~~.L1yf-.J C-..

~ <..c.., J"-\041 c.

B,*~'C'b ~

Electrostatic
A. Thszyllski

Properties of Tubulin and Microtubules

J. A. Brownl, E. J. Carpenter and M. L. A. Nip2 E. Crawfordl

ABSTRACT. In this paper we present the results of our ~o~lar dynamics calculations of electrostatic properties of tubulin. Values of net charge, charge distribution and dipole moment components are given for the tubulin heterodimer. Physical consequences of these results are discussed for micro/ tubules.

I. Microtubules ameter, tubulin rials are cilia the body, are 25 nm. protein about among and cell the anchored the the flagella membrane majority at form There are polymers and proteins (MTs) They [I]. cell most are MTs and rigid are the

Introduction filaments of all on the cytoskeleton cells proteins the They either cell also move with and the are may shape form largest polymers carry since the matethey core along the of cell MTs of diof

found serve

in

nearly

eukaryotic motor maintain cell. to

as tracks

which to a typical

serve

as a scaffolding within

structures beat in

which or

a co-ordinated the emanate cell through from

manner its

objects Within

to

propel MTs

environment. The Perhaps that pull negative

of the these the

a centriole. centres. cells

ends

microtubule mitotic be no spindles cell

organizing of dividing

most

importantly, apart

microtubules in anaphase. MTs teins, [3]. to

chromosomes

can

division from with

without two a mass related

microtubules. largely of homologous globular 55 they basic an outer kilodaltons together which of pro-

formed {3-tubulin are known the MT. very

mainly each closely

a-tubulin These two

approximately and is the with

structurally This dimmer tube tube that to form and shown are

bind subunit

form

a heterodimer to an form inner polymers, to form

as a{3-tubulin. The of MT 15 nm as that 5 nm al. [9] this

polymerizes 25 nm bound lateral are and linear bonds

is a hollow (Fig. 1) .The

diameter of via strongly weaker

diameter known a sheet

is composed connected Tubulin height. a and

protofilaments, is in has small In curled diameter indeed

a tube. 8 nm that. on t.he the so in t.he

molecules The electron

roughly

cylindrical of ident.ical. of

globules Nogales et

crystallography are t.he nearly possibilit.y Moving have this

{3 monomers level MT A of In have allows and the the B A B

However, several around lattice the MT

difference particular,

monomer called

t.ypes. in of

latt.i,{~s. latti,p lattice,

a left.-handed nm relat.ive the 1 a

sense, to and their

protofilaments neighbours.

a longitudinal offset is only

shift. O.92hm

4.92

because

{3 monomers

switched

I'v

A.

TuszvNsKI

ET

AL

125 1 15

nm nm--1

8nm

FIGURE 1. A section of a typical microtubule demonstrating the helical nature of its construction and the hollow interior which is fillpd with l,vtoplasm. Each vertical column is known as a protofilament and the typi(;al MT has 13 protofilaments.

rolcs in alt~rnating protofilaments. This change results in the development of a structural discontinuity in the B lattice known as the seam [2]. Each tubulin monomer is comprised of approximately 450 amino acids and in thc neighbourhood of 7000 atoms. A detailed examination shows that each monomer is formf'ct by a core of two (j-sheets that are surrounded by a-helices. The monomer structure is very compact, but can be divided into three functional domains: an amino-terminal domain containing the nucleotide-binding region. an

ELEC'TROSTATIC'

PROPERTIES

OF

TUBULIN

AND

MICROTUBULES

intermediate domain containing the taxol-binding site and a carboxy-terminal domain (C-terminus), which probably constitutes the binding surface for motor proteins [8]. This is the one portion of the tubulin dimmer which was not imaged due to it.~ freedom to move following formation of the tubulin sheet in the crystallographic samples. This tail of the molecule has been described by Sackett [11]. 2. Thbulin's Charge Distribution

lising the atomic resolution strurturc of tubulin fr'oli the Protein Dat,a Bank (PDB) [4] provided by the electron diffraction crystallography of Nogales et al. [9], we have attempted to determine the electronic charge and the dipole moment. Calculations of the potential energy were done with the aid of a molecular dynamics package, TINKER [6]. This computer program serves as a platform for molecular dynamics simulations and includes the facility to use several force-field parameter sets, some of which are protein specific. The most common of these parameter sets for proteins are AMBER [10] and CHARMM [5]. AMBER was selected over CHARMM on the rational that it is a more up-to-date parameter set. The overall performance of the program gave us confidence that the results it provided for tubulin were meaningful. It was determined using TINKER that tubulin is quite highly negatively charged at physiological pH and that much of the charge is concentrated on the C-terminus. (Perhaps as much as 40% of the overall charge.) In Table 1 we have summarized the results of our calculations for tubulin dimmers without the C-terminal tails. Fig. 2 shqws a titration curve for the net charge on the tubulin a(:Jheterodimer (with the C-termini) as a function of pH. We have also been able to obtain a detailed map of the electric charge distribution on the surface of the tubulin dimmer (see Fig.3). It is clear that the C-termini which extend outward carry a significant electric charge. At neutral pH, the negative charge on the carboxy-terminus causes it to remain extended due to the electrostatic repulsion within the tail. Under more acidic conditions, the negative charge of the carboxy-terminal region is reduced by associated hydrogen ions. The effect is to allow the tail to acquire a more compact form by folding (see Fig.4). Although this is probably the largest structural change which occurs due to changes in the cell's pH, we expect that other structural changes, perhaps the result of post-translational modification in the process of microtubule assembly, can similarly affect the electrostatics of the tubulin dimmer . TABLE 1. Tubulin's Electrostatic Properties ( tail region excluded)

TUSZYNSKI

ET

AL

5 0
~ bJ) ~

-5

..I:= .-10 u ~ z -15 -20 -25 -30

6 pH

FIGURE 2. Tubulin titration curve for the tubulin 0(3 heterodimer as a function of pH; obtained with no salt and no intra-molecular charge compensation. Figure courtesy of Dr. D. Sackett [12].

FIGURE 3. A map of the electric charge distribution on the surface of a tubulin dimmerwith C-t.ermini tails present.. Figure prepared using MOLMoL [7].

ELECTROSTATIC'

PROPERTIES

OF

TUBULIN

AND

MICROTUBULES

+
-

+
Low pH (acidic)

FIGURE 4. Cross-section of a MT including the carboxy-termini of the tubulin subunits. The folding shown of the carboxy-termini of the tubulin dimmerdemonstrates the change in the geometry of the molecule with pH. Neutral pH is shown on top, the tail folds at lower pH as the negative charges are screened. 3. The Dipole Moment of Tubulin and Microtubules

In Table 1 we listed the dipole moment of the dimmerwithout the tails (see Fig.5a). The story here, however, is more complicated. As shown in Fig.5, there are several additional sources of dipoles when tubulin is present in an ionic solution. When two dimmers bound within a protofilament, their positively and negatively are

A.

TuszvNsKI

ET

AL

+++

(a)

(b)

(c)

(d)

FIGURE 5. The various contributions to the dipole moment of a tubulin dimer: (a) the intrinsic dipole moment of the globular protein, (b) the double layer formed when two dimers are bound in a protofilament, (c) a possible internal dipole created by electronic transitions in the hydrophobic pocket and (d) a double layer formed by counter ions surrounding the C-termini tails.

charged ends form a double layer with a net dipole moment along the protofilament axis (Fig.5b). Beside each tubulinmonomer there is a hydrophobic pocket that may develop a double well structure. This can give rise to an internal (switchable) dipole moment due to electronic transitions on this positive background (see Fig.5c). Finally, as is shown in Fig.5d the C-termini which are negatively charged are surrounded by counter ions in solution leading to the formation of double layers. The principal contribution to the dipole moment of a tubulin dimer comes, however, from the location of partial charges on the constituent amino acids. When this is brought to bear on the structure of the microtubule, we predict the emergence of a very interesting anti-ferroelectric structure with permanent dipoles placed almost perpendicular to the surface of the MT cylinders and almost cancelling each other due to rotational symmetry (see Fig.6).

4. Discussion

and Conclusions

In this paper we have summarized our calculations regarding the net charges alld dipole 1IIoments of tubulin and microtubules. These calculations were performed using an atomic resolution structure of tubulin in conjunction with molecular dynamics simulations. We found a large negative charge concentrated on the outer surface of the protein, almost half of which is located on flexible peptide tails attached to the outer face. This charge is undoubtably instrumental in tubulintubulin interactions as well as the interactions of tubulin with motor proteins such as kinesin. Fig. 7 shows the results of our force field calculations were two tubulin dimmers placed in each other's neighbourhood and allowed to interact electrostatare ically. The brush strokes represent the direction of the Coulomb force of attraction.

r:I,~:{'THOSTATI('

PROPERTIES

Of

TUBULI1\1

AND

MIC'ROTliBULES

FIGURE 6. The arrows indicate the orientation of the permanent dipole moments of individual tubulin dimmerswith respect to the surface of a microtubule. Figure prepared using MOLMoL [7].

It is clear that a hexagonal lattice is likely to emerge in the process of tubulin assembly into a microtubule. We have performed detailed calculations of the tubulintubulill electrostatic potelltial along various directions and they strongly support the creation of the known lattice structures. These and related results obtained in our study will be published elsewhere [14]. Finally, we should comment on the role of electrostatic screening due to ions in the cytoplasm or a buffer solution in which tubulin and microtubules may be found. Beyond a certain distance, the so-called Debye length, we can ignore electrostatic effects due to screening. This is the most important length scale as the Bjerrum length that compares electrostatic effects with thermal effects is larger. In our case, we expect that beyond 2.0 nm, illteractions can be neglected; be they of a charge(;hargp, charge-dipole, or van der Waal's nature. The potential is gradually switched off in the calculation so there are no discontinuities in the electrostatic potential, 4>. This is in fact close to the true situation since ions in the surrounding solution will screen any surface charges. Some of the results presented in this paper represent "vacuum" results givell that the solvent is not explicitly taken into account. If the surroUllding l1Iixturp of iolls is considered, then the potential due to a point charge rlo~'~ IlOt f..ul off sillll)I~, as l/r, but illstead as
(/)'X -r ,.

1(

where [(-1 i~ thl' Debyl' lpllgth, Since Wf' (;onsirler locatioll~ withill

t.\'pi(;all.,. 0,6 nm ulldpr plly~iologi(;al (;ollditioll~, 1.0 nm of the MT surfa(;e, they are not screen by

r\.-;Z\~SI,

FIGURE 7. A view of the attractive regions about a tubulin dimmer as would be experienced by another diller. The smallest principal moment of inertia of the dimmers perpendicular the page, the is middle one is aligned vertically, the largest principal moment horizont.ally. See text for more details. Figure prepared using RASMoL [13]. the ions of the solution as there is not sufficient room for even water to be located iI] thf' intl'r\'l'Iling spacE'. 5. Acknowledgments This r('.~par(.h was supported by gral1t.~ from ~SERC and MITACS-Ml'vlPD Discussions \\ith Dr. Dan Sa(;kett of NIH are gratefully acknowledged.
References 1. Bruce Alberts, Dennis Bray, .1ulian Lewis, Martin Raff, Keith Roberts, and .1ames D. Watson, Molecular biology of the cell, 3rd ed., Garland Publishing, London, 1993. 2 Linda ..\. l\mo:;, Thc microtubule lattice -20 years 011., Trerlds il! Cell Biology 5 {1995), no.2, 48-51. 3. Lil!da A. Arnos arId W. Bradshaw Amos, Molecule.~ of the cytoskeleton, 1st ed., MacMillan Press, London, 1991. J Hl'll'n \" Berrrlan, .1ohn V\'('stbrook, Zukal!g [.',~ng, (:ary Gillilal!d, Talapady N. Bhat, Helge

Wei:;:;ig, Ilya N. Shil!dyalov, ,lnd Phil E. Bourne, The p1'otei71 data bank, Nucleic Acids Re:;earch 28 {2000), 1!0. 1, 2:35242 5. Berrlard R. Brook:;, [{obert E. Brurcoleri, Barr.y D. Olaf:;ol!, David .1.State:;, S. Swamil!athal!, and rvlartin Karplus, CHARMM~ A r1'ofJ1'am tor rnacrornolecula1' energy, minimization, arId (j d""amirs calculation.~, .1ollrllill of ('orll[Jutational Chemistry 4 ( 191;:3), no.2, 187217. \,lil~l,a{~1 .1. !)Udl'k illld .Jav \\ ['olld~r, Ac,urat" rnodelirl9 of the iru1'amolecular electrostatic enf'rqy ()fp1'ou;in.~. .1 or ()lllrlllalional ('llcmi:;try 16 {1995). rlo. 7,791-816.

E;LEC'TROSTATIC'

PROPERTIES

OF

TUBULIN

AND

MIC'ROTUBULES

7. Ret,o Koradi, Martin Billet,er, and Kurt Wiithrich, MoiMoi: A program for display and arlalysis of 7nacrornolecular structures, J Mol Graphics 14 (1996),51-,55. 8. Eckhard Mandelkow a!ld Eva-Maria Ma!ldelkow, Microtubule,~ artd rnic7'otubule-associated proteirts, Current Opinion in Cell Biology 7 (1995), !lO. 1, 72-81. 9. Eva Nogales, Sharon G, Wolf, and Kenneth H. Downing, Structure of the alpha-beta tubulin dirner by electron crystalography, Nature 393 (1998), !lo. 6663, 199-203. 10. David A. Pearlman, David A. Case, James. W. Caldwell, Wilson S. Ross, Thomas E. Cheatham, Stephen E. DeEolt, David M. Ferguson, George L. Seibel, and Peter A. Kollman, AMBER, a package of computer programs for applying molecular mechanics, normal mode analysis, molecular dynamics and free energy calculations to simulate the structural and energetic properties of molecules, Computer Physics Communications 91 (1995), 1-41. 11. Dan L. Sackett, Subcellular biochemistry, Proteins: Structure, Function and Engineering, vol. 24, ch. Structure and Function in the Tubulin Dimmer and the Role of the Acidic Carboxyl Terminus, Plenum Press, New York, 1995. 12. -, personal communication, 2000. 13. Roger Sayle and E. James Milner-White, RasMol:

Biomolecular

graphics for all, Trends

in

Biochemical Sciences (TIES) 20 (1995), no.9, 374. 14. Jack A. Tuszyiiski, James A. Brown, Eric J. Carpenter, Electrostatics properties of tubulin and microtubules,

Ellen Crawford, submitted

and M. L. Alex Nip,

to Proc. Nat. Acad. Sci.

1 DEPARTMENT 2DI;:P,\RTMENT Current rvlontreal, QC

OF PHYSICS, OF PHYSIC'S, Faculty H3C 3J7

UNIVERSITY UNIVERSITY

OF ALBERTA, OF ALBERTA, Department

EDMONTON EDMONTON

AB AB

CANADA, CANADA, University

T6G T6C;

2Jl 2J 1

addr'ess: Canada,

of Medicine,

of Biochemistry,

of Montreal,