Microtubule Proteins, 1st Edition

Visit the link below to download the full version of this book:

https://medipdf.com/product/microtubule-proteins-1st-edition/

Click Download Now

 PREFACE
Microtubules are involved in several basic and essential cell functions such as chro-
mosome segregation, intracellular organization, axonal transport, motility, and determination
of cell shape. All of these functions are based on the capacity of tubulin, the main component
of microtubules, to polymerize and depolymerize.
In this book we have summarized our knowledge of tubulin structure and function, with
a special emphasis on the analysis of polymerization mechanism. In this way, the factors
that promote or decrease tubulin assembly are indicated.
This book is published for postgraduate students and researchers studying microtubules
and the poisons which prevent their polymerization. All the contributors would be glad if
this book would help to a further development of the study of microtubules.

J. Avila
December 1988
 THE EDITOR
Jesus Avlla, Ph.D., has been the Director of Centro de Biologia Molecular in Madrid,
and currently is Research Professor of Spanish Research Council and Professor of the
Department of Molecular Biology at the Universidad Aut6noma de Madrid.
Dr. Avila received his Ph.D. from Complutense University in Madrid in 1971. After
doing postdoctoral work at the National Institutes of Health, Bethesda, Maryland, he was
appointed as a staff member of the Centro de Biologia Molecular.
Dr. Avila has published more than 100 papers. His current research interest includes
the structure-function relationship to microtubule proteins.
 CONTRIBUTORS

Jesus Avila, Ph.D. Ernest Hamel, M.D., Ph.D.

Professor Senior Investigator
Centro Biologia Molecular Laboratory of Biochemical Pharmacology
Consejo Superior de Investigaciones Cientfficas Developmental Therapeutics Program
Universidad Aut6noma de Madrid Division of Cancer Treatment
Madrid, Spain National Cancer Institute
National Institutes of Health
Bethesda, Maryland
Nicholas J. Cowan, D.Phii.
Professor Maria A. Hernandez, Ph.D.
Department of Biochemistry Centro de Biologia Molecular
New York University Medical Center Consejo Superior de Investigaciones Cientfficas
New York, New York Universidad Aut6noma de Madrid
Madrid, Spain

Javier Diaz-Nido, Ph.D. Sally A. Lewis, M.A.

Centro de Biologia Molecular Associate Research Scientist
Consejo Superior de Investigaciones Cientfficas Department of Biochemistry
Universidad Aut6noma de Madrid New York University Medical Center
Madrid, Spain New York, New York

Luis Serrano, Ph.D.

Yves Engelborghs, Ph.D. Centro de Biologia Molecular
Department of Chemistry Consejo Superior de Investigaciones Cientificas
University of Leuven Universidad Aut6noma de Madrid
Leuven, Belgium Madrid, Spain
 TABLE OF CONTENTS

Chapter 1
Dynamic Aspects of Microtubule Assembly .............................................. 1
Yves Engelborghs

Chapter 2
Tubulin Genes: Structure, Expression, and Regulation .................................. 37
Sally A. Lewis and Nicholas J. Cowan

Chapter 3
Structure and Function of Tubulin Regions .............................................. 67
Luis Serrano and Jesus Avila

Chapter 4
Interactions of Tubulin with Small Ligands ............................................. 89
Ernest Hamel

Chapter 5
Microtubule Proteins in Neuronal Cells ................................................ 193
Javier Diaz-Nido, Maria A. Hernandez, and Jesus Avila

Index ................................................................................... 259

 1

Chapter 1

DYNAMIC ASPECTS OF MICROTUBULE ASSEMBLY

Yves EngeJborgbs

TABLE OF CONTENTS
I. Introduction ....................................................................... 2

II. Components of the System ........................................................ 2

A. Tubulin .................................................................... 2
B. Microtubule-Associated Proteins .......................................... 2
C. Microtubules .............................................................. 3
D. Rings ...................................................................... 4

III. Assembly Conditions ............................................................. 6

IV. Polymorphism .................................................................... 7

V. Polymerization Analysis According to Oosawa ................................... 7

A. Equilibrium Analysis ...................................................... 7
B. Kinetics ................................................................... 8

VI. Kinetics of Microtubule Assembly ............................................... 10

A. Nucleation ................................................................ 10
B. Intermediates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... .11
..........
C. Kinetics of Growth ....................................................... 11
D. Disassembly Kinetics ..................................................... 12
E. Relaxation Kinetics ....................................................... 14
F. Discriminating the Two Ends ............................................. 15

VII. Kinetic Events at Steady State ................................................... 17

A. Incorporation of Protein or Nucleotide ................................... 17
B. Hydrolysis ................................................................ 18
C. Disassembly .............................................................. 19
D. Role of T-GDP ........................................................... 20
E. Dynamic Instability ....................................................... 21
F. Annealing ................................................................ 24

VIII. Oscillations ...................................................................... 24

IX. In Vivo Dynamics ................................................................ 26

X. Conclusions ...................................................................... 28

Acknowledgments ....................................................................... 29

References ............................................................................... 29
2 Microtubule Proteins

I. INTRODUCTION

This chapter on microtubule dynamics is conceived as a review of the major paradigms

that have governed this research field. In this way I hope to transmit some of the thrill of
the field, as well as an important consideration: whenever a measurement was made directly
instead of indirectly, a surprising observation was made and totally new aspects of the system
appeared.
The dynamics of the formation of microtubules can be studied by a variety of techniques.
All of them reveal a different aspect of the total process, and a full picture will be obtained
when all these pieces can be integrated into a single mechanism. In this chapter the mechanism
of microtubule assembly will be divided in sequential conceptual steps. The contribution of
the different techniques to the understanding of these steps will be given.
The initial analysis of microtubule assembly in vitro is based on the theory of Oosawa, I
which had been successfully applied to the study of actin polymerization. It remains the
basic framework to which regular reference will be made. Therefore, it will be discussed
in some detail. Aspects not directly related to microtubule dynamics will only be discussed
very briefly and can be found more extensively in excellent reviews. 2~5

II. COMPONENTS OF THE SYSTEM

Since the structure of the final polymeric product is a source of inspiration for the
pathway of assembly, we will describe the microtubule structure here, as well as the basic
components of the system.

A. TUBULIN
The major constituent of the microtubule system is the protein tubulin. It is the association
product of two different subunits: a and 13 tubulin. Both have a molecular weight of 50,000
Da and are highly homologous. 6 .7 The word tubulin always refers to the a-13 heterodimer.
This is usually considered as one unit, although the association is only due to noncovalent
interactions. Detrich and Williams 8 determined the dissociation constant by equilibrium
sedimentation and found a value of 8( ± 3) x 10- 7 M. This implies that at a tubulin
concentration of 2 !JM about 24% of the protein is dissociated into the two subunits. This
measurement has, however, only been done in one particular buffer condition: in 0.1 M
PIPES buffer, 1 rnM MgS04 at pH 6.9. It is conceivable that the different buffer conditions
that influence polymerization also influence dimer formation. This remains to be investigated.
In relation to polymerization it is important to mention that tubulin has two binding sites
for GTP, only one of them allows free exchange with nucleotide in solution (the E-site).
At this E-site, GTP is hydrolyzed during assembly.9,lo It has been shown that the E-site is
situated on the 13 subunit with a photoaffinity analogue of GTP, II or by direct UV illumination
of GTP-tubulin (T-GTP).12 Tubulin is a rather unstable protein. At room temperature its
polymerizability is lost within a couple of hours.
Electrophoretic analysis indicates that tubulin consists of many isoforms. 13 Some of
them show tissue specificity and the genetic regulation of their expression is the subject of
intensive research (see other chapters in this book). Tubulin is also subjected to post-
translational modifications, e.g., detyrosination.1 4

B. MICROTUBULE-ASSOCIATED PROTEINS
During the process of the isolation of microtubules, by repeated cycles of polymerization-
depolymerization, certain proteins copurify. These proteins are called the microtubule-as-
sociated proteins (MAPs).15~17 They can be separated from tubulin by ion exchange chro-
matography,17 whereby care has to be taken not to remove the necessary Mg2 + ions.IS
 3

Originally these proteins were considered to be impurities of the preparation. However, their
role as cofactor for polymerization became soon undisputable. 19 Their association to micro-
tubules in vivo was further proven by the use of specific antibodies. 20
The mixture of MAPs has been divided into two classes on the basis of their molecular
weight: the T-groUP with a relative molar mass between 58 and 70 kDa and the high molecular
weight (HMW) group, subdivided in the HMWI (or MAPI) and the HMW2 (MAP2) group
with a molar weight around 350 and 270 kDa, respectively. 21-23 More MAPs have been
isolated since then, and their properties and biological distribution are reviewed by Matus. 23
These associated proteins stabilize microtubules and therefore increase the yield of poly-
merization. They are supposed to interact with other cell organelles. Their affinity for
microtubules can be modulated by phosphorylation. IS Different MAPs can create a different
morphology of microtubules in electron microscopic pictures. 24 ,2S The properties of, for
example, MAP2 are extensively described in Reference 26,

C. MICROTUBULES
Microtubules are formed by the reversible association of either pure tubulin, or of the
microtubule protein mixture (MTP), i.e., the mixture of tubulin and the MAPs. The earlier
structural studies based on electron microscopy and image reconstruction is presented in an
excellent review by Amos.27 A tubular structure is proposed with an outer and inner diameter
of about 24 and 14 nm, respectively. Hydrated samples give a slightly larger outer diameter
of 30 nm. 28 The length varies considerably according to the polymerization conditions and
can reach several micrometers.
In the microtubules two binding domains can be considered: (1) the longitudinal asso-
ciations which lead to the formation of so-called protofilaments, in which the tubulin mol-
ecules are aligned according to their own length axis. (2) In the cylinder these protofilaments
are laterally associated while the tubulin subunits are slightly staggered. In this way a closed
cylinder is formed with 13 protofilaments. Electron microscopic studies on flagellar outer
doublet microtubules, revealed that the whole structure can be described by a series of
helices, e.g., a left- handed three-start helix. 29 Two lattices can be defined dependent on
the distribution of the subunits in the three-start helix (see Figure I). In the A lattice, the
three-start helix is formed by alternating a and ~ subunits, while in the B lattice, it is formed
by pure a-a or ~-~ sequences. The flagellar doublets consist of a complete microtubule
with A lattice, and a C-shaped sheet (with B lattice) attached to it. 29 In vitro. the number
of protofilaments is variable, with 14 being predominating. Here the B lattice is usually
found, but the spiral of identical subunits is not continuous; instead a seam is present. 30,31
This seam is not necessarily continuous over the whole length of the microtubule.
When MAPs are present, these are organized on the surface of the microtubules in a
kind of superlattice. 24,27
All evidence points to a microtubule structure where all the tubulin molecules are arranged
with the same polarity. Therefore, the microtubule itself shows the same overall polarity
with one end having only a subunits exposed to the solvent, the other end only ~ subunits
(except for the lateral interfaces of the subunits). Taking the spiral organization of the
microtubule into account, a variety of different configurations can be imagined for the detailed
structure of the end. Erikson 32 coined the notion of cosy comers: these are the sites on the
end where an associating monomer can make multiple contacts at once. Weisenberg33 pre-
sented a theory where the exact form of the ends plays an important role. It is clear that the
structure of the ends continuously fluctuates, such that the rate constants measured represent
a weighted average of the different configurations. This average can however differ when
growth is fast or slow, depending on the time available for the units on the ends to rearrange
or not, and in this way the rate constant of growth and dissociation can be influenced. 34
4 Microtubule Proteins

13/3 14/3

FIGURE I. Models of possible lattices for microtubules made from 13

and 14 protofilaments. Tubulin (al3) is shown as a black and empty circle .
The A-lattice is shown in the lower part, the B-Iattice in the upper part of
the models . In the A-lattice the a and the 13 subunits alternate in the three-
start helix. In the B-Iattice , a-a or 1313 sequences are found, except for
the discontinuity. (From Mandelkow, E.-M., Schultheiss, R., Rapp, R. ,
Muller, M. , and Mandelkow , E., 1. Cell. BioI. , 102,1067,1986. With
permission. )

D. RINGS
In conditions where the formation of microtubules is prevented, e.g., low temperature,
presence of calcium ions, or absence of GTP, ring-like oligomers are formed. The size and
the shape of these rings varies considerably according to the solvent conditions: pH, ionic
strength, and according to the tubuliniMAPs ratio. A detailed electron microscopy and
sedimentation study of the rings formed in the presence of MAPs was published by Borisy
and co-workers. 35. 38 The most prominent form of these rings has a sedimentation coefficient
of 30 S. The structure of these double rings, as proposed by the authors , is shown in Figure
2. However, also 18 S rings can be formed at the expense of the 30 S rings, dependent on
the pH. Higher aggregates start to appear at pH 6.3 and increase in concentration with
decreasing pH . Around pH 6.4 almost uniquely 30 S oligomers are in equilibrium with the
dimers and MAPs. Ionic strength is also very important and 0.2 M KCI is sufficient to
dissociate the 30 S rings completely. Rings have been crystallized by Voter and Erick-
 5

- - - 39 nm - - --+-1

HGURE 2. A schematic drawing of the ring structure proposed by Scheele

and Borisy. Each ovoid represents Otj3 tubulin; 29 tubulin dimers are ar-
ranged in one tum of a heli x. The hatched areas are zones where MAPs
are supposed to bind. Six tubulin dimers contribute to one MAPs binding
domain . (From Scheele , R. B. and Borisy, G. G., J . Bioi. Chern ., 253 ,
2834, 1978. With pennission.)

son . 39 Rings are in dynamic equilibrium with the constituents, tubulin, and MAPs. Pantaloni
et al. 40 measured the half-life for the equilibration of radiolabeled tubulin, with unlabeled
tubulin present in the rings . The radioactive tubulin was added in such small concentrations
that the dimer ring equilibrium was not disturbed . At O°C the half-life was 100 min , while
this was reduced to lO min at 2rc. Zeeberg et al. 41 studied nucleotide exchange with rings ,
and found that the rate-limiting step for exchange of nucleotide is the exchange of tubulin
dimers itself. This indicates that nucleotide exchange in the ring state is relatively slow .
The rings are very pressure sensitive, and therefore a fast-pressure perturbation allows the
study of relaxation behavior of the system.42 After a fast-pressure drop from 200 to 1 atm
the reequilibration can be followed with 90° light scattering. Three relaxation steps are
observed in this way for the system at pH 6.4 Two of them are dependent on protein
concentration. The identification of these relaxation steps is unfortunately not obvious. The
two protein concentration-dependent steps are probably due to the association of tubulin and
MAPs with ring intermediates. The concentration-independent step is interpreted as due to
ring opening and closure. Indeed, the particle size factor for 90° scattering drops by about
20% upon opening of a single ring , and by 54 % upon the opening of a double ring into the
single straight rod . Ring formation is slightly exothermic: rings are partially broken down
upon warming.
Marcum and Borisy 38 studied the dependence of ring formation on the total microtubule
protein concentration . These authors propose a stoichiometry of 5 MAPs to 30 tubulin
molecules. However, the system does not follow the description with an all or none equi-
librium model :

5 MAPs + 30 T ~ T30 . MAPss (I)

The concentration dependence of the real ring system is much less pronounced than the
theoretical model of full cooperativity predicts.
The process of ring formation with pure tubulin was s tudied by Frigon and Timasheff43
for T-GTP and by Howard and Timasheff4s fot T-GDP. Double rings are formed, the stability
of which depends very strongly on the Mg2 + concentration, but also on the presence of
GOP or GTP. From the concentration dependence of the s edimentation behavior, a model
6 Microtubule Proteins

TABLE 1
Equilibrium Constants for Ring
Formation by Pure Tubulin43 -45

MgCI2 (mM) 1" 8 nb

For T-GTP

Log(Kp) 91.7 109.35 19.6

Log(K z) 3.37 4.09 0.76
Log(r) 2.41 7.11 5.2

For T-GDP

Log(Kp) 78.9 119.6 44.8

Log (K2) 3.84 4.16 0.36
Log(r) -16.2 15.61 34.7

• The data at I mM Mg2+ are our own extrapolation

on the basis of a linear log(K) vs. log [Mgz+] plot.
For GTP only two concentrations are available.
b Number of MgZ + ions, obtained from the slopes of
these plots, involved in the overall ring formation
(from Kp), in dimerization (from K2) and in ring clo-
sure (from 0.

for ring fonnation could be deduced: an isodesmic oligomerization is assumed up to the

point of ring closure with 26 monomers involved. At high Mg2+ ion concentration, e.g., 8
roM and higher, rings are much more stable in GDP than in GTP. The difference in stability
is largely due to the effect of Mg2+ on the step of ring closure. From the concentration
dependence of ring closure, it can be deduced that about 35 magnesium ions are taken up.
At 1 roM Mg2 + ring closure is extremely unfavorable with GDP (see Table 1).
A very important conclusion is that oligomer as well as ring fonnation by pure tubulin
is negligible in 1 roM Mg2+. However, it becomes very important whenever the magnesium
ion concentration is raised to more than 10 roM. This is often done to increase the yield of
microtubule fonnation in the absence of MAPs. It should be clear that the mechanism of
assembly and disassembly can be different in view of this difference in ring stability (see
further for discussion).

III. ASSEMBLY CONDITIONS

In 1968 and 1972 Weisenberg et al. 9,46 detennined the essential features for the poly-
merization conditions:

1. GTP is necessary for polymerization, and it is hydrolyzed. It can be replaced by

nonhydrolysable analogues.
2. GDP remains finnly bound in the polymer, while Pi is released.
3. GDP itself is a strong inhibitor.
4. Microtubules are very sensitive to Ca2+. Weisenberg's discovery of this calcium
sensitivity is one of the major steps forward in the high-yield preparation of microtubules.

Although it was originally claimed that tubulin did not polymerize on its own, it became
soon clear that this largely depends on the solution conditions. Pure tubulin can assemble
 7

into microtubules but only at rather high protein concentrations. 47 This can be lowered by
the addition of sucrose or glycerol,48 usually 25% v/v (= 3.4 M) of the latter is used. Also
Me 2SO (optimally around 8%) stimulates assembly. 49 Different buffers give different results,
e. g., PIPES buffers give a higher yield of polymerization as compared to the MES buffer. 50
Glutamate (1 M) also stabilizes the polymers very strongly while dissociating the MAPs
from the microtubules. 51 The effect of solution components like glycerol or zwitterionic
substances on the assembly and the stability of proteins is discussed by Gekko and Timasheffi 2
and Arakawa and Timasheff. 53 Another important parameter is the concentration of mag-
nesium ionsY These vary from 0.5 to 16 mM. High concentrations of Mg2+ certainly
stimulate assembly, but next to microtubules, other polymers, mostly with S-shaped sheets
are formed. 54 This is not necessarily problematic, since the propagation constant of the S-
sheets and the microtubules may be very similar. Assembly in phosphate buffer is possible,
although interference might be expected from the fact that Pi is liberated during assembly
(see further).55 This wide variety of conditions make comparison of results of different
laboratories sometimes difficult.
Ca2+ ions dissociate microtubules, and all buffers contain 0.5 to 1 mM EGTA to sequester
these ions. 46 The ionic strength is very important, and 0.2 M is sufficient to prevent any
association.
As already mentioned GTP is necessary for assembly. Its concentration varies between
0.1 to a few millimolar. Higher concentrations may be inhibitory.56 During normal assembly
GTP is hydrolyzed and GDP accumulates. Since this is inhibitory, it is better to use a
regenerating system, e.g., acetate kinase and acetylphosphate, to keep GTP constant and
reduce the GDP concentration to a negligible level. 57 Neither acetylphosphate nor acetate
itself have any influence on the polymerization mechanism. Other regenerating systems can
be imagined but the absence of interference by the components should always be checked.
GTP can be formed from GDP and ATP, since the protein preparation is usually contaminated
with NDP-kinase. 58 .59 The MAPs can be replaced by other cofactors like polycations. The
charge density is very important, and if it is too high a second layer of tubulin can be
wrapped around the microtubule to form a so-called duplex microtubule. 60 •61
Tubulin has many binding sites for inhibitors, e.g., colchicine and its analogues, vinca
alkaloids, podophyllotoxin, nocodazole, etc. These molecules bind very specifically to tub-
ulin and prevent the formation of correct microtubules. 62 -66 One stimulator of assembly with
very high affinity is known: taxol. 67 It polymerizes tubulin very efficiently, even without
MAPs.

IV. POLYMORPHISM

Depending on the solution conditions, different types of polymers can be formed. Next
to the forms already described, tubulin hoops have to be mentioned. 68 .69 These large curved
sheets are formed in conditions of low protein concentrations. In the presence of Zn2+ ions
large flat sheets are formed. 70 All these different types of polymers made from the same
protein prove that the different solvent conditions result in small variations in surface structure
of the protein which are strongly amplified by the high degree of polymerization.
An interesting form of polymer is the structure formed when C-shaped sheets associate
laterally to microtubules to form the so-called hooks. 71.72 These structures allow the deter-
mination of the structural polarity of the microtubules and can be applied to biological
samples. 73

V. POLYMERIZATION ANALYSIS ACCORDING TO OOSAWA

A. EQUILIBRIUM ANALYSIS
Only the essence of the theory derived by Oosawa l is presented here inasmuch as it
8 Microtubule Proteins

defines the tenninology used. In an open polymerization system, e.g., actin or tubulin, a
whole spectrum of polymeric species is fonned. The concentration of an i-mer can be defined
as a function of the monomer concentration (C I) as follows:

(2)

This relation can be understood intuitively: i monomers are involved, which implies i-
I interactions. Except for a small number of initial steps, these are all assumed to be identical,
and are therefore all described by the individual association constant Kp. The factor A
corrects for the instability of the initial oligomers up to the nucleus (n-mer). The total number
concentration (C N = IC) and mass concentration (C M = ILC j ) of polymers can be cal-
culated by applying two series expansions on this relation. The following final equations
are obtained:

CM = A' C/(l - Kp . C I)2 (3)

CN = A' C/(l - Kp . C I) (4)

A consideration of Relation 3 shows that CMincreases very strongly when C Iapproaches

1I~. This is therefore the upper limit for the monomer concentration and it is called the
critical concentration. When the parameter A is extremely small, indicating a very difficult
nucleation, a sharp increase in CM is obtained when the total protein concentration exceeds
the critical concentration. The maximum value of A (= 1) is obtained for an isodesmic
polymerization, which shows a very gradual increase of CMwith total concentration. Inter-
estingly, the parameter A also describes the destabilization of the ends of the polymer. The
number concentration of polymers will be much smaller when nucleation is difficult, as
compared to an isodesmic polymerization (see Fonnula 4). Attempts to detennine the pa-
rameter A for microtubule nucleation in the presence of MAPs have been made. 74 A value
around 0.1 was found at temperatures higher than 30°C, but A decreases to 0.01 around
lOoC, indicating that nucleation is not extremely difficult.
It is important to realize that not only an equilibrium mass concentration is defined but
also an equilibrium number concentration. The ratio of both concentrations gives the average
degree of polymerization: <i> = CM/CN'
Fonnula 2 also allows the calculation of the length distribution. The ratio C j + /C j
~CI < 1 is constant and therefore an exponential length distribution is described.

B. KINETICS
The rate of fonnation and disappearance of the i-mer can be described as follows:

In this relation the rate constant of association (k+) and of dissociation (L) is assumed to
be the same for all species, except for the nucleus. Of course it is impossible to follow the
fonnation of each i-mer separately. Generally a parameter proportional to the mass concen-
tration is measured (e.g., turbidity in certain conditions). The time dependence of the mass
concentration can easily be obtained by summing up Equation 5 for all the species:

dCM/dt = dIiC/dt = iIdC/dt

= k+ . C 1 • C N - L . CN + n . dCn/dt
= -dC/dt (6)

where Cn is the concentration of the nucleus.

 9

Nucleus fonnation is usually simplified to a single global process of order n. The rate
constants k n and k-n stand for nucleus fonnation and disappearance, respectively:

(7)

The number concentration eN is itself the solution of the following rate equation:

(8)

This equation shows that nucleation is the only way for a new polymer molecule (or a new
end) to be fonned, or to disappear.
Additional mechanisms for the creation of ends, not considered by Oosawa, can be
imagined, e.g., the breakage of long polymers. Wegner and Savko7s assumed such a process
of fragmentation to occur in actine, and suggested to describe it by the following additional
tenn in Equation 9:

(9)

The dependence on eM implies that fragmentation can occur at each bond with the same
probability. It seems more likely that very long polymers have a higher chance to fragment
than shorter ones. When shearing long polymers like actin or microtubules, or subjecting
them to sonication, the process of fragmentation is considerably enhanced.
Another process than can influence eN is annealing of filaments, i.e., the longitudinal
association of filaments. This tenn has been described as

(10)

which implies that all i-mers can associate with equal intrinsic probability. This is an
oversimplication, as a smaller i-mer will certainly diffuse faster and have a higher chance
to associate. 76 It is to be expected that this process is going to be important in all conditions
where a very high number concentration is reached, e.g., immediately after shearing or
sonication.
A full description of the whole polymerization curve necessitates the integration of the
rate equations of all individual species. However, this leads to an infinite set of simultaneous
rate equations which cannot be solved. Therefore the simplication was introduced that the
growth rate constant (k+) and the dissociation rate constant (L) is the same for all the
species with their size greater than n. The number of equations is now limited to n species
plus one for eM' Frieden and Godette 77 obtained a good simulation of the polymerization
curves of actin with a numerical integration of such a set of six equations. Wegner and
EngeF8 reduced the set to two equations, one for the time dependence of eN and one for
eM' by assuming a steady state between the initial small oligomers. A similar equation was
obtained by Tobacman and Korn,79 assuming a preequilibrium for the nucleus. Even these
two simultaneous equations have to be integrated numerically in order to stimulate the
observed polymerization curve. Only by neglecting all the dissociation rate constants, Oos-
awa could obtain an analytical solution. This equation has therefore only limited applicability.
The theory of one-dimensional polymerization has been further elaborated by De Levie
and co-workers ,80.81 and analytical solutions have been obtained in particular situations
without neglecting the dissociation rates.
10 Microtubule Proteins

VI. KINETICS OF MICROTUBULE ASSEMBLY

A. NUCLEATION
Although spontaneous nucleation is apparently a process to be avoided in the cell, it
occurs in vitro and it is interesting in its own. As discussed before, it can be defined as the
formation of early oligomers, up to the point where their association rate constants become
independent on their size. Nucleation is thus the formation of the first complete motive that
is propagated by the addition of a new building block. Is this motive the smallest microtubule
that can be formed? This depends on the question of whether the closure of a sheet into the
cylinder is a necessary step to obtain the motive. This is not necessarily true. According to
Erickson and Pantaloni,82 it is sufficient to have comparable rates of association in the
longitudinal and the lateral direction to obtain fast two-dimensional growth. Therefore, since
lateral associations are less stable than longitudinal ones, it is sufficient to make small
oligomers, long enough to have an equal probability of asssociation in the lateral and
longitudinal direction.
Although the process of growth can be studied separately, whenever nucleation is neg-
ligible, e.g., in the approach to steady state, after spontaneous nucleation, or after seeded
assembly, this is never possible for nucleation. During spontaneous polymerization, nu-
cleation slows down very rapidly due to its high power dependence on the concentration of
the monomer pool, which is rapidly depleted by the growth process. The number concen-
tration of polymers finally formed is the result of integration over time of this competition
between growth and nucleation, and it is the most concrete information available about
nucleation.
For actin it is relatively easy to imagine how the nucleus looks, because only the
association of the first three molecules differs from the subsequent ones. Therefore a stim-
ulation of the full polymerization curve was possible. For tubulin, however, nucleation must
be a rather complicated process. As a consequence, most authors determined only the size
of the nucleus, from the cooperativity number after spontaneous nucleation. This coopera-
tivity number is very similar to the Hill coefficient for ligand binding. In the case of a
polymerizing system, it is the apparent stoichiometry coefficient of the nucleus, and it is
derived from the concentration dependence of the lag phase or, of the slope of the linear
phase. (This was shown to be the maximal growth rate at the inflexion point of the poly-
merization curve).42 The cooperativity number can also simply be derived from the de-
pendence on total protein concentration of the first-order rate constant of approach to steady
state, since this reflects the number concentration of ends. For polymerization of tubulin in
the presence of MAPs, the cooperativity number was found to be about 2, a surprisingly
small number. 83 Pure tubulin in glycerol showed a value as high as 10 to 12.54 The difference
was assumed to be due to the preexistence of oligomers when MAPs are present. The rate-
limiting step of nucleus formation is thus the association of two such oligomers. Pure tubulin
in Me 2SO also gave a low value. 84 This was interpreted as due to the transient formation
of oligomers at the mixing interphases, due to the local high concentrations of Me 2SO. A
recent study, however, shows that higher values of n are obtained in Me 2SO, when direct
plots or 10g(Kobs ) vs. 10g(C,o,) are used. 85
High-speed sedimentation of microtubule protein resulted in a MAP-free protein fraction
that was not able to form microtubules. 86 From this it was concluded that rings are necessary
for nucleation. A similar conclusion came from experiments with predissociated rings. 83
When rings were dissociated with the thiol reagent diamide,87 and when subsequently an
excess of mercaptoethanol was added, this protein material was correctly taken up into the
microtubules, but was not able to contribute to nucleation. The same results were obtained
by increasing and subsequently decreasing the ionic strength.
The X-ray scattering experiments of Mandelkow et al., 88.89 however, showed that rings
 11

are first broken down into smaller fragments prior to the assembly of microtubules. Therefore,
it can be concluded that in the presence of MAPs, nucleation starts from small oligomeric
fractions, which are formed by the partial dissociation of the rings. Rings are therefore often
supposed to be the products of a side path of assembly, whenever microtubules are not
formed. Many experiements, however, suggest that rings are the direct products of the
dissociation of microtubules. 41.90 In this way they are intermediates in repeated cycles of
assembly and disassembly. Especially in the presence of GDP and high Mg2+ rings are
shown to be rather stable, as mentioned before. 45

B. INTERMEDIATES
If the nucleus would be the smallest microtubule that can be imagined, only a few
species would have to be described.
However, if the nucleus is the association product of two small oligomers, many inter-
mediates can be imagined between this nucleus and complete microtubules. Several attempts
were made to visualize these intermediate structures via electron microscopy, carefully trying
to preserve the structures in solution. Kirschner and co-workers 91 observed the formation of
curly ribbons, with a lateral width smaller than 13 protofilaments. Since the curvature of
rings is perpendicular to the curvature of the microtubule surface, a curled spiral could be
an intermediate with a gradual change-over of this curvature. An interesting question is
whether the sheets grow in the lateral direction and stop at the right width, or whether
closure itself is the stop signal for lateral growth. In vitro, some flexibility exists about the
number of proto filaments involved in a microtubule, suggesting that closure is responsible
for arresting growth in the lateral direction.

C. KINETICS OF GROWTH
As indicated before, the kinetics of growth can be studied separately and therefore much
more information is available about this process. When microtubule seeds are added to a
solution of tubulin, which is made competent to polymerize at the same time, a first-order
process of microtubule growth is observed. In terms of the theory of Oosawa, this can be
analyzed in a rather straightforward way:

(II)

This is Equation 6 where the nucleation term has been dropped. This equation is written
with the assumption that a polymer has a constant number of growth sites, irrespective of
its size. This factor is included in k+. If CN is constant, this equation can be integrated to:

(12)

The observed first-order rate constant is thus k+ .CN • The average length and the mass
concentration allow the calculation of CN (see earlier). Since it was shown that turbidity is
a measure of the mass concentration of microtubules, polymerization kinetics can be followed
by this simple technique. 92 First-order growth kinetics were described by Lee et al. 93 These
authors demonstrated the sensitivity of the system to ionic strength and the existence of a
pH optimum around 6.2 to 6.4 Bryan94 studied growth of microtubules seeds in the presence
of polyanions to sequester the MAPs. His study experimentally demonstrated the essential
features of the theory presented above: approach to steady state could be described by a
single exponential, the rate constant of which was dependent on the concentration of added
seeds, while the extent of polymerization was independent of it. In this way he obtained
the value for k + = 4 X 106 M - I • S - 1 at 30°C (extrapolated to zero polyanion concentration).
An extensive study of the different features of this condensation mechanism was done in