Interactions of microtubule-associated protein-2 with microtubules and neurofilaments


Material Information

Interactions of microtubule-associated protein-2 with microtubules and neurofilaments
Physical Description:
x, 125 leaves : ill. ; 29 cm.
Joly, John Charles, 1963-
Publication Date:


Subjects / Keywords:
Research   ( mesh )
Microtubule Associated Proteins -- physiology   ( mesh )
Microtubules -- physiology   ( mesh )
Cytoskeleton -- physiology   ( mesh )
Neurofilament Proteins   ( mesh )
Department of Biochemistry and Molecular Biology thesis Ph.D   ( mesh )
Dissertations, Academic -- College of Medicine -- Department of Biochemistry and Molecular Biology -- UF   ( mesh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph.D.)--University of Florida, 1990.
Bibliography: leaves 115-124.
Statement of Responsibility:
by John Charles Joly.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001484001
oclc - 22503298
notis - AGZ6060
System ID:

This item is only available as the following downloads:

Full Text








This work was made possible due to the love and support

of my parents. Their support through my undergraduate and

graduate studies enabled me to complete this long journey.

I must thank members of the Purich lab; especially Greg

Flynn for giving me my start in the lab and collaborating on

the work in Chapter 2, and Jim Angelastro for constantly

patient help with the flipping HPLC. Also Alexandra

Ainzstein for carrying out one of the digestion experiments

in Chapter 3. I am also greatly indebted to Dan Purich for

support through these years, for tolerating my antics and

for his gracious help in aiding my postdoctoral hunt.

Without his help I am sure I would not have enjoyed this

last year as much as I did.



ACKNOWLEDGEMENTS ................................... ii

LIST OF TABLES .... ........... ...... ........ ......... v

LIST OF FIGURES ....................................... vi

ABBREVIATIONS ....................................... viii

ABSTRACT ................... .......................... ix


1 INTRODUCTION ...... ............. ............... 1

An Overview of the Cytoskeleton ................. 1
Microtubules .................................... 2
Neurofilaments .................. ............... 6
Microtubule-associated Proteins ................. 12
Proposal ............. .. ........ ................. 26


Introduction .............................. ...... 29
Materials and Methods .......................... 30
Results .... .... ................. ..... ............. 33
Discussion ................................... ... .... ...... 42


Introduction ...................... ......... ... 46
Materials and Methods ........................... 47
Results .. .................. .. .... .......... ...... 51
Discussion ............... ...................... 62


Introduction .. .. .............................. 66
Materials and Methods ........................... 68
Results ... ..................................... 74
Discussion ................................. .... 97



Interactions with Neurofilaments ............... 103
Structure of MAP-2 ........................... 107
MAP-2 Sequence Interactions with Microtubules .. 108

REFERENCES .......................................... 114

BIOGRAPHICAL SKETCH .................................. 124


Table page

1-1 Major classes of microtubule-associated
proteins .................................. 13

3-1 Amino acid composition of the 28 kDa
MAP-2 fragment ............................ 55

3-2 Amino terminal sequence analysis of the
28 kDa fragment .......................... 57


Figure page

1-1 Comparison of the carboxyl termini of
both a- and B-tubulin from chick brain .... 7

1-2 Structural organization of the neuro-
filament triplet protein .................. 10

1-3 Comparison of the carboxyl termini of
murine MAP-2 and murine tau protein ...... 25

1-4 Summary of MAP-2 structure in relation to
a microtubule ............................ 27

2-1 Thrombin digestion of radiolabeled MAP-2 .. 35

2-2 Binding of MAP-2 or MAP-2 fragments to
neurofilaments or tubulin ................. 37

2-3 Autoradiogram of MAP-2 and MAP-2 fragments
binding to cytoskeletal protein ........... 38

2-4 MAP-2 binding to purified neurofilament
triplet protein or the L subunit of
neurofilaments ............................ 40

2-5 Autoradiogram of MAP-2 and MAP-2 fragment
binding to neurofilament triplet protein
and L subunit ............................. 41

2-6 Determination of the isoelectric point of
the 28 kDa fragment of MAP-2 .............. 42

3-1 Purification of heat-stable microtubule-
binding fragment and tau .................. 52

3-2 HPLC purification of the microtubule-binding
fragment of MAP-2 digested with microtubules
present .................... ......... ..... 54

3-3 Comparison of proteolytic fragmentation
patterns and the amino terminal sequences of
the microtubule binding fragments of MAP-2
and tau .................................... 58

3-4 HPLC purification of the microtubule-binding
fragment of MAP-2 digested initially without
microtubules .............................. 62

4-1 Stimulation of microtubule assembly with
synthetic peptides ......................... 76

4-2 Time course of peptide induced assembly .... 77

4-3 Electron micrograph of peptide induced
assembly ................................... 79

4-4 Immunofluorescence of microtubules polymer-
ized without and with m2 peptide ........... 80

4-5 Critical concentration plot of peptide
induced tubulin polymerization ............. 81

4-6 Seeded assembly of tubulin with synthetic
peptides .................................. 83

4-7 Effects of MAP-2 peptides on MAP binding to
microtubules ............................... 84

4-8 Densitometry of the coomassie blue stained
gel ................ ..................... .. 86

4-9 Comparison of the stimulation of tubulin
polymerized by peptides m2 and m2' ......... 87

4-10 Effects of increasing the mn2 concentration
on high-molecular-weight MAP binding to
microtubules ............................... 89

4-11 32P-MAP-2 binding to taxol-stabilized
microtubules .............................. 91

4-12 Displacement of trace phosphorylated MAP-2
from taxol-stabilized microtubules by
unlabeled MAP-2 ............................ 92

4-13 Displacement of trace phosphorylated MAP-2
from taxol-stabilized microtubules by
peptide m2'. ................................ 94

4-14 Radiolabeled MAP-2 binding to taxol-
stabilized microtubules in the presence and
absence of 1.5 mM m2' peptide .............. 95

4-15 Double reciprocal plot of MAP-2 binding .... 96

5-1 The octadecapeptide repeats of murine MAP-2,
murine tau, bovine tau, and a corresponding
sequence of the 190 kDa adrenal gland MAP .. 112



ATP, adenosine triphosphate

cAMP, cyclic adenosine monophosphate

cDNA, complementary DNA

DEAE, diethylaminoethyl

GTP, guanosine triphosphate

HPLC, high pressure liquid chromatography

MAP, microtubule-associated protein

mRNA, messenger RNA

NF, neurofilament

pI, isoelectric point

PMSF, phenylmethanesulfonyl fluoride

SDS, sodium dodecyl sulfate


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



John Charles Joly

May, 1990

Chairman: Dr. D. L. Purich
Major Department: Biochemistry and Molecular Biology

Bovine brain microtubule-associated protein-2 (MAP-2)

is a 280 kDa protein that binds to microtubules and

neurofilaments. MAP-2 was fragmented by thrombin into a 240

kDa projection domain and a 28 kDa microtubule-binding

domain. The 28 kDa cleavage fragment possessed a

neurofilament binding site for the L subunit of

neurofilaments. The thrombin cleavage site of MAP-2 was

very similar to the chymotryptic cleavage site of the

microtubule-associated protein tau. The 28 kDa microtubule-

binding fragment was derived from the carboxyl terminus of

the intact protein and the 240 kDa projection domain was

from the amino terminus. The microtubule-binding fragment

was very rich in lysine and arginine residues and its

isoelectric point was approximately 10.0. The microtubule-

binding fragment contained three octadecapeptide imperfect

repeats located fifty residues from the thrombin cleavage

site. These sequences were chemically synthesized to assay


for promotion of tubulin polymerization. Only the second

promoted tubulin polymerization in vitro, yielding

microtubules of normal morphology. The time course of

peptide-induced microtubule assembly was similar to

microtubule-protein in vitro. This octadecapeptide

displaced MAP-lb from MAP-containing microtubules. The

addition of the next three amino acids in the MAP-2 sequence

to the carboxyl terminus of the peptide increased its

ability to promote tubulin polymerization at lower

concentrations and displaced all high-molecular-weight MAPs,

MAP-la,b and MAP-2a,b. This extended peptide displayed

competitive binding with radiolabeled MAP-2 to taxol-

stabilized MAP-free microtubules, suggesting the peptide

bound to the same site on microtubules as MAP-2. The

dissociation constant for MAP-2 binding was 3.4 MM in the

absence of the extended peptide and 14 MM in the presence of

1.5 mM peptide. The estimated inhibition constant for the

extended peptide is 0.5 mM, about 100 times lower than for

the Km of MAP-2. These observations suggested that the

second repeated sequence of MAP-2 represents an important

recognition site for MAP-2 binding to microtubules and that

other structural features within MAP-2 may reinforce the

strength of MAP-microtubule interactions.


An Overview of the Cytoskeleton

The cytoskeleton of eukaryotic cells is a dynamic

organelle responsible for maintaining cell shape and

rigidity, cell motility, and intracellular vesicle transport

and trafficking. The cytoskeleton consists of three main

types of filaments: microfilaments, intermediate filaments,

and microtubules. Microfilaments are seven nanometers in

diameter and are composed primarily of actin but also

contain actin-binding proteins. Intermediate filaments are

ten nanometers in diameter and vary in composition depending

on the particular cell type. In brain tissue, the neuronal

cell intermediate filaments are made of neurofilament

proteins while in glial cells, the intermediate filaments

are composed of glial fibrillary acidic protein.

Microtubules are 24 nanometer diameter structures composed

mainly of the heterodimer tubulin. In addition to tubulin,

there are cell specific microtubule-associated proteins

(MAPs) that bind to microtubules.

When specific cytoskeletal filaments are mixed in

vitro, they interact with each other. If bovine spinal cord

neurofilaments are mixed with bovine brain microtubules, the

viscosity of the resulting solution increases greatly and a

gel is produced (Runge et al., 1981). This is thought to be

a relevant and physiological interaction based on

microscopical examination of neuronal axoplasm which shows

neurofilaments and microtubules running parallel to each

other in neurite processes. The work of Hirokawa and his

coworkers over the years has demonstrated through microscopy

techniques that projections exist between the filamentous

structures (Hirokawa et al., 1985; Hirokawa 1982). These

projections could be cross-bridges connecting the two types

of filaments. Possible candidates for the cross-bridges are

MAPs. Both microtubule-associated protein-2 (MAP-2) and tau

proteins are MAPs and can bind to neurofilaments as well as

microtubules (Letterier et al., 1982; Heimann et al., 1985;

Miyata et al., 1986). It is the role of MAPs, specifically

MAP-2 and its interactions with microtubules and

neurofilaments, that is the focus of this thesis.

Microtubule Structure and Function

Microtubules are the most dynamic component of the

cytoskeleton and exist in all eukaryotic cells except

enucleated erythrocytes. The microtubule core is formed

from alpha and beta tubulin heterodimers arranged into

parallel rows, extending the length of the tubule, termed

protofilaments. Isolated microtubules and those observed in

sectioned cell specimens contain thirteen protofilaments,

while most microtubules assembled in vitro possess fourteen

protofilaments (McEwen and Edelstein, 1980). The tubulin

subunits arrange themselves in a head-to-tail fashion along

the protofilaments yielding the distinct polarity in


microtubules. This polarity was first observed by Rosenbaum

and Child (1976) and Witman (1975), who demonstrated biased

addition to microtubules in vitro. One end of the

microtubule polymer displays an increased rate of addition

compared to the other end. This characteristic polarity of

microtubules can be determined by interactions with

microtubule-binding proteins. One method is based on the

interaction of dynein with assembled microtubules (Haimo et

al., 1979; Haimo, 1982). Dynein is a large protein found in

flagella of Tetrahymena that binds to microtubules and

hydrolyzes adenosine triphosphate to produce the whip-like

motion during movement. The main globular head of the

dynein molecule tilts at an angle of 550 in the direction of

the end of preferred growth of the microtubule. Another

method to distinguish microtubule polarity takes advantage

of special in vitro solution conditions that favor formation

of microtubule walls decorated with additional

protofilaments. The extra protofilaments align themselves

into hook-shaped sheets and curve in one specific direction

depending on the polarity; either clockwise or

counterclockwise (Burton and Himes, 1978).

This intrinsic polarity of microtubules enables

microtubules to perform vectorial functions. When

intracellular vesicles travel along a microtubule in an

axon, the vesicle generally moves in one direction. Rarely

will a vesicle change its direction after starting its

journey, and most vesicles move toward the cell body in a

neuron (Hollenbeck and Bray, 1987). The polarity of a


microtubule is clearly a vital property for interactions

with other cytoskeletal components as well as for their

proper function. In a non-neuronal cell, microtubules are

capped at their minus end near the centrioles and extend the

plus end to the cell margin, yielding an overall radial

polarity in the cell.

Some other interesting properties of microtubules are

that they self-assemble in vitro in the presence of GTP at

warm temperatures and that they depolymerize at cold

temperatures or upon the addition of calcium ions. Each

alpha/beta dimer binds two moles of GTP per mole tubulin,

but each heterodimer has two types of nucleotide binding

sites (Weisenberg et al., 1968; Berry and Shelanski, 1972).

The beta subunit has an exchangeable nucleotide binding site

that readily exchanges GDP for GTP. The alpha subunit has a

nonexchangeable binding site that exchanges GDP for GTP very

slowly. Guanosine triphosphate nucleotides are hydrolyzed

only upon or after heterodimer incorporation into the

polymeric tubule.

Both in vitro and in vivo microtubule dynamics have

been monitored through the use of tubulin modified with a

fluorescent tag. This tag can be a direct covalent

attachment of a fluorochrome or by using an anti-tubulin

antibody with a secondary antibody labeled with a

fluorescent probe. In most cases examined the microtubules

have been shown to be highly dynamic structures rapidly

polymerizing and depolymerizing depending on the surrounding

conditions (Mitchison and Kirschner, 1984; Kristofferson et


al., 1986; Sammak et al., 1987). In vitro, the addition of

MAPs reduces microtubule dynamics (P.S. Yamauchi, personal


Microtubules are essential for proper mitotic function.

When the mitotic spindle poison colchicine is added to

cells, mitotic arrest is seen; particularly at metaphase

prior to microtubule depolymerization in anaphase which

achieves chromosome movement toward the spindle poles. The

chromosomes are connected to the microtubules through

kinetochores which cap the microtubules at one end, while

the opposite end is capped by the spindle pole or centriole.

During interphase the centrioles seem to act as microtubule

organizing centers and are located just outside the nucleus.

Some other important functions of microtubules are the

maintenance of cell anisometry and promotion of cell shape

changes. Nerve axons and retinal rod cells rely on the

microtubule network to support their distinctive cell

morphology, especially the neurite processes (Heidemann et

al., 1986). Conversely, the activation of platelets and the

action of polymorphonuclear leukocytes depend on changes in

the microtubule-cytoskeleton (Malawista, 1986).

Microtubules also provide the basic framework for the cell

motility machinery as they are essential components of

flagella and cilia.


The main component of microtubules is tubulin. This

protein exists as a heterodimer of 100 kDa with each subunit

possessing a molecular mass of approximately 50 kDa. There

is considerable homology between the a and B subunits

suggesting a common ancestral origin (Valenzuela et al.,

1981). Both a and B tubulin are very conserved across

species lines suggesting that there are stringent

requirements on the structure of tubulins over a reasonably

great phylogenetic range (Cleveland and Sullivan, 1985).

Both subunits contain a glutamate-rich carboxyl terminus

that is thought to be responsible for MAP binding to

microtubules (see Fig. 1-1). When the carboxyl termini of

both subunits are proteolytically removed with subtilisin,

the tubulin can self-assemble without MAP binding (Serrano

et al., 1984). In fact, when these tubules are sedimented

and the pellet fraction analyzed, the MAPs are found only in

the supernatant fraction. The presence of high

concentrations of salt (0.4 M -0.6 M) disrupts MAP binding

to microtubules indicating the interaction of MAPs with

microtubules is dependent on the ionic interactions. Thus

the glutamate rich carboxyl termini are thought to be the

binding sites for MAPs which in turn contain positively

charged residues (Littauer et al., 1986).


Neurofilament Structure

Neurofilaments are one member of the intermediate

filament protein family of which there are five main

components: (1) acid keratins, (2) basic keratins, (3) glial

fibrillary acidic protein, vimentin, peripherin, and

desmin, (4) neurofilaments and alpha-internexin, and (5)

lamins. Keratins are found in epithelial cells and their







0 )

4C -


0 0


0 0




in o


derivatives such as skin and nails (Steinert and Roop,

1988). Desmin filaments are found mostly in muscle cells

while vimentin filaments are located in mesenchymal cells

(Steinert and Roop, 1988). Glial fibrillary acidic protein

is the basic building block of glial filaments which are

found in glial cells (Steinert and Roop, 1988). Peripherin

is located in the neurons of the peripheral nervous system

(Portier et al., 1984) while neurofilament proteins are

found in most neuronal cells (Steinert and Roop, 1988).

These filaments are composed of three subunits. The main

core of the filament is made of the L subunit, a 70 kDa

protein. The other two subunits are the M subunit and H

subunit which are 150 kDa and 210 kDa respectively. Both of

these subunits contain multiple phosphates on serine and

threonine residues.

The neurofilament proteins have some common structural

features with each other and all other intermediate filament

proteins. Each has a 40 kDa conserved rod domain believed

to be derived from a common ancestral gene (Weber et al.,

1983). This common rod-shaped region is rich in alpha

helical content and possesses a very conserved epitope that

reacts with a mouse monoclonal antibody that recognizes all

intermediate filament proteins (Pruss et al., 1981). This

40 kDa region has been conserved and is essential for

intermediate filament assembly (Steinert et al., 1981;

Geisler and Weber, 1981). Each intermediate filament

protein contains hypervariable regions that flank the

central rod region. These hypervariable regions form the


amino and carboxyl termini of the proteins also known as the

head and tail regions respectively. It is the variation in

these regions that distinguishes each individual

intermediate filament protein. The difference between

neurofilament proteins and other intermediate filament

proteins like desmin and vimentin is the tail region.

Normally the tail region is only approximately 5 kDa in mass

but is over 55 kDa in the H subunit of neurofilaments and 50

kDa in the M subunit. It is the tail regions of the M and H

subunits (see Fig. 1-2) that contain many of the

phosphorylation sites found in these proteins (Julien and

Mushynski, 1983). The tail region of the L subunit is

considerably shorter than the M or H subunits and is rich in

glutamate content which may be important in MAP binding

similar to the glutamate rich carboxyl termini of alpha and

beta tubulin. The tail regions are considerably less

conserved across species lines.

Neurofilament Function

The neurofilament proteins are found mostly in axons as

opposed to the cell body and dendrites. Once assembled into

polymers, neurofilaments do not easily dissociate (Giesler

and Weber, 1981; Moon et al., 1981). In fact, they are

insoluble in aqueous buffers and their purification relies

on solubilization with high concentrations of urea

(Tokutake, 1984). Upon removal of the urea by dialysis, the

filaments reassemble. Filament formation after dialysis

from urea is not restricted to the neurofilament triplet

(i.e. neurofilaments composed solely L, M, and H subunits),

N \-c

Coilla Coilib

Coil 2 Tail a E segment


Coil I

Coil 2 Tail El KSPI E2 KSP2 KE including SP segments


Coil I

Coil 2 Tail a E and KSP segments

KEP segment

Fig. 1-2 Structural organization of the neurofilament
triplet proteins. All three proteins contain coil-coil
regions common to all intermediate filament proteins.
Differences are seen in the amino terminal heads and
carboxyl terminal tails. The phosphorylation sites in NF-M
and NF-L are in the KSP-rich sequences.





but for each individual subunit when purified from the other

two subunits can also be assembled by this method although

the filaments are not as structurally intact for the M and H

subunits (Tokutake et al., 1984). Unlike the case of facile

microtubule assembly/disassembly, intermediate filament

proteins form long-lived structures that provide the only

known function of these proteins, namely stabilization and

maintenance of the axon's structural integrity and caliber.

Neurofilaments increase the size of an axon by adding to the

volume occupied by the cytoskeleton (Lasek et al., 1983).

The contribution to axonal diameter or caliber, has a direct

effect on axonal function because the larger the axonal

diameter, the faster the action potential travels down the

axon. Also, this increase in size permits specialization of

nerve cells because the largest axons have the ability to

excite several target nerve cells at one time (Zucker,

1972). This can synchronize an entire population of cells

to act coordinately as with muscle cells to produce

movement. Axon size is very important to organisms which

must coordinate large muscles to produce rapid and essential

processes necessary to organismal survival. With a minimum

number of junctions, axon size is very important and

expression of the neurofilament genes can be crucial for

proper organismal behavior and survival (Bullock and

Horridge, 1965). Interestingly, some axons contain little

neurofilament proteins but still function properly

apparently as a result of increased microtubule content that


can replace neurofilaments in increasing the size of the

axon (Morris and Lasek, 1984).

While neurofilaments are required architecturally for

proper axon function, they are not required for such axonal

dynamics as axonal transport which is dependent on

microtubules. When an axon is disrupted by the neurotoxin

B,8'-iminodipropionitrile, the axonal inner structure is

rearranged with neurofilaments segregating to the outer

regions of the axon diameter and microtubules and organelles

segregating toward the center (Papasozomenos et al., 1981).

This model system has been used to examine interactions with

microtubules and neurofilaments where it has been shown that

cross-links exist between microtubules and neurofilaments as

well as within neurofilament networks themselves. An

antibody specific for MAP-2 has been found to co-localize

with both groups of filaments in iminodipropionitrile-

treated neurons implicating MAP-2 as a cross-linker within

and between these cytoskeletal systems (Papasozomenos et

al., 1985).

Microtubule-Associated Proteins


MAPs are a varied class of proteins that have been

classified on the basis of their binding or modification of

microtubules. When one considers the diversity of

microtubule functions, it is not surprising that there

exists a great number of proteins that regulate temporal,

spatial, and metabolic controls of microtubule processes

(see Table 1-1). The first attempt to characterize proteins

*1) I ( I
I r0
IHO 0 0
0 I e al >-,0

- *- Q04 10 0 0 0
r o 4-) u
S* 0 *- n ON 0 M
00 0I9MA
o w k 0 '--H 024 c
*H0r H 'D *Hri

40) 4.) 0 1 -r W -I 0 O
>4 00
0U a40 4-


0 V
00 r

0 -u

H 0
-4 0-H

rl 6
H -H


00 C
0 *.

0 W-4
WU) (Q .. ,.Q

r. 04
) a) -as

1 WV r0 o 3
P4 -H 0 r I
04) U AI


, r.
*4 *

n N


I 4

HV a)

2 C) -H
0I. (0


+j -H


4- U
(0 4)
*- -*r-
0 0
m i









*-4 4

I t0
0 c C 4) I
0 0 r4 C (0 ti k t
o 0 4.) 0 > 90
o0 t -r. .- .,-i r-4
r. N M 4J 0'
4 4-) rq a) 0 0 t
I1 r* -H *t Hi(T 4 j 0- 1
S0) 0 > U I 0 0AQ
4 Hl H- r-4 4 4
V.0 40 A H 04 V
o w o .to
O 4 0 +3 A ) ) 4) >0e
0 0 0 VO -H r-I i

W > in l >+
4 t P4 M ca

41 40 to t ro 0 4-
rql .U -,- 4 "

,0 to .r. 0 r H r
o-o ,.-4 t-co
V ,0 j Oq C
N3Q to r O P4 *O *

k! ( <0 to i in a1 4 a) i* 4 )

*+1 4. 0 *-i "I AMAI

C wO mr1i! S4 4 H I H A k I I
r r4 0I O MU- 4 0 In g l 0

0 0 *4 4. 4) o

H r)

E- M -O
H I r
I nI I H I I

II n
N VN N c 0 H o 'q
to 04M

0 0 4 H
O 00
In 0t

4 0 P4 4 tv
04 to ) X

04 HO U) U) ix


associated with microtubule-networks was by Gibbons (1965)

who demonstrated that dynein could be selectively released

and then rebound to axonemes of Tetrahymena. Later,

Weisenberg (1972) observed that microtubules can be

assembled in vitro from crude brain extracts, and this

observation led to the identification of proteins associated

with polymerized microtubules. Through the use of

repetitive cycles of temperature-dependent assembly and

disassembly (Shelanski et al., 1973), quantitative MAPs or

MAPs that bind to microtubules in a defined molar ratio were

identified. The two most obvious classes of MAPs in this

category were the high-molecular-weight MAPs, MAP-1 and MAP-

2, and a family of polypeptides known as tau which has a

molecular mass of 55 to 68 kDa.

Unfortunately, the use of temperature-dependent cycles

has obscured the discovery of many less abundant, but

potentially very significant MAPs. During the course of

several temperature-dependent cycles much protein is lost,

both tubulin and MAPs. Only approximately fifteen percent of

the microtubule-protein is yielded from bovine brain tissue

after three cycles. Also the use of nucleotides, both

adenosine- and guanosine triphosphate, increases the yield

of microtubule-protein by increasing the amount of tubulin

that polymerizes, but also influences MAP associations with

microtubules. For instance, the inclusion of adenosine

triphosphate in the first warming of the brain tissue

extract releases MAP-1C, a cytoplasmic retrograde dynein,

from the microtubule-lattice (Paschal et al., 1987b). This

protein was not characterized as a microtubule-associated

protein for years because it was routinely discarded during

the cycling of microtubule-protein.

One of the high-molecular-weight MAPs already mentioned

is MAP-1 which is actually a family of polypeptides at

approximately 350 kDa. There are two closely spaced bands on

denaturing polyacrylamide gels that are MAP-la and MAP-1b.

Even though they are similar in molecular weight they

exhibit different monoclonal antibody reactivities and

produce different peptide digest patterns. These proteins

exhibit no preferential localization in brain tissue but

their purification from white matter as a family takes

advantage of the fact that other high-molecular-weight

contaminants (i.e. MAP-2) are enriched in gray matter

(Vallee, 1986). There is no available purification to date

for separating the subspecies of this family. MAP-lC on the

other hand has been purified to homogeneity and found to be

a microtubule-activated ATPase that can translocate

microtubules on glass slides and can translocate vesicles

along microtubules in an ATP-dependent fashion (Shpetner et

al., 1988). This very large protein has a native molecular

mass of approximately 450 kDa and scanning transmission

electron microscopy revealed that MAP-lC has a morphology

and mass of a two-headed dynein (Vallee et al., 1988).

Additionally, ultraviolet irradiation in the presence of

vanadate cleaved the protein into two fragments of about the

same size as those produced from flagellar dynein (Paschal

et al., 1987a).

Another microtubule-activated ATPase that can

translocate microtubules as well as intracellular vesicles

is kinesin. Kinesin was first discovered during studies of

organelle transport in the giant squid axon, where

microtubules serve as tracks for the movements of organelles

(Vale et al., 1985a). Kinesin isolated from squid axoplasm

can induce movement of carboxylated latex beads along

purified microtubules or gliding of microtubules on glass in

the presence of ATP (Vale et al., 1985b). The direction of

movement along the microtubules was anterograde or towards

the nerve terminal in an axon. This is exactly the opposite

direction of movement as MAP-1C (Paschal and Vallee, 1987).

Kinesin has since been found in a variety of organisms and

cell types from mammalian brain tissue (Brady, 1985) to

Drosophila melanogaster (Saxton et al., 1988). This wide

distribution of kinesin suggests it may be involved in a

variety of microtubule-based motility systems in different

cell types. Immunolocalization studies in sea urchin eggs

and some mammalian cultured cells have demonstrated kinesin

is located in the mitotic spindle suggesting a role for

kinesin in mitosis (Scholey et al., 1985).

Kinesin is composed of two heavy chains of 120-124 kDa

and two light chains of 62-64 kDa and is a highly elongated

molecule with an axial ratio of approximately 20:1 (Bloom et

al., 1988). Rotary shadowing of kinesin shows it is a rod-

shaped molecule approximately 80 nanometers long. One end

of each kinesin molecule contains a pair of globular domains

while the opposite end is fan shaped (Hirokawa et al.,


1989). Monoclonal antibodies against the heavy chains stain

the globular structures while antibodies versus the light

chains stain the fan-shaped end (Hirokawa et al., 1989). A

60 kDa amino terminal section of the heavy chain corresponds

to the globular head region and contains the nucleotide-

dependent microtubule binding activitity and is thought to

the motor domain (Yang et al., 1989).

Among the lesser known MAPs is the MAP-4 class which

consists of a series of MAPs with molecular masses around

200-240 kDa depending on the species. It is found in almost

all cultured mammalian cells, in mouse tissues, and in

Drosophila melanogaster. MAP-4 shares the property of

thermostablity with MAP-2, a feature that is rare in

proteins of their size. The overall structure and binding

to microtubules of a few MAP-4 types are just starting to be

analyzed. Another group of not very well characterized MAPs

are STOPs which confer cold-stability on microtubules

(Margolis and Rauch, 1981). Normally, microtubules

depolymerize when exposed to cold temperatures; in the

presence of STOPs, the microtubules are stable. These

tubules can be depolymerized by calmodulin and low calcium

concentrations, and it has been shown that these proteins

are all retained on a calmodulin affinity resin (Job et al.,


Tau Proteins

Among the best characterized MAPs to date are the tau

proteins, a family of closely related polypeptides. On

denaturing polyacrylamide gels the proteins exhibit

molecular masses of 55 to 70 kDa with usually four or five

distinct bands appearing. Tau is a phosphoprotein with

series and threonines primarily phosphorylated. It is

located in neuronal tissues and restricted to axons of

neurons. Tau has been found in paired helical filaments and

in plaques from Alzheimer brain patients (Goedert et al.,

1988). It has the remarkable properties of being soluble in

2.5% (v/v) perchloric acid and being insoluble in 25% (v/v)

glycerol (Lindwall and Cole, 1984). Tau also is heat-

stable, and its purification takes advantage of this


The source of heterogeneity in tau was unclear until

recently when it has been demonstrated that the different

tau polypeptides are the result of alternative splicing from

one mRNA transcript (Himmler, 1989). By the selective

splicing of specific exons from the transcript, a specific

tau translation product is synthesized. This work was

performed with bovine brain mRNA. The first tau sequence to

be identified was the murine system by Lee et al. (1988) who

showed two tau transcripts were made in vitro with differing

carboxyl termini. An interesting structural feature of the

transcripts noticed in the murine tau investigation was the

presence of three imperfect octadecapeptide repeats in the

center of the molecule. Each repeated sequence ended with a

proline followed by three glycines. All the repeats

contained series and threonines, and they were rich in the

basic amino acids arginine and lysine. These investigators

hypothesized the repeats were important for tau binding to


Another group of investigators showed a proteolytic

fragment of tau could bind to microtubules (Aizawa et al.,

1988). A 14 kDa chymotryptic fragment present in each

bovine tau polypeptide was found to bind to microtubules.

The amino terminus of this fragment was determined by Edman

sequencing which localized where the fragment was derived

from after comparison with the murine cDNA predicted

sequence. This proteolytic fragment contained two of the

three repeats believed to be involved in microtubule

binding. This was the first indication that not all the

repeats were necessary for tau binding to microtubules. An

interesting feature of bovine tau protein not elucidated by

the protein chemistry work of Aizawa et al. (1988) was shown

by the cDNA sequencing work of Himmler et al. (1989). They

determined that all the bovine forms of the tau polypeptides

contain four repeats rather than three as seen in the murine

forms. It is unclear what the significance of having four

repeats versus three repeats is although tighter or better

binding could be the result.

Some interesting work with the repeats of tau has been

done in vitro. Ennulat et al. (1989) has shown that

synthetic peptides corresponding to the first and second

repeats of murine tau protein can promote the polymerization

of tubulin. The third repeat has also been tested but

failed to promote tubulin polymerization into microtubules.

Himmler et al. (1989) also has demonstrated that a


polypeptide consisting of the four repeats can cosediment

with taxol-stabilized microtubules and he demonstrated that

just two repeats fused together also possessed this ability.

The importance of the cosedimentation data cannot be

overlooked, but the work of Ennulat et al., (1989) is more

significant because this was the first demonstration of a

synthetic peptide of a microtubule-associated protein

performing the same function as a MAP.

Interestingly, Aizawa et al. (1989) found a sequence

that is similar to one of the repeats of tau in the 190 kDa

adrenal gland-specific microtubule-associated protein. This

sequence could also promote the polymerization of tubulin in

vitro. An emerging theme in microtubule cytoskeletal

research is that a group of similar sequences in some MAPs

are responsible for promoting MAP-tubule interactions.

Microtubule-Associated Protein-2

Microtubule-associated protein-2 (MAP-2) is a very

large protein specific for neuronal tissue and restricted to

dendrites. Its molecular mass on denaturing gels is

approximately 280 kDa but the predicted mass from its cDNA

is only about 200 kDa (Lewis et al., 1988). MAP-2 is very

similar to tau in that it is heat-stable, a phosphoprotein,

and the murine form contains a trio of imperfect

octadecapeptide repeats. MAP-2 also shows heterogeneity on

denaturing gels splitting into two high-molecular-weight

forms, MAP-2a and MAP-2b. It is unknown what the cause of

this heterogeneity is. The only known function of MAP-2 is

to polymerize tubulin into microtubules.


MAP-2 can be phosphorylated by a variety of kinases.

The cAMP-dependent protein kinase, calmodulin dependent

protein kinase, calcium/phospholipid-dependent protein

kinase, and protein kinase C have all been shown to use MAP-

2 as a substrate (Goldenring et al., 1985; Tsuyama et al.,

1986; Akiyama et al., 1986). MAP-2 can also be

phosphorylated by non-neuronal specific kinases such as the

insulin receptor kinase and an epidermal growth factor

stimulated kinase (Kadowaki et al., 1985; Hoshi et al.,

1988). When MAP-2 is isolated by standard cycling

procedures from brains it contains about 10 moles phosphate

per mole of MAP-2 (Tsuyama et al., 1986). About ten more

phosphates can be added with exogenous cAMP-dependent

protein kinase. Rat brain MAP-2, isolated immediately

following rapid in vivo heat-treatment, contains

approximately 46 moles of phosphate per mole of MAP-2

(Tsuyama et al., 1987). Such microwave treatment reduces

the activity of phosphoprotein phosphatases. The

phosphorylation state of MAP-2 has a direct influence on its

function. The more phosphorylated the protein is, the less

its affinity for tubulin is and the ability to promote

polymerization is reduced (Murthy and Flavin, 1983).

MAP-2 has very little organized secondary structure.

Circular dichroic measurements of MAP-2 revealed it

contained very little alpha-helix or beta-sheet (Hernandez

et al., 1986). This same study determined that MAP-2

possessed a highly elongated structure as analyzed by gel

filtration chromatography and analytical


ultracentrifugation. In addition, the predicted structure

from the cDNA studies of Lewis et al., (1988) also revealed

little organized secondary structure.

Surprisingly, when MAP-2 is digested with various

endoproteases, only two major fragments are usually

produced: the first, a small fragment approximately 28-36

kDa in mass and the second, a large 240 kDa fragment. The

small fragment contains the microtubule-binding site of MAP-

2 (Vallee, 1980; Flynn et al., 1987), and the large fragment

is known as the projection domain because it is seen in

electron micrographs before protease treatment protruding

from the microtubule wall but is absent after protease

digestion (Vallee and Borisy, 1978). This limited digest

pattern with several proteases suggests some higher order of

structure. A completely random structure would give greater

heterogeniety in protease digests. If trypsin or

chymotrypsin are employed, the digest patterns are more

complex but the microtubule-binding products are 34-36 kDa.

The projection domain contains a majority of the

phosphorylation sites. The significance of these

phosphorylation sites in the projection domain remains

unclear. One known function of the projection domain is

that it contains a binding site for the regulatory subunit

of cAMP-dependent protein kinase (Vallee, 1986). This

suggests that MAP-2 may be associated with a protein kinase.

This kinase could phosphorylate the microtubule-binding

domain thereby modulating the affinity of MAP-2 for


The entire cDNA structure of murine MAP-2 recently was

reported by Lewis et al. (1988), and it showed that MAP-2

possessed three imperfect octadecapeptide repeats similar to

murine tau. Secondary structure predictions from the cDNA

confirmed the findings of Hernandez et al. (1986) showing

MAP-2 contained little alpha-helical content or beta-pleated

sheet content. A synthetic polypeptide of 100 amino acids

of MAP-2 sequence containing the first and second repeats

was shown to bind to MAP-stabilized microtubules and could

cycle with these microtubules (Lewis et al., 1988). This

report revealed extensive homology of MAP-2 with tau

especially in the carboxyl termini of both proteins as

depicted in Fig. 1-3. It will be interesting to know the

bovine MAP-2 sequence to see if it contains three

octadecapeptide repeats like murine MAP-2 or four repeats as

in the bovine tau proteins.

Recently a third form of MAP-2 has been found in

addition to MAP-2a and b (Garner et al., 1988). This

protein is a 70 kDa heat-stable MAP that cross-reacts with

MAP-2 antibodies but is only expressed in neonatal and

juvenile rats. Northern blots of different developmental

stages in rat show a 6 kilobase mRNA in neonatal brain

tissue and a 9 kilobase mRNA in adult brain tissue when

probed with a MAP-2 specific cDNA (Garner and Matus, 1988).

When this 70 kDa protein, now termed MAP-2c, was sequenced,

the sequence was the same as MAP-2a and b except for a 1,372

amino acid deletion corresponding to the central section of

the adult form (Papandrikopoulou et al., 1989). This was

K f
4 ct


-i ~


-K >

n E





A ^

> 0
U) e
U) U)
< >

t o
0 (
H >



( *







e *
a *

U *
M *
Mn *

> 4
# *

0 *



Q *
(0 *
U *

QH *
Q *

04 0'

0 ni

E-44 E-4
Q t
S 0Q
>< > <

m)- E-o

MS 0

U *

i a


0) 44
*dr 0
p r4

: *.- *

0 0 U

0) (1 )
m0 C U

(4 4r.I 'd 4


W04 0
$4 U
40 0
'- 1 4 ) -4-
4- 0

0 0 0

-'-4 V Q.0
4-)+J4 04-

0I OH u

W 0.

0 PV
w 14 0 IA
-) $4 4)
C 0 $4
0 4)0
1 04

t 0

(4 0 W 4.)

0 00

* *




OK 0
04' 0
ch 6
. .

ce *
v K


due to alternative splicing of the mRNA from one gene of

MAP-2. The protein still contained the carboxyl terminus

ofadult MAP-2 with the triad of imperfect repeats. It is

currently unclear what the significance of the embryonic and

adult forms is, although a reduced degree of cytoskeletal

cross-linking during axonal and dendritic growth with the

embryonic form is a possibility (Papandrikopoulou et al.,


Altogether, the findings on the structural organization

of MAP-2 can be depicted as shown in Fig. 1-4. This figure

shows the section spliced out of mature MAP-2 creating the

MAP-2c form. The binding site for the regulatory subunit of

cAMP-dependent protein kinase is located at the amino

terminus and the proposed microtubule-bundling domain is

located at the extreme carboxyl terminus of the molecule.

When this bundling sequence is removed, MAP-2 loses its

ability to cause microtubule bundling (Lewis et al., 1989).

This sequence contains homology to leucine zipper proteins

and can be functionally replaced by these leucine zippers

such as the one contained in GCN4. It has been hypothesized

that the bundling domain at the carboxyl terminus of MAP-2

interacts with another MAP-2 carboxyl terminus causing the

microtubules to bundle.


The aim of this study was to determine how MAP-2

interacted with two different cytoskeletal components,

neurofilaments and microtubules. This was to be

accomplished through the use of proteolytic digests to

MAP-2 Structure:



Protein Kinase
Binding Site

Hinge Region

Absent in MAP-2C ,

Microtubule Projection
Binding Arm
Bundling Domain


Fig. 1-4 Summary of MAP-2 structure in relation to a
microtubule. The amino terminus of MAP-2 contains a binding
site for the regulatory (R) subunit of cAMP-dependent
protein kinase. The carboxyl terminus contains the
imperfect repeats as well as a proposed bundling sequence.
The hinge region contains a protease sensitive site.

determine which fragments) of MAP-2 could interact with the

different polymers. Once the fragment was determined, a

structural characterization of its biochemical properties

was to be carried out. This included a determination of its

isoelectric point, amino acid composition, and amino

terminal analysis. With the availability of the entire

sequence of murine MAP-2, internal sequences responsible for

promoting tubulin polymerization in vitro were to be

identified. In addition to tubulin polymerization, MAP-2

sequences responsible for displacement of high-molecular-

weight MAPs from microtubules were to be identified.



Neuronal cytoplasm is highly organized, and both

microtubules and neurofilaments run parallel with respect to

the axon's longitudinal axis in a manner suggesting

microtubule-to-neurofilament cross-linking (Wuerker and

Palay, 1969; Ellisman and Porter, 1980; Hirokawa, 1982). In

vitro observations indicate that microtubules interact with

neurofilaments, and MAPs can enhance the attainment of high

solution viscosity and/or gelation (Runge et al., 1981;

Aamodt and Williams, 1984b; Minami and Sakai, 1983;

Letterier et al., 1982). Aamodt and Williams (1984a) used

falling-ball viscosmetry to demonstrate the occurrence of an

optimal MAP level in plots of viscosity/gelation versus MAP

concentration; they likened this MAP concentration profile

to that of bivalent antibody cross-linking in

immunoprecipitin formation. Previous studies have traced

this apparently optimal MAP profile to the presence of

endogenous GTPase activity, and the inhibition of cross-

linking/gelation at high MAPs can be eliminated with a GTP-

regenerating system (Flynn and Purich, 1987). The

requirement of GTP is to maintain microtubule stability as

microtubules will disassemble after all the GTP is converted

to GDP. Nonetheless, high-molecular-weight MAPs do bind to

microtubules (Kim et al., 1979; Vallee, 1982; Purich and

Kristofferson, 1984) and neurofilaments (Runge et al., 1981;

Aamodt and Williams, 1984b; Minami and Sakai, 1983;

Letterier et al., 1982), and some interactions of the 280

kDa neuronal MAP-2 have been explored by limited proteolytic

fragmentation. Tubule binding is restricted to a 34-36 kDa

tryptic or chymotryptic fragment of MAP-2, and the remaining

240 kDa component corresponds to the lateral projections

observed in electron micrographs of microtubules decorated

with MAP-2 (Kim et al., 1979; Vallee and Borisy, 1977).

During the course of studies on neurofilament-

microtubule-MAP-2 interactions I sought to localize the

site(s) of neurofilament binding with respect to the tubule-

binding and -projection domains of MAP-2. A thrombin

cleavage technique was developed to obtain these MAP-2

fragments in higher yields than that obtained with trypsin

or chymotrypsin. Interestingly, a 28 kDa tubule-binding

domain was found to contain a neurofilament-binding site.

My studies also suggest that this binding interaction has

considerable ionic character, as suggested by isoelectric

point determinations of the MAP-2 fragment.

Materials and Methods

Bovine thrombin (catalog number, T 4648) and the

catalytic subunit of cAMP-dependent protein kinase were

purchased from Sigma. Ultrapure ammonium sulfate and urea

were purchased from Schwarz-Mann, and carboxymethyl Sephadex

from Calbiochem. [3P]ATP (specific activity > 7000


Curies/mmol) was an ICN product, and Ampholines were

obtained from LKB. Assembly buffer for preparation of

microtubule-protein contained 0.1 M piperazine-N, N'-bis[2-

ethanesulfonic acid], 1 mM ethyleneglycol-bis[B-aminoethyl

ether]-N,N,N',N'-tetracetic acid, and 1 mM magnesium


Preparation of proteins

Bovine brain microtubule-protein was prepared by the

procedure of Shelanski et al., (1973). Neurofilaments were

prepared from fresh bovine spinal cord by the method of

Delacourte et al., (1980) as modified by Letterier et al.,

(1982). Neurofilament triplet protein was prepared as

described by Tokutake et al., (1983), and the NF-L subunit

of neurofilaments was purified according to the method of

Geisler and Weber (1981). Tubulin was separated from MAPs

by the phosphocellulose method of Weingarten et al., (1975).

MAP-2 preparation and phosphorylation

MAP-2 was purified by the method of Herzog and Weber

(1978), concentrated by ammonium sulfate precipitation, and

phosphorylated by the catalytic subunit of cAMP-dependent

protein kinase prior to gel filtration chromatography.

Typically 500 units of kinase was dissolved in 0.025 ml

dithiothreitol (50 mg/ml), incubated at room temperature for

10 minutes, and used immediately with 1.5 millicuries

[32P]ATP, 0.02 mM unlabeled ATP, for 30 minutes at 37C in

the presence of approximately 40 mg heat-stable MAPs. The

MAPs were separated on a BioGel A-1.5M column with the MAP-

2 fractions pooled and concentrated in a dialysis bag

against dry carboxymethyl Sephadex at 4C. This purified

[ 32P]MAP-2 was clarified by centrifugation at 130,000 x g

for 25 minutes in a Beckman Airfuge prior to digestion and

incubations with cytoskeletal proteins.

Digestion of MAP-2 with thrombin

Purified and radiolabeled MAP-2 (50,000 CPM/Ag) was

incubated at 0.4 mg/ml with 4 units/ml thrombin. To

determine the optimum time of digestion, aliquots were taken

at 5 minute intervals from 0 to 30 minutes. Once an optimum

time of 30 minutes was determined, all digestions of MAP-2

prior to incubation with cytoskeletal proteins were

conducted for 30 minutes and quenched by the addition of 1

mM phenylmethylsulfonyl fluoride and incubation on ice.

Sedimentation of Dolymerized protein

Thrombin-digested or undigested MAP-2 was incubated

with neurofilaments for 10 minutes at 4C, or microtubules

for 30 minutes at 37C. The incubations were then layered

over 20% (w/v) sucrose in assembly buffer and centrifuged at

130,000 x g for 20 minutes. Supernatant fractions were

discarded and the pellet fractions were washed with 1 mg/ml

bovine serum albumin and 0.1% (v/v) Triton X-100 and

resuspended in 8 M urea. Aliquots of both supernatant and

pellet fractions were analyzed for radioactivity by liquid

scintillation spectrophotometry, and an equal number of

counts were loaded on 7-17% (w/v) polyacrylamide gels.

Polyacrylamide gel electrophoresis

Sodium dodecyl sulfate gel electrophoresis was carried

out as described by Laemmli (1970), and nonequilibrium pH


gel electrophoresis (NEPHGE) was performed by the method of

Roberts et al., (1984). For the NEPHGE gels radiolabeled

MAP-2 was digested with thrombin and the digestion was

quenched by the addition of an equal volume of 9.3 M urea,

0.5% (w/v) dithiothreitol and 2% (v/v) Nonidet NP-40 in 5 mM

potassium carbonate. The digests were loaded onto "NEPHGE"

tube gels with 1% (v/v) Ampholine 9-11 and 2% (v/v)

Ampholine 3.5-10, and run for 2000 volt-hours. Gels were

stained with coomassie blue, destined, dried under vacuum,

and exposed to Kodak X-AR5 film at -800C.


Binding to microtubules and neurofilaments

In view of the high molecular weight of MAP-2,

proteolytic fragmentation by trypsin or chymotrypsin has

proven to be useful in defining the MAP-2 domain(s)

interacting with other cytoskeletal components (Olmsted,

1986). Vallee (1980) first demonstrated that MAP-2 can be

fragmented into 35 and 240 kDa components by chymotrypsin or

trypsin. The smaller fragment contains the microtubule-

binding domain, and the larger is designated as the

projection-arm domain. While these protease cleavage

products have been very useful in many investigations of

microtubule self-assembly, chymotryptic and tryptic cleavage

do not yield stable limit polypeptides, and the stability of

such proteolytic fragments is quite limited, leading to the

loss of the initially cleaved domains and the ability of

these fragments to stimulate microtubule assembly. In a

survey of the action of other proteases, I observed that


thrombin, an arginine-specific serine protease,

predominantly yielded MAP-2 fragments of 28 kDa and 240 kDa.

MAP-2 cleavage can be readily assessed by SDS gel

electrophoresis of [32P]MAP-2 because this protein is

extensively phosphorylated (Theurkauf and Vallee, 1983). As

seen in Fig. 2-1 this 28 kDa fragment is very stable and

very resistant to further digestion over the time points

shown. Indeed, phenylmethanesulfonyl fluoride at 1.0 mM

final concentration blocked any further degradation over a

five to seven day period at 4C.

This development has allowed for probing with much

greater ease the interactions of MAP-2 fragments with

neurofilaments and microtubules. The basic approach is to

determine which proteins or protein fragments cosediment

with assembled microtubules or neurofilaments using

ultracentrifugation and subsequent electrophoretic analysis.

First MAP-2 was enzymatically phosphorylated with the cAMP-

dependent protein kinase to a level of 1 mole of added

phosphate per mole of MAP-2 based on the conditions of

Letterier et al., (1982). After purification of the

radiolabeled protein by gel filtration chromatography, the

MAP-2 was concentrated, and aliquots were digested with

thrombin. Next, the thrombin-digested fragments (indicated

by the plus sign) or undigested MAP-2 (indicated by the

minus sign) were incubated with microtubules or

neurofilaments under the conditions listed in fig. 2-2. The

polymerized and unpolymerized cytoskeletal proteins were

separated into pellet [p] and supernatant [s] fractions by

0 5 10 15 20 30


- Intact MAP-2 &
Projection fragment

- 28 Kd fragment


Fig. 2-1 Thrombin digestion of radiolabeled MAP-2.
[ 32P]MAP-2 was incubated at 37*C with 4 units/ml thrombin
for the indicated time in minutes. The digestion was
quenched by heating at 1000C for 5 minutes in the presence
of sodium dodecyl sulfate and the products resolved on a 15%
(w/v) polyacrylamide gel. The gel was then dried under
vacuum and expose to Kodak X-AR 5 film.

ultracentrifugation. Because the projection and tubule-

binding domains do not contain identical phosphorylation

sites, a constant total amount of radioactivity was applied

for each electrophoretic sample. Lanes 1-4 of the coomassie

stained gel in Fig. 2-2 and the corresponding lanes of the

autoradiogram in Fig. 2-3 demonstrate that only the 28 kDa

thrombin-produced fragment of MAP-2 binds to neurofilaments.

The next four lanes in both Fig. 2-2 and Eig. 2-3

demonstrate that this thrombin fragment behaves as the so-

called microtubule-binding domain of MAP-2 as it binds to 1

mg/ml of taxol-stabilized microtubules composed solely of

tubulin. In the absence of neurofilaments or microtubules

the 28 kDa fragment remained in the supernatant fraction

even after ultracentrifugation as seen in lanes 9-12 of Fig.

2-3. Indeed, the entire pellet fraction was used for the

electrophoretic analysis in lanes 10 and 12 of Eig. 2-3, and

virtually no high-molecular-weight or fragmented MAP-2

cosedimented without neurofilaments present. These

observations verified that the fragment is only sedimentable

as a result of interactions with either neurofilaments or


Next, I sought to determine the neurofilament

protein(s) interacting with MAP-2 or the 28 kDa fragment.

Earlier work by Miyata et al., (1986) and Heimann et al.,

(1985) demonstrated MAP binding to the low-molecular-weight

subunit of neurofilaments. A second series of binding

assays were conducted and Fig. 2-4 shows the coomassie blue

staining pattern and Fig. 2-5 shows the corresponding

Protein Tubulin or Tb
M /+ +- + -V+ +-- M

\s pspsp s psp s p

I 2 3 4 5 6 7 8 9 10 II 12

Fig. 2-2 Binding of MAP-2 or MAP-2 fragments to
neurofilaments or tubulin. A coomassie blue stained
gradient 7-17% (w/v) polyacrylamide gel is shown.
Radiolabeled MAP-2 was incubated with (+) or without (-)
thrombin as described in Fig. 2-1, quenched with 1 mM
phenylmethylsulfonyl fluoride and incubated with 2 mg/ml
neurofilament protein or 1 mg/ml taxol-stabilized tubulin
with 1 mM guanosine triphosphate. Samples were then handled
as described in "Methods". M refers to molecular weight
markers, s to supernatant fraction, and p to pellet

Protein Tubulin or Tb
:/++ A +I--\ A-+ -
1-Pg-p q i




\s p s ps p s ps ps p/

Fig. 2-3 Autoradiogram of MAP-2 and MAP-2 fragment binding
to cytoskeletal protein. The coomassie blue stained
polyacrylamide gel shown in Fig. 2-2 was soaked in 25% (v/v)
glycerol for 30 minutes after destaining and dried under
vacuum. The dried gel was exposed to x-ray film for 4 hours
without intensifying screens.


autoradiogram. Lanes 3, 4 and 7, 8 of Eig. 2-5 verified

that uncleaved MAP-2 binds to the neurofilament triplet and

to filaments composed solely of L subunit. The data shown

in lanes 1, 2 and 5, 6 of Fig. 2-5 extend the earlier

observations by clearly demonstrating the binding of the 28

kDa fragment to the neurofilament triplet and L component of

neurofilaments. It should be noted that only trace levels

of tubulin are evident in the 50-55 kDa molecular weight

range in the gels shown in Fig. 2-2. Much higher levels of

tubulin are required for binding of MAP-2 to assembled

tubules. This suggests that binding depends on presence of

neurofilaments and does not require tubulin or assembled

microtubules for binding.

Determination of the isoelectric point of the 28 kDa

Tubulin and neurofilament proteins contain negatively

charged regions that may be important in MAP binding

(Olmsted, 1986). The data presented in the preceding

section demonstrate that the 28 kDa fragment of MAP-2 binds

to both cytoskeletal organelles. Accordingly, I tried to

use conventional isoelectric focusing to determine the

isoelectric point of the radiolabeled fragment. This

consistently failed at all ranges of ampholytes, and the

fragment never migrated into the first dimension of a two

dimensional gel. This is indicative of very basic proteins.

In order to estimate the isoelectric point of very basic

neurofilament subunit proteins the Nonequilibrium pH gel

electrophoresis technique is used (Roberts et al., 1984).

When this technique was tried, the fragment readily migrated

Triplet Low No NF
M ATw+ /+ -+ +



\s p s pAs p s p/\s p s p
1 2 3 4 5 6 7 8 9 10 11 12

Fig. 2-4 MAP-2 binding to purified neurofilament triplet
protein or the L subunit of neurofilaments. A coomassie
blue stained gradient 7-17% (w/v) polyacrylamide gel is
shown. Radiolabeled MAP-2 was incubated with thrombin as
described in Fig. 2-2 and in "Methods", and incubated with
either 2.2 mg/ml neurofilament triplet or 1.1 mg/ml L
subunit at 4*C for 10 minutes. Nomenclature is the same as
in fig. 2-2.

Triplet Low NF

I.u'm:j V



\S p sps p s ps p s p
1 2 3 4 5 6 7 8 9 10 11 12

Fig. 2-5 Autoradiogram of MAP-2 and MAP-2 fragment binding
to Neurofilament triplet protein and L subunit. The
coomassie blue gel shown in Fig. 2-4 was soaked in 25% (v/v)
glycerol for 30 minutes after destaining and dried under
vacuum. The dried gel was exposed to x-ray film for 4 hours
without intensifying screens.


into the first dimension of the gel as seen in Fig. 2-6, and

it migrated between two very basic isoelectric point

markers, ribonuclease (pi=9.3) and lysozyme (pI=10.5-11.0).

The thick arrow denotes the position of the 28 kDa fragment

in the gel. From this migration pattern, the isoelectric

point of the 28 kDa fragment of MAP-2 was estimated to be

approximately 10. The migration of the trace-labeled

fragment in both dimensions also indicated that the

phosphorylation conditions do not lead to significant

heterogeneity in overall ionic charge or molecular weight.

Furthermore, the dephosphorylated form will necessarily

display an even higher isoelectric point. The large smear

at the top right side of the gel is the projection domain

and further breakdown products of this domain. These

fragments are acidic and do not readily migrate into the



The findings presented in this chapter indicate that

neurofilaments, a specific class of intermediate filaments,

interact with MAP-2 in the region corresponding to the 28

kDa microtubule-binding domain. In earlier work with actin,

Sattilaro (1986) reached a similar conclusion about the

binding of chymotryptic fragments of MAP-2 to polymerized

actin. Likwise, this same fragment of MAP-2 constitutes the

microtubule-binding domain (Vallee, 1980). Thus, all three

major cytoskeletal classes (i.e. microfilaments,

intermediate filaments, and microtubules) interact with a

common structural region of MAP-2. In view of our estimated



C0 _
5 4


Fig. 2-6 Determination of the isoelectric point of the 28
kDa fragment of MAP-2. An autoradiogram of a two
dimensional gel is shown. The first dimension is
Nonequilibrium pH gel electrophoresis and the second
dimension is sodium dodecyl sulfate polyacrylamide gel
electrophoresis with a 7-17% (w/v) gradient of acrylamide.
Protein molecular weight/isoelectric point reference
standards were: 1) phosphorylase b; 2) bovine serum albumin;
3) carbonic anhydrase; 4) soybean trypsin inhibitor; 5)
ribonuclease and 6) lysozyme.


isoelectric point value of 10 for the 28 kDa component, all

of these MAP interactions must be strongly influenced by

electrostatic charge. Vallee (1982) has clearly

demonstrated that 0.35 M sodium chloride can remove MAPs

from taxol-stabilized microtubules, and this observation is

also in harmony with the notion of ionic interactions

between tubules and MAPs. Furthermore, it has been shown

that increased phosphorylation of MAP-2 reduces its ability

to promote tubulin polymerization (Murthy and Flavin, 1983).

My use of trace labeling of heat-stable MAPs with

[ P]ATP and protein kinase reinforces the general utility

of this method as first applied by Heimann et al., (1985).

The radiolabeled low-molecular-weight thrombin fragment of

MAP-2 is more stable than that obtained with trypsin and

chymotrypsin, and this stability may also facilitate studies

of the stoichiometry and dissociation constants for fragment

binding to microtubules, neurofilaments, or actin. Other

more approximate methods using intact MAPs and either

densitometry or pelleting of assembled cytoskeletal elements

(Miyata et al., 1986; Heimann et al., 1985) still require

refinement and/or verification.

Finally, the thrombin results underscore the facility

and generality of serine protease of MAP-2 into low- and

high-molecular-weight fragments. This behavior is

reminiscent of proteolytic action on myosin. Furthermore,

intact MAP-2 has an isoelectric point of 5.4 (Berkowitz et

al., 1977), whereas I found that the tubule/filament-binding

domain has a value of 10. This suggests that there must be


significant acidic and basic charge localization in the

high- and low-molecular-weight fragments, respectively, of

MAP-2. Nonetheless, the biological significance of this

protein design feature awaits further understanding of the

role of MAP-2 in the cytomatrix.



The thrombin protease digestion of MAP-2 shown in Fig.

2-1 revealed a very stable 28 kDa fragment. This fragment

possessed the sequences responsible for cosedimentation of

MAP-2 with microtubules as well as neurofilaments. Since

this fragment contained the "active site" of MAP-2, I wanted

to learn more about its biochemical properties. It was

already known the fragment had a very basic isoelectric

point and it was suspected that the basic residues were

involved in the binding to the anionic termini of both alpha

and beta tubulin. A second interesting point concerning the

structure of the fragment was the specificity and stability

of the protease digestion products. The autoradiogram of

Eig. 2-1 showed no heterogeneity in the 28 kDa microtubule-

binding fragment and limited heterogeneity in the very

protease-sensitive 240 kDa domain. However, limited

heterogeneity in the larger fragment could be observed by

the low resolving power of the 15% (w/v) polyacrylamide gel

for large molecular weight proteins. Nevertheless, the

specificity of the digestion in producing the 28 kDa

fragment was impressive. This specificity implied some

higher order structure in the MAP-2 molecule yet a previous


report on the circular dichroism spectrum of MAP-2 revealed

little alpha-helical or beta-pleated sheet content. If

there was no higher order structure, I would expect to see a

far more complex pattern of digestion. In order to resove

this paradox a large scale purification of the microtubule-

binding fragment of MAP-2 was undertaken.

In addition, the structure of MAP-2 was probed by

protease digests with and without the presence of

microtubules. The low-molecular-weight digestion products

were isolated and their amino terminal composition checked

by Edman sequencing.

Materials and Methods


32 P]ATP (7000 Ci/mmol) was purchased from ICN along

with ultrapure grades of ammonium sulfate, sodium dodecyl

sulfate (SDS), acrylamide, and bis-acrylamide. Immobilon

was obtained from Millipore Corporation and coomassie

brilliant blue R-250 was from Serva. DEAE-Sephadex A-50 was

purchased from Pharmacia; and bovine thrombin,

trifluoroacetic acid, Mes buffer, and phenylmethane-

sulfonylfluoride (PMSF) were from Sigma.

Preparation of proteins

Bovine brain microtubule protein was prepared by the

procedure of Shelanski et al. (1973). MAP-2 was purified by

the method of Herzog and Weber (1978) and radiolabeled as

previously described (Flynn et al. 1987).

Preparation of heat-stable microtubule-binding fragments

Heat-stable microtubule-binding fragments were prepared

according to Vallee (1986) with the following modifications.

Thrombin was used instead of chymotrypsin at 8 U/ml and 37*C

for 30 minutes to digest thrice cycled bovine microtubule-

protein at a concentration of 5 mg/ml. PMSF was added to 1

mM at the end of the digestion to stop proteolysis. The

assembled tubules were sedimented at 100,000 x g for 75

minutes at 300C and the pellet was resuspended in 0.75 M

NaCl and 1 mM dithiothreitol. After homogenization and

incubation on ice for 30 minutes the protein was heated in a

boiling water bath for nine minutes followed by cooling on

ice for 20 minutes. The resulting slurry was centrifuged

for 30 minutes at 15,000 x g at 40C. The supernatant

fraction contained heat-stable microtubule-binding fragments

from tau and MAP-2. These heat-stable binding fragments

were concentrated by ammonium sulfate precipitation and then

dialyzed against microtubule assembly buffer (100 mM Mes, pH

6.8, 1 mM EGTA, and 1 mM magnesium sulfate) at 4C with 1 mM

PMSF, and passed over a 1 ml DEAE-Sephadex A-50 column

equilibrated in the same buffer. The breakthrough fractions

were pooled and precipitated with 60% (w/v) ammonium

sulfate. After sedimentation, the precipitate was

resuspended in assembly buffer and used for HPLC analysis or

Immobilon blotting.


Preparation of the 28 kDa fragment of MAP-2 without
microtubules present during digestion

Purified MAP-2 corresponding to 0.6 mg at a

concentration of 2.0 mg/ml was digested with 5 U/ml thrombin

for 30 minutes at 37*C. Digestion was quenched with 2 mM

PMSF and cooling on ice for 10 minutes. The digestion

products were passed over a 1 ml DEAE-Sephadex A-50 column

and handled as described in the preceding section.

HPLC purification of the microtubule-binding fragment of
MAP-2 and its amino acid analysis

High performance liquid chromatography was carried out

on a Hewlett Packard Model 1090a chromatograph, equipped

with a diode array detector. The ammonium sulfate

concentrated, microtubule-binding fragments were clarified

by centrifugation at 3000 x g for 5 minutes prior to loading

on a Waters Associates C-18 column, equilibrated in 0.1%

(v/v) trifluoroacetic acid. The protein was eluted with a

linear gradient from 0-50% (v/v) acetonitrile with 0.1%

(v/v) trifluoroacetic acid at a flow rate of 0.5 ml/min and

1 ml fractions were collected. The elution profile was

monitored at a wavelength of 220 nm because the fragments

are very low in aromatic amino acid content. Initially,

polyacrylamide gel electrophoresis was used to check the

composition of material in each peak. Fractions containing

the MAP-2 microtubule-binding domain were pooled, dialyzed

against 100 mM ammonium bicarbonate, lyophilized, and then

hydrolyzed in 6 N HC1 for 24 hours at 1100C. Samples were

analyzed with a Beckman Model 6300 Amino Acid Analyzer.

SDS electrophoresis and blotting

Immobilon was handled according to the manufacturer's

instructions prior to electroblotting. The

polyvinylidenedifluoride membrane was wetted in 100% (v/v)

methanol for 5 minutes followed by soaking in distilled

water for another 5 minutes and then it was allowed to dry.

Ultrapure grades of SDS, acrylamide, and bis-acrylamide were

used to avoid blocking the N-terminus. The electrophoretic

samples for sequencing were heated to 80*C for 5 minutes

after adding Laemmli sample buffer which contained ultrapure

SDS but no bromophenol blue dye. As an indicator of when

the electrophoresis was finished molecular weight markers

were run in adjacent lanes with dye present. To scavenge

any radicals that could possibly react with the samples,

0.1% (v/v) thioglycolate was added to the top chamber

buffer. A 12% (w/v) acrylamide SDS-gel containing heat-

stable microtubule-binding fragments was electrophoretically

transferred to the membrane in 10 mM CAPs, pH 10.0, 10%

(v/v) methanol for 6 hours at 70 volts. The membrane was

stained with Coomassie Brilliant Blue R-250, destined in

50% (v/v) methanol-10% (v/v) acetic acid. The blot was air

dried and stored at -200C in the dark until the sequencer

was available. The band of interest was excised with a

razor blade and sequenced in a gas-phase protein sequencer

(Applied Biosystems 470A protein sequencer) with on-line

phenylthiohydantoin analyzer at the Protein Chemistry core



Site of thrombin cleavage

In order to gain more information about this site of

facile thrombin cleavage, a high-yield isolation method was

developed for amino acid analysis and sequencing

experiments. I again employed thrombin, but digested

assembled three-cycle microtubule-protein (i.e., tubulin and

MAPs), followed by heat-treatment of the resulting pellet to

remove all tubulin and heat-labile MAPs. Then, DEAE-

Sephadex ion-exchange chromatography was used to separate

the heat-stable microtubule-binding fragments from any high

molecular weight digestion products. This ion-exchange

chromatography step was a key part of the purification and

took advantage of the basicity of the microtubule-binding

fragments as they did not interact with the resin while any

incompletely digested MAPs and any contaminating projection

domain did interact with the resin very strongly. This is

shown in Fig. 3-1. The breakthrough fractions from the

DEAE-Sephadex chromatography contained the microtubule-

binding fragments of the heat-stable proteins MAP-2 and tau

(see lane 5) resulting in greater heterogeneity than that

seen in Fig. 2-1 where purified MAP-2 was digested. I also

found that a tau monoclonal antibody recognized the upper

bands in lane 5, identifying them as putative digestion

products of tau.

At this point in the purification scheme, I observed

four closely spaced bands on a gel ranging from 28 to 36

97 --
66 -

31 -a


I ':F


- -

M 1 2 3 4 5

Fig. 3-1 Purification of heat-stable microtubule-binding
fragment of MAP-2 and tau. Microtubule-protein was digested
with thrombin as described under "Methods". The coomassie
blue stain of a 12% (w/v) polyacrylamide gel shows the
purification during its various stages. Lane 1 is before
cleavage; 2, after cleavage; 3, supernatant after
centrifugation; 4, heat-stable protein; 5, after DEAE-
Sephadex chromatography; 6, a larger amount of sample 5. M
corresponds to molecular weight markers.


",, 1


kDa. Many different ion-exchange resins were tried to

separate these bands such as strong and weak cation

exchangers, and hydroxyapatite chromatography, but all were

unsuccessful. A two-dimensional gel using the NEPHGE

technique in the first dimension was run and all four

fragments were found to be very basic but each fragment was

slightly less basic with increasing molecular weight. This

indicated that the use of ion-exchange resins to separate

the fragments was probably futile and another means of

separation was necessary. The microtubule-binding fragments

in lane 5 of fig. 3-1 can be readily separated by reverse-

phase HPLC as shown in Eig. 3-2 where SDS gel

electrophoresis revealed that peak C corresponded to the 28

kDa component. This lane also revealed that there is only

minor contamination by a faster migrating component.

Fraction C was subjected to acid-catalyzed hydrolysis and

amino acid analysis, and these results are listed in Table

3-1. I had previously reported this fragment had an

unusually high isoelectric point in comparison with intact

MAP-2 (Flynn et al. 1987), and the amino acid analysis

confirmed this observation. The fragment is comprised of

nearly 14 mole percent in lysyl and arginyl residues.

Curiously it contains a higher than usual proline content.

The analysis also confirmed the fact that this fragment was

low in aromatic amino acids as it contained only two

tyrosine residues and three phenylalanine residues. This

low content of aromatic residues makes it very difficult to

monitor its purification at 280 nm as is usually done for

cA B C -A-

Q -28

0 10 20 30 40 50 60

Time after Injection, min

Fig. 3-2 HPLC purification of the microtubule-binding
fragment of MAP-2 digested with microtubules present. The
MAP-fragments seen in lane 5 of Fig. 3-1 were separated by
reverse phase HPLC and the fractions resolved by a 12% (w/v)
polyacrylamide gel (see inset) which is coomassie stained.


Amino acid composition of the 28 kDa MAP-2 fragment

Amino Acid 24 h Hydrolysate Estimated
(mol%) Residues/mol


































proteins. Also the analysis was rich in its glutamate +

glutamine content as well as its aspartate + asparagine

content even though it has already been established the

fragment was very basic. This meant that most of these

amino acids were in the amide form rather than the

carboxylic acid form.

Amino-terminal sequence of the microtubule-binding fragment

Microsequencing techniques were employed with the 28

kDa fragment electroblotted from SDS-polyacrylamide gels to

a derivatized nylon screen (Immobilon). This allowed

further reduction in any contamination by other MAP

fragments such as that observed in the HPLC preparation.

The amino terminal sequence obtained was, Thr-Pro-His-Thr-


Lys, based on the results listed in Table 3-2. In the

absence of thrombin treatment, identical sequence

experiments with either electroblotted MAP-2, as well as

MAP-2 in solution, did not yield any phenylthiohydantoin

derivatized amino acids at detectable levels. Likewise,

experiments with the immobilized 240 kDa projection-arm

fragment yielded no sequence data. Protein samples failing

to yield detectable levels of amino acid derivatives were

subjected to acid-catalyzed hydrolysis and amino acid

analysis to assure that sufficient levels of protein for

sequencing had been employed. In all cases adequate levels

of protein were present above the detection limits for


Amino-terminal sequence analysis of the 28 kDa
MAP-2 fragment

Cycle No. Residue (pmol)a Cycle No. Residue (pmol)

1 Thr 74 11 Ala 103

2 Pro 131 12 Ile 74

3 His 15 13 Leu 144

4 Thr 42 14 Val 82

5 Pro 122 15 Pro 102

6 Gly 84 16 Ser 65

7 Thr 54 17 Glu 39

8 Pro 105 18 Lys 105

9 Lys 141 19 Lys 78

10 Ser 85

a Values reported correspond to the sum of the major,
leading, and trailing cycle yields for each amino acid.

W0 C ,C
U i d P *-4
u a) to (A

a) A 4 P 4- -i
V -I 04-

01 0 0 1

rOE-4 J 4
10 a) U o rto

C Oe 1

,-I -, -3
.) 4H I

-.4 C Q
.0 D *uM U-

0 WO C

J0 004O
to4)V 0 0i

W 0 0
. -4 .0 r 0
$4 04 0 4 Vr-

01 i0 )t

S4 0 -4 *

*4 04 M C04

p 0 ( 0 0
H 0 P I '-O 4) 4)

v 0 -,-40 ,
0 C0 C0 Z 0"

0 oil I v (1 0

,g0 l0 u
0 -4 k rO (0 V
IQ (0 C *r -O u
u 1c4 (0 V (0
0 0 0 0
'-4 H- kC *
r (0 C 0I 0CO

Q)0 4 .0 r
ON 0 O 4J

1 tp H0 V 0 Li
VG & 0) 10

( l 0" 0 1 <4


I I I 1 1

0 I
I \6
0- I i o
o c? I

\I ( 0 /i
o i If) '
I & II ,'

I :I\
I -dl \
I \ I

E. I

c *~c

0, u

^ =3~

sequencing. These findings suggested that the amino-

terminus of MAP-2 is blocked and that the 240 kDa fragment

is derived from the amino-terminus whereas the 28 kDa

fragment resided at the carboxyl end. These observations

are in accord with the findings by Kosik et al. (1988) who

reported that the N-terminus of MAP-2 appears to be blocked;

moreover, Lewis et al. (1988) reported the entire derived

amino acid sequence using murine MAP-2 cDNA clones. The

primary sequence data with bovine brain MAP-2 correspond to

the murine sequence spanning residues 1626 to 1644 with only

three exceptions. As shown in Fig. 3-3, there is a similar

protease-accessible sequence in the microtubule-binding

fragment of bovine tau protein. In that case, however,

fragments were generated by chymotryptic cleavage (Aizawa et

al., 1988). Both of these cleavage-site sequences reside

approximately thirty-to-forty residues toward the N-terminal

side of the first of three nonidentical octadecapeptide

repeats (indicated schematically by the arrowheads) found in

both MAP-2 and tau (Lewis et al., 1988; Lee et al., 1988).

Amino-terminal sequence of the microtubule-binding fragment
after digestion in the absence of microtubules

My initial experiments to determine if the microtubule-

binding fragment of MAP-2 also bound to neurofilaments,

digested purified MAP-2 alone in solution. In the preceding

section the amino terminal sequence of the binding fragment

was determined but the fragment was generated by digestion

of microtubule-protein at 37C in the presence of 1 mM

guanosine triphosphate. In this regard, MAP-2 was digested

in the presence of microtubules, a condition that might have


influenced the cleavage point of thrombin were the preferred

site of digestion hindered by interaction with the

microtubule lattice. To determine whether the cleavage

point is the same for the microtubule-binding fragment under

the two different conditions, the amino terminal sequence

was checked once again but now in the absence of

microtubules during the digestion. Purified MAP-2 was

cleaved alone in solution, and passed over a DEAE ion-

exchange column to remove the projection domain. The

breakthrough fractions containing the microtubule-binding

domain were pooled, electrophoresed through a polyacrylamide

gel, transferred to a polyvinylidenedifluoride membrane, and

sequenced by automated Edman chemistry. Ten cycles were

performed yielding ten residues that were exactly the same

as the first ten residues in Table 3-2 and Eig. 3-3. This

indicated that thrombin cleaved MAP-2 at the same argininyl

residue in the presence and absence of microtubules.

Additionally, the breakthrough fractions from the DEAE-

Sephadex column were subjected to reverse phase HPLC similar

to the fragments generated in the presence of microtubules.

When loaded on a C-18 column and eluted with the same

gradient as mentioned in the preceding section, one major

peak corresponding to the 28 kDa fragment was seen (see Fig.

3-4) and one minor peak corresponding to peak A in Fig. 3-2

was seen. The elution time for the 28 kDa fragment was

unchanged from Fig. 3-2 and corresponded to peak C from this




Time after Injection,min

Fig. 3-4 HPLC purification of the microtubule-binding
fragment of MAP-2 digested initially without microtubules.
The inset shows a 12% (w/v) polyacrylamide gel before and
after the HPLC purification: lane 1, digest before
purification; lane 2, fraction 22; lane 3, fraction 23; lane
4, fraction 30; lane 5, fraction 31. Fractions were
collected every two minutes and were 1 ml in volume.



The experiments described in this chapter were

designed to gain further insight about the microtubule-

binding fragment of MAP-2. There is now general agreement

that initial proteolytic cleavage of MAP-2 yields two

fragments (Vallee 1980; Flynn et al. 1987). With thrombin,

these initial cleavage products corresponding to values of

240 kDa and 28 kDa based on electrophoresis, are quite

stable with regard to further degradation. All efforts to

sequence Immobilon-linked MAP-2 and the similarly

immobilized 240 kDa projection arm fragment consistently

failed to yield any detectable levels of PTH-amino acids.

Nevertheless, acid hydrolysis and subsequent amino acid

analysis of these immobilized proteins demonstrated that

sufficient levels were clearly present for detection in the

gas-phase sequencer. This observation led me to believe

that intact MAP-2 showed no evidence of a free N-terminus,

and another group recently reported the same difficulty in

attempts to sequence MAP-2 (Kosik et al. 1988). These

observations suggest that the MAP-2 amino acid sequence, as

derived from nucleotide sequence data (Lewis et al. 1988),

does not provide a complete account of the MAP-2 primary

structure, and further work will be required to establish

the nature of the N-terminal modification blocking Edman


A striking common structural feature in MAP-2 and tau

emerges from the combined findings of Aizawa et al. (1988)

and these studies. The former found that chymotryptic

cleavage of the bovine tau proteins yielded a microtubule-

binding fragment with the N-terminal sequence shown in fig.

3-3, and I have now demonstrated that thrombin attacks at a

similarly accessible region in bovine MAP-2 (See also Fig.

3-3). It should be noted that both of these cytomatrix

proteins have four proline residues in exact registration,

and with the exception of the occurrence of a val-pro in the

MAP-2 sequence, each of the prolines in both cleavage sites

was preceded by a hydroxy-amino acid. Efforts to survey

other known sequences in the GenBank database have indicated

the uniqueness of these protease-accessible regions in tau

and MAP-2; however, Earnshaw et al. (1987) described a

centromere-binding protein containing three prolines in

exactly corresponding positions with little other structural

relatedness to tau and MAP-2. Also it should be noted that

the NF-M sequence in chicken contains a proline at every

third residue for 102 residues in the repeated sequence

(EXPXSP)17 (Zopf et al., 1987). The circular dichroism

spectral data of Hernandez et al. (1986) indicates that

uncleaved MAP-2 contains little, if any, alpha helical or

pleated-sheet secondary structure; yet, the preferential

action of the endoprotease thrombin at a single site

suggests that MAP-2 may display some "hinge-point" behavior

akin to the protease-accessible region of myosin. This

region may permit the projection arm to swing away from the

microtubule surface. Certainly, the observed sedimentation

coefficient of 4.5 (Hernandez et al. 1986) also suggests

that MAP-2 has extended a flexible structure. The roughly

spherical hemoglobin molecule, itself only one-third the

molecular weight of MAP-2, has an almost identical

sedimentation coefficient (Sanders et al. 1981).

Chymotryptic cleavage between Tyr-128 and Ser-129 in the tau

proteins may reflect a corresponding protease-accessible

site of structural discontinuity between microtubule-binding

and projection domains. These hinge point regions may be

very important in their presentation of the microtubule-

binding sequences that actually interact with tubulin.



Microtubule-associated proteins (MAPs) exhibit one of

several properties: the ability to copolymerize with

tubulin during microtubule assembly, the capacity to utilize

tubulin or another MAP as substrates for enzyme-catalyzed

modification, or the use of microtubules as the

architectural framework for motility (Olmsted, 1986; Purich

and Kristofferson, 1984). The first property is shared by

the high-molecular-weight proteins (MAP-1 and MAP-2) as well

as the tau proteins, and these proteins remain associated

with reassembled microtubules during the course of

microtubule-protein purification. Recently, the cDNA-

derived amino acid sequences of the murine MAP-2 (Lewis et

al., 1988) and the murine tau (Lee et al., 1988) proteins

have been defined, and these proteins were both found to

contain a related triad of imperfectly repeated

octadecapeptide sequences in their tubule-binding regions.

Oligopeptide analogues of the repeated sequences in murine

tau and a 190 kD bovine adrenal gland MAP can promote

microtubule assembly as monitored by light scattering

techniques (Ennulat et al., 1989; Aizawa et al., 1989). I

wished to investigate whether the triad of imperfect


octadecapeptide repeats of murine MAP-2 and a MAP-2 sequence

from the protease-accessible hinge region could promote

microtubule polymerization and mimic the action of MAPs.

While several peptides corresponding to sequences in

fibrous MAPs can stimulate microtubule assembly, very little

is known about whether these synthetic peptides constitute

the entire site necessary for the MAP binding to

microtubules. At the time little information on MAP-2

sequences responsible for binding to microtubules or

promotion of tubulin polymerization was available. The

report of Lewis et al. (1988) showed a 100 residue

polypeptide consisting of the first two imperfect repeats

plus flanking sequences could cosediment with MAP-stabilized

microtubules. No information on the binding of small

peptides or promotion of tubulin polymerization was known.

The experiments described in this chapter attempt to define

an "active site" of MAP-2 by testing for sequences promoting

tubulin polymerization. The most likely candidates were the

repeated sequences since two of them were in the 100 residue

polypeptide of Lewis et al. (1988). These peptides were

chemically synthesized along with a hinge-region sequence

and tested for stimulation of tubulin polymerization.

If the repeated sequences are indeed the primary sites

of interaction, then those promoting tubule assembly in the

absence of MAPs may also displace MAPs from microtubules or

block their binding to microtubules. Moreover, I was

motivated to learn whether a particular peptide and MAP

display competitive binding behavior that would indicate the


peptide(s) binding to the same site as MAPs on the

microtubule. The effectiveness of all three MAP-2 repeated

peptide analogues in terms of MAP displacement from

microtubules was also checked. These experiments show that

peptides corresponding to the second repeated sequence of

MAP-2 can promote microtubule polymerization and displace

MAP-1 and MAP-2 from recycled microtubule-protein.

Materials and Methods


[32 P]ATP (7000 Ci/mmol) and [3H]GTP (18 Ci/mmol) were

purchased from ICN along with ultrapure grades of ammonium

sulfate, sodium dodecyl sulfate (SDS), acrylamide, and bis-

acrylamide. Liquid scintillation cocktail 3a70 was obtained

from Research Products International. Acetate kinase was a

Boehringer Mannheim product, while phosphocellulose resin

and GF/F glass fiber filters were from Whatman. Anti-8

tubulin antibody was purchased from Amersham. DEAE-Sephadex

A-50 was purchased from Pharmacia; and bovine thrombin,

trifluoroacetic acid, Mes, Pipes, and Tris buffers,

dithiothreitol, guanosine triphosphate, bovine serum

albumin, EGTA, Triton X-100, phenylmethanesulfonylfluoride,

catalytic subunit of cAMP-dependent protein kinase, and goat

anti-murine IgG Texas red conjugate were from Sigma. t-BOC

amino acids and the phenylacetamidomethyl resin were from

Applied Biosystems International. Taxol was a gift supplied

by Dr. Matthew Suffness at the National Cancer Institute,

Bethesda Md.

Preparation of proteins

Isotonic bovine brain microtubule-protein was isolated

according to the method of Karr et al. (1979) and stored at

-80*C after two cycles of assembly-disassembly. Hypotonic

bovine brain microtubule-protein was isolated by the method

of Shelanski et al. (1973). Tubulin was prepared according

to the method of Kristofferson et al. (1986). [ P]MAP-2

was purified by the procedure of Herzog and Weber (1978) as

modified by Flynn et al. (1987) except the purified protein

was concentrated by ammonium sulfate precipitation after gel

filtration chromatography. Unlabeled MAP-2 was prepared

identically except the phosphorylation reaction was omitted

prior to the gel filtration column.

Preparation of synthetic peptides

All peptides were made with an Applied Biosystems

synthesizer model 430A according to the method of Erickson

and Merrifield (1976) with t-BOC protected amino acids and

starting with a phenylacetamidomethyl resin. Peptides were

cleaved and deprotected using a mixture of hydrogen

fluoride, anisole, and dimethyl sulfide in a 9:1:0.3 ratio

(v/v) at -10*C for 2.5 hours. After evaporation the resin

was washed with cold diethyl ether and extracted into 1 M

acetic acid and then freeze dried. Purity was tested by

HPLC profile or by gas phase microsequencing. The peptides

were stored at -20*C as a lyophilized powder. All synthetic

peptides except the mN peptide were made by Dr. Jan Pohl of

the microchemical facility at Emory University. The m.


peptide was synthesized by the Protein chemistry core

facility at the University of Florida.

Microtubule assembly with synthetic peptides

All assembly experiments were done with a GTP-

regenerating system (MacNeal et al., 1977) consisting of 2

units/ml of acetate kinase, 10 mM acetyl phosphate, and 0.1

mM [3H]GTP (20 Ci/ml). All assay mixes also contained 1 mM

dithiothreitol to maintain reduced sulfhydryls in the

peptides. The peptides were weighed out just before use and

dissolved in 100 mM Pipes, pH 6.8, 1 mM EGTA, 1 mM magnesium

sulfate with 1 mM dithiothreitol. Varying concentrations of

each peptide were added to 1.6 mg/ml pure tubulin and 0.4

mg/ml three-cycle microtubule-protein and incubated at 30C

for 30 minutes. The extent of microtubule assembly was

monitored by the rapid filtration assay of Maccioni and

Seeds (1978) as modified by Wilson et al. (1982).

Microtubules were diluted 20X into 100 mM Pipes, pH 6.8, 1

mM EGTA, 1 mM magnesium sulfate, 1% (v/v) glutaraldehyde,

10% (v/v) dimethylsulfoxide, 25% (v/v) glycerol, and 1 mM

ATP and kept at 300C until ready to assay. The diluted

mixture was then applied to Whatman GF/F filters on a vacuum

filtration device presoaked in the same buffer except no

glutaraldehyde was used. Each filter was then washed with

15 ml of the same buffer and the radioactivity was

solubilized in 1.5 ml 0.1 N NaOH for 30 minutes followed by

addition of scintillation cocktail.


Preparation of microtubule seeds and elongation assay

Seeds were prepared according to Kristofferson et al.

(1986). Tubulin at 5 mg/ml was assembled with 1 mM GTP in

80 mM Pipes, pH 6.8, 10 mM magnesium chloride, 1 mM EGTA, in

30% (v/v) glycerol at 370C for 30 minutes. The microtubules

were then sheared with 3 passes through a 22 gauge needle to

produce microtubule seeds. The seeds were diluted 100X into

0.5 mg/ml tubulin and varying concentrations of peptides.

After 30 minutes at 37C the samples were handled as

described in the preceding section for measuring tritiated

guanine nucleotides.

Electron microscopy

Microtubules were diluted 10X into a warmed solution of

1% (v/v) glutaraldehyde in microtubule-assembly buffer and

fixed for one minute at room temperature. The fixed samples

were placed on a carbon coated Formvar 400 mesh grid and

negatively stained with 1% (w/v) uranyl acetate. Grids were

observed on a JEOL 100 CX microscope at 50,OOOX

magnification. Samples with and without peptide were

processed identically.

Fluorescence microscopy

Microtubules formed in the presence of synthetic

peptides were diluted 10X into a warmed solution of 1% (v/v)

glutaraldehyde and 0.1% (v/v) Triton X-100 in microtubule-

assembly buffer and fixed for 2 minutes at room temperature.

The microtubules were diluted to 50,000 times their original

concentration and an aliquot of 100 gl was centrifuged on to

a glass coverslip at 30 psi in a Beckman airfuge for 20

minutes. The coverslip was then fixed in -20*C methanol for

4 minutes and blocked with 10 mg/ml bovine serum albumin in

phosphate buffered saline, pH 7.3, with 0.1% (v/v) Triton X-

100 for 10 minutes. The coverslip was then stained with a

murine anti-tubulin antibody at a dilution of 1:200 for 30

minutes followed by washing in the same buffer. A goat

anti-murine IgG secondary antibody conjugated with Texas red

fluorochrome was used at a dilution of 1:80 followed by

washing in phosphate buffered saline, pH 7.3, with 0.1%

(v/v) Triton X-100. The coverslips were mounted in 20 mM

Tris, pH 7.9, with 90% (w/v) glycerol and viewed with a

Zeiss epifluorescence microscope at 1000X power.

Isotonic microtubule experiments

Before use the isotonic bovine brain microtubule-

protein was carried through a third cycle of

assembly/disassembly, and the concentration of protein was

determined by the method of Bradford (1976). Synthetic

peptides were weighed out just prior to use and dissolved in

PEM buffer (100 mM Pipes, pH 6.8, 1mM EGTA, 1 mM MgS04)

containing 1 mM dithiothreitol. Peptides were added at the

indicated concentrations to 0.8 mg/ml isotonic microtubule-

protein with 0.5 mM GTP and 1 mM dithiothreitol and

incubated at 37C for 20 minutes. The microtubules were

subsequently stabilized with 10 pM taxol for 10 minutes at

370C. The samples, 2504l final volume, were then

centrifuged for 8 minutes at 300,000 x g, 37*C in a Beckman

TL 100.2 rotor. The pellets were dissolved in 8 M urea and

analyzed by gel electrophoresis.


Competition with radiolabeled MAP-2

All radiolabeled MAP-2 experiments were performed with

polypropionate airfuge tubes which were coated with 10 mg/ml

bovine serum albumin for 5 minutes and rinsed with PEM

buffer just prior to use. This treatment reduces

nonspecific binding of proteins to the walls of the

centrifuge tubes. The radiolabeled MAP-2 was clarified

prior to use for 20 minutes at 130,000 x g in a Beckman

airfuge to remove any aggregated or denatured protein.

Phosphocellulose-purified tubulin was incubated at 5 mg/ml,

37*C with 1 mM GTP for 20 minutes and subsequently

stabilized with 50 AM taxol for an additional 10 minutes.

The microtubules were then diluted twenty-fold into a

solution containing 3 AM radiolabeled MAP-2 with either

unlabeled MAP-2 or synthetic peptides for 20 minutes at

37C. The solution also contained 10 MM taxol and 1 mM GTP

to maintain microtubule stability. The samples, final

volume of 100l, were then carefully loaded into coated

airfuge tubes with the aid of a microcapillary pipetter onto

a 50 Al layer of 20% (w/v) sucrose in PEM buffer warmed to

37*C. The samples were centrifuged for 30 minutes at

130,000 x g and the supernatants removed and discarded. The

pellets were washed with 100 Ai of 10 mg/ml bovine serum

albumin in phosphate buffered saline, pH 7.3, containing

0.1% Triton X-100 and resuspended in 100 Al 8 M urea.

Aliquots of 25 pl were taken for liquid scintillation



Polvacrylamide gel electrophoresis

Polyacrylamide gel electrophoresis was done by the

method of Bloom et al. (1985) omitting sodium dodecyl

sulfate in the separating and stacking gels and adding 2 M

urea to the separating gel. Gels were stained with

coomassie Brilliant Blue R-250 and scanned with an LKB

ultrascan densitometer.


Peptide interactions with tubulin and microtubule-protein

To further analyze sequence(s) responsible for MAP-2

binding to tubulin within the 28 kDa fragment, four peptides

were synthesized. The first (m, = TPHTPGTPK) corresponded

to the N-terminus of the 28 kDa fragment from bovine MAP-2

that was determined previously by microsequencing (see Table

3-2). The others corresponded to the three octadecapeptide


m3 = AQAKVGSLDNAHHVPGGG). Peptide Nm was based on the

bovine sequence data while the murine MAP-2 sequence data

was used for ml, m2, and m3, because no such data are yet

available for the bovine MAP-2. The high state of purity of

each peptide was confirmed on the basis of HPLC elution

profile analysis or gas-phase microsequencing.

First, I sought to determine whether any of these

peptides would influence the assembly of microtubule-protein

that contained both tubulin and MAPs. I worked with

recycled microtubule-protein to which sufficient pure

tubulin was added to lower the content of MAPs to about one-

fifth their normal level. This final composition was


approximately 5% MAPs and 95% tubulin by weight. This ratio

was chosen to accentuate any stimulatory effects of the

peptides on the assembly process, and no microtubule

polymerization occurred at the protein concentrations used

without peptide addition. To assay the extent of

microtubule assembly at different levels of peptides mn, ml,

m2, and m3, [3H]guanine nucleotide uptake was measured with

the glass fiber filter assay of Maccioni and Seeds (1978) as

modified by Wilson et al. (1982). Only peptide m2,

corresponding to the second repeat in MAP-2, stimulated

microtubule-assembly as evidenced by the data shown in Fig.

4-1. The level of m2 peptide required for polymerization

was very high. Concentrations greater than 250 AM were

needed to stimulate microtubule polymerization. When

peptides Nm, ml, or m3 were employed individually, no

incorporation of guanine nucleotide was observed above

background levels. Moreover, in companion experiments, I

found that none of these peptides mixed individually with

peptide m2 altered the stimulation of microtubule assembly

by peptide m2. A range of concentrations from 0-1 mM was

tried for peptides mi and m3 with 0.5 mM or 0.75 mM m2

peptide but none showed any effect on the extent of

microtubule polymerization.

I also found that assembly of pure tubulin could be

stimulated by m2 only. Indeed, assembly with tubulin and m2

exhibits a typical time-course for the polymerization

process as shown in Fig. 4-2. An initial lag phase

indicative of subunit nucleation was seen followed by a


o a O

o 8

E 4 *
o /

0 4--

0 5 10 15 20 25 30

Fig. 4-2 Time course of peptide induced assembly.
Phosphocellulose-purified tubulin (1.0 mg/ml) was incubated
with m, peptide (1.0 mM) for 30 min. at 37*C. At the times
indicated, the amount of GTP incorporation was determined as
described in "methods".


rapid polymerization phase that plateaued around 30 minutes.

Without addition of m2 peptide, no tritium label is retained

on the glass fiber filters. I verified that the observed

polymerization resulted in microtubules by using electron

microscopy (see Eig. 4-3) and immunofluorescence microscopy

(see Fig. 4-4). The electron micrographs revealed a

morphology typical of microtubules composed solely of

tubulin. When tubulin (1 mg/ml) was incubated with and

without peptide m2 (1 mM), microtubules were observed only

in those micrographs of samples to which this peptide had

been added. The same was seen for the immunofluorescence

micrographs in Fig. 4-4 where panel A had the same

concentrations of protein and peptide as the electron

micrograph in fig. 4-3 and panel B was without added

peptide. Panel A shows typical in vitro microtubules

stained with an anti-tubulin antibody. The concentration of

tubulin used was 1 mg/ml because it was clearly above the

critical concentration for peptide m2 induced assembly while

for tubulin alone it was just at the lower limit for

polymerization (see Fig. 4-5). Any molecule that shifts the

x-intercept to the left is a microtubule-stabilizer and any

molecule that shifts it to the right is a microtubule-

destabilizer. Clearly, m2 is a stabilizer of microtubule

polymerization. I also tested the action of several common

inhibitors of microtubule assembly to learn whether peptide

m2 induced assembly in a manner akin to normal assembly of

brain microtubules. Inclusion of colchicine (0.1 mM),







Fig. 4-3 Electron micrograph of peptide induced assembly.
Tubulin (1.0 mg/ml) and m2 peptide (1.0 mM) were incubated
for 30 min. at 370C and then diluted into 1% glutaraldehyde
in microtubule assembly buffer warmed to 370C. After
fixation for one minute, the sample was processed for
electron microscopy. An identical sample without m2 peptide
was also done but showed no microtubules. Magnification is
50,OOOX and the bar equals 0.25 Am. Formvar coated grids,
uranyl acetate, and photographs were supplied by the
Electron Microscopy core facility

Fig. 4-4 Immunofluorescence of microtubules polymerized with
and without m2 peptide. Panel A shows the same tubulin and
peptide concentrations as Fig. 4-3 and panel B shows just
tubulin with no peptide addition. Both samples were stained
with anti-beta-tubulin followed by an anti-mouse IgG
conjugated with Texas red fluorochrome.


0 5.0


L 2.5-- (-)M2

0.0 1 I I
0 1 2 3 4
mg/ml Tubulin

Fig. 4-5 Critical concentration plot of peptide induced
tubulin polymerization. Varying concentrations of tubulin
plus microtubule seeds were mixed with or without m2 peptide
(1.0 mM) and assayed for GTP incorporation after 30 min. at


calcium ion (2 mM), or podophyllotoxin (0.1 mM) resulted in

complete inhibition of peptide m2-induced assembly.

These observations indicate that only the peptide m2,

with a sequence corresponding to the second repeated region

of the microtubule-binding fragment MAP-2 could stimulate

tubulin assembly. Nonetheless, the possibility remained

that the other peptides could still promote elongation, but

not nucleation, of microtubule assembly. To investigate

this possibility, I added pre-formed microtubule seeds to

tubulin (0.5 mg/ml) and [3H]GTP in the presence or absence

of the peptides. Without any peptide additions, only a

minimal increase in guanine nucleotide incorporation was

observed; however, upon addition of peptide m2, significant

assembly was again observed. By contrast, peptides mi and

m3 failed to cause any significant increase of labeled

guanine nucleotide incorporation into microtubules beyond

background levels (see Fig. 4-6). Thus, m2 is the only

peptide that can stimulate nucleation and elongation.

Displacement of MAPs from recycled microtubule-protein by
MAP-2 repeated sequence peptides

While only peptide-m2 promoted microtubule self-

assembly, I was interested in determining whether mi and m3

might also bind to assembled tubules and displace MAP-2.

Therefore, the ability of the MAP-2 repeated sequence

peptides to displace high-molecular-weight MAPs from

assembled microtubules was assessed. For this purpose,

microtubule-protein isolated by the isotonic extraction

method (Karr et al., 1979) was used because such protein as

*-0 M2
E A-A M3


e -
2- 5.0

0.0 -
0.0 0.5 1.0 1.5 2.0
mM peptide

Fig. 4-6 Seeded assembly of tubulin with synthetic peptides.
Microtubule seeds were added to a solution containing 0.5
mg/ml tubulin (a level below the critical concentration).
Varying amounts of m,, m and m3, were added and
polymerization initiated by warming to 37*C. The plotted
values correspond to radiolabel incorporation over a 30
minute period.


isolated from the gray matter of the brain is rich in both

MAP-1 and MAP-2. The isotonic microtubule-protein was

stabilized with taxol after 20 minutes of 1 mM GTP at 370C.

This method reduced the amount of abnormal microtubule

structures common with taxol-induced polymerization where

taxol is added at the start of polymerization. In this

experiment, taxol was added after 20 minutes of microtubule

assembly. After a 10 minute incubation, the peptides were

added and the samples kept at 370C for an additional 20

minutes. In addition to the three repeats of MAP-2, two

other peptides were tested. The first (m,*) was a glycine

substitution for a lysine in the m, peptide converting the

carboxyl terminal sequence to that of all the other repeats.

The second (m2') was the same as m2 plus the next three

residues present in the MAP-2 sequence. These additional

residues were RVK which made the peptide more electro-

positive as well as possibly adding more structural

conformation. As seen in Fig. 4-7, SDS gel

electropherograms of the microtubule after assembly and

centrifugation, indicate that MAP-2 was only susceptible to

displacement by a 21-amino acid peptide m2' corresponding to

the m2 sequence above plus residues RVK at the C-terminus.

Interestingly, MAP-lb was selectively displaced by peptide

m2, and all high-molecular-weight MAPs were removed from

microtubules in the presence of peptide m2'. Densitometry

tracings of lanes 2, 5, and 6 indicate the profiles of MAPs

with m2, m2', and no peptide, respectively (Fig. 4-8).

Because the 21-amino acid peptide was more effective in

MAP-1 b
MA P-2a, b

m2 m3

m1 m2 none

Fig. 4-7 Effects of MAP-2 peptides on MAP binding to
microtubules. Coomassie Blue staining of proteins in
pelleted microtubule fractions after electrophoresis on a 4%
polyacrylamide gel: (Lane 1) ml; (lane 2) m2; (lane 3) m3;
(lane 4) m 1'; (lane 5) mi2'; (lane 6) no peptide. All
peptides were added to a concentration of 2.0 mM.

C,, I I
z i

S|1 1 I
none l /I. i,
"- 'l iil

b a b a ,,


Fig. 4-8 Densitometry of the coomassie blue stained gel.
Upper trace represents lane 6, middle trace lane 2, and the
lower trace is lane 5 of the gel depicted in Fig. 4-7.

displacing MAPs, it was compared to the m2-peptide with

regard to the promotion of microtubule assembly using

tritiated GTP incorporation as a measure of polymerized

protein (see fig. 4-9). I found that tubulin polymerization

was considerably more effective in the presence of m2',

especially at lower peptide concentrations. At the greatest

concentration of peptide, the level of tritiated GTP

incorporation was the same for both m2 and m2'. Together,

these observations suggest that only peptides corresponding

to the second repeated sequence can displace MAPs from

assembled microtubules. The data also indicate that both

MAP-1 and MAP-2 can be displaced by peptide-m2', suggesting

further that this peptide may bind to common, or closely

overlapping, sites on microtubules. Peptides m,, m2, and m3

were otherwise without effect, as was an analogue of mI

containing a gly in place of lys toward the C-terminus.

This mi analogue was synthesized because the lysine residue

was disrupting a possible beta turn structure. In nm2 and m3

as well as all three repeats of tau, the carboxyl termini

are proline-(glycine)3. This structure was hypothesized to

be important for microtubule-assembly since both the first

two tau repeats contained this carboxyl terminal tail and

could promote tubule polymerization. The first repeat of

MAP-2 was very similar to the first repeat of tau except for

the glycine difference; however, the m, analogue failed to

promote tubulin polymerization and also failed to displace

high-molecular-weight MAPs.



f ,0







Fig. 4-9 Comparison of the stimulation of tubulin
polymerization by peptides m, and m2'. The incorporation of
tritiated GTP into microtubule-polymer was measured to
examine the assembly promoting activity of m, and the
extended analogue m2'3. Phosphocellulose-purified tubulin
was used at 1 mg/ml. The open circles represent m2 and the
closed circles m21.

0 4-

_ T I


* I

0 1


A gradient of m2' peptide was used to determine the

effective concentration range of displacing the MAPs. The

results shown in Fig. 4-10 demonstrate that the extended

second repeated sequence peptide m*2 removes MAP-1 and MAP-

2 from microtubules in a concentration-dependent manner.

Again, densitometry was used to gauge the extent of MAP

depletion in the assembled tubule fraction, and the

concentration of peptide-m2' that displaces 50% of MAP-2 was

about 1.5-2.0 mM (data not shown). This level of peptide-

g2' is about four times the concentration needed to promote

tubulin polymerization in the absence of MAPs or

microtubule-stabilizing agents.

Radiolabeled MAP-2 binding to microtubules

While the findings presented in Figs. 4-7 and 4-10

provide clear evidence of MAP displacement, a quantitative

displacement/binding assay was developed to more accurately

measure the MAP displacement. MAP-2 was incubated with

3',5'-cyclic-AMP-stimulated protein kinase and [c- P]ATP

under conditions that have been found to result in the

incorporation of about 1-1.5 phosphoryl groups per MAP-2

molecule (Flynn et al., 1987). First I determined the

concentrations of tubulin and MAP-2 necessary for saturation

of binding to microtubules. The level of tubulin used was

0.25 mg/ml which was taxol-stabilized like the isotonic

microtubule-protein described in the previous section. At

this low level of tubulin, 1-5 AM of 32P-MAP-2 was seen to

saturate the available binding sites on the microtubules.

This is depicted in fig. 4-11.




0 0.5

1.0 1.5 2.0

Fig. 4-10 Effects of increasing the m2' concentration on
high-molecular-weight MAP binding to microtubules. Peptide
m2' was added to isotonic microtubule-protein to the final
mM concentration indicated at the bottom of each lane.
After centrifugation, the pellet fractions were analyzed by
gel electrophoresis. The coomassie Blue staining of a 4%
polyacrylamide gel is shown.



Because extensive MAP-2 phosphorylation can alter the

affinity of MAP-2 to microtubules (Murthy and Flavin, 1983;

Hoshi et al., 1988), I wanted to compare the binding

behavior of the trace phosphorylated MAP-2 with unmodified

MAP-2 isolated by the standard recycling preparation

protocol (Shelanski et al., 1973; Herzog and Weber, 1978).

Accordingly, taxol-stabilized microtubules were incubated

with the phosphorylated MAP-2 in the presence of several

concentrations of unphosphorylated MAP-2 (Fig. 4-12). The

data points show that the amount of radiolabeled MAP-2 bound

to microtubules decreases at increasing concentrations of

the unlabeled MAP-2; the solid line is the theoretical curve

calculated on the basis of isotopic dilution, using the

ratio R = [MAP-2*]/([MAP-2] + [MAP-2*]), where labeled and

unlabeled protein are MAP-2* and MAP-2, respectively. The

data normalized with respect to R (Fig. 4-12 inset) indicate

that the relative affinities of both MAP-2 species are the

same within experimental error.

Displacement of labeled MAP-2 by a second repeated sequence

To gain a more quantitative view of MAP-2 displacement,

the amount of microtubule-bound [32P]MAP-2 as a function of

the concentration of peptide-m2' was examined. As shown in

Fig. 4-13, MAP-2 is displaced by this peptide, but the

desorption process is not described by a typical hyperbolic

dissociation curve. The basis of the slight stimulation of

MAP-2 binding at 0.5 mM peptide-m2, is unclear, but careful

inspection of the MAP-2 band in lane 2 of Fig. 4-10 revealed

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd