The Microtubule-binding region of microtubule-associated protein-2


Material Information

The Microtubule-binding region of microtubule-associated protein-2 investigations with site-directed mutagenesis
Alternate title:
Investigations with site-directed mutagenesis
Physical Description:
xii, 178 leaves : ill. ; 29 cm.
Coffey, Richard Lawrence, 1967-
Publication Date:


Subjects / Keywords:
Research   ( mesh )
Microtubule-Associated Proteins -- genetics   ( mesh )
Microtubule-Associated Proteins -- isolation & purification   ( mesh )
Microtubule-Associated Proteins -- physiology   ( mesh )
Microtubules   ( mesh )
Protein Binding   ( mesh )
Mutagenesis, Site-Directed   ( mesh )
Cattle   ( mesh )
Rats   ( mesh )
Mice   ( mesh )
Molecular Sequence Data   ( mesh )
Base Sequence   ( mesh )
Amino Acid Sequence   ( 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, 1994.
Bibliography: leaves 168-177.
Statement of Responsibility:
by Richard Lawrence Coffey.
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 - 002324371
oclc - 50514317
notis - ALS7923
System ID:

Full Text



Richard Lawrence Coffey





I thank my parents for their support and love, which

made this endeavor possible. I am grateful to Dr. Dan

Purich for his guidance and support throughout my graduate

studies. I also wish to thank the members of the Purich lab

for their friendship and technical assistance. I am

particularly indebted to Farzin Foruhari for his superb

help. I wish to thank the members of my committee for their

helpful suggestions and advice, and I thank Dr. Brian Cain

for help with site-directed mutagenesis. I am grateful to

Dr. Michael Bubb for his time and effort in performing the

analytical ultracentrifugation described in Chapter 3. I

also thank the members of the Department of Biochemistry and

Molecular Biology for their help and friendship over the

past five years, especially Rich Schnizer, Phillip Hartzog,

Sue Boehlein, Holly Gray, and Ian Hornstra.





ABSTRACT . . .. x



General Overview of the Cytoskeleton . 1
Microtubules . .. 2
Tubulin . . 5
Microtubule-Associated Proteins . 13
Introduction. . 13
Microtubule-Based Motor Proteins 14
Fibrous Microtubule-Associated Proteins 15
General features of MAP-2, MAP-4, and tau 15
MAP-2 ... . 21
Proposal . . 32


Introduction . . 34
Materials and Methods . 36
Results . . 47
Discussion . . 85


Introduction . . 90
Materials and Methods . ... 92
Results . . 99
Discussion . . 116



Introduction . . 124
Materials and Methods . 126
Results . . 131
Discussion . . 146


Discussion . . 152
Future Directions . .. .. 159

REFERENCE LIST. ..... . .168



1-1 Carboxyl-terminal acidic regions of mammalian
alpha and beta tubulins . 10

1-2 Schematic diagram of the fibrous, heat-stable
MAPs MAP-2, MAP-4, and tau . 17

1-3 Nonidentical 18 amino acid repeats found in
fibrous, heat-stable MAPs . 26

2-1 General structural organization of brain MAP-2
including nucleotide and deduced amino acid
sequence of the bovine microtubule-binding
region (MTBR) . 49

2-2 SDS gel electrophoretic analysis of bacterially
expressed MTBR1, at various stages of
isolation . ... 52

2-3 HPLC elution profile of the recombinant MAP-2 MTBR
subjected to CM, reverse phase chromatography 54

2-4 Microtubule assembly induced by the recombinant
MAP-2 microtubule-binding region . 57

2-5 Extent of tubulin polymerization in the presence
of bacterially expressed MAP-2 fragments .. 59

2-6 Displacement of MAP-2 MTBR from taxol-stabilized
microtubules by a 21-amino acid peptide (m1')
corresponding to the second nonidentical repeat
sequence of MAP-2 . .... 62

2-7 Indirect immunofluorescence microscopy of
microtubules assembled with the recombinant
MT-binding fragment containing the N-terminal T7-
genel0 epitope sequence (MTBRI,,) ... 64

2-8 Circular dichroism spectroscopy of the recombinant
MAP-2 MT-binding fragment . 67

2-9 Two models for repeated sequence interactions with
microtubules . . 71

2-10 Displacement of MTBR1, from taxol stabilized
MTs by synthetic 18 amino acid peptides
corresponding to m2-like sequences substituted
for the first and third repeats in the expression
plasmids described in the text . 73

2-11 Displacement of MT-binding domain mutants and
wild-type by peptide-m2' . 75

2-12 Extent and rate of tubulin polymerization induced
by mutant and wild-type forms of MTBRI,6 79

2-13 Displacement of wild-type and mutants
MTBR,1[m~2-m2-m32] and MTBR1m12'-m2-m3' ]
by peptide-m2' . ... 82

2-14 Comparison of extent and rate of tubulin
polymerization induced by wild-type MTBR16a,
mutant MTBR1n[m12-m2-m32], and mutant
MTBR16[m12'-m2-m3'] o. ................ 84

3-1 Schematic diagram of MAP-2 and four microtubule-
binding domain (MTBR) fragments . 102

3-2 Scatchard binding plots for MAP-2 microtubule-
binding fragment interactions with taxol-
stabilized microtubules in a cosedimentation
assay . . 104

3-3 Binding of MTBRn to taxol-stabilized MTs in
the presence of 2 AM tau protein . 107

3-4 Scatchard plot for the MT binding of
MTBRS1[mym3m3] mutant lacking the second
repeat sequence . . 110

3-5 Scatchard plots for binding of MTBRQ and a
MTBR162m12M2m I] mutant containing additional
copies of m2-like sequences in place of mI and m3 113

3-6 Sedimentation equilibrium behavior of the mutant
MTBR,[mi2m2my] and wild-type MTBR,1 115

4-1 Extent and rate of tubulin polymerization induced
by wild-type MTBR, mutant MTBR1mlm13mMy], and
mutant MTBRa[mym ] . 133

4-2 Nonidentical repeated sequences from fibrous,
heat-stable MAPs which have been tested as
synthetic peptides for their ability to promote
microtubule assembly . 136

4-3 Extent and rate of tubulin polymerization induced
by wild-type MTBR1, and mutants with second
repeat substitutions K4A, K8A, and K8D 138

4-4 Extent and rate of tubulin polymerization induced
by wild-type MTBR6. and mutants with second
repeat substitutions T2K, C5A, G6A, and G6E 140

4-5 Extent and rate of tubulin polymerization induced
by wild-type MTBRI, and mutants with second
repeat substitutions IllD and IllE 143

4-6 Extent and rate of tubulin polymerization induced
by wild-type MTBR,, and mutants with second
repeat substitutions P15V/G16I and
G16A/G17A/G18A . . 145

4-7 Extent and rate of tubulin polymerization induced
by wild-type MTBRI6 and mutants with second
repeat substitutions R12G/R14G, R12G, R14Q,
R14V, R14L and R14E . 148

5-1 A closed thermodynamic cycle for MT assembly
stimulated by the MT-binding region (MTBR) of
MAP-2 . . 163



Abs, absorbance

ATP, adenosine triphosphate

ADP, adenosine diphosphate

BSA, bovine serum albumin

Ci, Curie

CPM, counts per minute

DTT, dithiothreitol

GTP, guanosine triphosphate

GDP, guanosine diphosphate

EGTA, ethyleneglycol-bis-(b-amino-ethyl ether) N, N'-

tetraacetic acid

HPLC, high performance liquid chromatography

IF, intermediate filament

IPTG, isopropyl thiogalactopyranoside

MAP, microtubule-associated protein

MT, microtubule

kD, kilodalton

MTBR, microtubule-binding region

NMR, nuclear magnetic resonance

PAGE, polyacrylamide gel electrophoresis

PCR, polymerase chain reaction

PHF, paired-helical filament


PMSF, phenylmethylsulfonyl 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



Richard Lawrence Coffey

August, 1994

Chairperson: Daniel L. Purich
Major Department: Biochemistry and Molecular Biology

Microtubule-associated protein-2 (MAP-2) promotes

microtubule (MT) assembly and is thought to crosslink and

stabilize the neuronal microtubule cytoskeleton. A bovine

cDNA clone coding for the microtubule-binding region (MTBR)

of MAP-2 was sequenced, revealing remarkable preservation of

primary structure compared to rat, mouse, and human MAP-2.

The MTBR was expressed in bacteria and purified.

Initial characterization demonstrated interaction with MTs

by immunofluorescence microscopy. The recombinant MTBR was

indistinguishable from native MAP-2 in its ability to

promote MT polymerization and to lower the critical

concentration for MT assembly and bound to MTs with a

maximum stoichiometry of one per tubulin dimer. Circular

dichroism spectra were similar to published spectra for

whole MAP-2 and tau and indicated very little alpha or beta


structure. Analytical ultracentrifugation revealed that the

MTBR was monomeric in solution.

Mutagenesis experiments were designed to address the

presentation of the repeated sequences to the MT surface.

Previous experiments using synthetic peptides demonstrated

that only a second repeat peptide interacted with MTs. The

second repeat in the MTBR was altered such that an

additional copy of the third repeat was introduced in this

position. This mutation abolished the ability to promote

assembly and reduced the apparent affinity for MTs. Further

mutation to reintroduce the second repeat in the first

repeat position restored the assembly-promoting properties

to wild-type levels. These observations further suggested

that the second repeat was dominant with respect to the

first and third repeats in MT-binding and that the activity

of the second repeat was position independent. Wild-type

MTBR was also mutated in the first and third repeats such

that these sequences would more closely resemble the second

repeat. A modest increase in affinity was observed

suggesting an increase in the concentration of interacting

repeats rather than simultaneous interaction at all three


Sub-site interactions within the second repeat were

investigated with amino acid substitutions at 12 of the 18

residues in this sequence. Removing positive charge or

introducing negative charge had the most deleterious effects

on promoting MT-assembly, particularly at second repeat

residues K8 (1713), R12 (1716), and R14 (1718).



General Overview of the Cytoskeleton

The cytoskeleton is a complex and dynamic network of

protein filaments in eukaryotic cells. This network is

necessary for cells to assume and maintain specific shapes

and for the spatial organization and directed movement of

intracellular components. The cytoskeleton is responsible

for force generation in cell motility and organelle

transport, organizing cell anisometry and providing

structural rigidity. The three principal types of protein

filaments are microtubules, actin filaments, and

intermediate filaments. Various associated proteins

interact with these filaments to modulate their dynamic

properties, crosslink cytoskeletal components, produce

mechanical force, and attach other cellular structures to

the cytoskeleton.

Actin filaments, or microfilaments, are helical

polymers of actin seven nanometers in diameter that have

intrinsic polarity because the subunits are all oriented the

same way in the polymer lattice. Actin is an ATPase, and

ATP-bound actin polymerizes much more readily than ADP-bound

actin. ATP is hydrolyzed predominately in the polymer as


hydrolysis is much faster in polymeric actin than monomeric

actin (Carlier, 1989). Actin composes the thin filaments of

muscle fibers but is also abundant in (a) the cortical

regions of nonmuscle cells, (b) focal contacts, and (c)

specialized structures such as microvilli.

Intermediate filaments consist of a heterogeneous

population of proteins numbering thirty or more in each

mammalian species (Steinert and Roop, 1988). The

distinguishing feature of an intermediate filament protein

is a central alpha-helical "rod" domain of 310-356 amino

acids with variable amino-terminal and carboxyl-terminal

domains. The assembly of intermediate filaments involves

the formation of alpha-helical coiled-coils between

intermediate filament proteins and then the oligomerization

of these species into filaments 10-15 nanometers in

diameter. Intermediate filaments are abundant in the

cytoplasm of most eukaryotic cells as well as in the

nucleus, forming the inner surface of the nuclear envelope.


Microtubules are polymers, 24 nanometer in diameter, of

tubulin. Tubulin is found as a heterodimer of alpha- and

beta-tubulin proteins, which assembles with uniform polarity

into microtubules. Tubulin heterodimers are arranged in

linear strands called protofilaments, which extend

lengthwise along the microtubule. Most microtubules contain


either thirteen or fourteen protofilaments (McEwen and

Edelstein, 1980) leading to a hollow cylindrical structure.

Each tubulin heterodimer occupies a surface in the polymer

measuring 8 nanometers longitudinally and 5 nanometers in

width (Amos, 1979).

The intrinsic polarity of subunit addition to the

polymer leads to different assembly properties at the two

ends of a microtubule. This is due to the different types

of interactions between subunits when tubulin dimers bind to

the two different microtubule ends. The more favorable end

for subunit addition is termed the plus end while the less

favorable is the minus end. In vitro, pure tubulin has a

critical concentration for assembly of approximately 10 MM

of tubulin heterodimer (Lee and Timasheff, 1975). This

concentration represents a macroscopic constant for net

polymerization that depends on the microscopic behavior at

both the plus and minus ends. The polarity of a microtubule

can be determined by interaction with the microtubule-

associated protein, dynein (Haimo, 1982). When bound to

microtubules, the globular head domain of dynein tilts at an

angle of 55" in the direction of the plus end, a feature

that is easily seen by electron microscopy.

Microtubule organizing centers serve to nucleate

microtubules in cells. In most cells, the organizing center

is the centrosome near the nucleus (Brinkley, 1985). Thus,

the minus end is capped by the centrosome and microtubules

extend out radially with plus ends toward the plasma

membrane. The anisometry created by this arrangement

provides directional information for microtubule-motor

dependent transport. In specialized cell structures such as

axons, a uniform polarity is also seen where parallel

microtubules run the length of the axon with their plus ends

distal to the cell body.

Microtubules are dynamic polymers and many cellular

processes are dependent upon rapid shifts in the monomer-

polymer equilibrium. Microtubule stability can be modulated

in vivo by factors such as the abundance of microtubule

associated proteins and the rate of tubulin synthesis. In

vitro, the assembly/disassembly properties are termed

dynamic instability (Mitchison and Kirschner, 1984). Under

steady state conditions when the amount of polymer is

constant, there are two populations of microtubules, growing

and shrinking. Interconversions between these two states

are rare but do occur. A change from growing to shrinking

is known as a "catastrophe" while a change from shrinking to

growing is a "rescue." Since these changes are rare, most

microtubules either grow to become very long or depolymerize

completely. Over time, this situation leads to a reduction

in the number of microtubules but an increase in the average


Many functions of microtubules were discovered by using

drugs that target microtubules causing depolymerization,

such as colchicine, or enhanced stability, such as taxol.

Microtubules are critical to cell motility as they form a

crucial part of the mechanical structure in cilia and

flagella. The mitotic spindle is composed of microtubules,

and chromosome movement is microtubule dependent. Certain

types of vesicle transport are microtubule dependent, and

microtubules are involved in the endocytic pathway (Pierre

et al., 1992). Microtubules also serve to position cellular

structures in the cytoplasm such as the endoplasmic

reticulum and golgi apparatus. Moreover, microtubules are

critical to the formation and maintenance of specialized

cellular morphologies such as neurite processes (Heidemann

et al., 1986).


The tubulin heterodimer, the basic unit for microtubule

assembly, is composed of alpha- and beta-polypeptides.

These protein subunits are approximately 50 kD in molecular

weight and share 36-42% homology (Little and Seehaus, 1988).

Mammals have at least five alpha and six beta isotypes coded

by the tubulin gene superfamily (Villasant et al., 1986;

Wang et al., 1986; Lopata and Cleland, 1987). The

functional significance of having many tubulin isotypes

remains uncertain. However, in Aspergillus, Doshi et al.

(1991) show that disruption of one alpha gene blocks nuclear

division while disruption of a different alpha gene leads to


abnormal cell morphology but normal division. Different

isotypes may, therefore, play specific roles in the cell,

perhaps by interacting differently with MAPs or having

different assembly properties.

Another tubulin isotype termed gamma-tubulin has

recently been described (Zheng et al., 1991; Stearns et al.,

1991). Comparing the sequence of gamma-tubulin in

Aspergillus, yeast, Drosophila, and human reveals 66%

homology between species and 35% homology to both alpha- and

beta-tubulin. Gamma-tubulin is found in the centrosome, a

microtubule organizing center, and not in assembled

microtubules. Therefore, gamma-tubulin may serve as a

nucleation site for microtubule assembly as well as a minus

end capping protein.

In the heterodimer, alpha- and beta-tubulin bind 1 mole

of guanine nucleotide each. Weisenberg et al. (1968) first

showed that these sites are nonidentical. Alpha-tubulin

contains a nonexchangeable GTP-binding site, or N-site,

which traps GTP at the alpha-beta interface (Gaehlin and

Haley, 1977, 1979). Spiegelman et al. (1977) demonstrated

that there is no significant N-site exchange in vivo by

comparing simultaneously the half-lives of VP-labeled GTP

at the N-site and 3S-labeled tubulin. The N-site has also

been shown to contain deoxy-GTP in vivo in PC12 cells after

treatment with nerve growth factor (Angelastro and Purich,

1992). Beta-tubulin contains an exchangeable guanine

nucleotide binding site, the E-site. However, this site is

only exchangeable in the free tubulin dimer and not in

assembled microtubules (Weisenberg et al., 1976). GTP and

GDP bind at the E-site with dissociation constants of 2.2 x

108 M and 6.1 x 10-" M, respectively (Zeeberg and Caplow,

1979). Tubulin with GTP at the E-site polymerizes much more

readily than GDP-tubulin, and E-site GTP is hydrolyzed to

GDP during polymerization (Kobayashi, 1975). Carlier and

Panteloni (1981) demonstrated that there can be a lag in GTP

hydrolysis during assembly. This phenomenon could lead to a

GTP-tubulin boundary at the end of a polymerizing

microtubule, thought to confer stability against

depolymerization (Karr et al., 1979). A stochastic loss of

the GTP-tubulin cap of a growing MT could explain the

conversion of growing to shrinking microtubules in the

aforementioned dynamic instability model (Mitchison and

Kirschner, 1984). If a growing microtubule were to lose its

GTP cap, GDP-tubulin would be exposed at the end causing

rapid depolymerization. The free GDP-tubulin released could

undergo exchange to become GTP-tubulin and be incorporated

into existing microtubules leading to a bulk redistribution

in polymer lengths. The guanine nucleotide binding regions

in tubulin have been proposed based upon photoaffinity

labeling studies and sequence comparison with other GTP-

binding proteins (Linse and Mandelkow, 1988). Indeed, these

regions contain some of the most conserved residues in

tubulin (Luduena et al., 1992).

The extreme carboxyl-terminal regions of alpha- and

beta-tubulin are quite variable among tubulin isotypes.

However, these regions are very rich in glutamic acid and

are therefore highly acidic as shown in Figure 1-1. These

domains are thought to interact with various MAPs through

their basic microtubule binding regions (Littauer et al.,

1986), as discussed in the following section on MAPs. In

fact, removing these regions by subtilisin digestion greatly

reduces the critical concentration for MT assembly, thereby

abolishing the requirement for MAPs to promote assembly

(Serrano et al., 1984). Furthermore, Mejillano and Himes

(1991) neutralized exposed carboxylates in this region by

chemical modification and found MT-assembly was enhanced

while the interaction with MAPs was reduced. Thus, the

acidic carboxyl-terminus may serve as an internal inhibitor

of assembly, while basic regions of MAPs could abolish this

inhibition by effectively neutralizing the glutamic acid

side chains through ionic interactions.

Tubulin undergoes several forms of posttranslational

covalent modification. Piperno et al. (1987) showed that

lysine 40 in alpha-tubulin can be acetylated in vivo on the

e-amino group. Acetylated tubulin is associated with

Figure 1-1: Carboxyl-terminal acidic regions of mammalian
alpha and beta tubulins. The sequences represent the
extreme C-terminal glutamate rich portion of tubulin thought
to be involved in MAP-2 binding. The C-termini of all
mammalian tubulins in the PIR database are shown.

human alpha
pig alpha
rat alpha
hamster alphal
hamster alpha2
mouse alpha2

human beta
human beta5
rat betal5
mouse beta
mouse beta3
mouse beta4
mouse beta5
hamster betal6
hamster beta3
pig beta
















populations of microtubules with increased stability, but

the mechanism of stabilization is unknown. The neuron

specific beta(III) tubulin isotype is phosphorylated at

serine 444

and tyrosine 437 (Alexander et al., 1991) in the extreme

carboxyl-terminus. Serine 444 is phosphorylated to a

stoichiometry of one (Luduena et al., 1992) while the

stoichiometry at position 437 is low (Alexander et al.,

1991). The significance of this phosphorylation is unknown

although it adds negative charge to this already acidic

domain. Adding further negative charge to this portion of

beta(III) tubulin is the phenomenon of gamma-glutamylation

at glutamate 445 and glutamate 438 (Alexander et al., 1991).

Mass spectrometry was used to show that up to six glutamates

are added in an isopeptide bond to the gamma-carboxyl groups

at positions 445 and 438 leading to "forked" polypeptide

chains. This modification increases the anionic nature of

the carboxyl-terminal region although the significance is

not known. The final type of modification known to occur to

tubulin is tyrosination and detyrosination of alpha-tubulin.

Most tubulin isotypes are synthesized with a tyrosine at

their carboxyl-ends. Subsequently, this tyrosine can be

cleaved from tubulin by carboxypeptidase B and then

reintroduced by the enzyme tubulin-tyrosine ligase

(Thompson, 1982). The functional significance of tyrosine

removal and addition is unknown although detyrosinated


tubulin is associated with more stable microtubules in vivo

(Gundersen and Bulinski, 1986).

Tubulin and MAPs can be readily purified through

multiple cycles of assembly and disassembly with intervening

centrifugation steps (Weisenberg, 1972; Shelanski et al.,

1973). At 4*C, tubulin and MAPs do not assemble and

insoluble contaminants can be cleared by

ultracentrifugation. Upon warming to 37"C in the presence

of EGTA, Mge, and GTP, polymerization occurs and the

microtubules and MAPs can be pelleted to separate them from

soluble contaminants. Glycerol is included to enhance

microtubule assembly. MAPs and tubulin can be separated by

cation-exchange chromatography using phosphocellulose where

tubulin does not bind the resin while MAPs remain bound and

can later be eluted with 0.6 M KC1 (Weingarten et al.,

1975). In vitro, microtubule assembly is most easily

analyzed turbidimetrically using a spectrophotometer. Since

microtubules are essentially long rods, they will scatter

light in a manner proportional only to polymer mass and

independent of polymer length as long as the wavelength of

incident light is much smaller than the length of the

microtubules (Gaskin et al., 1974). Using light of 350 nm

wavelength therefore ensures that the observed absorbance

change is proportional to total polymer weight since

microtubules polymerized in vitro are of the order of 2 to 4

microns long. When determining the ability of a MAP to


stimulate MT-assembly, a concentration of free tubulin is

used that will not self-assemble in the absence of MAP.

Thus, when the MAP is added, the magnitude and rate of MAP-

induced assembly can be measured. To assess the affinity

and stoichiometry of MAP-binding to microtubules,

ultracentrifugation can be employed to separate free MAP

from MT-bound MAP (Joly and Purich, 1990). Microtubules can

first be polymerized in the presence of taxol to ensure that

all of the tubulin is incorporated into the polymer followed

by incubation with the MAP of interest. Because of the

tremendous difference in sedimentation properties between

MAPs and assembled MTs, centrifugation will separate the MT-

bound MAP from free MAP.

Microtubule-Associated Proteins


A large family of proteins are defined functionally as

MAPs based on their ability to interact with microtubules.

Such proteins include fibrous MAPs, which serve to stabilize

and crosslink microtubules, force-generating enzymes known

as motors proteins, and a variety of proteins found in

centrosomes and kinetochores. The family of known MAPs is

increasing rapidly, in part because taxol-stabilized

microtubules can be used as an affinity matrix to purify

MAPs from extracts of various tissues and species. One such

study recently identified twenty five MAPs in yeast (Barnes

et al., 1992).

Microtubule-Based Motor Proteins

Axonemal dyneins were the only proteins known to

generate force through interaction with microtubules for

many years. The development of video-enhanced differential

interference contrast microscopy allowed real-time

visualization of vesicle movement and individual

microtubules in a variety of systems (Allen et al., 1982,

1985). Subsequently, two classes of motor proteins have

been identified, cytoplasmic dynein and kinesin. Kinesin is

an ATPase composed of two heavy chains (110-135 kD) and two

light chains (60-75 kD) (Kuznetov et al., 1988; Bloom et

al., 1988). Electron microscopy revealed an overall

structure of two globular head domains followed by fibrous

rods that connect in a feathered tail domain. (Amos, 1987;

Hirokawa et al., 1989; Hisanga et al., 1989; Scholey et al.,

1989; Sheetz, 1989). Kinesin remains tightly bound to MTs

in the presence of the nonhydrolyzable ATP analog AMPPNP but

is released in the presence of ATP. ATP-dependent force

generation due to kinesin has been demonstrated in vitro as

gliding of MTs over glass surfaces, gliding of adjacent MTs

in solution, and transporting of anionic latex beads along

immobilized MTs (Vale et al., 1985; Scholey et al., 1985;

Cohn et al., 1987; Kachar et al., 1987; Porter et al., 1987;


Paschal and Vallee, 1987; Saxton et al., 1988; McCaffrey and

Vale, 1989; von Massow et al., 1989).

Cytoplasmic dynein is also a multi-subunit MT-dependent

ATPase. Heavy chains are thought to form two globular heads

(King and Whitman, 1989), while various light chains compose

fibrous stems and a base that connects the head regions

(Vallee et al., 1989). In vitro, dynein has been

demonstrated to cause the movement of glass beads and

vesicles along MTs as well as to promote gliding of MTs over

dynein coated glass slides (Lye et al., 1987; Paschal et

al., 1987; Amos, 1989; Schnapp and Reese, 1989).

Fibrous Microtubule-Associated Proteins

Among the best characterized nonmotor MAPs are MAP-1A,

MAP-1B, MAP-2, MAP-4, and tau. MAP-1A and 1B are high

molecular weight MAPs (>300 kD) abundant in brain and show

developmental regulation (Matus, 1990). MAP-1B (also known

as MAP-1.2, MAP-5, and MAP-X) is the earliest high molecular

weight MAP to appear during brain development suggesting a

role in neuronal differentiation (Bloom et al., 1988).

Indeed, Brugg et al. (1993) showed that blocking MAP1B

expression inhibits neurite outgrowth in cultured PC12


General features of MAP-2. MAP-4, and tau

MAP-2, MAP-4, and tau compose a distinct class of

fibrous MAPs possessing the hallmark features shown

schematically in Figure 1-2. These MAPs share a highly

Figure 1-2: Schematic diagram of the fibrous, heat-stable
MAPs MAP-2, MAP-4, and tau. The relative size and
organization are shown for each protein. The acidic N-
terminal projection arms are to the left and the basic C-
terminal MT-binding domains are to the right. The
projection arms are unrelated in sequence whereas the MT-
binding domains are highly conserved. The non-identical
repeated sequences are shown as vertical bars in the MT-
binding region (MAP-2 has three; MAP-4 has four; and tau has
either three or four).

Fibrous, Heat-Stable MAPs

projection arms *-


conserved MT-
- binding region

lji 200 kD

J118 kD

Tau 39-49 kD


conserved carboxyl-terminal microtubule binding domain

containing three or four imperfectly repeated 18 amino acid

sequences. Immediately preceding the microtubule binding

domains of these proteins are protease accessible regions

rich in proline. The amino-termini of these proteins,

although unrelated in primary structure, are termed

projection arms as they can be seen as projections from the

microtubule surface using electron microscopy. The

projection arms differ dramatically in ionic properties from

the microtubule binding domains, the latter being quite

basic while the former are acidic. Furthermore, these MAPs

are heat stable, maintaining solubility and the ability to

promote MT-assembly even after extended incubation at 1000C.

This class of MAPs is the focus of the remainder of this


MAP-2. MAP-2 migrates as a doublet of MAP-2a and MAP-

2b on SDS-polyacrylamide gels at approximately 300,000

molecular weight (Kim et al., 1979). Molecular cloning,

however, revealed that a single gene codes for MAP-2 protein

with a molecular weight of 200,000 in mouse (Lewis et al.,

1988). The basis of the heterogeneity is not known but the

anomalous migration is reminiscent of other fibrous MAPs.

The complete sequence for rat MAP-2 has also been determined

by Kindler et al. (1990) and shows striking conservation in

the microtubule binding region and the binding site for the

regulatory subunit of cAMP-dependent protein kinase at the


amino-terminus, with substantial divergence in the

intervening sequence. As shown in Figure 1-2, MAP-2 has an

amino-terminal projection arm that does not interact with

microtubules (Flynn et al., 1987) and a carboxyl-terminal

microtubule binding region that comprises approximately ten

percent of the molecule. A protease accessible hinge region

separates the MT-binding and projection domains (Vallee,

1980; Flynn et al., 1987; Aizawa et al., 1987). These

domains differ dramatically in their ionic properties with

isoelectric points of 10.5 and 4.8 for the projection arm

and MT-binding domain, respectively (Flynn et al., 1987).

MAP-2 shares a homologous microtubule binding domain with

tau and MAP-4 comprising the last 185 amino acids of each of

these proteins. There is 67% amino acid identity between

these regions in tau and MAP-2 (Lewis et al., 1988).

MAP-2 expression is neuron specific and the protein is

restricted to dendrites (Bernhardt and Matus, 1984; Miller

et al., 1982; Vallee, 1982; Lewis et al., 1986; Garner et

al., 1988; Okabe and Hirokawa, 1989). However, this

distribution holds true only in adult tissue. MAP-2c is a

70 kD product derived from the same gene through alternative

RNA splicing that lacks most of the internal portion of the

projection arm (see Figure 1-2). This form of MAP-2 is

found only in embryonic neurons, most abundantly in axons

(Rierder and Matus, 1985; Garner et al., 1988).

Tau protein. Tau describes a family of closely related

neuronal proteins with molecular weights of 55-65 kD as

judged by electrophoretic mobility (Cleveland et al., 1977).

Alternative splicing leads to at least a dozen cDNAs of

different sizes (Goedert et al., 1991; Lee, 1990). Further

electrophoretic heterogeneity results from various extents

of protein phosphorylation (Lindwall and Cole, 1984). Tau

proteins vary in their microtubule binding domain,

containing either three or four nonidentical repeats.

Various insertions also cause heterogeneity in the amino-

terminal portion. Recently, a larger tau isoform was

described at 110 kD molecular weight with a long insert near

the amino-terminus (Goedert et al., 1992; Couchie et al.,

1992). This isoform was found only in the peripheral

nervous system in both fetal and adult tissues, whereas the

smaller isoforms were found in both areas in fetal stages

but only in the central nervous system in the adult. In

certain neurodegenerative diseases, tau is found as the

major component of insoluble aggregates called paired-

helical filaments (PHF) (Lee et al., 1991). Abnormal

phosphorylation is thought to play a role in the formation

of PHFs (Kosik, 1990, 1991) by inhibiting tau's interaction

with microtubules and promoting assembly into a filamentous


MAP-4. This MAP is widely distributed unlike MAP-2 and

tau, which are neuron specific. West et al. (1991) assert


that a variety of MAPs that have apparent molecular weights

in the 200 kD range are all MAP-4. These proteins include

the HeLa cell 200 kD MAP, mouse MAP-4, rat 200 kD MAP,

bovine 190 kD MAP, and MAP-U. The function of this MAP is

not known, although MAP-4 appears to be associated with

interphase MTs and mitotic spindle fibers (Olmsted et al.,

1986). Cloning of bovine MAP-4 revealed a molecular weight

of 100 kD, far short of that estimated by SDS

electrophoresis (Aizawa et al., 1990). This study revealed

a carboxyl-terminal basic domain, homologous to the same

regions of MAP-2 and tau, consisting of a proline-rich

region followed by four nonidentical peptide repeats. When

proteolytically cleaved, this portion of MAP-4 stimulated

microtubule assembly while the amino-terminal portion did

not (Aizawa et al., 1987). The amino-terminal portion bears

no similarity to MAP-2 or tau and contains a unique 14 amino

acid motif that is repeated 18 times.


Ultrastructural and biophysical properties. Using

electron microscopy, MAP-2 is visualized as long, thin, and

flexible with a mean length of 147 nm, varying between 90

and 180 nm (Voter and Erickson, 1982; Gottlieb and Murphy,

1985). Hernandez et al. (1986) reported a frictional ratio

of 3.7 for MAP-2, consistent with either an elongated or a

flexible, expanded structure. In this same study,

analytical ultracentrifugation yielded a sedimentation


coefficient of 3.5S and a molecular weight of 220,000

indicating the majority of the protein was monomeric.

Consistent with being flexible and unordered, circular

dichroism spectra indicate less than 5% alpha or beta

structure in MAP-2. Interestingly, these structural

parameters were unchanged whether or not 100"C treatment was

included in the purification protocol (Hernandez et al.,

1986). When bound to microtubules assembled in vitro, MAP-2

projected out from the microtubule wall up to 80-90 nm

(Voter and Erickson, 1982). Hirokawa's group showed that

MAP-2 composes crossbridges between MTs of 50 to 120 nm in

length (Chen et al., 1992).

Interactions with other cellular components. The best

characterized binding partner for MAP-2 except for tubulin

is the regulatory subunit of cAMP-dependent protein kinase

(Vallee et al., 1981). This protein binds with high

affinity (a dissociation constant in the nanomolar range) to

MAP-2 at the amino-terminus between amino acids 83 and 113

(Rubino et al., 1989). This interaction has also been found

in vivo using immunohistochemical techniques (Miller et al.,

1982). Because of the two-fold axis of symmetry of this

dimer, the regulatory subunit could serve to mediate

crosslinking by binding two molecules of MAP-2 on adjacent

microtubules (Ainzstein and Purich, 1992).

MAP-2 has been shown to interact with other

cytoskeletal components, suggesting a role in linking


different parts of the cytoskeleton to microtubules. Using

in vitro assays, Flynn et al. (1987) showed that the MT-

binding region of MAP-2 binds neurofilaments, while

Satillaro (1986) demonstrated that this portion of MAP-2

interacts with actin filaments. Furthermore, calmodulin can

bind to the microtubule binding region (Lee and Wolff,

1984). Yamauchi and Purich (1987) found that

phosphatidylinositol interacts with MAP-2 and reduces

binding of this MAP to MTs.

Posttranslational modifications. MAP-2 can be readily

phosphorylated by a large number of protein kinases, mostly

on serine/threonine residues. At least thirty moles of

phosphate per MAP-2 can be incorporated in vitro when MAP-2

has been previously dephosphorylated (Tsuyama et al., 1986).

MAP-2 is also highly phosphorylated in vivo. In fact, there

may be two pools of MAP-2 in rat brain, one with 46

phosphoryls that does not associate with MTs and another

with 10-16 phosphoryls that is MT-bound (Tsuyama et al.,

1987). Increased phosphorylation correspondingly reduces

the ability of MAP-2 to promote tubulin polymerization in

vitro (Jameson and Caplow, 1981; Burns et al., 1984) and

will inhibit MAP-2 binding to preformed microtubules (Murthy

and Flavin, 1983). Interestingly, one study showed that

dephosphorylated MAP-2 bound less well to microtubules in

vivo when microinjected than did moderately phosphorylated

MAP-2 (Bruggs and Matus, 1991). However, heavily


phosphorylated MAP-2 did not associate with microtubules in

this study. These results indicate that phosphorylation at

certain sites may be required for optimal MT-binding in vivo

while further phosphorylation reduces binding. Recently,

Ainzstein and Purich (submitted) used site-directed

mutagenesis and mass spectrometry to identify three serine

residues involved in the regulation of MAP-2 binding to

microtubules. In these studies, protein kinase C was used

to phosphorylate a recombinant MAP-2 fragment in vitro that

corresponded to the microtubule binding region. Two serine

residues were rapidly modified. One serine was between the

first and second repeated sequences while the other was

within the second repeat. Subsequently, a serine between

the second and third repeats was phosphorylated, which

abolished MT binding as long as one of the first two series

was phosphorylated.

Interaction with microtubules. Vallee (1980) showed

that trypsin and chymotrypsin cleaved MAP-2, producing a 35

kD fragment that bound to MTs. Subsequently, Flynn et al.

(1987) used thrombin to produce an Mr 28 kD MT-binding

fragment that was localized to the carboxyl-terminus of MAP-

2. Similar studies in tau (Aizawa et al., 1988) and MAP-4

(Aizawa et al., 1989) led to the isolation of homologous MT-

binding regions at the carboxyl-termini of these proteins

(see Figure 1-2).

Figure 1-3: Non-identical 18 amino acid repeats found in
fibrous, heat-stable MAPs. The basic repeated sequences
found in the MT-binding regions of MAP-2, MAP-4, and tau are





MAP-2 (1)
MAP-4 (1)
Tau (1)

MAP-4 (2)
Tau (2)

MAP-2 (2)
MAP-4 (3)
Tau (3)

MAP-2 (3)
MAP-4 (4)
Tau (4)


cDNA sequencing of MAP-2, tau, and MAP-4 revealed

highly conserved, although nonidentical, 18 amino acid

repeats in the MT-binding region (Figure 1-3). Joly et al.

(1989) showed that a peptide corresponding to the second

repeat in MAP-2, m2, promotes MT-assembly while the first

and third repeats do not. Furthermore, this peptide is a

competitive inhibitor of MAP-2 binding to microtubules,

although its affinity is approximately two orders of

magnitude lower than the intact protein, which is thought to

be in the micromolar range (Joly and Purich, 1990). This

peptide also reduces microtubule dynamics in vitro,

indicating stabilization of the polymer (Yamauchi et al.,

1993). In MAP-4, a peptide corresponding to the first

repeated sequence promotes MT-assembly and binds with a

dissociation constant of 180 AM (Aizawa et al., 1989).

Ennulat et al. (1989) demonstrated that first and third

repeat peptides from tau (t, and t3) promote MT assembly

while t4 does not. Further analysis by Maccioni et al.

(1989) revealed t, and t3 binding to microtubules in the 20-

40 AM range while Melki et al. (1991) reported t3 binding in

the 100 AM range. A recent report implicates the inter-

repeat region between the first and second repeats in tau in

microtubule binding (Goode and Feinstein, 1994). These

authors used a short, basic peptide corresponding to this

region (KVQIINKKLD) and showed that it promotes MT-assembly

and competes with intact tau for MT-binding. This region in


MAP-2 is well conserved compared to the tau sequence and

therefore may also play a role in MT interaction.

Lewis et al. (1988) proposed that the spacing between

the repeated sequences may be large enough in an extended

configuration for each repeat to interact with an adjacent

tubulin subunit on a microtubule. This orientation could

provide stability to the polymer by anchoring subunits

together through the MAP. However, there are several lines

of evidence against this hypothesis. First, Yamauchi et al.

(1993) found that a peptide corresponding to a single MAP-2

repeat could stabilize microtubules in vitro, causing a

dramatic reduction in dynamics. Therefore, covalently

linked repeats do not seem to be necessary to stabilize

microtubules. Dingus et al. (1991) raised an antibody to

the third repeat in MAP-2, m3, and MAP-2 still promoted

assembly in the presence of this antibody. Furthermore, the

antibody bound to MAP-2, which was already bound to

microtubules, indicating that m3 does not tightly associate

with tubulin and is not critical for promoting assembly.

Wallis et al. (1993) examined the equilibrium binding

of MAP-2 to MTs and found a high degree of positive

cooperativity, consistent with their finding patches of MAP-

2 on microtubules by electron microscopy. However,

cooperativity has never been reported with respect to MAP-

2's ability to promote assembly. Furthermore, other

investigators have not observed cooperativity in direct

binding studies (Joly and Purich, 1990) or when observing

the affect of MAP-2 on the dynamics of individual

microtubules (Kowalski and Williams, 1993). Also, direct

binding studies with tau showed no evidence of cooperativity

(Butner and Kirschner, 1991; Biernat et al., 1993).

Cooperativity was not seen when the MT-binding properties of

synthetic peptide analogs of the repeats from tau and MAP-4

were analyzed (Melki et al., 1991; Maccioni et al., 1989;

Aizawa et al., 1989).

The stoichiometry of MAP-2, tau, and MAP-4 binding to

microtubules is of considerable interest as this could

provide insights into the presentation of the microtubule-

binding region to the polymer lattice. If three repeats

interacted simultaneously with adjacent tubulin subunits,

then stoichiometry should be limited to 1 MAP per three

tubulin molecules. A higher stoichiometry might be achieved

if only one repeat was interacting. However, the presumably

flexible and acidic projection arms of these MAPs may cause

steric hindrance to higher stoichiometries when bound along

the microtubule wall. Furthermore, it is not known if these

MAPs interact with alpha-tubulin, beta-tubulin, or both.

Biernat et al. (1993) reported binding of approximately 1

tau per 2 tubulin dimers while Joly and Purich (1990) showed

a stoichiometry of MAP-2 binding of one per 3 tubulin

dimers. Experiments showed that synthetic peptides

corresponding to one repeat sequence bound with higher

stoichiometries of 2 per tubulin dimer for t3 (Melki et al.,

1991), 1.9 and 1.7 for t, and t3, respectively (Maccioni et

al., 1989), and 1.2 for the first repeat in MAP-4 (Aizawa et

al., 1989).

MAP-2 is thought to bind to the variable acidic region

of tubulin composing the carboxyl-terminal 15-20 residues

(see Figure 1-1). Serrano et al. (1984) found that

selective removal of this region abolished the requirement

for MAP-2 in promoting MT assembly and resulted in a loss

binding of MAP-2 to microtubules. It is not clear if these

sequences in both alpha- and beta-tubulin are involved.

Littauer et al. (1986) showed preferential binding to beta-

tubulin peptides. However, Paschal et al. (1989) assert

that the sequence EGEE, present in most alpha- and beta-

tubulins is the site of interaction. Cross et al. (1991)

show that beta-tubulin peptides containing this sequence as

well as peptides corresponding to adjacent beta-tubulin

acidic sequences interact with MAP-2. However, they do not

address the interaction of alpha-tubulin with MAP-2.

Providing evidence of MAP-alpha-tubulin binding, Melki et

al. (1991) show that a tau peptide corresponding to the

first repeat interacts with both alpha- and beta-tubulin.

Direct observation of the effects of MAPs on the

dynamics of single microtubules can be seen using

differential interference contrast microscopy. As described

above, microtubules exhibit dynamic instability where

individual MTs are either growing or shrinking with

infrequent transitions between these two states (Mitchison

and Kirschner, 1984). MAP-2 appears to affect the dynamics

of single MTs in several ways leading to the net growth

observed by light-scattering in bulk solution (Kowalski and

Williams, 1993; Pryer et al., 1992). The rate and duration

of MT growth phases were increased by the presence of MAP-2.

Furthermore, MAP-2 caused a decrease in shortening rates and

reduced the number of growing to shrinking transitions

(catastrophes) while increasing the shortening to growing

transitions (rescues). Dreschel et al. (1992) found that

tau had similar effects on MT dynamics. Interestingly, they

also showed that in vitro phosphorylation of tau reduced its

ability to suppress growing to shrinking transitions.

MAP-2 and cellular morphology. Several recent studies

have shed light on the role of MAP-2 in morphogenesis either

by introducing the protein into cells where it is not

normally found or inhibiting its synthesis in neurons.

Dinsmore and Solomon (1991) reduced MAP-2 protein levels by

antisense RNA expression in embryonal carcinoma cells

undergoing neuronal differentiation. The lack of MAP-2

synthesis in these cells caused a failure to produce neurite

extensions and a failure to withdraw from the cell cycle.

Cultured cerebellar macroneurons treated with MAP-2

antisense oligonucleotides also failed to produce neurites

and later became rounded (Caceres et al., 1992). MAP-2

transfection into non-neuronal cells, including HeLa, CHO,

and NIH 3T3 induced the dramatic formation of microtubule

bundles which did not originate at microtubule organizing

centers (Lewis et al., 1989). Also, MAP-2c expression in

Sf9 cells caused these normally rounded cells to extend long

processes containing densely bundled MTs (LeClerq et al.,

1993). Taken together, these findings are consistent with

MAP-2's postulated role in neuronal differentiation through

stabilization and crosslinking of the microtubule



The research described in this dissertation focused on

the interaction of the microtubule-binding region of MAP-2

(MTBR) with microtubules. The primary structure was deduced

from the cDNA fragment coding for the microtubule-binding

region of bovine MAP-2. A scheme for bacterial expression

and purification of this portion of MAP-2 was developed.

The recombinant product was characterized for its assembly

promoting and microtubule binding properties. Site directed

mutagenesis was used to alter the repeated sequences in the

MTBR such that the order and number of copies of individual

repeats were changed. The effects of these mutations on

affinity, stoichiometry, and assembly promoting properties

were assessed. The importance of various amino acid side


chains in the second repeat for promoting MT assembly was

investigated using site directed mutagenesis.



Microtubule-associated protein-2, a cytomatrix

component of neuronal dendritic processes, interacts with

microtubules (or MTs) and is thought to stabilize the

neuronal cytoskeleton by forming cross-links between

microtubules (Olmsted, 1986). A single MAP-2 gene appears

to serve as a precursor for several different protein

species that arise from alternative splicing of a common

MAP-2 mRNA precursor. MAP-2ab are high-molecular-weight

forms appearing in dendrites of adult neurons

(Papandrikopoulou et al., 1989), and MAP-2c is a low-

molecular-weight counterpart containing the same extreme

amino- and carboxyl-termini found in the adult forms. MAP-

2ab can be readily cleaved by serine proteases (Vallee and

Borisy, 1977) into a N-terminal projection-arm domain and a

C-terminal microtubule-binding region, the latter designated

here as MTBR. Unlike chymotrypsin and trypsin, however, the

arginine-specific protease thrombin produces stable limit

digestion products after cleavage at the protease-accessible

hinge region of MAP-2 (Flynn et al., 1987). The MT-binding



fragment is quite basic with an isoelectric point of about

10.3, and the projection-arm is acidic with a corresponding

isoelectric point of approximately 4.8 (Flynn et al., 1987).

The 203-amino acid residue MT-binding fragment displays the

same apparent affinity as MAP-2ab for microtubules (Joly et

al., 1989). This region includes a triad of regularly

spaced nonidentical octadecapeptide repeats (Lewis et al.,

1988), and peptide analogues corresponding to the second-

repeated sequence promote tubulin polymerization (Joly et

al., 1989) and displace MAP-2 from assembled microtubules

(Joly and Purich, 1990).

Most of the biochemical characterization of MAP-2 has

been achieved through characterization of the heat-stable

bovine or porcine protein, while the amino acid sequence was

deduced using mouse or rat MAP-2 cDNA (Lewis et al., 1988;

Kindler et al., 1990). MAP-2 phosphorylation is known to

alter MT assembly properties (Jameson and Caplow, 1980), and

recent studies have shown that phosphorylation enhances

dynamic instability of brain microtubules in vitro

(Raffaelli et al., 1992). My interest in analyzing MAP

interactions with microtubules through site-directed

mutagenesis led me to sequence the corresponding 1.1

kilobase fragment of bovine brain MAP-2 cDNA previously

isolated in our laboratory (John Joly, personal

communication). I also developed an efficient bacterial

expression system and purification method to isolate this


microtubule-binding region of MAP-2, which provide for

further characterization of the microtubule-binding

properties by using site-directed mutagenesis. This allowed

me to obtain new insights regarding the interactions of the

nonidentical peptide repeats with microtubules.

Materials and Methods

[a-"P]dATP, SDS, and sucrose were ICN products.

Amersham, Inc. supplied the [a-35S]dATP and their

"Oligonucleotide-Directed in vitro Mutagenesis System"

(Version 2.1). Reverse transcriptase was from Promega, and

Klenow and T4 DNA ligase were from New England Biolabs.

Restriction endonucleases were purchased from New England

Biolabs or Promega. The bovine brain cDNA library was

obtained from Clonetech, Inc., and the Sequenase (Version

2.0) DNA Sequencing Kit was a U.S. Biochemicals product.

Expression vector pETh-3b, a derivative of pET-3b described

by Studier et al. (1990), was a gift from Dr. Donald R.

McCarty of the Department of Vegetable Crops, University of

Florida. Chloramphenicol and acetate kinase were purchased

from Boehringer-Mannheim. Ampicillin, isopropyl-p-

thiogalactopyranoside (IPTG), 2-[N-Morpholino]ethanesulfonic

acid (MES), Piperazine-N,N'-bis[2-ethanesulfonic acid]

(PIPES), EGTA, MgC12, dithiothreitol, protease inhibitors,

lysozyme, tris-HCl, trifluoroacetic acid and GTP were from

Sigma. Phosphocellulose-P11 was a Whatman product. Taxol


was obtained from Dr. Matthew Suffness of the National

Cancer Institute, Bethesda, MD.

The mouse cDNA sequence data of Lewis et al. (1988)

served as the basis for preparing oligonucleotide polymerase

chain reaction (PCR) primers that spanned the mouse MAP-2

amino acid sequences 5359-5388 (sense-primer) and 5643-5665

(antisense-primer). Bovine brain RNA was isolated and

transcribed into cDNA with reverse transcriptase using

random hexamer primers. The polymerase chain reaction was

then carried out and the products were fractionated on an

agarose gel. After thirty cycles of amplification, bands

were observable at positions corresponding to 300 base

pairs. These bands were electroeluted and treated with

mung-bean nuclease to remove single-stranded ends, then

ligated into the Sma I site of M13mpl9 and sequenced. A

MAP-2 clone was identified and subsequently radiolabeled for

use as a probe in the screening of a lambda gtll bovine

brain cDNA library. Approximately three hundred thousand

plaques were initially screened using duplicate filter

lifts. Positives were picked and plated; two additional

rounds were then performed, and two sets of clones were

plaque purified. One clone contained a 1.1 kilobase insert

after a lambda liquid lysate preparation and EcoR I

digestion. This 1.1 kilobase fragment was introduced into

pUC19 for subcloning and sequencing. The resulting plasmid,

pJJl, was grown in large quantity, digested with EcoR I, and


the 1.1 kilobase insert purified from an agarose gel.

Several subclones were generated by digestion with Hae III

and Alu I and subsequently sequenced. This approach allowed

the location of two unique internal restriction sites (BamH

I and Sph I) from which three fragments spanning the entire

1.1 kilobase MT-binding region were generated. I subcloned

these fragments into both M13mpl8 and M13mpl9 and sequenced

them in both orientations using the Sequenase 2.1 dideoxy-

sequencing kit.

I digested plasmid pJJ1 with Sac I and Xba I which

served to remove a fragment between a Sac I site in the 3'-

untranslated region of the clone and an Xba I site in the

pUC19 polylinker. A short synthetic double-stranded

oligonucleotide with Sac I and Xba I ends was ligated into

the vector to recreate a circular double-stranded plasmid

lacking the EcoR I site at the 3' end of the MAP-2 clone.

The resulting plasmid was digested with EcoR I (at the 5'

end of the clone) and Hinc II (adjacent to the Xba I site in

the pUC19 polylinker) to liberate the 1.1 kilobase MAP-2

clone. This was then ligated into the EcoR I and EcoR V

sites in the polylinker of the expression vector pETh-3b

(McCarty et al., 1991 ) which had been previously digested

with Sph I and Pvu II, the ends made blunt with T4 DNA

polymerase, and ligated. Removal of this Sph I-Pvu II

fragment destroyed the Sph I site and removed two PpuM I

sites so that these sites in the MAP-2 cDNA fragment could


be subsequently employed. The resulting plasmid, pRC10,

allows for the expression of the entire 323 amino acid C-

terminal fragment of MAP-2 fused to the first eleven amino

acids of the T7-genel0 product. This protein is termed

MTBR15, as it corresponds to the carboxyl-terminus of MAP-2

starting at position 1509 (in the mouse sequence) and

contains the T7-leader peptide. A plasmid expressing a

shorter product beginning at position 1629 (in the mouse

sequence) was generated by digestion of pRC10 with EcoR I

(at the beginning of the MAP-2 sequence) and PpuM I (at the

thrombin cleavage site in the corresponding MAP-2 amino acid

sequence) and lighting a short synthetic oligonucleotide.

The resulting plasmid, pRC9, expressed the 203 amino acid

carboxyl-terminal portion of MAP-2 (defined as the thrombin

cleavage product by Joly et al. (1989)) with the T7-epitope

sequence at the N-terminus and is termed MTBR1,6. Plasmid

pRC8 expressing the C-terminal 203-amino acid microtubule

binding domain without the T7-leader sequence was prepared

by digestion with PpuM I (at the thrombin cleavage site in

the MAP-2 amino acid sequence) and Nhe I (adjacent to the

start codon), rendering the ends blunt with Klenow, followed

by ligation. The resulting sequence was in-frame and coded

for the expression of the 203 amino acid C-terminal MAP-2

fragment preceded by the sequence Met-Ala-Arg (the arginine

is found in the native MAP-2 sequence).


Plasmids expressing the MTBR,69[m12-2-m-3], MTBRI629[ml-m2-

m32], and MTBRi629[m12-2-m]3 mutants where the first and third

repeats were made to resemble the second repeat were created

using the Amersham kit for oligonucleotide-directed

mutagenesis. A BamH I-Sph I fragment from the MAP-2 cDNA

(containing the m2, m3, and part of the m, repeated

sequences) was ligated into the polylinker of M13mpl9 so

that a single stranded DNA template could be produced.

Synthetic oligonucleotides complementary to the m, and m3

regions but containing the desired substitutions were used

to carry out the mutagenesis. The m, mutant oligonucleotide


introduced a unique BsrF I site. The m3 mutant

oligonucleotide ( 5'-TCGCTTAAAAATGCGCGCCACCGTCCC) allowed

the introduction of a unique BssH II site. Putative mutants

were screened for the presence of the new restriction sites

and positives were sequenced to insure that only the desired

changes were made. The mutated fragments were then ligated

back into the expression vector, pRC8, previously digested

with BamH I and Sph I. The resultant vectors (pRC11 coding

for MTBR,2[m12-m2-m3]; pRC12 coding for MTBR1629[ml-m-m32]; and

pRC13 coding for MTBRg9[mi2--m2-32]) were screened for the

presence of the new restriction sites before expression and

purification of the mutant proteins. Addition changes were

made to pRC13 by oligonucleotide directed mutagenesis to

include the sequence RVK, which is present after the second


repeat, after the m12 and m2 repeats. The resulting

plasmid, pRC22, encodes a further modified MAP-2 fragment

termed MTBR,6[m12'-m2-m32']. Plasmid pRC22 was later

modified by PRC mutagenesis in the regions coding for the

first and third repeats to render the amino acid sequence of

these repeats identical to the second repeat, generating

pRC24. The protein product encoded by this plasmid was

termed MTBRI[m2-m2-m2]. Plasmid pRC34, coding for an MT-

binding domain with an extra identical copy of the second

repeat in the third repeat position (termed MTBR,,[ml-m2-

m2]) was constructed by introducing an Afl II Pst I

fragment from pRC24, containing the second and third

repeats, into pRC8. Plasmid pRC35, coding for an MT-binding

region with an extra identical copy of the second repeat in

the first repeat position (termed MTBR1,[m2-m2-m3]) was

similarly constructed by introducing the Afl II Pst I

fragment from pRC8 into pRC24. All constructs were

sequenced prior to expression.

Expression plasmids were transformed into E. coli

BL21(DE3) and grown on LB plates containing 50 Ag/ml

ampicillin and 50 Ag/ml chloramphenicol. A single colony

was used to inoculate a 5 ml LB culture containing 50 Ag/ml

ampicillin and 50 Ag/ml chloramphenicol and grown overnight

at 37C. This culture was diluted into 500 ml LB containing

50 Ag/ml ampicillin and 50 sg/ml chloramphenicol and grown

at 37'C to an optical density of 0.4 (600 nm). IPTG was


added at a concentration of 1 mM and growth continued for

two hours. Cells were harvested by centrifugation at 5000 x

g for 5 min, resuspended in 20 ml buffer (80 mM PIPES, 1 mM

EGTA, and 1 mM MgCl2, pH 6.8). They were pelleted again at

5000 x g for 5 min and resuspended in 4 ml of a buffer

containing 100 mM Tris-HCl pH 8.0, 50 mM NaCi, 1.0 mM MgCl2,

1.0 mM dithiothreitol, 40 Ag/ml DNAse, 40 Ag/ml RNAse, 100

Ag/ml N-tosyl-L-phenylalanine chloromethyl ketone, 100 Ag/ml

N-a-p-tosyl-L-lysine chloromethyl ketone, and 0.1 mM

phenylmethylsulfonyl fluoride. This suspension was

subjected to two passes through a French pressure cell at

950 atm. Lysozyme was added to 1 mg/ml, and the solution

was incubated at room temperature for 15 min with occasional

stirring. This fluid was then layered onto 12 ml 25%

sucrose (w/v) containing 100 mM Tris-HCl (pH8.0) and

centrifuged at 14000 x g for 20 min at 20"C. The resulting

pellet containing insoluble inclusion bodies was rinsed with

a 50 mM Tris-HC1 buffer (pH 8.0) and resuspended using a

Dounce homogenizer in 1 ml 50 mM Tris-HCl (pH 8.0) buffer,

containing 200 mM NaCI and 1 mM dithiothreitol. This

suspension was then rapidly heated to a final temperature of

80"C and held for 10 min to solubilize the heat-stable MAP-2

microtubule-binding fragment. Subsequent incubation on ice

for 20 min and centrifugation at 400,000 x g for 6 min

served to remove most of the aggregated contaminating

proteins. The supernatant predominantly contained intact


MAP-2 MT-binding fragment, and this was passed through a

0.22 pm filter and subjected to reverse-phase chromatography

using a Brownlee large-bore C18 reverse-phase column. A

linear, two-step gradient from pure H20 to 36% (v/v)

acetonitrile in water after 45 min, and to 50% (v/v)

acetonitrile in water after 75 min caused the pure product

to elute at approximately 52 min after the gradient

commenced. (Trifluoroacetic acid (0.1% v/v) was present

throughout the gradient.) Fractions containing pure MAP-2

microtubule-binding domain were pooled and dried in a Savant

Speedvac concentrator, with subsequent resuspension in 1 ml

MEM buffer (100 mM MES, 1 mM EGTA, 1 mM MgCl2, pH 6.8). The

solution was aliquoted, frozen in liquid nitrogen, and

stored at -70*C. Immediately before use, the recombinant

MTBR was rapidly thawed in a 37'C water bath and clarified

by centrifugation at 100,000 x g for 30 min at 37"C.

Bovine brain microtubule protein (tubulin and MAPs) was

purified by cycles of assembly and disassembly with

intervening centrifugation according to the standard

procedure of Shelanski et al. (1973). Tubulin was purified

to homogeneity by the method of Kristofferson et al. (1986)

using phosphocellulose chromatography followed by an

additional assembly/disassembly cycle to remove denatured

protein. Purified tubulin was stored at -70"C until use.

Standard turbidimetric techniques were used to monitor

the assembly process (Gaskin et al., 1974). First, all


components except tubulin were mixed at final concentrations

of 100 mM MES (pH 6.8), 1 mM EGTA, 60 mM KC1, 1 mM MgCl2, 1

mM DTT, 1 mM GTP, and the indicated amount of MAP, and kept

on ice. Immediately before the assay, phosphocellulose

purified tubulin was thawed and added to the cold mixture at

a concentration of 11 pM. This solution was rapidly mixed

by gentle pipetting and quickly transferred to a cuvette

pre-warmed to 37*C in a Cary 210 spectrophotometer so that a

baseline value could be obtained. The assembly process,

maintained at 37*C, was monitored at 350 nm and recorded by

a chart recorder. The extent of assembly was measured as

the total absorbance change between the initial value and

the value after a plateau was reached. The rate of assembly

was determined from the slope of the early linear portion of

the curve.

To assess changes in the apparent affinity of tubulin

for microtubules in the absence or presence of the

recombinant MT-binding fragment, critical concentration

determinations were made as follows. Tubulin was assembled

at 20 MM with 3.4 pM recombinant MT-binding fragment, as

described above except that a GTP regenerating system was

used consisting of 2 units/ml acetate kinase and 10 mM

acetyl phosphate. Samples were allowed to reach a plateau

level of polymerization, and the apparent absorbance at 350

nm was recorded. Tubulin was serially diluted to 15, 10,

and 5 AM in the presence of a constant level of MT-binding


domain (3.4 AM). The absorbance was recorded at each

dilution point only after steady-state polymerization was


Microtubule assembly was carried out with tubulin and

"epitope-tagged" MT-binding domain containing an additional

eleven amino acids from the T7-genel0 product at the N-

terminus (MTBRI,,). Tubulin (10 AM) and epitope-tagged

binding domain (10 AM) were incubated in PEM buffer (80 mM

PIPES, 1 mM EGTA, 1 mM MgCl2, pH 6.8) containing 1 mM GTP at

37*C for 15 minutes. Microtubules were fixed with

glutaraldehyde, and prepared for immunofluorescence

microscopy as described by Kristofferson et al. (1986). All

antibodies were used at a 1:50 dilution. Rabbit antiserum

directed against the T7 peptide fused to the MT-binding

domain was used first, followed by Texas Red-conjugated

anti-rabbit IgG. Then, a mouse monoclonal beta-tubulin

antibody was added, followed by fluorescein-conjugated anti-

mouse IgG. Microtubules were visualized using rhodamine and

fluorescein filters on a Zeiss Axiophot microscope.

CD spectra were collected from 180 to 260 nm at room

temperature using a Jasco J500C spectropolarimeter. The

instrument was calibrated with camphor sulphonic acid at 290

nm. A concentration of 45 MM MTBR1, was used in 10 mM

sodium phosphate at pH 6.9. The average of ten spectra is



To investigate displacement of the recombinant MT-

binding domain by repeated sequence peptides, I used a

competitive binding assay described by Joly et al. (1990).

Tubulin (50 AM) was permitted to self-assemble at 37*C for

20 min followed by the addition of 50 AM taxol for an

additional 10 min at 37*C. Microtubules were diluted to 2.5

AM in PEM buffer with 1.0 mM GTP, 10 pM taxol, 1.0 mM

dithiothreitol, 10 MM recombinant MT-binding fragment, and

the indicated amounts of competitor peptide. After 20 min

incubation at 37*C, samples were centrifuged in a Beckman

Airfuge through a 20% (w/v) sucrose cushion at 100,000 x g

for 30 min at 37*C. Pellets and supernatants were analyzed

by electrophoresis according to the procedure of Laemmli

(1970) and visualized after Coomassie staining. In Figures

2-11 and 2-13, the amount of MTBR in the pellet fractions

was determined by densitometry on the Coomassie stained


Protein concentrations were determined by the method of

Bradford (1976) with bovine serum albumin used as the

protein standard. SDS-polyacrylamide gel electrophoresis

was performed according to the method of Laemmli (1970).


To gain further insight concerning the sub-site interactions

of the three nonidentical repeated sequences and possibly

other regions within the bovine brain MAP-2 microtubule-

binding domain, molecular cloning techniques were utilized


to isolate a 1.1 kilobase region corresponding to the C-

terminus of MAP-2. For this purpose, a bovine brain cDNA

library was screened with a bovine probe obtained by PCR

amplification of bovine brain cDNA using oligonucleotide

primers based on the mouse brain MAP-2 cDNA sequence. With

this approach, sequence artifacts arising from the PCR

technique were avoided and hybridization was maximized by

using the bovine cDNA probe with a bovine library. Because

the cDNA library had an average insert size of 1-2

kilobases, and because the carboxyl-terminal microtubule-

binding region of MAP-2 only spans 600-700 base pairs, this

strategy promised to provide an intact cDNA fragment

corresponding to the entire MT-binding region. Indeed, as

shown in Figure 2-1, dideoxy sequencing confirmed that this

was the case, and the fragment contained the C-terminal

binding domain beginning at a position corresponding to

amino acid 1509 in the murine sequence and extending past

the stop codon into the 3' untranslated region. Joly et al.

(1989) previously demonstrated that a thrombin cleavage

fragment of MAP-2 fully retains the tubule-binding

properties of intact MAP-2, and I can now assign this

sequence as that spanning the region from residues 1629 to

Figure 2-1: General structural organization of brain MAP-2,
including nucleotide and deduced amino acid sequence of the
bovine microtubule-binding region (MTBR). The schematic
above includes the N-terminal cAMP protein kinase regulatory
subunit binding site, the C-terminal microtubule-binding
region, and the three imperfectly repeated sequences. The
latter are indicated in the schematic as vertical bars and
in the sequence as boxed areas. The mouse amino acid
sequence (top row) of the microtubule-binding region serves
as a means to align the corresponding bovine MAP-2 sequence
(middle row). Any exceptions to exact amino acid identity
are indicated in dashed boxes. The corresponding bovine
MAP-2 cDNA sequence is indicated on the bottom row.
Confirming gas-phase amino acid sequence data for positions
1526-1544 and 1629-1645, obtained with chymotryptic (Dingus
et al., 1991) and thrombin (Joly et al., 1989) cleavage
products are hyphenated in the sequence.


Projection tMT-Binding
Armnn Region

Absent in Embryonic MAP-2c -
I R, 147 1519 182i
Binding Sit _________ Thnranbin

1S09 K K T T A A (Ir-L" A Q A P r5A- F K Q A K D K V M1 D 0 fT-l K
R K T T A A IG A 1 SI A Q A P IS VI F K--Q--A-- K--D--K--V-tS -D--G-(V--*K--
1539 S P K R S S L PR P S S I L PP t V S D Ri lt I 8
S--P.----K--R--S S L P a P S S I L P P R R 0 V S 0 D t Il N S P
1S69 S L N S S I S A R T T R S P 1 R t A 0 K S 0 T S T P T
1599 T P 0 S T A I T P 0 T P P S Y S S R T P 0 T P 0 T P S Y P R
1629 T P O T P K S 3 I L V P S K K V A II R T P P K S P A T P
*T--P G--T--P--K--S-4tW-I--L-- V--P--S--B--K--K V A I I R T P P K S P A T P

1659KQ L R LI N Q P L PDL K N E K S K I GS T 1 K Q P
169 K 0 Q V Q I V T K K I D L H V T K CO L K IXK fR
171 o 5 G R V K I I S V K L D IK S K A Q A K L oV L D 5NA Hm
1749 H V P 0 G G N V K I D S Q K L N F R 5 H A K A R V D H 0 A 3
1779 I I T Q S P ISI R S S V A S P R R L S N V S S S L S I H L 8
1009 S P Q L A T L A E D V T A A L A R Q 0 L (stop)
S P Q L A T L A 3 D V T A A L A K Q G L (stop)


1828 of the mouse sequence. The only differences in the

bovine sequence compared to the murine are three additional

amino acids in the bovine sequence (1631-1633), an alanine-

for-glycine substitution at 1639, and a glycine-for-serine

change at position 1788. The additional HTP sequence near

the thrombin cleavage site was also confirmed by Edman

microsequencing (Joly et al., 1989). Dingus et al. (1991)

also carried out Edman analysis of a chymotrypsin cleavage

fragment of bovine MAP-2, and their fragment can now be

definitively localized at positions corresponding to amino

acid 1526 to 1544. Significantly, the three nonidentical

repeated sequences which are boxed in Figure 2-1 are

identical in bovine and murine sequences.

To achieve efficient expression of the microtubule-

binding domain of MAP-2 in E. coli, I utilized the pETh-3b

expression vector. This plasmid has the advantage that

protein expression is under the control of the T7 promoter,

and expression is carried out in the E. coli variant BL21

(DE3) which contains an integrated copy of the T7 RNA

polymerase gene under control of the lacUV5 promoter

(Studier et al., 1990). Accordingly, expression commences

upon the addition of IPTG to the bacterial culture medium.

As shown in Figure 2-2, efficient biosynthesis of the MT-

binding fragment of MAP-2 is achieved (see lane 1), and this

component can be readily purified to homogeneity in three

steps. Lane 2 illustrates the state of purification

Figure 2-2: SDS gel electrophoretic analysis of bacterially-
expressed MTBRlA, at various stages of isolation: cell-free
extract (lane 1); insoluble fraction collected by
centrifugation through a sucrose cushion and resuspended by
homogenization (lane 2); protein after heat treatment and
clarification by ultracentrifugation (lane 3); and, MTBR1a
after reverse-phase HPLC fractionation (lane 4). (Note:
Molecular weight standards are shown at left.)






MW 1 2 3 4

Figure 2-3: HPLC elution profile of the recombinant MAP-2
MTBR subjected to C, reverse phase chromatography. MTBR,
eluted at 52 min un er the conditions described in the text,
with minor protein contaminants trailing as indicated.


20 40 60


Time (min)






obtained upon centrifugation through a sucrose cushion to

obtain presumptive protein inclusion bodies rich in the MT-

binding fragment. Lane 3 shows the level of purity after

heat-treatment at 80C for 10 min. At this point, the

microtubule-binding fragment became soluble, and lane 4 of

Figure 2-2 demonstrates that I have achieved homogeneity

using HPLC separation (Figure 2-3) with a reverse-phase C18


Sloboda et al. (1976) demonstrated that MAP-2 readily

promotes tubulin self-assembly with half-maximal stimulation

occurring in the 2-5 MM range, and later work with thrombin

cleavage to produce the microtubule-binding fragment of MAP-

2 (Joly et al., 1989) showed that this fragment and intact

high-molecular-weight MAP-2 bound to microtubules with

virtually identical affinity (Kd = 2-5 MM). The data shown

in Figure 2-4 indicate that the bacterially expressed

microtubule-binding fragment MTBR162 stimulates tubulin

polymerization. Indeed, while self-assembly of pure tubulin

(10 MM) proceeded very slowly under these conditions, the

polymerization rate was increased by more than a factor of

50 at 2.4 MM MTBRa,, and the extent of polymerization was

also greatly enhanced. As shown in Figure 2-5a, the initial

plateau values for polymerization depended upon the

concentration of MTBR1, in a linear fashion. In companion

experiments (Figure 2-5b) the tubulin concentration was

varied in the presence of the bacterial expression product

Figure 2-4: Microtubule assembly induced by the recombinant
MAP-2 microtubule-binding region. Assemblies were carried
out with 10 AM tubulin at 37 C and followed by turbidity at
350 nm. The recombinant MAP-2 fragment was at
concentrations ranging from 0 to 2.4 AM.


A. 0.2


0 3 6 9

12 15

Figure 2-5: Extent of tubulin polymerization in the presence
of bacterially expressed MAP-2 fragments. (A) Plot of the
overall amplitude of microtubule assembly data (from Fig. 3)
versus the concentration of microtubule-binding domain. (B)
Critical concentration behavior of microtubules assembled
with MAP-2 MTBR. Samples of assembled microtubules were
subjected to various extents of dilution in the presence of
the recombinant protein, and turbidity measurements were
made to estimate the amount of microtubule polymer remaining
assembled. 20 AM Tubulin (2.0 mg/ml) was assembled by
addition of MAP-2 fragment maintained at 3.4 AM, and
dilutions were made as indicated while holding the binding
domain concentration at 3.4 pM.

[MT-binding domain], /M

0.5 1.0







to obtain estimates of the critical tubulin concentration

required for microtubule polymerization. I found that

microtubules polymerized in the presence of MTBR1,

exhibited a critical concentration of 0.2 mg/ml tubulin. It

should be noted that 3.5 AM levels of the recombinant

protein were used in these experiments, and such

concentrations would correspond to that observed in 2.5

mg/ml recycled microtubule protein (i.e., crude tubulin and


As an additional test of the reversible binding of

MTBR,9, I also conducted competitive binding experiments

with the MAP-2 second repeated sequence peptide analogue

(VTSKCGSLKNIRHRPGGGRVK) using taxol-stabilized microtubules

and subsequent ultracentrifugation to separate microtubule-

bound and free MAP (Joly and Purich, 1990). As shown in

lanes 1 and 2 of Figure 2-6, MTBR1, was present in both

supernatant and pellet fractions, as would be expected when

MAPs are in excess concentration relative to tubulin (i.e.,

10 MM and 2.5 MM, respectively). However, at 0.4 mM (lanes

3 and 4) and 2.0 mM (lanes 5 and 6) of the repeated sequence

analogue, virtually all of the microtubule-binding domain

can be displaced from assembled microtubules.

In order to more fully characterize the bacterially

expressed microtubule-binding domain interactions with

assembled microtubules, I conducted indirect

immunofluorescence microscopy experiments. Antibodies were

Figure 2-6: Displacement of MAP-2 MTBR from taxol-stabilized
microtubules by a 21-amino acid peptide (m2') corresponding
to the second non-identical repeat sequence of MAP-2. The
amount of microtubule-binding domain remaining associated
with microtubules was visualized by SDS-gel electrophoresis
after separation of microtubule pellet and supernatant
fractions obtained by ultracentrifugation through a sucrose
cushion (20% w/v). Molecular weight markers are indicated
at left, and lanes 1-2, 3-4, and 5-6 correspond to pair-wise
supernatant and pellet fractions obtained in the presence of
0, 0.4, and 2.0 mM peptide-m2', respectively.

97 -
43 ..

MW 1

2 3 4 5 6

Figure 2-7: Indirect immunofluorescence microscopy of
microtubules assembled with the recombinant MT-binding
fragment containing the N-terminal T7-genel0 epitope
sequence (MTBRI6,). (A) Fluorescence images observed upon
incubation with anti-beta-tubulin antibody. (B) Fluorescence
images observed upon incubation with antibody to the T7
capsid antigen peptide fragment described in the text.


- - -


used which were directed against beta-tubulin as well as

antibodies raised against the 12 amino acid gene-ten leader

peptide in the MTBRIne that is the natural amino terminal

end of the T7 major capsid protein (Luts-Freyermuth et al.,

1990). All assembled microtubules were uniformly labeled

with the recombinant MAP-2 fragment, as indicated by the

correspondence of fluorescence labeling in Figure 2-7a

(anti-tubulin) and Figure 2-7b (anti-genelO epitope). As an

additional control, I used the bacterially expressed wild-

type MT-binding domain in place of the epitope-tagged form,

and no fluorescence was observed when incubated with the

rabbit anti-genelo antisera.

The CD spectrum for MTBR162 is shown in Fig. 2-8. This

result is consistent with an unordered structure,

particularly low in alpha and beta structure as evidenced by

the lack of positive signal in the 190-200 nm range. This

spectrum is very similar to spectra reported for whole MAP-2

(Hernandez et al., 1986; Voter and Erickson, 1982) even

though the MT-binding region used here comprises only 10% of

the primary structure of MAP-2. Wille et al. (1992) also

reported a CD spectrum for tau which is virtually

indistinguishable from Figure 2-8.

Earlier findings (Joly et al., 1989; Joly and Purich,

1990; Yamauchi et al., 1993) suggest that the second

repeated sequence may be the primary site on MAP-2 for MT

binding. Nonetheless, those earlier studies with 18- and

Figure 2-8: Circular dichroism spectroscopy of the
recombinant MAP-2 MT-binding fragment (MTBR16) The
average of ten spectra is shown for 45 pM MTBR,6, (1 mg/ml)
in 10 mM sodium phosphate buffer.



-10.0 10 200 210 2202, ,40
190 200 210 220 230 240

wavelength (nm)


21-amino acid peptide analogues were inconclusive because

the analogues of the second repeat displayed 60-100x higher

dissociation constants for MT binding when compared to MAP-2

or its thrombin-cleaved MT-binding region. Accordingly,

site-directed mutagenesis experiments were conducted in

which the first and third repeats were modified to resemble

the second repeat. If one considers the interactions

diagrammatically presented in Figure 2-9, then MAP-2 binding

to MTs can be depicted as a single associative interaction

between the second repeat (or nm2) and a single MT site or as

a multi-site interaction involving the additive effects of

all three repeated peptide sequences (i.e., ml, m2, and m3).

If the single-site case applies, then no significant

improvement in the strength of MAP binding should occur as a

result of mutations in the first or third repeats, even if

they are modified to more closely resemble the second

repeated sequence. If the multi-site case is operant, then

a substantial increase in MAP affinity should attend the

conversion of either the first alone, the third alone, or

their combined replacement.

Accordingly, the following mutant MT-binding regions

were constructed and expressed: MTBRiUg[m,2-m2-m3],

MTBR1,[ml-m2-m32], and MTBR,9[m12-m2-m3], and each has the

entire sequence from position-1629 (thrombin site) to

position-1828 (C-terminus). The subscripts "12" and "32"

indicate that the m, and m3 sequences have been modified

(with specific substitutions shown in bold letters) to

resemble m2 as follows:






Joly et al., 1989 showed that while peptides m, and m3

failed to stimulate in vitro tubulin polymerization, the m2

peptide was effective. I did not initially make complete

replacements of mI and m3 sequences by m2, and therefore

conducted a series of experiments using synthetic peptide

analogues corresponding to peptides m12 and m32 to confirm

that they behaved in a manner analogous to m2. As shown in

Figure 2-10, peptides m2 (lanes 3 and 4 for supernatant and

pellet fractions), m12 (lanes 5 and 6) and m32 (lanes 7 and

8) all displace the MAP-2 MT-binding domain from assembled

microtubules with virtually indistinguishable efficiency.

With these results confirming the feasibility of using

the above substitutions, I proceeded to characterize the MT-

binding properties of aforementioned mutant MT-binding

domains (Figure 2-11). In these experiments, the strength

of microtubule binding was evaluated by determining the

amount of MTBR displacement from taxol-stabilized

microtubules as a function of the concentration of twenty-

one amino acid m2' peptide VTSKCGSLKNIRHRPGGGRVK (Joly and

Figure 2-9: Two models for repeated sequence interactions
with microtubules. Tubulin a and P subunits are shown as
darkened and open circles, respectively. The binding scheme
at the left indicates multisubunit interaction, and the
scheme at the right involves a single interaction with the
second repeat.




Region....... m m3


Figure 2-10: Displacement of MTBRI, from taxol stabilized
MTs by synthetic 18 amino acid peptides corresponding to m2-
like sequences substituted for the first and third repeats
in the expression plasmids described in the text (peptide-
mg and peptide-m2, respectively). Lanes 1-2, 3-4, 5-6, 7-8
correspond to supernatant and pellet fractions after
separation by ultracentrifugation without peptide, with 2.5
mM peptide-m2, with 2.5 mM peptide-m12, and with 2.5 mM
peptide-m,,, respectively. Molecular weight markers are
shown at left.

P S P s P S P


2 3 4 5 6 7 8




MW 1

0" 1 .

Figure 2-11: Displacement of MT-binding domain mutants and
wild-type MTBR1e (m,-m2-m3) by peptide-mn'. Mutants m -m2-m3
m,-mR2-M, and m2-m2-m3, correspond to changes in the first
and third nonidentical repeats to resemble the second repeat
(see text). Tubule-bound and unbound MTBR were separated by
ultracentrifugation. Pellet fractions were subjected to
SDS-polyacrylamide gel electrophoresis followed by Coomassie
staining. The amount of MT-binding domain in the pellet
fraction was measured by densitometry of the gels and is
normalized relative to the density of the MTBR band obtained
in the absence of peptide-m2'.

I ^ 0--0 m1 -m2-m3
O-0n -m12-m2-m3
- 0.75 V- m -m2-m32
0.75 --A m- -mA-m32
2 A 1232

o 0.25-- A

0.00 -
0.00 0.25 0.50 0.75 1.00
[peptide-m2'] mM


Purich, 1990). The wild-type binding fragment displayed

weaker binding than MTBR19[m12-m2-m3] or MTBR1,[m1-m2-m32],

and all three bound less strongly than the combined mutant

MTBR,6[m2m2-m-m32]. In terms of the concentration of

peptide-m2' required to displace half of the binding

fragment from taxol-stabilized microtubules, the wild-type

and combined mutant MTBR1629[m12-m-m32] gave values of 0.13

and 0.6 mM, respectively. This degree of enhanced binding

falls considerably short of that expected if all three

repeated sequences bind simultaneously on a microtubule, and

these data suggest that the second repeated sequence is the

primary site of interaction. In support of this conclusion

is my separate finding that complete substitution of the m2

region with m3, using cassette mutagenesis, greatly reduced

the extent of binding as discussed in Chapter 3. Indeed, in

assembly studies with comparable levels of wild-type and

this mutant binding domain the extents of assembly after 25

min in the presence of 10 AM tubulin were 0.205 and 0.028,

respectively. The latter value is only slightly above that

which we typically observe for tubulin self-assembly in the

absence of microtubule-associated protein.

Sloboda, Dentler, and Rosenbaum (1976) demonstrated

that microtubule-associated proteins can greatly stimulate

tubule polymerization, presumably by stabilizing nucleation

as well as by reducing the critical concentration for

microtubule self-assembly. Both the rate of polymerization


as well as the extent of polymerization, respectively, can

be used as a indices for evaluating MAP interactions with

microtubules. Thus, as an independent test of the

contribution of each peptide repeat to MAP binding, the

ability of mutant forms of the MT-binding region to

stimulate microtubule assembly was studied. Shown in Figure

2-12a are the rates and extents of MAP-stimulated tubulin

polymerization for wild type, MTBR16[m12-m2-m2] and

MTBR162[m,-m2-m.]. These data demonstrate that conversion of

the first repeat or third repeat to resemble the second

repeat has little effect on the ability of MAPs to stimulate

assembly. However, the nature of the differences in rates

of polymerization is not known. Furthermore, I tested two

additional mutant forms: the first, designated as

MTBR1[im29-m2-m3], has the first repeat replaced by an exact

m2 octadecapeptide repeat; and the second mutant form,

designated MTBR,6,[m-m2-m2], has the third repeat replaced

by an identical ma sequence. The findings presented in

Figure 2-12b indicate that increasing the number of second

peptide repeats has relatively little influence on either

the rate or extent of tubulin polymerization. As stated

earlier, if multiple binding of m2-like sequences occurred,

one should anticipate that the additivity of free energies

of binding should be reflected in a substantial

multiplicative change in the apparent affinity of a MAP for

microtubule-binding sites. In such a case, mutant MAP forms

Figure 2-12: Extent and rate of tubulin polymerization
induced by mutant and wild-type forms of MTBR, (A) Wild-
type and mutants with either the first or third repeat
changed to resemble the second repeat. (B) Wild-type and
mutants with the first or third repeat replaced with an
identical copy of the second repeat. Duplicate assembly
runs were made for each at 37*C with 1.1 mg/ml tubulin and
0.73 JM MTBR. The error bars represent the range. Extent
(indicated by the open bars) and rate (indicated by shaded
bars) of tubulin polymerization are averages of these

A 100
-50 3

0.10.- 0
0.10 25

0.00 0
ild m 12m2m3 mm2m32 m2m2m32

0.25 B

0.20 .;z

o 0.15 --o

0.10 --
25 x
0.05- o

0.00 0
wild m2m2m3 m1m2m2


containing extra copies of the second repeat sequence would

be expected to be far more effective in promoting tubule

assembly, and the data in Figure 2-12ab clearly indicate

that the contrary is true.

Joly and Purich (1990) showed that inclusion of three

extra amino acids (RVK), found just after the second repeat,

into a peptide analogue of this repeat greatly improved the

assembly promoting capability. With this in mind, I changed

the three amino acids after the first and third repeats in

MTBR62[m12-m2-m32] to RVK. This new mutant was designated

MTBR,1,[m,2'-m2-m32']. As shown in Figure 2-13 these

mutations did not further enhance the affinity for

microtubules in a peptide-displacement assay. Furthermore,

the extent of assembly was not increased by inclusion of the

RVK sequence after the m12 and m. repeats in Figure 2-14.

However, the rate of assembly was increased substantially by

altering the first and third repeats to resemble the second

repeat and further increased by including the RVK sequence

after the m2-like sequences.

Ideally, I would have liked to assess the affinity of a

construct with all three repeats as exact copies of the m2

sequence. However, this product was very insoluble in

buffers containing less than 0.5 M NaCl and therefore

readily pelleted even in the absence of microtubules.

Figure 2-13: Displacement of wild-type MTBR1, and mutants
MTBR,[m12-m2-m2] and MTBR1[m '-m2-m-m' ] by peptide-m2'.
Mutant m12-m,--m corresponds to changes in the first and
third repeats to resemble the second repeat while mutant
m1'-m2-m32' corresponds to further modification to introduce
the sequence RVK after the altered first and third repeats.
Tubule-bound and unbound MTBR were separated by
ultracentrifugation. Pellet fractions were subjected to SDS
polyacrylamide gel electrophoresis followed by Coomassie
staining. The amount of MTBR in the pellet was measure by
densitometry of the gels and is normalized relative to the
density of the MTBR band obtained in the absence of peptide-


of* 1.08
mA-A m12-m2-m32
S\0 0-0 m12'-m2-m32
.D 0.8 \ \



0.4 A O


0.0 I I
0.0 0.5 1.0 1.5

[peptide-m2] mM

Figure 2-14: Comparison of extent and rate of tubulin
polymerization induced by wild-type MTBR16,, mutant
MTBR1,[m2-m2-m2_ ], and mutant MTBR16g[m12'-m2-m32']. Mutant
m12-m2-m3 has the first and third repeats changed to
resemble the second repeat while mutant m1,'-m2-m' was
further modified to include the sequence RVK after the
altered first and third repeats as described in the text.
Duplicate runs were made for each at 37*C with 11 gM tubulin
and 0.73 pM MTBR. The error bars represent the range.
Extent (indicated by the open bars) and rate (indicated by
shaded bars) of tubulin polymerization are averages of the

m1m2m3 m12m2m32 m12'm2m32'








125 2

75 3



The purification and preliminary characterization of

bacterially expressed MAP-2 MT-binding region starting from

the thrombin cleavage site (numbered 1629 in the mouse

sequence) to the C-terminus was described. The focus on the

microtubule-binding region of this brain-specific MAP

relates to a interest in defining the sub-site interactions

that stabilize fibrous MAP binding to microtubules. This

region retains all of the binding properties of adult MAP-

2ab (Joly et al., 1989), and bacterial expression and

purification of wild-type and mutant binding regions now

open the way for a more detailed biochemical and structural

inquiry. Likewise these developments should facilitate

future experiments with living cells to understand

structure/function relationships affecting the molecular

design of this protein.

Earlier sequencing of mouse and rat MAP-2 cDNA,

respectively revealed that both proteins have highly

conserved regions in their extreme amino and carboxyl

regions (Lewis et al., 1988; Kindler et al., 1990), and

studies with bovine MAP-2, a protein from an evolutionarily

more distant organism, indicate the remarkable preservation

of sequence throughout the microtubule-binding domain.

Indeed, from position 1538 to C-terminus at 1828 there are


only two substitutions (i.e., an D/E at 1564 and a G/S at

1785) and a three amino acid HTP insert in the bovine

sequence immediately following the thrombin-cleavage site at

1629. Such extensive preservation of sequence has been

observed in only a few cases; most notable is the HSP70

multigene family of heat shock proteins (Hendrick and Hartl,

1993), including the endoplasmic reticulum protein BiP and

the major constitutively expressed hsc70 proteins. While

these are arguably the most conserved sequences known,

differing only by one or two substitutions in 650 residues,

my observations with the microtubule-binding region also

point to strong evolutionary pressure to maintain structure

in this region of MAP-2. One might expect that such

conservation reflects the stability against mutation

displayed by other proteins with which MAP-2 interacts, and

the tubulins are widely recognized for their evolutionary

stability. Studies with MAP-2 and related work with tau and

MAP-4 indicate that the tubule-binding domain is primarily

confined to the repeated sequence motif; so this does not

account for conservation outside this motif. However,

Brandt and Lee (1992) showed that a small stretch of amino

acids immediately preceding the repeated sequence region may

allow tau to bundle microtubules. Of course, proteins other

than tubulin may interact with MAP-2 in this domain. Flynn

et al. (1987), for example, observed that the MT-binding

region binds to neurofilaments.


In this present investigation, I have studied the

microtubule binding properties of site-specific mutant forms

of the MAP-2 MT-binding region. Mutations in the first and

third repeated peptide sequences can lead to somewhat higher

affinity microtubule binding when such sequences are made to

resemble the m2 sub-site. It should be noted, however, that

Goode and Feinstein (1994) recently demonstrated the

importance of so-called inter-repeat regions that intervene

between each of the nonidentical repeats in tau protein, and

future work should address the importance of corresponding

elements in MAP-2.

In terms of the purification scheme adopted in this

study, the fact that greater than 95% of the bacterially

expressed protein was found to be in the form of insoluble

inclusion bodies has proven advantageous. Because fibrous

MAPs are heat-stable and completely retain microtubule-

binding characteristics even after incubation for 10 min at

100*C, the protein can be obtained in a soluble form after

initial harvesting of inclusion bodies by centrifugation and

subsequent heat treatment. Further purification is also

facilitated by the fact that the fragment has an overall

isoelectric point of around 10.5, and I can obtain high

levels of purification using either high performance liquid

chromatography or phosphocellulose and gel permeation

chromatography. Experience with approximately thirty


different site-directed mutant forms indicates that the

expression/purification protocol works equally well.

All available evidence points to the dominant role of

the second repeated sequence in MAP-2 binding to

microtubules, which in Figure 2-9 is illustrated in terms of

P-subunit binding based on the findings of Cross et al.

(1990). First, peptide analogues of this sub-site promote

tubulin polymerization, whereas analogues to the other

repeats do not (Joly et al., 1989). Second, only m2-peptide

analogues displace MAP-2 (Joly and Purich, 1990). Third,

Dingus et al. (1991) demonstrated that antibody binding to

the third repeat does not interfere with MAP-2 binding to

microtubules, indicating that the third repeated sequence

makes no significant contribution to the binding energy.

While improvement in microtubule binding affinity can be

accomplished by conversion of the first and third sequences

to resemble m2, the modest gain can be accounted for by

assuming that only one sequence is actually engaged in

microtubule binding and that I have effectively increased

the molarity of m2-like sequences by a factor of three. If

the m2 and the additional m2-like sequences had interacted

at three MT sites simultaneously, additivity in binding

energy would have been expected to result in a far greater,

multiplicative, increase in the association equilibrium

constant for MAP-2 binding to microtubules. In summary, the

single m2-interaction model in Figure 2-9 seems most