MSU

LIBRARIES
“

 

 

RETURNING MATERIALS:
PIace in book drop to
remove this checkout from
your record. FINES wiII
be charged if book is
returned after the date
stamped beIow.

 

 

 

 

 

A HICROTUBULE ASSOCIATED
NUCLBOSIDE DIPHOSPRATE KINASE

by

Jeffrey Alan Nickerson

A DISSERTATION

Sublitted to
Michigan State University
in partial fulfill-cut of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Depart-ent of Biochelistry

1985

I

:i'fi‘ :- f
C \._,)\"'. /

. C 9 ~

ABSTRACT

A MICROTUBULE ASSOCIATED

NUCLBOSIDE DIPHOSPHATB KINASE

BY

Jeffrey Alan Nickerson

Isolated microtubule protein contains a nucleoside
diphosphate kinase (NDP kinase) activity which may
catalyze a transphosphorylation of tubulin bound GDP to
61?. NDP kinase binds to ig_vitro assembled bovine brain
microtubules with the association strongest at early
stages of assembly.

NDP kinase was isolated from bovine brain microtubule
protein preparations in nanogram amounts. The enzyme is
a hexamer of two types of subunits of about 18.000
daltons. The enzyme was isolated. in larger quantities,

from bovine brain cytosol without a preliminary

microtubule protein isolation. Both enzymes had the sane
molecular weights, subunit molecular weights. isoelectric
points, catalytic properties, and thermal denaturation
profiles and are. therefore. identical or very similiar
enzymes. Both subunits of NDP kinase could be reversibly
phosphorylated by ATP. Phosphorylation of the enzyne
created sultiple. more acidic forms that retained
activity.

A rabbit anti-NDP kinase antibody was raised to a
bovine brain membrane associated NDP kinase. This
antibody cross-reacted with the microtubule protein NDP
kinase and with the NDP kinase of mouse 3T3 fibroblasts.
The antibody did not cross-react with tubulin or with the
saJor microtubule associated proteins. The antibody
stained fibrous structures resembling microtubules in 3T3
cells. It is likely. therefore, that NDP kinase is
associated with uicrotubules in cells.

The guanine nucleotide binding site of tubulin was
affinity labeled with periodate oxidized GDP. This
covalently attached nucleotide could be phosphorylated
showing that a true transphosphorylation. which might be
catalyzed by NDP kinase. mechaniss does exist. This
experimental design should permit good mechanistic
studies of transphosphorylation and of the role

nucleotides play in microtubule assembly.

For My Father

Janes Warren Nickerson

ACKNOWLEDGEMENTS

I thank William Wells for his support and advice over the
years and those residents of Wells laboratory who have
Iade the years richer. especially. James Kurtz. Christine
Collins, Rita Ray. and Zhong-ru Gan. I would like to
acknowledge some of the people with whom I have
collaborated: Louis Burton, Richard Wagner, Roselyn
Zator. and Ruichi Hasegawa. Without the assistance of
Betty Brazier I would never have made it through the maze
of university regulations. For this and for her
encouragement I an grateful. I would like to thank all
the following people for the contributions that they made
to my research or to my life in the Biochemistry
Depart-ent: Wary Weidenan. Richard Luecke. Charles Sllth.
Clarence Suelter, Steve Brooks, Clare Rodgers, Mark
Seyfred. Wasahiro Sakai. Greg Cheng, Larry Pischel, Greg
Doll. my friends fros Shelagh Ferguson-Killers

laboratory, and others too numerous to name.

TABLE OF CONTENTS

page

LIST OF TABLES ..................................... vii

LIST OF FIGURES .................................... viii
INTRODUCTION ....................................... 1
CHAPTER 1. LITERATURE REVIEW ....................... 3
WICROTUBULES ................................... 4
Nicrotubule reconstitution in vitro ....... 5
Nicrotubule associated proteins ........... 7
Nicrotubule associated enzymes ............ 9

Rings ..................................... 13
NICROTUBULES AND NUCLEOTIDES ................... 16
NUCLEOSIDE DIPEOSPBATE KINASE .................. 27
GTP AS A REGULATORY MOLECULE ................... 31
REFERENCES ..................................... 34
CHAPTER II. PRELIMINARY STUDIES .................... 43
INTRODUCTION ................................... 44
EXPERIMENTAL PROCEDURES ........................ 45
Tubulin preparation ....................... 45
Nucleoside diphosphate kinase assays ...... 46
In vitro microtubule assenbly ............. 46

Sucrose density gradient centrifugation... 47‘
RESULTS AND DISCUSSION ......................... 48

REFERENCES ..................................... 55

iv

CHAPTER III. ISOLATION AND CHARACTERIZATION
OF THE NICROTUBULE ASSOCIATED

NUCLEOSIDE DIPHOSPEATE KINASE .................... 56
INTRODUCTION ................................... 57
EXPERIMENTAL PROCEDURES ........................ 59

Materials .......... . ...................... 59
Preparation of licrotubule protein ........ 59
Preparation of microtubule NDP kinase ..... 61
Isolation of NDP kinase
directly fron brain ..................... 62
Electrophoresis ........................... 63
Detersination of kinetic constants ........ 64
Assays .................................... 65
Sedimentation equilibrium detersination
of native enzyne nolecular weight ....... 65
RESULTS ........................................ 66
Isolation of microtubule NDP kinase ....... 63
Deternination of native enzyme
molecular weight ........................ 33
Isolation of NDP kinase
directly from brain ..................... 84
Comparisons of the two NDP kinases ........ 92
Phosphorylation of NDP kinase ............. 94
DISCUSSION ..................................... 100
REFERENCES ..................................... 105

CHAPTER IV. THE ASSOCIATION WITH NICROTUBULES
OF A NUCLEOSIDE DIPHOSPHATE KINASE HAVING

TRANSPNOSPNORYLASE ACTIVITY ...................... 107
INTRODUCTION ................................... 108
EXPERIMENTAL PROCEDURES ........................ 111

Naterials ................................. 111
Preparation of microtubule protein ........ 112
Isolation of NDP kinase

froa bovine brain ....................... 112
Assays .................................... 113
Electrophoresis techniques ................ 113’
Illunochenical techniques.. ............... 114

Nucleotide affinity labeling of tubulin... 115
Tissue culture and
imuunofluorescent staining .............. 117

RESULTS ........................................ 119
The association of NDP kinase with
oligoneric licrotubule protein rings.... 119
The association of NDP kinase with

microtubules in vitro ................... 124
The association of NDP kinase with
microtubules in 3T3 cells ............... 131
Transphosphorylation catalyzed
by NDP kinase ........................... 145
DISCUSSION ..................................... 150
REFERENCES ..................................... 159

vi

LIST OF TABLES

page
CHAPTER III
Table I Sedinentation equilibriua
deternination of NDP kinase
molecular weight .................. 88
Table II Isolation of NDP kinase directly
fron bovine brain ................. 90
Table III Conparison of kinetic properties
of NDP kinases ................... 93
CHAPTER IV
Table I Cosedinentation of NDP kinase
with microtubules .................. 130

vii

LIST OF FIGURES

page
CHAPTER II
Figure 1
Assembly of microtubules in_vitro
with ATP ..................................... 49
Figure 2
The association of NDP kinase with
sedimenting microtubules ..................... 53
CHAPTER III
Figure 1
Carboxymethyl Sepharose chromatography
of 2 cycle sicrotubule protein ............... 69
Figure 2
Sephadex G-75 chromatography of Carboxymethyl
Sepharose Isolated NDP kinase ................ 72
Figure 3
Blue Sepharose chromatography of Sephadex G-75
purified NDP kinase .......................... 74
Figure 4
Chronatofocusing profiles of NDP kinase ........ 76
Figure 5
Sodiul dodecylsulfate polyacrylaaide gel
electrophoresis of NDP kinase ................ 81
Figure 6
Sedimentation equilibrium profiles
of NDP kinase isolated fros .
microtubule protein ......................... 85
Figure 7
Heat denaturation of NDP kinase ................ 96
Figure 8 32
Autoradiograph of P-labeled NDP kinase ..... 98

viii

CHAPTER IV

Figure 1
Binding of NDP kinase to oligoneric
nicrotubuleprotein rings ..................... 120

Figure 2
The association of NDP kinase with
microtubules during asselbly ................. 126

Figure 3
SDS—PAGE of bovine brain particulate
NDP kinase ................................... 133

Figure 4
Characterization of anti—NDP kinase IgG ........ 136

Figure 5
Indirect inmunofluorescent staining
of 3T3 cells for NDP kinase

and cytoskeletal proteins .................... 143
Figure 6
Transphosphorylation of oxGDP-labeled
tubulin ...................................... 146

ix

INTRODUCTION

The traditional concept of the cell in biochemistry
is a liquid cytoplasmic soup in which menbrane—delineated
organelles float. The chesistry of the cell, enzynatic
and otherwise. is assuled to be solution chemistry with
substrates. products. enzyles. and regulators all
bouncing around randonly. obeying the simple rules of
solution kinetics. This view still is a necessary
simplification because it allows us to write sinple
equations modeling the behavior of enzymes, but we must
remember that it is only an assuned simplification.

Recent advances in our understanding of the cell show
it to be a much lore ordered and structured entity than
we had ever ilagined. It see-s sore and more likely that
even netabolisn has a structure, with the enzymes of
metabolic pathways held in specific positions and
orientations in a cytoplasn that is a dense array of
fibers and fiber associated structures. Perhaps solid
state chemistry is a more appropriate paradign for
cellular chemistry than is solution chemistry. It is
because of this change in our thinking that I believe

that the biochemistry of cytoplasmic structure will come

2
to dominate our thinking and research about cellular
metabolism in the not-too-distant future.

For my thesis research. I have chosen to study one of
the best characterized structural systems in the
cytoplasn, the nicrotubule. The principle microtubule
protein, tubulin. tightly binds two guanine nucleotides
whose functions are not known. In early studies on
nicrotubules we discovered. or perhaps re-discovered, the
presence of nucleoside diphosphate kinase in preparations
of sicrotubule protein. It had previously been proposed
that this enzyle might catalyze a reaction
phosphorylating a tubulin bound GDP to GTP in_gi£g_and by
this mechanisl it might regulate some aspect of
microtubule assembly or function. We decided to study
this enzyle. We wished to isolate and characterize the
NDP kinase from microtubule protein preparations and to
develop lethods of producing it in sufficient quantities
for further biochemical study. We wanted to deternine
whether this enzyse actually binds to microtubules both
;g_vitro and in the cell. Finally, we wanted to know
whether this enzyse might catalyze the phosphorylation of
tubulin bound GDP. We saw this work as a foundation and
hoped that it would pave the way to detailed biochelical
studies of the role NDP kinase night play in the assembly

or function of microtubules.

CHAPTER 1

LITERATURE REVIEW

HICROTUBULES

Electron aicroscopists and cell biologists have, over
the last few decades. identified a filasentous structure
within the cytoplasl of eukaryotic cells consisting of at
least four filament types. each distinguished by a
characteristic diameter in electron licrographs of cell
cross-sections. The ssallest of these are the F-actin
containing microfilaments with a diameter of 60 A (1).
Intermediate filaments all have a disaster of 100 A but
biochenical characterization has shown that there are at
least 5 different types, each with a unique protein
composition (2). Nyosin filanents are distinguished by
their diameter of 150 A (3). Largest. first discovered,
and lost important of all for the purposes of this Thesis
are the microtubules measuring 250 A. The development of
lore advanced methods for extracting. fixing. and
observing the cytoskeleton have revealed a such more
complex structure than had been imagined. containing many
lore types of filaments and structures than the four that
we have been discussing (4.5). Nevertheless. these four
filament types reuain the only biochemically well
characterized cytoplasnic structural elements.

Wicrotubules are involved in cell notility.
intracellular motility. and in the saintenance of cell

structure. The properties of microtubules and their role

5
in these various phenomena have been extensively reviewed
(6-8). The main component of nicrotubules is tubulin. a
protein heterodiser of two sililiar subunits of 52.000 to
55.000 daltons (9). Nicrotubules also contain a variety
of other proteins, the identity of which lay vary from
cell type to cell type. If these proteins can be
co-isolated with tubulin. then they are refered to as
licrotubule associated proteins or NAPs. The structure
and properties of microtubules vary in different cell
types and even between different classes of nicrotubules
in the sane cell. Some of these differences in
aicrotubule properties include changes in ultrastructure.
for exanple the unusual and characteristic double-tubule
structure of outer doublet microtubules in cilia and
flagella (10). and differences in cold stability (11).
with some microtubules undergoing depolymerization at
4°C. Since tubulin is the major coaponent of all these
microtubules it seess likely that these differences in

properties are caused by the NAPs.

Nigrotubule Reconstitution In Vitro — Nicrotubule proteins

have been isolated free a wide variety of species and
tissues. but most biocheaical studies have been done on
the licrotubule proteins of either bovine or porcine brain
isolated by cycles of assembly-disassenbly. Nicrotubule
proteins can be assembled into licrotubules ;g_gi;£2_by

incubation at pH 6.3-7.0 and at tenperatures greater than

6

250C in the presence of GTP and in the absence of
Ca2+. These microtubules can then be disassembled by
decreasing the temperature to less than 20°C. Taking
advantage of these properties. microtubule proteins can be
isolated from brain by cycles of assenbly-disassenbly
(12-14). Nicrotubules are assenbled from a brain high
speed supernatant by incubation with GTP and the Ca2+
chelator EGTA at tenperatures above 25°C. Nicrotubules
are harvested by wars centrifugation. resuspended in the
cold to disasselble microtubules. and centrifuged in the
cold to remove aggregated material. This represents a
single cycle of assembly-dissassenbly. Two or three
cycles are generally used to isolate microtubule protein
for biochemical studies.

These isolation procedures depend on the ability to
reconstitute licrotubules fros their constituent
proteins. Nicrotubule asseably is not a siuple
polymerization. but is instead a two step
initiation—elongation process (15-17). In the first step,
initiation conplexes or nucleation centers are for-ed fro-
aicrotubule proteins. In the second step of assembly.
elongation. microtubules grow from these nucleation
centers. As predicted on theoretical grounds (18). there
is a critical concentration of aicrotubule protein
required for assembly. At protein concentrations below
this critical value assembly will not occur.

Nicrotubule assesbly Lg_vitro does not reach

7
equilibrium; it reaches steady state. At one end of the
sicrotubule, polymerization is occuring. while at the
opposite end. the licrotubule is depolymerizing. At
steady state. the rate of assenbly at one end is equal to
the rate of disassembly at the other. This opposite end
asselbly-disassenbly is called ‘treadmilling' and was
first described by Leslie Wilson and his coworkers
(19,20). It is not known whether tread-illing occurs in_
1112, The fact that microtubules in the cell are anchored
in nicrotubule organizing centers suggests that the
disassesbling end of the sicrotubule light not be free to

engage in disassembly reactions (21-23).

Wigrotubule Associated Proteins — Proteins that co-isolate

with tubulin through cycles of asseably-disassenbly are
operationally defined as microtubule associated proteins
(WAPs). To prove that they are true licrotubule proteins.
that is structural couponents of microtubules. it should
be shown that these proteins are associated with
microtubules in cells. This can generally be done by
iI-unocytocheaical sethods (24-26).

Nicrotubule protein isolated fron bovine or porcine
brain by cycles of assembly-disassembly consists of
75-90 x tubulin by mass and 10-25 x use. (12-14. 27.28).
The lost prominent of the NAPs are the high solecular
weight proteins WAPl and WAP2 with molecular sasses of

lore than 300,000 daltons. These proteins. both of which

8
have been isolated and characterized. associate with the
sicrotubule as projections protruding from the surface at
regular intervals along its length (29—31). A second
group of NAPs are a group of proteins of 70.000 to 80.000
daltons known as tau proteins (29.32.33). One of these.
known as the tubulin assembly protein. has been isolated
and can also be shown to associate with licrotubules In
cells (26). although its exact structural location on the
microtubule is unknown.

The function of the high solscular WAPs and tau
proteins is not well understood. These proteins have been
shown to promote the ig_vitro assenbly of tubulin
(29.26.34) although they are not necessary for asselbly
(35,36). It has been suggested that the NAPs may be the
binding sites for organelle attachment to nicrotubules
(37). NAPI and WAP2 are the lost likely candidates for
filling this role because. as we have discussed. they
exist as projections from the aicrotubule surface. It has
also been suggested the NAPs may be the cross-bridges
between aicrotubules and other cytoskeletal structural
elenents such as actin filaments (38) or inter-ediate
filaments (39). There is actually some evidence for this
latter speculation. Both WAP2 isolated without a heating
step and several of the tan proteins cross-link F-actin in,
vitro forling a gel of actin filaaents (38). WAP2
isolated with a heating step organizes purified actin

filaaents into discrete bundles which are conposed of

9
20 s NAP2 and 80 X actin (40). Since IAP2 binds to both
microtubules and to F—actin it might easily cross-link the
two in the cell. Both NAP] and NAP 2 bind to
neurofilaments. the intermediate filaments of neurons
(41). but neither bind to the internediate filaments of
astroglial cells (41). Co-localization in cells of NAPz
and the intermediate filament protein vimentin has been
found in primary cultures of rat brain (39). As a result
of these studies it seems likely that the most commonly
studied WAPs. tau and the high molecular weight NAPs. are
structural elements mediating the interaction of the

microtubule with other cytoskeletal elements.

Nigrotubule Associated Enzymes - There is a second group

of proteins that have been found in preparations of brain
microtubule protein isolated by cycles of assembly—
disassembly. the microtubule associated enzymes. Since
co-persistance with tubulin through such cycles is
circumstantial evidence of ig_gigg_association with
microtubules these proteins deserve mention in any
discussion of microtubule biochemistry. but because the
microtubule associated enzynes are generally present at
much lower stoichiometric ratios with tubulin they are
more difficult to study and. as a result. frequently
overlooked. Among these enzymes are diglyceride kinase
(42). a cAWP-dependent protein kinase (42—45).

phospholipase C (46). an ATPase (42.47.48). adenylate

lO
kinase(49). and nucleoside diphosphate kinase (49-51).

The microtubule associated enzymes have m variety of
catalytic activities and they might be microtubule
associated for one of at least two reasons. First. their
catalytic activities could be involved in microtubule
function or in the regulation of microtubule function.
Thus. a protein kinase associated with microtubules might
phosphorylate a nicrotubule protein. changing its
structure or function. The second possibility is that
binding to microtubules is a method for regulating the
function of the enzyme. Numerous examples of ambiquitous
enzymes . enzymes that are regulated by changing their
subcellular localization. are known (52). Binding to
cellular organelles may change their catalytic properties
or their access to substrates.

The best characterized of the microtubule associated
enzymes is the cANP-dependent protein kinase. whose
structure and activity may provide us with clues about
what to expect from other microtubule associated enzymes.
The protein kinase exists as a complex of three
polypeptide chains. a catalytic subunit. a regulatory and
cANP-binding subunit. and NAP2 (53). It is likely that
the catalytic and regulatory subunits are located on the
arm of WAP2 that projects from the microtubule surface
(53.54). The principal microtubule protein substrate for
phosphorylation by this protein kinase is NAP2 itself

(43-46.53). The phosphorylation of microtubule proteins by

11
this protein kinase may inhibit the rate of microtubule
‘treadailling'. that is polymerization at one end of the
microtubule and depolymerization at the other at
steady—state (19.20). In summary. the microtubule
associated CARP-dependent protein kinase exists as a
complex with a structural coaponent of the microtubule-
WAP2. which also serves as its principal substrate. and
may regulate aicrotubule assembly or disassembly events.

There is sufficient evidence in the case of the
protein kinase to show that it is in fact a microtubule
associated protein and that it is probably important in
microtubule function. In the case of the other enzymatic
activities found in microtubule protein preparations. I
believe there is as yet insufficient evidence to classify
them as truely microtubule associated. To better focus
our studies on these enzymes we should establish criteria
for concluding that a particular enzyme is microtubule
associated. More importantly. we should decide whether
the in_vitro demonstration of microtubule association is
either sufficient or necessary for establishing the
importance of an enzyme in microtubule function.

The criterion that is most coamonly used is that of
co-purification of the protein with tubulin through many
cycles of assembly-disassembly. Proteins which bind tor
microtubules tightly. that is with very low dissociation
constants. will. after several cycles. pass through

subsequent cycles at an apparent constant stoichiometry

12
with tubulin. But. let us imagine the existance of a
microtubule protein that binds to microtubules. although
less tightly. In this case the stoichiometric ratio of
the protein to tubulin may decrease from assembly cycle to
cycle. If the binding constant is sufficiently low the
amount of the protein that is present may become
unmeasureably low after several cycles. I see no reason
to assume that any protein that is important in
microtubule function must bind tightly enough to
co-purify through cycles of asseably—disassembly.
especially at the sub-physiological protein concentrations
used in cycling. This operational definition of
microtubule associatedness was useful in the early days of
microtubule research since it helped to identify the most
likely and easily studied WAPs. but it does not represent
a logical imperative.

What guidelines should we use to prove that an enzyme
is microtubule associated? I believe that there are three
criteria to be satisfied. A microtubule associated enzyme
should bind to microtubules ig_vitro even if the binding
is low affinity or at a low stoichiometry to tubulin.
Second. it should be demonstrated that the enzyme
associates with microtubules in cells. Third. there
should be some functional significance to the association
with microtubules. Either. binding to microtubules should
affect the catalytic properties of the enzyme. or it

should have soae role in microtubule function.

13

Rings and Things - Preparations of microtubule protein
which are made by cycles of assembly-dissassembly contain
tubulin in two forms. as 63 dimers and as ring—shaped
oligomers of 30—36s (this is reviewed in 55). Oligomeric
aicrotubule protein rings also contain NAPs. especially
the high molecular weight proteins NAPI and NAP2. and
differ in structure according to whether the microtubule
protein was prepared in the presence or in the absence of
glycerol (27.55). Nicrotubule protein prepared in the
presence of glycerol contains rings that are 38s
double—walled structures containing 42 tubulin subunits
(55.56). while microtubule protein prepared in the absence
of glycerol contains 18s rings with 29 tubulin subunits
and 30s rings that consist of two stacked 18s rings
(55,57). These 18s and 308 rings are interconvertable and
exist in equilibrium. Typically. 25-75% of the protein in
a microtubule protein preparation isolated by two cycles
of assembly exists in oligomeric rings. depending on
solution conditions and protein concentration. While NAPs
are present in rings they are not necessary. Pure tubulin
can form rings that resemble the 38s rings discussed above
(58). although. of course. they contain no NAPs.

It has been demonstrated that the tubulin of rings and
the tubulin that exists as free dimers are not quickly
interconverted. nor are they in a simple dimer ring

equilibrium (reviewed in 55). Weingarten et al. (32)

14

showed that in microtubule protein preparations isolated
in the presence of glycerol. 6s and 363 forms of tubulin
did not interconvert even during a complete cycle of
assembly-disassembly. Thus. tubulin that was in the Ge
form before assembly was again in the 8s form after
disassembly and tubulin that started as 36s was again 36s
after disassembly. The Ca and 36s forms could be
interconverted only at high ionic strength. although Gs
tubulin could be converted to 38s by the addition of
excess tau protein. Other studies by Zeeberg g£_gl; (59)
and by Pantaloni g£_gl; (60) show that there is a very
slow exchange of subunits between 368 rings and 6s
dimers. 36s rings are direct. cold induced disassembly
products of sicrotubules; they are not formed from
microtubule released dimers (59). but once rings and
dimers are produced by cold disassembly they exchange
subunits with a half tiae of about 90 minutes (60).

Studies of ;g_vitro microtubule assembly by
synchrotron radiation scattering have shown that
oligomeric rings are not nucleation centers for assesbly;
rather. they are disassembled into protofilament fragaents
which nucleate assembly (61-63). The tubulin subunits
which appear in microtubules during the early stages of
assembly come preferentially from 36s rings (60). This
has been shown with assembly experiments starting with
[125IJ-tubulin in dimers and unlabeled tubulin in rings

(60). It seems unlikely that rings are directly involved

15
in microtubule elongation events (81-63) althdugh ring
fragments could again be involved. It is not known

whether oligoaeric rings exist ;n_vivo'or. if they do.

 

what their role might be in the cell. Since rings are
cold induced depolymerization products that are broken
into smaller fragments upon warning (GO-63.78). I
speculate that ring fragments may exist in_gigg_where they

may be involved in microubule assembly reactions.

16

HLCROTUBULES AND NUCLEOTIQES

Tubulin prepared from brain by either the
chromatographic procedures of Weisenberg et al. (64) or by
cycles of assembly—disassembly (12-14) contains a mixture
of bound guanine nucleotides. mostly GTP and GDP. There
are two distinguishable nucleotide binding sites on each
tubulin dimer (84-67). At the ‘non-exchangeable' site
(N-site) the nucleotide is very tightly. though
non-covalently. bound and does not exchange with guanine
nucleotides in solution. At the ‘exchangeable' site
(E-site) the nucleotide is less tightly bound and may
exchange with GTP or GDP in solution. although. as we
shall see. only under some conditions.

Early estimates of the stoichiometry of nucleotide
binding to isolated tubulin (66—68) were 1 mole of guanine
nucleotide per mole of tubulin dimer. but these were
unreliable estimates because the methods used to extract
nucleotides did so incompletely. removing nucleotide only
from the E—site (69.70). Subsequent work by Weisenberg et
al. (89,70) showed that isolated microtubules contain 2
moles of guanine nucleotide per mole of tubulin dimer.

All of the E-site nucleotide is unremovable under
non-denaturing conditions and is GDP after microtubule
disassembly (88.69.71). lost of the N-site nucleotide on
the tubulin diner is GTP (49.68). while on the assembling

microtubule the E-site nucleotide starts as GTP. but is

17
hydrolyzed to GDP as assembly proceeds (71-73).
The binding of GTP to the E-site of the dimer requires
a free. reduced sulfhydryl group at the binding site (74)
and requires a metal ion: Wg2+, Ca2+, or Cr3+
(74.75). This suggests that the species binding to the

microtubule is a GTP-metal ion complex. forexample

HEZ+-GTP. The dissociation constant for GTP binding to
the E-site on the tubulin dimer has been measured and most
estimates cluster around 0.5 x 10-5 W (74.76.77). The
dissociation constant for GDP binding to the dimer E-site
may be 88 high as 2.0 x 10-6 W (74) and as low as 0.15 x
10-6 W (77). The guanine nucleotide bound to the E-site
on microtubule protein oligomeric rings is. however. much
more tightly bound and does not detectably exchange with
nucleotides in solution (59.69). In fact. Zeeberg g£_al;
(59) have shown that the rate of nucleotide exchange on
36s rings is the same as the rate of tubulin subunit
exchange and. as we have seen. this occurs with a half
time of about 90 minutes (60). The molecular basis for
this difference in exchangeability between oligomeric and
dimer binding sites is not known.

Many studies have suggested that the E-site is
specific for the binding of guanine nucleotides and their
analogues (49.86.74). In a more thorough investigation.
Penningroth and Kirshner (79) have shown that. when

tightly bound GDP is removed. the E-site has a broad

specificity for both ribo— and deoxyribo- nucleotides.

18
The much higher affinity of guanine nucleotide binding.
however. makes it likely that they are the only
nucleotides bound in_gigg;

The ig_vitro assembly of microtubule protein in
nucleotide free buffers has only been achieved with a high
concentration of glycerol or sucrose (12.80). when the
E-site has been preloaded with GTP (80). or when a GTP
regenerating system has been included in the
polymerization mixture (80-83). It seems clear that a
nucleoside triphosphate must occupy the tubulin E-site at
some point in the assembly process in order to support in
vitro microtubule assembly.

The most fundamental question about nucleotides and
microtubules is: why is a nucleotide required at all? It
has been reported that tubulin dimers with E-site GDP can
participate in sicrotubule elongation but not in assembly
initiation events (84). although the rate of elongation in
the presence of GDP is 100 fold slower than with GTP
(84). Penningroth and Kirschner (79) have reported that
assenbly proceeds with sub-stoichioaetric amounts of
E-site nucleotide. suggesting that tubulin can participate
in elongation even in the absence of any E-site
nucleotide.

Hydrolysis of E-site GTP with the release of inorganic
phosphate accompanies assembly (49.68.69.71.72.83.85). but
studies with non—hydrolyzable analogues of GTP show that

the energy of hydrolysis is not necessary for assembly.

19
Both guanosine 5'-[p.X-methylenetriphosphate] or GNP-PCP
(49.86) and guanosine 5'-[fi.r-imido- triphosphate] or
GNP-PNP (69.70.86-88) have been used to induce assembly.
It has frequently been reported that microtubules formed
with these non-hydrolyzable GTP analogues are resistant to
Ca2+ induced disassembly (69.87.89). but in these
studies microtubule assembly and disassembly have been
measured by by changes in turbidity. Penningroth and
Kirschner (49) studied this aore carefully and found that
although the turbidity did not decrease much after
treating analogue assembled microtubules with Ca2+, no
Ca2+ stable microtubules could be found by electron
microscopy. It may be that turbidity measurements have
given an artifactually high estimate of microtubule
stability in these studies. Wicrotubules assembled with
GNP-PCP (49) or with the non-hydrolyzable analogue
Cr3+GTP (75.90) are more cold stable than those
assembled with GTP and this has been verified by electron
microscopy (49). Nicrotubules assembled with
non-hydrolyzable analogues are also aore resistant to
dilution- (91) and podophyllotoxin- (86) induced
disassembly. These studies show that the energy of GTP
hydrolysis is not necessary for assembly. Despite soae
methodological problems. I think it can be concluded that
GTP hydrolysis somewhat destabilizes the microtubule.
perhaps by increasing the rate of reactions at the

disassembling end.‘

20

The best idea so far presented for why GTP and GTP
hydrolysis are used in the mechanism of assembly has been
advanced by Kirschner (92.see also 93.94). He has
proposed that GTP hydrolysis during the asseably process
is part of the mechanism by which the cell selectively
stabilizes properly anchored microtubules and prevents the
formation of spontaneously assembled microtubules. thus
providing the cell with a way to control its cytoskeletal
geometry. The evidence in support of this hypothesis is
only theoretical. An unanchored aicrotubule at steady
state is ‘treadmilling'. that is assembling at one end and
disassembling at the other. At the assembling end tubulin
molecules contain GTP. while at the disassembling end they
contain GDP. Thus the equilibrium constants for tubulin
addition to the microtubule can be different at each end
without violating the statistical mechanical principle of
microscopic reversibility. The result of this. as shown
by Rirschner (92). is that the critical concentration of
tubulin required for assembly is different at both ends of
the microtubule; so. there exists a range of tubulin
concentrations at which disassembly at one end will be
faster than assembly at the other end. By maintaining a
free tubulin concentration in this range. the cell can
prevent the spontaneous formation of unanchored
microtubules. leaving only those microtubules whose
disassesbling end is anchored in an appropriate organizing

center.

21
There are two mechanisms by which E-site GTP might be
generated to support microtubule assembly: nucleotide
exchange and the transphosphorylation of E-site GDP to GTP

without exchange.

Exchange:

Tubulin—GDP + ‘GTPzTubulin-J'GTP + GDP

Transphosphorylation:

Tubulin-GDP + 'NTP;===Tubulin-GTP + ‘NDP

A transphosphorylation mechanism could be driven by any
nucleoside triphosphate (NTP) as pictured above.

Under conditions of pH. ionic strength. and metal ion
concentration favoring assembly. ATP and every other
nucleoside triphosphate tried will induce microtubule
assembly (12.49.89.77). even though binding of these
nucleotides to tubulin cannot be detected. Nichael Jacobs
and his coworkers (77.95) were the first to suggest that a
transphorylase activity in microtubule protein
preparations transfers the gamma phosphate of NTP to a
tightly bound tubulin-GDP. leaving a GTP bound to the
tubulin E-site. This tubulin-GTP would then be capable of
participating in assembly reactions requiring GTP.

It has been suggested that the proposed

22
transphosphorylation reaction might be catalyzed by a
nucleoside diphosphate kinase (49.50) and the two terms
have soaetines been used interchangeably. I think that it
is important to make a distinction between the two terms.
Thus nucleoside diphosphate kinase (NDP kinase) catalyzes
the phosphorylation of a free NDP in solution to NTP while
transphorylase uses tubulin-GDP as a substrate.

phosphorylating GDP in situ to GTP. An NDP kinase could

 

support ATP induced assembly by phosphorylating GDP in
solution to GTP which would exchange with a GDP bound to

the tubulin dimer.

Mechanism 1

NDP Kinase: ATP + GDP =ADP + GTP

Exchange: GTP + Tubulin-‘onr:‘cnp + Tubulin-GTP

A transphosphorylase could support ATP induced assembly by

phosphorylating tubulin bound GDP ;g_s;§u.

Main—2.

Transphosphorylation:

ATP + Tubulin-GDP:ADP + Tubulin-GTP

23

It is possible that transphosphorylation. if it actually
occurs. and nucleoside diphosphate kinase could be two
activities of the same enzyme but there is no reason to
assuae that they are g_pgiori.

Penningroth and Kirshner (49) demonstrated that
microtubule protein preparations contain a true
transphorylase activity which can phosphorylate tubulin
E-site bound GDP in_gitg; They incubated tubulin
containing bound GDP with a mixture of [3H] GTP and
[gamma-32P] UTP under conditions that would support
assembly. They found that the 32p labeled nucleotide
appeared on the tubulin dimer almost 10 times faster than
the 3E labeled nucleotide. Thus. they concluded that
transphosphorylation is occuring 10 times faster than
simple nucleotide exchange. The validity of this
experiment has been challenged by Brylawski and Caplow
(96) who have suggested that it is impossible to remove
free nucleotide from microtubule protein preparations and
so the results of the double label experiment are
complicated by the presence of unlabeled guanine
nucleotides diluting the specific radioactivity of the
[3B] GTP pool. Jacobs and Euitorel (50) have performed,
a similiar double label experiment with [3H] GDP and
[gamma-32P] GTP. Tubulin was preloaded with [3N] GDP

and then incubated with [gamma-32P] GTP. The nucleotide

24
bound to the E-sites on dimers was almost all 32p
labeled while the E-sites on 36s rings was both 35 and
32P labeled. This experiment suggests that E-site GTP
on dimers is produced by exchange while E-site GTP on
rings is produced by transphosphorylation. This result
cannot be dismissed as an artifact of isotope dilution.
but. does not finally settle the debate. NDP kinase is
associated with the 368 rings and for reasons of simple
propinquity may preferentially phosphorylate GDP being
released from nearby ring E-sites.

Penningroth and Rirshner (49) presented evidence that
the transphosphorylase activity is not intrinsic to the
tubulin dimer. The transphosphorylase activity can be
separated from the commonly studied WAPs. the high
solecular weight NAPs and tau proteins. by chromatographic
procedures (49). It seems likely that microtubule protein
preparations contain a previously unstudied NAP or
microtubule associated enzyme which is the transphorylase.

Some early work suggested that the N-site GDP was
phosphorylated by the transphorylase (66) but the aethod
used to distinguish E-site and N-site nucleotide did not
work. It has now been convincingly demonstrated that
phosphorylated nucleotide ends up on the E-site (49,50).
Thus. the transphosphorylase. if it exists. is
phosphorylating E-site GDP.

The issue of whether a transphosphorylase exists is

further complicated by the fact that simple exchange of

25
E-site nucleotide can support microtubule assembly. For
example. the non-hydrolyzable nucleotide analogues GNP-PCP
and GNP-PNP (49.69.70.88-88) and the non-hydrolyzable
Cr3+GTP complex (75.90) can both support ig_vitro
microtubule assembly. Further. the rate of GDP release
froa the E-site during asseably may be fast enough to
support ATP catalyzed assembly by an NDP kinase mechanism.
without transphosphorylation (97). If the Penningroth and
Kirshner (49) experiments discussed above are valid then
transphosphorylase generated GTP can appear at the E-site
an order of magnitude more rapidly than GTP in solution
can exchange onto the E-site. It is possible that. while
simple exchange ggg_support microtubule assembly.
transphosphorylation is the normal method for generating
E-site GTP for assembly. I suggest that the
transphosphorylation may exist for the generation of
E-site GTP on rings and other oligomers that may be
polymerization intermediates such as protofilaaent
fragaents (61-63.84). where the nucleotide is not readily
exchangeable (50.69). Whether this mechanisa has any
relationship to microtubule function in the cell is
unknown and. given our current state of ignorance.
practically beyond speculation. We can only fall back on
a basic tenet of the faith and believe that any process or
structure that we can deaonstrate in_vitro must have some
relevance ig_gigg;

This discussion shows that NDP kinase might be the

26
putative transphosphorylase in microtubule protein
preparations and might play an important role in
microtubule assembly. This discussion highlights the
important questions to be answered about microtubule
preparation NDP kinase. Is NDP kinase truely microtubule
associated? Does a transphosphorylation reaction actually
occur? Is that reaction catalyzed by NDP kinase? If we
could answer these questions we might know something
important about the role of nucleotides in aicrotubule

assembly.

27

NUCLEOSIDE DIPHOSPHATE KINASE

Nucleoside diphosphate kinase (NDP kinase) has been
studied from a wide variety of sources: plant. animal. and
prokaryotic (for a review see 98). where they are usually
cytosolic (99-105). mitochondrial(105-109). or associated
with the plasma membrane (110.111). These enzymes
transfer a phosphate from a nucleoside triphosphate to a
nucleoside diphosphate and are not very substrate specific
for either nucleotide. NDP kinase from most sources has
subunits of 16.000 to 20.000 daltons and exists either as
a tetramer of about 80.000 daltons or as hexamers of about
100,000 daltons. The only NDP kinase that doesn't quite
fit into this size range is the 2s enzyme of Chlamydomonas
flagella (112). which. judging by its small sedimentation
coefficient. is considerably smaller.

Those NDP kinases whose catalytic properties have been
studied catalyze double displacesent (ping-pong) reactions
(98.99.102.108.113). Based on these kinetic results we
would predict that NDP kinases should have phosphorylated
intermediates. high energy phosphate-enzyme complexes. In
fact. such phosphorylated foras of NDP kinase have been
detected (98.101.114-117) and the phosphorylated amino
acid is a histidine (98.101.114.115).

The nucleotide specificity of NDP kinases is not very
high and varies from enzyme to enzyme (98). All NDP

kinases have a requirement for a divalent metal ion.

28

989001811Y "22+ and Nn+. so it seems likely that the
actual substrates for NDP kinases are Ng2+NTP and
H22+NDP (98,100,102).

NDP kinases from most sources come in multiple
isozymes (100.102.105.118). Since. the enzymes are not
known to contain carbohydrate it has been proposed that
the different isozymes represent different polypeptide
sequences (118). Studies showing multiple isozymes have
frequently been done by isoelectric focusing of tissue
homogenates or crude extracts (105.118). It has not
previously been suggested that some of the isozymic forms
of NDP kinase could be phosphorylated intermediates.so let
me suggest it here. A single hexameric NDP kinase
molecule contains 8 active sites and so could be
phosphorylated at 6 positions. The addition of 8
phosphates to the molecule could profoundly lower the
molecules pI. Nany phosphorylation permutations are
possible. containing 0 to 6 phosphates. and could
represent many isozymes.

A bovine brain NDP kinase has been isolated and
studied by Robinson g;_a;: (119). This enzyme is isolated
from the particulate fraction of brain. is apparently
menbrane associated. and is solubilized in the isolation
procedure by the detergent BRIJ 56. The isolation
procedure is novel. relying on an unusual property of this
NDP kinase. its ability to compete with BRIJ 58 sicelles

for binding to the ligand of Blue Sepharose. This NDP

29
kinase is a hexamer of apparently identical 16.600 dalton
subunits. functions with a phosphorylated intermediate.
and has a pI of 8.4. The membrane localization of this
NDP kinase suggests that it may not be the NDP kinase
which is present in microtubule protein preparations.

As discussed earlier in this chapter. NDP kinase
activity has been found in preparations of microtubule
protein prepared by cycles of assembly-disassembly
(49.50.97). It has been shown that this NDP kinase is not
of mitochondrial origin and not released from organelles
by homogenization (120).

Nucleoside diphosphate kinase activity has been found
in proximity to microtubules in cilia and flagella. NDP
kinase activity is associated with the outer fiber
microtubules of sea urchin sperm flagella (121) and
Tetrahymena cilia (122). NDP kinase associated with
Chlamydomonas flagella microtubules can be seperated into
a 2s enzyme that is exclusively flagellar and several 4.3a
isozymes that are present both in the flagella and in the
cell body (112).

The most endearing of NDP kinase's characteristics is
its memory. Studies of the bovine liver mitochondrial NDP
kinase (123) show that the conformations of the enzyme
phosphorylated intermediate vary according to the identity
of the donor nucleotide. Thus. the affinity of the enzyme
intermediate for a particular NDP depends on the identity

of the NTP which provided the high energy phosphate.

30
These affinities for NDPs may vary by as much as 10-fold.
This is important to keep in mind when studying the
kinetic properties of NDP kinases: kinetic constants may
be meaningless alone and must be seen as partially

dependent upon the catalytic history of the enzyme.

31

GTP AS A REGULATORY MOLECULE

I have suggested that NDP kinase may regulate
microtubule assembly or function- a regulation mediated by
guanine nucleotides. While GTP can serve as a substitute
for ATP in some reactions. providing energy or high energy
phosphates. it can also function as a regulator of
metabolic processes. Studies in recent years of such
systems suggest a principle of nucleotide specificity;
when the cell wants a nucleotide for energy it uses ATP.
but when it wants a nucleotide for regulation it uses
GTP. While there are certainly exceptions to this
principle. it is valid often enough to be useful.

Probably the best studied use of GTP in a regulatory
system is with the G proteins in plasma membrane. These G
proteins are guanine nucleotide binding membrane proteins
that mediate hormone dependent inhibition or stimulation
of adenylate cyclase (124-128) and a different GTP binding
protein. transducin. regulates the light-dependent
cGNP-specific phosphodiesterase in retina
(124.125.129.130). These GTP binding proteins exist as
multisubunit complexes when they bind GDP. They remain
inactive until they bind GTP and dissociate into subunits
which are active regulators of adenylate cyclase and V
phospodiesterase.

The regulation of chain elongation during protein

synthesis involves a factor. Tu. which binds GTP and is

32

activated by that binding (131). GTP binding proteins may

regulate 382* channels in mast cells (132) and may also
mediate the effects of insulin-receptor binding on a
plasma membrane protein kinase (133.134).

GTP strongly stimulates glycosylation of microsomal
membrane proteins (135-137). even though it is not a
substrate in any of the reactions. Incubation with GTP.
but not other nucleotides. causes isolated microsomes to
fuse and form larger vesicles (138) and increases the
permeability of the microsomal membrane to proteins and
sugar nucleotides (139). It has been suggested that the
increases in microsomal membrane protein glycosylation are
caused by this increased permeability of the membranes to
sugar phosphates (139).

The protein products of the ras oncogene family are
GTP binding proteins of 21.000 daltons named p21 (140).
There are oncogenic forms of p21 and non-oncogenic for-s
found in normal cells. The normal cellular forms of p21
hydrolyze GTP to GDP. but the oncogenic forms have lost
most of their GTPase activity (141) suggesting that GTP
hydrolysis may be the biochemical event that attenuates
the growth promoting effects of p21 in the normal cell.

These examples show that microtubule assembly is not
the only process that might be regulated by GTP and GTP
hydrolysis. Since GTP in the cell is most easily made by
NDP kinase and since transphosphorylation might occur. not
just with tubulin-GDP. but with any guanine nucleotide

binding protein. NDP kinase and protein-GDP

33
transphosphorylases might play a role in the regulation of
cellular processes. There is circumstantial evidence that
NDP kinase may help to regulate plasma membrane adenylate
cyclase by providing GTP for the G proteins. There is a
clear correlation between the amount of NDP kinase
associated with the plasma membrane and the membrane's
adenylate cyclase activity (111). Since this study was
performed by a former graduate student from this
department it must obviously be very well done.

I think that I have shown that there is enough
rational to study further the role that NDP kinases might
play in the GTP mediated regulation of cellular
processes. The presence of NDP kinase in microtubule
protein preparations makes microtubules a good system in

which to begin that study.

10.

11.

12.

13.

14.

15.

18.

17.

34

REFERENCES

Lazarides. E. (1976) J. Cell Biol. 68) 202-219.
Lazarides. E. (1980) Nature (London) 283, 249-256.

Rorn. E. (1978) Proc. Natl. Acad. Sci. U.S.A. 1§J
588-599.

Wolosewick. J. J. and Porter. K. R. (1979) J. Cell
Biol. 82, 114-139.

Lenk. R. and Penman. S. (1979) Cell 164 289-301.

Roberts. R. and Nyams. J. S. (eds.) (1979)
Nicrotubules, Academic Press. New York.

Snyder. J. A. and McIntosh. J. R. (1976) Ann. Rev.
Biochem. 4__5_, 699-720.

Dustin. P. (1979) Wicrotubules, Springer-Verlag.
New York.

Luduena. R. F. (1979) in Microtubules (Roberts. R.
and Nyams. J. S.. eds.) Academic Press. New York.
pp. 65-116.

Gibbons. I. R. (1981) J. Cell Biol. 9;; 107s-124s.

Webb. B. C. and Wilson. L. (1980) Biochemistry 19,
1993-2001.

Shelanski. N. L.. Gaskin. F.. and Cantor. C. R.
(1973) Proc. Natl. Acad. Sci. U.S.A. 19, 765-788.

Borisy. G. G.. Narcum. J. R.. Olmsted. J. B..
Nurphy. D. B.. and Johnson.R. A. (1975) Ann. N. Y.
Acad. Sci. 253, 107—132.

Nurphy. D. B. and Niebsch. R. R. (1979) Anal.
Biochem. 994 225-235.

Johnson. K. A. and Borisy. G. G. (1977) J. Nol.
Biol. 117, 1-31.

Bryan. J. (1978) J. Cell Biol. 114 749-787.

Gaskin. F.. Cantor. C. R.. and Shelanski. W. L.
(1974) J. Wol. Biol. 894 737-758.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

30.

31.

32.

33.

34.

35

Oosawa. F. and Eigashi. S. (1967) Prog. Theoret.
Biol. 14 79—164.

Wargolis. R. L. and Wilson. L. (1978) Cell 13, 1-8.

Farrell. K. W.. Kassis. J. A.. and Wilson. L.
(1979) Biocheaistry 18, 2642-2647.

Spiegelman. B. W.. Lopata. W. A.. and Kirschner. N.
W. (1979) Cell 16, 239-252.

Brinkley. B. R.. Cox. S. N.. Pepper. D. A.. Wible.
L.. Brenner. S. L.. and Pardue. R. L. (1981) J.
Cell Biol. g9, 554-562.

Tucker. J. B. (1979) in Wicrotubules (Roberts. R.
and Eyams. J. S.. eds.). Academic Press. New York.
pp. 315-358.

Sherline. P. and Sciavone. K. (1977) Science 198.
1038-1040.

Watus. A.. Bernhardt. R.. and Hugh-Jones. T. (1981)
Proc. Natl. Acad. Sci. U.S.A. 184 3010-3014.

Lockwood. A. E. (1976) Cell 134 613-627.

Scheele. R. B. and Borisy. G. G. (1976) Biochem.
Biophys. Res. Commun. 194 1-7.

Asnes. C. F. and Wilson. L. (1979) Anal. Biochem.
98, 64-73.

Berzog. W. and Weber. K. (1978) Eur. J. Biochem.
92, 1-8.

Kim. B.. Binder. L. I.. and Rosenbaum. J. L. (1979)
J. Cell Biol. 89, 266—276.

Vallee. R. B.. and Davis. S. (1983) Proc. Natl.
Acad. Sci. U.S.A. ng 1342-1346.

Weingarten. W. D.. Lockwood. A. R.. Bwo. S. Y.. and
Rirschner. W. W. (1975) Proc. Natl. Acad. Sci.
U.S.A. 13, 1858-1662.

Witman. G. B.. Cleveland. D. W.. Weingarten. W. D..
and Kirschner. W. W. (1976) Proc. Natl. Acad. Sci.
U.S.A. 134 4070-4074.

Wurphy. D. B.. Vallee. R. B.. and Borisy. G. G.
(1977) Biochemistry 194 2598-2605.

35.

38.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

36
Lee. J. C.. Tweedy. N.. and Timasheff. S. N. (1978)
Biochemistry 11, 2783-2790.

Berzog. W. and Weber. R. (1978) Eur. J. Biochea.
8;, 249-254.

Sherline. F.. Lee. Y. C.. and Jacobs. L. S. (1977)
J. Cell Biol. 13, sso-sss.

Griffith. L. and Pollard. T. D. (1978) J. Cell.
Biol. 18, 958-965.

Bloom. G. S. and Vallee. R. B. (1963) J. Cell Biol.
98, 1523-1531.

Sattilaro. R. F.. Dentler. W. L.. and LeCluyse. E.
L. (1981) J. Cell Biol. 88, 467-473.

Letterier. J. F.. Liem. R. R. B.. and Shelanski. W.
L. (1982) J. Cell. Biol. 88, 982-986.

Daleo. G. R.. Piras. N. N.. and Piras. R. (1976)
Eur. J. Biochem. 88, 339-346.

Sloboda. R. D.. Rudolph. S. A.. Rosenbaum. J. L..
and Greengard. P. (1975) Proc. Natl. Acad. Sci.
U.S.A. 18, 177-181.

Sheterline. P. (1977) Biochem J. 168. 533-539.

Sandoval. I. V. and Cuatrecasas. P. (1976)
Biochemistry 88, 3424-3432.

Quinn. P. J. (1973) Biochem. J. 133, 273-381.

Webb. B. C. (1979) Arch. Biochem. Biophys. 198,
296-303.

Tominaga. S.. Eirosawa. R.. and Naziro. Y. (1982)
FEBS Lett. 144, 112-116.

Penningroth. S. N. and Kirschner. W. W. (1977) J.
Wol. Biol. 115, 643-673.

Jacobs. W. and Buitorel. P. (1979) Eur. J. Biochem.
99, 613-622.

Burns. G. and Islam.R. (1981) Eur. J. Biochem. 117,
515-519.

Wilson. J. E. (1976) Trends Biochem. Sci. 8,
124-125.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

69.

37
Theurkauf. W. E. and Vallee. R. B. (1982) J. Biol.
Chem 257, 3284-3290.

Vallee. R. (1980) Proc. Natl. Acad. Sci. U.S.A. 11,
3206-3210.

Scheele. R. B. and Borisy. G. G. (1979) in

Wicrotubules (Roberts. R. and Nyams. J. S.. eds.).

Academic Press. New York. pp. 175—254.

Weingarten. W. D.. Suter. W. W.. Littman. D. R..
and Nirschner. N. W. (1974) Biochemistry 18,
5529-5537.

Scheele. R. B. and Borisy. G. G. (1978) J. Biol.
Chem 253, 2846-2851.

Frigon. R. P.. Valenzuela. W. 8.. and Timasheff. S.
N. (1974) Arch. Biochem. Biophys. 165. 442-443.

Zeeberg. B.. Cheek. J.. and Caplow. W. (1980)
Biochemistry 18, 5078-5086.

Pantaloni. D.. Carlier. W. F.. Simon. C.. and
Batelier. G. (1981) Biochemistry 88, 4709-4716.

Nandelkow. E. N.. Earmsen. A.. Nandelkow. E.. and
Bordas. J. (1980) Nature (London) 287, 595-599.

Bordas. J.. Nandelkow. E. N.. and Nandelkow. E.
(1983) J. Wol. Biol. 164, 89-135.

Wandelkow. E.. Nandelkow. E. N.. and Bordas. J.
(1983) Trends Biochem. Sci. 8, 374-377.

Weisenberg. R. C.. Borisy. G. G.. and Taylor. E. W.
(1968) Biochemistry 1, 4466-4479.

Berry. R. and Shelanski. W. L. (1972) J. Nol. Biol.
71. 71-80.

Jacobs. W.. Smith. B.. and Taylor. E. W. (1974)
J. Wol. Biol. 82, 455-468.

Jacobs. W. (1979) in Wicrotubules (Roberts. R. and
Eyams. J. S.. eds.) Acadeaic Press. New York. pp.
255-278.

Robayashi. T. (1975) J. Biochem. (Japan) 11,
1193-1197.

Weisenberg. R. C.. Deery. W. J.. and Dickenson. P.
J. (1976) Biochemistry 18, 4248-4254.

70e

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

38

Weisenberg. R. C.. Deery. W. J.. and Dickenson. P.
(1976) in Cell Notility (Goldman. R.. Pollard. T..
and Rosenbaum. J.. eds.) Cold Spring Harbor
Laboratory. New York. pp. 1123-1131.

David-Pfeuty. T.. Erickson. H. P.. and Pantaloni.
D. (1977) Proc. Natl. Acad. Sci. U.S.A. 11,
5372-5376.

Bonne. D. and Pantaloni. D. (1982) Biocheaistry 81,
1075-1081.

Carlier. N. -F. (1982) Nolecular and Cellular
Biochemistry 81, 97-113.

Arai. T.. Ihara. Y.. Arai. R. -I.. and Kaziro. Y.
(1975) J. Biochem. (Japan) 11, 647-658.

NacNeal. R. N. and Purich. D. L. (1978) Arch.
Biochem. Biophys. 191, 233-243.

Levi. A.. Cimino. N.. Nercanti. D.. and Calissano.
P. (1974) Biochem. Biophys. Acta 365, 450-453.

Jacobs. N. and Caplow. N. (1976) Biochem. Biophys.
Res. Commun. 88, 127-135.

Kirschner. N. W.. Nonig. L. S.. and Williams. R. C.
(1975) J. Nol. Biol. 88, 263-276.

Penningroth. S. N. and Kirschner. N. W. (1978)
Biochemistry 11, 734—740.

Kobayashi. T. and Simizu. T. (1976) J. Biochem.
(Japan) 18, 1357-1364.

NacNeal. R. R.. Webb. B. C.. and Purich. D. L.
(1977) Biochem. Biophys. Res. Commun. 18, 440-447.

NacNeal. R. K. and Purich. D. L. (1977) J. Biol.
Chem. 252, 4440-4442.

NacNeal. R. R. and Purich. D. L. (1978) J. Biol.
Chem. 853, 4683-4687.

Zacharoff. R. V.. Deery. W. J.. and Weisenberg. R.
C. (1980) J. Nol. Biol. 139. 641-659.

Naccioni. R. and Seeds. N. W. (1977) Proc. Natl.
Acad. Sci. U.S.A. 18, 462-466.

87.

89.

90.

91.

92.

93.

94.

95.

96.

97.

100.

101.

102.

103.

39
Terry. B. J. and Purich. D. L. (1980) J. Biol.
Chem. 255, 10532-10536.

Arai. T. and Kaziro. Y. (1976) Biochem. Biophys.
Res. Commun. 88, 369-376.

Arai. T. and Kaziro. Y. (1977) J. Biochem. (Japan)
82, 1063-1071.

Sutherland. J. (1976) Biochem. Biophys. Res.
Commun. 18, 933-938.

Gaskin. F. (1981) Biochemistry 88, 1318-1322.

Weisenberg. R. C. and Deery. W. J. (1976) Nature
(London) 263, 792-793.

Kirschner. N. W. (1980) J. Cell Biol. 88, 330-334.
Weisenberg. R. C. (1981) Cell Notility 1, 485-497.

Bill. T. L. and Kirschner. N. W. (1982) Proc. Natl.
Acad. Sci. U.S.A. 19, 490-494.

Jacobs. N. (1975) Ann. New York Acad. Sci. 253,
562-572.

Brylawski. B. P. and Caplow. N. (1983) J. Biol.
Chem. 258, 760-763.

Terry. B. J. and Purich. D. L. (1979) J. Biol.
Chem. 254, 9469-9476.

Parks. R. E.. Jr. and Agarwal. R. P. (1973) in The
Enzymes (Boyer. P. D.. ed.) Volume VIII. Academic
Press. New York. Pp. 307-334.

Colomb. N. G.. Cheruy. A.. and Vignais. P. V.
(1972) Biochemistry 11, 3370-3378.

Colomb. N. 6.. Cheruy. A.. and Vignais. P. V.
(1972) Biochemistry 11, 3378-3386.

Colomb. N. 6.. Cheruy. A.. and Vignais. P. V.
(1974) Biochemistry 18, 2269-2277.

Agarwal. R. P. and Parks. R. E.. Jr. (1971) J.
Biol. Chem. 246, 2258-2264.

Nakamura. N. and Sugino. Y. (1966) J. Biol. Chea.
241, 4917-4922.

104.

105.

106.

107.

108.

109.

110.

111.

112.

113.

114.

115.

116.

117.

118.

119.

120.

121.

40
Ratliff. R. L.. Weaver. R. R.. Lardy. H. A.. and
Ruby. 8. A. (1964) J. Biol. Chem. 239, 301-309.

Cheng. Y. -C.. Robison. B.. and Parks. R. E.. Jr.
(1973) Biochemistry 18, 5-11.

Glaze. R. P.. and Wadkins. C. L. (1967) J. Biol.
Chem. 242, 2139-2150.

Schnaitman. C. and Greenawalt. J. W. (1968) J. Cell
Biol. 88, 158-175.

Goffeau. A.. Pedersen. P. L.. and Lshninger. A. L.
(1968) J. Biol. Chem. 243, 1685-1691.

Jacobus. W. E. and Evans. J. J. (1977) J. Biol.
Chem. 252, 4232-4241.

Chao. H. N. (1976) J. Biol. Chem. 251, 2330-2333.

Kimura. N. and Johnson. G. S. (1983) J. Biol. Chem.
12609-12617.

Watanabe. T. and Flavin. N. (1976) J. Biol. Chen.
251, 182-192.

Colomb. N. G.. Cherny. A.. and Vignais. P. V.
(1969) Biochemistry 8, 1926-1939.

Norman. A. W.. Wedding. R. T.. and Black. N. F.
(1965) Biophys. Biophys. Res. Commun. 88, 909-915.

Pedersen. P. (1968) J. Biol. Chem. 243, 4305-4311.

Walinder. 0.. Zetterqvist. 0.. and Engstrom. L.
(1969) J. Biol. Chem. 244, 1062-1064.

Lascu. 1.. Pop. R. D.. Porumb. E.. Presecan. E..
and Proinov. I. (1963) Eur. J. Biochem. 135,
497-503.

Cheng. Y. -C.. Agarwal. R. P.. and Parks. R. E..
Jr. (1971) Biochemistry 11, 2139-2143.

Robinson. J. B.. Jr.. Brems. D. N.. and Stellwagen.’
E. (1981) J. Biol. Chem. 256, 10769-10773.

Rarr. T. L.. White. B. D.. and Purich. D. L. (1979)
J. Biol. Chem. 254, 6107-6111.

Yanagisawa. T.. Easegawa. S.. and Nohri. N. (1968)
Exp. Cell. Res. 88, 86-100.

122.

123.

124.

125.

126.

127.

128.

129.

130.

131.

132.

133.

134.

135.

136.

137.

138.

41

Robayashi. T.. Raji. R.. and Sakai. N. (1976) J.
Biochem. (Japan) 18, 413-418.

Katz. N. and Westley. J. (1980) Arch. Biochem.
Biophys. 204, 464-470.

Gilman. A. G. (1984) Cell 88, 577-579.

Nouslay. N. D. (1984) Trends Biochem. Sci. 8,
39-40.

Northrup. J. R.. Sternweis. P. C.. Smigel. N. D..
Schleifer. L. 8.. Ross. E. N.. and Gilman. A. G.
(1980) Proc. Natl. Acad. Sci. U.S.A. 11, 6516-6520.

Northrup. J. R.. Smigel. N. D.. and Gilman. A. G.
(1982) J. Biol. Chem. 257, 11416-11423.

Smigel. N. D.. Northrup. J. N.. and Gilman. A. G.
(1982) Recent Prog. Norm. Res. 88, 601-622.

Yamazaki. A.. Stein. P. J.. Chernoff. N.. and
Bitensky. N. W. (1983) J. Biol. Chem. 258,
8188-8194.

Fung. B. K. K. (1983) J. Biol. Chem. 258,
10495-10502.

Raziro. Y. (1978) Biochim. Biophys. Acta. 505,
95-127.

Gomperts. B. D. (1983) Nature (London) 306, 64-66.

Neyworth. C. N.. Rawal. S.. and Nouslay. N. D.
(1983) FEBS Lett. 154, 87-91.

Neyworth. C. N.. Wallace. A. V.. and Nouslay. N. D.
(1983) Biochem. J. 214, 99-110.

Godelaine. D.. Beaufay. B.. and Wibo. N. (1977)
Proc. Natl. Acad. Sci. 0. S. A. 18, 1095-1099.

Godelaine. D.. Beaufay. N.. Wibo. N.. and
Amar-Costesec. A. (1979) Eur. J. Biochem. 88,
17-26.

Godelaine. D.. Beaufay. B.. and Wibo. N. (1979)
Eur. J. Biochem. 88, 27-34.

Paiement. J.. Beaufay. B.. and Godelaine. D. (1980)
J. Cell Biol. 88, 29-37.

42

139. Godelaine. D.. Beaufay. N.. Wibo. N.. and Ravoet.
A. -N. (1983) J. Cell Biol. 81, 340-350.

140. Taparowski. E.. Shimizu. R.. Goldfarb. N.. and
Wigler. N. (1983) Cell 88, 581-586.

141. Gibbs. J. B.. Sigal. I. S.. Poe. N.. and Scolnick.
E. N. (1984) Proc. Natl. Acad. Sci. U.S.A. 81,
5704-5708.

CHAPTER II

PRELIMINARY STUDIES

43

44

INTRODUCTION

Nicrotubule protein prepared by the assembly-
disassembly method of Shelanski 88_81, (1) contains
several enzymes that may be microtubule associated and
that might affect microtubule structure or function. The
best studied of these microtubule associated enzymes is a
cANP dependent protein kinase which is capable of
phosphorylating microtubule proteins (2-4).

The tubulin dimer has two distinguishable guanine
nucleotide binding sites. one readily exchanging it‘s
nucleotides with guanine nucleotides in solution (E-site)
and the other (N-site) binding it's nucleotides to
tightly for exchange to readily occur (reviewed in 5). A
transphosphorylation mechanism has been proposed whereby
E-site GDP might be phosphorylated to GTP 18_8118_(6).

It is by this mechanism that ATP and other NTPs might
support microtubule assembly without actually binding to
the tubulin E-site. It is possible that such a reaction
might be catalyzed by a nucleoside diphosphate kinase
(NDP kinase) making this an important enzyme to study.

We were interested in studying a microtubule
associated enzyme and it's role in microtubule function.
The studies presented in this chapter were performed in
order to determine whether NDP kinase is microtubule
associated and whether it might be worth the effort of

further study.

45
EXPERINENTA_ PROCEDURES
The sodium salt of dTDP. pyruvate kinase. and lactate
dehydrogenase were purchased from Boehringer-Nannheim.
All other biochemicals were purchased from the Sigma

Chemical Company.

Tubulin Pregaration - Tubulin was prepared by the method
of Shelanski et al. (1) as modified in this laboratory.
Bovine brain was homogenized with a Waring Blender in 0.5
ml per gram of brain in Assembly Buffer (20 mN NES. pH
6.4. 70 mN NaCl. 1 mN EGTA. and 0.5 mN NgClz). The
homogenate was centrifuged for 45 minutes at 25.000 x g
and then for 45 minutes at 78.000 x g at 40c, The
supernatant was mixed with an equal volume of Assembly
Buffer containing 8 N glycerol. 0.5 mN ATP. and 0.5 mN
GTP and incubated at 37oc for 45 minutes before
centrifuging at 78.000 x g for 45 minutes at 30°C. The
pellets were resuspended in AB and were 1 cycle
microtubule protein. All further cycles were done
similiarly except that the centrifugation steps were

done at 105,000 x g for 1 hour. Nicrotubule protein
prepared by 2 cycles of assembly-disassembly is about

90 t tubulin and 10 s microtubule associated proteins as

estimated by sodium dodecyl sulfate-polyacrylamide gel

electrophoresis and Coomassie blue staining (7). Protein

46
concentrations were measured by the method of Lowry EL.

2.1... (8).

Nucleoside Diphosghate Kinase Assays - NDP kinase was

assayed by a coupled enzyme system similiar to that used
by Bergmeyer (9). The assay mixture contained 83.3 mN

triethanolamine. pH 7.5. 16.7 mN NgCl 67 mN RC], 1.1

2.
mN phospoenolpyruvate. 5 mN ATP. 0.2 mN NADN. 0.2
units/ml pyruvate kinase. 9 units/ml lactic
dehydrogenase. and 0.7 mN dTDP (or 0.32 mN dGDP).

Isotope assays were done by the method of Nakai and
Glinsman (10). Nicrotubule protein was incubated with
[gamma-32P] ATP and GDP or dTDP. The NDP kinase
reaction was stopped by the addition of trichloroacetic
acid to a final concentration of 10 s (w/v). Protein was
pelleted at 4°C and the supernatant was extracted 3
times with 3 volumes of diethyl ether and lyophylized.
The nucleotides were dissolved in water and spotted onto
PEI cellulose plates which were developed in 4 N sodium
formats. pH 3.4. Nucleotides were located under

ultraviolet light. scraped off of the plates and

quantified by liquid scintillation counting.

18 Vitro Nicrot8881e Assembly - 18_vitro assembly of
microtubules was performed in Assembly Buffer with 2
cycle microtubule protein at a final concentration of 2-6

mg/ml. Before use. cycle 2 microtubule protein was

47
centrifuged at 6.500 xg for 5 minutes at 4°C to remove
dust and particulate matter in order to reduce background
light scattering. Samples were kept at 4°C until they
were placed in a Gilford 2400 spectrophotometer with a
water jacketed 4 sample chamber maintained at 37°C.
Nucleotide additions were made just before placing the
samples in the spectrophotometer. Assembly was followed

by measuring light scattering at 350 nm (11).

Sucrose Density Gr881ent Cegtrifugation - Linear
gradients of 30-60 3 sucrose in 10 mN NES. pH 6.5. and 1
mN NgCl2 were formed at 30°C. Nicrotubule protein

was assembled in 10 mN NES. p8 6.5. I mN GTP. 10 s
sucrose. and 1 mN NgCl2 at 30°C. Samples (0.5 ml)

were layered on gradients and centrifuged at 38,000 rpm
in a Beckman SW41 rotor for 1 hour at 30°C. Immediatly
after the run. 0.5 ml fractions were removed sucessively
from the top of each gradient with a syringe. Each
fraction was assayed for protein concentration and for

NDP kinase activity.

48

RESULTS AND DISCUSSION

The results of 18_vitro polymerization of 2 cycle
microtubule protein are shown in Figure 1. Nicrotubule
assembly can apparently be supported better by ATP than
it can by GTP. If the nucleotide is added at the same

time as the saaple is warmed to 37oc, then ATp

supported assembly follows GTP supported assembly after a
lag of several minutes (Figure 1A). If the nucleotides
are preincubated with microtubule protein for 15 minutes
at 15°C. a temperature at which there is no assembly.
then the lag is eliminated and GTP and ATP supported
assembly occur with similiar kinetics. This result
suggests that there may be an extra reaction for ATP
supported assembly that does not occur with GTP supported
assembly and that this reaction may be rate limiting
during early stages of microtubule assembly. The most
logical candidate for this rate limiting step would be an
NDP kinase or transphosphorylase.

We were able to confirm the presence of NDP kinase
activity in 2 cycle microtubule protein preparations by
both a coupled enzyme assay and an isotopically labeled
nucleotide assay as described in "EXPERINENTAL
PROCEDURES”. Nicrotubule protein prepared by two cycles
of assembly-disassembly contains about 0.05 units of NDP
kinase per mg of protein. that is the amount required to

transfer 0.05 umoles of phosphate per minute per mg

49

Figure 1. Assembly of Nicrotubules 18_vitro with ATP.
The 18_vitro assembly of microtubules was measured by
changes in light scattering at 350 nm and 37°C as
described in ”EXPERIMENTAL PROCEDURES". In the
experiment shown in Figure 1A. the nucleotide was
added to the assembly mixture just as assembly was
begun. In the experiment shown in Figure 16. the
nucleotides were added and incubated with the
assembly mixture at 15°C for 15 minutes before the
assembly was begun. In both experiments assembly was
begun by placing the samples at 37°C and all

samples contained 5.6 mg/ml microtubule protein.

50

Figure 1A

 

ole- A

O.l2 —

 

0.06 -

Light Scattering (A350)

0.00—

 

 

 

 

Time (min)

Light Scattering (A350)

51

 

 

 

      

 

Figure 1B
one B
012 [lmM ATP ,,,,,,,
lmM GTP
_/
0.06 '/
e/
I" '/
' :NoAdditifls./
000 ./.
‘ 1 1, l l i l
5 no :5 20 25 so 35

Time (min)

 

52
protein at 300C.

The sucrose density gradient centrifugation of LB.
vitro assembled microtubules is shown in Figure 2. There
are two major protein bands in sucrose density
gradients. The protein band at the top of the gradients
should correspond to tubulin dimers and other
non-assembled microtubule proteins. The faster
sedimenting bands should correspond to intact
microtubules. although we have not verified this by
electron microscopy. Boehringer-Nannheim bovine liver
NDP kinase all sedimented in the top few fractions of an
identical gradient (data not shown). In assembled
microtubules. most NDP kinase sedimented in the top
fractions of the gradient. but the NDP kinase specific
activity was highest in the microtubule fractions. This
suggests that NDP kinase might be a microtubule protein.
an integral constituent of the microtubule.

These early experiments convinced me that NDP kinase
might be a microtubule protein. that NDP kinase might be
involved in assembly reactions. and that microtubule
protein NDP kinase would be an important enzyme to study

further in my thesis research.

53

Figure 2. The association of NDP kinase with sedimenting
microtubules. Assembled microtubules were sedimented
through linear 30-60 3 sucrose gradients in a Beckman
SW41 rotor at 38,000 rpm for 1 hour at 30°C.
Fractions of 0.5 ml were removed successively from
the top of the gradients and assayed for protein

concentration and NDP kinase.

54

Figure 2

 

.._\
.A

Tubulin Region Microtubule Region

\.

_ . A
20 — A \A
.— \.\ . \ . .\A
T A
op A/ \tiij. \kbe—A
l

1
l l l

l 5 IO IS 20
Fraction (0.5 ml)

L4.

40

l

l

(D

—4

 

 

 

—A N DPK ActivityinmoI/mm/ml)

69“

Protein mg m

10.

11.

55

REFERENCES

Shelanski. N. L.. Gaskin. F.. and Cantor. C. R.
(1973) Proc. Natl. Acad. Sci. U.S.A. 18, 765-768.

Sloboda. R. D.. Rudolph. S. A.. Rosembaum. J. L..
and Greengard. P. (1975) Proc. Natl. Acad. Sci.
U.S.A. 18_177-181.

Sheterline. P. (1977) Biochem. J. 133, 273-381.

Sandoval. I. V. and Cuatrecasas. P. (1976)
Biochemistry 18, 3424-3432.

Jacobs. N. (1979) in Nicrotubules (Roberts. K. and
Nyams. J. S.. eds.) Academic Press. New York. pp.
255-278.

Penningroth. S. N. and Kirschner. N. W. (1977) J.
Nol. Biol. 115, 643-673.

Weber. K. and Osborn. N. (1969) J. Biol. Chem. 244,
4406-4412.

Lowry. 0. B.. Rosebrough. N. J.. Farr. A. L.. and
Randall. R. J. (1951) J. Biol. Chem. 193, 265-275.

Bergmeyer. N. U. (1974) Nethogg of Enzymatic
Analysis, Academic Press. New York. pp. 488-489.

Nakai. C. and Glinsman. W. (1977) Biochem. Biophys.
Res. Commun. 18, 1419-1425.

Berne. B. J. (1974) J. Nol. Biol. 88, 755-758.

CHAPTER III

ISOLATION AND CHARACTERIZATION

OF THE

NICROTUBULE ASSOCIATED

NUCLEOSIDE DIPHOSPHATE KINASE

56

57

INTRODUCTION

The tubulin dimer has two distinguishable guanine
nucleotide binding sites (1). At the "non-exchangeable"
(N-aite) the nucleotide is tightly. though not
covalently. bound and does not exchange with guanine
nucleotides in solution (2.3). At the ”exchangeable
site" (E-site) the nucleotide is less tightly bound and
may exchange with free GTP or GDP under certain
conditions (2.3). A GTP molecule must be bound at the
E-site for 18_vitro microtubule assembly to occur (4) and
this GTP is hydrolyzed to GDP during the microtubule
elongation phase of assembly (5.6). After microtubule
depolymerization there are two mechanisms by which E-site
GTP can be regenerated so that the dimer can participate
in further assembly reactions: 1) nucleotide exchange
with GTP in solution and 2) the transphosphorylation of
E-site GDP to GTP without exchange (7-9). Jacobs and
Buitorel (9) have shown that the a transphosphorylation
catalyzed by some enzyme present in microtubule
preparations generates GTP on the dimers of multimeric
microtubule protein rings without detectable exchange
(9).

Nicrotubule protein prepared by cycles of assembly-

disassembly contains a nucleoside diphosphate kinase (NDP

58
kinase) (EC 2.7.4.6) activity (7.9-11). This NDP kinase
co-assembles with microtubule protein through 5 cycles of
assemblydisassembly. co-sediments with microtubules in
sucrose density gradients and. therefore. may be an
integral part of the microtubule. Under non-polymerizing
conditions. NDP kinase may be an integral part of
multimeric microtubule rings (9.11).

While there is some evidence that microtubule
associated NDP kinase catalyzes transphosphorylation at
the tubulin E-site (7.9). this has not been
satisfactorily established. As a first step toward
studying the role of NDP kinase in transphorylation
reactions. we have isolated the NDP kinase from bovine
brain microtubule preparations. We have been able to
isolate only very small amounts of this enzyme and so. as
an aid to future studies. we have isolated the seas NDP
kinase directly from bovine brain cytosol. In this
report. we document the properties of the enzyme from

both sources.

59

EXPERIMENTAL PROCEDURES

Naterials - Rabbit muscle pyruvate kinase. hog muscle
lactate dehydrogenase. phosphoenolpyruvate. and ferritin

were purchased from Boeringer Nannheim. 3

N20. bovine
serum [140] albumin (BSAI). and NEN-963 mixture were

New England Nuclear products. SOS-PAGE standards were
obtained from Bio-Rad and Boehringer Nannheim. All other
chemicals for electrophoresis were obtained from

Bio-Rad. Blue Dextran 2000. Pharmalytes. Sephadex IEF.
and all chromatography column components were Pharmacia
products. except for Fractogel TSK-55 (superfine) which
was purchased from Pierce. Nost other reagents were
purchased from the Sigma Chemical Company. [J-szP] ATP

was synthesized by the method of Glynn and Chappel (12)

as modified by Walsh 81_81, (13).

Preparatiogjof Nicrotubule Protein - Nicrotubule protein
was prepared by the method of Borisy 88_81, (14) as
modified for use with bovine brain by Nurphy and Niebsch
(15). In a typical preparation. 700 grams of bovine
brain. placed on ice within 20 minutes of slaughter. was
trimmed of blood clots and meninges and was homogenized
in PNED buffer (1 ml/gram of tissue) with a Tekmar
Tissuemizer. PNED buffer contained 0.1 N PIPES. pH 6.6.
1 II NzCIZ. 1 mN EGTA. and 2 mN dithiothreitol. The

homogenate was centrifuged in a Sorvall GSA rotor at

60
12.000 rpm for 30 minutes at 4°C. The supernatant was
centrifuged in a Sorvall SS-34 rotor at 18.000 rpm for 45
minutes at 4°C. The supernatant fraction resulting
from this second spin was warmed to 35°C. GTP was added
to a final concentration on 1 mN. and the mixture was
allowed to polymerize for 30 minutes. Nicrotubules were
then pelleted by centrifuging in the SS-34 rotor at
18.000 rpm and 25°C for 50 minutes. The pellets were
resuspended in 40 ml PNED at 4°C and recentrifuged at
18.000 rpm for 60 minutes at 4°C in the SS-34 rotor.
The supernatant portion was removed and was warmed to
35°C. GTP was added to a final concentration of 1 mN.
and polymerization was allowed to continue for 20
minutes.

Nicrotubules were harvested by centrifugation in the
ss-34 rotor at 16,000 rpm for so minutes at 25°C. The
pelleted microtubules were quickly frozen in a dry
ice-methanol constant temperature bath and stored at
-80°C. When required for experiments. these
microtubule pellets were thawed to 4°C. resuspended in
a buffer appropriate to the experiment. and centrifuged
in the SS-34 rotor at 18.000 rpm for 60 minutes at
4°C. This starting material was designated ”2 cycle
microtubule protein" and was used in all the isolations
of NDP kinase from microtubule protein discussed in this

paper.

61
Preparatigg of Nigrot8881e188f K18888_- All steps in this
isolation procedure were performed at 4°C. A
carboxymethyl-Sepharose CL-GB column (3.6xl8 cm) was
equilibrated with NND (50 mN NES. pN6.2. 1 mN NgClz,
and 2 mN dithiothreitol). Nicrotubule protein in NND was
loaded onto the column and the column was washed in NND.
The NDP kinase bound to the column was eluted with
0.75 N NaCl in NND. The NDP kinase containing fractions
were pooled and concentrated to 3 ml with a Nillipore
CX-30 vacuum ultrafilter and loaded on to a Sephadex G-75
(fine) column (2.5 x 18.5 cm) which had been equilibrated
in 50 mN Tris-NCI. pH 7.3. 2 mN dithiothreitol. NDP
kinase was eluted with this same buffer. The NDP kinase
containing fractions were pooled and loaded onto a small
(1-5 ml) column of Blue Sepharose CL-6B which had been
equilibrated in 50 mN Tris-NCI. pH 7.3. 2 mN
dithiothreitol. The column was washed in the same buffer
and NDP kinase was eluted in 2 mN ATP and 2 mN NgClz.
In some preparations a linear gradient of ATP and NgCl2
was used to elute the enzyme from Blue Sepharose. NDP
kinase containing fractions were pooled. concentrationed
with a cx-so ultrafilter. and chromatographed on a column
of TSK-55 (3.2x38 cm) in the same Tris-dithiothreitol
buffer. NDP kinase fractions from TSN chromatography
were then chromatofocused on a PBE-94 chromatofocusing
column (Pharmacia) with a 9.0-6.0 pH gradient. The start

buffer was 50 mN ethanolamine which had been titrated to

62
pH 9.2 with acetic acid. The gradient was formed with
Polybuffer 96 (Pharmacia) diluted 1 to 10. supplemented
to 2 mN with dithiothreitol. and titrated to pH 6.0 with
acetic acid. NDP kinase containing fractions were again
concentrated with a cx-so ultrafilter. frozen in a dry
ice-methanol constant temperature bath and stored at

-80°C until use.

18818t108,0f 88P Niggsegyirectly from Brain - NDP kinase
was isolated directly from brain by the same procedures
used to isolate it from microtubule protein. with the
addition of several preliminary steps. Bovine brain
(1000-1600 grams) was obtained within 20 minutes of
slaughter and immediatly placed on ice. All subsequent
steps were performed at 4°C. The brains were cleaned

of blood clots. minced. and homogenized with a Tekmar
Tissuemizer in 1 ml/gram of 0.1 N PIPES. p86.62. 2 mN
dithiothreitol. The homogenate was centrifuged in a
Sorvall GSA rotor for 60 minutes at 10.000 rpm. We then
performed a 10 s to 27 s polyethylene glycol-8000
(PEG-8000) cut on the supernatant. The supernatant was
supplemented to 10 x PEG-8000 by adding an appropriate
volume of 50 s PEG-8000 slowly and with vigorous
stirring. After 20 minutes of stirring this was
centrifuged in a GSA rotor at 10.000 rpm for 20 minutes.
The supernatant of this spin was supplemented to 27 s

PEG-8000 and after 30 minutes was centrifuged as before.

63
The supernatant of this spin was discarded and the
pellets were frozen at -80°C until they were used.
These pellets were resuspended in 200-300 ml of 50 mN
NES. pH 6.2. 2 mN dithiothreitol by gentle Tekmar
homogenization and the resuspended pellets were
centrifuged in a GSA rotor at 12,000 rpm for 30 minutes.
The supernatant of this spin was used as starting
material for the further isolation of NDP kinase by the
same procedure that had been used for the isolation of
the microtubule protein preparation NDP kinase. This
procedure consisted of 5 chromatography steps:
CN-Sepharose. Sephadex G-75. Blue Sepharose. TSK-55. and
chromatofocusing on PBE-94. For some preparations the
Sephadex G-75 step was eliminated. The final product was

stored at -80°C.

818ctrophore818_- Discontinuous SDS-PAGE was done by the
method of Laemmli (16) with some modifications when the
gels were to be silver stained. For silver staining the
gels were 0.75 mm gradient gels of 9-16 x acrylamide with
a bis-acrylamide to acrylamide ratio of 0.053 which is
twice the usual ratio. The gel solutions were passed
through a 0.2 uN Nillipore filter before polymerization
to remove insoluble material. Samples were prepared by
placing in a boiling water bath for 2 minutes and mixing
with two volumes of 0.25 N Tris ECI. pH 6.8. 5 s 808.

3 s sucrose. and 10 mN dithiothreitol. They were placed

64
in a boiling water bath again for 3 minutes and when they
had cooled to room temperature they were alkylated with
0.06 N iodoacetamide in the dark for 1 hour as suggested
by Nashimoto 81_81, (17) in order to eliminate the
artifact bands appearing on the stained gel. Immediatly
after electrophoresis the gels were silverstained by the
method of Nerrill (18). This method was modified by
adding two 15 minute water washes between the oxidizer
and silver steps and a 1 minute water wash between the
silver and developer steps and by using 5 s acetic acid

to stop development.

Determinatiog of Kinetic Constants - Kinetic data for
both NDP kinases were analyzed using a computer program
that performed a weighted non-linear regression-fitting
of the data to the Wichaelis-Nenton equation as described
by Duggleby (19). In all cases we assumed that the
standard deviations of velocity measurements increased
proportionatly with velocity. Nucleotide concentrations
in all kinetic experiments were determined by optical
density measurements at pH 7.0 using the following
extinction coefficients (all in cmzlumole):

(25.7-14.7 for ATP. €256.12“! for GTP. €282.10'0 for

UTP. (”o-12.8 for on. and 6237'9-9 for dTDP.

65
Assays — NDP kinase was assayed by the pyruvate kinase-
lactate dehydrogenase coupled enzyme system described by
Bergmeyer (20). These assays were performed with 25-200
81 of sample in a final volume of 500 N1 at 30°C.
Adenylate kinase assays were done in the same way except
that 1.6 mg/ml ANP replaced the dTDP. For each assay.
one unit of activity is defined as the amount of enzyme
causing the oxidation of 1 umole of NADN per minute.
Protein content was determined by the method of Bradford
(21) because polybuffers did not interfere with this

assay.

Sedimentation Eguilibrium Determination of Native Enzyme
Nolecular Weight - The molecular weight of NDP kinase was

determined by a modification of the method of Pollet El.
81, (22) which we have developed. Protein samples were
dissolved in 20 aN PIPES. pH 7.0. 2 mN dithiothreitol.
1 I! lzClz. 2 s PEG-4000. and 20 mg/ml Dextran T40
(Pharmacia). The Dextran T40 is added to form a
stabilizing gradient so that minimal mixing occurs during
rotor deceleration. The samples also contained 4 uCi of
820 per ml so that the volume of fractions removed
after centrifugation could be measured by liquid
scintillation counting. Protein samples of 150 ul were
centrifuged in the Beckman Airfuge for 24 to 48 hours at
4°C. [14C1-bovine serum albumin was included in

every run as an internal standard since we did not have a

66
method for accuratly determining the angular velocity of
the rotor. At the end of the centrifugation period. the
tubes were carefully removed from the rotor and 10-30 Nl
fractions were removed from the meniscus with a Beckman
Nicrotubue Fractionator. Fractions were assayed for NDP
kinase activity and then liquid scintillation counted
for 38 in a Beckman CPN-100 with NEN-963 mixture.
Fractions from [14C1-bovine serum albumin containing

38 and 14C in the same

samples were counted for both
way. The average radius from the center of rotation. r.
of each fraction was read from a table created by a
computer program that numerically integrated an
expression for the volume V(r). between the meniscus and
a horizontal plane through a point at a distance. r. from

2

the axis of rotation (22). When centrifuging at

equilibrium.

(1) M—Eraiiifll.

drz 231

where C is the protein concentration: Nr 13 the
molecular weight: V is the partial specific volume: '2
is the angular velocity of the rotor: p is the density of

the solution: T is the temperature: and R is the gas

constant .

67
Thus. if we assume that V(NDP kinase)'v(BSA)’

then:

fill£_£l_ (NDP kinase)

dra nr(snp kinase)

The quantity d(ln C)/dr2 was determined for each

protein by least-squares fitting of In C versus r2 to a
line where C is in counts/minute for BSA and in units/ml
for NDP kinase. Equation 2 could then be solved for

Nr(NDP kinase). the molecular weight of NDP kinase.

RESULTS

18818tiog of Nicrotubule-associated NDP Kiggge - We have
been able to isolate the NDP kinase form 2 cycle
microtubule protein by a 5-step procedure: CN-Sepharose.
Sephadex G-75. Blue Sepharose. Fractogel TSR-55
(superfine). and chromatofocusing on PBE-94. This method
produces a homogeneous preparation of NDP kinase but only
in very small quantities. Using silver staining
intensity to approximatly measure protein concentrations.
we would estimate that from 1 gram of microtubule protein
we can isolate only about 1-5 ng of NDP kinase. even
though the yield of activity through the procedure is
15-20 8.

Figure 1 shows the chromatography of 2 cycle
microtubule protein on CN-Sepharose. Both peaks I and II
contained tubulin and high molecular weight proteins as
shown by SOS-PAGE and. therefore. appear to represent
different oligomers of tubulin and microtubule associated
proteins. Peak III did not contain detectable protein
and was probably nucleotide that was present in the
microtubule protein starting material. Batch elution of
the the column with 0.75 N NaCl yielded a NDP kinase
preparation that was contaminated with adenylate kinase
activity (1 unit/10 units of NDP kinase). Salt gradient
elutions of NDP kinase were attempted. but NDP kinase

eluted in three or more separate peaks. This might be

Figure 1. Carboxymethyl Sepharose chromatography of 2
cycle microtubule protein. Bovine brain microtubule
protein (20 mg in 3 ml) was loaded on a column (3.6 x
16 cm) which had been pre-equilibrated in NND at
4°C. The column was washed with NND and then with
0.75 N NaCl in NND while 4 ml fractions were
collected. The absorbance of each fraction at 280 nm
was measured and each fraction was assayed for NDP
kinase activity. The NDP kinase peak eluted in 0.75
N NaCl contained 1.86 units of NDP kinase and 0.12

units of adenylate kinase activity.

Figure 1

~_E\DEVVEQZ all.

 

 

 

 

 

 

2m '
nv nu nXU
_ _ 6
O
I
/
ole O
O 4
N
M
5
7. o
O ! 3
.\
\I Q
m a\e\a .- O
s.
n my \
e\\e\ ”I nlv
\
I eiii... eisi .
— P
O 5. 00

E: 0mm woz<mmomm< ellls

FRACTION NUMBERM ml)

70

71
due to NDP kinase binding to other microtubule proteins
or to the presence of multiple apparent isozymes.

The second step in the isolation procedure.
chromatography on Sephadex G-75 (Figure 2). was performed
in order to change the buffer and to remove the adenylate
kinase activity which could not be separated from the
preparation by the subsequent Blue Sepharose
chromatography step in the procedure. Sephadex G—75-
purified NDP kinase was then chromatographed on Blue
Sepharose. which has been used to isolate a number of
nucleotide requiring enzymes. NDP kinase bound to Blue
Sepharose and could be eluted by ATP: either 2 mN ATP in
a batch elution or a linear 0-7 mN ATP gradient as shown
in Figure 3. NDP kinase could not be eluted from Blue
Sepharose with UTP or GTP. although both of these
nucleotides serve as phosphate donors in the NDP kinase
reaction. If 2 cycle microtubule protein was
chromatographed directly on Blue Sepharose without a
preceding ion-exchange step. the NDP kinase was not
retained by the coluan material. There may be sufficient
nucleotide in 2 cycle microtubule protein to prevent NDP
kinase binding to Blue Sepharose.

NDP kinase chromatofocused as a single peak or
activity at a pH of 8.0 (Figure 4A). As we will discuss
below. the NDP kinase sample for chromatofocusing was
first chromatographed on Fractogel TSR-55. which removed

ATP from the Blue sepharose elution buffer. and was then

72

Figure 2. Sephadex G-75 chromatography of Carboxymethyl
Sepharose isolated NDP kinase. The NDP kinase
fractions eluted from a Carboxymethyl Sepharose
column as shown in Figure 1 were concentrated to 3 ml
with a Nillipore CX-iO ultrafilter and loaded on a
2.5 x 18.5 cm column of Sephadex G-75 fine. The
column had been equilibrated and was developed with
50 mN Tris-8C1. pH 7.3. containing 2 mN
dithiothreitol at 40C. Fractions of 3 ml were
collected and assayed for absorbancy at 280 nm. NDP
kinase activity. and adenylate kinase (AR) activity.
This step was included in the isolation procedure to
desalt the preparation and to reaove adenylate

kinase.

73

Figure 2

O. x _E\fl_c3 x<
8 6.

_E\m::3 xdoz
2

 

 

 

 

 

 

 

 

 

 

4.
O O O O O
_ . _ . _ . _ _
5 h w
_.I K
G DA\°\LTI
w a
A
H
P O
E 2
S
W
D elo
o N
8
2 Q
D iiiiiiii i
nu mlnnHlIIl.AY. IIIIIII lllllll .QY mw
'5 "Hui-1.1 illil'liil'l
iiiiinIiiPiiiiii
5
_ . _ . _ _
8. 5 4. 00
O O O

 

e: 08 8249584

FRACTION (3mI)

74

Figure 3. Blue Sepharose chromatography of Sephadex G-75
purified NDP kinase. Sephadex G-75 fractions
containing NDP kinase activity were pooled and loaded
on a 1 ml column of Blue Sepharose CL-6B which had
been equilibrated in 50 mN Tris-HCl. pH 7.3.
containing 2 mN dithiothreitol at 40c. The coluln
was washed first with 60 ml of the equilibration
buffer and then with a 0-7 mN linear gradient of ATP
in the same buffer. Fractions of 3 ml were collected
and assayed for NDP kinase activity. Activity
containing fractions were pooled. concentrated to
2 ml with a CX-iO. and they were either stored at
'300C or used immediatly for the chromatofocusing

step in the isolation procedure.

0.6

,0
.b

NDPK Units/ml

,0
N

75

Figure 3

 

 

 

 

BLUE SEPHAROSE '6
NDPK '
ATP

' - 4
i 4“"
- E..

- 2

. 4
O J O

 

 

«5 2o 2%) 40 so
VOLUNEIGnD

76

Figure 4. Chromatofocusing profiles of NDP kinase.
Chromatofocusing was performed as described in
”EXPERIMENTAL PROCEDURES".

A. Chromatofocusing of NDP kinase during its
isolation from 2 cycle microtubule protein. The NDP
kinase sample had previously been chromatographed on
TSK-55s and had been supplemented to 0.6 mN with ADP
15 minutes before loading it on a 1 x 28 cm column of
PBE 94.
B. Chromatofocusing of NDP kinase during its
isolation directly from bovine brain. This step was
performed just as described in part A.
C. Chromatofocusing of NDP kinase without a
pretreatment with ADP during its isolation from
microtubule protein. The NDP kinase sample had just
been eluted from Blue Sepharose with 2 mN ATP and
2 mN I3012. The ATP was not removed by a gel
filtration step and the sample was not pretreated
with ADP as in parts A and B. The sample was
chromatofocused on a 1.2 x 30 cm column of PBE 94

which had been equilibrated at pH 8.4.

77

Figure 4A

+9.0

‘ b8.0

‘ 70
96.0

 

 

P
5
AV

 

. b

4 s
O O

.E\m._..23 wm<z§ aoz

 

 

30 40

20
FRACTION (3mII

 

N DP KI NASE UNITS/ml

78

Figure 4B

 

 

IO

1

so
FRACTION (3mli

   

 

40

E“
o

'0!

NDP KINASE UNITS lml
'o

,0
0

79

Figure AC

 

<3
(3

 

 

  

210 ‘ s‘o
FRACTION (SIM)

 

9.0

:8.0

~10

 

   

80
treated with 2 mN ADP. This was done to remove the
phosphate from any NDP kinase-phosphorylated
interaediates that might be present and that might have a
more negative charge than unphosphorylated NDP kinase.

This isolation procedure produced an NDP kinase that
was free of detectable ATPase. GTPase. adenylate kinase.
and protein kinase activities. The yield of NDP kinase
from 2 cycle microtubule protein was 15-208. but the
-fold purification could not be determined since there
was too little protein to measure.

We have examined our isolated NDP kinase by
SOS-PAGE. Coomassie Blue R-250 staining after
electrophoresis was not sensitive enough to reveal
protein bands. so gels were stained by the silver
staining method of method of Nerrill 25.21; (18).
Several problems had to be overcome before protein bands
could be detected by this procedure. Polybuffers in the
final enzyme product interfered with the silver staining
of NDP kinase in our initial gels which were 123
acrylamide. Polybuffers stained as diffuse yellow bands
that obscured the lower molecular weight protein bands.
For this reason. we decided to use 9-168 acrylamide
gradient gels to seperate polybuffers as much as possible
from protein hands. When the gels were fixed overnight
in 10% acetic acid. 403 methanol. NDP kinase bands were
not visible after staining. Bashimoto 81_81, (17) had

discovered that proteins as large as 17,200 daltons were

81

Figure 5. Sodium dodeclysulfate polyacrylamide gel
electrophoresis of NDP kinase. Electrophoresis and
staining were done as described in ”EXPERINENTAL
PROCEDURES”. The standards used to estimate NDP
kinase molecular weights were: lysozyme. 14.400:
myoglobin. 17.200; soybean trypsin inhibitor. 21.500:
carbonic anhydrase. 31.000: RNA polymerase. 39.000:
ovalbumin. 45.000: bovine serum albumin. 66.200;
phosphorylase b. 92.500: and RNA polymerase. 155.000
and 165.000 daltons.
A. This gel shows the NDP kinase isolated from 2
cycle microtubule protein. After the initial silver
staining this gel was cycled through silver nitrate
and developer twice more. The background was
lightened as described by Nerrill 88_81, (18) and the
gel was toned by the method of Berson (31). The gel
was photographed with Polaroid type 55 P/N film and
the negatives were printed on a variable contrast
paper (Kodak Polycontrast Rapid II RC) at high
contrast.
B. This gel shows both the NDP kinase isolated from
microtubule protein (lane 1) and isolated directly
from brain (lane 2). The preparations were run as
far down the gel as possible so as to amplify any
molecular weight differences that might exist. The

gel was stained as for gel A.

82

Figure 5

155 — . 155-

92 r 92 —
- 68-
39 -

39 r
31 -

21.5 "

Ir- ' I:

17.2 '- 21.5-

 

14.4 — . -

83
eluted from their gels after soaking in similiar fixing
solutions and that this elution was reduced by doubling
the bisacrylamide to acrylamide ratio. Accordingly. we
increased this ratio to 0.053 for our gradient gels and
stained the gels immediatly after electrophoresis to
minimize the elution of lower molecular weight proteins.
A protein alkylation procedure (17) using iodoacetamide
was performed on samples before electrophoresis to
eliminate the artifact bands that frequently appear after
silver staining. As shown in Figure 5. when using this
electrophoresis and staining protocol. the NDP kinase
isolated from 2 cycle microtubule protein was found to
consist of a doublet of protein bands with apparent
molecular masses of about 18.000 daltons. No other

protein bands were visible.

Determinatiog of Native Enzyme Nolegular Weight - We

 

could isolate only small amounts of NDP kinase from 2
cycle microtubule protein. and it lost activity as it
became more dilute. Therefore. to determine the native
enzyme molecular weight. we needed a technique that would
alow us to use very small amounts of protein with a
minimum of enzyme dilution. We chose a method deveIOped
by Pollet 81_81, (22) for sedimentation equilibrium
determination of protein molecular weights using the
Beckman Airfuge. We modified this technique as described

in ”EXPERINENTAL PROCEDURES" for easier determination of

84

fraction volume and for determination of rotor velocity

lith a [14C]-bovine serum albumin internal standard.

The results of such an experiment are shown in
Figure 6. Figure 6A shows the sedimentation equilibrium
profile plotted as NDP kinase activity vs. volume from
the initial meniscus. r2 was determined from the
volume data with the use of a computer program and the
data was replotted as shown in Figure 6B. we could see
that In C was a linear function of r2, indicating that
sedimentation equilibrium had in fact been achieved.
Table I summarizes the results of three sedimentation
equilibrium experiments with NDP kinase isolated from
microtubule protein. The mean molecular weight of NDP
kinase was 103,000 +/- 7.000. This result. along with
the SOS-PAGE data discussed above. is consistant with an
NDP kinase structure of six subunits. each of about

18.000 daltons.

18818tiogcof188P Rinase Directly from Brain - We were
able to isolate only very small quantities of NDP kinase
from 2 cycle microtubule protein. We hoped to make an
antibody to this enzyme and to do detailed molecular
studies of NDP kinase function in microtubule events:
however. these quantities were too small for these
purposes. To obtain larger quantities of enzyme we
decided to isolate it directly from bovine brain without

a preliminary microtubule protein isolation. Nost tissues

85

Figure 6. Sedimentation equilibrium profiles of NDP
kinase isolated from microtubule protein. The
experiment was performed as described in
"EXPERIMENTAL PROCEDURES".

A. Samples (150 ul) of NDP kinase or [14c] bovine

serum albumin were centrifuged in the Airfuge at

40C and 7 p.s.i. for 24 hours. At the end of the

run 10 to 30 ul fractions were sucessively removed

from the meniscus and assayed for NDP kinase activity

or 14C radioactivity. Each fraction was also

assayed for 3“20 radioactivity to determine its
volume.

B. In this figure the sedimentation equilibrium
profile is plotted as In C vs. r2, The values of

r2 were determined from the volume between each
fraction and the meniscus by a computer program as
described in the text. From the ratio of the slopes
of the two lines. the molecular weight of NDP kinase

can be calculated to be 104.200.

86

Figure 6A

 

A

NDPK mU/ill

N

 

40 60

 

I
80

BSA

1L
l00

 

I
I20

VOLUME FROM MENISCUS (pl)

NDPK

J
I40

 

.n
C)
C)

["tjssn CPM/hl

NDPK mU/pl

87

Figure 6B

 

500

200

3

["c] BSA CPM/ul

(fl
(3

 

 

 

 

TABLE I

Sedimentation equilibrium determination of

NDP kinase molecular weight

This is a summary of the results of three
sedimentation equilibrium determinations of the molecular
weight of the NDP kinase isolated form microtubule
protein. All three experiments were performed as
described in ”EXPERINENTAL PROCEDURES". Experiment 3 is

depicted in Figure 6.

Experiment Time Airfuge Approximate Nolecular
pressure velocity weight
h p.s.i. rpm
1 48 6 43.000 110.000
2 48 6 43.000 96.000
3 24 7 46.000 104.000

Nean +/- SE 103,200 +/- 6.800

89

have been reported to contain multiple isozymes of NDP
kinase so. to avoid confusion. we decided to isolate NDP
kinase from whole-brain by the same chromatographic
procedures used to isolate the enzyme from microtubule
protein preparations. In this way. if multiple forms
were found at any step. then we would know which form to
choose for further isolation. We needed a preliminary
isolation step to decrease the volumes and the amount of
protein. With this goal in mind. we produced a PEG-8000
precipitation profile of the NDP kinase from microtubule
protein (data not shown) and found that this NDP kinase
precipitated between 10 and 27s (w/v) PEG-8000.
Accordingly we used this same 10 to 27s PEG-8000
differential precipitation step before the chromatography
steps in our isolation of NDP kinase directly from brain.

A typical isolation of NDP kinase from brain is
summarized in Table II. The procedure consisted of a
10-278 PEG-8000 fractionation. CN-Sepharose
chromatography. dialysis. Blue Sepharose and Fractogel
TSK-55 chromatography. and chromatofocusing from pH 9 to
6. Nost of the NDP kinase in bovine brain was
particulate and was pelleted in the first centrifugation
step. Because the enzyme we were trying to isolate was
cytosolic and not organellar. we used the first
supernatant as a reference for calculating yield and

-fold purification. The losses in NDP kinase yield

9O

m.e o.~vme .ooe An.o ecu-ooOCOuoso»:o
~.n o.nqn n.5n o.n amazes
m.v o.on~ n.v~ AA esoueenem code
e.o n.n Av.o one secsogaonisu
o.~n o.~ on.o ooo.n can wen
o.mm o.~ on.o ooo.ee one you
ooe o.~ ee.o oov.o~ accuses-non
a Ducal mE\wu«co we
>ufi>au0<
0A.“? COuUCOuuuhq-m OuquOQm c~000um 000m

.cmeuaveooum
deucefiuuenxm: cw vsnwuuuefl we fleEuOuuen 0903 wxewee was couueHOmu ssh

canon sau>on EOLu >~u0euau ewecux mnz no couueaoma

HH HQQ‘B

91
between the 10 and 27 8 PEG-8000 steps probably represent
storage losses of activity. The enzyme was routinely
stored at -800C when it was in the 27 s PEG-8000
precipitation pellet. NDP kinase was typically isolated
directly from brain with an overall yield of 1-5 8 and
with a purification of 1500 to 2000-fold.

The chromatofocusing step in the isolation is shown
in Figure 4B. NDP kinase focused at a pH of 8. the same
as for the NDP kinase isolated from microtubule protein.
In prelimianary attempts at isolationfrom both
microtubule protein and directly from brain.
chromatofocusing was done without a preceeding gel
filtration step and without an incubation with ADP. A
typical result is shown in Figure 4C. NDP kinase focused
in multipke peaks at pEs below 8. NDP kinase from other
sources have been reported to with phosphorylated
intermediates (23-25). An NDP kinase hexamer could
contain as many as six phosphates per molecule and the
various phosphorylated forms would focus at more acidic
pNs. By gel filtration on Fractogel TSR-55. we removed
from the preparation the ATP. which had been used to
elute the NDP kinase from Blue Sepharose. We could then
remove the phosphate from the intermediate by incubating
the enzyme with excess ADP with the result that the
enzyme then focused as a single peak at a pH of 8

(Figures 4A and 4B).

92

SDS-PAGE of whole-brain NDP kinase showed that this
enzyme. like the microtubule protein enzyme. consisted of
two proteins of about 18.000 daltons. As shown in Figure
5B. these are the same subunit molecular masses present
in the NDP kinase from microtubule protein. NDP kinase
isolated directly from brain sometimes. but not
consistantly. had a contaminant polypeptide band at
36.000-38.000 daltons (Figure 5B). This may be a dimer

of NDP kinase subunits formed during sample processing.

Comparisons of_:8e Two NDP Rinases - Both NDP kinases.
isolated from microtubule protein and isolated directly
from bovine brain. appeared to be identical. They were
isolated by the same method and chromatographed in the
same positions during all steps of the isolation
procedure. Both preparations appeared identical on an
SDS-PAGE gel (Figure 5B) and they had the same pl
(Figure 4). To further test the hypothesis that the two
NDP kinases were identical. we did two more experiments.
First. we measured the kinetic constants. km and
relative Vmax. for the two enzymes. As shown in
Table III. the kinetic constants measured for both
enzymes were similiar. The Rm values for the two NDP
kinases that we had prepared were very different from
those reported by Robinson 88_81, (26) for the bovine

brain particulate NDP kinase. This result is consistant

93

.25 0A «0 cowueuucoocoo efiauoeaosc e ue Deusaee:

nfluiwl c0e:«n01 ha Cannons» sesne> ecu sue seesaw
.caeusveOOLL deuceEuuenxn: seven Denuuwmefl

we Hancoao :«eun e:«>on Ecuu >Auoeuufi veueaoww amenax aaz ecu ma mash
.cweuobeOOLQ deuceeuueaxu: have: venauusefi ee moodueueneun daemown

eusnouOLOAE e~0>0 ~ cueun ec«>on EOuu DeveHOau ese:«x an: en» mu suave

.AoNv

0N.o om.o hm.o LOB”
H.H nn.o Do.m $9.5 mFU
n.~ mN.o fin.o boa m9:
0v no.0 5n AN.o vn aH.o GEO
00H o~.o cod N.A cod h.d mflt
sauce :8 sauce :8 sauce :6
----...ufl::m.._u.fl-::---mm........----...muu.wi-iiflfliimmfiu... ................
iiuumm... .......... may“ ........... Mam“ ...... 3.53.5
sueasuuuuem ~0a0u>0 eusnsuouo«t

.eeasce cues uOu mad uOu xe£> any 0» e>uueaeu aes~e> xee>
as» one scan: mesae> xee> e>aueaeu we Deuuoaeu sue wesae> wee> .cmeusveOOLa
deuceeuuenxm: news: venuuuweb we snow one: ecouuerOAeu use aaeme<

wesech no: «0 weuuueOOLQ ouuecux uO coauueneoo

HHH man‘9

94
with the view that the NDP kinase isolated from
microtubule protein preparations and the whole-brain NDP
kinase are identical but that both are different from the
brain particulate enzyme.

Next. we compared the thermal denaturation
characteristics of the two NDP kinase that we had
isolated. Both enzymes were heated at 600C for various
times in the presence of 1 mg/ml bovine serum albumin.
The rates of activity loss were measured (Figure 7) and
were the same for both enzymes. For both enzymes. 50% of
the initial activity was lost in 9 minutes and 75% was
lost in 19 minutes. This result is consistant with the

idea that both enzymes are identical.

Phosphorylation of NDP Rinase - The behavior of NDP

kinase on chromatofocusing columns (Figure 4) suggusted
that this enzyme could be phosphorylated. changing its
ionic properties. and that this phosphorylation was
readily reversable. covalently bound phosphate being
removed by ADP. We initially thought that the
faster-migrating subunit observed after SOS-PAGE might be
a phosphorylated form of the slower-migrating subunit.

To test this view. we attempted to phosphorylate NDP
kinase with [ -32P]ATP. The enzyme was then
electrophoresed in SDS and an autoradiograph of the dried
gel was prepared. As shown in lane A of Figure 8. both

subunit bands were phosphorylated. When NDP kinase that

95
had been phosphorylated in this way was subsequently
treated with a 10-fold excess of ADP. most of the
phosphate label was removed. This is shown in lane B of
Figure 8 and suggests that the phosphorylated enzyme is
an intermediate in the reaction. and not a more stable

phosphate ester.

96

Figure 7. Neat denaturation of NDP kinase. Samples of
NDP kinase isolated either from microtubule protein
prearations or directly from bovine brain as
described under "Experimental Procedures" was
incubated in 0.1 N PIPES. pH 6.6. with 2 mg/ml bovine
serum albumin in a 60°C water bath for various
times. The incubation was terminated by placing the
saaples on ice. -4|—-: the microtubule protein NDP

kinase.-—C}— : the NDP kinase isolated directly from

bovine brain.

% ACTIVITY REMAINING

100

75

50

25

97

Figure 7

 

  

  

 

,
0|
0
J l
0 1o 20 30
TIME at 60‘C (on...)

 

Figure 8. Autoradiograph of 32P-labeled unp kjheee,
Samples of NDP kinase isolated directly from bovine
brain as described under "Experimental Procedures"
(115 ng in 5 ul) were incubated with 5 uCi of
[Ea-Ia-32P1ATP (1100 cpm/pmole) in a total volume
of 10 ul for 2 minutes at room temperature. The
sample for lane A was mixed with 25 ul of SDS-PAGE
sample buffer which contained 5: SDS. and then with
10 ul of ADP. The sample for lane B was mixed with
10 ul of ADP. incubated for 5 minutes at room
tenperature. and then mixed with 25 ul of SDS-PAGE
sample buffer. Two lanes of standards; the sane as
shown in Figure 5: were run on the same gel. After
electrophoresis. the two radioactive lanes were
separated. dried onto filter paper. and
autoradiographs were prepared. The lanes of
standards were fixed in 12.58 (w/v) trichloroacetic
acid. stained with 0.13 (w/v) Serva blue W overnight.
and destained in water. A doublet of bands was
phosphorylated with molecular weights of about
18.000. As can be seen by comparing the two lanes.
the 5 minute chase with nonradioactive ADP removed
much of the labeled phosphate for the sample in lane

99

Figure 8

 

<—92.5

‘ e—os

4—31

<—-21

<—17

' <—14.4

100

DISCUSSION

We have isolated the NDP kinase from 2 cycle
microtubule protein in relatively high yield but only in
nanogram quantities. This enzyme has a molecular mass of
103,000 +/- 7.000 daltons as determined by sedimentation
equilibrium ezperinents (Table I). On silver stained
SDS-PAGE gels. this NDP kinase has a doublet of protein
bands at about 18.000 daltons (Figure 5). We therefore
conclude that microtubule protein NDP kinase is a
hexamer. but we cannot determine the exact composition of
subunits. This microtubule protein NDP kinase
chromatofocused as a single isozyme with a pl of 8.0
after removal of ATP and dephosphorylation with ADP
(Figure 4).

In order to obtain larger quantities of this enzyme.
we isolated it directly from brain without a preliminary
microtubule protein isolation. To insure that we would
isolate the same isozyme of NDP kinase directly from
brain that we had isolated from 2 cycle microtubule
protein. we used an identical isolation procedure with
the addition of a 10-278 PEG-8000 differential
precipitation (Table II). This whole-brain NDP kinase
chromatofocused at a pH of 8.0 (Figure 4) and contained‘
at least two different subunits of about 18.000 daltons.
We think that both of the enzymes we have isolated. the

one isolated from microtubule protein and the one

101
isolated directly from brain. are identical. They have
the same pIs. They are isolated by the same
chromatographic procedures and behave similiarly during
each chromatographic step. The two enzymes have similiar
catalytic behavior. similiar Km values and relative Vmax
values. and they thermally denature at the same rates.
We propose that they are the same enzyme.

Robinson 88_81, (26) have isolated a NDP kinase from
the particulate fraction of bovine brain. This NDP
kinase was a hexamer of 16.600 dalton subunits and did
not consist of two separate subunits as revealed by
SDS-PAGE in contrast to the NDP kinase isolated in this
study. The particulate enzyme had a p! of 8.4. similiar
to the microtubule protein enzyme. The particulate
enzyme could be eluted from Blue Sepharose by GTP. while
under similiar conditions. the microtubule protein enzyme
could not. even though it would use GTP for a substrate.
The particulate and microtubule protein NDP kinases had
different catalytic properties (Table III). Finally. the
microtubule enzyme would not compete with detergent
micelles for Blue Sepharose dye binding as would the
particulate form and could. therefore. not be isolated by
the method of Robinson 88_81, (26). The NDP kinases that
have been studied have been principally cytoplasmic or
mitochondrial enzymes (27) although. in rat liver. 12* of
the total NDP kinase activity is nuclear (28). Narr EL.

81, (29) have presented convincing evidence that the NDP

102

kinase of bovine brain microtubule protein preparations
is not of mitochondrial origin. We suggest that the
particulate NDP kinase from bovine brain and the NDP
kinase that we have isolated in this study are different.

There are several possible explanations for the two
subunit bands present after SDS-PAGE of NDP kinase from
microtubule protein peparations. One is that they are
different gene products. A second possibility is that
they are created by some sort of post-translational
modification of a single gene product. At first. we
thought that the faster-running species might be a
phosphorylated intermediate. but this can be ruled out.
since. as seen in Figure 8. both bands are
phosphorylated. We cannot. however. discount the
possibility that the faster running species is
phosphorylated creating a stable phosphate ester. not at
the active site. Robinson 88_81, (26) found that
treatmentof the bovine brain particulate NDP kinase with
various proteases produced both a loss of activity and
the appearance of multiple isozymes. electrofocusing at
lower pNs. Although our NDP kinase was a single isozyme.
it is still possible that renoving a few amino acid
residues by protease action could produce the lower
molecular weight form without substantially altering the

enzymes pl.

103

The NDP kinase isolated directly from brain could be
phosphorylated to a high-energy phosphate intermediate as
shown in Figure 8 which could be dephosphorylated by the
addition of ADP. Both subunits could be phosphorylated.
suggesting that they are both catalytic subunits. We
believe that many of the multiple forms of NDP kinase
that we encountered during early attempts at isolation
were caused by this phosphorylation. For example. when
NDP kinase was chromatofocused without the removal of ATP
and pretreatment with ADP. multiple. more acidic forms of
the enzyme were observed (Figure 4C). In retrospect.
this is not surprising: as many as six phosphates could
be present on any one NDP kinase molecule and this would
substantially change its charge properties.

It has been suggested by Penningroth and Kirschner
(7) that a microtubule-associated NDP kinase might
catalyze the transphosphorylation of tubulin E-site GDP
to GTP without this nucleotide leaving its binding site.
This guanine nucleotide is tightly bound at the E-site
when tubulin is in oligomeric microtubule protein rings
(8) and may not freely exchange with GTP. Of course. it
is still not known whether such rings exist 12.21224 but
they or some similiar form of microtubule protein
oligomer may exist in the cell and require NDP kinase to
regenerate E-site GTP. In this way. NDP kinase could
control microtubule assembly or function. The idea that

NDP kinase could regulate a GTP-dependent process is not

104
unprecedented. The activity of plasma membrane guanine
nucleotide-dependent adenylate cyclase may be regulated
by the binding of NDP kinase to nearby sites in the
plasma membrane (30). Now that we have isolated the NDP
kinase from microtubule protein preparations we have a
tool for studying the role that this enzyme may play in

microtubule assembly and function.

10.

11.

12.

13.

14.

15.

16.

105

REFERENCES

Weisenberg. R.. Borisy. G.. and Taylor. E. (1968)
Biochemistry 1, 4466-4479.

Levi. A.. Cimino. N.. Nercanti. D.. and Calissano. P.
(1974) Biochim. Biophys. Acta 365, 450-453.

Penningroth. S. N. and Kirschner. W. W. (1978)
Biochemistry 11, 734-740.

Robayashi. T. and Simizu. T. (1976) J. Biochem. 18,
1357-1364..

David-Pfeuty. T.. Erickson. N. P.. and Pantaloni. D.
(1977) Proc. Natl. Acad. Sci. U.S.A. 18, 5372-5376.

Bonne. D. and Pantaloni. D. (1982) Biochemistry 81,
1075-1081.

Penningroth. S. N. and Rirschner. N. W. (1977) J.
Nol. Biol. 115, 643-673.

Weisenberg. R. C.. Deery. W. J.. and Dickinson. P. J.
(1976) Biochemistry 18, 4248-4254.

Jacobs. N. and Buitorel. P. (1979) Eur. J. Biochem.
99, 631-622.

Nickerson. J. A. and Wells. W. W. (1978) Biochem.
Biophys. Res. Commun. 88, 820-826.

Burns. R. G. and Islam. K. (1981) Eur. J. Biochem.
117, 515-519.

Glynn. I. N. and Chappell. J. B. (1964) Biochem. J.
90, 147-149.

Walsh. D. A.. Perkins. J. P.. Brostrom. C. 0.. Bo. E.
S.. and Nrebbs. E. G. (1971) J. Biol. Chem. 246,
1968-1976.

Borisy. G. G.. Narcum. J. N.. Olmsted. J. B.. Nurphy.
D. B.. and Johnson. K. A. (1975) Ann. N. Y. Acad.
Sci. 253, 107-132.

Nurphy. D. B. ans Biebsch. R. R. (1979) Anal.
Biochem. 88, 225-235.

Laemmli. u. x. (1970) Nature (London) 222. eso-ssz.

17.

18.

19

20.

21.

22.

23.

24.

25.

27.

80.

31.

106
Bashimoto. F.. Borigome. T.. Kangayashi. N.. Yoshida.
N.. and Sugano. N. (1983) Anal. Biochem. 129,
I92-199.

Nerrill. C. R.. Goldman. D.. and Van Keuren. N. L.
(1982) Electrophoresis 8, 17-23.

Duggleby. R. G. (1981) Anal. Biochem. 110, 9-18.

Bergmeyer. N. (1974) Nethods of Enzymatic Analysis,
pp. 488-489. Academic Press. New York.

Bradford. N. (1976) Anal. Biochem. 18, 248-254.

Pollet, R. J.. Baase. B. A.. and Standaert. N. L.
(1979) J. Biol. Chem. 254, 30-33.

Norman. A. W.. Wedding. R. T.. and Black. N. R.
(1965) Biochem. Biophys. Res. Commun. 88, 703-709.

Nourad. N. and Parks. R. E. (1966) J. Biol. Chem.
244, 3838-3844.

Colomb. N. G.. Cheruy. A.. and Vignais. P. V. (1974)
Biochemistry 18, 2269-2277.

Robinson. J. B.. Brems. D. N.. and Stellwagen. E.
(1981) J. Biol. Chem. 256, 10769-10773.

Parks. R. and Agarwal. R. (1973) The Enzymes (P. O.
Boyer. ed.) Vol. 8. pp. 307-334. Academic Press. New
York.

Cheng. Y. C.. Robison. B.. and Parks. R. E. (1973)
Biochemistry 18, 5-11.

Rarr. T. L.. White. B. D.. and Purich. D. L. (1979)
J. Biol. Chem. 254, 6107-6111.

Narimichi. N. ans Johnson. G. S. (1983) J. Biol.
Chem. 258, 12609-12617.

Berson. G. (1983) Anal. Biochem. 134, 230-234.

CHAPTER IV

THE ASSOCIATION WITH NICROTUBULES

OF A

NUCLEOSIDE DIPHOSHATE RINASE

HAVING

TRANSPHOSPHORYLASE ACTIVITY

107

108

INTRODUCTION

Nucleoside diphosphate kinase (NDP kinase) activity
has been found in preparations of brain microtubule
protein isolated by cycles of assembly-dissassembly
(1-3). It has been shown that this NDP kinase is not of
mitochondrial origin. not released from organelles by
homogenization. and. therefore. it is available for
interaction with microtubules 1p_y1yg_(4). NDP kinase
has also been found in proximity to the microtubules of
flagella in sea urchin sperm (5) and Chlamydomonas (6)
and to the microtubules of Tetrahymena cilia (7).

The tubulin dimer contains two distinguishable
guanine nucleotide binding sites (8-11). At one of these
sites called the exchangeable site (E-site) the
nucleotide must usually be GTP in order to support
microtubule assembly and this GTP is hydrolyzed to GDP as
assembly proceeds (12-14). There are at least two
mechanisms by which E-site GTP may be regenerated so that
the dimer may participate in further assembly reactions:

transphosphorylation and nucleotide exchange.

109

Exchange:

Tubulin-GDP + '6”: Tubulin-‘GTP + one

Transphosphorylation:

Tubulin-GDP + 'GTP:Tubulin—GTP + *GDP

It has been proposed that NDP kinase may catalyze the
proposed transphosphorylation reaction by which GDP bound
to the tubulin E-site may be phosphorylated without ever
leaving that site (1.2). If this is true. then NDP
kinase might play an important regulatory role in the
regulation of microtubule assembly and function. The
work of Penningroth and Kirschner (1) showed that when
microtubule protein was incubated with [3N] GTP and
[gamma-32P] GTP under conditions that would support
microtubule assembly. 32P labelled nucleotide appeared
on the tubulin dimer almost 10 times faster than did 3N
labeled nucleotide. Thus. they concluded that
transphosphorylation is occuring 10 times faster than
simple nucleotide exchange. A similiar experiment and
result has been reported by Jacobs and Buitorel (2) but
the validity of these experiments has been challenged by
Brylawski and Caplow (15) who have suggested that the
results are artifactually caused by the presence of pools

of free nucleotide that are difficult to remove.

110

As a result of these studies we believe that there
are important questions to be answered about NDP kinase
and microtubules. We want to know whether there is
really a transphorylation mechanism for the generation of
E-site GTP and whether that reaction is catalyzed by NDP
kinase? Next. we want to know whether NDP kinase binds
to microtubules and to microtubule protein oligomers both
_8_vitro and in the cell.

In recent studies (16) we have isolated and
characterized the NDP kinase from bovine brain
microtubule protein. Nost of our findings about this NDP
kinase have been confirmed for the porcine brain
microtubule NDP kinase by the subsequent work of Nuitorel
88_81, (17). In this paper we will begin answering
questions about the role of NDP kinase in microtubule
function and about the possible binding of NDP kinase to
microtubules and microtubule protein oligomers and we

will be discussing the development of techniques and

strategies that will permit their ultimate resolution

111

EXPERINENTAL PROCEDURES

Naterials - Rabbit muscle pyruvate kinase. hog muscle
lactate dehydrogenase. phosphoenolpyruvate. and ferritin
were purchased from Boeringer Nannheim. Bovine serum
[14C] albumin (BSAl) and NEN-963 liquid scintillation
mixture were New England Nuclear products. SOS-PAGE
standards were obtained from Bio-Rad and Boehringer
Nannheim. All other chemicals for electrophoresis were
obtained from Bio-Rad except for Serva blue stains which
were a product of Serva Fine Biochemicals. Blue Dextran
2000 and all chromatography column components were
Pharmacia products. except for Fractogel TSN-55
(superfine) which was purchased from Pierce. Nost other
reagents were purchased from the Sigma Chemical Company.
[gamma-329] ATP was synthesized by the method of Glynn
and Chappel (18) as modified by Walsh 81_81, (19).
Rabbit antiserum to tubulin and actin was purchased from
Niles-Veda. FITC-goat antirabbit IgG was obtained from
Niles-Yeda and from Boeringer-Nannheim. Norse radish
peroxidase coupled goat anti-rabbit IgG and peroxidase
color development reagent were products of Bio-Rad and
rabbit muscle actin was from Sigma. All solutions for

tissue culture were obtained from Gibco.

112

Preparatiop:of Nicrotubule Protein - Nicrotubule protein
was prepared as previously described (16) except that the
first assembly reaction was initiated by the addition of

2 mN ATP and 0.2 mN GTP instead of with 1 mN GTP.

Isolptiop of N8P kinase from Bovine Brain - NDP kinase
was isolated from bovine brain cytosol by the method of
Nickerson and Wells (16). Previous work has shown that
this is the same NDP kinase present in bovine brain
microtubule protein preparations (16).

The NDP kinase from the particulate fraction of
bovine brain was isolated by the method of Robinson EL.
81, (20) as modified in this laboratory. Bovine brain
(1000 g) was homogenized in 1 mI/gram of tissue of 0.1 N
PIPES. pH 6.62. 2 mN Dithiothreitol with a Tekmar
Tissuemizer. This and all subsequent operations were
done at 4°C. The homogenate was centrifuged at 12.000
rpm in a Sorvall GSA rotor for 1 hour. The pellets were
extracted with 2 liters of 40 mN Tris ECI. pH 7.5. 10 x
(v/v) ethanol. 2 s (w/v) BRIJ 56. 5 mN NgSO4. and 1 .u
EDTA. The pellets were resuspended in this buffer with

the Tekmar Tissuemizer. The resuspended pellets were

centrifuged as before for 2 hours. The supernatant
fraction was mixed with 40 ml of swollen Blue Sepharose
6B and this mixture was gently agitated for 18 hours

before filtering out the Blue Sepharose with a coarse

113
scintered glass funnel. The Blue Sepharose was poured
into a column which was then washed with 1500 ml of 10 mN

7’1“ 3C1: 9” 7'5' 5 '" NgClz before the enzyme was

eluted with 250 mL of the same buffer containing 0.2 mN
GTP. The eluate was concentrated to 7 ml with a Whatman
CX-3O ultrafilter. Portions of this preparation (2 ml)
were chromatographed on a 3.2 x 38 on column of Fractogel
TSK-55s in the same buffer. The NDP kinase containing
fractions were concentrated with a CX-3O ultrafilter and

frozen in aliquots at -800C.

Assays - Protein concentrations were determined by the

method of Lowry et al. (21) except during the isolation
of brain cytosol NDP kinase when the method of Bradford
(22) was used. because the polybuffers employed in the

procedure interfered with the former assay. NDP kinase
was assayed by coupled enzyme assay as previously

described (16).

Eleptrophoresis Technigues - The identity and purity of
protein preparations was determined by SDS-PAGEI using
the method of Laemmli 81_81, (23) except that in some
experiments the gels used were 0.75 mm 9-168 (w/v)
acrylamide gels with an acrylamide to bis-acrylamide
ratio of 19. Gels were fixed in 15 2 (w/v)
trichloroacetic acid. washed twice with water. stained

with 0.1 s (w/v) Serve Blue W. and destained in water.

114

Western blots were prepared by electrobotting samples
from SDS-PAGE gels onto nitrocellulose membranes with a
BioRad Transblot apparatus operated at 60 volts for 4
hours at 10°C. Electroblotting was performed in 25 mN
Tris. 192 mN glycine. 10 8 (v/v) methanol. and 0.01 x
(w/v) sodium dodecyl sulfate. Gels were prepared for
autoradiography by drying them onto Whatman 3NN paper at

high temperature and under vacumn.

Immunochemipal Tephnigues - Antibodies were raised in New
Zealand white female rabbits. After preliminary
bleedings to obtain pre-immune serum. rabbits were
immunized with 500 ug of NDP kinase in Freunds complete
adjuvant (Sigma). Injections were repeated at 10 day
intervals with 500 ug of protein in Freunds incomplete
adjuvant (Calbiochem) until antibody was detected.
Thereafter. 150 ug injections in Freunds incomplete
adjuvant were made at convenient intervals. Animals were
bled from an ear vein 4 to 7 days after each injection.
Serum was stored at -8OOC.

The presence of antibody was detected by the method
of Ouchterlony (24) except that the gels were 1 s agarose
containing 25 mN Tricine. pH 8.6. Ouchterlony plates
were stained by washing for several days against many
changes of phosphate buffered saline. staining with 0.1 x

Serva Blue W. and destaining with water.

115

The IgG fraction of rabbit serum was isolated on a
3 ml column of Protein A Sepharose. The column was
equilibrated with 0.1 N NOPS. pH 7.0. Serum was passed
through the column and it was washed with 20 ml of
equilibration buffer. IgG was eluted with 0.1 N glycine.
pH 3.1. Fractions of 1 ml were collected and immediatly
neutralized by the addition of 75 ul of 1 N Tris base.
The fractions containing eluted protein were pooled.
dialyzed against 1 liter of phosphate buffered saline and
concentrated with a Nillipore CX-30 ultrafilter.

Aliquots were frozen at -800C.

NDP kinase was detected on nitrocellulose after
electroblotting with the use of anti- NDP kinase IgG and
the procedure of Towbin et al. (25). All steps were
performed in Tris buffered saline (TBS: 20 mN Tris. pH
7.5. 500 mN NaCl). the blots were blocked in 3 3 (w/v)
gelatin in T83. and the antibody incubations were done in

1 3 (w/v) gelatin in TBS.

88cleotige Affinity Labelegrof Tubulin - The tubulin
E-site was affinity labeled with periodate oxidized GDP
by the method of Naccioni and Seeds (26). Periodate
oxidized GTP was synthesized by a modification of the
Naccioni and Seeds (26) method. One micromole of GTP.
sometimes containing [3B] GTP. was brought to a final
volume of 180 uL in water to which was added 20 ul of

50 mN sodium periodate (pH 5.5-6). The reaction was

116

allowed to proceed at 40c for so minutes when 10 ul
more of periodate was added for a further 30 minute
incubation. Excess periodate was removed by adding 10 ul
of 50 mN glycerol. The product. oxGTP. was isolated by a
modification of the method of Palmer and Avruch (27).
The reaction products were loaded on a 3 ml column of
DEAE Sepharose which had been equilibrated in saturated
sodium bicarbonate and then washed with 30 ml of water.
oxGTP was eluted with a 0.1- 1.5 N linear gradient of
triethylamine carbonate buffer. A 2 N stock solution of
this buffer was made by bubbling 002 through 28 x (v/v)
triethylamine until the pH decreased to 7.5. Nucleotides
were detected by absorbancy at 260 nm. Peaks were
pooled. frozen. and lyophylized. The purity of oxGTP was
checked by thin layer chromatography on PEI-cellulose
plates developed in 1.2 N LiCl.

Nicrotubule protein (2 cycle) was treated with Norit
A charcoal (0.1 mg per mg of protein) which had been
extensively washed with 2 x (w/v) bovine serum albumin
and then with water. The microtubule protein was
incubated at 30°C for 20 minutes and centrifuged for 15
minutes at 175,000 x g in the Beckman Airfuge. The
microtubule pellets were resuspended in 10 IN Tris BCl.
pH 6.8. 20 mN sodium borohydride and incubated for 30
minutes on ice before dialysis against 2x1 liter changes

of 0.05 N PIPES. pH 6.82. Dialyzed protein was frozen at

117
-800C in aliquots. This microtubule protein. because
it has been through an assembly-disassembly cycle. is

affinity labeled with oxGDP.

Tissue Culture and Immunofluorescent Staining - 3T3 mouse

fibroblasts were obtained from Dr. John Wang of this
department. They were cultured in Eagles Nodified
Essential Nedium containing 100 ug/ml streptomycin. 100
U/ml penicillin. and 10 3 (v/v) newborn calf serum at
37°C in a water saturated atmosphere of 10 3 (v/v)
002. Cells were passed by treatment with 0.25 3 (w/v)
trypsin in phosphate buffered saline (PBS) at 35°C for
3 minutes and then with 0.025 8 (w/v) EGTAI in pas at
35°C for 20 minutes. For immunochemical localization
studies. cells were grown on glass coverslips which had
been washed in 100 8 ethanol and then autoclaved.

Nethods for immunofluorescent staining were adapted
from the the procedures of Weber and Osborn (28). Cells
grown on coverslips were washed for 10 seconds in PBS at
37°C and then fixed in pure methanol at -10°C for 6
minutes. The cells were rehydrated by immersion in PBS
at room temperature for 30 seconds and were then ready
for staining.

The coverslips were placed in a humidity chamber for
incubations with antibody. They were coated with an
appropriate dilution of rabbit anti-NDP kinase.

anti-tubulin. or anti-actin or with the same dilution of

118
preimmune serum and incubated for 30-60 minutes. The
coverslips were rinsed in 4-6 changes of PBS and covered
with an appropriate dilution of FITC conjugated goat
anti-rabbit IgG and incubated for 30-45 minutes. After a
further 4-6 washes in PBS the coverslips were mounted on
slides in a drop of mounting medium containing 10 3 (v/v)
0.1 W sodium phosphate. pH 8.5. and 90 8 (v/v) glycerol.
They were then viewed and photographed with a Leitz
Orthoplan fluorescence microscope with optics optimized
for FITC fluorescence. Appropriate dilutions of all
antibodies were chosen empirically and varied

considerably from batch to batch of the same antibody.

119

RESULTS

The Assopiation of N8P Kinase wit8 Oligomeric Nicrotubule

Proteip R1pg8_-Previous studies have suggested that most
of the NDP kinase from microtubule protein preparations
binds to multimeric microtubule protein rings under
non-assembling conditions (2.29). This is consistent
with a role for NDP kinase in transphosphorylation since
ring E-site nucleotide is not readily exchangeable and
might require transphosphorylation to regenerate GTP
(2.30). This suggested that the association of NDP
kinase with rings might be an important phenomena. so we
attempted to confirm that NDP kinase was associated with
rings. Oligomeric rings were separated from
non-aggregated protein by chromatography on columns of
Sepharose CL-8B. As shown in Figure 18. only 16 N of the
NDP kinase eluted with oligomeric rings in the void
volume. Previous studies. which have shown a stronger
association of NDP kinase with rings. had been done with
high concentrations of glycerol in the buffers and done
with microtubule protein that had been isolated in the
presence of 6 N glycerol. When our experiment was
repeated in the same buffer but this time containing 6 N
glycerol. a much higher percentage of NDP kinase
associated with rings (Figure 10). The addition of 8 N
glycerol to isolated NDP kinase did not cause isolated

NDP kinase to aggregate and elute in the void volume

120

Figure 1. Binding of NDP kip8se to oligomerip
migrotubule protein rings. One of two identical

pellets of microtubule protein was resuspended in 50

 

mN NES. pN6.6. 1 mN NgClz. 1 mN EGTA. 1 IN
dithiothreitol and the second was resuspended in 2.5
ml of the same buffer containing 4 N glycerol. The
resuspended pellets were centrifuged at 18.000 rpm at
4°C in a Sorvall SS-34 rotor. Two 1 ml samples

(4.4 mg/ml. 50 m0 NDP kinase/mg) of the resuspended
microtubule protein were chromatographed on similiar
2.8 x 17 cm columns of Sepharose 6B preequilibrated
in the same buffer (part B) or in the same buffer at
pH 6.2 (Part A). A 1 ml sample of 4 N glycerol
resuspended microtubule protein (5.0 mg/ml. 55 mU NDP
kinase/mg) was chromatographed on a similiar column
that had been equilibrated in the same buffer
containing 4 N glycerol (part C). Chromatography was
performed at 4°C at a flow rate of 36 ml/hour and 3
ml fractions were collected. assayed for absorbancy
at 280 nm. and assayed for NDP kinase activity. The
column was calibrated with the following standards:
Vo. Blue Dextran 2000 (2.000 kilodaltons):
thyroglobulin (670 kdaltons); ferritin (480
kdaltons); gamma-globulin (158 kdaltons); ovalbumin

(44 kdaltons); myoglobin (17 kdaltons); and vitamin

B12 (1.3 kdaltons).

121

Figure 1A

 

 

E

 

 

 

 

:E\DvaaoZ all-
m 5 0
d - a — Ky
Iv
2
6
H O
\\\\ ma
\
wunw nrlln.
/ 5
T 0/0 2
I“ /.
iim\m\‘.m O
I. .IIi/‘I \.\ l 2
O I.
1% Tm/
in
6
IV0

FRACTION NUMBERI3m|I

122

Figure 18

 

 

 

 

 

 

 

:E\Dvaaoz e s
.0: 5 O
_ _ _ _ w
6.
6 e\
H e\ O
D. e\ 3
Is
only /e/e/
I “As "a
I“. \I\
.\\-
8/‘I I”
I” III
F
I. w
[m O\§$m~
So e“\.e.\\\\-\IA\-\ My
TV in
5
B
_ b _ _
2. J o
O 0
Some mozqmmommc. .II.

FRACTION NUMBER(3mII

123

Figure 1C

 

.0
m

0—0 ABSORBANCE 280 nm
0
on

 

O

I I T I I II
\g 570 ISO 44 |.3
1.

./\ pH6.6, 4M Glycerol

,sa

“I0

f \\ H _
\3 7: X!’ 5

IO i5 20 25 so 35
FRACTION NUMBER(3mlI

 

L
h)
0

I NDPK (mU/mII

 

124
(data not shown) and so we conclude that glycerol
promotes the association of NDP kinase with rings.

As shown in Figure 1A decreasing the pH to 6.2
decreases the amount of NDP kinase which is ring
associated by 50 x. By repeating a gel filtration step
twice at pH 6.2. it is possible to almost eliminate NDP
kinase activity from oligomeric rings. This may be a
convenient technique for removing NDP kinase from

assembly-competent microtubule protein.

The Assopiatiop of,§8P Kip88e with Nicrotp881es Ip Vitro
In previous studies (31) we found that in the presence of
6 N glycerol NDP kinase persisted in microtubule protein
preparations through 5 cycles of assembly-disassembly.
remaining at a constant specific activity through the
last 2 cycles. We proposed that NDP kinase might be a
microtubule associated protein because of this behavior.
Nore recently Burns and Islam (29) have gel-filtered
assembled microtubules on Sepharose 2B columns and shown
that microtubules can be separated from NDP kinase
activity under these conditions. As a result they have
concluded that NDP kinase is not microtubule associated.
that it is merely trapped in the microtubule pellets
during microtubule protein isolation. To resolve this'
conflict we have done two types of experiments.

First. we isolated microtubule protein by the method

of Sheterline (32) in which microtubules are pelleted

125

through a cushion of 30 x sucrose at 30°C and 40,000 x
g for 60 minutes. This reduces the possibility that NDP
kinase is merely being trapped in microtubule pellets.
The NDP kinase specific activity of microtubule protein
prepared in this way was 0.15 units/mg of protein
compared with 0.05 units/mg for 2 cycle microtubule
protein prepared as described under "Experimental
Procedures". supporting the view that NDP kinase binds to
microtubules and remains bound as they pellet through
sucrose.

In other experiments we followed the time course of
NDP kinase incorporation into microtubules (Figure 2).
At various times during 18_vitro assembly at 30° or
35°C. microtubules were pelleted by 3 minute
centrifugations at 175.000 x g in a Beckman Airfuge at
room temperature. The amount of protein and NDP kinase
activity was determined in the resuspended pellets for
each time point. As can be seen in Figure 2. microtubule
assembly reached a steady state level by 6 minutes at
35°C and by 8 minutes at 30°C. NDP kinase appeared
in the pellets more rapidly; the ratio of NDP kinase
activity to the amount of protein in pelleted
microtubules decreased after 1 minute. This demonstrated
that the incorporation of NDP kinase into microtubules'
occured early in assembly. This was especially obvious
when assembly was slowed down by doing the experiment at

30°C (Figure 28). The higher NDP kinase activity early

126

Figure 2. The associption of N88,kinase with

microtubules during assembly. Samples of microubule

protein were incubated with 1 mN GTP in an airfuge
rotor that was kept at 30° or 35°C in a water

bath. Cold (4°C) microtubule protein was added to
GTP immediatly before the sample was warned in the
rotor. At the end of the chosen assembly interval.
the rotor was centrifuged at 195,000 x g for 3
minutes at room temperature in the Beckman Airfuge.
Supernatants were removed with a syringe and the
pellets were resuspended for assay of protein or NDP
kinase activity. In the 35°C experiment 110 ul
samples assembled contained 5.1 mg/ml microtubule
protein and 57 mU NDP kinase/mg). while in the 30°C
experiment 150 ul samples contained 3.0 mg/ml

microtubule protein with 46 m0 NDP kinase/mg.

127

Figure 2A

 

«east-:5 xaoz 3.3.“. s

0.
3 2 I. O
_

 

 

A.35°C

 

 

 

«3333 522a .ozoa oils

 

TIME (min)

128

Figure 23

 

322355.22 3:8 .
5

 

 

5 c
_

LT _ _ _ _ _

 

 

l O. 0

l0

 

  

 

_
nu
nu

«23:33 5305 3:9... oils

50-
0

TIME (min)

129

in assembly suggests that NDP kinase may have a role in
initiation and early elongation events during assembly.

The results shown in Figure 2 suggest that NDP kinase
is binding to microtubules. but there are still two
alternative explanations of the data which must be
eliminated in order to validate that conclusion. The
first possibility is that NDP kinase is aggregating and
pelleting independently of microtubule formation. The
second possibility is that NDP kinase is merely being
trapped in the microtubule pellets without actually
binding to microtubules. The experiment summarized in
Table I was performed in order to examine these two
possibilities.

Nicrotubules were assembled at 35°C with 2 mN GTP
for 0 to 10 minutes after which they were pelleted. as
before. by 3 minute centrifugations at 175,000 x g in the
Beckman Airfuge at room temperature. After 10 minutes of
assembly. 228 ug of protein and 1.8 m0 of NDP kinase
activity were pelleted. When 83 uN colchicine was
included in the assembly buffer. no protein and no NDP
kinase activity could be pelleted. If NDP kinase was
aggregating and pelleting independently of microtubules.
then we would not expect colchicine. which prevents
microtubule assembly. to affect the amount of pelleted.
NDP kinase. From this experiment we conclude that the
increase in pellet NDP kinase activity which parallels

microtubule assembly is dependent upon microtubule

130

.eHnEew egu cu ucsoee gave» one no eoeuceouen

e we wusunen ecu c“ >0w>uu0e no casuohn no acoOIe us: as» one wesue> omega .0
.wxceun as» u°u ewosu cezu sendeEw one see» wesae> an newseo sue wesae> e>uueuez .n
\ecauuzoaoo unocuuz waves“: 0 uOu essensocu

weanEew any you yes» e>one >uu>uu°e so cueuOLQ as: ecu ucewehaeu wesae> ~a< .e

 

 

 

 

 

 

 

 

 

o o.o fl o.A n.o a o.o scuohevoo :5 no ea

m.o « m.~ ~.H H ~.o o.o « A.mv on

o n.o s o.o n.o « n.o o
huw>«0060«fl0h (mm CNOwa moz CAOUOHQ COuUufifl‘ «Cufiv OE“?
¢O«UQn:OCH

UOUOHAOG Cw Man>u00¢ HO CuOUOul A6008 UO #

o.mn « o.n o~.o « n~.o n.~ « n.v scwofinoaoo to no on

wood m.v~ « .men an.o s n~.~ A.v « ~.m- cu

«mod ”.5 a o no.0 « e~.o o.~ « A.” o
(ma—ova. sue:

«on i noe.o h oH.oi no.4 a n.oi ochoweouoo 2: no oa

i no.o « no.o no.0 « n.nei oceohzoeoo :5 no 0

new i no.0 « no.e 0.8 « n.o- on

yea i oo.o « o ~.A « o o
NHO>OOOK (mm END OIGCux £02 DE C«.u0um mm ¢O«u«0fi< Rafiev OEwF
seen“: 402 c0«ue0:0:~

 

ewueaaem :« auu>uuot no sueOOLm uez

.wcouuecAEueuev eue0wunuuu uOu wcouuea>ev vuevceuw i\+ aces! ecu sue
_ ~E\«U: m~.o mec«eucoo weaneew

C303“ n.3dfl> .55

.50”

e be: meanEew ssh

.Acu690Ae Eouew ena>onv
.ewecflx an: no HE\= Nv~.o one AE\OS o.m «0 amdueuuceocou cueuOuo saonouowuue
.UOmn ue ~ eusmuu Cu menuhowev we snow we: assauuenxe swab

(mm—U

weasnsuOLo«z cu“: seesax an: no couueuceeuoewou

H fldfl<fi

131
formation.

In the second part of the experiment (Table 1B) we
included 1.25 ug / ml [140] bovine serum albumin in the
assembly mixture. Since bovine serum albumin does not
bind to microtubules. the amount of [140] serum albumin
that could be pelleted gave us a quantitative estimate of
the amount of soluble protein trapped in the microtubule
pellets. If there was any association of serum albumin
with microtubules this value would be an overestimate. so
that the percent of serum albumin found in the pellets is
a maximum estimate or upper limit to the percent of
soluble protein that might be trapped in the pellets
without binding. As shown in Table 1B. after 10 minutes
of assembly. 2.5 x of the serum albumin was recovered in
the microtubule pellets. If NDP kinase did not bind to
microtubules we would expect that 2.5 x or less of the
NDP kinase in the samples would be recovered in the
pellets. In fact 9.2 s of the sample NDP kinase was
found in the pellets. We believe. therefore. that NDP

kinase is binding to and pelleting with microtubules.

The Association of NDP Kinase with Nicrotubules in 3T3

Cells - In experiments such as those shown above we and

 

others have sought to determine whether NDP kinase binds
to microtubules and to other microtubule protein
structures 1p_vitro, but what we really want to know is

whether NDP kinase binds to microtubules in the cell. We

132
decided to attempt indirect immunoflourescent staining of
cells for NDP kinase. This required that we produce an
antibody against NDP kinase. Our attempts to produce an
antiserum against bovine brain cytosol NDP kinase
isolated by the method of Nickerson and Wells (16) were
unsucessfull so we isolated the particulate fraction NDP
kinase from bovine brain as discussed under ”Experimental
Procedures" and were able to produce rabbit anti-serum to
this enzyme which is distinct from but similiar to the
microtubule NDP kinase (16).

As shown in Figure 3. SDS-PAGE examination of the
isolated NDP kinase antigen showed it to be a homogeneous
doublet of bands with molecular weights of about 18.000.
Robinson 88_81, (20) using a very similiar isolation
procedure detected only one type of subunit in
particulate NDP kinase. but they were using a SDS-PAGE
system with inadequate resolving power to see the
doublet. Our gels were 9-16 8 linear gradient gels with
an acrylamide to bis-acrylamide ratio of 19 used with the
Laemmli (23) system of buffers. We have previously
reported (16) that the NDP kinases from bovine brain
microtubule protein and from bovine brain cytosol
examined with this same SOS-PAGE system also contained
two different subunits with molecular weights of about
18.000.

Rabbit anti-NDP kinase serum was used for the

isolation of rabbit anti-NDP kinase IgG by chromatography

133

Figure 3. 88$-PA88,of,8ovine brain partigulate NDP

kinase. Samples from each step in the isolation

 

procedures were electrophoresed on 9-16 x gradient
gels as described under "Experimental Procedures".
Lane A is the bovine brain first supernatant. Lane 9
is the Brij 56 extract from the pellets of the first
centrifugation step. Lane C is the NDP kinase eluted
from Blue Sepharose. Lane D is the Blue Sepharose
NDP kinase containing fractions concentrated with a
CX-3O ultrafilter. Lane E is NDP kinase after
chromatography on TSK 55s. This is the NDP kinase
which was used as an antigen for antibody

production. NDP kinase consists of a doublet of

bands with molecular weights of about 18.000 daltons.

134

Figure 3

 

135
on Protein A Sepharose. As shown in Figure 4A. anti-NDP
kinase activity was present in both serum and isolated
IgG. but it was absent in preimmune serum.

The antibody had been raised against a form of bovine
brain NDP kinase which is unlikely to associate with
microtubules since it is membrane associated (20). As
shown in Figure 4B. the antibody which had been raised to
bovine brain particulate NDP kinase cross-reacted with
the bovine brain microtubule NDP kinase isolated from
cytosol by the method of Nickerson and Wells (16). Thus.
we would expect the antibody to recognize the form of NDP
kinase that may associate with microtubules.

We wanted to use an antibody raised against bovine
brain NDP kinase to stain mouse fibroblasts. so it was
important to see whether our IgG would cross-react with
NDP kinase from these cells. A flask of 3T3 cells was
homogenized and 150 ug of the homogenate was loaded on a
gel for SDS-PAGE. After electrophoresis. the separated
proteins were blotted onto nitrocellulose and the blot
was probed with anti-NDP kinase IgG as described in
”Experimental Procedures". As shown in Figure 4C. only a
band of about 18.000 daltons is recognized by our
antibody. suggesting that. in 3T3 cells. our antibody is
monospecific for NDP kinase. It is important to note
that with this SDS-PAGE system and blotting protocol the
ability to resolve the NDP kinase doublet would be lost.

so we do not know whether 3T3 cell NDP kinase has one or

136

Figure 4. Character1888iop,of,8pti-N8P kinase IgG.
A. On this Ouchterlony plate the center well (C)
contained 20 ul of NDP kinase (1.5 mg/ml). Well 1
contained 20 ul of preimmune serum. Well 2 contained
30 ul anti-NDP kinase serum. Well 3 contained 20 ul
of anti-NDP kinase IgG (0.6 mg/ml). Well 4 contained
an affinity labeled anti-NDP kinase which was not
used in this study because it was only produced with
very small yields. This plate shows that the
anti-NDP kinase antibody was present in both
immunized rabbit serum and in the IgG isolated from

that serum.

 

137

Figure 4A

 

138

Figure 4. 88888cter188tion ofvpnti-NDP kip8gei1g8,
B. This immunoblot was stained with anti-NDP kinase
IgG. Samples were prepared for electrophoresis by
treatment with 20 mN iodoacetamide and then
denaturation in sodium dodecyl sulfate. Samples to
be electrophoresed on Gel A were supplemented to
100 mN with -mercaptoethanol before electrophoresis
while the samples for gel B were not reduced. After
electrophoresis. the protein bands were blotted onto
nitrocellulose and stained with anti-NDP kinase IgG
as described under "Experimental Procedures”.
Lane 1. Nicrotubule NDP kinase isolated from
bovine brain cytosol as described by Nickerson
and Wells (16).
Lane 2. NDP kinase isolated form the
particulate fraction of bovine brain as
described under "Experimental Procedures”. This
is the antigen that was used to produce the

anti-NDP kinase IgG.

139

Figure 4B

TOP-
92.5-
66.2-

45"

31-

140

 

Figure 4. Characterization of guti-NQP kippse Ig8,
C. This immunoblot was prepared and stained with
anti-NDP kinase. Lane 1 contained 2 ug of the

antigen. bovine brain NDP kinase: lane 2 contained 2

 

ug of actin; lane 3 contained 2 ug of microtubule l
protein: lane 4 contained 100 ug of 3T3 cell

homogenate. From this experiment we can conclude

that our anti-NDP kinase IgG was. in 3T3 cells.

monospecific for NDP kinase and we can conclude that

our antibody did not recognize either actin or the
cytoskeletal proteins in 2 cycle microtubule

protein. The standards for this gel were lysozyme

(14.4 kdaltons). soybean trypsin inhibitor (21.5).

ovalbumin (45). and bovine serum albumin (86.2).

 

141

Figure 40

1 2 3 4
"-66.2
“-45
-21.5

- 14.4

I.

In

J iiii‘.

142
two distinguishable subunits. The same experiment was
done with isolated microtubule protein and with purified
actin. As can be seen in Figure 4C. neither actin nor
any microtubule protein present in sufficient quantity
can be recognized by our anti- NDP kinase IgG. If 3T3
cell microtubules are stained by this antibody. we know
that it is NDP kinase that is being stained and not other
microtubule proteins.

The use of anti-NDP kinase for the staining of 3T3
cells is shown in Figure 5. Anti-NDP kinase serum and
IgG stain filamentous structures in the cell which appear
to radiate from the nucleus. The staining is very
dependent on antibody concentration. At low
concentrations. there is little staining. while at high
concentrations the background stains too intensely to see
filaments. An optimal concentration of antibody must be
determined for each batch of serum or IgG. The staining
pattern for NDP kinase is distinctly different from the
pattern of actin localization. so we can be sure that NDP
kinase is not associated with actin filaments. NDP
kinase localization is similiar to tubulin localization
in 3T3 cells except that there is less cytoplasmic
staining so that the filaments stand out more clearly
against the background. This may be caused by the
presence of a larger percentage of unpolymerized tubulin
in cytoplasm then the percentage of non-filamentous NDP

kinase.

 

143

Figure 5. 1pdirect immunofluorespent staining of 3T3

 

cells for §8P kinase and pytopkeletal proteins.

 

Indirect immunofluorescent staining was done as
described in "Experimental Procedures". The bar on
each picture represents 1 um.

A. In this picture the 3T3 cell was stained with
preimmune serum.

B. This 3T3 cell was stained with a 1:100 dilution
of rabbit anti-actin serum.

C. This shows the staining of a 3T3 cell with a
1:100 dilution of rabbit anti-tubulin serum.

0. This 3T3 cell was stained with a 1:50 dilution of
rabbit anti-NDP kinase serum.

E. This shows the staining of a 3T3 cell with 50

ug/ml anti-NDP kinase IgG.

ll' III!

ith

nun

144

Figure 5

 

 

145

When cytoskeletons were extracted by the method of
Osborn 88_81, (33) before protein fixing. microtubules
remained intact but all NDP kinase staining was lost. We
believe that the Triton x-1oo in the extraction buffer
breaks down the interaction of NDP kinase with
microtubules. Neither microtubules nor filamentous NDP
kinase could be detected in cells which had been
incubated in 100 uN colchicine in tissue culture medium
for 10-30 minutes before the cells were fixed (data not
shown).

These immunochemical localization studies suggest
that NDP kinase may be associated with microtubules in
3T3 cells but. as discussed below. this must be treated
as a preliminary result. We would like to improve our
localization techniques my microinjection studies with

fluoescently labeled NDP kinase.

Transphosphorylation Catalyzed by NDP Rinase - As had
been reported by Naccioni and Seeds (26) tubulin could be

affinity labeled with oxGTP. This tubulin was capable of
undergoing a cycle of assembly-disassembly after which
the covalently attached E-site nucleotide should be
oxGDP. We were able to achieve a stoichiometry of about
0.4 moles of oxGDP incorporated per mole of tubulin.

This is substantially below the theoretical maximum of 1

mole per mole of tubulin. but it was good enough for use

146

Figure 6. Transphosphorylgtippcof oxGDP labeled tubulin
Samples of microtubule protein or oxGDP affinity
labeled microtubule protein prepared as described
under ”Experimental Procedures" were brought to 20 mN
PIPES. pH6.6. 0.8 mN NgCl2 and 75 ug/ml of bovine
brain NDP kinase was added. [gamma-32P] ATP was
added to a final concentration of 1 uN and the sample
was incubated for 3 minutes at 30°C. The
incubation was terminated by the addition of SOS-PAGE
sample buffer and the sample was electrophoresed with
BioRad low molecular weight standards. An
autoradiographs was prepared from the dried gel from
which a lane of standards had been cut and stained.
Lane A shows the phosphorylation of microtubule
protein. The only phosphorylated band detected was
the high molecular weight protein NAP2 which is a
substrate for the endogenous protein kinase. Lane B
shows the phosphorylation of microtubule protein
which had been affinity labeled with oxGDP.

Phosphate is incorporated into the tubulin bands for
this sample even though. as seen in Lane A. tubulin
is not a substrate for the protein kinase. We
suggest that the phosphate is being incorporated into

the covalently bound nucleotide on tubulin.

147

Figure 6

 

148
in transphosphorylation experiments performed as
discussed below.

We incubated microtubule protein or affinity labeled
microtubule protein with [gamma-32P] ATP and bovine
brain NDP kinase for 3 minutes. after which the sample
was treated with SDS-PAGE sample preparation buffer and
then electrophoresed in our SDS-PAGE system as discussed
in "Experimental Procedures". After electrophoresis the
gels were dried and autoradiographs were prepared. Any
nucleotide which was not covalently attached to the
tubulin molecule is released by treatment with SDS and
migrates slightly ahead of the dye front in our SDS-PAGE
system. When [38] oxGTP was used for the affinity
labeling of tubulin and the tubulin was not treated with
[gamma-3°P] ATP. the tubulin bands could be cut out of
the gel and counted for 3N to estimate the amount of
affinity label incorporated.

The microtubule protein preparations used in this
experiment contain a protein kinase activity which is
known (34-36) to phosphorylate the high molecular weight
microtubule associated protein NAP2. As shown in Figure
6. NAP2 is the only protein which incorporates labeled
phosphate in normal microtubule protein. When the
tubulin molecules in the same preparations are affinity
labeled with oxGDP. tubulin also incorporates labeled
phosphate. Since tubulin is. as we can see. not a

substrate for the endogenous protein kinase. we can

 

149

32P label is incorporated into the

conclude that the
covalently attached nucleotide. Since that nucleotide is
covalently attached. we know that it did not exchange
onto tubulin. but that it must have been phosphorylated
by ATP without ever leaving it's tubulin binding site.
This phosphorylation of tubulin bound nucleotide suggests
that a true transphosphorylation mechanism exists.
catalyzed by the NDP kinase added to this preparation.

We are currently using this experimental design to

explore the nature of the transphosphorylation

mechanism.

150

DISCUSSION

There are two mechanisms by which E-site GTP might be
generated to support microtubule assembly: nucleotide
exchange and the transphosphorylation of E-site GDP to

GTP without exchange.

Exchange:
s s
Tubulin-GDP + GTP Tubulin- GTP + GDP
Transphosphorylation:
I t
Tubulin-GDP + NTP Tubulin-GTP + NDP

A transphosphorylation mechanism could be driven by any
nucleoside triphosphate (NTP) as pictured above.

Under conditions of pH. ionic strength. and metal ion
concentration favoring assembly. ATP and every other
nucleoside triphosphate tried will induce microtubule
assembly (1.30.37.38). even though binding of these
nucleotides to tubulin cannot be detected. Jacobs and
his colleagues (38.39) were the first to suggest that a
transphorylase activity in microtubule protein
preparations transfers the gamma phosphate of NTP to a
tightly bound tubulin-GDP. leaving a GTP bound to the

tubulin E—site. This tubulin-GTP would then be capable

 

151
of participating in assembly reactions requiring GTP.

It has been suggested that the proposed
transphosphorylation reaction might be catalyzed by a
nucleoside diphosphate kinase (1.2).

Penningroth and Rirshner (1) demonstrated that
microtubule protein preparations contain a true
transphorylase activity phosphorylating tubulin E-site
bound GDP 18_8188, They incubated tubulin containing

32

bound GDP with a mixture of [3B] GTP and [gamma- P]

UTP under conditions that would support assembly. They

32

found that the P labeled nucleotide appeared on the

3n

tubulin dimer almost 10 times faster than the
labeled nucleotide and concluded that
transphosphorylation is occuring 10 times faster than
simple nucleotide exchange. The validity of this
experiment has been challenged by Brylawski and Caplow
(15) who have suggested that it is impossible to remove
free nucleotide from microtubule protein preparations and
so the results of the double label experiment are
complicated by isotope dilution. Jacobs and Nuitorel (2)
have performed a similiar double label experiment with
[3N] GDP and [gamma-32F] GTP. Tubulin was preloaded

with [38] GOP and then incubated with [gamma-°2P]
GTP. The nucleotide bound to the E-sites on dimers was'

32P labeled while the E—sites on 36s rings

32

almost all
was both 3E and P labeled. This experiment

suggests that E-site GTP on dimers is produced by

152

exchange while E-site GTP on rings is produced by
transphosphorylation. This result cannot be dismissed as
an artifact of isotope dilution. but. does not finally
settle the debate. NDP kinase is associated with the 35s
rings and for reasons of simple propinquity may
preferentially phosphorylate GDP being released from
nearby ring E-sitss.

In this paper we report an experimental protocol that

will permit the unambiguous study of

 

transphosphorylation. we have affinity labeled the
E-site of tubulin in assembly competent microtubule
protein with an analogue of GDP. If this GDP can be
phosphorylated to GTP it must be done by a
transphosphorylation mechanism since the nucleotide is
covalently attached.

We have taken advantage of the affinity labeling
procedure of Naccioni and Seeds (26). who have shown that
oxGTP affinity labels the E-site of tubulin specifically
and that the resulting complex is assembly competent. By
taking this tubulin through an assembly-disassembly cycle
we are able to generate a tubulin dimer affinity labeled
with oxGDP. As shown in Figure 6. microtubule protein
preparations contain a protein kinase that phosphorylates
principly NAP2. Tubulin could not be phosphorylated by’
[gamma-32F] ATP but oxGDP affinity labeled tubulin
could. From this experiment we conclude that the tubulin

protein is not being phosphorylated. since the endogenous

153
protein kinase is not using tubulin as a substrate. so
the phosphorylation of the tubulin is due to
phosphorylation of the affinity bound nucleotide. This
shows that a transphosphorylation mechanism exists since
the covalently bound nucleotide cannot be exchanging with
nucleotide in solution. Nore intensive studies using
this protocol should allow us to study the kinetics and
mechanism of this mechanism.

In our experiments (Figure 1). we found that a
maximum of 15 N of the NDP kinase in microtubule protein
preparations was associated with rings. but that glycerol
could increase this to 50 N. The association of NDP
kinase with rings is an important phenomena since Jacobs
and Ruitorel (2) have shown that transphosphorylation. if
it occurs. occurs on the rings.

Preparations of microtubule protein which are made by
cycles of assembly-dissasssmbly contain tubulin in two
forms. as 6s dimers and as ring-shaped oligomers of
30-36s (this is reviewed in 40). Oligomeric microtubule
protein rings also contain NAPs. especially the high
molecular weight NAPI and NAP2. and differ in structure
according to whether the microtubule protein was prepared
in the presence or in the absence of glycerol (40.41).
Nicrotubule protein prepared in the presence of glycerol
contains rings that are 36s double-walled structures
(40.41). while microtubule protein prepared in the

absence of glycerol contains 18s rings and 30s rings

154
which resemble two stacked 18s rings (40,43). These 18s
and 30s rings are interconvertable and exist in
equilibrium. Typically. 25-758 of the protein in a
microtubule protein preparation isolated by two cycles of
assembly exists in oligomeric rings. depending on
solution conditions and protein concentration. It may be
that NDP kinase associates with the 36s rings which are
present in the presence of glycerol. but not with the 30s
rings present without glycerol. There are two views that
we might take of the glycerol dependence of NDP
kinase-ring association. First. the high degree of ring
association may be artifactually caused by glycerol

concentrations that do not exist 1p_vivo. Second.

 

glycerol may mimic the very high protein concentrations
which exist 1p_y1yp, but which are not achievable in 1p_
81888_sxpsriments. We are inclined to accept the second
interpretation of these results. since if a
transphosphorylase activity actually exists and if it
acts on ring tubulin. the association of NDP kinase with
rings might be expected.

The tubulin of rings and the tubulin that exists as
free dimers are not quickly interconverted. nor are they
in a simple dimer-ring equilibrium (reviewed in 40); the
6s and 36s forms of tubulin did not interconvert even ‘
during a complete cycle of assembly-disassembly (44).
Other studies by Zeeberg 88_81, (45) and by Pantaloni 25.

81, (45) show that there is a very slow exchange of

 

155
subunits between 36s rings and 6s dimers. 36s rings are
direct cold induced disassembly products of microtubules;
they are not formed from microtubule released dimers
(45). but once rings and dimers are produced by cold
disassembly they exchange subunits with a half time of
about 90 minutes (46).

Studies of 1p_y18;p_microtubule assembly by
synchrotron radiation scattering have shown that
oligomeric rings are not nucleation centers for assembly;
rather. they are disassembled into protofilament
fragments which nucleate assembly (47-49). The tubulin
subunits which appear in microtubules during the early
stages of assembly come preferentially from 36s rings
(46). A transphosphorylation mechanism for the
generation of E-sits GTP would be more important for
tubulin in rings because. as we have seen. the nucleotide
on ring tubulin is not readily exchangeable and dimer and
ring forms of tubulin do not readily exchange.

In previous studies (16) conducted in the presence of
glycerol. we found that NDP kinase associated with
microtubules 1p_y1888, NDP kinase co-psrsisted with
microtubules through 5 cycles of assembly-disassembly and
co-sedimentsd with assembled microtubules through sucrose
gradients. we wanted to confirm this association in the
absence of glycerol and by a protocol that would be more
useful for further studies. To this end microtubules.

were assembled in 150 ul Airfuge tubes and pelleted at

166
various times during assembly by very rapid (1-3 minute)
centrifugations. This study was only made possible by
the very high g forces and small dimensions of the
Beckman Airfuge, which allowed us to freeze the action
very quickly at various times during assembly. As seen
in Iigure 2. the NDP kinase content of microtubules was
highest early in assembly and decreased during the later
elongation stage. This implies that if NDP kinase plays
a role in assembly, it does so during initiation or early
elongation events.

To rule out the possibility that NDP kinase was
aggregating and pelleting in a microtubule independent
manner we showed that colchicine prevented the pelleting
of both microtubule protein and NDP kinase (Table I).
Since NDP kinase might be trapped in the microtubule
pellets without binding to microtubules. we did an
experiment to obtain an upper limit estimate of the
amount of such trapping that might occur (Table I).
Nicrotubules were assembled and pelleted in the presence

C] bovine serum albumin. Only 2.5 8 of the serum
albumin was recovered in the microtubule pellets, while
9.2 X of the NDP kinase was pelleted. From this
experiment we would expect that 2.5 x or less of the
total NDP kinase would be trapped, so we conclude that‘

the rest of the pellet NDP kinase was microtubule bound.

157
The stoichiometry of NDP kinase binding to microtubules
was very low. It may be that we have not found optimal
conditions for binding.

le have proposed that NDP kinase catalyzes a
transphosphorylation reaction generating tubulin B-site
GTP. There is nothing about this proposal that requires
NDP kinase to bind to microtubules once they have been
assembled. Nevertheless, NDP kinase does bind to
microtubules ;g_vitro. It seemed important. therefore.
to determine whether NDP kinase might bind to
microtubules ig_glgg; This was done by indirect
immunofluorescent staining of 3T3 cells with rabbit
anti-NDP kinase IgG.

As seen in Figure 5. antibody against NDP kinase
decorates filaments in the cytoplasm of 3T3 cells. These
filaments originate near the nucleus. They are not
observed in cells that are pretreated with colchicine,
which suggests that they are microtubules or that they
depend on microtubules for structure. The antibody
against NDP kinase used in this study was specific for
NDP kinase (Figure 4). It did not recognize tubulin or
any other 3T3 cell protein. We conclude that NDP kinase
is associated with filamentous structures in 3T3 cells
that are most likely microtubules.

Immunofluorescent staining studies are not an ideal
way to localize proteins. It may be that fixation

procedures in organic solvents artifactually condenses

158
proteins into filamentous structures and it is probable
that fixation changes the availability of various
antigenic determinants for antibody binding. Despite
these difficulties. immunochemical localization may be
the best available technique for locating proteins in the
cell. We are contemplating microinjection studies with
fluorescently tagged NDP kinase to confirm that this
enzyme is microtubule associated in the cell. we believe
that the use of both techniques together solves the

problems of validity that each technique presents alone.

10.

11.

12.

13.

14.

15.

18.

159

REFERENCES

Penningroth, S. I. and Kirschner. I. I. (1977) J.
Hol. Biol. 115, 643-673.

Jacobs. H. and fluitorel. P. (1979) Eur. J. Biochem.
99, 613-622.

Terry, B. J. and Purich, D. L. (1979) J. Biol.
Chem. 254, 9469-9476.

Karr, T. L.. lhite, H. D.. and Purich. D. L. (1979)
J. Biol. Chem. 254, 6107-6111.

Yanagisawa. T.. Basegawa. 8.. and Nohri. H. (1968)
Exp. Cell. Res. ggJ 66-100.

Watanabe. T. and Flavin. I. (1976) J. Biol. Chew.
251, 162-192.

Kobayashi. T.. Kaji. R.. and Sakai. B. (1976) J.
Biochem. (Japan) 1g, 413-416.

Heisenberg, R. C.. Borisy, G. C.. and Taylor, B. I.
(1968) Biochemistry 14 4466-4479.

Berry. R. and Shelanski. I. L. (1972) J. Nol. Biol.
71, 71-60.

Jacobs. R.. Smith. B.. and Taylor. E. w. (1974)
J. lol. Biol. ggJ 455-468.

Jacobs, I. (1979) in Nicrotubules (Roberts. R. and
Byams, J. 8., eds.) Academic Press. New York. pp.
255-278.

David-Pfeuty, T.. Erickson. H. P.. and Pantaloni.
D. (1977) Proc. Natl. Acad. Sci.
U.S.A. 154 5372-6376.

Bonne. D. and Pantaloni. D. (1962) Biochemistry 2;,
1075—1081.

Carlier. I. -F. (1982) Molecular and Cellular
Biochemistry 514 97-113.

Brylawski. B. P. and Caplow. I. (1963) J. Biol.
Chem. 266, 760-763.

Nickerson, J. A. and Wells. W. W. (1964) J. Biol.
Chem. 269, 11297-11304.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

160

Buitorel. P.. Simon. C>. and Pantaloni. D. (1964)
Bur. J. Biochem. 144, 233-241.

Glynn. I. M. and Chappell. J. B. (1964) Biochem. J.
90, 147-149.

Walsh. D. A.. Perkins. J. P.. Brostrom. C. 0.. No.
B. S.. and Krebs. B. G. (1971) J. Biol. Chem. 246,
1966-1976.

Robinson. J. B.. Jr.. Brems. D. N.. and Stellwagen.
B. (1981) J. Biol. Chem. 256, 10769-10773.

Lowry. 0. B.. Rosebrough. N. J.. Farr. A. L.. and
Randall. R. J. (1951) J. Biol. Chem. 193, 266-275.

Bradford. N. (1976) Anal. Biochem. ng 246-254
Laemmli. D. K. (1970) Nature (London) 227, 660-665.

Ouchterlony. 0. (1956) Progress in Allergy. S.
Karger. Basel. vol.5, pp.1-76.

Towbin. B.. Staehlin. T.. and Gordon. J. (1979)
Proc. Natl. Acad. Sci. U.S.A. 19, 4350-

Naccioni. R. B. and Seeds. N. N. (1963)
Biochemistry 22, 1572-1679.

Palmer. J. L. and Avruch. J. (1981) Anal. Biochem.
116, 372-373.

Osborn. l. and Weber. K. (1962) Methods in Cell
Biology (L. Wilson. ed.). Academic Press. New York.
vol. 24. pp. 96-132.

Burns. G. and Islam.K. (1961) Eur. J. Biochem. 117,
515-519.

Heisenberg. R. C.. Deery. N. J.. and Dickenson. P.
J. (1976) Biochemistry ig, 4246-4254.

Nickerson. J. A. and Hells. N. N. (1976) Biochem.
Biophys. Res. Commun. §§J 620-626.

Sheterline. P. (1977) Biochem. J. 166, 533-539.
Nebster. R. B.. Henderson. D.. Osborn. N.. and
Neber. R. (1976) Proc. Natl. Acad. Sci. U.S.A. 1g,
6611-5615.

Daleo. G. R.. Piras. l. N.. and Piras. R. (1976)
Bur. J. Biochem. g2, 639-346.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

161

Sloboda. R. D.. Rudolph. S. A.. Rosenbaum. J. L..
and Greengard. P. (1975) Proc. Natl. Acad. Sci.
U.S.A. 134 177-161.

Sandoval, I. V. and Cuatrecasas. P. (1976)
Biochemistry 19, 3424-3432.

Shelanski. N. L.. Gaskin. F.. and Cantor. C. R.
(1973) Proc. Natl. Acad. Sci. U.S.A. 124 765-766.

Jacobs. H. and Caplow. N. (1976) Biochem. Biophys.
Res. Commun. 66, 127-135.

Jacobs. N. (1975) Ann. New York Acad. Sci. 253,
562-572.

 

Scheele. R. B. and Borisy. G. G. (1979) in
Nicrotubules (Roberts. K. and Hyams. J. S.. eds.).
Academic Press, New York. Pp. 175-254.

Scheele. R. B. and Borisy. 6. G. (1976) Biochem.
Biophys. Res. Commun. 12; 1-7.

Weingarten. N. D.. Suter, I. N.. Littman. D. R..
and Rirschner. N. N. (1974) Biochemistry 16,
6529-5537.

Scheele. R. B. and Borisy. G. G. (1976) J. Biol.
Chem 253, 2646-2651.

Weingarten. N. D.. Lockwood. A. B.. Bwo. S. Y.. and
Kirschner. N. N. (1975) Proc. Natl. Acad. Sci.
U.S.A. 134 1656-1662.

Zeeberg. B.. Cheek. J.. and Caplow. N. (1960)
Biochemistry 194 5076-5066.

Pantaloni. D.. Carlier. N. P.. Simon. C.. and
Batelier. G. (1961) Biochemistry 224 4709-4716.

Nandelkow. B. N.. Narmsen. A.. Nandelkow. B.. and
Bordas. J. (1960) Nature (London) 267, 595-599.

Bordas. J.. Nandelkow. B. N.. and Nandelkow. B.
(1963) J. Mol. Biol. 164, 69-135.

Nandelkow. B.. landelkow. B. N.. and Bordas. J.
(1963) Trends Biochem. Sci. 64 .