position of tryptophan synthase subunit. Proc.
Natl. Acad. Sci. USA , 84, 4441–4444.
Zaks, A. and Klibanov, A. (1984) Enzymatic catalysis in organic media at 100°C.
Science , 224, 1249–1251.
 9.
BIOCATALYSIS IN NON-CONVENTIONAL
MEDIA
PATRICK ADLERCREUTZ
Dept. of Biotechnology, Center for Chemistry and Chemical Engineering, Lund
University, P.O. Box 124, S-221 00 Lund, Sweden
Telephone: +46 46 2224842; Fax: +46462224713;
Email: [email protected]

ABSTRACT
The scope and limitations of biocatalysis in non-conventional media are
described. First, different kinds of non-conventional reaction media, such as
organic solvents, supercritical fluids, gaseous media and solvent-free systems,
are treated. Second, enzyme preparations suitable for use in these media are
described. In several cases the enzyme is present as a solid phase but there are
methods to solubilise enzymes in non-conventional media, as well. Third,
important reaction parameters for biocatalysis in non-conventional media are
discussed. The water content is of large importance in all non-conventional
systems. The effects of the reaction medium on enzyme activity, stability and on
reaction yield are described. Finally, a few applications are briefly presented.

9.1 INTRODUCTION

9.1.1 What is Meant by Non-conventional Media?

In living cells, water is the predominant solvent. It is therefore not surprising that
scientific studies of enzymes have been carried out mainly in aqueous media. Often quite
dilute solutions of substrates and enzymes in aqueous buffers have been studied.
However, one should bear in mind that high concentrations of proteins, other
biopolymers and low molecular weight compounds are present around the enzymes in
living cells. Furthermore, some enzymes are associated with membrane structures
containing mainly hydrophobic lipids. Accordingly, some of the “non-conventional”
 Applied biocatalysis 340
reaction media used recently for enzymatic reactions may be as natural for the enzymes
as dilute aqueous solutions. Media normally referred to as non-conventional for
enzymatic reactions are those containing mainly organic substances (solvents, substrates,
products, etc.) or supercritical fluids (Table 9.1). Another type of non-conventional
systems uses gaseous reaction media. Common for all these media is that the water
content is reduced compared to “conventional media” (aqueous solutions).

9.1.2 Why are Non-conventional Media Used?

Most enzymes are able to catalyse reactions in both directions. The enzymes are just
catalysts so they can never determine in which direction a reaction is going, that is

Table 9.1 Non-conventional media for biocatalysis.

Organic solvents
Supercritical fluids
Solvent-free systems
liquid systems
gaseous systems

determined by the reaction conditions and the equilibrium position. Non-conventional

media are of special interest for hydrolases. In aqueous media these enzymes catalyse
hydrolysis, but most non-conventional media shift the equilibrium position so that the
enzymes can be used for reversed hydrolytic reactions and transferase reactions. This
widens the scope of hydrolases considerably.
Another important advantage with many non-conventional media is that they can
solubilise hydrophobic compounds which are poorly soluble in water. Thereby the
conversion of these important substrates is facilitated. Further advantages are that the risk
of microbial contamination is much lower in non-conventional media, and under
optimized conditions the stability of enzymes is often higher than in aqueous solutions.

9.2 TYPES OF NON-CONVENTIONAL MEDIA

9.2.1 Organic Solvents

Organic solvents constitute the most commonly used non-conventional media for
biocatalysis. They are commonly employed to change equilibrium positions of reactions
 Biocatalyst in non-conventional media 341
catalysed by hydrolytic enzymes or to increase the solubility of hydrophobic compounds.
Some organic solvents are miscible with water in all proportions. Often used solvents in
this group are alcohols, acetone, dimethyl sulphoxide and dimethyl formamide. Often the
enzyme is dissolved in the solvent-water mixture so that a homogeneous system is
formed. This is advantageous since there are no mass transfer limitations. However, if a
high concentration of solvent is needed, poor operational stability of the enzyme is often
observed. The amount of solvent that can be added without negative effects on the
enzyme varies depending on the enzyme and the solvent (Khmelnitsky et al., 1988).
Often the enzyme stability can be improved by using a suitable water-immiscible
solvent instead of a water-miscible one. Two-phase systems are obtained with the
enzyme and other hydrophilic substances present in the aqueous phase while hydrophobic
substrates and products mainly partition to the organic phase (Figure 9.1). Water
immiscible solvents often used for enzymatic reactions are hydrocarbons, ethers and
esters; further details on solvents are found in the section “9.5 Selection of solvents”,
below. In order for the bioconversion to occur, the substrates must be transferred to the
enzyme in the aqueous phase; after the reaction hydrophobic

Figure 9.1 Schematic presentation of an enzymatic conversion in a two-phase

system. S= substrate, P=product and E=enzyme.

products are transferred back to the organic phase. Because of these transport processes,
it is important that the interfacial area is large enough or else mass transfer will limit the
bioconversion rate. However, there is also a risk for enzyme inactivation at the interface.
The distribution of the reactants in the aqueous/organic two-phase system can be
controlled by choosing a suitable solvent and to some extent by manipulations of the
aqueous phase, for example by changing pH (of course the pH must be compatible with
the enzyme). The partitioning of substrates and products to the organic phase is a great
advantage when substrate and/or product inhibition is a problem in homogeneous
 Applied biocatalysis 342
systems.
The proportions of water and organic solvent can be varied from pure water to almost
pure organic solvent. In order to retain enzymatic activity there seems to be a need for a
little water. However, this minimal amount of water is sometimes considerably less than
a monolayer of water around the enzyme molecules. The rest of the medium can be an
organic solvent. The effects of water on biocatalysis in non-conventional media are
treated below.

9.2.2 Supercritical Fluids

All materials have critical points which determine their phase behaviour. The critical
temperature is that above which a gas cannot be liquified by increasing the pressure.
Instead, a supercritical fluid with properties between gaseous and liquid phases is formed.
Enzymes can express activity in supercritical and near supercritical fluids (Nakamura,
1990). The most commonly used fluid in this group is supercritical carbon dioxide. The
main advantages with this reaction medium are that it is nontoxic and that it can be
removed easily after the reaction. Furthermore, diffusion rates are considerably higher in
supercritical media compared to normal liquid media. This could increase the overall
bioconversion rate in some cases. As solvent, supercritical carbon dioxide resembles
hexane. It is thus water-immiscible and is a good solvent for very hydrophobic
compounds. The solubility of compounds of intermediate polarity can be improved by the
addition of organic cosolvents. The main drawback with supercritical reaction media is
that high pressures must be used, which requires special reactors and other equipment.

9.2.3 Gaseous Media

It has been shown possible to convert gaseous substrates with enzymes (Russell and
Yang, 1996; Lamare and Legoy, 1993). The enzyme is present as a solid phase which is
passed by a stream of gaseous substrate. There is no need for a liquid phase. An example
of such a reaction is the oxidation of ethanol by alcohol oxidase. Relatively high reaction
temperatures are used to keep the substrates in the gas phase. Consequently, it is
favourable to use thermostable enzymes for these applications.

9.2.4 Solvent-free Reaction Media

From the industrial point of view, it is advantageous to work with minimal amounts of
solvents. This minimises the reactor size and thereby reactor cost. The extreme case is to
omit the solvent completely. The use of solvent-free systems is attractive also because
solvents can cause many problems (for example fire hazards, environmental problems
and high costs). It has been proven possible in many cases to carry out bioconversions in
solvent-free mixtures of substrates. Since most substrates are organic compounds, these
mixtures behave like systems containing organic solvents and the advantages mentioned
 Biocatalyst in non-conventional media 343
above are applicable for the solvent-free case as well. It is for example possible to reverse
hydrolytic reactions in solvent-free systems.
In order for the reaction to occur with an acceptable rate it is normally required that at
least part of the substrates are present in a liquid form (or gaseous, for gas phase
reactions). In many cases this is easily achieved but for high-melting substrates there is
often a need for elevated temperatures, which makes it necessary to use thermostable
enzymes. In some cases the substrates can be chosen so that they form an eutectic
mixture, which is characterised by a considerably lower melting point than the individual
substrates. In some case the melting point has been further lowered by the addition of
small amounts of extra components, often polar organic solvents (Gill and Vulfson,
1994).

9.3 ENZYME PREPARATIONS SUITABLE FOR NON-CONVENTIONAL

MEDIA

Different types of enzyme preparations are shown in Figure 9.2.

Figure 9.2 A schematic presentation of different types of enzyme preparations

used in non conventional media, a: enzyme powder; b: enzyme
crystals; c: enzyme on a porous support; d: covalently modified
enzyme dissolved in the solvent; e: enzyme solubilised by surfactant;
 Applied biocatalysis 344
f: enzyme solubilised by polymer; g: enzyme solubilised in a microemulsion.

9.3.1 Enzyme Powders

The most straightforward way of using solid enzymes in organic media is to suspend the
solid enzyme directly in the solvent. If one wants to get quick results from a
bioconversion and does not want to optimise the efficiency of the enzyme, this method is
the obvious first choice. There are many example in the literature where enzymes have
been used successfully in synthesis just as powders directly from the enzyme
manufacturer. Sometimes there is a need to dissolve the enzyme powder and re-lyophilise
it from a buffer with a more suitable composition, see the section “9.6 pH control in non-
conventional media”.
Most enzyme powders are prepared by lyophilisation (freeze drying). However, the
lyophilization procedure might inactivate the enzyme to some extent. To avoid this and
thereby increase the activity of lyophilized enzymes in dry organic solvents, the
lyophilization can be carried out in the presence of lyoprotectants such as sorbitol
(Dabulis and Klibanov, 1993). The inactivation is believed to be caused at least partly by
a reversible conformational change in the enzyme. This process can be reversed and the
enzyme reactivated by the addition of small amounts of water (Dabulis and Klibanov,
1993).

9.3.2 Enzyme Crystals

Crystals constitute the most concentrated form of enzyme and they can therefore be
attractive as catalysts. In the crystallisation of a crude enzyme preparation a considerable
purification of the enzyme can be achieved which is a further advantage. In order to
stabilise the enzyme crystals to make them useful catalysts they are often crosslinked
with bifunctional reagents such as glutaraldehyde. Very high catalytic activity and
stability has been reported for these crosslinked enzyme crystals, some of which are
commercially available (Margolin, 1996).

9.3.3 Enzymes on Supports

In many applications of enzymes in non-conventional media, the enzymes are used in the
immobilised form on support materials which often are porous. These immobilised
preparations usually express considerably higher specific activity (moles substrate
converted per unit enzyme and time) than enzyme powders. The reason can be facilitation
of mass transfer either by the spreading of the enzyme on a large surface area or by better
suspension of the catalyst particles (enzyme powders often tend to aggregate). An
alternative explanation is that the support might protect the enzyme from inactivation
during the drying of the preparation.
 Biocatalyst in non-conventional media 345

Methods to immobilise enzyme on porous supports

Immobilisation methods are treated in detail in chapter 6. Most enzyme immobilisation

methods used in connection with non-conventional media rely on noncovalent
interactions between the support and the enzyme. The reason why this works well in
many cases is that enzymes normally have a low tendency to dissolve in the reaction
media used. Adsorption or deposition on porous supports are often used methods. It is
important to remember that other substances (for example salts and other polar
substances) are often “immobilised” on the support because they are present during the
immobilisation procedure and not soluble in the reaction medium. Those substances
influence the microenvironment of the enzyme and thereby its catalytic activity.

Support characteristics

General characteristics of the support to be considered are discussed in chapter 6 and

those of relevance for the use of enzymes in non-conventional media are listed in Table
9.2. In addition, the specific surface area of the support is of special importance for the
applications in non-conventional media. Inactivation can occur if the surface area is too
large in relation to the amount of enzyme (Figure 9.3). In order to avoid inactivation at
least a monolayer of enzyme molecules should be formed on the accessible surface
(Wehtje, Adlercreutz and Mattiasson, 1993). For enzymes of high purity and high
activity, the amount of enzyme needed is sometimes quite small. In such cases the
enzyme can be mixed with a protecting protein before immobilisation to achieve at least a
monolayer of protein and thereby avoid

Table 9.2 Factors to consider when choosing the support for an enzyme to be used in a
non-conventional medium.

Mechanical properties
tendency to break
compressibility
Influence on enzymatic activity
morphological characteristics
particle size
pore size
specific surface area
 Applied biocatalysis 346

chemical properties
water partition characteristics
direct effects on the enzyme
Cost

inactivation. On the other hand, mass transfer limitations can start to limit the reaction
rate if too much enzyme is immobilised in relation to the surface area (Wehtje,
Adlercreutz and Mattiasson, 1993).
In the section “9.4 Effects of water” below, it is described how the support can
influence the distribution of water in the system and thereby influence the activity

Figure 9.3 The effect of the enzyme loading on a porous support on the
catalytic activity expressed. At low loadings partial inactivation of
the enzyme often occurs. At high loadings mass transfer limitations
reduce the observed activity. The extent of decrease at low and high
loading depends on the enzyme, the support and the reaction.

expressed by the enzyme. Furthermore, substrates and products can interact with the
support. This can influence the rate of the reaction. Furthermore, if the product is
adsorbed on the support, its extraction after complete reaction can damage the enzyme or
 Biocatalyst in non-conventional media 347
even extract it from the support if it is not covalently bound.

9.3.4 Solubilised Enzymes

The methods described below have been used for enzyme solubilisation in organic media
but they should be applicable to supercritical media and solvent-free systems, as well.

Covalently modified enzymes soluble in organic media

Enzymes can be made soluble in organic media by covalent attachment of polymers. The
most common method is to couple polyethylene glycol chains to the amino groups of the
enzyme (Inada et al., 1986). As the enzyme becomes soluble in the reaction medium
(usually an aromatic or chlorinated hydrocarbon), mass transfer will normally not limit
the reaction rate. The enzyme can be reused following its precipitation from the reaction
mixture with a nonpolar solvent like hexane. A drawback with the method is that
inactivation of the enzyme can occur during the derivatization procedure. Often one must
make a compromise between high solubility (high number of PEG chains added) and
high remaining activity (low number of PEG chains). Other polymers used for enzyme
solubilisation in organic media include polystyrene and polyacrylates.

Preparation of non-covalent enzyme-polymer complexes

Enzymes can be solubilised in organic media in the form of noncovalent enzyme-polymer

complexes (Otamiri, Adlercreutz and Mattiasson, 1992). Using this technique, the
enzyme inactivation occuring during the covalent coupling of polymers to enzymes
(described above) is avoided. The non-covalent complexes are formed simply by
lyophilisation of a solution containing both the enzyme and the polymer followed by
dissolution in the organic medium. Inorganic salts in the lyophilisate seem to play an
important role in the complex formation. Polymers used in this method include ethyl
cellulose and poly(methyl methacrylate).

Solubilisation with surfactants

Surfactants have been used to solubilise lipases in organic solvents (Okahata and Mori,
1997). One method starts with mixing aqueous solutions of the surfactant and the
enzyme. The enzyme-surfactant complex precipitates and can subsequently be dissolved
in organic media. Several surfactants have been tested and especially good results have
been obtained with dialkyl glucosyl glutamates. In one case it was shown that the
complex consisted of one enzyme molecule surrounded by approximately 150 surfactant
molecules.
The surfactant Aerosol OT (dioctylsulfosuccinate, sodium salt), which is widely used
 Applied biocatalysis 348
in the preparation of microemulsions (reversed micelles, see below) has been used to
solubilize chymotrypsin in isooctane (Paradkar and Dordick, 1994). A solution of the
surfactant in the organic solvent was used to extract the enzyme from an aqueous
solution. The number of surfactant molecules participating in the solubilisation was only
about 30 per enzyme molecule which is much less compared to the case with reversed
micelles.

Enzymes in microemulsions

Microemulsions are macroscopically homogeneous and isotropic mixtures of water,

organic solvent and surfactant. Since microemulsions are thermodynamically stable, they
can be formed by just mixing the components. One way to solubilise an enzyme in a
microemulsion is to add a small amount of an aqueous solution of the enzyme to a
solution of the surfactant in the organic solvent. After mixing, a transparent solution is
formed and the enzyme often expresses high catalytic activity (Martinek et al., 1989). On
a microscopic scale microemulsions are structured into aqueous and oil microdomains
separated by a surfactant rich film. The enzyme is normally present in the aqueous
domains. Under some conditions reversed micelles (water droplets surrounded by a
surfactant film and dispersed in the bulk organic phase; see Figure 9.2) are formed and
with other proportions of the components both the aqueous and the organic phase are
continuous. The microemulsions can be considered as organic-aqueous two-phase
systems with a very large interfacial area which minimizes mass transfer limitations.
Under some conditions, large amounts of water can be included in the microemulsions,
which is an advantage in reactions involving both hydrophilic and hydrophobic reactants.
Drawbacks are that the surfactant can inactivate enzymes and cause problems in the
isolation of the reaction product.

9.3.5 Solid Versus Solubilised Preparations

For practical purposes it is often beneficial to use a heterogeneous system with the
enzyme as a solid preparation which easily can be separated from the product in the
liquid phase. Solid enzyme preparations can conveniently be used in packed bed and
stirred tank reactors. As in other cases with heterogeneous catalysis, mass transfer
limitations can reduce the overall reaction rate, but usually this is no major problem.
Solubilised enzyme preparations are well suited for many fundamental studies, for
example spectroscopic investigations requiring transparent solutions. When the
solubilised preparations are used as catalysts it is an advantage that mass transfer
limitations are normally absent, but product isolation and enzyme recovery are usually
more difficult than with solid enzyme preparations. Methods used to separate the enzyme
from the product solution include precipitation of the enzyme, and the use of membranes
which retain the enzyme but not the product.
 Biocatalyst in non-conventional media 349

Figure 9.4 Typical dependence of the specific activity of an enzyme on the

water activity.

9.4 EFFECTS OF WATER

Even in a non-conventional medium which mainly consists of an organic phase, a gas

phase or a supercritical fluid, the amount of water in the system influences the catalytic
activity of the enzyme to a large extent. The hydration level of the enzyme is the key
parameter. If the water content is kept as low as possible, the enzymatic activity is
usually very low. Normally the enzymatic activity increases with increasing enzyme
hydration, this is often explained by water acting like “lubricant” to increase the internal
enzyme flexibility. Sometimes there is a maximum in enzymatic activity at a certain
amount of water in the system (Figure 9.4). One possible reason of the decreased activity
is mass transfer limitations due to substrate transport through an aqueous phase or to
aggregation of the catalyst particles. Furthermore, enzyme stability often decreases with
increasing water content. Water participates in many mechanisms which cause
inactivation of enzymes. Finally, water acts as substrate in reactions catalysed by
hydrolytic enzymes and thereby decreases the yield in reversed hydrolytic reactions and
transferase reactions.
 Applied biocatalysis 350

9.4.1 Quantification of Water

The amount of water in the reaction mixture can be quantified in different ways. The
most common way is to use the water concentration (in mol/1 or % by volume).
However, the water concentration does not give much information on the key parameter:
enzyme hydration. In order to have a parameter which is better correlated with enzyme
hydration, researchers have started to use the water activity to quantify the amount of
water in non-conventional reaction media (Halling, 1984; Bell et al., 1995). For a
detailed description of the term activity (thermodynamic activity), please look in a
textbook in physical chemistry. Activities are often very useful when studying chemical
equilibria and chemical reactions of all kinds, but since they are often difficult to measure
they are not used as much as concentrations. Normally, the water activity is defined so
that it is 1.0 in pure water and 0.0 in a completely dry system. Thus, dilute aqueous
solutions have water activities close to 1 while non-conventional media are found in the
whole range of water activities between 0 and 1. There is a good correlation between the
water activity and enzyme hydration and thus enzyme activity. An advantage with the
activity parameter is that the activity of a component is the same in all phases at
equilibrium. The water activity is most conveniently measured in the gas phase with a
special sensor. The water activity in a liquid phase can thus be measured in the gas phase
above the liquid after equilibration.
Water activity can be used to quantify water in all non-conventional media and is used
to an increasing extent. However, in microemulsions containing reversed micelles, water
is still often quantified as the molar ratio of water to surfactant (abbreviated as w0 or R).
This parameter is used because it is correlated with the size of the reversed micelles,
which in turn is correlated with enzymatic activity.

9.4.2 Distribution of Water in Non-conventional Media

When reactions are carried out in non-conventional media, water is distributed between
the different phases present. Some water is bound to the enzyme and thereby has a large
influence on the catalytic activity. Some water is dissolved in the solvent and if supports,
polymers or other substances are present these bind water as well.
It is thus beneficial to work at fixed water activity in studies of the influence of
solvents, supports or other substances on enzymatic catalysis. Otherwise the effects due
to differences in enzyme hydration will strongly influence the results and mask the
effects sought. A typical example of this was seen when reaction rates were compared for
the same reaction carried out in different solvents at varying water concentrations. In the
different solvents, maximal reaction rate was observed at widely different water
concentrations. However, when water was quantified in terms of water activity the
optimum was observed at about the same water activity in all solvents (Valivety, Halling
and Macrae, 1992) (Figure 9.5).
 Biocatalyst in non-conventional media 351

9.4.3 Water Activity Control

When studying the influence of the water activity on the rate of an enzymatic reaction in
organic solvents one can pre-equilibrate both the enzyme preparation and the substrate
solution in atmospheres of controlled water activity. After mixing, the rate at this water
activity can be measured. Atmospheres with controlled water

Figure 9.5 Distribution of water in mixtures containing an enzyme on a support

suspended in two different organic solvents. The solubility of water is
higher in solvent B than in solvent A. When the solvents are
compared at fixed amount of water, different amounts of water are
bound to the enzyme. However, at fixed water activity, the same
amount of water is bound to the enzyme in the two solvents and a
good evaluation of other solvent effects can be made.

activity can easily be obtained using saturated salt solutions. Small containers with
enzyme or substrate solution can be put into larger containers with saturated salt solutions
in the bottom (Figure 9.6). By using different salts, a range of water activities can be
obtained (Table 9.3). During the reaction, the water activity may change, especially if
water is formed or consumed in the reaction. If the reaction is slow, the equilibration
through the gas phase described above can be used to maintain the water activity, but if
large amounts of water must be removed or added to the reaction at fixed water activity, a
more efficient system is needed. One way to achieve this is to pass the saturated salt
 Applied biocatalysis 352
solution through silicon tubings which are immersed in the reactor (Figure 9.6) (Wehtje
et al., 1993). The surface area of the silicone tubing should be chosen considering the
water transport capacity required.
An alternative method is based on the fact that salt hydrates containing different
numbers of water molecules are interconverted at fixed water activities. The first salt
hydrate used was Na2CO3 10 H2O. This is converted to Na2CO3 7 H2O at a water activity
of 0.74 at 24°C. The salt hydrates act as a buffer of the water activity. As long as both
salt hydrates are present the water activity remains at 0.74. The salt hydrates can be
added directly to the organic reaction mixture. The pair of salt hydrates should be chosen
to give a water activity suitable for the enzymatic conversion (Zacharis et al., 1997).

Figure 9.6 Control of water activity by equilibration via the gas phase or via
silicone tubing. 1: gas phase; 2: immobilised enzyme; 3: reaction
medium (substrate in organic solvent); 4: stirring bar; 5: saturated salt
solution; 6: salt crystals; 7: silicone tubing; 8: organic solvent (to
keep the salt solution solvent saturated).

Table 9.3 Saturated salt solutions suitable for water activity control. Values are given for
25°C.
 Biocatalyst in non-conventional media 353

Salt water activity

LiCl 0.113
K-acetate 0.225
MgCl2 0.328
K2CO3 0.432
Mg(NO3)2 0.529
SrCl2 0.708
KCl 0.843
KNO3 0.936
K2SO4 0.973

Table 9.4 Log P values of some common solvents. From Laane et al. (1987).

Solvent logP
dimethylsulfoxide −1.3
dioxane −1.1
N,N-dimethylforamide −1.0
methanol −0.76
acetonitrile −0.33
ethanol −0.24
acetone −0.23
propanol 0.28
ethyl acetate 0.68
butanol 0.80
diethyl ether 0.85
butyl acetate 1.7
dipropyl ether 1.9
chloroform 2.0
benzene 2.0
toluene 2.5
dibutyl ether 2.9
 Applied biocatalysis 354

tetrachloromethane 3.0
pentane 3.0
cyclohexane 3.2
hexane 3.5
octane 4.5
dodecanol 5.0
dodecane 6.6
dioctyl phthalate 9.6

9.5 SELECTION OF SOLVENT

In all the solvent-containing biocatalytic systems, the nature of the solvent influences the
reaction to a large extent. The solvent can affect the activity and the stability of the
enzyme and the maximal yield in the reaction.

9.5.1 Effects on Enzyme Stability

Solvents can cause enzyme inactivation (decrease the number of active enzyme
molecules). The exact mechanisms are not so well known, but it is clear that solvent
polarity plays an important role. Several solvent parameters have been used to try to
rationalise the influence of solvents on enzymes. The parameter which has been used
most for this purpose is the log P value, which is defined as the logarithm of the partition
coefficient of a substance in the standard 1-octanol/water two-phase system (Table 9.4).
Log P values can be determined experimentally by measuring the partitioning of the
solvent between octanol and water. Alternatively, log P values
 Biocatalyst in non-conventional media 355

Figure 9.7 Activity retention of immobilised cells catalysing an epoxidation

reaction after exposure to organic solvents with different log P
values. Data from Brink and Tramper (1985), plotted according to
Laane et al. (1987).

can be calculated on the basis of the molecular structure of the solvent using the
hydrophobic fragmental constants of Rekker (Rekker, 1977). These constants represent
the contribution of each molecular fragment to the log P value of the substance; they are
based on large numbers of experimental data.
Most systematic investigations of solvent effects on biocatalysts have been carried out
in water-immiscible solvents. Normally solvents in this group with high log P values (>4)
(hydrophobic solvents) cause less inactivation of biocatalysts than more polar solvents
(Laane et al., 1987) (Figure 9.7). Among the solvents with low log P values, which are
mainly alcohols and polyols, there is a clear trend that the solvents with the highest log P
values cause the largest reduction in reaction rate (Khmelnitsky et al., 1988).
Accordingly, considering the whole range of solvents, there seems to be a group of
solvents with intermediate log P values (between 0 and 2) which inactivate enzymes
more than solvents with both lower and higher log P values. For some applications,
solvents from this group are needed to dissolve the substrates and in these cases a
compromise taking into consideration both enzyme stability and substrate solubility must
often be made.
 Applied biocatalysis 356

9.5.2 Effects on Enzyme Activity

The solvation of the substrates by the solvent influences their tendency to react. In an
analogous way to water activity it has been proposed that one should use substrate

Figure 9.8 Schematic presentation of an enzymatic esterification reaction in a

homogeneous system which may contain a water-miscible solvent to
decrease the water activity and in a two phase system containing one
aqueous phase and one organic phase. The enzyme is present in the
aqueous phase.

activities instead of substrate concentrations when dealing with reaction rates and
equilibria. If the substrate has a high solubility in a solvent its activity is lower in that
solvent than in another one in which the substrate is less soluble. The reaction rate is then
supposed to correlate with the substrate activity. There are clear indications that this way
of reasoning is correct: the apparent Km values of substrates in “good” solvents (solvents
in which the substrate is highly soluble) are normally higher than in “bad” solvents (Bell
et al., 1995).
Solvents can act as inhibitors of enzymes and thereby decrease their activity. At least
in some cases the inhibition is competitive and due to binding of solvent molecules in the
active site of the enzyme.
Finally, solvents can dissolve different amounts of water and thereby indirectly
influence the enzyme activity. However, this effect can be compensated for by
controlling the water activity (see above).

9.5.3 Effects on the Maximal Yield in the Reaction

As mentioned above, non-conventional media are often used to shift the equilibrium
position so that reversed hydrolytic reactions can be carried out. The equilibrium shift can
be achieved by different mechanisms. A schematic presentation of two systems for ester
synthesis are shown in Figure 9.8. Let us first consider the homogeneous system which
 Biocatalyst in non-conventional media 357
contains water and may contain a water miscible organic solvent. The position of the
chemical equilibrium is determined by the equilibrium constant:

Eq. 9.1

Ideally, thermodynamic activities of the reactants should be used in the equation, but
since concentrations are normally easier to measure these are often used instead. The use
of the activity of water (which can be measured fairly easily) and the concentrations of
the other reactants has been recommended for studies of enzyme catalyzed reactions in
organic media (Halling, 1984). In order to increase the synthesis of the ester, the water
concentration (or activity) should be reduced. This can be achieved by replacing part of
the water with a water miscible solvent.
If instead a water-immiscible solvent is used, the enzymatic esterification occurs in the
aqueous phase. The esterification is favored by a low water activity as in the
homogeneous system, but another mechanism is usually more important in these two-
phase systems. The ester produced is extracted into the organic phase, thus lowering the
product concentration in the aqueous phase which contains the enzyme. The advantage of
using two-phase systems for shifting the equilibrium is not limited to biocatalytic
reactions. It is a generally observed that in reactions involving two substrates and two
products it is advantageous if the products partition to different phases (Semenov et al.,
1987). When the two products are mainly present in different phases, the reaction
between the two (the reversal of the desired reaction) is greatly suppressed. In the case
with ester synthesis, the ester partitions mainly to the organic phase while the other
reaction product, water is present almost exclusively in the aqueous phase. This means
that equilibrium constants based on total concentrations (involving both phases) can
increase by more than a factor of 10 000 compared to one-phase systems of either
medium (water or organic solvent) (Semenov et al., 1987).

9.6 “pH CONTROL” IN ORGANIC MEDIA

It is always important that the ionisation state of the enzyme is suitable for catalysis.
Since protonation and deprotonation of the enzyme seldom occurs to any appreciable
extent in non-conventional media, the ionisation state of the enzyme must be suitable
already before exposure to the organic medium. This is done by adjusting the pH value of
the aqueous enzyme solution used for immobilisation, lyophilisation or other procedures
for making the enzyme preparation. The enzyme keeps its ionisation state from the
aqueous solution. This has been called the “pH memory” of enzymes in organic media
(Zaks and Klibanov, 1985). To increase the buffering capacity of the system, buffer salts
are often present in the enzyme preparation. Even better buffering capacity can be
obtained by dissolving buffering substances in the organic phase. Examples of useful
 Applied biocatalysis 358
substances are trisoctylamine and triphenyl acetic acid (Blackwood et al., 1994). Even
though the buffering substances are primarily present in the organic phase, they are able
to control the pH in the aqueous phase.

9.7 APPLICATIONS

9.7.1 Ester Synthesis

Esters are common components in cosmetics and skin-care products. They can be
synthesized from fatty acids and alcohols using either chemical or enzymatic reactions.
The chemical reactions are normally catalysed by acid catalysts. Enzymatic synthesis is
carried out under milder conditions and therefore it provides products of very high purity.
A range of esters such as isopropyl palmitate and isopropyl myristate are now produced
industrially using enzymatic synthesis. The reactions are carried out in solvent-free
systems using an immobilised lipase as catalyst. In order to get high yields in the
reactions, water is removed continuously.

9.7.2 Synthesis of Chiral Building Blocks by Enzymatic Resolution

Enzymatic synthesis of chiral building blocks has been mentioned several times in this
book. A common synthetic route is to synthesize a racemic mixture which can be
resolved using a hydrolytic reaction catalysed by a hydrolase. The synthesis of glycidyl
butanoate (section 4.14) is a typical example. One main problem in the development of a
process of this kind is to find an enzyme with high enough stereoselectivity for the
desired resolution. An alternative to the screening of many enzymes is the screening of
many reaction conditions. The most complete screening is achieved if many enzymes are
tested under different reaction conditions. The limit is set by the number of experiments
which can be carried out. Anyway, the use of non-conventional reaction media makes it
possible to carry out reversed hydrolytic reactions and transferase reactions in a good
way and this widens the scope of enzymatic resolution considerably. In many cases the
stereoselectivity of the enzyme is quite different in an esterification reaction in an organic
medium compared to a hydrolytic reaction in water. There are no general rules to predict
which kind of reaction will give the highest stereoselectivity in a particular example.

9.7.3 Peptide Synthesis

Proteases can be used for the synthesis of peptides in a way analogous to the ester
synthesis catalysed by lipases. The most successful industrial example of enzymatic
peptide synthesis is described in section 4.6: aspartame synthesis. In the industrial
process in Europe the equilibrium position is shifted towards synthesis because the
 Biocatalyst in non-conventional media 359
reaction conditions (in aqueous solution) are chosen so that the product precipitates. An
alternative way is to use an organic solvent to shift the equilibrium position.
There are highly developed methods for chemical peptide synthesis, both solid phase
methods and solution methods. This makes it rather difficult for the enzymatic methods
to compete. However, the aspartame example shows that for peptides which have a big
market it can be worthwhile to develop an enzymatic process.

9.7.4 Production of Cocoa Butter Substitutes

Lipases can be used in transferase reactions to exchange fatty acids in fats. This is of
considerable interest to the food industry. The enzymatic production of cocoa butter
substitutes is the most well-known example. Cocoa butter is the fat component in
chocolate. It melts in the range between room temperature and body temperature because
its triglyceride molecules contain certain combinations of fatty acids. Natural

Figure 9.9 Schematic presentation of the conversion of 2-oleyl-1, 3-

dipalmitoylglycerol to 2 oleyl-3-palmitoyl-1-stearoylglycerol and 2-
oleyl-1, 3-distearoylglycerol in a lipase-catalysed reaction. The
triglycerides produced are the main components in cocoa butter.

cocoa butter is rather expensive and the supply varies considerably depending on the
harvest, so there is an incentive to prepare substitutes from cheaper, readily available raw
materials. Fats with melting behaviour virtually identical to that of natural cocoa butter
are now produced industrially starting from cheap palm oil fractions. The catalysts used
are lipases which are specific for the 1- and 3-positions in the triglyceride molecules
(Figure 9.9). Chemical catalysts do not show the same specificity. Other substances
which can be prepared enzymatically in non-conventional media are mono- and
diglycerides, fatty acids and phospholipids (Adlercreutz, 1994).

9.7.5 Enzymatic Synthesis of Chiral Hydroxynitriles

Chiral hydroxynitriles are useful synthetic intermediates. They can be prepared using
enzymatic synthesis in reactions between aldehydes or ketones and hydrogen cyanide
(Griengl et al., 1997) (Figure 9.10). There are different kinds of enzymes (hydroxynitrile
lyases) catalysing the formation of the (R)- and the (S)-enantiomers, respectively. It is a
problem that the reactants can react spontaneously as well,
 Applied biocatalysis 360

Figure 9.10 Enzymatic synthesis of (R)-mandelonitrile is complicated by the

spontaneous reaction which gives a racemic product.

because the product in the spontaneous reaction is racemic, and the desired product
should contain only one enantiomer (Figure 9.10). It is therefore important to suppress
the spontaneous reaction as much as possible. One way to achieve this is to use a low pH
value (around 3) but this reduces the enzyme activity. Another way is to carry out the
reaction in an organic solvent. A further advantage with an organic solvent is that it helps
to solubilise the hydrophobic aldehydes and ketones used as substrates.

9.8 QUESTIONS

9.1. What are the advantages and drawbacks with water quantification using
concentrations and water activities, respectively?

9.2. The catalytic activity of enzymes is often found to be considerably lower in

organic media than in aqueous solutions. This is normally caused by a combination of
different effects. Suggest a few possible effects !

9.3. You carry out a reaction in an organic solvent. The reaction is catalysed by a lipase
immobilised by adsorption on a porous support. The reaction virtually stops long before
equilibrium is reached and you suspect that this can be due to the formation of an acidic
reaction product. What can you do to increase the buffering capacity of the system?
 Biocatalyst in non-conventional media 361

9.4. In aqueous solution the substrate S has a solubility of 10 mM and the Km value is
5 mM. In order to obtain more concentrated substrate solutions in the reactor, organic
solvents are used.

a) A water-miscible solvent is used. What effect do you expect this to have on the
apparent Km value?
b) A water-immiscible solvent is used (equal volumes of water and organic solvent).
What effect do you expect this to have on the apparent Km value if the calculations
are based on substrate concentrations in the aqueous phase and if the substrate
concentrations are based on the total volume, respectively?

9.5. You have earlier used the enzyme E immobilised on support X. The reaction
worked well at a water content of 0.5 % (by vol.) in toluene. However, support X is no
longer available so you try to use support Y instead. Under the conditions found optimal
with support X the reaction rate is only 10 % of what it was before. Suggest possible
explanations!

9.6. The solubilities of water in a few different solvents at 25°C are as follows: hexane:
7 l/1; ethyl acetate: 30 ml/1; diisopropyl ether: 4 ml/1. Assume that the water activity is
proportional to the water concentration in each solvent (in reality there are large
deviations from this ideal behaviour). If you add 3.0 l water to 1.00 ml of each dry
solvent, which water activities will you get at equilibrium in a closed container without
gas phase?

9.7. Under which conditions in non-conventional media do enzymes normally have the
best stability?

9.9 HINTS AND ANSWERS

9.1. See the section “Quantification of water”.

9.2. A few different effects are discussed in this chapter. For a more complete
treatment of this topic, please see (Klibanov, 1997).

* The enzyme might work at too low water activity in the organic medium.
* The solvent might solvate the substrate too well so that the apparent Km value is
increased drastically.
* If a solid enzyme preparation is used, mass transfer may limit the reaction rate.
* The enzyme might have been partly inactivated during the catalyst preparation
(drying, etc.).
 Applied biocatalysis 362

9.3. Directly after adsorption of the enzyme on the support you can add aqueous buffer
which is then dried onto the support. Probably a better way is to add a buffer which is
soluble in the organic phase.

9.4

a. The apparent Km is expected in increase. See the section “Effects on enzyme

activity” in “Selection of solvent”.
b. Based on aqueous phase concentrations, the apparent Km is not expected to change
much unless the solvent dissolves to a large extent in the aqueous phase and works
as a water miscible solvent (see above) or if it acts as an inhibitor. If concentrations
are calculated for the total system, the apparent Km value can be expected to
increase.

9.5. The enzyme loading on support Y might not be suitable. There might have been
enzyme inactivation or mass transfer limitations. Support Y might bind more or less
water than support X. It is better to work at fixed water activity. There can be specific
effects reducing the activity of this particular enzyme on this particular support.

9.6. Hexane: 1.0 (a two-phase system is formed); ethyl acetate: 0.1; diisopropyl ether:
0.75.

9.7. In highly hydrophobic media and at low water activity. This is a general guideline
but there are individual variations.

9.10 REFERENCES AND SUGGESTED FURTHER READING

Halling (1994). A more thorough description of the fundamental principles of

biocatalysis in non-conventional media.
Koskinen and Klibanov (1996). More details concerning different modes to use
enzymes in organic media and several chapters on synthetic applications.
In the reference list below, there are several review articles and research articles on
specific topics.
Adlercreutz, P. (1994) Enzyme-catalysed lipid modification. Biotechnol. Gen. Eng. Rev. ,
12 , 231–254.
Bell, G., Halling, P.J., Moore, B.D., Partridge, J. and Rees, D.G. (1995) Biocatalyst
behaviour in low-water systems. Trends Biotechnol. , 13 , 468–473.
Blackwood, A.D., Curran, L.J., Moore, B.D. and Halling, P.J. (1994) “Organic phase
buffers” control biocatalyst activity independent of initial aqueous pH. Biochim.
Biophys. Acta , 1206, 161–165.
 Biocatalyst in non-conventional media 363
Brink, L.E.S. and Tramper, J. (1985) Optimization of organic solvent in multiphase
biocatalysis. Biotechnol. Bioeng. , 27, 1258–1269.
Dabulis, K. and Klibanov, A.M. (1993) Dramatic enhancement of enzymatic activity in
organic solvents by lyoprotectants. Biotechnol. Bioeng. , 41, 566–571.
Gill, I. and Vulfson, E. (1994) Enzymatic catalysis in heterogeneous eutectic mixtures of
substrates. Trends Biotechnol. , 12, 118–122.
Griengl, H., Hickel, A., Johnson, D.V., Kratky, C., Schmidt, M. and Schwab, H. (1997)
Enzymatic cleavage and formation of cyanohydrins: a reaction of biological and
synthetic relevance. Chem. Commun. , 1997, 1933–1940.
Halling, P.J. (1984) Effects of water on equilibria catalysed by hydrolytic enzymes in
biphasic reaction systems. Enzyme Microb. Technol. , 6, 513–516.
Hallling, P.J. (1994) Thermodynamic predictions for biocatalysis in non-conventional
media: theory, tests and recommendations for experimental design and analysis. Enz.
Microb. Technol. , 16, 178–206.
Inada, Y., Takahashi, K., Yoshimoto, T., Ajima, A., Matsushima, A. and Saito, Y. (1986)
Application of polyethylene-glycol-modified enzymes in biotechnological processes:
organic solvent-soluble enzymes. Trends Biotechnol. , 4, 190–194.
Khmelnitsky, Y., Levashov, A., Klyachko, N. and Martinek, K. (1988) Engineering
biocatalytic systems in organic media with low water content. Enzyme Microb.
Technol. , 10, 710–724.
Klibanov, A.M. (1997) Why are enzymes less active in organic solvents than in water?
Trends. Biotechnol. , 15, 97–101.
Koskinen, A.M.P. and Klibanov, A.M. (1996) Enzymatic Reactions in Organic Media .
London: Chapman & Hall.
Laane, C., Boeren, S., Vos, K. and Veeger, C. (1987) Rules for optimization of
biocatalysis in organic solvents. Biotechnol. Bioeng. , 30, 81–87.
Lamare, S. and Legoy, M.-D. (1993) Biocatalysis in the gas phase. Trends Biotechnol. ,
11, 413–418.
Margolin, A.L. (1996) Novel crystalline catalysts. Trends Biotechnol. , 14, 223–230.
Martinek, K., Klyachko, N.L., Kabanov, A.V., Khmelnitsky, Y.L. and Levashov, A.V.
(1989) Micellar enzymology: its relation to membranology. Biochim. Biophys. Acta ,
981, 161–172.
Nakamura, K. (1990) Biochemical reactions in supercritical fluids. Trends Biotechnol. ,
8, 288–292.
Okahata, Y. and Mori, T. (1997) Lipid-coated enzymes as efficient catalysts in organic
media. Trends Biotechnol. , 15, 50–54.
Otamiri, M., Adlercreutz, P. and Mattiasson, B. (1992) Complex formation between
chymotrypsin and ethyl cellulose as a means to solubilize the enzyme in active form in
toluene. Biocatalysis , 6, 291–305.
Paradkar, V.M. and Dordick, J.S. (1994) Aqueous-like activity of -chymotrypsin
dissolved in nearly anhydrous solvents. J. Am. Chem. Soc. , 116, 5009–5010.
Rekker, R.F. (1977) The hydrophobic fragmental constant . Amsterdam: Elsevier.
Russell, A.J. and Yang, F.X. (1996) Catalyze gas-phase reactions with enzymes.
Chemtech. , 1996, 24–27.
Semenov, A.N., Khmelnitsky, Y.L., Berezin, I.V. and Martinek, K. (1987) Water-organic
solvent systems as media for biocatalytic reactions: the potential for shifting chemical
equilibria towards higher yield of end products . Biocatalysis , 1, 3–8.
 Applied biocatalysis 364
Valivety, R.H., Halling, P.J. and Macrae, A.R. (1992) Reaction rate with suspended
lipase catalyst shows similar dependence on water activity in different organic
solvents. Biochim. Biophys. Acta , 1118, 218–222.
Wehtje, E., Adlercreutz, P. and Mattiasson, B. (1993) Improved activity retention of
enzymes deposited on solid supports. Biotechnol. Bioeng. , 41, 171–178.
Wehtje, E., Svensson, I., Adlercreutz, P. and Mattiasson, B. (1993) Continuous control of
water activity during biocatalysis in organic media. Biotechnol. Techniques , 7, 873–
878.
Zacharis, E., Omar, I.C., Partridge, J. and Robb, D.A. (1997) Selection of salt hydrate
pairs for use in water control in enzyme catalysis in organic solvents. Biotechnol.
Bioeng. , 55, 367–374.
Zaks, A. and Klibanov, A.M. (1985) Enzyme-catalyzed processes in organic solvents.
Proc. Natl. Acad. Sci. USA . 82, 3192–3196.
 10.
PROCESS CONCEPTS FOR
BIOCONVERSIONS
ADRIE J.J.STRAATHOF1 and VOLKER KASCHE2
1 Kluyver Laboratory for Biotechnology, Delft University of Technology,
Julianalaan 67, 2628 BC Delft, The Netherlands. Tel: +31–15–2782330; Fax:
+31–15–2782355; e-mail: [email protected]
2 Technische Universität Hamburg-Harburg, Denickestr. 15, D-21071
Hamburg, Germany.
Tel: +49–40–7718–2127; Fax: +49–40–771–821–27; e-mail: kasche@tu-
harburg.de

ABSTRACT
A 100% yield of an enzymatic reaction is usually not obtained easily. The
thermodynamic equilibrium may be unfavorable, competing reactions may
occur, or the co-reactants may require recycling by an additional reaction. Some
of the most important concepts that are used to optimize the yield are described
in this chapter. These concepts are of major importance for
• Reverse hydrolysis reactions
• Enantioselective reactions
• Redox reactions
In most of the applied bioconversions one or more of such reactions occur.

10.1 THERMODYNAMICALLY AND KINETICALLY CONTROLLED

REACTIONS
For biotechnological purposes enzymes are used as biocatalysts to accelerate a desired
reaction to its end point. As catalysts enzymes can catalyze a reaction in both directions.
Unfortunately the name of an enzyme given by the Enzyme Commission for the
classification of enzymes, and biochemistry text books, stresses the function as a catalyst
in one direction only Hydrolases, a group of enzymes of considerable importance in
 Applied biocatalysis 366
biotechnology, also catalyze the reverse condensation reaction, a function with important
synthetic applications. Proteases are used as catalysts for the condensation of peptides
with other peptides or amino acids (insulin semi-synthesis; synthesis of other peptide
hormones); lipases are used to catalyze the formation of lipids from alcohols and fatty
acids. In these processes the enzyme is used to increase the rate of reaction to the end
point of the reaction, which can be calculated from the equilibrium constant. The latter
cannot be influenced by the enzyme.
This may be illustrated by the following process, catalyzed by penicillin amidases (EC
3.5.1.1 1) from various sources

(10.1)

Figure 10.1 Product concentration as a function of time in thermodynamically

controlled (end point B) and kinetically controlled (end point A)
processes catalyzed by enzymes.
 Process concepts for bioconversions 367
In this reaction an acyl-enzyme is formed as a reactive intermediate. This is deacylated in
a nucleophilic attack on the carbonyl carbon atom in the ester linkage of the acylenzyme
(Enz−O−CO-CH2Ph) by H2O or other nucleophiles (R’-OH; R”-NH2).
When H2O deacetylates the acyl-enzyme, phenylacetic acid is formed. When
nucleophiles other than H2O deacylate the acyl-enzyme, a new condensation product, in
this case phenylacetyl-O-R’ or phenylacetyl-NH-R” is formed. By definition the
hydrolysis of these condensation products can be catalyzed by the same enzyme that
catalyzes their formation in equation 10.1. Thus, when the acyl-enzyme is formed from
phenylacetyl-glycine or phenylacetyl-O-Me, this gives rise to an alternative process to
produce Penicillin G, in addition to the thermodynamically controlled (= equilibrium
controlled) condensation of phenylacetic acid and 6-aminopenicillanic acid (6-APA).
This reaction that involves an activated side chain is a kinetically controlled (= rate
controlled) process where the hydrolase acts as a transferase (Kasche, 1986; 1989).
In Figure 10.1 the time course of thermodynamically and kinetically controlled
processes catalysed by biocatalysts are compared. The product yield at the maximum or
end point is influenced by pH, temperature, ionic strength, and the solubility of the
product. In the kinetically controlled process (but not in the thermodynamically
controlled process) the maximum yield also depends on the properties of the enzyme (see
next sections). In both processes the enzyme properties determine the time required to
reach the desired end point. The conditions under which maximum product yields are
obtained do not generally coincide with the conditions where the enzyme has its optimal
kinetic properties or stability. The primary objective is to obtain maximum yields. For
this aim it is not sufficient to know the kinetic properties of the enzyme as functions of
various parameters. It is also necessary to know how the thermodynamically or the
kinetically controlled maximum is influenced by pH, temperature and ionic strength, and
how this may be influenced by the immobilization of the biocatalysts on different
supports.

10.1.1 End-points of Thermodynamically Controlled Processes Catalyzed

by Biocatalysts
The product yield of a thermodynamically controlled reaction depends on pH when acids
and bases participate in the reaction. This pH-dependence can be analyzed using known
values of p K a -values of the acidic and basic groups of the reactants and the products.
For thermodynamically controlled processes the apparent equilibrium constant for the
product yield in condensation reactions, K app , must be determined. This equilibrium
constant is defined by the following equation:

(10.2)
 Applied biocatalysis 368
where the sum covers all ionic form present in the solution. It is related to K th , the
thermodynamic equilibrium constant (where all activities are unity), which is
independent of pH. Such a thermodynamic equilibrium constant can only be given for a
reference reaction equation. This is a reaction equation where the charges are not taken
into account, e.g. where all species are in their neutral form. When the neutral form does
not occur at the conditions of interest (as for 6-APA, which is either anionic, cationic or
zwitterionic) , it is more convenient to choose a reference reaction with the same
functional charged groups in both the substrates and the products, e.g. for penicillin
synthesis:

For a given pH, the pKa values of the substrates and products determine which fraction
(F, ranging from 0 to 1) will bear the charge of the reference state. These fractions relate
the apparent and true equilibrium constants.

(10.3)

Figure 10.2 Calculated yield of the thermodynamically controlled Penicillin G

 Process concepts for bioconversions 369
synthesis. For each curve the initial concentration of 6-aminopenicillanic acid
is given in the graph; the initial concentration of phenylacetic acid
has the same value. The values used in the calculation are given in
Question 1 in this chapter.

The yield of Penicillin G can be calculated from K app . (see Figure 10.2, and Question 1).
The optimum yield of a condensation product is obtained at the pH where Ka p p has a
maximum. For peptide synthesis with serine proteases this coincides with the pH where
the enzyme kinetic properties have their maxima. For the synthesis of penicillins with
penicillin amidase, or esters with serine proteases or esterases, the pH of maximum
product yield is much lower than the pH optimum of the enzymes. For penicillin amidase
the pH stability is also markedly reduced at pH 4–5. Thus, in these cases,
thermodynamically controlled processes for the synthesis of the condensation products
are not favorable. When these enzymes are used as catalysts in thermodynamically
controlled hydrolysis reactions an increase in pH increases the product yield. Penicillin
hydrolysis is generally carried out at pH about 8.0, where the enzyme has its optimum. At
this pH the equilibrium yield of hydrolysis product is about 97%. It could be further
increased by increasing the pH. Due to the limited stability of the enzyme and the product
6-aminopenicillanic acid at pH>8, a higher pH is not used in the biotechnological
process.
The temperature dependence of the equilibrium concentration of a product in a
thermodynamically controlled process is determined by the heat (enthalpy change) of the
catalyzed reaction. For an exothermic process an increase in temperature

Table 10.1 Thermodynamic data for some processes where enzymes are used as
biocatalysts to catalyze thermodyamically controlled processes. (End
point=equilibrium)

Process Enthalpy Process With increasing temperature the

change is yield
Hydrolysis of:
Penicillin G >0 endothermic increases
Penicillin V <0 exothermic decreases
Sucrose ? ? ?
Lactose >0 endothermic increases
Maltose <0 exothermic decreases
Isomaltose >0 endothermic increases
Cellulose <0 exothermic decreases
 Applied biocatalysis 370

Starch <0 exothermic decreases

Lipids ? ? ?
Peptides >0 endothermic increases
Acyl-amino acid >0 endothermic increases
Isomerization of >0 endothermic increases
glucose

leads to an increase in the reaction rate. The product yield, however, decreases with
temperature. For an endothermic process increasing the temperature results in a higher
product yield at the end point. This illustrates the need to know the thermodynamics
(enthalpy and free energy change) of the process in order to optimize the yields. For most
processes of biotechnological interest detailed thermodynamic data are still not available.
Some existing data are given in Table 10.1 (Kasche, 1986; Tewari, 1990).
In some cases, the process may require the use of high temperatures, even if this
reduces yield, such as in the hydrolysis of starch and cellulose.
In biotechnological processes involving ions at concentrations higher than normal
(>0.1 M), the influence of the ionic strength and the non-ideality of the solution on the
yield must also be considered. In case, where attractive ion-ion interactions are involved,
a higher ionic strength reduces the yield, whereas the opposite applies for repulsive
interactions (Kasche, 1986).

10.1.2 Maximum Yields in Kinetically Controlled Processes

The rate of the synthesis of Penicillin G from phenylacetyl-glycine and 6-amino-
penicillanic acid (6-APA, see Figure 10.3) decreases during the reaction because of the
decreasing activated side chain concentration. At the same time the rate of hydrolysis of
the product increases. This yields a time dependent maximum in the product
concentration—much larger than the final equilibrium concentration—as illustrated in
Figure 10.3. Such kinetically controlled processes are important in living systems in the
biosynthesis of condensation products (protein, nucleic acids, oligo- and polysaccharides)
at concentrations much greater than equilibrium concentrations. In biotechnology they
offer an alternative method for the synthesis of condensation products, where the
equilibrium constants are unfavorable for the
 Process concepts for bioconversions 371

Figure 10.3 Kinetically controlled ((i), (ii), and (iii)) and thermodynamically
controlled (ii) synthesis of penicillin G at pH 6.0 and 25°C (Ionic
strength I=0.2). Initial concentrations of substrates 10 mM. The
enzyme concentrations in U/ml are given above the kinetically
controlled maxima of each curve.

equilibrium controlled process (Kasche, 1986). Detailed analysis of kinetically controlled

processes have shown that the nucteophile must be bound in the active site of the enzyme
before it can deacylate the acyl-enzyme. Hydrolases that catalyze hydrolytic reactions
without the formation of a covalent intermediate (acyl-, phosphoryl-, glucosyl-enzyme)
 Applied biocatalysis 372
cannot act as transferases to catalyze kinetically controlled processes (exception:
RNAse). Thus, metallo-proteases (thermolysin) or carboxyl-proteases (pepsin) can only
be used to catalyze the synthesis of peptides in thermodynamically controlled processes.
The product yield and concentration at the kinetically controlled maximum depend on the
kinetic properties of the

Figure 10.4 Time dependence and maximal yield in kinetically controlled

synthesis of condensation products AN. (i) Product hydrolysis
rate=synthesis rate ( -lactam antibiotics); (ii) Product hydrolysis
rate << synthesis rate (peptide synthesis from amino acid esters with
proteases, whose esterase activity >> amidase activity).

enzyme (Figure 10.4). The ratio of the apparent rate constant for the transferase function,
k T , to the corresponding rate for the hydrolase function, k H , can by easily determined
from the initial rates of the formation of the transferase product (AN), and the hydrolase
product (AOH), v AOH .
 Process concepts for bioconversions 373

(10.4)

Equation 10.4 shows that this ratio increases with the nucleophile content. Experi-mental
data for some transferases and hydrolases are given in Table 10.2. Using these data and a
water concentration of 55 M, Equation 10.4 predicts that a high product yield, in
kinetically controlled processes catalyzed by hydrolases, require nucleophile
concentrations >0.1 M. These are much higher than substrate concentrations in vivo.
The pH-dependence of the maximum product concentration in kinetically

Table 10.2 The ratio of the apparent transferase to hydrolase rate constants (k T /k H )app,
for some enzymes that can be used for kinetically controlled synthesis.

Enzyme nucleophile (k T /k H )
2=TRANSFERASES
DNA-polymerase DNA 107
hexokinase glucose 106
3=HYDROLASES
lipase alcohols 10
alkaline phosphatase TRIS 102–103
RNAse I alcohols 10
nucleosides 102–103

glycosidases ( -galactosidase) alcohols 102

lactose 10–102
serine-, thiol-proteases amino acids 10–102
amino acid esters 102–104
amino acid amides 102–105
D-amino acid amides 10–103
alcohols −102
TRIS −102
amidases (penicillin amidase) 6-aminopenicillanic acid 103–104
amino acids −103
 Applied biocatalysis 374

alcohols 1

controlled processes is not easily determined from the relationships in Figure 10.4.
Generally the pH-dependence is difficult to analyze. The pH optimum must be
determined experimentally. It must be higher than the pKa-value of the nucleophile,
otherwise the nucleophile will be protonated. For the kinetically controlled synthesis of
peptides pH values >9 have been shown to be optimal. The corresponding values for -
lactam antibiotics have been found to be in the range 5–8.
For the temperature dependence of the maximum yield in kinetically controlled
processes, detailed data are generally not available. In all cases, where this has been
studied experimentally, a decrease in the maximum product yield with increasing
temperature is observed (Kasche, 1986).
Several kinetically controlled processes are already used on an industrial scale, such as
the conversion of porcine or recombinant proinsulin to human insulin and the conversion
of sucrose to palatinose. In the field of -lactam antibiotics (cephalosporins and
penicillins), the semisynthesis of products with side chains other than phenylacetyl, such
as D-phenylglycyl (ampicillin, cephalexin) or other -lactam structures are currently being
developed (Bruggink, Roos and De Vroom, 1998).

10.2 ASYMMETRIC SYNTHESIS AND KINETIC RESOLUTION

In Chapter 2, the two methods of enantioselective biocatalysis have been introduced:

Asymmetric synthesis and kinetic resolution.
Asymmetric synthesis starts with a prochiral compound. This is a compound which is
not chiral, but can be converted into a chiral compound by a chiral (bio) catalyst.
Subsequently, two types of prochiral compounds can be distinguished: The first one has a
stereoheterotopic face (which usually is a double bond) to which an addition reaction
takes place. An example is the conversion of the prochiral compound propene into 1,2-
epoxypropane (which has two enantiomers, of which one may be preferentially formed
using an enantioselective catalyst). The second type of prochiral compound has two so-
called enantiotopic atoms or groups. If one of these is converted, the compound becomes
chiral. Meso-compounds belong to this class. Figure 10.5 and 10.6 show some examples
of the different types of asymmetric catalysis with prochiral compounds.
Enantioselective catalysis by kinetic resolution does not start with a prochiral
compound but with a mixture of two enantiomers. The chiral catalyst should convert one
of the substrate enantiomers, leaving the other one pure. When the enantioselectivity of
the catalyst is absolute, the remaining substrate as well as the product may reach an
enantiomeric excess of 100%. So the remaining substrate as well as the product may be
the target compound, but the maximum yield for either one is only 50%. Therefore
 Process concepts for bioconversions 375
racemization and/or recycling of the other 50% is required.

10.2.1 Enantiomeric Excess and Yields

Usually reaction engineering concentrates on maximizing the yield and minimizing the
residence time in a reactor. For enantioselective catalysis processes, maximizing the
(enantiomeric) purity if even more important. The chemical purity of a chiral compound
AR is defined as the percentage AR+AS in the product. A high chemical purity is not
essential, since contaminations may be removed by conventional purification techniques
in additional steps. More important is the enantiomeric purity (enantiopurity), which
indicates the extent of contamination by the distomer AS. Removal of this enantiomer is
very troublesome. Enantiomeric contamination of a chiral starting compound of a
synthetic route is likely to result in enantiomeric contamination of the end-product.
Enantiomeric purity is most often expressed as enantiomeric excess (ee); for AR it is
defined by:

(10.5)

and are the concentrations of the enantiomers of compound A. The enantiomeric

excess is 0 (0%) for a racemate and 1 (100%) for an enantiopure R-compound. In this
definition for enantiopure S-compound, then
For commercial application, an enantiomeric excess of >95% is required, but in
many cases >99% is desirable.
When a kinetic resolution process is carried out, and no racemization occurs, one can
derive from the mass balances of R- and S-enantiomers that there is a relation
Applied biocatalysis 376
 Process concepts for bioconversions 377

Figure 10.5 Some examples of the formation of chiral compounds from

prochiral starting substrates with a double bond.

Figure 10.6 Example of the fonmtion of chiral compound from prochiral meso-
substrates.

between the yields and the enantiomeric excesses of remaining substrate and product:

(10.6)

This can be checked using the relations between the yields and extent of conversion X

(10.7)

Note that in process technology the symbol X or is used to indicate the extent of
conversion (see Chapter 11), whereas in bioorganic chemistry the symbol c is common
(see Chapter 2).
In order to extend the comparison of enzymatic kinetic resolution and asymmetric
catalysis of chapter 2 and identify the potential improvements of these processes, the
derivation of the equations that determine the enantiomeric excess and yield are given in
detail.

10.2.2 Kinetics of Enzymatic Kinetic Resolution

If two enantiomers, AR and As, compete to form a Michaelis complex with the enzyme,
the rate equations can be written as
 Applied biocatalysis 378

(10.8)

(10.9)

In a batch reactor, the ratio of conversion rates of the enantiomers is given by

(10.10)

This can be simplified to

(10.11)

with

(10.12)

This is the definition of the enantiomeric ratio E, which indicates the selectivity of the
enzyme for the fast reacting enantiomer (AR) relative to the slowly reacting enantiomer
(As).
Integration with the boundary condition when gives an expression
which relates the concentration of the two enantiomers.
 Process concepts for bioconversions 379

(10.13)

Since it is customary to use the dimensionless quantities X and ee rather than

concentrations, this equation is modified. Usually the kinetic resolution is started with a
racemate, so , and using Eqs. (10.5) and (10.7) this can be transformed into the
Chen-equation for substrate (Chen et al., 1982) (cf Chapter 2):

(10.14)

Figure 2.8 shows the enantiomeric excess of the remaining substrate As as a function of
the conversion, for several values of the enantiomeric ratio, according to the Chen
equation. When the selectivity of the enzyme is high (E=100), the remaining substrate is
enantiopure from 50% conversion on. For lower selectivities, the yield of enantiopure
substrate is lower. But sooner or later an enantiomeric excess of virtually 100% will be
reached for the remaining substrate. At the moment that a sufficiently high ee has been
obtained, the reaction should be terminated, because otherwise this enantiopure substrate
will be converted further until racemic product remains.
When one is interested in the enantiomeric excess of the product the situation is
different. Another Chen-relation must be derived, by substituting the relation between the
enantiomeric excess of substrate and product that has been given before. This leads to

(10.15)

The enantiomeric excess of the product according to this second Chen equation (Figure
2.8) shows a very unpleasant feature: The first percentages of product already show an
enantiomeric excess <100%, and during the conversion a decrease occurs. Only for very
high values of E 100% enantiomeric excess is obtained. Typical practical values of E are
5–50. When >98% enantiomeric excess is desired, it is therefore better to concentrate on
compounds which are remaining substrates of a reaction than on compounds which are
products.
To improve the kinetic resolution, the E-value can be improved by changing the
enzyme, substrate, or solvent, e.g. (see section 2.3.6) or the process can be modified (see
section 10.2.4).
 Applied biocatalysis 380

10.2.3 Kinetics of Asymmetric Synthesis from Prochiral Compounds

The kinetics of the enzymatic conversions of a prochiral compound are simple. A single
Michaelis complex with the prochiral substrate is formed. This complex may react to
product PR with maximum rate or to PS with . Thus the enantiomeric ratio is a

(10.16)

The enantiomeric excess of the product in a batch reactor now can easily be calculated
because :

(10.17)

This means that the enantiomeric excess of the product is at a constant value, which is
independent of the conversion. Therefore, the reaction should be allowed to proceed to
full conversion, in order to obtain 100% yield. If the enantiomeric ratio is high enough,
this kind of process is very advantageous and should be prefeffed to a resolution process
which has a maximum of 50% yield. But suppose E= 20, then , which is
useless and cannot be improved easily, as will be shown in the next section. In such a
case a kinetic resolution process is much more flexible, because then ee=100% always
can be reached.

Table 10.3 Effect of deviations of the standard enantioselective enzymatic conversion on

the enantiomeric excess. The standard conversion is a single irreversible batch
reaction in a homogeneous solution starting form racemic or prochiral
substrate. (+)=positive effect, (−)=negative effect, (o)=no effect.

Effect on ee s Effect on ee p Effect on ee p in Reference

Deviation from in kinetic in kinetic asymmetric
normal resolution resolution synthesis
situation
equilibration − − − Chapter 2
double resolution + + o (Guo, 1993)
 Process concepts for bioconversions 381

(product
recycling)
background o − − (Straathof and
reaction Jongejan, 1997)
racemization − (substrate rac.) + (substrate rac.) − Chapter 2
o (product rac.) - (product rac.)
diffusion − − o (Straathof and
limitation Jongejan, 1997)
substrate ± ± o (Straathof et al.,
dissolution 1998)
plug flow reactor 0 0 0 (Rakels et al.,
1994)
CSTR/fed batch − − 0 (Rakels et al.,
reactor 1994)

10.2.4 Influence of Process Strategy on the Enantiomeric Excess

Usually the enantiomeric excess is calculated for a standard conversion process: a single
irreversible batch reaction in a homogeneous solution starting from racemic or prochiral
substrate. However, if the assumptions that were used for the derivation of Eqns. (10.14),
(10.15) and (10.17) do not hold, different equations apply, and the enantiomeric excess
may be higher or lower. Table 10.3 shows an overview of some modifications, including
some potential improvements (substrate racemization) and problems (equilibration) that
were treated in Chapter 2. Clearly, many modifications will lead to a decrease rather than
to an increase of the enantiomeric excess.
In multiphase reactors, e.g. immobilized enzyme reactors or reactors with partly
undissolved substrate, one has to account for mass transfer limitations that may occur
(see chapter 11). Mass transfer limitations slow down the reaction rate, and this
retardation is more serious for the fast-reacting than for the slowly-reacting substrate
enantiomer. The result is that the enzyme actually “sees” a relatively large proportion of
the slowly-reacting enantiomer. If the enzyme gets the opportunity to convert much of
this enantiomer, the enzyme seems to be not very enantioselective and the performance of
a kinetic resolution becomes poor. In order to obtain the maximum yield of an
enantiopure compound in a kinetic resolution process, one should avoid mass transfer
limitations, e.g. by using smaller immobilized enzyme particles or by increasing the
stirring rate. If the mass transfer cannot be increased easily, one may simply decrease the
enzyme reaction rate by using less enzyme.
 Applied biocatalysis 382

Table 10.4 Function of some coenzymes in enzymatic reactions.

Coenzyme Reaction
NAD, NADP, FAD Redox (electron transfer)
UDP Glucose transfer
ATP Phosphoryl transfer
TTP (thiamine) Aldehyde transfer
CoA Acyl transfer
SAM Methyl transfer

It can easily be understood that mass transfer limitations will not affect the enantiomeric
excess that can be reached in an asymmetric synthesis, but just slow down the reaction.

10.3 COENZYME REGENERATION

Many reactions require coenzymes, e.g. NAD+/NADH is required for many

oxidations/reductions (see Table 10.4). An example is the reduction of the steroid
cortisone to 20-dihydrocortisone with the enzyme 20- -hydroxysteroid dehydrogenase.
A stoichiometric amount of NADH is consumed by this reaction. Because of the high
price of NADH (about $27000/kg), the NAD+ that is formed has to be regenerated to
NADH by a second reaction. With a regeneration system that is able to carry out n
regenerations, the cost of NADH drops to $27000/n * (M NADH/M product) per kg of
product (M is the molecular mass). So then n >>> 1000 is required. In general, the
demands for a good regeneration system are:

1. The source of reducing equivalents should be cheap.

2. The thermodynamics should be favorable ( rG<0)
3. The kinetics should be favorable (the regeneration should be fast and the regeneration
reaction should be selective, thus requiring a suitable catalyst)
4. The stability should be good (e.g., NADH should not decompose during the
regeneration).

Regeneration reactions may be (Simon et al., 1985):

• (electro) chemically catalyzed

• enzymatically catalyzed
• “catalyzed” by living cells

These possibilities will be treated subsequently.

 Process concepts for bioconversions 383

10.3.1 Electrochemical Regeneration

When electrochemical regeneration is used, the regeneration reaction equation is:

Electrons are cheap, and by adjusting the voltage of the electrode, a favorable overall
rG can be obtained. Besides, the method in principle is clean, because no other reagents
are added which have to be removed after the reaction. The electrode surface has to act as
the (heterogeneous chemical) catalyst. Research has led to sufficiently active but not to
sufficiently selective carbon electrodes. Some dimerization of NAD occurs, and this
limits the number of regenerations n.
By using a mediator such as methylviologen (MV) this situation can be improved.
Methylviologen is reduced selectively on the electrode surface, and subsequently used by
an enzyme (e.g. lipoamide dehydrogenase) to reduce NAD+:

Instead of using an electrochemical and enzymatic reaction simultaneously for the

regeneration, one can also use an enzymatic reaction only, as shown below.

10.3.2 Enzymatic Regeneration

An example of coenzyme regeneration with isolated enzymes is L-alanine production
from pyruvate with an NADH-dependent alanine dehydrogenase (AlaDH):

A regeneration reaction for NADH that may be used is the conversion of pyruvate with a
NADH-dependent lactate dehydrogenase (LDH):

These two reactions have to be carried out simultaneously, so the overall reaction will be
 Applied biocatalysis 384

The advantage of this reaction is that lactate is much cheaper than pyruvate, so the
reaction is not only useful for the regeneration of NADH but also for the production of
pyruvate. The disadvantage is that lactate dehydrogenases are enantiospecific, so a D-
lactate as well as an L-lactate dehydrogenase are required for a full conversion of racemic
lactate (Figure 10.7).

Figure 10.7 L-Alanine production from DL-lactate via pyruvate with cofactor
regeneration (Wandrey, 1984).

The thermodynamics of the reaction are favorable if a large excess of ammonium is used.
In a batch reactor the number of regenerations of NADH may be very high, but then the
reaction should be started with a very small amount of it and the reaction rate would be
very low. A sufficiently high rate is obtained if the concentration of NAD(H) is around its
K m value, but at those concentrations it will be necessary to recover it after the reaction.
A solution to this problem is the enzyme membrane reactor (Figure 10.8). This is a
kind of CSTR (continuous stirred tank reactor), with retains the enzyme and the cofactor
using an ultrafiltration membrane. This membrane has a cut-off of about 10000. Enzymes
usually have a molecular mass of 25000–250000, but the molecular mass of NAD(H) is
much too low for retention. Therefore it is first derivatized with polyethylene glycol
(PEG 20000). The reactivity of NAD(H) is hardly affected by the derivatization with this
soluble polymer. Alanine can now be produced continuously by high concentrations of
both enzymes and of NAD (H) in this reactor.
In order to obtain a sufficiently large membrane area on an industrial scale, hollow
 Process concepts for bioconversions 385
fiber ultrafiltration modules are applied (similar types as used for protein purification).
There are many other examples of cofactor regeneration reactions and/or of reactions
which may be performed in an enzyme membrane reactor. An important example is the
regeneration of NADH by formate dehydrogenase (FDH), starting with formate
(Wichmann et al., 1981). The advantage of this reaction is that it is irreversible because
carbon dioxide is liberated, while formate is a relatively cheap electron donor.

Figure 10.8 Enzyme membrane reactor for production of L-alanine from DL-
lactate (Wandrey, 1984).

Since FDH from Candida boidinii now can be produced at pilot scale, this reaction can
be generally used for NADH-regeneration. Recently, the same concept has been used for
NADPH regeneration. This became possible because a NADPH-dependent FDH has
been obtained by multipoint site-directed mutagenesis of the gene coding the enzyme
from the bacterium Pseudomonas sp. 101. (Seelbach et al., 1996).
For ATP regeneration, a similar concept has been used (Berke et al., 1988). ATP can
be regenerated from ADP using acetyl phosphate and the enzyme acetate kinase, upon
release of acetate. The reaction is irreversible and acetyl phosphate is a relatively cheap
phosphate donor. Thus, in an enzyme membrane reactor, PEGderivatized ATP was
consumed by a phosphorylase or a synthetase in a reaction leading to a product of
interest, and the ATP was regenerated by the acetate kinase (Figure 10.9).
 Applied biocatalysis 386

Figure 10.9 Reaction scheme for the synthesis of glucose-6-phosphate.

10.3.3 Regeneration with Living Cells

Although regeneration of coenzymes using isolated enzymes is feasible, it may not
always be worthwhile to find and produce the required enzymes, and to build a dedicated
reactor. Instead, regeneration is often carried out with living cells, requiring only one
fermentation to obtain the desired cells. Then, regeneration can be carried out with a
cheap substrate and the enzymes present in the whole cells, such as alcohol
dehydrogenase in the following reaction:

Usually a large number of other reactions will occur simultaneously, some of them being
beneficial for the coenzyme regeneration, whereas others lead to undesired byproducts.
Also, the substrate and product of the main reaction may get involved in undesirable side-
reactions. Therefore, whole cell reactions may be cheaper and simpler to carry out than
reactions using isolated enzymes, but they are less easily controlled, less reproducible and
yield more waste. The most widely studied class of reactions using whole cells are
reductions catalyzed by baker’s yeast, which is cheap and widely available (Sybesma et
al., 1998).

10.4 QUESTIONS AND HINTS

Question 1

At pH 5, 25°C, 57.8 mM benzylpenicillin (A−) is hydrolyzed to an equilibrium mixture of

 Process concepts for bioconversions 387

17.1 mmol/1 6-APA (P ± and its conjugated acid P−) and 17.1 mmol/1 phenylacetic acid
(Q° and its conjugated base Q−) by the enzyme penicillin amidase.

The pH is kept at 5 during the reaction by the addition of sodium hydroxide.

Clearly, it is also possible to synthesize benzylpenicillin by the reverse reaction. Two
(relevant) acid dissociation equilibria are involved:

(10.18)

(10.19)

(a) What would be the optimum pH for synthesis from a thermodynamic point of
view? And what would be the equilibrium conversion, when the reaction is started with
0.2 mol/1 6-APA and 0.4 mol/1 phenylacetic acid?

(b) At what pH must the hydrolysis of 0.2 mol/1 penicillin G be carried out to reach
>90% hydrolysis yield?

Question 2

The compound (S)-2-chloropropionic acid is an important building block for herbicides.

Suppose two enzymatic reactions are available for its production:

(a) Racemic 2-chloropropionic acid may be used as the substrate; a dehalogenase

preferentially converts the (R)-enantiomer into lactic acid (E=10).

(b) Racemic ethyl ester of 2-chloropropionic acid may be used as the substrate; a lipase
preferentially converts the (S)-enantiomer into 2-chloropropionic acid (E=75).

The required enantiomeric excess of (S)-chloropropionic acid is 98%. Which of the

two reactions will give the highest yield of the target compound? Both enzymatic
reactions can be described with irreversible Michaelis-Menten kinetics.
 Applied biocatalysis 388

Solution question 1a

Options for the reference reaction are

The equilibrium constant for the latter reaction is calculated here, although the same final
result will be obtained if one of the other reactions is taken.

(10.20)

The fractions are calculated from the dissociation equilibria. At pH 5:

(10.21)

(10.22)

Since at equilibrium 0.0171 mol/1 6-APA and phenylacetic acid are present

(10.23)

For the reverse reaction, C A is to be maximized. Thus the minimum of is to be

found from
 Process concepts for bioconversions 389

(10.24)

A minimum is reached at

(10.25)

Thus the optimum pH for synthesis is

(10.26)

When the reaction is started with c P0 =0.2 mol/1 and CQ 0 =0.4 mol/1 the equilibrium
concentration of P can be calculated from the equilibrium constant equation with the
actual values of the fractions:

(10.27)

Since and are both 0.387 at pH 4.40, cp = 0.00448 mol/1 at equilibrium according
to this equation. The conversion of P is 98%.

Solution question 1b

If the yield for hydrolysis of 0.2 mol/1 penicillin G should be >90%, the yield of its
synthesis (which is the reverse hydrolysis) should be <10%. From Figure 10.2 one can
read that for initial concentrations of 0.2 mol/1 this yield will be obtained at pH >7.5.

Solution question 2

In the first reaction, the remaining substrate is the target compound. The Chen formula
for remaining substrate should be used, substituting E=10 and ee=0.98:
 Applied biocatalysis 390

(10.28)

The solution of this equation is X=0.68. This value can be found iteratively, or by writing
X in this equation explicitly, or from Figure 2.8. If 68% is converted, only 32% substrate
remains which is sufficiently enantiopure.
In the second reaction, the product is the target compound. The Chen formula for
formed product should be used, substituting E=75 and ee=0.98:

(10.29)

The solution of this equation is X=0.12, which can also be estimated from Figure 2.8 .
This means that 12% of product can be formed which is sufficiently enantiopure.
In conclusion, the first reaction yields much more (S)-2-chloropropionic acid, although
the enantiomeric ratio of this reaction is much lower.

10.5 REFERENCES AND FURTHER READING

Berke, W., Schüz, H.J., Wandrey, C., Morr, M., Denda, G. and Kula, M.R. (1988)
Continuous regeneration of ATP in enzyme membrane reactor for enzymatic
syntheses. Biotechnol. Bioeng. , 32, 130–139.
Bruggink, A., Roos, E.C. and De Vroom, E. (1998) Penicillin acylase in the industrial
production of -lactam antibiotics. Organic Process Research & Development , 2, 128–
133.
Chen, C.S., Fujimoto, Y., Girdaukas, G. and Sih, C.J. (1982) Quantitative analyses of
biochemical kinetic resolutions of enantiomers. J. Am. Chem. Soc. , 104, 7294–7299.
Guo, Z.W. (1993) Novel plots of data from combined multistep enzymatic resolutions of
enantiomers. J. Org. Chem. , 58, 5748–5752.
Kasche, V. (1986) Mechanism and yields in enzyme catalysed equilibrium and
kinetically controlled synthesis of -lactam antibiotics, peptides and other condensation
products. Enzyme Microb. Technol. , 8, 4–16.
Kasche, V. (1989) Proteases in peptide synthesis. In Proteases a Practical Approach ,
edited by U.Bond and R.Beynon, pp. 125–143. Oxford: IRL Press.
Rakels, J.L.L., Paffen, H.T., Straathof, A.J.J. and Heijnen, J.J. (1994) Comparison of
enzymatic kinetic resolution in a batch reactor and a CSTR. Enzyme & Microbial
Technology , 16, 791–794.
Seelbach, K., Riebel, B., Hummel, W., Kula, M.R., Tishkov, V.I., Egorov, A.M., et al.
(1996) A novel efficient regeneration method of NADPH using a new formate
dehydrogenase. Tetrahedron Lett. , 37, 1377–1380.
 Process concepts for bioconversions 391
Simon, H., Bader, J., Guenther, H., Neumann, S. and Thanos, J. (1985) Chiral
compounds synthesized by biocatalytic reductions. Angew. Chem., Int. Ed. Engl. , 24,
539–553.
Straathof, A.J.J. andJongejan, J.A. (1997) The enantiomeric ratio—origin, determination
and prediction. Enzyme & Microbial Technology , 21, 559–571.
Straathof, A.J.J., Wolff, A. and Heijnen, J.J. (1998) Solid-to-solid kinetic resolution—
determination of the enantiomeric ratio. Journal of Molecular Catalysis B-Enzymatic ,
5, 55–61.
Sybesma, W.F.H., Straathof, A.J.J., Jongejan, J.A., Pronk,J.T. and Heijnen, J.J. (1998)
Reductions of 3-oxo esters by baker’s yeast: Current status. Biocatalysis and
Biotransformations , 16, 95–134.
Tewari, Y.B. (1990) Thermodynamics of industrially-important, enzyme-catalyzed
reactions. Appl. Biochem. Biotechnol. , 23 , 187–203.
Wandrey, C. (1984) Production of L-amino acids from a-hydroxy acids. Biotech.
Europe , 391–404.
Wichmann, R., Wandrey, C., Bückmann, A.F. and Kula, M.R. (1981) Biotechnol.
Bioeng. , 23, 2789–2802.
 11.
BIOREACTOR DESIGN
JOAQUIM M.S.CABRAL1 and JOHANNES TRAMPER2
1 Laboratório de Engenharia Bioquímica, Centro de Engenharia Biológica e
Química, Institute Superior Técnico, 1000 Lisboa, Portugal Telephone: +351 1
8419063;
Fax:+351 1 8419062; Email: [email protected]
2 Food and Bioprocess Engineering Group, Wageningen Agricultural
University, PO Box 8129, 6700 EV Wageningen, The Netherlands. Telephone:
+31 317 483204;
Fax: +31 317 482237; Email: [email protected]

ABSTRACT
This chapter describes the different types of batch and continuous bioreactors.
The basic reactor concepts are described as well as the respective basic
bioreactors design equations. The comparison of enzyme reactors is performed
taking into account the enzyme kinetics. The modelling and design of real
reactors is discussed based on the several factors which influence their
performance: the immobilized biocatalyst kinetics, the external and internal
mass transfer effects, the axial dispersion effects, and the operational stability of
the immobilized biocatalyst.

11.1 INTRODUCTION TO BIOREACTOR DESIGN

11.1.1 Defining the Subject

Bioreactor design is an integral part of biotechnology. Especially when designing
bioreactors, integration of biological and engineering principles is essential. The
bioreactor should be designed such that specific biological and technological demands of
a process are met. Naturally, quality and price of the product are decisive for commercial
realization. The aim of bioreactor design can thus be defined as “minimization of the
costs of the pertinent product while retaining the desired quality, and this within the
biological and technological constraints.” This does not mean a priori that minimizing
 Bioreactor design 393
the costs of the bioreactor also means minimizing the costs of the integral process. This
depends largely on the cost-determining part(s) of the process. If running the bioreactor is
cost determining, then maximization of the overall volumetric productivity of the
bioreactor is, in general, the rational approach. If, on the other hand, downstream
processing is cost determining, then maximization of the product concentration in the
bioreactor is, in general, the rational thing to do. However, here again integration is the
keyword. Bioreactor design should be an integral part of the overall process design. In the
following sections of this chapter the bioreactor will be defined with respect to reactor
concepts and types and to tools in bioreactor design. In the bioreactor the actual
conversion is accomplished by the biocatalyst. In this chapter biocatalyst means either an
enzyme, an enzyme complex, a cell organelle or a non-viable whole cell. Furthermore, a
biocatalyst can be free or immobilized, which has far-reaching consequences with respect
to mass transfer. Integration of mass transfer and biokinetics is essential in the description
(microkinetics) of immobilized biocatalysts. The source of biocatalysts can be of either
microbial, plant of animal origin.

11.1.2 Productivity and Product Concentration

Overall volumetric productivity

Overall volumetric productivity Qp (mol.m−3s− 1) (it is also common to use a yearly

basis) is the average production capacity per unit volume and time of the bioreactor. The
overall volumetric productivity is confined, on the one hand, by physical constraints,
such as mass and heat transfer, and, on the other hand, by biocatalyst concentration Cx
(mol.m−3) and activity of the biocatalyst, expressed as substrate consumption rate
(mol.m−3.s−1). Maximization of the overall volumetric productivity of the bioreactor in
principle means minimization of the costs of investment, because one can suffice with
smaller equipment. It usually also means lower operating costs. In general, it means too
that it is desirable to operate the bioreactor as close as possible to the physical constraints,
the horizontal dotted line in Figure 11.1.
 Applied biocatalysis 394

Figure 11.1 Constraints of overall volumetric productivity. (Adapted from

Cooney, 1983).

This physical limitation is the result of mass and heat transfer limitations, which are
stoichiometrically related to product formation. The vertical dotted line in Figure 11.1
symbolizes the limitation which is a consequence of the fact that the concentration of the
biocatalyst is bound to certain defined limits, for instance solubility in case of isolated
enzymes and “space” in case of suspended cells. Figure 11.1 also shows that the
biocatalyst should have a minimum specific activity to be able to operate the bioreactor
close to its physical ceiling.

Overall biocatalyst productivity

In addition to limitations by mass and heat transfer and concentration of biocatalyst, the
overall volumetric productivity of the bioreactor is determined by the overall productivity
of the biocatalyst, Prpx (−), defined as the total moles of product which are produced by 1
mol of biocatalyst during its operational lifetime t1(s). Prpx is related to the specific
product production rate qp (s−1) (moles of product produced per mol of biocatalyst per
second) as:
 Bioreactor design 395

(11.1)

The definition of the overall yield of product on substrate (total moles of product
produced per total mol of substrate), leads to:

(11.2a)

(11.2b)

The time needed to empty, clean, refill, restart, etc., the bioreactor between two
operations is the so-called down-time, which is symbolized by td (S). In case td is
relevant it can be introduced in Equation (2) by replacing 1/t1, preceding the integral, by
1/(t1+td). In addition to the molar productivity used above, the mass productivity (kg
product instead of mol) is also quite commonly used in engineering (conversion from one
to the other by means of the molecular weights). It is also common practice to use hour,
day or year as unit of time.
The search for and the development of a useful biocatalyst with a suitable yield,
specific activity and stability is, in the beginning, the task of microbiologists,
biochemists, molecular biologists, protein engineers, etc. However, especially with
respect to stability, the process engineer also has means available, among others
immobilization, to improve the stability of biocatalysts.

Product concentration

The effect of the composition of the product stream leaving the bioreactor on the costs of
the downstream processing is large. Therefore, it is essential to take this into account
when designing the bioreactor and when selecting the feed stock. This often means in
practice that the bioreactor is designed such that the concentration of product is as high as
possible. The end concentration of product Cp (mol.m−3) in the bioreactor depends on
, and the residence time in the bioreactor. For a batch reactor, with tb(s) as the
time that the batch lasts, this leads to:

(11.3)
 Applied biocatalysis 396

and for a continuous reactor with a liquid throughflow F1 (m3.s−1) and a volume V (m3):

(11.4)

Concentration of product is especially a key-parameter when downstream processing is

the cost-determining part of the integrated process. Product recovery is often a laborious
and expensive operation, especially when diluted aqueous solutions are involved, such as
we usually encounter in biotechnology. However, it has become clear that the aqueous
reaction medium, which was for a long time supposed to be essential for biocatalysis, can
be replaced to a large extent by a suitable organic solvent (Laane, de Bont and Tramper,
1987). Obviously the purity of the raw materials also determines downstream processing
to a large extent. The choice of the industrial substrate should therefore also be a rational
one, considering the integrated process.

11.1.3 Bioreactor Types

The stirred vessel

In Figure 11.2 a schematic view of a stirred vessel is given. The vessel is cylindrical with
a height Hv (m) and a diameter Tv (m). Usually Hv is equal to or greater than 2 Tv. It is
equipped with a stirrer in the lower compartment. This stirrer is mounted near the bottom,
usually at a distance equal to the stirrer diameter. At a lower position the stirrer and
bottom interact, leading to a decrease in power consumption. At a higher position liquid
circulation problems can occur because, at increased gas flow rate in case of aeration, the
bubbles will not be recirculated in the lower compartment. Sometimes the upper
compartment (s) are also equipped with a stirrer. The vessel is equipped with baffles to
prevent rotation of the contents as a whole. For aeration an air sparger is mounted below
the stirrer. For mass transfer
 Bioreactor design 397

Figure 11.2 Schematic representation of a strirred bioreactor.

its construction is generally not relevant, so it is chosen on the basis of sterility and
cleaning criteria.
Figure 11.3 shows a number of stirrers that are used. The turbine stirrer, being easy to
construct and having a high power number, is the most widely used. The other types are
less intensively applied. A detailed description of all types of stirrers can be found in
Zlokarnik (1972).
 Applied biocatalysis 398

Figure 11.3 Schematic representation of a number of stirrers. (Adapted from

Zlokarnik, 1972).

Figure 11.4 Schematic representation of the continuous-flow stirred-tank

reactors.

Information about special design considerations like stirrer drives, sealings, and also self-
aerating stirrers can be found in Sittig (1983). Of all reactors the stirred-tank reactor is the
most versatile to carry out different jobs.

The continuous-flow stirred-tank reactor (CSTR)

As its name suggests, this is a refinement of the simple stirred-tank reactor which has
been adapted so that product can be withdrawn and fresh substrate added on a continuous
basis. The main modification is the introduction of two ports to provide for fluid
transfers, but further controls are needed to ensure that the inflow and outflow are
balanced (Figure 11.4). The requirements for efficient mixing are similar to those already
 Bioreactor design 399
discussed but with the added restriction that streaming of substrate straight through the
reactor must be avoided. This is best done by mounting the two ports as far apart.
Continuous flow also places greater reliance on the efficiency of maintaining biocatalyst
in the reactor or of biocatalyst removal from the product stream. Not only must the latter
be done on a continuous basis but the particles must also be returned to the main reactor
continuously.
Although the system is fully mixed, a particular element of fluid can be considered to
have an average residence time V/F1, within the reactor vessel. This residence time is a
function of the main reactor volume divided by the flow rate through the reactor and, in
general, the longer the residence time the higher the conversion of the product emerging
from the reactor (Equation 11.4). All the biocatalyst in the CSTR is subject to the same
conditions but, unlike the batch reactor, these conditions do not change with time.
Consequently it is important to choose an equilibrium point where the biocatalyst can
operate at a maximum advantage. Under most operating conditions the solution in the
tank is rich in product and poor in substrate. This makes a CSTR inappropriate in
situations where the product is toxic or inhibitory, but very useful where the substrate has
an adverse effect on the kinetics or stability.
Although in many cases, as we will see, CSTRs are theoretically less efficient than
other continuous reactors, Lilly and Dunnill (1976) showed that in practice they can be
more efficient than packed beds (discussion in next paragraph). This results

Figure 11.5 Schematic representation of a packed bed reactor.

from the relative efficiency with which compressible particles such as cellulose can be
stirred in a CSTR compared with the difficulty of achieving the required velocity of fluid
through a packed bed. This advantage is unlikely to be sustained with rigid porous beads
but the use of CSTRs with compressible or finely divided particles is a serious option in
many processes. The capacity of the system to accommodate solids and gases is a further
 Applied biocatalysis 400
factor favouring their choice. However, it should be remembered that the catalytic
intensity of a CSTR is low so the size of reactor is likely to be large and the residence
time longer than for static-bed reactors.

The packed-bed reactor

In this type of bioreactor the substrate solution passes through a settled bed of particles
held in a column and product emerges continuously at the far end. The degree of
conversion is determined by the time the fluid remains in the bed and this contact time is
determined by the free volume in the reactor bed divided by the flow rate through the
column V/F1.
Fluid can pass upwards or downwards through a vertically mounted bed (Figure 11.5),
while horizontal cylinders can also be used. The latter are rarely appropriate since settling
of the beads leads to channels in the roof of the tube, permitting fluid to by-pass the
particles. The advantage of an up-flow column is that the bed can expand slightly,
preventing pockets of gas or suspended solids from accumulating and disturbing the flow
pattern. However, this expansion also results in a lower activity per unit volume, greater
hold-up of liquid and grading of particles with the associated risk of channelling. These
problems are more apparent with organic beads at high fluid velocities where downflow
columns are usually preferred. In this mode, the design of the column end-piece is
simplified, since the column packing itself often acts as an efficient distributor, giving an
even flow of substrate across the whole diameter of the bed.
It is essential that the column is packed evenly, since the flow of fluid as a steady front
through the bed is a major determinant of column efficiency. Irregularities in packing,
contact with the walls or differences in particle size can lead to the development of
channels which rapidly propagate, allowing substrate to stream through the bed along a
low-pressure path, by-passing the bulk of the biocatalyst.
Unlike stirred tanks, the particles in a packed bed are static and the fluid moves past
the beads. This largely prevents attritional damage to the beads and increases biocatalytic
density in the reactor but places greater emphasis on the quality of the support and the
fluid dynamics in the bed. A greater relative velocity can only be achieved by pumping
fluid through the bed at high volumetric flow and this can seriously limit the choice of
particles which can be efficiently utilised in anything other than a small packed-bed
reactor. The most serious problem occurs with highly active biocatalyst held in
compressible gels or particles with a poor compression strength.
Even with no fluid flow, the weight of bed packing can cause compaction and a
restriction of the interparticle channels. However, for downflow systems, as the flow of
fluid increases there is concomitant rise in pressure across a column and compressible
gels suffer further compaction, increasing the back-pressure and restricting fluid
velocities. For the upflow situation, the pressure drop rises up to a limit, when the liquid
velocity equals the falling velocity of the bed particles, under these conditions a fluidized
bed is obtained.
A packed-bed reactor operating under plug-flow conditions is theoretically the most
 Bioreactor design 401
efficient method of utilising a particulate biocatalysts for simple enzyme conversion.
Since there is no back-mixing an element of substrate solution passes through the bed,
being progressively converted to product without further dilution by fresh substrate.
When operating continuously at steady state each particle in a bed is subject to
constant conditions but the concentration of reagents changes with the position in the
column. When substrate is converted to product in a single pass the pattern of conversion
down the bed resembles that seen when the same reaction is followed with respect to time
in a batch reactor. This stems from the fact that distance travelled through the column is
equivalent to processing with an equal concentration of biocatalyst in the batch reactor
for the period of the column contact time.
The differential of conditions across the bed can cause problems if the substrate is
inhibitory or the product. In the former case the reaction proceeds very slowly, in contrast
to a CSTR, and conversion within the whole bed is poor, while in the latter case the
differential half-life of the sections makes control difficult.
In practice even single-pass reactors deviate from ideal plug flow and, under some
operating conditions, drift into a situation where film diffusional effects reduce
efficiency. Back-mixing also occurs as a result of the porous structure of many supports,
back-diffusion at low flow rates and channelling or inhomogeneities in the flow. Low
flow rates can result from progressive compaction of the bed, the build-up of back-
mixing may be forced on the operator by a decrease in column activity. Despite this, for
many industrial applications the compactness and flexibility of packed beds gives the
high biocatalytic activity and speed of response

Figure 11.6 Schematic representation and operation of a fluidized bed reactor.

required for processing. Restrictions in the equipment usually limit operating pressures to
105 Pa (Buchholz, 1979) although this is sufficient to allow columns of over one meter in
 Applied biocatalysis 402
depth to be operated with residence of a few minutes when using rigid particles.

The fluidized-bed reactor

As the flow of liquid upwards through a packed bed increases the pressure drop rises.
When this pressure equals the weight per unit area of the bed the particles become
bouyant in the liquid and the bed fluidizes, taking on the dynamics of a single fluid.
Following fluidization there is no further increase in back-pressure as flow rate increases
and the system is stable until the flow reaches a rate where particles are washed out of the
bed (Figure 11.6). There is very little contact between the particles so that attrition is low.
Some back-mixing occurs due to turbulence, although in practice this may not be a
serious problem and a fluidzied bed can perform in a similar way to a porous packed bed.
One major advantage of fluidization is the ease with particulate debris and gas bubbles
can be accommodated without causing blockage or by-passing. Bubbles rise freely
through the bed (Allen, Charles and Coughlin, 1979) while gelatinous or colloidal
particles are often sufficiently buoyant to be swept out of the reactor in the main stream.
Fluidized beds have an advantage over CSTRs in the ease with which even very fine
catalytic particles can be retained and returned continuously to the reactor. If the velocity
is reduced the particles will be deposited as a sediment, since their buoyancy is related to
fluid velocity. Thus fluidized beds are designed with a wide section at the top. In this
region the fluid velocity falls so that particles are no longer kept in suspension by the
moving liquid and so fall back into the main part of the reactor.

Figure 11.7 Schematic representation of a bubble-column reactor.

 Bioreactor design 403

The principal drawback of a fluidized bed is the restricted range of low rates and, as a
consequence, the limited contact-time range in the bed during single-pass operation.
Relatively dense particles are particularly useful since they also have a greater differential
in buoyancy compared to organic debris. An increase in particle size provides additional
operational flexibility but, in general, fluidized beds are most useful for finely divided,
high-activity particles. The contact time can be increased by using the reactor in a recycle
mode, with the option of continuous product removal and, under these circumstances, the
system operates like a stirred tank.

The bubble-column reactor

When oxygen is needed in the biocatalytic reaction the bubble column is an attractive
alternative to the stirred tank. A schematic representation of this simple reactor is given
in Figure 11.7. Usually it is engineered with Hv≥2 Tv. At the bottom a sparger is
mounted. To prevent too heterogeneous flow patterns in the lower compartment, the
sparger nozzles have to be distributed over the cross section of the bottom. Therefore, one
ring or a small number of parallel pipes or a starlike construction of pipes is commonly
used. In the pipes holes are drilled. Complicated spargers or very small holes merely have
disadvantages for most applications. Like the stirred tank, it can be run continuous or
batch wise, either with free or immobilized biocatalyst.

Figure 11.8 Schematic representation of the air-lift reactor. A. Internal-loop

reactor B. External-loop reactor.
 Applied biocatalysis 404

The air-lift loop reactor

In contrast to the bubble column, the air-lift loop reactor has a hydrodynamic flow pattern
which can be well described and which is controlled by the gas flow.
The air-lift consists of two pipes, interconnected at top and bottom. In one of the pipes
(the riser) air is sparged at the bottom. The air rises and escapes at the top. Therefore,
under most circumstances there is no air present in the other pipe (the downcomer). The
density difference between riser and downcomer causes an intensive liquid circulation.
Two designs can be used, i.e., the internal (Figure 11.8A) and the external loop reactor
(Figure 11.8B).
When an internal loop reactor is built underground, we refer to this as a deep shaft.
Volumes can be up to thousands of in m3. Hv generally is much larger than Tv, usually of
the order of 10 Tv, but for the deep shaft up to 100 Tv. In this case it is especially
advantageous to use multiple feedstock inlets.

Novel multiphase bioreactors

In the literature many examples of more or less exotic bioreactors can be found. Few
actually are applied, outside the laboratory. Here two novel designs, the membrane and
the liquid-impelled loop reactor, are discussed briefly. These two reactors are simple to
use and, to a certain extent, liable to scale-up and both integrate the actual biocatalysis
with part of the down-stream processing.
Membranes have the property to retain one or more components of a liquid mixture,
whereas others may pass. The size of the biocatalyst often differs considerally from the
size of the product molecules. The pore size of the membrane must thus be chosen in
such a way that the product can pass the membrane, while the biocatalyst is retained.
Usually membrane bioreactors consist of ultrafiltration
 Bioreactor design 405

Figure 11.9 Different arrangements and modes of operation for membrane

bioreactors: Continuous Stirred Tank Reactor (CSTR) with
recirculation arrangement (a), dead-end cell (b), tubular with
entrapped enzyme (c).

membranes and a hollow fiber membrane module is preferred (Figure 11.9), due to the
high surface area/volume ratio.
Conversions in two-liquid-phase systems are promising. Although these reactions can
be performed in a stirred emulsion system, the use of membrane bioreactors can be
advantageous. In addition to retaining the biocatalyst in the reactor, the membrane also
serves as a separator between aqueous and organic phase, thus avoiding energy-
demanding phase separations (Prazeres and Cabral, 1994).
A novel bioreactor, especially designed to work with two liquid phases, is the liquid-
impelled loop reactor (Figure 11.10), in which the advantages of air lifts and
 Applied biocatalysis 406

Figure 11.10 Schematic representation of a liquid-impelled loop reactor.

Figure 11.11 The mass balance over the bioreactor.

of organic solvents are integrated. In this reactor, mixing under controlled-flow

conditions and low shear forces are combined with high solubilities of components
poorly soluble in water, increasing reaction rates and facilitating product recovery by
extraction. The principle of the air lift is used in the liquid-impelled loop reactor. Instead
of a gas phase, a dispersed liquid phase induces the circulation of the continuous liquid
phase in this type of reactor. Several configurations are possible: internal or external
loop, up flow or down flow of the dispersed phase, or a combination of these. It is also
possible to combine liquid injection with gas injection (Van Sonsbeek, De Blank and
Tramper, 1992).
 Bioreactor design 407

11.1.4 Reactor Concepts

Introduction

The bioreactor has been introduced in general terms in the previous section. In this
section the basic bioreactor concepts, i.e., the batch, the fed-batch, the continuous-flow
stirred-tank reactor (CSTR), the cascade of CSTRs and the plug-flow reactor, will be
described.
Integration with the (micro) kinetics, in other words the kinetics of the pertinent free
biocatalysts or of the immobilized biocatalysts including mass transfer, yields the overall
reactor description or macrokinetics in later sections. In order to come up with these
descriptions, a mass balance over the bioreactor should be drawn up (Figure 11.11).
In words:
The accumulation of a compound A in the reactor, with concentration CAr, is equal to
the amount of A that comes in minus the amount that goes out, and augmented by the
amount that is produced.
In formula:

(11.5)

Figure 11.12 The batch reactor.

In this equation V is the liquid volume in the bioreactor (m3), CA the concentration of A
(mol.m−3) by which the subscripts i and o refer to the concentration in the influent and
effluent, respectively, t is time (s), Fi and Fo the flow (m3.s−1) of the in and outgoing
stream, respectively, and the production rate per unit volume (mol.m−3.s−1) of A.

The batch reactor

 Applied biocatalysis 408
The most-prominent characteristic of the batch reactor is the fact that there are no in- and
outgoing flow (Figure 11.12). This means that all that is produced is accumulated. The
mass balance [Equation (11.5)] thus simplifies to assuming ideal mixing:

(11.6)

When the volume V remains constant, Equation (11.6) further simplifies to:

(11.7)

With boundary conditions CAr=CAr (0) at t=0 and CAr=CAr (tb) at t=tb, the time one batch
lasts, separation of variables and integration leads to:

(11.8)

Substitution of the pertinent rate equation yields the time a run should last to obtain a
desired conversion.

Figure 11.13 The fed-batch reactor.

The fed-batch reactor

The distinguishing feature of the fed-batch reactor is that there is only an ingoing flow
and no outgoing flow (Figure 11.13). Assuming ideal mixing equation (11.5) thus
 Bioreactor design 409
becomes:

(11.9)

or:

(11.10)

This equation cannot be solved analytically without further simplification and data.

The continuous-flow stirred-tank reactor (CSTR)

For the CSTR defined here (Figure 11.14) the volume V is constant and ideally mixed
and the inflow equals the outflow, i.e. Fi=F0=F. For solving the mass balance [Equation
(11.5)] of a CSTR it is assumed that the reactor essentially is in a steady state:

(11.11)

Equation (11.5) thus becomes:

(11.12)

Figure 11.14 The CSTR.

 Applied biocatalysis 410
or:

(11.13)

in which CSTR is the average residence time (s). As in an ideally mixed vessel, the
concentration in the reactor is equal to CAo. and this equation can be solved immediately
by substituting for the appropriate rate equation with CAr=CAo .
For n CSTRs in series (Figure 11.15) the same assumptions are made for each vessel as
for the single CSTR above. For each vessel in the series Equation (11.13) thus holds:

(11.14)

The subscript j refers to the j-th vessel. In the following sections the optimum design for
n CSTRs in series containing (immobilized) biocatalyst of constant activity and following
Michaelis-Menten kinetics will be worked out.

Figure 11.15 The cascade of CSTRs.

 Bioreactor design 411

Figure 11.16 The plug-flow reactor (PFR).

The plug-flow reactor (PFR)

In the ideal plug-flow reactor (Figure 11.16) the continuous phase flows as a plug
through the reactor; i.e., there is no mixing or, in other words, no axial dispersion.
Consequently, if a compound is consumed or produced, a concentration gradient will
exist in the direction of flow. The mass balance is therefore first set up over an infinite
small slice perpendicular to the direction of the flow with volume dV of the bioreactor.
Assuming steady state and Fi=F0=F, Equation (11.5) then is reduced to:

(11.15)

Rewriting gives:

(11.16)

Integration over the whole reactor:

(11.17)

or:

(11.18)

in which pf (s) is the residence time of the plug-flow reactor. This equation is, in
principle, the same as that of the batch reactor and integration with the (micro) kinetics is
identical.

Discussion of the reactor concepts

 Applied biocatalysis 412
A comparison of the various types of reactor concepts, in a general sense, is actually only
possible between the batch, the CSTR and the PFR. The cascade of CSTRs, depending on
the number of vessels n in the series, more or less behaves as an ideal mixer for n→1 or
an ideal plug flow for n→ . The fed-batch reactor is more difficult to situate. Although
the concentration of compounds important for the rate of reaction can be controlled
optimally during the whole fed period, the reactor volume is only partially utilized,
especially in the beginning. Nevertheless, this reactor concept certainly has decisive
advantages in many cases, as shown by the fact that it is one of the most widely used.
For the batch, the CSTR and the PFR the following equations for the “residence time”
have been derived, respectively:

(11.8)

(11.13)

(11.18)

In general, the rate of reaction decreases if the reactant concentration decreases (reaction-
order > 0). For a CSTR this means that the rate of reaction is low for the whole reactor as
it is determined by the low concentration in the reactor, equal to that of the outflow. For
the other two reactor types the conversion takes place at concentrations less than the
higher incoming concentration. This means that in case of “ordinary” kinetics
, in other words that the CSTR should be larger than the batch and the
PFR in order to accomplish the same degree of conversion. Naturally, the down-time of
the batch reactor is not taken into account here.
In an autocatalytic reaction, i.e. the more product the faster the rate of reaction, or
when there is substrate inhibition the versus CAr curves can look quite different. In
that case . For these types of kinetics it is thus advantageous to use a
CSTR. Only when zero-order kinetics are involved there are no difference in “residence
times” and thus in the volume, of the three types of reactors to accomplish a certain
conversion.

11.2 DESIGN OF BIOCATALYST REACTORS

 Bioreactor design 413

11.2.1 Introduction
In chapter 8 the most generally used kinetic equations for describing the consumption of
substrate as a result of biocatalysis have been given and/or derived. In biocatalysis, in the
absence of limitation of the rate of consumption by diffusion of substrate, the Michaelis-
Menten equation usually is a good description:

(11.19)

in which

= substrate consumption rate per unit volume (mol.m−3.s−1)

vmax = maximum substrate conversion rate (mol.m−3.s−1)

Cs = substrate concentration (mol.m−3)
Km = Michaelis-Menten constant (mol.m−3)

On the basis of a mass balance the following general equations describing the basic
reactor concepts have been derived in the last section. For the batch reactor the time tb (s)
needed for a conversion of a component i from Ci (end) at t=0 to Ci (0) is given by:

(11.20)

The mass balance over a continuous stirred-tank reactor (CSTR) in the steady state yields
for the average residence CSTR (S):

(11.21)

in which

Cii = concentration of i in the inflow (mol.m−3)

Cio = concentration of i in the outflow (mol.m−3)
 Applied biocatalysis 414
For a CSTR which is assumed to be ideally mixed, the concentration in the reactor, Cir, is
equal to the concentration in the outflow, Cio.
Similarly, for the j-th reactor in a cascade of n CSTRs:

(11.22)

Finally, for the PFR it was derived that:

(11.23)

These are the equations which will be integrated below to come up with the basic overall
bioreactor models.

11.2.2 The Batch Bioreactor

Michaelis-Menten kinetics

Substitution of Equation (11.19) in Equation (11.20) gives for the substrate s:

(11.24)

Working this equation out gives:

(11.25)

When the kinetic constants, the initial concentration of substrate and the desired
conversion are known, the required batch time tb can thus easily be calculated.

11.2.3 The Continuous-flow Stirred-tank Reactor (CSTR)

 Bioreactor design 415

Michaelis-Menten kinetics

Substitution of Equation (11.19) in the general equation for the CSTR [Equation (11.21)]
yields:

(11.26)

with Csr=Cso for an ideally mixed solution. For any desired conversion the required
residence time can thus be calculated directly.

11.2.4 The Cascade of n CSTRs

Introduction

Much attention has been given in the past to reactor systems consisting of a series of
well-stirred tanks, because of the relative simplicity and the great importance of these
systems. In standard textbooks on chemical reaction engineering, like Levenspiel (1972),
general concepts of reactor design are treated and mostly illustrated with n-th order
reaction kinetics. The optimization of a series of CSTRs is usually executed by defining
the optimum as the smallest total reactor size (holding time) to perform a specific
conversion. This definition is also applicable in the derivation of the following sections.
Finding the optimal design thus amounts to finding the minimum total holding time,
which is a function of all intermediate substrate concentrations, i.e., the concentration of
substrate in the first, the second, etc., until the last-but-one reactor in the series.
Mathematically formulated, this involves finding the intermediate substrate concentration
values subject to the following equation [derived from Equation (11.22)]:

(11.27)

This set of (n) equations has to be solved simultaneously for the (n) intermediate
concentrations Csr. The choice of the total number of reactors n is naturally dictated by
the economics of the process. Usually, as discussed by Reusser (1961), only about two to
four reactors are justified. Below it will be shown that the introduction of a second
reactor considerably reduces the total reactor volume. The case of MichaelisMenten
kinetics will be worked out in detail.
 Applied biocatalysis 416

Michaelis-Menten kinetics

A simple analytical expression has been derived by Luyben and Tramper (1982) for the
optimal design of CSTRs in series, assuming a constant activity of the biocatalyst in the
reactor. The optimum is defined as the smallest total reactor size (holding time) to
perform a specific conversion. The resulting total holding time can also be used as a good
approximation for the total holding time of equal-sized CSTRs. The mathematically more
complex case of n equal-sized CSTRs will be illustrated by an example. Consider n
CSTRs in series with an inlet concentration of substrate of Csi,1 (mol.m−3) for the first
reactor (Figure 11.15).
Introducing Michaelis-Menten kinetics Equation (11.19) in Equation (11.22) gives:

(11.28)

This equation can be written in dimensionless form by introducing the following

variables:

(11.29)

(11.30)

and:

(11.31)

Substitution of Equations (11.29–11.31) in Equation (11.28) leads to:

(11.32)
 Bioreactor design 417
Equation (11.32) is a general relation that gives the dimensionless holding time required
in the j-th reactor to obtain a dimensionless concentration j starting from j−1,
assuming Michaelis-Menten kinetics. From j immediately follows the volume of the j-
th reactor for a given maximum reaction rate vmax, volumetric flow rate Fi and initial
substrate concentration Csr,1 (Equations (11.31) and (11.22)).
Finding the optimal design according to the above definition amounts to finding the
minimum of the total holding time, which is a function of all j’s. Mathematically
formulated, this involves finding the intermediate j -values subject to the following
equation:

(11.33)

Only two terms of this summation contain j, leading to:

(11.34)

Differentiation and rearranging gives:

(11.35)

This simple result relates the conversion (1j— ) in the j-th reactor to the conversion in
the (j−1)th and the (j+2)th reactor for a series of CSTRs, in which a reaction takes place
following Michaelis-Menten kinetics. Important to note is that K c dropped out of the
relation. This means that the intermediate substrate concentrations for a series of
perfectly mixed tank reactors are independent of the Michaelis-Menten constant Km.
Equation (11.35) consists of a set of (n−1) relations which can be solved directly for a
given total conversion (1— n) as follows. Writing the set of equations from Equation
(11.35):
 Applied biocatalysis 418

(11.36)

(11.37)

(11.38)

Substituting from the bottom up gives:

(11.39)

The conversion (1− j) is related to the inlet concentration of the first tank, 0, which is
equal to 1 by definition of Equation (11.29) and Equation (11.39) thus simplifies to:

(11.40)

Some results of Equation (11.39) and subsequent use of Equation (11.32) are presented in
Table 11.1. This table gives the dimensionless concentrations and holding times in the
mixed reactors for an initial concentration of ten times the

Table 11.1 Dimensionless concentrations and holding times in the mixed reactors for a n
=0.01 and κ=0.1.

j (independent t of K) j (dependent of of k)
n 1 2 3 4 5 1 2 3 4 5
1 0.01 10.89
2 0.1 0.01 1.80 0.99
3 0.215 0.046 0.01 1.149 0.533 0.401
4 0.316 0.100 0.0316 0.010 0.900 0.432 0.285 0.238
 Bioreactor design 419

5 0.398 0.158 0.0631 0.025 0.01 0.753 0.391 0.247 0.189 0.166
Source: Adapted from Luyben and Tramper, 1982.

Table 11.2 Dimensionless total holding times for optimal and equal-sized mixed
reactors, for α n=0.01 and two values of κ.

k= 0.1 k= 1
n

1 10.890 10.890 99.990 99.990

2 2.790 2.917 18.990 19.006
3 2.082 2.234 11.915 11.947
4 1.855 1.983 9.639 9.677
5 1.746 1.854 8.549 8.588
6 1.683 1.776 7.197 7.954
7 1.641 1.723 7.505 7.540
8 1.613 1.684 7.216 7.249
9 1.591 1.656 7.003 7.033
10 1.575 1.633 6.839 6.867
1.451 5.595

Source: Adapted from Luyben and Tramper, 1982.

Michaelis-Menten constant and a conversion of 99%. The data show that the difference in
holding time between two subsequent reactors is largest for low values of n, especially
between the first two reactors and becomes smaller as n increases.
Table 11.2 gives the total holding times for two values of K, both for a series of CSTRs
with minimal total volume and for a series of equal-sized mixed reactors. Total holding
times for equal-sized mixed reactors have been calculated using a zero finding routine.
The last value in Table 11.2 is the dimensionless holding time for a PFR reactor with
Michaelis-Menten kinetics, calculated by means of the following equation:

(11.41)
 Applied biocatalysis 420
An important observation from Table 11.2 is the considerable difference going from one
to two or more CSTRs. For the conditions studied, there is only a minor difference (less
than 10%) between the total holding time for optimal and equal-sized mixed reactors.
Even in extreme cases, i.e., for very low values of κ and , this difference remains
relatively small (37% for n=3, κ=10−3 and =10−4).
Furthermore, it can be shown that, in the limiting cases of first-order kinetics [Equation
(11.35) also holds for this case] and zero-order kinetics, the equal and optimal sizes are
exactly the same. As shown, the optimal holding times can be calculated very simply by
means of Equation (11.40) and the sum of these can thus be used as a good
approximation for the total holding time of equal-sized CSTRs. This makes Equation
(11.31) an even more valuable tool for design equations. The restrictions are imposed by
the assumption that the biocatalytic activity is constant in the reactors. Especially in the
case of soluble enzymes, for which ordinary Michaelis-Menten kinetics in particular
apply, special measures have to be taken. Continuous supply of relatively stable enzyme
to the first tank in the series is a possibility, though in general expensive. A more
attractive alternative is the application of a series of membrane reactors.

11.2.5 The Plug-flow Reactor (PFR)

Michaelis-Menten kinetics

Substitution of Equation (11.19) in Equation (11.23) for the PFR gives an equation
similar to the one found for the batch reactor:

(11.42)

or

(11.43)

Analogous to the batch reactor with Michaelis-Menten kinetics, this equation for the
residence time of the PFR can be solved directly when the kinetic constants, the inlet
concentration of substrate and the desired conversion are known.

11.2.6 Comparison of Enzyme Reactors

 Bioreactor design 421
When the biochemical reactors are kinetically controlled, the batch bioreactors and the
PFR are described by the same design equations (Equations (11.25) and (11.28)) and
show a better performance than the CSTR in most cases, except for substrate inhibition
kinetics.
Figure 11.17 compares the substrate conversion degrees obtained in a PFR and

Figure 11.17 Comparison of a CSTR and a PFR, for the Michaelis-Menten

kinetics.

in a CSTR with the same residence time in both types of bioreactors for the Michaelis-
Menten kinetics.
For zero-order reaction, both reactors show the same performance:

(11.44)

where X is the conversion degree

For the first-order reaction rates, the PFR displays a higher performance than the
CSTR:

(11.45)

Obviously, Michaelis-Menten kinetics, having as extremes zero-order (very high

 Applied biocatalysis 422
substrate concentrations) and first-order (very low substrate concentrations) kinetics,
shows a behaviour in between.

11.3 DESIGN OF REAL REACTORS

11.3.1 Introduction
The design equations described in the last section are only valid for ideal reactions when
the reactions are kinetically controlled. However, the modelling of enzyme reactions
should take into account several factors which influence their performance. These factors
are:

(a) The (immobilized) biocatalyst kinetics;

(b) The external and internal mass transfer effects;
(c) The axial disperson (back-mixing) effects;
(d) The heat transfer effects; and
(e) The operational stability of the immobilized biocatalyst.

The design of real reactors, taking into account the diffusion, axial dispersion and
enzyme inactivation effects, is described in the following sections, considering
Michaelis-Menten kinetics as a model. These models are very important in predicting and
simulating bioreactor performance and in modeling future processes. Also, for control
purposes they are indispensable.

11.3.2 Internal and External Mass-transfer Effects

When a biocatalyst is immobilized on or within a solid matrix, mass transfer effects may
exist because the substrate must diffuse from the bulk solution to the immobilized
biocatalyst. If the biocatalyst is attached to non-porous supports there are only external
mass transfer effects on the catalytically active outer surface; in the reaction solution, the
supports are surrounded by a stagnant film and substrate and product are transported
across this Nernst layer by diffusion. The driving force for this diffusion is the
concentration difference between the surface and the bulk concentration of substrate and
product.
For instance, the rate of flow of substrate F from the bulk solution to the biocatalyst
surface in one m3 reactor volume is given by:

(11.46)
 Bioreactor design 423

where ks1 is the mass transfer coefficient, A1 is the particle surface area per unit of
volume and Csb and Csi are the bulk and surface concentrations of the substrate,
respectively.
In a surface reaction, the flow of substrate to the biocatalyst surface and the reaction
take place consecutively. At steady state the rate of external mass transfer of substrate,
will be equal to its internal removal by reaction. Hence, for a biocatalyst reaction,
which obeys Michaelis-Menten kinetics, the overall rate of reaction, vobs will be:

(11.47)

This equation may be solved for Csi if ksiA1 and the kinetic constants are known
(Mosbach, 1976) or Csi may be obtained graphically (Mosbach, 1976) using the
following dimensionless equation:

(11.48)

where is the dimensionless substrate concentration and Da is an adapted

Damkohler number, . The dependence of on for
different values of Da is shown in Figure 11.18.
The external mass-transfer effects on the activity of an immobilized biocatalyst can be
expressed quantitatively by the external effectiveness factor ee, defined as the ratio of
the observed reaction rate to the rate (Csb) which would be observed if all the
biocatalyst would be surrounded by the bulk concentration:

(11.49)

Figure 11.19 shows the dependence of the external effectiveness factor ee on and Da.
Similar plots have been obtained by other authors (Mosbach, 1976).
 Applied biocatalysis 424

Figure 11.18 against the dimensionless bulk concentration b for

different values of the substrate modulus Da for external diffusion.
Adapted from C.Horvath and J.M. Engasser. Biotechnol. Bioeng., 16,
909 (1974).
 Bioreactor design 425

Figure 11.19 Plots of the external effectiveness factor ee as a function of the

substrate modulus Da for different values of the dimensionless bulk
substrate concentration b. is the limiting first-order effectiveness
factor attained at sufficiently low concentrations. Adapted from
C.Horvath and J.M.Engasser. Biotechnol.Bioeng., 16, 909 (1974).
 Applied biocatalysis 426

Figure 11.20 Schematic plot of the overall rate of reaction catalyzed by a

surface-bound biocatalyst against the bulk substrate concentration.
Adapted from C.Horvath and J.M. Engasser. Biotechnol.Bioeng., 16,
909 (1974).

For the first order reaction, the external effectiveness factor has an analytical solution,
which is given by:

(11.50)

The rate flow of substrate or the rate of biocatalysts reaction (Cs) may play a
predominant role, depending on their relative magnitudes, as the lower rate step will be
the controlling step.
As can be seen in Figure 11.20, at high bulk substrate concentrations when the reaction
is zero order, will always approach vmax and the reaction is kinetically controlled. At
lower bulk substrate concentrations the reaction can be both kinetically or diffusionally
controlled depending on the ratio of ks1A1 and Vmax/Km. When ks1A1 >> vmax/Km, mass
transfer is much faster than the biocatalytic reaction, but when ks1A1 << vmax/Km, the
biocatalytic reaction is much faster than the diffusion of substrate.
 Bioreactor design 427
Decreasing resistance to external mass transfer is achieved with an increase of linear
velocity of fluid, as this velocity reduces the resistance to a point that Csi approaches Csb
and the reaction rate (Csb) will be the controlling step.
When a biocatalyst is immobilized within a porous support, in addition to possible
external mass-transfer effects there could also exist resistance to internal diffusion of
substrate, as this must diffuse through the pores in order to reach the biocatalyst.
Consequently a substrate concentration gradient is established within the pores, resulting
in a concentration decreasing with increased distance (in depth) from the surface of
immobilized biocatalyst preparation. Similarly, a corresponding product concentration
gradient is obtained in the opposite direction.
Unlike external diffusion, internal mass transfer occurs simultaneously with the
biocatalyst reaction and takes into account the depletion of substrate within the pores
with increasing distance from the surface of the support. The rate of reaction will also
decrease, for the same reason. The reaction is dependent on the substrate concentration
and thus the distance from the outside support surface.
The usual way to study this problem is by considering that there is a coupled reaction
diffusion process that can be solved, at the steady state, when the rate of internal
diffusion and biocatalytic reaction are equal, using appropriate differential equations for
the various geometries considered and isothermal conditions (Mosbach, 1976):

(11.51)

where Cs is the substrate concentration; x is the distance from the outer surface; p is a
geometrical factor with the value of +1 for spherical pellets, 0 for cylindrical pellets, and
−1 for rectangular membranes; and |Deff is the effective diffusivity of the substrate inside
the support, given by:

(11.52)

where |Dsl is the substrate diffusivity in the liquid; , is the void fraction of the porous
support and is a tortuosity factor that takes into account the pore geometry and, by
definition, is larger than unity.
The analytical solutions of these equations are easily obtained for first-order or
constant-order reactions, but numerical solutions are required for Michaelis-Menten type
reactions. The above equations, in these cases, are usually rewritten in terms of
dimensionless variables. For a spherical pellet this equation is:
 Applied biocatalysis 428

(11.53)

with the boundary conditions:

Figure 11.21 Normalized overall rate as a function of the dimensionless bulk

substrate concentration b for different values of , the substrate
modulus for internal diffusion in a membrane. Adapted from
C.Horvath and J.M.Engasser. Biotechnol. Bioeng., 16, 909 (1974).

In this equation is the dimensionless substrate concentration, Z is the dimensionless

position in the porous support given by Z=r/R, R is the radius of the spherical pellet, and
 Bioreactor design 429

is the first-order kinetics Thiele modulus defined by:

(11.54)

Numerical integration yields the effective rate of reaction, as a function of the

concentration with the modulus (Figure 11.21). The same results can also be
represented in the form of graphics of effectiveness factor ei against this Thiele
modulus (Figure 11.22).
The internal mass-transfer effects can be reduced, however, by decreasing the particle
dimensions of the porous support containing the biocatalyst. Particle-diameter decrease
results in a reduction of the distance from the outer support surface that the substrate
must cross and, consequently, also results in a decrease of the substrate concentration
gradient.
When external and internal diffusion resistances affect simultaneously the rate of the
biocatalytic reaction, the relative contributions of each effect must be estimated

Figure 11.22 Graph illustrating the effectiveness factor ei as a function of

the Thiele modulus with the dimensionless surface concentration
as the parameter. The effectiveness factor for spherical particles
and membranes are represented by solid and broken lines,
 Applied biocatalysis 430
respectively. Adapted from C.Horvath and J.M.Engasser. J. Theor. Biol., 4 2,
437 (1973).

separately and quantified by the corresponding effectiveness factors. Hence, the overall
reaction rate is given by:

(11.55)

11.3.3 Effects of Mass Transfer on the Performance of Immobilized-

biocatalyst Reactors
The design equations previously described are only valid when there are no factors which
modify the kinetics of the immobilized biocatalyst (partition effects, heat and mass
transfer effects and decay of biological activity) and the hydrodynamic characteristics of
the reactor (back-mixing). Thus, the kinetic constants Km and vmax used in those
equations are intrinsic values obtained in the absence of those factors, being only
dependent on the conformational and stereochemical effects inherent in the
immobilization procedure used.
When the mass transfer (diffusional) effects are significant, the design equations for
immobilized biocatalyst reactors can be modified, by introducing the concept of the
overall effectiveness factor in the equation of the reaction rate:

(11.56)

The following sections describe the reactor performance of the three main types of
reactors in the presence of mass transfer effects integrated with Michaelis-Menten
kinetics.

The batch bioreactor

Substitution of Equation (11.56) in Equation (11.20) gives:

(11.57)
 Bioreactor design 431

This equation cannot be integrated immediately as is a function of the substrate

concentration Cs. A possible procedure is calculation of the overall effectiveness factor
at n substrate concentrations in the range of to Through this n ( , Cs)
data pairs an n-th order polynomial is fitted.
This polynomial is substituted in Equation (11.57) such that integration is possible.

The continuous-flow stirred-tank reactor (CSTR)

Analogous to the above ordinary Michaelis-Menten kinetics, substitution of Equation

(11.56) in Equation (11.21) yields a formula which allows straightforward calculation of
the residence time required for a specific conversion:

(11.58)

again with Csr=Cso.

For CSTR, because the component concentrations are constant, the effectiveness factor
does not change in the reactor.

The plug flow reactor (PFR)

Also in case of diffusion-limited reactions where the overall effectiveness factor is used
to describe the effect of diffusion on the rate of biocatalysis, the mathematics are the
same as in the case of the batch reactor. Substitution of Equation (11.56) in Equation
(11.23) thus yields:

(11.59)

This equation, again, cannot be integrated immediately as is a function of the

substrate concentration Csr, which changes when going from the entrace to the exit of the
plug-flow reactor. The procedure to solve this equation is the calculation of the overall
effectiveness factor at n substrate concentration in the interval Csi to Cso. Through this n (
,Csr)data pairs an n-th order polynomial is fitted. This polynomial is substituted in
Equation (11.59) followed by a numerical solution.
 Applied biocatalysis 432

11.3.4 Effects of Back-mixing on the Performance of Immobilized-

biocatalyst Reactors
In real tubular (or column) reactors there is, usually, a back-mixing effect which
influences the performance of the ideal plug-flow reactor. This axial dispersion is higher
for fluidized-bed reactors than for packed-bed reactors, although comparatively lower
than for continuous-feed stirred-tank reactors, where the mixing is complete.
The modeling of real immobilized-enzyme column reactors, mainly the fluidized-bed
type, has been described (Emery and Cardoso, 1978; Allen, Charles and Coughlin, 1979;
Kobayashi and Moo-Young, 1971) by mathematical models based on the dispersion
concept (Levenspiel, 1972), by incorporation of an additional term to account for back-
mixing in the ideal plug-flow reactor. This term describes the non-ideal effects in terms
of a dispersion coefficient.
The design equation is:

(11.60)

where |Del (m2.s−1) is the axial dispersion coefficient in the liquid, v1s (m.s−1) is the
superficial fluid velocity, L the reactor length or bed height and Z (= x/L) the normalized
distance along the reactor coordinate x, |Del/(v1sL) is the dispersion number or the inverse
of the Bodenstein number. When there are simultaneously dispersion and mass-transfer
effects, Equation (11.60) involves the overall effectiveness factor

(11.61)

11.3.5 Effects of Enzyme Inactivation on the Performance of Immobilized

Biocatalyst Reactors
The performance of immobilized biocatalyst (enzyme) reactors is influenced by enzyme
inactivation during operation, mainly due to thermal denaturation, desorption of the
biocatalyst from the solid support, disintegration or solubilisation of the support and
microbial attack.
The overall effect of these factors can be determined experimentally and it is
convenient to reduce them in order to increase the operational stability of the biocatalyst.
The prediction of the loss of the performance of an immobilized enzyme reactor due to
thermal denaturation can be quantified, taking into account the enzyme-inactivation
 Bioreactor design 433
kinetics models. The most used is the exponential decay model:

(11.62)

where E is the effective enzyme activity in the reactor at time t and kd is the first-order
decay constant.
This equation can also be written as:

(11.63)

where E0 is the initial enzyme activity at time t=0.

Also the enzyme activity can be correlated with the maximum activity, vmax through
the following equation:

(11.64)

where k2 is the dissociation constant of the enzyme-substrate complex into product,

which is dependent in the temperature.
Substitution of Equation (11.63) in Equation (11.64) yields:

(11.65)

or

(11.66)

where vmax0 is the initial maximum velocity at time t=0.

Substituting this model in the reactor-performance equations for Michaelis-Menten
kinetics, the following equations can be obtained for the three main types of reactions:
Batch reactor:
 Applied biocatalysis 434

(11.67)

CSTR:

(11.68)

PFR:

(11.69)

From these equations it can be deduced that the immobilized biocatalyst deactivates more
slowly in a CSTR than in a batch or a PFR.

11.4 EXERCISES

11.4.1 Glucoamylase in a concentration of 0.5 kg.m−3 catalyzes a desired conversion

=1 kg/kg in a batch reactor at a constant rate of 1 g of product (glucose) formed per
second per kg of biocatalyst. One batch takes 1 week overall and there are 40 runs per
year.

a) What is the overall productivity of the biocatalyst and the overall volumetric
productivity of the reactor?

b) What is the end product concentration if the actual biocatalysis is 4 days?

Solution: a) Pr =346 kg/kg and Q =2.19. 10− 4 kg.m−3.s− 1
px p
b) Cp=173 kg.m−3

11.4.2 The initial substrate concentration in exercise 1. is 192 kg.m−3. From the point
of view of economics the batch has to be continued until 99% conversion.

a) What is the additional batch time needed to accomplish this?

 Bioreactor design 435

b) First-order kinetics can be assumed in this last traject. Use the rate of 90%
conversion to estimate the first-order rate constant.

c) What is the overall (volumetric) productivity of the biocatalyst and bioreactor, for
40 runs/year?
Solution: a) 24.6 h
b) 5.2.10−5.s−1
c) Pr =380.16 kg/kg and Q =7603.2 kg.m− 3.year− 1
px p

11.4. Invertase is immobilized by adsorption on a non-porous glass support with a bead

diameter of 2 mm for the hydrolysis of 100 mM sucrose at pH 4.5 and 45°C. Calculate:

a) The external effectiveness factor;

b) The observed reaction rate;

c) The substrate concentration at the support surface.

Data:
Intrinsic kinetic constants: v =0.10 mol.m−3.s−1 and K =10 mM Mass transfer
max m
coefficient, ksl=2.10− 7 m.s− 1

Solution: a) ee=0.7;

b)=0.0636 mol.m−3.s−1;

c) C =17.5 mM
si

11.4.4 Invertase is immobilized in a porous support with a bead diameter of 3 mm for

the hydrolysis of 250 mM sucrose at pH 4.5 and 45°C. Calculate:

a) The overall effectiveness factor;

b) The overall effectiveness factor when the bead size is 1mm;

c) The overall effectiveness factor when the substrate concentration is 10 mM;

 Applied biocatalysis 436

d) The overall effectiveness factor when the enzyme concentration is 10 times higher;

e) The effectiveness factor after 90 days of operation, when the half life of the
immobilized invertase is 30 days and this enzyme preparation follows a first-order decay.

Data
Intrinsic kinetic constants: v =0.10 mol.m− 3.s− 1 and K =10 mM
max m
Support porosity, i=0.30
Support tortuosity, i=3
Diffusion coefficient of sucrose in solution, D 3.10− 10 m2.s− 1
sl=
Mass transfer coefficient, k =2.10−7 m.s−1
sl

Solution: a) =0.27;

b) =1;

c) =0.015;

d) =0.0072;

e) =1

11.4.5 Sucrose (1 M concentration) is hydrolyzed with a conversion degree of 98% by

invertase immobilized in a porous support with a bead size of 3 mm in a continous
operation with a flow rate of 5 1.min−1.
Calculate:

a) The ideal volumes for a PFR and a CSTR type of reactors;

b) The real volumes for both type of continuous reactors;

c) The time of operation, in both type of reactors, after which the conversion degree
decreases to 90%.

Data
Intrinsic kinetic constants: v =0.10 mol.m− 3.s− 1 and Km=10 mM
max
Support porosity, i=0.30
 Bioreactor design 437
Support tortuosity, i=3
Diffusion coefficient of sucrose in solution, D =3.10− 10 m2.s− 1
sl
Mass transfer coefficient, k =2.10−7 m.s−1
sl
Operational half-life of the immobilized invertase=30 days (following a first-order
decay)

Solution: a) V =0.849 m3 and V 3

PFR CSTR=1.225 m ;

b) V m3 and VCSTR=74.2 m3;

PFR=3.54

c) t =167 days and t

PFR CSTR=195 days.

11.4.6 A lipase from Rhizomucor miehei is immobilized onto a porous polyamide

support with a bead diameter of 0.3 mm and used for the hydrolysis of 100 mM caprylin
at pH 7.0 and 30°C, in a continous plug flow reactor. Calculate the amount of
immobilized lipase needed to achieve a conversion degree of 97% with an inlet flow rate
of 1 1.min−1.
Data:
Intrinsic kinetic constants: v =10 mmol.mg prot− 1. min− 1 and K =200 mM
max m
Enzyme concentration=24 mg prot./g support
Effective diffusivity, D 3.10−11 m2.s− 1
eff=
Mass transfer coefficient, k =2.5.10−6 m.s−1
sl
Apparent density of the support=0.6 g.dm−3

Solution: 27.7 g of immobilized lipase

11.4.7 A Chromobacterium viscosum lipase is microencapsulated in AOT reversed

micelles in isooctane with a Wo=24 and used in the controlled hydrolysis of 50 mM
triolein at pH 7.0 and 35°C, in a continous stirred membrane reactor, with a flow rate of 1
1.min− 1. Design the reactor in order to achieve 95% of conversion.
Data:
Apparent kinetic constants: v =0.810 mmol.mg prot.−1 min− 1 and Km= 210 mM
max
Enzyme concentration=0.035 g.dm−3

Solution: 427 1

11.4.8 To prevent accummulation of ammonium ions in consumption fish production

ponds the water must be refreshed continously at a rather high rate. One process to
remove the ammonium ions is the use of immobilized bacteria in porous particles with a
particle size of 2 mm. Calculate the size of the plug flow reactor with a bed porosity of
0.5 by which 1 m3 of pond water can be treated per minute.
 Applied biocatalysis 438
Data:
Intrinsic kinetic constants: v =0.10 mol.s− 1.m− 3 particles and Km=100 mM
max
Effective diffusivity, D =10− 9 m2.s− 1
eff
Mass transfer coefficient, k =2.10− 4 m.s− 1
sl
Inlet concentration, C =54 gNH m− 3
si 4+
Outlet concentration, C =18 gNH + m− 3
so 4

Solution: 39 m3

11.5 REFERENCES AND SUGGESTED FURTHER READING

For those who want a more detailed description of bioreactor design, modelling and
operation, the following three general references are suggested:
Van’t Riet, K. and Tramper, J. (1991) Basic Bioreactor Design . New York: Marcel
Dekker.
Rosevear, A., Kennedy, J.F. and Cabral, J.M.S. (1987) Immobilized enzymes and cells .
Bristol: Adam Hilger.
Blanch, H. and Clark, D.S. (1996) Biochemical Engineering . New York: Marcel Dekker.
The refences below include review articles and research articles on specific topics.
len, B.R., Charles, M. and Coughlin, R.W. (1979) Fluidized immobilized enzyme reactor
for the hydrolysis of cornstarch to glucose. Biotechnol. Bioeng. , 21, 689–706.
uchholz, K. (1979) Characterization of immobilized biocatalysts. Dechema-Monograph ,
84, 208–223.
ooney, C.L. (1983) Bioreactors design and operation. Science , 219 , 728–734.
mery, A.N. and Cardoso, J.P. (1978) Parameter evaluation and performance studies in a
fluidized-bed immobilized enzyme reactor. Biotechnol. Bioeng. , 20 , 1903–1929.
obayashi, J. and Moo-Young, M. (1971) Backmixing and mass transfer in the design of
immobilized enzyme reactors. Biotechnol. Bioeng. , 13 , 893–910.
aane, N.C.M., de Bont, J.A.M. and Tramper, J. (1987) Biocatalysis in nonaqueous media .
Amsterdam: Elsevier.
evenspiel, O. (1972) Chemical reaction engineering . New York: John Wiley.
lly, M.D. and Dunnill, P. (1976) Immobilized enzyme reactors. Meth. Enzymol. , 44, 717–
738.
uyben, K.Ch.A.M. and Tramper, J. (1982) Optimal design for continuous stirred tank
reactors in series using Michaelis-Menten kinetics. Biotechnol. Bioeng. , 24, 1217–1220.
osbach, K. (1976) Immobilized Enzymes. Methods in Enzymology , 44, 397–450.
azeres, D.M. and Cabral, J.M.S. (1994) Enzymatic membrane bioreactors and their
applications. Enzyme Microb. Technol. , 16 , 1–13.
eusser, E (1961) Theoretical design of continuous antibiotic fermentation units. Appl.
Microbiol. , 9 , 361.
ttig, W. (1982) The present state of fermentation reactors. J. Chem. Technol. Biotechnol. ,
32 , 47–58.
an Sonsbeek, H.M., De Blank, H. and Tramper, J. (1992) Oxygen transfer in liquid-
impelled loop reactors using perfluorocarbon liquids. Biotechnol. Bioeng. , 40, 713–718.
 Bioreactor design 439
Zlokarnik, M. (1972) Ruhrtechnik. Ullmann Encyklopädie der Technischen Chemie , p.
259. Weinheim: Band II, Verlag Chemie.

NOTATION

A1 particle surface area per unit of volume (m−1)

CA concentration of species A (mol.m−3)
CP product concentration (mol.m−3)
Cs substrate concentration (mol.m−3)
Csb bulk substrate concentration (mol.m−3)
Csi surface substrate concentration (mol.m−3)
Cx biocatalyst concentration (mol.m−3)
Da Damkohler number (−)
|Del axial dispersion coefficient (m2.s−1)
|Deff effective diffusivity (m2.s−1)
E enzyme activity (Units)
E0 initial enzyme activity (Units)
F volumetric flow rate (m3.s−1)
Fi inlet flow rate (m3.s−1)
Fo outlet flow rate (m3.s−1)
Hv height of a vessel (m)
Km Michaelis-Menten constant (mol.m−3)
ksl mass transfer coefficient of the substrate (m.s−1)
k2 dissociation constant of the enzyme-substrate complex
(mol. mol−1. s−1)
L reactor length (m)
Prpx overall productivity of the biocatalyst (−)
QP overall volumetric productivity (mol product.mol
substrate−1.s−1)
qs specific product production rate (mol product.mol
biocatalyst−1.s−1)
R radius of a spherical particle (m)
substrate consumption rate (mol.m−3.s−1)
Tv diameter of a vessel (m)
tb residence time in a batch reactor (s)
tl operational life time (s)
 Applied biocatalysis 440

V volume (m3)
vls superficial fluid velocity (m.s−1)
vmax maximum substrate conversion rate (mol.m−3.s−1)
vobs overall reaction rate (mol.m−3.s−1)
X conversion degree (−)
Yov overall yield of product on substrate (−)
Zps dimensionless position of the substrate in a porous support
(−)
dimensionless concentration (−)
dimensionless substrate concentration (−)
dimensionless residence time (−)
κ dimensionless Michaelis-Menten constant (−)
normalized residence time (Units.m−3)
Thiele modulus (−)
effectiveness factor (−)
overall effectiveness factor (−)
 12.
PATENT ASPECTS OF BIOCATALYSIS
PETER S.J.CHEETHAM1 and PHILIP J.D.THOMAS2
1 Zylepsis Limited, 6 Highpoint, Henwood Business Estate, Ashford, Kent, UK
Tel: +44 1233 660555; Fax: +44 1233 660777
2 Eric Potter Clarkson, Park View House, 58 The Ropewalk, Nottingham, UK
Tel: +44 115 955 2211; Fax: +44 115 955 2201; Email: pthomas@eric-
potter.com

ABSTRACT
In this chapter we discuss the rationale for patent systems, the nature of patents
and the procedures by which they are obtained for biotechnological products
and processes. Hopefully this chapter will serve to increase patent awareness
amongst bioscientists, dispel some common misconceptions and highlight some
problems which patenting biotechnological inventions can pose.

12.1 INTRODUCTION

In addition to the obvious important scientific contribution that must be made, the
successful commercialisation of new technological products depends on essential
contributions from other, more commercial and less scientific disciplines (Figure 12.1).
These include patenting procedures. Patent law therefore has wide implications for
research and development (R&D), trade and other corporate and government policy areas
in general.

12.1.1 The Legal Basis of Patents

The Patent System was first established in Europe during Renaissance times. The first
recorded patent was in 1449 to John of Utyjnam for a novel glass making process
 Applied biocatalysis 442

Figure 12.1 Success-contributing factors.

and the first legislation was the British Statute of Monopolies in 1623–4, which defined
an invention as “a manner of new manufacture”. The word “patent” is derived from the
latin “pateo” meaning “to lie open,” i.e. to make information generally available, or more
generally “litterae patentas”=“an open letter,” and the verb “to invent” is derived from the
latin “to find.” The first “biotech” patent was probably to Louis Pasteur in 1873 (US
Patent No. 141,072) who claimed “a yeast free from of organic germs of disease as an
article of manufacture.”
The basic principle underlying the patent system is that the patent holder is granted the
legal right to enforce a monopoly, excluding others from commercially exploiting an
invention without the patent-holder’s permission, for a limited period of time, in return
for the complete disclosure of the invention to the public. However, the granting of a
patent does not imply that the invention does not infringe other patents. Once the period
of protection allowed by the patent law has expired anyone can then exploit the invention
commercially. A patent does not however prevent the use of the patented invention for
private and non-commercial, or purely experimental purposes. Also, if a patented
invention is not being exploited, a third party may apply for a license to exploit it
themselves.

12.1.2 Ownership
In general, an invention made by an employee in the course of his normal duties belongs
to his employer, but the inventor has the right to be named in any patent for the invention.

12.2 WHY PATENT?

 Patent aspects of biocatalysis 443
The patent system aims to promote innovation by giving to the patent owner a monopoly
right for a fixed period. Whilst the patent is in force the patent owner can charge a
premium for its products, free of competition. For patent claims to products, the
manufacture, use or sale in, or importation into, the territory (where the (valid) patent is
in force) of the product is generally a direct infringement of the patent. For claims to a
process, the carrying out of the process in the territory, or the manufacture, use or sale in,
or importation into, the territory of the direct product of the claimed process is generally a
direct infringement. Indirect or “contributory” infringement may occur by supplying an
essential component of the invention even though this component itself is not the subject
of any claim. For example, contributory infringement may occur when a reagent for use
in a patented process is sold specifically for the purpose of carrying out the patented
process even though the reagent may be used for other, non-patented purposes.
It is important to emphasise that a patent does not give the owner or his exclusive
licensee a positive right to use the invention; for example, to use most medical inventions
some regulatory approval is required.
Effective patent protection is essential for high tech industries such as the
biotechnology industry because of the long periods of expensive and unpredictable R&D
necessary to create the next generations of products and the associated economic and
social benefits. For instance, it has been estimated that it costs some $4–500 million to
bring a new drug to market, requiring the synthesis of over 5,000 molecules to discover
one with the right properties, and with an investment of 15–20% of a company’s sales
revenues over many years to finance this R&D. If an effective patent system does not
exist, as soon as an inventor company launches its new product then competitors could
start to sell their own “me too” products, without making any R&D investment. They can
therefore sell their product at a much lower price, placing the original inventor company
at a very grave competitive disadvantage.
Given that pharmaceutical companies have to aim at new products with annual sales of
at about 1 billion US dollars per annum in order to recover all their costs and make the
return on investment expected of them, then the magnitude of the problem caused by any
competitor that has major advantages as regards little or no R&D costs is very apparent.
This is reflected in the performance of countries with markedly different national
patenting systems that offer significantly different levels of protection to innovative
companies. The need for an effective patenting system has seen various countries
reforming their national systems, from Italy in 1978 to Brazil in 1986, and with China
introducing a patent system for the first time in 1985 and strengthening it in 1993. These
reforms have been followed by increases in R&D expenditure, foreign investment, high
tech employment and exports etc. For instance, in 1978 when Italy reformed its system
R&D expenditure was only 123 billion Lira and with only one Italian company was in the
top 100 pharmaceutical companies, but by 1987, R&D expenditure in Italy had grown to
600 billion Lira with seven Italian companies in the top 100 pharmaceutical companies.
By contrast, Canada had an effective patent system until 1969, when compulsory
licensing provisions that gave only nominal royalties to inventors were introduced for
drugs the percentage of sales spent on R&D in Canada fell from 6% in 1969 to just 2.7%
 Applied biocatalysis 444
in 1979. Subsequently, corrective legislation was passed in 1987 and 1993, which had the
desired effects, with R&D spending rising to 7% in 1988, and then to 12% in 1995.

12.2.1 The Importance of Patents to the Biocatalysis Company

For the successful commercialisation of almost all biocatalysis products, patent
protection is essential since without patent protection, and the prospect of a period of
enforceable monopoly in which to recoup investment and make a profit, it is unlikely that
the substantial investment required for bringing a product process to the market will
occur. Frequently, the company making the invention does not wish to, or is not able to,
bring the invention to the market itself. Instead, it enters into an alliance with a partner
which is better able to do this. In this case, and especially in the period where it is still
uncertain that the product will be useful or commercially successful, patents and patent
applications, and the know-how associated with the product or processes, may be the
most valuable asset of the company.
Thus, the presence of patent protection, and the proper protection of proprietary know-
how, are important not least because the value of a company, and its ability to raise
finance, relies heavily on these factors. This success of companies who do not intend to
bring a product to market themselves depend on licensing or assignment of their patent
rights. Typically, a license giving a third party the right to use a patented invention will
include up-front payments to the patent owner with milestone payments and royalties on
eventual sales, whereas an assignment (in which ownership of the patent or patent
application is transferred) may involve a one-off payment to acquire all of the rights in
the invention.
In some cases, the technology may be used to identify new products, but the company
itself may not wish to do this. Having potentially enforceable patent protection which
may be licensed out or assigned may be the only way in which this type of company can
properly exploit its inventions.
Patents can also be of great importance to a biocatalysis company as an information
resource. Often patent publications are the first means by which scientific developments
are reported publicly. Indeed, the UK Patent Office have estimated that one third of
current research work in Europe has already been carried out and published in a patent
filing. Hence, a company can potentially reduce its R&D costs by monitoring relevant
patent publications and, if necessary, coming to an agreement with the patent holder to
allow it to use the new technology described. It can also identify third party patents which
might pose a threat to its own commercial plans at an early date i.e. before significant
investments have been made.

12.2.2 Secret Know-how in Preference to Patent Protection

As part of the bargain between the patent applicant and the state, the patent applicant
must disclose his invention so as to allow a suitably-qualified third party to carry it out.
 Patent aspects of biocatalysis 445
Furthermore, in the USA the “best mode” of carrying out the invention must be disclosed.
Thus, potentially, a patent applicant has to disclose to its competitors what it is doing.
In some circumstances secret “know-how” may be more desirable as a form of
protection than patents. For example, processes for making a product where the process
cannot readily be reverse-engineered, and where there is patent protection for the product,
may better be kept as know-how. Similarly, if exclusivity can be obtained by keeping
secret a particular cell-line or virus or other form of biological material and the prospect
of a competitor being able to make a copy of the material is very low, it may be better to
keep this as know how. The key question to be considered is whether, as a practical
matter, the know how can be kept secret: employees may leave to work for competitors,
or publish scientific papers which inadvertently reveal know how.

12.2.3 Prior Users Rights

An interesting special set of circumstances governing patents apply in the case of so
called “prior users rights.” These rights may arise in some countries to companies who
can prove they were secretly operating a process or making a product, or had made
effective and serious preparations to do so, prior to the patenting of that

Table 12.1 Requirements for a patentable invention.

Novelty
Generally, if the invention has been disclosed to the public, even by one of the inventors, before
the filing date of the patent application the invention is no longer novel and so a patent cannot be
granted. However, major exceptions to this are in the USA and Canada where a one-year grace
period for inventor disclosures applies. Some, more restricted, grace periods apply in other
countries such as Japan and Australia.
Inventive step or non-obviousness
The invention must not be regarded by a person of average skill in the art as a trivial or routine
variant of, or follow plainly and logically from, something which has been disclosed publicly.
Industrial applicability or utility
For an invention to be patentable it must be useful.
Not excluded
Certain inventions are specifically excluded by law from being the subject of a patent. At the EPO
these include animal varieties, and methods of medical treatment (but compounds for use in such
methods are patentable).

process or product by a second party. The prior user right may allow the company to
 Applied biocatalysis 446
continue the specific act which would otherwise be an infringement of the later patent.
However, as the scope of the prior user rights is very uncertain the option of not patenting
inventions and relying on prior user rights is extremely dangerous.

12.3 REQUIREMENTS FOR A PATENTABLE INVENTION

In general, a patent can be granted for an invention which is new (the “novelty”
requirement), is not obvious (the “inventive step” requirement), is commercially or
industrially useful (the “utility” requirement) and is not otherwise barred by law from
being the subject of a patent (for example, at least at the European Patent Office (EPO)
plant and animal varieties are not patentable, and inventions the publication or
exploitation of which are “contrary to morality” are not patentable). The requirements for
a patentable invention are outlined in Table 12.1.

12.3.1 The Patentability of Biocatalysis Products and Processes

In general, any new and non-obvious biotechnological innovation that leads to an
industrially useful product or process is patentable unless it is specifically excluded from
patentability by law. Thus, a novel microorganism as polynucleotide is potentially
patentable; a novel protein or peptide is potentially patentable; a novel method of
preparing a known protein or peptide is potentially patentable; and a new use for a known
microorganism polynucleotide, protein or peptide is potentially patentable.
Patent claims in general, and in the biotechnology field in particular, fall into two
classes: those for “products” (or “compositions of matter”) and those for “processes” (or
“methods” or “uses”) typical examples of product patents include those with claims to taq
polymerase, erythropoietin, tissue plasminogen activator, and nucleic and amino acids
encoding various enzymes. Process patents fall into two categories: those which claim
methods of making a product (i.e. methods of manufacture) and those that claim methods
of using a product (e.g. methods of treatment). Typical method of manufacture patents
are those which, for example, claim specific host cell culture systems for producing a
recombinant protein, methods of making or extracting modified plant materials and
methods of isolating a protein to a particular degree of purity. It is not unusual for a
patent to have claims in all of these categories.

12.3.2 Novelty
The purpose of the novelty requirement is to prevent a monopoly being granted on
something which the public already has in its possession.
A product or process described in a patent application is novel if it has not been
disclosed to the public before the filing date of a patent application that contains that
 Patent aspects of biocatalysis 447
subject matter. (However, grace periods exist in the USA, Canada and certain other
countries such that the inventor’s own disclosure of his invention may not count against
him in certain circumstances.) The novelty requirement is very strict in the sense that for
a prior disclosure to be novelty-destroying it must contain each and every element of the
invention claimed, and it must be a disclosure sufficient to allow the notional “skilled
person” to carry out the invention (i.e. merely stating that the invention has been carried
out previously without giving sufficient details of how to carry it out is not novelty-
destroying).

Products of nature

It is now established patent practice to recognise novelty for a natural substance which
has been isolated for the first time and which had no previously recognised existence.
This is because the isolated form in which the product is claimed has not existed
previously in nature and the product was not available to the world in any practical or
useful sense. Thus, a polynucleotide cloned into a vector, or a cDNA copy of a mRNA,
are novel over the natural situation since these molecules do not exist in nature. Similarly,
a protein expressed heterologously from a polynucleotide and isolated will be novel in
circumstances where, for example, (a) no protein product of the polynucleotide has
previously been identified (this is often the case when polynucleotides are cloned by
virtue of their similarity to other polynucleotides), and (b) where the protein has been
purified from nature previously, but the heterologous expression leads to a different
physical form of the protein (such as modified by glycosylation, or lack of it). It follows
that recombinantly-produced proteins are not novel over their naturally-occurring
counterparts merely by virtue of their method of production: if a protein has been purified
from nature previously, and the recombinantly-produced counterpart is identical to that
isolated from nature, the recombinant protein itself is not novel, although its method of
production, and the polynucleotide encoding the protein, may be novel.
Claims to products defined by a particular state of purity may cause difficulties at the
European Patent Office where it has been held that “a known product does not
necessarily acquire novelty merely by virtue of the fact that it is prepared in a purer
form”. However, if the claim is directed to product defined in terms of a technical feature
e.g. specific activity not present in the less pure known product, the potential difficulty
may be overcome. The situation may be different in the USA where it has been held that
purification of a protein to homogeneity was enough to distinguish a claim over a prior
disclosure of only unpurified mixtures of native protein.
In the case of polynucleotides, it is well established that a previously-known DNA
library containing a cloned polynucleotide does not destroy the novelty of a specific
polynucleotide isolated from that library. Similarly a previously unknown microorganism
present in soil would be novel.
The advent of rapid DNA cloning and sequencing techniques, and the public
availability of nucleotide sequence information from the likes of the human genome
sequencing project, various “expressed sequence tag” (EST) databases, and various
 Applied biocatalysis 448
microbial pathogen sequencing projects, is having an impact on the patentability of
polynucleotides. Since the monopoly right afforded by a patent is defined by its claims, a
claim directed at a specific polynucleotide may not be especially useful if it does not
include the possibility of variants of that polynucleotide which may encode the same or a
functionally similar polypeptide, or which otherwise may be useful. Thus, a claim to a
polynucleotide per se often may take the form: “A polynucleotide which includes the
nucleotide sequence a-b-c-d-e or variants, fusions or fragments of said nucleotide
sequence”. It is now increasingly common to find that a polynucleotide that falls within
the scope of a claim of this form has been disclosed in a public database as a DNA
sequence (and therefore is novelty-destroying for the claim). In this case, it is necessary
to exclude the known polynucleotides from the patent claim in order to make the claimed
polynucleotide novel. Often no function has been shown for the polynucleotide described
in the database, and sometimes no homologous, functional sequences are known.
Whether or not the claimed polynucleotide is “non-obvious” in this scenario depends on
the facts of the case. Even if a polynucleotide is not novel per se because it has been
disclosed for a particular utility, a new and inventive use for the polynucleotide may still
be patentable.
For a prior disclosure to be novelty-destroying for a process, each and every step of the
process will have to have been disclosed. In cases where the prior disclosure does not
explicitly disclose a process step but it is implicit that the step has been carried out in the
prior disclosure, the step is considered to be disclosed. For example, if a prior disclosure
of fermentation process for the production of a recombinant protein in yeast omits to
mention that the fermentation is carried out at between 25°C and 30°C, but the “person
skilled in the art” (see below) would immediately know this from the context of the
disclosure of the fermentation process, this feature is deemed to be disclosed.
New uses of known products may be patented. Thus, if protein X is only known for use
as and anti-coagulant agent, claims directed to its use in an application based on the
finding that it has antioxidant activity, would be novel.

12.3.3 Inventive Step (or Non-obviousness)

The product or process invention claimed in a patent must have an inventive step; that is
to say the invention must not be “obvious to a person skilled in the art”. The rationale for
this is to prevent novel, but trivial or obvious, variants of a known product or process
from being given a monopoly. Inventive step is a separate requirement from novelty: the
question “Is there an inventive step?” is only relevant once the novelty of a claimed
invention has been established. At least at the EPO, for the purposes of inventive step,
and for the purposes of sufficiency of disclosure (see below) the fictional “person skilled
in the art” is deemed to have the same degree of knowledge, and for bicatalysis
inventions is likely to be a team of suitably qualified people e.g. one Ph.D. researcher
assisted by two laboratory technicians fully acquainted with the known techniques
relevant to the technology in question.
Since the mid-1980s the EPO has taken the view that, in the absence of any evidence to
 Patent aspects of biocatalysis 449
the contrary, a polynucleotide lacks an inventive step over a disclosure of a (partial)
sequence of the protein encoded by the polynucleotide, and probably also over a
disclosure of the isolated protein. The rationale is that in these circumstances the skilled
person can, using standard techniques, and without undue effort, obtain the
polynucleotide. Evidence which may indicate that, despite these circumstances, there is
an inventive step includes a showing that (1) there was no motivation on the part of the
skilled person to isolate the polynucleotide despite the presence of the information about
the protein, or (2) there were unforeseen technical difficulties in isolating the
polynucleotide, or (3) the isolated polynucleotide has unexpected, useful properties. The
EPO approach is therefore based on how the polynucleotide is obtained. In contrast, the
approach of the US courts has been based on the inherent non-obviousness of the
polynucleotide product. The US courts have stated that: “It may be true that, knowing the
structure of the protein, one can use the genetic code to hypothesise possible structures
for the corresponding gene and that one thus has the potential for obtaining that gene.
However, because of the degeneracy of the genetic code, there are a vast number of
nucleotide sequences that might code for a specific protein…. Therefore, given the nearly
infinite number of possibilities….he claimed sequences would not have been obvious”.
In general, the EPO consider inventive step on a so-called “problem-and solution”
basis: if the solution to a known problem is not obvious to the person skilled in the art
then there is an inventive step. The single closest prior disclosure is identified and the
obviousness or non-obviousness of the solution is assessed in relation to that disclosure.
The US courts have generally considered whether the prior disclosures would have
suggested to those of skill in the art that they should make the claimed composition, or
carry out the claimed process; and whether the prior disclosure would have revealed that
in making the claimed invention or carrying out the claimed process, those of skill in the
art would have a reasonable expectation of success.

12.3.4 Industrial Applicability (or Utility)

European patent law states that “An invention shall be considered as susceptible of
industrial application if it can be made or used in industry, including agriculture”.
However, the English Court of Appeal held recently that “Something that is useless is not
capable of industrial application. This is because, even though it might be capable of
being made, it cannot be made in industry, industry being concerned with making useful
things”. In addition, there has been a growing view that polynucleotides (or polypeptides)
for which no specific function has been found (such as those identified by essentially
random sequencing methods of cDNA libraries or entire bacterial genomes) are unlikely
to meet the requirement of susceptibility to industrial application: if it is not known what
a polynucleotides or polypeptide does, how can it be useful? This view may become
enshrined in the patent law of EU Member States if the draft European directive on the
legal protection of biotechnological inventions comes into force in its present form (see
below).
The “utility” requirement under US patent law has always been stricter, to the extent
 Applied biocatalysis 450
that many biopharmaceutical patent applications in which a claim was made to a method
of medical treatment using the biopharmaceutical were rejected unless there was
evidence that the method of treatment worked in humans (i.e. clinical evidence of utility).
This very difficult situation for patent applicants has been alleviated to some extent by
new guidelines for examination at the US patent office which stipulate that the examiner
must accept evidence that the specified utility (e.g. medical treatment) is at least credible
based on the in vitro or animal model data available.

12.3.5 Sufficient (Enabling) Disclosure

A patent application must disclose the claimed invention in sufficient detail to allow a
person skilled in the art to carry out the invention. This requirement stems from the
bargain that the patentee has with the state granting the patent: in return for a limited
monopoly, the patentee must disclose the invention to the public in his patent, so the
public can work the invention when the monopoly expires. In the USA the requirement is
even more stringent than in Europe because the applicant must disclose in his patent
application the best mode of carrying out the invention known to him as of when the
application is filed.
Clearly, it is of great importance that enough information is given in the patent
application, but since the fictional “skilled person” may be a team of three including one
Ph.D. scientist (see above), the level of background knowledge of the skilled person is
high. Care must be taken, especially when the practising of the invention requires the use
of biological material such as microorganisms, that cannot be adequately described in the
written description and which is not otherwise available to the public. In these
circumstances it is necessary to make a deposit of the biological material (approved for
patent purposes) in order to comply with the sufficiency requirement. The deposited
material will become available upon publication of the patent application and there is
therefore a risk that if patent protection is not obtained ultimately to cover the invention,
a third party will be given essentially free access to the biological material and invention.
The same is true if patent protection does not extend to all territories in which the
invention is likely to be used. It is possible to stipulate at the European Patent Office that
the deposited biological material is available only to an approved expert after publication
(the “expert” solution). However, this is not available after grant when the deposit
becomes available to all and the ability to prevent a third party from using the deposited
material comes from the patent.
It is not possible, once the patent application is filed, to add further information. The
applicant must therefore rely on the original disclosure which should be as complete and
accurate as possible. Problems may arise if this is not the case. If the application does not
describe the invention in sufficient detail, ideally across the full breadth of the monopoly
claimed, the patent claims may be completely bad or their breadth may have to be
reduced significantly. Thus, as much information as possible on how to carry out the
invention should be given in the patent application. For example, if the patent application
claims the production of a polypeptide in a host cell, and the only working example is its
 Patent aspects of biocatalysis 451
production in Escherichia coli it is highly desirable to include a disclosure of how to
cause its production in yeast cells, mammalian cells, insect cells and transgenic animals.
Furthermore, in general it is not possible to correct mistakes in amino acid and
nucleotide sequences given in patent applications because typically, the error and its
correction are not immediately evident from the application as filed. A surprisingly large
number of patent applications contain incorrectly-determined DNA sequences. This
potential problem can be reduced by depositing a DNA clone containing the relevant
polynucleotide and referring to the (correct) sequence in the deposited DNA without
having to give its sequence. It may also be alleviated by providing details of a method of
obtaining the (correct) DNA molecule and details of how to determine (correctly) its
sequence.
Also, during examination of a patent application the scope or breadth of the claims
often needs to be narrowed because of prior disclosures. Any amendments to the claims
must be based on the contents of the application and so, ideally, the patent application
contains many layers of “fall-back” positions which identify useful features or preferred
variants of the invention ranging from the full scope of the claimed invention which was
initially considered to be justified, down to the specific embodiments described.
Hence, the practice of filing an outline or “informal” patent application and adding
detailed information to “complete” the application is not recommended. The first
application should be as comprehensive as possible.
A general outline of the contents of a patent specification are given in Table 12.2.

12.4 EXAMPLES OF BIOTECHNOLOGY PATENTS

One of the first microorganisms to be patented that had been manipulated by man was the
famous “Chakrabarty” Pseudomonas strain. This was a genetically altered, but

Table 12.2 Contents of a patent specification.

Background
A description of the general field of the invention and of previous public disclosures often called
the “prior art”, it is common to identify the problem that the invention solves, and to indicate the
disadvantages of the “prior art” solutions and advantages of the invention described in the patent
specification.
Description
Description of the invention written to attempt to identify the essential elements and to define the
invention in as broad as possible terms. The description also contains narrower, fall-back
definitions of the invention in case the broadest invention is not new or is obvious from some
 Applied biocatalysis 452

previous public disclosure which was not known to the applicant on preparing the specification.
Figures
Figures are not essential in a patent specification, but it is often the simplest way of including
information such as a nucleotide sequence, or details of a purification process for a protein, or
details of a new gene therapy vector.
Examples
Worked examples are, where possible, included. If worked examples are not available it may be
desirable to give “prophetic” or “dummy” examples which, although they have not been worked,
describe how to carry out a particular aspect of the invention. The patent application must give a
sufficient disclosure of how to carry out the invention without undue burden (the “sufficiency” or
“enablement” requirement). In the USA it is essential to describe the best way of carrying out the
invention known to the inventors when the application is filed (the “best mode” requirement). For
inventions using certain biological material (such as host cells, vectors, hybridomas, bacteria and
the like) it may be necessary to make a special deposit of the material in order to fulfil the
sufficiency requirement.
Claims
The claims in a patent application define the scope of the monopoly protection sought. The claims
may be amended during the examination procedure (for example, their scope may be narrowed
because a previous public disclosure means that they lack novelty). The claims of a granted (or
“issued”) patent define the monopoly allowed to the patent owner. A third party making, using or
selling a product or using a process which falls within the scope of the claims of a granted patent
may be infringing the patent (and so could be sued).

not genetically engineered, Pseudomonad with at least two stable energy generating
plasmids providing separate hydrocarbon degrading pathways. The plasmids confer
catabolic activity towards hydrocarbons so as to enable the organism to degrade alkanes
in oil slicks, offering potential as an oil dispersal and antipollution agent. The
significance of this case is that it was the first confirmation by the US Courts that a
genetically modified microorganism was patentable. This decision disposed of the
objection to the patentability of mircroorganisms simply because they are living;
especially by recognising that this strain was not naturally occurring and that is creation
required deliberate human intervention. It is interesting to note that although the process
using the microorganism was patented in just two years, the patenting of the
mircoorganism itself became a test-case in the US courts and required a nine year legal
struggle before it was granted (US Patent No. 4,259,444).

Table 12.3 Example of the key information given on the front page of a patent.
 Patent aspects of biocatalysis 453

United States Patent [19] [11] Patent Number: 4,889,818

Gelfand et al. [45] Date of Patent: Dec. 26, 1989
[54] PURIFIED Rossi et al., 1986, System. Appl. Microbiol.7:337–341.
THERMOSTABLE
ENZYME
Air et al., 1974, FEBS (Feb.Eur.Biochem.Soc) Lett. 38
(3):277–281
[75] Inventors: David H.Gelfand,
Oakland;
Susanne Stoffel, Fabry et al., 1976. Biochim . Biophys. Acta 453 (3): 228–
El Cerrito; 235
Frances
C.Lawyer,
Oakland;
Randall K.Saiki, Kaledin et al., Sep. 1981, Biokhimiya 46(9):1576–1584
Richmond, all of
Calif.
(Russian text), and Biochem. 46:1247–1254 (English text) .
[73] Assignee: Cetus Kaledin et al., Nov. 1982, Biokhimiya 47(11):1785–1791
Corporation,
Emeryville, Calif.
[21] Appl. No: 63,509 (Russian text), and Biochem. 47:1515–1521 (English text).
[22] Filed: Jun. 17, 1987 Ruttiman et al., 1985, Eur. J. Biochem. 149:41–46.
Kaledin et al., Biokhimiya, (1980) 45:644–651
Related U.S. Application A.Chien et al., J. Bacteriol. (1976) 127:1550–1557.
Data
[63] Continuation-in-part of Ser. Primary Examiner-Thomas G.Wiseman
No. 899,241, Aug 22 1986,
abandoned. Assistant Examiner-Patrica Carson

[51] Int. Attorney, Agent, or Firm-Janet E.Hasak; Kevin R. Kaster;

Cl.4...................................... Albert P.Halliun
C12N 9/00
[52] U.S. Cl. [57] ABSTRACT
..........................435/194; A purified thermostable enzyme is obtained that has
135/183; 935/14 unique characteristics. Preferably the enzyme is
isolated from the Thermus aquaticus species and has a
[58] Field of Search........ 435/183,
molecular weight of about 86,000–90,000 daltons. The
194; 935/14
termostable enzyme may be used in a temperature-
[56] References Cited cycling chain reaction wherein at least one nucleic acid
 Applied biocatalysis 454

U.S. PATENT sequence is amplified in quantity from an existing

DOCUMENTS sequence with the aid of selected primers and
nucleotide triphosphates. The enzyme is preferably
4,683,194 7/1987 Saiki et al.
stored in a buffer of non-ionic detergents that lends
4,683,195 7/1987 Mullis et al. stability to the enzyme.
4,683,202 7/1987 Mullis
OTHER PUBLICATIONS
Kavoev et al., Jan. 1981, J.
Bact 145(1):21–26
Stenesh and Roe, 1972,
Biochim. Biophys. Acta.,
272:156–166.
Klimczack et al., 1986,
Biochem. 25(17): 4850–4855.
3 Claims, 2 Drawing Sheets

A high profile example of biocatalysis patenting is that of Taq polymerase, the

thermostable DNA polymerase enzyme isolated from the thermophilic microorganism
Thermus aquaticus.
The enzyme exists in nature within the microorganism and the invention was claimed
to lie in providing a purified form of the enzyme. The issued main claim of the 1989 US
Patent No. 4,889,818 (see Table 12.3) reads as follows:
Purified thermostable Thermus aquaticus DNA polymerase that migrates on a
denaturing polyacryl-amide gel faster than phosphorylase B and more slowly than does
bovine serum albumin and has an estimated molecular weight of 86,000–90,000 daltons
when compared with a phosphorylase B standard assigned a molecular weight of 92,500
daltons.
The term “purified” distinguishes the enzyme from the product as it exists in nature. In
an attempt to distinguish the claimed enzyme from enzymes which were known before
the patent was filed, the enzyme is defined as having certain characteristics i.e. DNA
polymerase activity, thermostability and a molecular weight within a specified range.
Taq polymerase is useful in the polymerase chain reaction (PCR) and the European
patent rights for the PCR process and Taq polymerase were sold by Cetus in the early
1990’s to Hoffmann La Roche for approximately 300 million US dollars.
The validity of the US and European patents covering Taq polymerase and the PCR
process have been challenged and those legal proceedings seem set to continue for a
number of years before they are finally decided.

12.5 THE PATENTING PROCESS

 Patent aspects of biocatalysis 455
The procedure for obtaining a patent is typically started by filing a national patent
application in a patent office of a country which is party to the Paris Convention (an
international treaty agreeing certain reciprocal patent rights). This application can provide
a so-called “priority date” for the invention disclosed in this “priority” application such
that the patentability of the invention is assessed as of that date. This priority date can
given to further patent applications filed in other Paris Convention territories provided
that these further patent applications are filed within 12 months of the first priority
application. A patent application must be filed before any public disclosure of the
invention since, in most territories, public disclosures before the priority date can be used
to attack the novelty and non-obviousness of an invention described in a patent
application.
In most territories where the patent application is published at 18 months it is possible
to monitor the progress of the examination at the patent office and it possible (at least at
the European Patent Office (EPO)) and the Japanese patent office) to file “third party
observations” to try to prevent a patent from being granted or to try to restrict the scope
of the claims which are likely to be granted. Thus, in most territories with an 18 month
publication regime it is possible to assess the possible impact of a third party patent
application on your own commercial activities and, if necessary, take appropriate steps
such as filing observations, preparing for an opposition or designing around the claims to
be granted. A significant exception is the USA where publication only occurs on grant of
the patent.
Search and examination of an application are necessary in most countries before a
patent can be granted. The purpose of the search is to identify documents (whether earlier
patent applications or journal articles) which are relevant in assessing whether the
invention claimed in the patent application is new or non-obvious; the purpose of the
examination is for a patent office examiner to assess whether the claimed invention meets
all the requirements of patentability and other requirements of patent law. The
examination process is an interactive procedure between the patent examiner and patent
applicant (or more usually his professional representative) in which the patent applicant
may have to put forward arguments and evidence to rebut objections that the patent
examiner may have to the patent application. During this process the patent applicant
may have to amend (i.e. redefine) the claims to his invention.
A patent has a limited territorial effect and so it is desirable to seek patent protection in
those territories where any product is to be made or sold and any process used. Various
international treaties and conventions exist which facilitate the acquisition of patent
protection in many countries (see below).
Once a patent is granted (or “issued”) it will remain in force subject to payment of
renewal fees. Typically, the duration of a patent is twenty years from the filing date. In
general the patent monopoly extends until the expiry of the patent. However, in the
European Union and some other European countries, and in the USA and elsewhere,
there are provisions for extending the term of protection for a specific approved medical
product which is covered by a patent when there have been delays in obtaining regulatory
approval. For example, in European Union countries a “supplementary protection
 Applied biocatalysis 456
certificate” may be available which extends the protection for a given named product
which has marketing authorisation and which is covered by the claim of a patent. In the
USA, “patent term restoration” extends the effective life of a patent. The rationale for
these extensions to the term of the protection is to compensate the patentee for delays in
being able to bring his product to market due to the regulatory authorities’ need for
product approval.
The validity of a granted patent may be challenged by a third party throughout the life
of the patent on the grounds that it does not satisfy certain statutory requirements (for
example, that it is not new or that it is obvious). Typically, the challenge may take place
before the relevant patent office or before the national courts.

12.5.1 Patenting Costs

Patenting can be expensive. Typical Patent Attorney fees for preparing and filing a bio
technological patent application in just one country are from £1,500 to £4,000, depending
on complexity and the cost of progressing to grant of the patent from £1,000 to £3,000
depending on complexity, making a rough average cost to grant of £5,000 per country.
Some countries, such as the USA and Japan, are very much more expensive. Thus, when
additional expenses such as translation or transliteration fees and renewal fees in
individual are added, it is easy to see how a six-figure sum can be required to obtain
patent protection in an effective range of countries.

12.5.2 Differences Between Patent Systems

The patent systems of different countries have important differences which have arisen
for various historical, political and cultural reasons. However, only the differences
between the most important patent granting countries are significant in practice. For
instance, ca 85% of all patent applications are made in just seven OECD countries and
the patents granted in the USA and Japan accounted for over 50% of all OECD patents in
1987.
USA operate a first to invent system rather than the first to file system operated in
other countries.
The US first to invent system allows a person to obtain a US patent if he can prove that
he made the invention prior to another party who filed an earlier patent application for the
same invention. Proof of invention can involve submission of lab note books, making
diligent record keeping extremely important to preserve US patent rights.
In the first to file system, the first person who makes a patent application at the Patent
Office wins the race for patent rights regardless of whether or not he was the first to make
the invention covered by the application.
In Europe and most other countries a disclosure of an invention by an inventor can
invalidate his subsequent patent application for that invention.
In some countries, most importantly USA, there exists a grace period with which an
 Patent aspects of biocatalysis 457
inventor can still make a valid patent application after public disclosure of his invention.
The grace period is beneficial to some applicants, for example it may allow someone to
identify promising research opportunities in the scientific literature and reach agreements
with the original author/inventor or his employer before the cost of patenting is incurred.
A problem with grace periods is that they make searching for third party patent filings
less conclusive since there may be a number of individuals or companies who have not
yet filed a patent application, but could do so under the grace period.

12.5.3 Harmonisation of Patent Procedures

The Patent Co-operation Treaty (PCT) came into existence in 1970 and became
operational in 1978.
The PCT system is a multi-lateral treaty amongst states (over 90) and is useful for
applicants who wish to retain the option of obtaining patent protection in several
countries, but who wish to defer the cost of filing separate national patent applications,
including translation costs. This is achieved by filing a single international PCT patent
application.
In 1995 40,000 international applications were filed and this has been estimated to be
the equivalent of 1,000,000 separate national applications.
The PCT system does not at present grant patents; this is still the preserve of the
national and regional patent offices.
The European Patent Office (EPO) is a supranational organisation for the grant of
European patents. A single European application can designate up to 19 states so that
when granted the European patent can be brought into force in the states by filing a
translation into a prescribed official language.
The contracting states of the European Patent Convention are:

Austria Germany Portugal

Belgium Greece Spain
Cyprus Ireland Sweden
Denmark Liechtenstein Switzerland
Finland Luxembourg United Kingdom
France Monaco
Italy Netherlands

An advantage of the system is that it avoids the administrative inconvenience of separate

national applications and defers translation costs until after grant, so an applicant can
choose not to incur translation costs if no worthwhile protection is obtained.
The number of European patent applications filed at the EPO has increased from
70,000 in 1992 to 86,500 in 1997 (EPO, 1996). In an attempt to encourage small and
 Applied biocatalysis 458
medium-sized companies to use the system, the EPO reduced its official fees. The EPO
claim that this reduces the average cost of a European patent by around 20%. However,
the biggest single obstacle to increased use of the system is probably the cost of
translation after grant, which accounts for nearly 40% of the total cost of a European
patent.
One proposal for reducing translation costs is to publish a detailed abstract and the
claims of the application, in the official language of each designated state so that a full
translation would only be required if the patent was to be enforced. This so-called
“package solution” is not as attractive as it seems because the number contracting states
is set to increase with the likely addition of Poland, Slovakia, the Czech Republic,
Hungary, Bulgaria and Romania.

12.5.4 The Future of Patenting: Draft European Directive on the Legal

Protection of Biotechnological Inventions
The European Commission first proposed a biotechnology directive in 1985. (A
“directive” must be incorporated into domestic law by virtue of appropriate national
legislation (the national legislature choosing how to implement it) whereas a “regulation”
is directly implemented). In 1985 it was believed that a directive was needed to clarify
the protection afforded to biotechnological inventions under European patent law. The
original draft directive was rejected by the European parliament in March 1995, and a
new draft was published in January 1996; a primary objective of the directive remains the
clarification of the protection for biotechnological inventions.
The draft directive was considered and amended by the European Parliament July 1997
and a Common Position was adopted by the European Union Council on 26 February
1998 and was presented to the European Parliament on 9 March 1998. According to the
present draft Member States of the European Union are supposed to bring into force
appropriate laws, regulations and administrative provisions, to effect the directive not
later 2 years from the publication of the final directive in the official journal of the
European Communities. Article 1 of the draft states that “Member states…shall, if
necessary , adjust their national patent law to take account of the provisions of this
Directive” (emphasis added), and so a Member State may consider that its law does not
need changing in order to comply with the directive. Furthermore, at present all European
Union Member States are also members of the European Patent Convention (EPC) (and
have national laws which are harmonised with the EPC), but not all EPC Contracting
States are members of the European Union, and these states will not be bound by the
directive. Thus, it is not clear what the position will be when the EPC is in conflict with
the directive, nor is it clear whether members of the EPC who are not EU members will
be able to get patents for biotechnological inventions under the EPC if that is more
favourable than the directive. In general, it is not clear that the directive, if enacted in its
present form, will change inventions can be protected via the EPO or in EU Member
States.
Article 2 states that “For the purposes of this directive, (a) “Biological material”
 Patent aspects of biocatalysis 459
means any material containing genetic information and capable of reproducing itself or
being reproduced in a biological system;…”. Thus, the draft directive relates primarily to
genetic material.
Article 5 stipulates that the human body, at the various stages of its formation and
development, and the simple discovery of one of its elements, including the sequence or
partial sequence of a gene, cannot constitute patentable inventions. However, an element
isolated from the human body or otherwise produced by means of a technical process,
including the sequence or partial sequence of a gene, may constitute a patentable
invention, even if the structure of that element is identical to that of a natural element
provided that industrial application of a sequence or a partial sequence of a gene must be
disclosed in the patent application. As mentioned above, this reflects generally the view
that is prevailing in relation to the lack of susceptibility to industrial application of
polynucleotides with no known function.
Article 6 indicates what shall be considered unpatentable on the basis that its
commercial exploitation would be contrary to public policy or morality, namely (1)
processes for cloning human beings (2) processes for modifying the germ line genetic
identity of human beings, (3) uses of human embryos for industrial or commercial
purposes and (4) processes for modifying the genetic identity of animals which are likely
to cause them suffering without any substantial medical benefit to man or animal, and
also animals resulting from such processes. Items (1) to (3) are, or are likely to be,
banned by legislation in any event, and item (4) seems to follow existing EPO case law.
Articles 8 and 9 should be particularly welcomed since they will put on a statutory
basis the scope of protection conferred by product and process patents involving
biological material (as defined). Essentially Article 8 indicates that the protection
conferred by a patent on a biological material possessing specific characteristics as a
result of the invention, or a patent on a process that enables such a biological material to
be produced, shall extend to any biological material derived from that biological material
through multiplication or propagation in an identical or divergent form and possessing
those same characteristics. Article 9 indicates that the protection conferred by a patent on
a product containing or consisting of genetic information shall extend to all material
(except the human body) in which the product is incorporated and in which the genetic
information is contained and expressed. Article 10, however, makes it clear that the
protection referred to in Articles 8 and 9 does not extend to biological material obtained
from the multiplication or propagation of biological material marketed in the territory of
a Member State by the holder of the patent, or with his consent, if the multiplication or
propagation necessarily results from the application for which the biological material was
marketed, provided that the obtained material is not subsequently used for other
multiplication or propagation.
The directive now has become law, but it remains to be seen whether any member state
will decide that, at least in relation to the points that would affect the catalysis industry,
their national laws need amending to effect the directive.
It is reasonable to expect that the significant steps taken recently towards
harmonisation of patent laws will continue in the future. Supranational organisations such
 Applied biocatalysis 460
as the EPO and WIPO have been extremely successful and will probably expand in the
future so that the significance of national patent offices diminishes.
In the main, the recent and expected future developments are favourable to patent
applicants. However, significant problems remain. For example, the EPO system requires
that granted European patents must be brought into force by translation into each official
language of each country where protection is sought. This is extremely costly to
applicants. Also, unlike the US Patent and Trademark Office, neither WIPO nor the EPO
encourage small companies and individuals by offering reduced official fees (small entity
fees) compared to the fees levied on larger companies.
The general adoption of a first to file system including a grace period for filing patent
applications following disclosure by an inventor, with small entity fee reductions would
assist smaller companies and may prevent them from losing patent rights to their
innovative products and processes.

12.6 CONCLUSIONS

In the original 1992 chapter a number of patenting problems were identified from the
point of view of an applied biological scientist involved in the commercialisation of
biotechnological products and processes (see Table 12.4). This chapter also includes
views from a European Patent Attorney specialising in patenting biotechnological
inventions.
The reader will see that a number of the previously identified problems have been
partially or wholly addressed. It is hoped that patent systems will continue to respond to
the needs of users, so that patent protection for new biotechnology is widely available in
a straightforward and cost effective manner. This will stimulate the investments needed
to transform today’s experimental results into the beneficial commercial products and
processes of tomorrow.

Table 12.4 Patenting problems in biotechnology identified in 1992.

• In most countries valid patent protection cannot be obtained once knowledge of the invention
has entered the public domain, by written or oral description, by use, or in any other way before
the date of filing the patent application.
• The time, cost, complexity and specialised nature of the procedures required to obtain patent
protection, and also the need for the patent owner to take the responsibility for policing
infringements by bringing cases in the civil courts, act as disincentives to potential patentees
particularly if they are individuals or small companies with very limited resources.
• Because of the international nature of the biotechnology, patent coverage may be necessary in a
number of countries, and this is made more difficult by the legal and procedural differences in
 Patent aspects of biocatalysis 461

patent law between different countries. In effect, these differences can act as non-tariff trade
barriers, especially in biotechnology cases, where the application of existing patent laws to
biotechnology is still in a state of flux, making Patent Office decisions more unpredictable than
in more established areas of science such as chemistry.
• In Europe, the maximum period of patent protection is limited to 20 years from the date of
filing the application for the patent. However, many biologically-based inventions involving
food or pharmaceutical applications require extensive and lengthy testing before being given
regulatory approval. Therefore the actual period for which a company can sell and make
profitable use of a product under patent protection is often much shorter in practice, thus
reducing the return on the research and investment development and, in the longer term, the
incentive to carry out further research and development.
• The need to deposit viable cultures of microorganisms to meet the legal requirement of
providing a sufficient or enabling disclosure of the invention represents a substantial business
risk of giving competitors a head start in their R&D if no worthwhile patent protection is
obtained ultimately. In other technical areas only information describing the invention is
disclosed, and not actual physical materials that either embody the invention or that are very
closely related to the invention.

12.7 QUESTIONS

12.1 What is a “grace period”, in which countries does it apply, and how long does it
last for?

12.2 Distinguish between patents protecting products, processes and applications

(uses).

12.3 Estimate the cost of patenting a new product in the form of three related patents,
each in 6 countries (excluding significant patent court fees).

12.4 Define the terms novelty, inventive step, verbal disclosure, obviousness, claim,
inventor, assignee, foreign equivalent, prior art, priority date, provisional (preliminary)
application, examination, opposition.

12.5 Compare the European and US patent systems; in particular giving two major
differences in patenting procedures.

12.6 Give three reasons why microorganisms are intrinsically difficult to patent, and
discuss whether mutated or genetically engineered microorganisms can be patented.
 Applied biocatalysis 462

12.7 What is the “discovery from nature” argument?

12.8 Who owns a patent, what legal rights are provided by ownership of a patent and
how can these rights be enforced?

12.9 Discuss the advantages and disadvantages of protecting a new product by

patenting it, or by maintaining it as a trade secret.

12.8 REFERENCES AND FURTHER READING

Biotech Patenting Directive No. 98/44/EEC (30 July 1998) Official Journal of the
European Communities No. 1998/L213.
Cook, A.G. (1989) Patents as non-tariff trade barriers. TIBTECH , 7 , 258–263.
Crawley, P. (1991) Patenting in Biotech, for non-specialist . London: Bio Ind. Assoc.
Crespi, R.S. (1982) Patenting in Biological Sciences . Chichester: J.Wiley & Sons.
Crespi, R.S. (1985) Biotechnology and Patent protection: an international review . Paris:
OECD.
Crespi, R.S. (1988) A Basic Guide to patenting in biotechnology . Cambridge (UK):
Cambridge Univ. Press.
EPO (1996) Annual Report.
Gallafent, R.J. et al. (1984) Intellectual Property, Law and Taxation .
Grubb, P.W. (1986) Patents in chemistry and Biochemistry , 2nd edn. Oxford: Clarendon
Press.
Hacking A.J. (1986) Economic Aspects of Biotechnology . Cambridge (UK): Cambridge
Univ. Press.
Miles Gaythwaite, D. (1991) Intellectual Property and Know-how. In Biotechnology: The
Science and the Business , edited by V.Moses and R.E.Cape, pp. 69–88. Chur:
Harwood.
STBS Ltd (1991) Protecting and Exploring Biotechnological Inventions . Reading (UK)
or New York.
UK Patent Office. How to Prepare a UK Patent Application , State House 66–71, High
Holborn, London, WC1.
UK Patent Office. Patent Protection . Ibid.
WIPO Publication No. 884(E) (1995) The last twenty five years of the PCT , ISBN 92–
805–0601–3, Geneva.
A more detailed discussion on EPO and US case law relating to biotechnological
inventions can be found in “The European Patent Office’s Case Law on the Patentability
of Biotechnology Inventions”, Jaenichen, H.-R., Carl Heymanns Verlag KG, Köln,
Germany.
Various Websites provide information on patents. These include the European Patent
Office site (www.european-patent-office.org) which includes a non-exhaustive directory
of other patent Websites; the MicroPatent® site (www.micropat.com) ; the IBM Patent
 Patent aspects of biocatalysis 463
Server (www.patent.womplex.ibm.com) which gives access to over 26 years of US patent
specifications; the UK Patent Office (www.patent.gov.uk); the US Patent and Trademark
Office (www.uspto.gov); and Biotechnology and US patents
(www.nal.usda.gov/bic/Biotech-Patents).
 13.
SOME COMMERCIAL AND FINANCIAL
ASPECTS OF BIOCATALYSIS RESEARCH
AND DEVELOPMENT PROJECTS
PETER S.J.CHEETHAM
Zylepsis Limited, 6 Highpoint, Henwood Business Estate, Ashford, Kent, UK
Tel: +44 1233 660555; Fax: +44 1233 660777

ABSTRACT
This chapter attempts to explain some of the economic and commercial
influences that affect the development of biocatalysts into successful
manufacturing processes and products. In this respect, biocatalysts are not
special, and are subject to just the same effects as any other developing
technology. In this chapter, I attempt to discuss many of these important
influences, such as process costing, sensitivity analysis and product profitability
appraisal. I would especially like to emphasise the following points:
This chapter is intended to be read in close conjunction with the more
detailed Case Studies of particular processes and products provided else where
in this book, as well as hopefully being of interest in its own right.
To succeed a bioprocess must make a product with substantial and sustained
competitive advantages compared to both its direct competitors, and to other
products that can be used to substitute for it. Therefore, it must have
characteristics that customers will value sufficiently highly for them to switch to
buying the new product and even pay more for it. This competitive advantage
may be a substantially lower price; a superior performance, or some other
advantage such as greatly increased convenience or marketing image.
However, even if the product really does offer such advantages, its price is
still very important. Hardly anyone will buy a product if it is too expensive,
however superior it is, irrespective of whether the customer is an industrial
company or a high street shopper. Similarly few customers value new
technology so much that they will pay more for it, unless it gives them some
really tangible benefits. Thus, it is very important to remember that “price is a
key part of the product” and has to be right (which usually means got
sufficiently low).
 Commercial and financial aspects of biocatalysis research 465
Then the market for the product has to be big enough to justify the costs of
R&D and subsequent commercialisation necessary to bring it to market. The
money spent (invested) must be regained from sales revenues in a reasonable
time just to reach break-even point. Only then can real profits be made.
The last point is that for success a wide range of quite different aspects of the
process need to be optimised. In addition to the science and technical basis of
the process and the product, raw materials supplies, manufacturing, safety and
regulatory, product formulation, marketing and many other essential aspects
need to be successfully accomplished. Failure in just one could mean failure for
the whole venture, irrespective of how successful the initial science had been.

Table 13.1 Some commercial driving forces for the use of biocatalysts.

• to produce purer products, especially stereochemically pure products

• to enable the synthesis of new molecules with high activity and functionality
• to develop lower cost processes and products
• to enable to development of “green” processes with reduced environmental impact
• to allow the use of labile reactants
• to produce “natural” products

13.1 INTRODUCTION

In the context of this chapter biocatalysis is the use of enzymes or enzymes still
associated with their parent cells, to carry out defined chemical reactions under controlled
conditions, so as to efficiently convert raw materials into commercially more valuable
products. Some of the commercial driving forces for the use of biocatalysts are listed in
Table 13.1.
Biocatalysis technology only aquires real credibility when it is able to create profitable
products and efficient processes. Therefore the following sections are an attempt to
outline some of the commercial and financial factors that are important in influencing the
success or failure of attempts to commercialise new processes and products based on
research and development in applied biocatalysis, such as the case-studies presented
earlier in this book (chapter 4). The use of financial appraisal techniques are a vital guide
to technologists to enable then to carry out R&D most effectively, concentrating research
effort on those aspects that can have the biggest impact on product and process costs and
quality.
In order to produce a new product, process or service starting from a new idea a
 Applied biocatalysis 466
company must carry out market evaluation, technical research, optimisation,
development, scale-up, application studies and also financial appraisal and market
development, before it can establish manufacturing, sales and distribution facilities
(Figure 13.1). As in any commercial or technological activity, new projects or
modifications of biotransformation projects require the investment of valuable resources
in the forms of labour, capital, etc., in the expectation of generating benefits in the form
of profits resulting from the sale of products, savings in the operating costs of the
process, or an improvement in the companies’ share of the market. These can be expected
to increase a company’s profits or to lead to an increase in the company’s share price.
Thus, biotechnology can be viewed as one method of “adding-value” to products or
services, and hence enhance the value of the companies supplying those products and
services.
Such activities may range from a comparatively minor alteration to an existing
production plant producing an existing product, right through to an entirely new product
produced by a novel process in a completely new and purpose built factory. An
interesting and important exception are projects which do not generate financial profits,
but which need to be carried out for some other reason, for instance, so as to meet
statutory pollution controls or safety measures. In these cases the objective is to minimise
the expenditure required in order to achieve the required standards.
Commercial and financial aspects of biocatalysis research 467
 Applied biocatalysis 468

Figure 13.1 Development of a new product.

Companies’ motives in conducting R&D must be considered, not least because many
companies choose to carry out no R&D at all! For instance some companies operate in
areas that are not “research sensitive” and where new technical developments are unlikely
to take place. A more frequent strategy is for a company to expand not by R&D-driven
innovation, but by the acquisition of other companies, thereby increasing in size, range of
products etc. This strategy can be especially prevalent in times of high inflation since the
“book value” of potential acquisition targets may become significantly undervalued,
making them more likely to be taken-over.
Successful R&D culminating in new processes and products can enable:

1. Existing products to be manufactured more cheaply and/or efficiently, thereby

increasing company profits and/or market-share; or the process can be improved, for
instance by developing less polluting methods and more environmentally friendly
processes.
2. Improved versions of existing products, for instance with additional functionality, can
be produced.
3. New products can be produced, either as a logical extension of the company’s existing
product range (core business), or as new ventures into types of product not currently
produced. In this respect, novel science and technology can be very valuable by
breaking down some of the prevailing commercial and technical limitations, and thus
provide new “entry points” for companies hoping to enter a field of business for the
first time.

However, in order to achieve these objectives a good interpretation of corporate business

needs into achievable R&D objectives is essential. In some cases a supplier of a raw
material, intermediate or component may choose to “forward-integrate” their business
into production of the finished product. In other cases a manufacturer may wish to
“backwards-integrate” into the production of the raw material(s) required for its current
product(s). In either case it is obvious that in order to achieve success a company must
make products available to its customers with useful functional properties; and at a price
that allows cost-effective use of that product in the customer’s own products. That is, the
product must give the customer’s products a competitive advantage.

13.2 APPRAISAL TECHNIQUES

In order to make the best possible use of valuable financial and other resources, and to
minimise the uncertainty involved in making irreversible decisions, appraisal techniques
are vitally important. These can be technical, legal, regulatory, commercial or financial in
nature, for instance market forecasting and patent protection. Critical path analysis is
 Commercial and financial aspects of biocatalysis research 469
especially useful to identify interdependent operations and rate-limiting steps, slack-time
periods, and to generally optimise the sequence of events and their timing so as to
minimise delays and expenditure and to maximise chances of success and profitability.
Analysis of competing and potentially competing technologies is also important so that
new products are not “pipped at the post”.
Appraisal techniques help to balance the elements of “technical push” and “market
pull”. Despite some of the Jargon used, many of these concepts are quite simple and so
these techniques were probably first established in rudimentary forms when the first
commercial activities began.
The objective is to use the limited data currently available to make accurate
predictions, and thereby minimise future risk and maximise future profits. These
appraisal techniques help to ensure that scientists are not required to achieve technically
impossible goals; and on the other hand, help to ensure that scientists do not pursue
products and processes that are unlikely to prove economic, or which do not fit likely
market requirements. Financial appraisal techniques serve to compare the resources
required to carry out the project, the likely benefits of the project if successful, and the
risk involved: that is the probability of successful obtaining these benefits. They can also
be used to discriminate between different alternative uses for capital, labour and other
resources, and also serve to make process (es) as efficient and profitable as possible. For
instance, investment of capital in money markets, especially when high interest rates
prevail, is a more flexible, safer and more easily assessed risk than investment in a new
factory, but such safe investments will often be less profitable than if the new product is
successful. Appraisal techniques can help to determine whether to proceed with a project,
determine how much resources to commit, and maximise the chance that potential profits
will be sufficiently great to justify the investment in R&D etc. needed to commercialise
the project. In short, the market for the product must be sufficiently large to justify all the
R&D costs, and also all the other necessary expenses such as safety testing and patenting
etc. that are needed to develop and commercialise the product and then provide good
profits over a comparatively short time period.
Many other factors also have to be taken into account, such as the likelihood of
achieving strong patent protection, the possibility of licensing technology from other
companies and the probability of achieving successful regulatory approval, especially in
the case of new food and pharmaceutical products. Another important factor is how well
a proposed new products fits into the companies’ existing business, whether difficulties
with competitors’ patents can be expected, consumer and market trends, taxation factors,
development grants, the availability of raw materials, the cost of distributing products,
the vulnerability of the process to interruption by breakdown or strikes, not to mention
consumer perception and political factors.

13.3 WHAT IS A PRODUCT?

It is very important to define properly what a product is in practical terms. A product is

 Applied biocatalysis 470

not steroids, enzymes, -blockers etc. These are classes of products. A true product has
to be defined in terms of specific end-use functions, i.e. as it is bought and used by
specific customers. This means that one material can be sold in different forms depending
on its application, for instance xanthan gum is sold in different grades,

Figure 13.2 Market segmentation. Product X, Y, Z etc. refer to the particular

tailored products targeted at specific market segments A-F, which
require individual applications, formulations etc., so as to optimise
performance for each target product.

of quite different purities and prices , depending on their various applications. These
include cheap impure material for tertiary oil recovery, and purer and more expensive
gum for food applications. Thus, when accurately defined, each particular product is
characterised by a number of listed descriptors such as price, chemical composition,
physical form, applications, etc. These depend critically on what is required by the
customer. Customer requirements can differ a lot, for instance depending on their socio-
economic status, nationality, culture, etc. For instance, the sales of toilet soap, fabric
softeners, hard soaps, etc. vary dramatically both in relative and absolute terms,
depending on the GNP of the country and also cultural factors.
Therefore the “real” product needs to be defined much more precisely to meet
customers requirements than does the “generic” product (Figure 13.3). In addition the
producer needs to consider the “augmented product”, in which additional features have
been added. In developing such augmented products the manufacturer is moving towards
 Commercial and financial aspects of biocatalysis research 471
the “potential product” which is the ultimate product that could be produced in the
context of current technology and know-how.
A market can be broken-down (segmented) according to various criteria so as to define
those sub-markets that are most easily targeted by sales and marketing experts (Figure
13.2). Segmentation can be according to geography, such as home country market,
Europe, N. America, Japan, etc.; according to price and specification, such as technical
grade, pharmaceutical or food grade, or according to end-use, often depending on the
precise target of the particular companies the product can be sold

Figure 13.3 The total product concept (from Levitt, T.; The Marketing
Imagination, Collier Macmillan, 1986). Note: The dots inside each
ring represent specific activities or tangible attributes. For example,
inside the “Expected Product” are delivery conditions, installation
services, postpurchase services, maintenance, spare parts, training,
packaging convenience, and the like.

to, for instance pharmaceutical or agrochemical uses. Similarly, the product can be
ranked according to whether a basic product is to be produced comparable to that already
produced by competitors, or whether a product with additional performance features or
functionalities is to be aimed at (Figure 13.3). It is also well worthwhile to conceptualise
what the ultimate product and its competitive advantages should be.
The three main factors to be considered are: the costs involved; the probability of
technical and commercial success; and the likely profitability of the venture. Formal
assessment is not of course a completely exact exercise and should be used judiciously to
support the technical and business judgement and intuition of experienced staff.
Obviously, the higher the cost of the project, and especially the greater the technical and
 Applied biocatalysis 472
commercial risk, then the greater the effort that must be put into the appraisal. Another
factor to be borne in mind is the strategic value of the process/product. Thus a project that
is very close to a company’s core business, will be examined in a very different manner
as compared to a second project which constitutes a more speculative diversification
away from the company’s existing activities. Also, if important new generic technology
is being developed that is of likely future use to the company then the potential value of
this new technical capability should also be taken into account. Hence the importance of
good market evaluation as well as the technical appraisal of projects before they are
started.

13.3.1 Fixed Capital and Variable Costs

Costs are usually divided. Firstly into fixed capital costs such as those generated by
building and buying equipment and its installation, and secondly into variable costs such
as raw materials. There is not always a clear distinction between fixed and variable costs,
thus for instance permanent staff tend can be regarded as a fixed cost, and contract staff
and overtime costs to be classed as variable costs. The capital costs of constructing new
plants are very considerable. Thus is it estimated that a corn wet mill of 500,000 tonne/a
grind capacity will cost well over $108 to build and equip. Capital (or fixed) costs can
also include large interest repayments on loans for building and equipment, although
these can alternatively be classed as a running expense, and therefore included in the
working capital. Capital expenses are depreciated over a finite period of time, so as to
spread costs. This period varies from company to company, but is usually 10–20 years,
although computer equipment is depreciated more quickly because rapid technical
innovation in this field makes equipment obsolete more quickly. Economies can be
achieved by operating versatile processes so that the equipment can be used to make a
number of different products. Thus the capital costs can be spread over several products.
Such an approach is particularly valuable when producing seasonal products or a range of
low tonnage speciality products. Ability to easily vary the scale of production to match
demand for the product is also very advantageous so that product does not have to be
overproduced in times of low demand and can be easily expanded when demand
increases.
Secondly, there is variable expenditure (production, recurrent, working or operating
costs) which includes the costs of supplies and stocks of raw materials, emphasising the
need for efficient monitoring of the mass-balance of the process, energy, labour and
overheads. Overheads include direct overheads such as rates, insurance, administrative
staff etc. and indirect overheads such as a share in head office salaries and upkeep. This is
an important distinction because whereas local management can have little direct
influence in overheads costs and capital and utility expenses, they can have a big control
over variable expenditure.
Variable costs become relatively more important, constituting an increasingly large
proportion of the total costs, as the scale of operation increases and especially as the
manufacturing process is simplified, thereby decreasing the requirement for expensive
 Commercial and financial aspects of biocatalysis research 473
equipment. Thus processes involving high throughputs of cheap substrate and carrying
out efficient conversions using simple and therefore cheap equipment, and with a
minimum of product isolation minimised by maximising the concentration of product
achieved and the degree of conversion of raw materials to products, and by minimising
side-product and impurity formation, and thus the need for expensive concentration,
separation and isolation technologies. These parameters are obviously most important for
low cost products. The product purity required therefore has a big effect on final product
costs because of the increased cost of down-stream purification steps. Energy costs can
be minimised by reducing energy intensive operations such as high temperature
operations, sterilisation, high-speed stirring, refrigeration, evaporation and drying.
Note that because fixed costs are proportionally more important for small-scale
operations the influence of research and development work on reducing the unit process
costs on smaller scale processes will often be significantly different than when applied to
larger scale operations. Also equipment will not necessarily be used under conditions
which maximise its performance, but rather under conditions that minimise cost, thus for
instance a fermenter will not be operated at maximum agitation rates as the power
required will be too expensive. Other savings can be made by finding uses for wastes,
such as the spent biomass from fermentations, and also for any out-of-specification
“product” that is made.
In calculations it is also important to take into account any income that may be
expected from development grants etc. tax liabilities and the depreciation that takes place
on capital items expenditure. This second sum needs to be set aside so that provision can
be made to replace capital items at the end of their lifetime, especially as the effects of
inflation may be different for the various items involved in the costing. Once the capital
equipment has been depreciated then profitability can rise, since only variable costs have
to be covered. Thus for instance the selling price of a product can be decreased, even
down to the production cost, in order for example, to force a new competitor out of
business.

13.3.2 Economies of Scale

Many factors act together to determine the optimum scale of a process. These include the
demand for the product, competitors share of the market, any technical limitations on the
size of operation and also economies of scale effects. There is an approximate
logarithmic relationship between the unit production costs for a product and the volume
of production, whereby considerable economies of scale can be achieved. If the costs of a
process of one size (C’) is known then the costs of larger or smaller factories (C’) can be
approximately obtained from the relationship C’= (or n0.67), where n is the
scale-up ratio, i.e. n=2 for a plant that is twice as big. Alternatively, a graph of log capital
costs vs. log of plant capacity gives a straight line with a slope equal to the scale-up
factor (n). The power term varies from case to case, but is invariably less than one. This
scale effect is one reason why unit production costs are inversely proportional to the scale
of manufacture. For example, most amino acids are expensive and can only be used in
 Applied biocatalysis 474
high value products, such as specialised diets for patients. However, a few amino acids
have found much larger markets, for instance monosodium glutamate (MSG) as a flavour
enhancer, aspartic acid and phenylalanine for aspartame manufacture, and lysine and
methionine in animal food supplements. As a consequence of these economies of scale
and successful research to improve these processes, bulk costs are much cheaper. MSG,
phenylalanine and lysine are produced by fermentation and L-methionine by enzymic
resolution of chemically synthesised racemic product. Therefore the size of the
production facility that will be most economic, and that will be least risky as a new
undertaking, will be a compromise between factors such as the expected demand for the
product, economies of scale, and technical factors that may limit the size of some pieces
of equipment etc.
Obviously large investments entail undertaking larger risks. Also the economies
achieved by large-scale operation are only useful if such large capacity plants can be used
to capacity, although seasonal variations in demand often have to be allowed for,
otherwise many overhead charges are incurred without sufficient profits from sales to
offset against them. However, with large capacity plants or high value products even
small improvements in processes can be very profitable. Moreover, a large-scale of
operation may make a product potentially less attractive to competitors, because of the
large investment they would have to make to enter the field and compete effectively. The
very large scale manufacturing plants for the High Fructose Corn syrups are a good
example of these effects.
The construction of large new plants can even have a big effect on supply-demand
relationships for a product and hence the price of that product. When production capacity
lags behind the demand for product, prices will be high. But if a big new plant is opened
then production capacity will be increased significantly so that supply now exceeds
supply and product prices will fall.

13.4 DEFINITIONS OF TERMS

Some of the other terms frequently encountered include the following:

Full costing. This form of costing takes all expenses into account, including direct
costs such as raw materials, and also indirect costs—such as head office expenses,
insurance costs, design fees, contingency provisions, services, maintenance and
depreciation.
Marginal costing. (Figure 13.4) This is a form of costing that takes into account only
the variable costs, which depend on the scale of operation such as the cost of raw
materials, and also the fixed costs such as the cost of equipment. Marginal costing is
especially useful in planning, enabling for instance, the establishment of cost of product
versus scale of production relationships. Marginal costing can also be used in times of
difficulty, for instance due to reduced demand, in which case the replacement of capital
equipment is not allowed. So the factory may be able to continue in business, but not to
generate the capital required to eventually replace equipment when it wears out or
 Commercial and financial aspects of biocatalysis research 475
becomes obsolete. The precise definition of a marginal cost is: The cost required to
produce just one more, or one less item of production. Thus the marginal cost can be very
low if it can be made on a production line operating under peak capacity and with
overhead costs already paid, or the marginal costs can be very high if the extra item of
production can only be produced by building a new factory because the existing plant is
already working to full capacity.
Raw materials (substrate or feedstock) costs are usually a significant cost element in
most biocatalytic processes. Their importance highlights the importance of good buying
policy and skills and, when possible, the effective use of forwards buying and “futures”
markets to obtain cheap raw materials etc. despite variabilities in their
 Applied biocatalysis 476

Figure 13.4 Marginal costing. From A.Hacking, 1986. The concepts of

 Commercial and financial aspects of biocatalysis research 477
variable costs, fixed costs & break-even point etc. can be expressed in two
different ways, depending on whether the fixed expenses are
accounted for first, or added pro rata to the variable expense. From:
Economic Aspects of Biotechnology. Reprinted with permission from
Cambridge University Press. UK.

prices and availability. Finding ways to use novel raw materials is sometimes a useful
role for R&D so as to enable diversification of a process away from complete reliance on
its traditional types of raw material. The cost of the raw materials relative to the product
value is also very important. Obviously if this difference is not large then the process
must either be very efficient and achieve high degrees of conversion of substrates into
products; or be carried out on a very large scale in order to achieve large enough profits.
Utilities include the costs of electricity, generating steam, cooling water etc., and
emphasises the importance of good equipment and efficient processing and “energy-
audits” for energy conservation and consequent financial savings.
Overheads include head-office expenses, insurance etc. which will usually be spread
over all the operations of a company (cost centres).
Capital costs. Industry is becoming increasingly capital intensive because of the use of
increasingly sophisticated equipment, hence the importance of this topic. Also
biotechnology requires expensive equipment, for instance for sterile or hygienic
processing, and high precision control and data acquisition and processing equipment are
becoming standard. Capital costs can have a very big influence on investment decisions.
This is because large sums must be spent well in advance of any profits being generated,
and interest must be paid on the borrowings made to finance capital expenditure for
several years before any revenues are reached.
Preliminary estimates of capital costs can be made by means of ratio estimating using
the so-called “Lang factors”. Thus for instance the installed and functioning cost of
process equipment is ca 1.5 times the purchase/delivered cost of new equipment. This
factor is virtually independent of plant size. Equipment costs are usually written-off
(depreciated) over about a ten-year period, but this varies with the type of equipment. For
instance, computers are usually written-off over a few years, because the rate of technical
innovation in computer hardware is so rapid. However, fermenters are often depreciated
over as long as 15 years or more. Since the purchase of new equipment is an irreversible
and expensive decision, the option of leasing should also be explored. In some cases there
is a balance between capital and recurrent costs, particularly when a more expensive
piece of equipment can operate more efficiently. For instance multiple effect evaporators
make much more effective use of energy, in the form of steam, than do cheaper single
effect evaporators, but are more expensive to buy.
Fixed capital investment is the sum of direct and indirect capital outlays.
Working Capital (capital employed) is the money that has to be spent in order to start a
new project or process, before any income from sales or product is received. This money
is tied up, and therefore unavailable for other uses, in the form of current liabilities such
as purchased raw materials and borrowings etc., less current assets such as unsold
 Applied biocatalysis 478
products and other stock and creditors. An important objective of any business is to
minimise its working capital. Therefore it is necessary to minimise the time taken to
convert raw materials, labour and overheads into products and therefore profitable
income, and also to minimise the time that raw materials and product are held in stock
prior to use or sale respectively, hence the importance of so-called “just-in-time”
production and distribution methods. The true value of the savings achieved by reducing
the working capital can be expressed by comparing them with the costs of achieving such
reductions in working capital, e.g. the costs of buying new equipment etc.
Joint products. Some processes produce more than one product, for instance semi-
synthetic antibiotic factories and starch processing plants. These joint products are not
usually accounted for separately until the stage of the process at which the separate
products are produced, which is termed the split-off point. Until the split-off point all the
costs incurred are usually allocated between the various products on a pro-rata basis.
By-product credits. Often large quantities of side-products of relatively low value are
produced, such as the biomass from fermentations, which is often used as cattle feed or as
a soil-conditioning agent. Corn refining operations are an excellent example, as nearly
every item in the mass balance is sold, from the corn steep liquor as a fermentation media
component to the CO2 generated from the fermentation of glucose syrup into ethanol
which is used to enhance the growth of tomatoes. The receipts from the sales of the by-
products (side-products) are simply subtracted from the feed-stock costs, or added as a
credit item at the end of the costing. A company’s ability to develop markets for a side-
product is of course limited, because the amounts of side-product available are dependent
on the amounts of the main product produced.
Capital is required to meet both running expenses and particularly new capital items.
Sources are usually retained profits, equity (share capital) and borrowings.
Venture Capital is finance provided by investors in return for an equity stake in the
company in order to initiate a new project or company. Finance is subject to varying
conditions, but this method of financing research is generally used for longer term, more
basic and thus speculative high risks research projects. Venture capital was the original
“seed-corn” money provided to set up many of the new US biotechnology companies
such as Amgen, Genentech and Cetus. Thus for some companies maintaining investor
confidence is often of more immediate importance than achieving profitable results.
Many venture capital investment companies will consider biotechnology proposals, some
even specialise in biotechnology investments.

13.4.1 Evaluation Parameters and Criteria

How can we estimate whether a project is likely to be profitable? How can we decide
which projects are likely to be most profitable? For instance a low return on a large
investment may given more profits than does a large return on a smaller investments.
 Commercial and financial aspects of biocatalysis research 479

Figure 13.5 Cumulative cash flow diagram. Adapted from D.H.Allen, 1980. A
Guide to the Economic Evaluation of projects. Reprinted with kind
permission from Institution of Chemical Engineers.

13.4.2 Pay-back Time

Pay-back time is the period required to recover the money initially invested in the project.
Thus, the pay-back time is represented by the period (1–6) in the cash-flow diagram
(Figure 13.5). The pay-back time is not just the period required for profits to be first
generated—sufficient profits then have to be made to cover the initial expenditures made.
 Applied biocatalysis 480
For instance, if new equipment is bought for £100,000, and enabled savings of 20,000 £/a
to be made, then the pay-back time on this investment is 5 years. Before this time the
investment project is in debt; and afterwards it is in profit. Thus the operational life-time
of the equipment is of importance; or in other projects it is important how long the
product continues to sell profitably (the product life-time). The pay-back time is
obviously only a guide. It certainly does not measure profitability, and ignores what
happens to a project after the pay-back time. Nevertheless it is a useful rough indicator,
so that by setting a benchmark pay-back time unsuitable project proposals can be
screened out, on the basis that the shorter the pay-back time, then the more attractive the
project is likely to be. Pay-back times are most useful for longer-term projects, where
experience indicates that a fairly good correlation with internal rate of return values (IRR,
see below) exists.

13.4.3 Return on Investment

Return on investment (ROI), Return on Capital Employed (ROCE) and Internal Rate of
Return (IRR) are terms all referring to the profitability of a project or product in terms of
the original investment made, i.e.:

(13.1)

Also,

(13.2)

where n=number of years the project is assessed over.

ROI values are heavily influenced by capital costs. When estimating ROI in order to
evaluate projects at an early stage in their development, it is important to obtain as
accurate an estimate as possible of the market for the product, which can be particularly
difficult in the case of new products which do not yet “exist” in the market place.
Comparison of the ROI with likely returns from alternative types of investment should
always be made. The ROI expected by different companies can vary greatly. For
instance, drugs and vaccines etc. are usually expected to generate a much higher ROI
than commodity chemical products or basic foods, not least because of the very high
R&D expenditure that must be recovered because of the greater risks involved, and
because the expectations of profitability of pharmaceutical companies are greater than,
for instance, manufacturers of commodity chemicals or bulk foods.
 Commercial and financial aspects of biocatalysis research 481

13.4.4 Present Value and Net Present Value

Other parameters are more difficult to calculate. However, they can be very much more
useful because they take into account two other important aspects. Firstly, the pattern of
flow of money into and out of the business (cash-flow) since items of income (revenue)
and expenditure (outgoings) can vary a lot with time. For instance, as can be seen in
Figure 13.5, initially expenditure greatly exceeds income, whereas later on once the plant
is operating and product is being sold income exceeds expenditure. Secondly there is the
“time value of money”, that is money received now is more valuable than the same sum
of money received in a years time, as it can be invested and interest will have accrued on
it by the end of the year. Hence, early income to a business is valuable as it can be
invested in the business and so saves having to borrow money to buy equipment, pay
wages, etc. Thus the value of money is not completely absolute, but is a function of time.
The future income likely to be generated from investing in a new project can justifiably
be compared with the interest that could alternatively be earned by investing with a bank
the same sum required to research and commercialise the new product. At an interest rate
of 10% a sum of $106 will be compounded, and will have a value of $1.1×106 after 1
year, $1.21×106 after 2 years, $1.33×106 after 3 years, etc. That is a value of $(1.1)n×106
after n years. Therefore the proposed new process or product must generate at least this
amount to be a competitive investment, especially as investment in a bank is
comparatively risk-free, whereas there is considerable risk associated with the successful
development of new technology, that is that it may prove unsuccessful and produce no, or
reduced profits. If the proposed project is estimated to have a future value of $X after n
years, its value in current terms (i.e. its present value) can be obtained by discounting
back from year n to the present. Thus at a discount rate of 10% the present value of the
project would be $X/(1.1) n , or more generally

(13.3)

(discount factor interest rate rate of return). Obviously, in this example if the present
value is only equal to or less than $1×106 then it is not a good use of financial resources.
In fact, in order to allow for the risk factor involved, a present value considerably higher
that $1×106 would have to be obtained, in order to justify investment in a new project.
In addition a project needs to repay its costs. Therefore it is more realistic to use cash-
flow values rather than income values to calculate present values. A series of present
values can be generated at time intervals by projecting the cash-flow of a project into the
future. Obviously the early ones will tend to be negative, but hopefully they will be more
than balanced by large positive present values later on when the product has become
successful. NPVs (net present values) are additive. Therefore, the net present value is the
summation of all the individual component cash-flow derived present values, and is
 Applied biocatalysis 482
obtained by adding together all the PVs obtained for each year’s cash-flow i.e.:

(13.4)

Figure 13.6 The vertical dotted lines indicate the capital recovery periods, that
is the number of years the project must run before initial capital is
fully recovered and has earned its minimum 8% net of tax return each
year on the capital balance outstanding. (Adapted from Franks, J.R.
& Scholofield, H.H. 1977 Corporate Financial Management 2nd edn.
 Commercial and financial aspects of biocatalysis research 483
Gower Press).

Discounting can be calculated either continuously, or at discrete intervals, which will

give slightly different results. As can be seen in Figure 13.6, the early cash-flows tend to
be negative and the later cash-flows positive. Thus net present values calculated early on
in a project life tend to be lower than those calculated later. Obviously the higher the rate
at which the project is discounted then the lower the net present value calculated. Usually
the discount rate used is the investors minimum required ROR for a project of a given
degree of risk. Therefore a project with a net present value greater than zero is
acceptable, and a project with a negative NPV is unacceptable: Of course different
standards will be applied if the aim of the project is just to minimise expenditure, e.g. in
meeting pollution control standards.
More complex calculation can, of course, be carried out to allow for longer periods,
changes in interest rates/inflation, additional investments and the effects of taxation etc.
Since these are all negative factors in assessing a potential investment a substantially
higher return must be likely before proceeding. Figure 13.6 shows the influences of costs
and income on the time taken to recover initial capital investment. NPVs are thus a good
direct estimate of likely profits. The NPV of a project can be thought of as the sum by
which the initial investments can be increased while still allowing the project to break
even.

13.4.5 Discounted Cash-Flow Rate (DCFR) or Discounted Cash-Flow Yield

(DCFY)
The discounted cash-flow rate (or internal rate of return) is the discount rate at which a
project NPV equals zero. Thus in equation 13.4 the DCFR is the value of i at which the
NPV is zero.
Like NPVs and unlike ROI, DCFR allows for the time value of money. That is the
money invested in a project only produces profits after that period of time needed to carry
out R&D, build and commission a manufacturing plant, establish a distribution network
and a sales organisation and make appreciable sales etc. Whereas if that money had been
invested in a bank throughout that period it would have been steadily earning interest at a
substantial rate with very little risk incurred.
DCFR values are an estimate of a projects earning potential. Obviously the higher the
DCFR of a project then the more attractive it is. The minimum acceptable DCFR value is
the value that equals the companies required rate of return. The DCFR can be thought of
as a “mortgage” rate such that if a project repays interest to its investors at a rate equal to
the projects DCFR then the project will break-even at the end of its life, and the
‘mortgage’ will be paid off.
Calculation of the interest rates at which NPV=0 can be tedious. Computer software
exists to do this job, but quite accurate estimates can be obtained simply by calculating
some benchmark values, plotting a graph of interest rate vs. NPV and then interpolating
to obtain the required value of i.
 Applied biocatalysis 484
Production cost. The cost of the manufactured product, including raw materials, labour
and utilities costs etc.; but excluding any element of the selling costs such as distribution,
advertising & packaging, and profit margin added to give the sale price.
Gearing is a term that represents the ratio of the funds raised by a company by
borrowing, to its funds generated from profits from sales (retained profits).
Added value. Note also the term “added value” which is the difference between the
price of the finished product and the cost of the raw material needed to make it. Thus the
term “minimum added value” represents the added value needed to cover capital and
operating costs, and to yield a minimum acceptable rate of profitability.
Explanations and details of these terms can be obtained in various books and
pamphlets, for instance, Allen (1980) and Liversey (1983). No one parameter is self-
sufficient, and alone gives a complete view of the financial status of a project or product.
Therefore a number of parameters are usually calculated, and used in conjunction. They
serve to give indications of how to minimise the period over which capital needs to be
committed to a specific project, to show changes in the cost-structure of the process with
time, and to reduce the time taken to convert the material, labour and overhead resources
(the working capital) into cash in the form of profits from sales. That is, to make the
cash-flow as favourable as possible. Each company will have different criteria as to what
constitutes an acceptable financial risk when evaluating a project. Obviously healthcare
product/pharmaceutical companies have quite different criteria to companies
manufacturing bulk chemicals.

13.4.6 Project Life Cycles and Cash Flow

The cumulative cash-flow graph takes into account all the various incomes and
expenditure and so provides a good basis for economic evaluation of a project whether in
terms of value created or rates of return on investment. At the start of the project, no cash
flows have taken place, i.e. no expenditure has been made and no income received. Then
during the initial stages during which research, development, design etc. take place,
expenditure and thus a negative cash flow occurs (Figure 13.5, phase 1–2). Pilot scale
development work is roughly ten times as expensive to carry out as the laboratory
research preceding it. Also any significant changes to the process will cause additional
toxicological testing and regulatory examination of the product and thus cause significant
delays and increased costs.
Expenditure increases dramatically once building takes place, production plant and
facilities are bought and constructed (Figure 13.5, phase 2–3), and commissioning is
undertaken (phase 3–5), until the expected (design) performance is reached at point 5.
This investment in buildings, equipment and other expenditure prior to the sale of
products is termed the capital investment; or if the project is more risky, the venture
capital. Obviously changes to the process, especially those involving plant re-design,
become progressively more expensive the more advanced the project becomes.
Production of the finished product(s) commences during commissioning (point 4).
Subsequent to successful commissioning, revenues from the sale of product can be
 Commercial and financial aspects of biocatalysis research 485
expected to make a positive contribution to the cash flow, and once these exceed the
earlier expenditure a positive cash flow is reached after the break-even point. This is the
time at which total expenditure is precisely balanced by income (point 6); that is when
the revenues generated minus costs incurred equals zero. This positive cash-flow will
hopefully then continue through the remaining life of the process resulting in the paying-
off of R&D and other costs, and then producing increasing cumulative profits.
Profit margins are often greatest when the product is fairly new on the market and
before the maximum volume of sales is achieved. Per annum profit margins may decline
after some time, for instance due to the need to replace capital equipment, or because of
internal or external competition from new and improved products

Figure 13.7 Product life cycle.

produced by the same company or its competitors (Figures 13.7, and 13.5 phase 7–8),
until the process is deemed to be no longer a viable means of using resources, and so can
be shut down (point 8). After this point, the only remaining influences on cash flow will
be very small positive contributions from the value of equipment from salvage, if the
equipment can be used elsewhere in the organisation; or the negative costs incurred by
demolition, site remediation etc.

13.4.7 Product Life-cycles

 Applied biocatalysis 486
Once the new product is launched into the market-place it is vitally important to
appreciate the concept of the “product life-cycle”. First there is the introduction stage in
which the new product is first marketed. Then there is the growth stage in which sales
increase. This can be quite rapid if there are few direct competitors, but often a new
product is accepted only quite slowly by customers. After a time sales will plateau but
large volumes of sales and high profit margins may continue for a considerable time until
decline starts, particularly once competing products enter the marketplace and gain
market-share and reduce profit margins. Movement of a product into the decline phase is
not inevitable. Repeated increases in sales and successful competition with imitator
products can be achieved by continual improvement of the product and/or finding new
applications for it A good example is nylon (Figure 13.8) which was originally
introduced into military applications,

Figure 13.8 Hypothetical life cycle—Nylon from Levett, T. (1986).

such as parachutes and ropes, subsequently found mass-market uses in knit wear and
women’s stockings, and then when sales appeared to have plateaued, in such diverse
products as rugs, tyres and bearings. Thus some products do not decline. Very often this
is because research has continued and resulted in cheaper and/or improved versions of the
original product. Products are rejuvenated not only by the discovery of new end-uses
applications, but also by the introduction of new manufacturing technology or the ability
to use cheaper raw materials etc., which bring obvious benefits in terms of lower cost
 Commercial and financial aspects of biocatalysis research 487
and/or higher quality products (Figure 13.8).
The initial negative cash flow is a distinctive feature of most new businesses and
products, however successful they eventually become, especially if delays occur at
expensive stages such as the commissioning of production facilities. Thus running out of
capital is a major factor in the failure of “start-up” companies, particularly when its first
sales are delayed, for instance due to difficulties in obtaining regulatory approval for
products, hence the need for good budgetary control, comparing expenditure with
planned costs.
It is also important to stress the incremental nature of cash-flow, and also that many
new projects are very research intensive, particularly when inter-disciplinary skills are
required. The many activities involved during the research development and
commercialisation of new products are best integrated by some form of critical path
analysis so as to minimise the time and expense involved.
 Applied biocatalysis 488

Figure 13.9 Supply curves & demand. From A.Hacking, 1986. Economic
Aspects of Biotechnology. Reprinted with permission from
Cambridge University Press, UK.

13.5 SUPPLY AND DEMAND

 Commercial and financial aspects of biocatalysis research 489
The price of a product is obviously a major determinant of the demand for it by
customers. The demand for the product then establishes the scale of production, i.e. the
supply of the product. The relationship between the price of a product, the amount of
product that a manufacturing company can profitably sell, and the amounts of product
that will be bought by customers can be summarised by means of supply and demand
curves (Figure 13.9). A supply curve gives the relationship

Figure 13.10 Price—Volume of use relationships for different classes of

commercially important enzymes (calculated from data in
Poldermans, 1989).

between the amounts of product that a company would be willing to sell at the various
prices, and the amounts of product likely to be bought at these prices. A demand curve
shows how the amounts of product likely to be bought by customers will vary with the
price of the product. The relationship between the prices of a product and the amount
produced (the supply) can be related in terms of the “elasticity” of the product. This is
because the price of a product is roughly proportional to its functionality and utility to the
user. Thus, insulin is obviously a much more functional and therefore expensive protein
than soy protein. Similarly, whereas glucoamylase used for the saccharification of starch
costs just a few pounds a litre, enzymes such as factor VIII used in the treatment of
thrombolytic conditions cost around $1,000 per gram. Thus enzymes intended for
analytical uses, for the manufacture of speciality chemicals, and for bulk products vary a
 Applied biocatalysis 490
lot in price (Figure 13.10).
There is a constant drive on the part of manufacturers to upgrade the value of their raw
materials by improving their end-use functionality. Highly elastic products will show no,
or very little decrease in price in response to an increase in the amount of product made
available to the customer, such as some food products, whereas a highly inelastic
product’s price will change dramatically in response to even small changes in supply,
such as newly launched drugs and other “high performance” products. Thus it would
appear to be much easier to justify research on products with inelastic behaviour.
These are obviously extreme examples, most products behave much more closely to
what is termed unit elasticity—which is defined as being when a 1% reduction in price
will cause a 1 % increase in demand (sales). The equilibrium price can then be defined as
the price at which supply=demand (Figure 13.9 b and c). Of course the actual price will
vary with the grade of product made and thus its end use, examples include
pharmaceutical or technical grade chemicals, or potable or fuel alcohols. Obviously the
price-demand behaviour of inelastic products demonstrates their vital importance to some
consumers, especially when unchallenged by other producer companies or competing
products. This class of products is obviously attractive, not least because of the relative
ease with which high R&D costs can be recovered. The demand, or lack of demand for
products and also other considerations, such as brand loyalty on the part of the consumer
and restrictions in supply due to companies exploitation of their patents tend to distort
precise descriptions of market behaviour.
An exception is negative elasticity, whereby the price for certain usually very
expensive luxury products markedly decreases as they become more available. Examples
include gem stones, scarce fine fragrance materials such as animal musks, and other
luxury items. Elastic behaviour is promoted by alternative competing sources of supply
of the product, the number of uses possible for the material and the utility and value of
the product to the customer.
During the research and development phases of a project financial appraisal can and
should be carried out frequently, in order to check that the process under review still
constitutes an efficient use of resources. For instance, evaluation of the results of basic
research or pilot scale versions of the process etc. are valuable exercises, provided
sufficiently accurate information can be generated to enable a confident decision to be
made concerning a positive commitment to the next stage of the project. Such
information is additionally important since it highlights any inter-relationships between
the various steps in the process and the final cost of the product, and indicates those
operations where further efforts in process development can give the greatest beneficial
effects on the cost of the product.
An interesting effect that has to be taken into account in assessing the potential of a
new process or product is the distinction between “performance” (or “speciality”) and
“commodity” (or “bulk”) chemicals. The former tend to be more expensive and are
assessed principally on the basis of the value of their functionality in specific end-use
applications. The later tend to be cheaper and to be valued in terms of their chemical
specification, such as composition and purity etc., and tend to have a wide range of uses,
 Commercial and financial aspects of biocatalysis research 491
often in very different applications, since they are often used as precursors in the
synthesis of other products.
The assessment techniques mentioned above are obviously most accurate when used
for improved or new versions of existing commercial products, for instance a cheaper
glucose isomerase or a new form of this enzyme with greater productivity. Entirely new
products are much more difficult to assess because in many ways they create their own
new markets, for instance by fulfilling a previously unmet customer need.

13.6 SENSITIVITY ANALYSIS

Another very valuable use of such financial information is sensitivity analysis, in which
the effect of changing process variables on the overall processing costs can be assessed,
to reveal which factors are crucial in determining the profitability of the product, and to
concentrate R&D resources on the cost-limiting steps in the process. For instance, the
effects of operating at different temperatures, of achieving greater product yields, of
using cheaper but perhaps lower quality substrates, carrying out batch, semi-batch or
continuous reactions, or of achieving a range of different conversion efficiencies, can be
determined and used to optimise the process, bearing in mind the subjective probability
(an informed guess) of achieving the desired technical requirements, and the magnitude
of the likely improvements. Sensitivity analysis also gives a good indication of the cost-
structure of the process and the extent to which the current process can be improved; that
is, what the technical and cost ceilings and constraints are. An important aspect of
sensitivity analysis is the relationship between the unit profitability of the product and the
amount (units) of product sold. That is once the threshold price is reached at which
production costs are covered (Figure 13.9a), then there is an increasing incentive for the
manufacturer to produce more product as sales increase. The supply and demand curves
overlap at what is termed the equilibrium price, at which supply and demand are equal.
Sensitivity analysis can, of course, be expressed in terms of the pay-back time, return on
investment, etc. By assuming possible process parameters in advance the risk undertaken
by committing resources to a new project can be roughly evaluated; for instance in terms
of the minimum selling price required to cover variable costs and the matrix of
relationships between price, volume of sales and product profitability. It is particularly
important to carry out systematic sensitivity analysis on lab scale results since it is an
order of magnitude more expensive to make process modifications once the project has
entered pilot-scale studies.

13.6.1 Other Business Factors

Innovation appears to usually take place by small incremental steps; and less commonly
by “break through” advances arising from major new discoveries. In either case several
factors affecting commercial success or failure can be identified (Figure 13.11). Often
 Applied biocatalysis 492
success is dependent on technological interdependence effects, whereby a new
commercial initiative requires inventions to be made in two or more often quite unrelated
areas of technology; for instance biosensor technology depends on advances in
mircroelectronics, enzyme immobilisation and monoclonal antibody technologies.
On the other hand many commercial successes are “low-tech” or even “no-tech” and
can be profitable without any significant technical breakthroughs. For instance, in the
food industry, typically only 0.3% of companies turnover is spent on R&D activities, as
compared with 10–15% by pharmaceutical companies. Large companies often have an
advantage as they are capable of investing in multi-disciplinary research

Figure 13.11 Technological Discontinuity.

projects with long lead times before profits are made. However, small companies are
often regarded as being intrinsically more innovative, perhaps because they can
successfully commercialise products with markets too small to be attractive to larger
companies. In this respect it is obvious that the revenues from a company’s 10% share of
a 1×106 £/a business is equivalent to a 1% share in a 107 £/a business, but the former
10% market share may be much harder to capture.
Legislation is an important factor. Hence a list of articles dealing with Regulatory &
safety aspects is provided (see section 13.9). Regulation may result in levies on raw
materials, such as in the EEC levy on corn based sweeteners to protect domestic sugar
beet farmers, or anti-pollution regulations governing waste disposal. The acceptable
balance between risks taken and possible future benefits can differ markedly between
different companies; for instance between company making bulk low-profit products, and
low bulk, high profit margin products, especially as there is the option that shares in such
 Commercial and financial aspects of biocatalysis research 493
later companies can be sold at a profit even before any real profit has been made,
provided investor’s confidence in the company and thus its share price is high. The
former company has the option of “defensive pricing”, whereby its products prices are
reduced temporarily so as to force a new competing product off the market.
In many applications several quite different technical criteria must be successfully
fulfilled before the product will “work”, and even then cost and scale-up and commercial
criteria must be met. A good example is detergent enzymes where scientists had to
discover proteases that would be active and stable under conditions of high pH and
temperature and in the presence of oxidising and surface active agents. The same criteria
exists for lipases for use in detergents. However, suitable lipases proved rather more
difficult to find than the B. subtilis proteases that are used

Table 13.2 A comparison of the costs of some common commercial enzymes (Adapted
from Poldermans, 1989).

Enzyme costs
Enzyme Use $/kg of enzyme $/kg of product
glucoamylase dextrins to glucose 3 2
-amylase starch to dextrins 5 2–3
glucose isomerase glucose to fructose 5 3–4
pectinase fruit to juice 25 4–10
invertase sucrose to glucose/fructose 6 10–15
lactose lactose to glucose/galactose 25 200

in very large tonnage world-wide, so that only recently have suitable lipases, produced
from Candida humicola strains, been made available to the detergents industry.

13.6.2 Market Sizes and Values

Enzymes sales are estimated to be about 109 $/a world-wide. Biocatalysts are subject to a
“derived-demand” arising from the industrial applications in which they are used, for
processing raw materials to products that can be sold in their own right. Therefore the
commercial importance of biocatalysts is very much greater than these sales of enzymes
would indicate. Such sales values are really only of importance to enzyme manufacturer
companies. Biocatalysts enable the value of raw materials to be enhanced considerably
and the cost of the biocatalyst often represents only a few percent of the final value of
product and the total processing costs (Table 13.2). Also, simple enzyme sales values do
not include the enzymes that are made and used in-house, as occurs both for many
 Applied biocatalysis 494
speciality enzymes and also for many enzymes use in bulk quantities, such as penicillin
acylase which are both produced and used by pharmaceutical companies, and -amylase
and glucoamylase which are both produced and used by corn processing companies.
Secondly, enzymes sales do not fully take into account enzymes used in cell-associated
forms. These include the Pseudomonas chloraphis cells with nitrile hydratase used to
make acrylamide, the Bacillus brevis and Agrobacterium radiobacter cells with
hydantoinase, and hydantoinase and carbamoylase activities respectively, used for
making D-p-hydroxyphenylglycine, and the Pseudomonas putida cells with dehalogenase
activity used for making (S)-2-chloropropanoic acid.
The concept of the “supply-chain” is useful at this stage (Figure 13.12). The supply
chain is the sequence of operations whereby raw materials are converted into
intermediates, then into active materials, and then to formulated products suitable for sale
to the public. Biocatalysts are particularly useful for the reaction of the raw materials and
the production of the active ingredients. At each stage of the supply chain the value of the
materials being processed is enhanced, but quite different skills are needed for success at
the different stages. Thus procurement of raw materials may require agricultural, mining
or extraction expertise and resources;

Figure 13.12 The Supply Chain—Whi ch Position is the Best? From Polastro,
E.T., Walber, A. and Teeuwen, H.W.A. (1989) Bio/technology, 7,
1238–1241.

whereas skills and resources in consumer marketing, advertising and distribution to retail
outlets etc. are required when trading in the final formulated consumer product.
Very often, despite their vitally important role in a process, the microorganisms or
enzymes used constitute only a small proportion of the total running costs of the process
(Table 13.2). This is especially true when research has led to the use of very cheap or
efficient forms of enzymes, such as immobilised enzymes, which are not only highly
active, but which can also be used continuously for long periods, or reused in a batch-
wise mode of operation. For example, in the manufacture of high fructose corn syrups,
the combined costs of the four enzymes involved, -amylase, glucoamylase plus
pullulanase, and glucose isomerase constitute only about 5% of the running costs, which
 Commercial and financial aspects of biocatalysis research 495
for example is less than the cost of the chemicals required for regenerating ion-exchange
resins and adjusting pH etc.! Similar low enzyme costs are also seen in the production of
6-aminopenicillinic acid by the hydrolysis of penicillin V or G using immobilised
penicillin acylase. Then, as processing costs typically represent only 10–20% of the sale
price of the final product, the cost contributions of the biocatalysts to the final sale price
could easily be less than 1 %.
-Amylase, glucoamylase and glucose isomerase etc. are all relatively cheap enzymes
because they are produced in large quantities, because extensive research has been
conducted into improving their production, and because enzymes of very similar
properties are available from a number of different manufacturers who compete largely
on terms of price. They are estimated to cost just few pounds treatment costs per tonne of
product respectively (Table 13.2) and together constitute over 10% of world enzyme
sales. By contrast invertase is relatively expensive. This is an important factor in making
high fructose corn syrups more competitive than sucrose derived invert syrups. In some
cases it has proved economic for some enzyme users such as corn starch processing
companies to produce their own enzymes, a good case of backwards integration of their
business. In fact this commitment into “in-house” provision of a technology is a good
indication of the true success of the technology!
For more expensive enzymes the continuous use of enzymes made possible by their
immobilisation can result in considerable savings. By comparison typical chemical
catalysts represent a smaller proportion of the total manufacturing costs. Thus the
catalysts used in ammonia, cyclohexane and styrene manufacture have been estimated to
cost approximately only 0.7, 0.6 and 0.6% of the total production costs respectively. Thus
biocatalysts are still in general comparatively expensive compared with chemical
catalysts.
Often the benefits of using enzymes are more indirect, i.e. in avoiding the high
temperatures and pressures commonly used in chemical catalysis, so that cheaper
equipment can be used in producing a purer product, thereby reducing purification costs,
especially when a stereospecific and/or regioselective reaction is required, or in providing
a new synthetic route in order to bypass existing patented chemical processes. It is often
difficult to express the value of the biocatalyst, whether in terms of the weight of product
produced, or number of chemical bonds broken (or synthesised) per unit weight of
biocatalyst. The real value is of course in terms of the value added to the raw material per
unit cost of biocatalyst used (Table 13.2) and the value of the new business opportunity
created. Thus for any new technology, such as gene cloning or biocatalyst
immobilisation, a crucial test of its genuine success is its eventual application to high
bulk, low unit cost products, rather than the more immediately attractive high cost low-
bulk products.
Some typical commercial biotechnology products are citric acid, semi-synthetic
penicillins and cephalosporins, and vitamin B12. World production volumes and bulk
prices show a considerable range of values. Prices tend to be inversely proportional to the
amount of product sold, that is, the scale of production, and to the concentration at which
it can be produced in the bioreactor. The importance of the concentration at which each
product is produced in determining the cost of purification and isolation, and thus the
 Applied biocatalysis 496
final price of the product is illustrated (Figure 13.13). Penicillins, cephalosporins and
vitamin B12 are usually produced at concentrations of 12, 3 and 0.002% w/v respectively
(Dunnill, 1981), and so require considerable concentration prior to packaging and sale.
Thus the ease of high-yielding production is crucial in determining the cost of a product.
Hence primary metabolites are intrinsically more likely to be produced cheaply as
compared with secondary metabolites or especially coenzymes, unless highly over-
producing strains can be produced by mutagenesis or genetic engineering. This has been
the case for penicillin in which the concentrations of penicillin produced in fermenters
has been vastly increased over the last 50 years, by a combination of strain improvements
and advances in biochemical engineering (Figure 13.14) and has resulted in a dramatic
reduction in price (Figure 13.15). Biotechnology can be successful in “refinery” type
processes in which a raw material, such as maize, is processed to give a range of different
products each with quite different applications such as high fructose corn syrups, gluten,
corn oil, germ, ethanol, starch etc. Microbial and enzyme processes are also especially
suitable for the detoxification of chemicals that could otherwise only be destroyed by
potentially dangerous and environmentally unfriendly methods such as incineration.
The largest applications of biocatalysts is in the production of cereal based sweeteners.
An important relationship exists between the cost of the enzyme used, and its cost as a
percentage of the cost of the product (Table 13.2). This data shows

Figure 13.13 Concentration in starting material (pre-purification) of various

substances and their selling prices. This graph, which dates from
1984 (see Dwyer, J.L. Bio/Technology 2:957, Nov. ‘84; from
 Commercial and financial aspects of biocatalysis research 497
J.Nystrom, A.D.Little reproduced with permission of Bio/Technology, Nature
America Inc.), emphasises the strong correlation between the crude
product concentration and the product’s production costs, over a
broad range of products. Some biotech products now on the market
lie beyond the range of products previously being sold, commanding
market prices 100–1,000-fold higher than predicted.

the cost advantages made possible by the continuous use of an immobilised enzyme,
rather than the more usual single use of enzymes (sacrificial use) that is common in
bioprocessing. Examples of process costs are seldom revealed by companies; therefore,
good published examples are rare.
For detailed examples of the use of costing techniques as applied to enzyme
technology see Pitcher, Ford & Weetall (1976), who examined an immobilised lactase
process, Poulsen and Zittan (1976) whose calculations are based on the production of
fructose syrups using immobilised glucose isomerase, and more recent data by Harrison
& Gibson (1984) for semi-synthetic penicillin synthesis, Daniels (1987) for immobilised
glucoamylase, lactase and invertase, Harritatos (1991) for fructose production, Tramper
(1985) who carried-out a process cost evaluation of p-hydroxyphenyglycine production,
Jones et al. (1988) who have made
 Applied biocatalysis 498

Figure 13.14 Changes in penicillin broth potencies with time. From

A.L.Demain (1971) Overproduction of microbial metabolites and
enzymes due to alteration of regulation. Adv. Biochem. Eng., 1, p.
113, Reproduced with kind permission from Springer-Verlag
(Heidelberg).

an interesting comparison of different methods of producing L-phenylalanine (Figure

13.16), and Dennis and Davidson (1998), who compared acid-catalysed and enzymatic
production of isopropyl myristate.
In most biological reactions the final (maximum) product concentration is reached
asymptotically. Therefore, although there is an approximate relationship between the
concentration of product achieved during production and the costs of the product (Figure
 Commercial and financial aspects of biocatalysis research 499
13.13) it is sometimes cheaper to terminate the reaction and process the product before
the maximum product concentration is reached, so as to save some of the costs of further
lengthy reactor operation. Continuous reactions are often attractive because the “down-
time”, the time during which the equipment is not being productively used in-between
batch reactions, is minimised. However,

Figure 13.15 The history of the cost of penicillins. How technical advances can
reduce the price of chemicals and so move them from specialities into
bulk markets. From King, P.P. (1982) Biotechnology—an industrial
view. J. Chem. Technol. Biotechnol., 32, 2–8. Reproduced with kind
permission from Society of Chemical Industry, UK.

the costs of the more expensive equipment required for continuous operation and the
penalties caused by breakdowns, either mechanical or due to microbial contamination,
are correspondingly higher. This is also one of the factors that limits the maximum
practical size of bioreactors. Another size limiting factor is the size of equipment that can
be easily transported to the site where it will be used.
Obvious goals in bioprocessing are to minimise raw material costs and reaction times,
and to maximise product concentrations, yields and purities. However the effect of
varying any one variable on the overall process must be accurately assessed. This can be
difficult because of the interdependence of many of these variables, making computer
models a very useful tool. Thus the use of a cheaper raw material may appear to be
attractive, but if this material is of significantly lower purity and so requires costly
purification prior to use, or if the product derived from it needs more extensive
 Applied biocatalysis 500
processing in order to achieve the required purity, then the overall effect may actually be
disadvantageous. Similarly high yields of a product may be achieved by fermentation, but
can the price of the product bear the high cost of the sterile production scale engineering
required?

13.7 CONCLUSIONS

The key objective of applied biocatalysis is to develop widely applicable, easily scaled-
up, reproducible and cost-effective processes and products. However, usually a large

SRI base cases for production of 2000 mt/year L-PHE

Fermentation Cinnamic acid Hydantoin/
routea routeb phenylpyruvic acid
routea
$/kg % $/kg % $/kg %
Materials and suplies
Major feedstock(s) 3.35d 6.20a 7.45f
Other 1.57 0.40 0.20
Total materials and 4.92 31.9 6.60 47.7 7.65
supplies
Labor 1.77 11.5 1.30 9.4 1.24
Utilities 0.51 3.3 0.35 2.5 0.45
Fixed costs 0.48 3.1 0.31 2.2 0.25
Total operating costs 7.68 49.8 8.56 61.8 9.59
Capital-related charges 7.75 50.2 5.28 38.2 4.32
and income tax
Total revenue 15.43 100 13.84 100 13.91
requirement
Sources of revenue
L-PHE 14.94 96.8 13.84 100 13.91
Byproducts 0.46 3.2 g
Estimated total capital
Investment. $ in −24 −17 −14
millions
aAssumptions: nine 120.00- L batch fermentors yield; 0.15 kg L-PHE/kg glucose: 80% recove
 Commercial and financial aspects of biocatalysis research 501

of L-PHE in process train downstream of fermenter or bioreactor.

bAssumptions: 95% utilization of cinnamic acid (CA); 20:1. molar ratio of ammonia to CA 75%
conversion per pass of CA; two-step crystallization for CA recovery and PHE recovery; 80%
recovery of L-PHE downstream of bioreactor, steam stripping for 99% recovery of ammonia a
CO2 12 packed bioreactors operating In six parallel trains; refilling of each btoreactor once eve
14 days on a rotating basis; enzyme loading in bioreactor of 1500 phenylalanine ammonia-lyas
(PAL) units/L for a conversion of 15 g PHE/L/h (at peak) (design of 50% activity basis).
cAssumptions: Synthesis of phenylpyruvic acid: Batch synthesis process for precursors; overall
yield of 95+% of theoretical to phenylhydantoin overall yield of 90+% of theoretical from
phenylhydantoin to phenylpyruvic acid; recovery and recycle of acetic acid; no byproduct cred
taken for acetic acid formed from acetic anhydride addition. Conversion of phenylpyruvic acid
and aspartic acid. Bioreactor productivity of −18 g PHE/L/h (four columns in parallel); 98%
overall conversion; no byproduct credit taken for pyruvic acid (recovery cost assumed to be of
by revenue from sale): 80% recovery of L-PHE downstream of bioreactor.
dGlucose at 40c/kg (18c/lb).
eCinnamic aacid at $5.25/kg (S2.40/lb).
fHydantoin at $4.40/kg ($2/lb); benz aldehyde at $1.65/kg (75c/lb); acetic anhydride at 96C/kg
(44c/lb); aspartic acid at $2.75/kg ($1.25/lb).
gSale value of pyruvic acid assumed equal to recovery cost.

Figure 13.16 A financial comparison of various bioprocesses for

manufacturing L-phenylalanine (Jones et al., 1988).

number of technical, and also important commercial and legislative criteria must be
satisfied in order to achieve commercial success (Figure 12.1). These factors help to
explain why it often takes so long to convert a technical “break through” into a
commercial success, and why companies that possess well integrated R&D, production,
marketing and sales function are more likely to be successful. Another important factor is
that usually the cost of the biologically active ingredient (that is the product of R&D)
often only constitutes a small proportion of the final cost of the product (Figure 13.17),
exceeded by sales and other costs.
Financial appraisal techniques are a vitally important aspect of commercial
biotechnology, but it is often very difficult to gather authentic data to illustrate the
 Applied biocatalysis 502

Figure 13.17 Active ingredient is usually only 11% of drug sales price. From
S.C.Stinson (1992) Chiral drugs. Reprinted with permission from
Chemical & Engineering News, 21(39), pp. 46–77. September 1992,
copyright American Chemical Society.

basic principles, due to quite understandable and justified commercial secrecy. The
financial techniques described in this chapter can make an important contribution to the
sound appraisal of the likely commercial success of projects, so that as informed a
recommendation as possible can be made as how to proceed. Considerable benefit can be
obtained by directing projects in the best directions; and also in avoiding carrying out
fruitless R&D work, by making optimal decisions rather than wrong decisions, or merely
acceptable decisions. Finally, it should be remembered that R&D can be a very good
“once and for all time” investment which, if it is successful in meeting product cost-
quality-supply criteria leads to a successful business that can generate good profits for a
very long time. However, multidisciplinary research is often expensive, slow and difficult
to manage so that unfortunately research time-scales often do not match business time-
scales.

13.8 QUESTIONS
 Commercial and financial aspects of biocatalysis research 503

1. Explain the comment of “product life-cycle” and give examples.

2. How can “economies of scale” affect the viability or otherwise of a new processes?

3. Compare estimates of “pay-back time” and “return on capital invested” as estimates

of a product’s profitability.

4. Define the terms “marginal-costing”, “monopoly supplier”, “joint-venture”,

“market-survey”, “by-products” and “retained profits”.

5. Compare forward (upwards) integration, backwards (downwards) integration, and

horizontal integration, and give examples of each.

6. Distinguish between “fixed (capital) costs” and “variable (renewable) costs” and
give examples of items that could be included in each category.

7. Explain the concept and uses of “discounted cash flow” and related terms such as
“net present value”.

8. Explain how sensitivity analysis can aid research, process development and product
development.

9. How can the purity of the active ingredient, as specified in the product specification
affect the type of processing used?

10. Compare elastic and non-elastic demands for products, give examples of each, and
define “unit elasticity” and “negative elasticity”.

11. Discuss how cash-flow is likely to vary during the research, development and
commercialisation of a new product.

13.9 REFERENCES AND FURTHER READING

Allen, D.H. (1980) A Guide to the Economic Evaluation of Projects , 3rd Edition. UK:
Inst. Chem. Engineers.
Chibata, I. and Tosa, T. (1977) Transformation of organic compounds by immobilized
microbial cells. Adv. Appl. Microbiol. , 22, 1–25.
Daniels, M.J. (1987) The industrial operation of immobilised enzymes. Meth. Enzymol. ,
 Applied biocatalysis 504
136, 371–379.
Demain, A.L. (1971) Overproduction of microbial metabolites and enzymes due to
alternation of regulation. Adv. Biochem. Eng. , 1, 113–142.
Dennis, J.S. and Davidson, A.A. (1998) Cost modelling of esterification processes. In
Advances in Industrial Biocatalysis , InBio Europe ’98 symposium proceedings.
Stockport, UK: Spring Innovations Ltd.
Dunnill, P. (1981) Chemistry and Industry , 204–247.
Fish, N.M. and Lilly, M.D. (1984) The interactions between fermentation and protein
recovery. Biotechnology , 2, 623–627.
Godfrey, T. and West, S. (Eds.) (1996) Industrial Enzymology , 2nd edition. London:
MacMillan Press Ltd.
Gotfredsen, S.E., Ingvorsen, K., Yde, B. and Andersen, O. (1985) The scope of
biocatalysis in organic chemical processing . In Biocatalysts in organic syntheses ,
edited by J.Tramper et al., pp. 3–18. Amsterdam: Elsevier.
Hacking, A.J. (1986) In Economic Aspects of Biotechnology , edited by J.Baddiley et al.
UK: Cambridge University Press.
Harrison F.G. and Gibson, E.D. (1984) Approaches for reducing the manufacturing costs
of 6-amino penicillamic acid. Process Biochemistry , 19 , 33–36.
Harritatos, N.I. (1991) Process Economics. In Biotechnology—The science and the
business , edited by V. Moses and R.E. Cape, pp. 225–248. Switzerland: Harwood
Academic Publishers.
Jones, J.L., Fong, W.S., Hall, P. and Cometta, S. (1988) Whether bioprocesses for
volume chemicals. CHEMTECH , 18, 304–309.
King, P. (1982) Biotechnology—an Industrial view. J. Chem. Technol. Biotechnol. , 32,
2–8.
Liversey, F. (1983) Economics for Business Decisions . UK: Macdonald and Evans,
Plymouth.
Pitcher, W.H., Ford, J.R. and Weetall, H.H. (1976) The preparation, characterisation and
scale-up of a lactase system immobilised to inorganic supports for hydrolysis of acid
whey. Meth. Enzymol. , 44 , 792–809.
Poldermans, B. (1989) Commerciële enzymen en koolhydraten: toekomstverwachting.
Carbohydrates in the Netherlands , August, 25–28.
Poulsen, P.B. and Zittan, L. (1976) Continuous production of high-fructose syrup by
cross-linked homogenates containing glucose isomerase. Meth. Enzymol. , 44 , 809–
821.
Riisgaard, S. (1990) The enzyme industry and modern biotechnology. In Proceedings
5thEuropean Conference on Biotechnology , edited by C.Christiansen et al., Vol. 1, pp.
31–40. Munksgaard.
Slater, N.K.M. (1988) Economic aspects of lipid biotechnology. In Proc. conf. biotech.
for fats and oils industry , edited by T.H.Applewhite. Am. Oil Chemists Soc. , 238–243.
Stinson, S.C. (1992) Chiral Drugs. Chem. Eng. News , 21, September, 46–77.
Tramper, J. (1985) Immobilizing biocatalysts for use in syntheses. TIBTECH , 3, 45–50.
Wandrey, C. and Flaschel, E. (1979) Process development and economic aspects of
enzyme engineering. Adv. Biochem. Eng. , 12 , 147–218.
 Commercial and financial aspects of biocatalysis research 505

Regulatory Aspects

Complex legal, technical and public interest issues are involved. For further details a
number of useful reviews are available including the following:
Appler, W.D. and Giamporcaro, D.E. (1991) Regulatory aspects of biotechnology—
produced ingredients for food applications. In Biotechnology and Food Ingredients ,
537–556.
Bass, R. (1988) Toxicological evaluation of biotechnology products: regulatory
viewpoint. In Proc. conf. biotech. for fats and oils industry , edited by T.H.Applewhite.
Am. Oil Chemists Soc. , 256–257.
Elias, P.S. (1988) General regulatory aspects of Biotechnology: Europe. Ibid. , 262–269.
McNamara, S.M. (1989) FDA regulation of food substances produced by new techniques
of biotechnology. In Biotechnology challenges for the flavour and food industry ,
edited by R.C.Lindsay and B.J.Willis, pp. 65–86. Elsevier.
Nicholas, R.B. and Ager, B. (1991) The regulation of biotechnology in the United States
& Europe. In Biotechnology—The science and the business , edited by V.Moses and
R.E.Cape, pp. 103–115. Switzerland: Harwood Academic Publishers.
Parish, W.E. (1988) Toxicological evaluation of biotechnology products. In Proc. conf.
biotech. for fats and oils industry , edited by T.H.Applewhite. Am. Oil Chemists Soc. ,
258–261.
 INDEX

abzymes, 182, 183

ACA, 111
Accurel EP100, 217
acetate kinase, 334
acetolactate decarboxylase, 64
Acinetobacter , 167
acquisition, 402
acrylamide, 135, 158, 166, 202, 206
actinomycete, 156
activation energy, 279
activators, 284
active site residues, 242
acyl donor, 30
acylase, 154
acyl-enzyme, 318
ADCA, 111
added value, 416
adsorption, 216
affinity chromatography, 204, 205
affinity tail, 193, 194
air-lift loop reactor, 349
Ajinomoto, 167
alanine, 120
alanine dehydrogenase, 332
alanine racemase, 123
alcohol dehydrogenase, 46, 335
alcohol industry, 65
aldolases, 42
alginate, 48, 178
alitame, 117
allosteric regulators, 284
Amgen, 179
amidase, 166, 171
amino acid dehydrogenase, 154
amino acid sequence, 233
amino acids, 23, 124, 154
 Index 507
amino transferase, 154
aminoacylase, 124
aminolysis, 262
aminopenicillanic acid, 278
aminopenicillinic acid, 108
aminopeptidase, 121, 154
ammonia-lyases, 42
amoxicillin, 108, 120
AMP, 204
ampicillin, 108, 324
amylases, 57, 61, 66, 69, 78, 105, 154, 180, 279
amyloglucosidase, 106
anthropogenic, 174
antibiotics, 324
antibodies, 158, 172, 181, 182, 183, 195, 205
APA, 108, 171
apparent equilibrium constant, 319
application studies, 400
appraisal techniques, 402
Archeae, 173
Arrhenius equation, 279
Arrhenius plot, 286
Arthrobacter aurescens , 172
aspartame, 26, 113, 126, 164
aspartate ammonia lyase, 120
aspartic acid, 114, 119
aspartic-type proteases, 277
Aspergillus , 133, 156, 192
assignment, 382
asymmetric synthesis, 20, 324
asymmetrization, 21
ATP regeneration, 334
augmented product, 404
autocatalytic reaction, 356
auxotrophic, 171
Azotobacter vinelandii , 178

B. subtilis , 173
Bacillus , 156, 164, 192
back-mixing, 372
backwards-integrate, 402
bacteria, 156, 180, 187
baker’s yeast, 44, 135, 335
baking processes, 66
barnase, 261
batch reactor, 352, 358
 Index 508
bating, 59
beer industry, 64
beer production, 225
bell-shaped curves, 277
benzoic acid, 162, 166
biocatalytic plastics, 223
biochemical modification, 289
bioconversions, 154, 161
biodiversity, 164, 173
biological material, 395
bio-polishing, 79
bioreactor concepts, 351
bioreactor design, 339
bioreactor types, 342
biosensors, 59
biotransformations, 154
bleach boosting, 61
bleaching, 81
Boehringer, 163
bread, 65
break-even point, 417
brewing process, 64
bromelain, 65
Brookhaven Protein Data Bank, 256
bubble-column reactor, 348
bulk chemicals, 422
by-product credits, 411

C-3 synthons, 131

CALB, 26
caloric sweeteners, 102
Candida antarctica , 26
capital, 411
capital costs, 410
capital employed, 410
capital investment, 417
carbamoylase, 121, 154, 172
carbocyclic nucleoside drugs, 112
carbohydrates, 39
carnitine, 133, 138
carvone, 19
casein, 168
cash-flow, 413, 415
cash-flow diagram, 412
catalase, 64
catalytic antibodies, 158, 183
 Index 509
catalytic efficiency, 273
catalytic triad, 265
CD, 248
cefadroxil, 120
cefatrizine, 120
cefuroxime, 117
cellobiohydrolases, 39
cellulases, 57, 69, 71, 80
cellulose, 39
centrifuge, 198
cephalexin, 112, 324
cephalosporin, 111, 117
Cetus, 391
CGTase, 194
champagne production, 225
charged residues, 241
cheese, 62
chemical modifications, 280, 289
chemo-enzymatic syntheses, 157
Chen-equation, 328
chill-proofing, 64
chitinase, 200
chloropropanoic acid, 129
Chou-Fasman method, 238
chromogenic substrates, 169
chromophors, 169
chymosin, 2, 62, 157
chymotrypsin, 265
circular dichroism, 248
cis, cis-muconic acid, 166
cis-diol, 177
classical strain improvement, 176
cleaning preparations, 57
clearing zones, 167
CLEC’s, 215
CMC, 22
CMP-N-acetylneuraminic, 41
CNBr, 193, 288
cocoa butter substitutes, 311
cofactor, 123, 158
cofactor regeneration, 331
cofactors, 43, 331
coffee industry, 62
commercial driving forces, 400
commissioning, 417
commodity chemicals, 422
 Index 510
complexation, 168, 170
conformational motility, 273
continuous processes, 213
continuous-flow stirred-tank reactor, 344
controlled pore glass, 217
core business, 402
corn starch, 102
correlation spectroscopy, 253
cortisone, 331
costs, 406
covalent chemical modifications, 280
covalent linking, 219
covalently modified enzymes, 301
cross-linked enzyme crystals, 299
crosslinking, 289
crosslinking of enzymes, 215
crystal of protein, 255
CSTR, 185, 344 , 353, 358
CSTRs in series, 354
Cultor, 100
culture collections, 161
culture conditions, 167, 193
cyanohydrins, 137
cysteine, 126, 167
cytochrome C, 233

dairy industry, 62
database mining, 159
databases, 160
DCFR, 416
de novo design of proteins, 264
dead-end filtration, 198
decalactone, 107
decarboxylase, 134
dehydrogenases, 44
demand curve, 421
denaturation, 263, 279, 372
depositing a DNA clone, 388
depreciation, 406
derived-demand, 425
designing proteins de novo , 264
desizing, 78, 79
detergent enzymes, 57
detergent industry, 67
detergents, 57
development, 400
 Index 511
dextranase, 62
diacetyl, 64
diastase, 3
Differential Scanning Calorimetry, 246
diffusion controlled reaction, 273
diffusion limitation effects, 278, 370
dihydrocatechol, 133
dihydroxyacetone phosphate, 42
dihydroxypyrimidinase, 120
Diltiazem, 132
di-oxygenases, 46
direct costs, 408
directed enzyme evolution, 180
directed evolution, 180
discount factor, 414
discounted cash-flow rate, 416
dissociation constant, 272, 283
distomer, 19
disulfide bridge, 260, 263
Diversa, 163, 174, 180
diversity, 163, 164, 173
DNA polymerase enzyme, 390
DNA sequences, 388
DOPA, 170
DOT, 185
down-stream processing, 190
DRASTIC, 174
drugs, 19
DSC, 246
DSM, 164
DSP, 190
Dyno-Mill, 199

E. coli , 173, 177, 181, 192, 195, 197

educt, 156, 158, 176
effectiveness factor, 367, 369
elasticity of a product, 421
electrochemical regeneration, 332
electrostatic effects, 256
ELISA, 59
enantioconvergence, 38
enantiomeric excess, 27, 325
enantiomeric ratio, 26, 328, 329
enantiopurity, 325
enantioselectivity, 163
enantiospecific, 26
 Index 512
endoglucanases, 39
energy diagram models, 273
energy levels, 244
engineering proteins, 229
enthalpy, 321
enzyme activity, 276
Enzyme Commission, 18
enzyme inactivation, 307
enzyme membrane reactor, 333
enzyme powders, 298
enzyme-polymer complexes, 301
ephedrine, 134
epimerase, 178
epimerization, 49
EPO, 393
epoxidation, 46
epoxide hydrolases, 36
epoxides, 102
equilibrium constant, 32
equilibrium control, 318
equilibrium price, 422
Erlenmeyer, 185
Escherichia coli , 183
E-selectin, 40
ester synthesis, 310
esterases, 21, 132, 154, 169
eudismic ratio, 19
Eupergit, 219
European Patent Office, 393
eutomer, 19
examination process, 392
exoglucanases, 39
expert, 388
exponential growth phase, 186
external mass-transfer effects, 365

false negative, 166, 169, 173

false positive, 166, 170
fatty acids, 12
FDH, 158, 334
fed-batch reactor, 353
feed additives, 72
fermentation, 183, 184, 188, 197
fermentation broth, 190
fermentation processes, 127
fermenter, 184
 Index 513
ferments, 3
filtration, 198, 205
fine chemicals industry, 154
first to file system, 393
first to invent system, 393
first-order reaction, 364
fixed capital costs, 406
fixed capital investment, 410
fixed costs, 408
fluidized-bed reactor, 347
Fluka, 163
fluorescein, 169
fluorescence spectroscopy, 248
folding, 249
food industry, 154, 156
formate dehydrogenase, 124, 333
forward-integrate, 402
free energy, 321
free enthaply change, 273
freeze drying, 298
fructose syrups, 98
fruit processing, 64
FT-IR, 174
full costing, 408
fumarase, 23
fumaric acid, 119
fungi, 156, 183, 187

galacto-mannanases, 62
galactosidase, 83, 224
gaseous media, 297
gearing, 416
gel entrapment, 222
gel filtration chromatography, 205
gene libraries, 173
gene shuffling, 180
Genencor, 90, 100, 179
generic product, 404
genetic engineering, 174, 176, 178, 181
genetically modified microorganism, 389
genome, 172, 173
genomics, 173
Gibbs free energy, 163
Gist-brocades, 90
glipizide, 162
glucanase, 64, 74
 Index 514
glucitol, 46
glucoamylase, 57, 103
glucose analysis, 59
glucose isomerase, 98, 153, 159, 180, 213
glucose oxidase, 57, 59
glucose-2-oxidase, 102
glutaraldehyde, 215
glycidol, 132
glycidyl butanoate, 131
glycosidases, 39, 166
glycosyl transferases, 39, 40
glycosylation, 289
GOR method, 239
grace period, 393
granulation, 221
GRAS, 156, 184
growth media, 161

Haldane equation, 276

hallucinogen, 19
halo formation, 167
haloalkane dehalogenase, 175
heat inactivation, 279
heavy atom derivative, 255
hemicellulases, 66
herbicides, 129
heterologous expression, 193
hexahistidine tag, 195
HFCS, 98
high fructose corn syrup, 98
high throughput screening, 165, 174
history, 1
Hoffmann La Roche, 391
Homo sapiens , 172
homologous overexpression, 193
homologue, 172
homophenylalanine, 126
HPOPS, 133
Hughes press, 199
hybridization, 173
hydantoin, 171
hydantoinase, 120, 154, 170, 172, 173
hydantoin-racemase, 172
hydratases, 42
hydration level of enzyme, 303
hydrogen-deuterium exchange, 254
 Index 515
hydrolases, 156, 158, 163, 169
hydrolytic enzymes, 21
hydrophobic cosolvents, 287
hydrophobic interaction chromatography, 204
hydrophobic support, 217
hydrophobicity plots, 236
hydroxyisobutyric acid, 133
hydroxylation, 46, 133, 138
hydroxynicotinic acid, 137
hydroxynitrile lyase, 137, 312
hydroxynitriles, 312
hydroxyphenylglycine, 120

ibuprofen, 132, 137, 167

ICI, 176
idiophase, 187
IEC, 203
IMAC, 195
immmunogold staining, 217
immobilization, 287
immobilization methods, 213
immobilized biocatalysts, 213, 364
immobilized enzymes, 278, 299, 330
immobilized glucose isomerase, 213
immobilized lipase, 214, 311
in situ , 154
inactivation, 372
inclusion bodies, 192
indicator strain, 171
indigo, 81, 178
indirect costs, 408
inducer, 167, 176
infrared spectroscopy, 250
inhibitors, 283
insulin, 117, 178, 194, 264, 324
interest rate, 414
interfacial activation, 23
internal diffusion, 367
internal diffusion resistance, 369
internal rate of return, 413, 416
internet, 160
inulin hydrolysis, 99
inventive step, 386
ion exchange chromatography, 203
ion exchange resins, 219
ionic strength, 321
 Index 516
IRR, 413
irreversible binding inhibitors, 283
isatin hydrolase, 180
iso-electric point, 241
isomaltulose synthase, 102
isomerases, 48, 156, 159
isomerisation, 98
isopropylmyristate, 132
isotope enrichment, 254

joint products, 411

kinetic resolution, 324

kinetically controlled process, 318, 321
KLa, 189
Kluyveromyces lactis , 192

laccase, 64, 81
lactamase, 110, 125
lactate dehydrogenase, 332
lactic acid, 130
lactose hydrolysis, 83
lag phase, 186
Lang factors, 410
Langmuir adsorption, 216
leather industry, 59
lecithin, 67
legislation, 424
Leloir glycosyltransferases, 40
leucine dehydrogenase, 124
license, 382
ligase, 159
Lineweaver-Burke plot, 283
linoleic acid, 108
lipases, 10, 21, 57, 60, 61, 62, 69, 79, 132, 163, 169, 214
lipoamide dehydrogenase, 332
liquid-impelled loop reactor, 350
log P value, 307
Lonza, 162
lyases, 42, 156, 158, 178
lyophilization of enzymes, 298
lyoprotectants, 298
lysine, 124
lysozyme, 200, 259, 276
 Index 517

maize starch, 103

maltose binding protein, 195
malyl-L-tyrosine, 116
Manton Gaulin homogenizer, 199
marginal costing, 408
market development, 400
market evaluation, 400
mass balance, 351
mass transfer coefficient, 189, 365
mass transfer limitations, 330, 364
maximal velocity, 272
Maxygen, 180
meat, 65
medical applications, 58
mega-diversity countries, 164
membrane proteins, 255
membrane reactors, 124, 224, 333, 349
Merrifield method, 26
meso-diester , 21
mesophilic, 159, 181
metabolic pathway engineering, 173, 178
methoxycyclohexanol, 132
methoxyphenylglycidate, 132
methylaspartase, 43
methylpyranozoic acid, 162
methyl-pyrazinecarboxylic acid, 138, 162
methylviologen, 332
Michaelis constant, 272
Michaelis-Menten equation, 272, 357
microemulsions, 302
milk replacers, 75
minimum added value, 416
mixed-type inhibitors, 283
monoclonal antibodies, 182
Monod equation, 187
mono-oxygenases, 46
monopoly right, 380
motifs, 236
MPE, 178
MULTIM, 242
multiphase reactors, 330
multipoint attachment, 221, 289
mutagenesis, 167, 176, 178, 180
myo-inositol hexaphosphate, 75
 Index 518

NAD, 158, 202, 204

NADH, 158, 331
NADPH, 158, 334
naphthalene dioxygenase, 178, 179
naproxen, 132
national patent application, 391
native enzyme, 285
Needleman-Wunsch algorithm, 235
net present values, 414
neural network, 239
nicotinamide, 137
nitrilase, 135, 166, 167
nitrile hydratase, 135, 166
nitrophenol, 169
Nitto, 166
NMR, 250
NOESY spectrum, 252
non-competitive inhibitors, 283
non-conventional media, 294
novelty for a natural substance, 384
novelty requirement, 384
Novo-Nordisk. 90, 100, 174
NPV, 414
nuclear magnetic relaxation, 251
nuclear magnetic resonance spectroscopy, 250
nuclear Overhauser effect, 253
nucleophile, 323
nucleoside sugars, 40

oils and fats industry, 66

oligosaccharides, 39, 154
olive-oil extraction, 67
on column loading, 219
optimisation, 400
optimum scale of a process, 407
organic solvents, 158, 159, 163, 201, 295
OTR, 189
overall effectiveness factor, 370
overheads, 410
oxamniquine, 132
oxidases, 46
oxidoreductases, 43, 156, 158
oxygen transfer rate, 189
oxygenases, 46
 Index 519
oxynitrilase, 42

P. putida , 176, 177, 180

PAC, 134
package solution, 394
packed-bed reactor, 345
palatinose, 102, 324
pancreatic protease, 60
pancreatin, 57
papain, 64, 65
paper and pulp industry, 60
Paris Convention, 391
Pasteur, 380
patent, 380
Patent Co-operation Treaty, 393
patent ownership, 380
patent specification, 388
patent term restoration, 392
patentability, 383
patenting costs, 392
patenting procedure, 391
patents, 10
pay-back time, 412
PCR, 173, 391
PCT, 393
PDB, 256
PDMB, 170
pectinases, 62, 64
penicillin, 108
penicillin amidases/acylases, 110, 213, 278, 317
penicillin synthesis, 319
pentosanase, 74
pepsin, 2
peptide bond, 114
peptide synthesis, 311
peptides, 26, 117
performance chemicals, 422
peroxidase, 59
person skilled in the art, 386
Pfeiffer’s rule, 19
PFR, 363
pH dependence of activity, 277
pH dependence of equilibrium, 319
pH memory of enzymes, 310
pH optima, 279
pH profile, 262
 Index 520
phenoxypropanoic acid, 129
phenylacetyl carbinol, 134, 162
phenylalanine, 23, 116, 126
phenylalanine ammonia lyase, 128
phenylglycine, 122
phosphatase, 75
phospholipase A-2, 67
phytases, 76
phytic acid, 75
pilot scale development work, 417
ping-pong bi-bi mechanism, 30, 276
pitch, 61
PLP, 204
plug-flow conditions, 346
plug-flow reactor, 355, 363
polymerase chain reaction, 173, 391
polyphenylene, 176
porcine pancreatic lipase, 132
pravastatin, 133
Prelog’s rule, 46
present value, 414
primary screen, 164, 165, 166, 174
prior users rights, 382
priority date, 391
process economics, 154
prochiral compound, 325
product concentration, 342
product definition, 403
product life-cycle, 418
production cost, 416
productivity, 340
progesterone, 133
PROSITE, 236
protease inhibitor, 193
proteases, 57, 62, 65, 68, 163, 164, 168, 193, 195, 200, 202
protein A, 194
protein and meat industries, 65
protein electrostatics, 256
protein engineering, 180, 229 , 290
protein engineering work-cycle, 258
protein stabilization, 289
protein surface, 241
Pseudomonas , 166, 175, 178, 192, 388
Pseudomonas putida , 162, 173
psychrophiles, 159
pullulanase, 103
 Index 521
pulping process, 60
PV, 414
pyridoxyl-5-phosphate, 25, 50
pyruvate decarboxylase, 134, 162

quenchers, 249
quenching, 169

rabbit muscle aldolase, 42

racemases, 50, 125
racemate, 328
racemic mixture, 20
racemization, 121, 154
rapid equilibrium mechanism, 272
rate controlled, 318
rate of return, 414
raw material costs, 408
reactor concepts, 351
reactor types, 342
recombinant, 174, 176, 177, 184, 192
regulation, 284
regulatory aspects, 424
rennet, 62
residence time, 355
resolution, 20
resonance fluorescence, 249
return on capital employed, 413
return on investment, 413
reverse condensation, 317
reverse hydrolysis, 40, 309
reversed micelles, 302
reversible binding inhibitors, 283
reversible reactions, 276
Rhizopus arrhizus , 98
rhodamine, 169
Rhodococcus rhodochrous , 166
ribonuclease, 261
ribozymes, 158, 183
ricinoleic acid, 107
ROCE, 413
ROI, 413

S. cerevisiae , 173
safety in handling, 68
salt bridge, 260
 Index 522
salt hydrates, 305
scale effect, 407
scale-up, 400
secondary screen, 166
secondary structure, 236
sedative, 19
segmentation of markets, 404
selenium cystein, 244
sensitivity analysis, 423
series of CSTRs, 359
Serratia marcescens , 171
sialyl Lewis X, 40
silica, 217
single chain antibodies, 183
sol-gel matrices, 223
solubilised enzymes, 301
solubilization, 192, 200
solvent effects, 35
solvent selection, 307
solvent-free reaction media, 297
sorbose, 46
space-time yield, 178
speciality chemicals, 422
specific growth rate, 186
specificity constants, 26
spectroscopy, 247
spin-spin coupling, 253
stability and stabilization, 284
stability index, 286
stability of proteins, 262
starch, 103
starch modification, 60
stationary phase, 186, 187, 203, 204
steady state, 355
stereoheterotopic face, 325
steroid modification, 98
stimulated fluorescence, 249
stirred vessel, 342
stonewashing, 80
storage stability, 181
Streptomyces , 174, 192
structural predictions, 235
substrate inhibition, 363
subtilases, 263
subtilisin, 122, 261
sucralose, 117
 Index 523
sugar industry, 61
sugars, 39
supercritical fluids, 296
supernatant, 190
supplementary enzymes in feed, 57
supplementary protection certificate, 392
supply chain, 144, 425
supply curve, 420
support characteristics, 299
sweeteners, 102

tagatose, 102
tags, 193, 194, 195
tangential flow filtration, 198
Taq polymerase, 391
technical research, 400
temperature effects, 279
temperature optimum, 280
teratogenic, 19
tert-leucine, 123
testing (engineered) proteins, 242
textile industry, 77
thalidomide, 19
therapeutic enzymes, 57
thermal transition temperature, 247
Thermoanaerobacter ethanolicus , 163
thermodynamic control, 318
thermodynamic equilibrium constant, 319
thermodynamically controlled reaction, 319
thermolysin, 26, 116, 164
thermophilic, 159, 163, 164, 181
thermostable enzyme, 181, 203
Thermus aquaticus , 390
Thiele modulus, 369
threshold price, 423
Tosoh, 164
toxicological testing, 417
transaminase, 121, 124
transesterification, 23
transferases, 156, 159
transglutaminase, 85
transglycosylation, 40
transition peak, 247
trophophase, 187
trypsin, 68
tyrosine, 25
 Index 524

ultrafiltration membrane, 3;
ultrafiltration reactor, 123
umbelliferone, 169
unfolding, 249, 285
unit elasticity, 422
unnatural amino acids, 124
utilities, 410
UV spectroscopy, 247

variable costs, 406, 408

venture capital, 411, 417
Vibrio cholerae , 154
vicinal diols, 36
vinyl esters, 33
vitamin C, 46
volumetric productivity, 340
volumetric transfer coefficient, 189

water activity, 35, 304

whole-cell biocatalysts, 158, 162, 178
whole-cell reactions, 335
working capital, 410

Xanthobacter autotrophicus , 175

X-press, 199
X-ray crystallography, 254

yeast, 107

Zeneca, 176
zero-order reaction, 364
zymase, 10
 Index 525

COLOUR PLATE I. See Steffen B.Petersen, Figure 7.1, page 233.

COLOUR PLATE II. See Steffen B.Petersen, Figure 7.2, page 233.
 Index 526

COLOUR PLATE III. See Steffen B.Petersen, Figure 7.4, page 237.
 Index 527

COLOUR PLATE IV. See Steffen B.Petersen, Figure 7.5, page 240.
 Index 528

COLOUR PLATE V. See Steffen B.Petersen, Figure 7.6, page 243.

 Index 529

COLOUR PLATE VI. See Steffen B.Petersen, Figure 7.7, page 244.
 Index 530

COLOUR PLATE VII. See Steffen B.Petersen, Figure 7.9, page 248.
 Index 531

COLOUR PLATE VIII. See Steffen B.Petersen, Figure 7.10, page 252.
Index 532
 Index 533

COLOUR PLATE IX. See Steffen B.Petersen, Figure 7.11, page 257.

COLOUR PLATE X. See Steffen B.Petersen, Figure 7.12, page 259.