Vol. 269,No. 38,Issue of September 23,pp. 23769-23775, 1994 Printed in U.S.A.
THEJOURNAL OF BIOL~GICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc
Tubulin GTP Hydrolysis Influences the Structure, Mechanical of Microtubules* Properties, and Kinesin-driven Transport (Received for publication, June 10,
1994)
Ronald D. ValeSOq, Chris M. Coppin$, Fady MalikS,F. Jon KullS, and RonaldA. Milliganll** From the Departments of Pharmacology and Biochemistry, University of California, Sun Francisco, California 94143, the §Marine Biological Laboratory, Woods Hole, Massachusetts 02543, and the lpepartment of Cell Biology, Scripps Research Institute, La Jolla, California 92037
Tubulin is a GTPase that hydrolyzes its bound nucle- bulin-GTP subunits at the endof the polymer; hydrolysisof the otide triphosphate after it becomes incorporated into anucleotide occursrelatively rapidly afterincorporation into the microtubule. Theonly known consequence of nucleotide polymer. The microtubule continues togrow as long as it has a hydrolysis is that it increases the dissociation rate of cap of tubulin-GTP subunits;however, if nucleotide hydrolysis tubulin from the end of the microtubule by 2 orders of exposes GDP-containing subunits at the end of the polymer, magnitude. In this study, we investigated whether mi- then tubulin-GDP subunits rapidly dissociate and themicrotucrotubules composed of tubulin-GMPCPP (guanylyl cy,@- bule enters a shortening phase. Current estimatesfor the size methylenediphosphate) (a very slowly hydrolyzed GTP of the GTP cap suggest that it may be smaller than 200 subanalog) or tubulin-GDP exhibit additional structural orunits (10). functional differences. We show that tubulin-GMPCPP Since microtubules contain a paucity of tubulin-GTP submicrotubules are significantly stiffer than tubulin-GDP units, nonhydrolyzable analogues of GTP serve asuseful tools microtubules and have a 21%shallower protofilament of microtubules comtwist angle. We also find that kinesin, a microtubule- for examining the structure and dynamics of tubulin with unhydrolyzed nucleotide. One posed entirely based motor protein, transports tubulin-GMPCPP misuch analogue is GMPCPP,’ which is hydrolyzed extremely -30% fasterratesthantubulin-GDP crotubulesat s-’) after being incorporatedintothe microtumicrotubules. These findings suggest that growing mi- slowly (4 x bule (11).Tubulin-GMPCPP has an on-rate similar to that of crotubule ends, which are thought to be composed of tubulin-GDP, and the analogue binds to tubulin with an affintubulin-GTP, may have different structural and mechanical properties from the remainder of the microtubule ity comparable tothat of GTP (11).These results indicate that GMPCPP is indeed a good analogue for studying theconformapolymer. tion and propertiesof tubulin withunhydrolyzed nucleotide. In accordance with the GTP cap hypothesis, microtubules comMicrotubules, cylindrical polymers of C U tubulin ,~ subunits, posed of tubulin-GMPCPP are extremely stable (12, 13) and are involved in various forms of cell motility including ciliary subunits dissociate at 5000-fold slower rates than those measbeating, movement of chromosomes during mitosis, and or- ured for tubulin-GDP (11). In this study, we have examined the structure, mechanical ganelle transport (1). Microtubules grow and shrinkby adding subunits to their ends, and they exhibitconsiderable fluctua- properties, and kinesin-induced motility of microtubules comtions in length even under steady conditions. state The growing posed of either tubulin-GMPCPPor tubulin-GDP. Kinesin is a and shrinking phasescorrespond to the net additionor loss of microtubule-basedmotor protein, composed of two120-kDa thousands of tubulin subunits, and transitions between these heavy chains andtwo 60-70-kDa light chains, that is thought t o transport membranous organelles along microtubules (14, two states are stochastic and generally abrupt. This unusual behavior, termed dynamic instability (2), has been visualized 15). We show here that the microtubulescomposed of tubulinfor microtubules in vitro (3, 4) and invivo (5-7). Dynamic GMPCPP are stiffer and have a shallower protofilament twist instability enables the cell’s microtubule network to turnover angle than tubulin-GDPmicrotubules. We also find that kinesin transports tubulin-GMPCPP microtubules at rates -30% rapidly and thereby change itsconfiguration. Dynamic instability, which does not occur with polymers that faster than those of tubulin-GDP microtubules, even though are in a simple binding equilibrium with a pool of monomers both typesof microtubules stimulate kinesin’s ATPase activity (81, is an energy-requiring phenomenon that is thought to be to the same extent. Together, these results suggest that GTP hydrolysis by tubulin significantly affects the structural and linked to GTP hydrolysis by tubulin. According to the most widely accepted model (the GTP cap hypothesis; reviewed in mechanical properties of microtubules. Ref. 91, microtubules grow whentubulincontainingGTP MATERIALS AND METHODS (bound at the exchangeable site on the p subunit) adds to tuPreparation of Microtubules and Motor Proteins-Tubulin from bo* This work was supported in part by Grants 38499 (to R. D. V.) and vine brain was prepared and modified with N-hydroxysuccinimide rhoAFt39155 (to R. A. M.) from the National Institutes of Health, a grant damine, as described byHyman et al. (11).Microtubules (>4mg/ml) from the Whitaker Foundation(to R. D. VI, a grant from the American were polymerized in BRB80 180 m M PIPES, pH 6.8, 1 mM MgCl,, 1 mM Heart Association (to R. D. V.), and a grant from the Muscular Dystro- EGTA) supplemented with 4 mM MgC1, in the presence of either 1 mM phy Association (to C. M. C.). The costs of publication of this article were GTP(Boehringer Mannheim) or 1 mM GMPCPP (a gift from Dr.T. defrayed in part by the payment of page charges. This article must Mitchison) for 15 min at 37 “C. Unless indicated, microtubules were therefore be hereby marked “aduertisement”in accordance with 18 subsequently stabilized by addition of 20 PM taxol (a gift from the U.S.C. Section 1734 solely to indicate this fact. National Cancer Institute). For the ATPase experiments, microtubules ll Established Investigator of the American Heart Association. To whom all correspondence should be addressed. Fax: 81-727-28-7033; E-mail:
[email protected]. Theabbreviations used are: GMPCPP,guanylyla$-methylene** Established Investigatorof the American Heart Association. diphosphate; PIPES,1,4-piperazinediethanesulfonicacid.
23769
23770
Properties of nbulin-GDP lbbulin-GMPCPP and
TABLE I Velocities of motor-driven transport of tubulin-GDP or tubulin-GMPCPP microtubules Motility assays were performed as described under "Materials and Methods." The human kinesin express in E. coli contained amino acids 1-560 of the heavy chain. The velocities of 60 microtubules were measured, and themean and standard deviation were determined. The ratio refers to the velocity of tubulin-GMPCPP microtubules divided by the velocity of tubulin-GDP microtubules. The p values were determined using a two-tailed t test. Motor
Velocity Tubulin-GDP
Tubulin-GMPCPP
Ratio
p Value
PM/S
Squid kinesin Human kinesin (E. coli expressed) 14 S Tetrahymena dynein
* *
0.50 0.04 0.51 0.03
0.67 0.02 0.67 0.04
1.34 1.32
0.15). For tubulin-GMPCPPmicrotu- were found. First, a higher percentage of 14-protofilament mibules, taxol decreased the mean microtubule curvature by a crotubules was observed when microtubules were polymerized
Properties of lhbulin-GDP and lhbulin-GMPCPPMicrotubules
23774
a
in vitro motility The difference was particularly striking in the assay; pronounced bends of tubulin-GDP microtubules were observed that were never seen withtubulin-GMPCPP microtubules. The most extreme curvatures,however, seem in excess of those resultingfrom Brownian motion of the microtubule end and may be due to kinesin-induced deformation of the microtubule (31).In contrast to theeffect of nucleotide on microtubule flexibility, we observed relatively little effect of taxol on microtubule curvature. Previous studies have obtained different resultsconcerning the effect of taxol of microtubule rigidity. In preliminary work described by Gittes et al. (331, taxol was reported to produce a small increase in microtubule rigidity, while Dye et al. (34) reportedthat taxol considerably decreases the stiffness of microtubules nucleated from axonemes. The variation in results may reflect differences in the assays used to perform the measurement, and further studies will be required to resolve this matter. Electron microscopic studies indicate that changes in the curvature of microtubules appear tobe accommodated primarily by increased twisting of protofilaments rather than by compressiordextension in the spacing between subunits within a protofilament (31). Thus, the greater stiffness of tubulinGMPCPP microtubules relative to that of tubulin-GDP polymers might be due to a stronger lateral interactions between protofilaments. In support of this idea, a considerable difference in theprotofilament twist angleof tubulin-GMPCPP versus tubulin-GDP microtubules was observed, which is suggestive of different lateral protofilament interactions in thesetwo nucleotide states. Thenotion of stronger lateral interactions is FIG.6. Cryo-electron microscopyof tubulin-GDP (a)and tubu- also consistent with the finding that disassembling microtulin-GMF'CPP (6) microtubules (both stabilized with taxol).Representative images of 14-protofilament microtubules areshown. While bules (presumablytubulin-GDP) have frayed and coiled protothe 40-A periodicityof microtubules assembled under either condition is filaments at their ends (35), while growing microtubules (preessentially the same, the 14-protofilamentmoire repeat of tubulin- sumably tubulin-GTP) and microtubule composed of tubulinGMPCPP is 21% longer on average. The characteristic 2-0-3-0 stripe GMPCPP have straight, nonfrayed protofilaments (11, 35). pattern of the moire repeat, which is indicative of 14-protofilament Native or bacterially expressed kinesin also transported tumicrotubules (38), is indicated by the black dots and is best seen by viewing the page at a glancing angle along the microtubule axis. Note bulin-GMPCPP microtubules faster than tubulin-GDP microthe clean background (presumably less free tubulin due to the lower tubules. The greater speed of movement was not associated critical concentration),and the presence of many protofilament number transitions (arrowheads)when microtubules areassembled in the pres- with an increase in ATPase rate. Although solution ATPase may notreflect the ATPase rate of motors that are attached to ence of GMPCPP (panel 6 ) . the glass surface and engaged in motility, these finding raise the possibility that kinesin,on average, travels farther per ATP with GMPCPP (-90%) compared with GTP (5040%). Also, hydrolyzed with tubulin-GMPCPP microtubules. Since the eltubulin-GMPCPP microtubules had many more protofilament transitions, perhaps indicating that they can anneal more emental step in the kinesin chemomechanical cycle involves (361, an arrowheads movement of kinesin to an adjacent tubulin dimer readily than their tubulin-GDP counterparts (see increase in the distance between dimers could yield a greater in Fig. 6b). The most striking difference observed, however, was step size. This model, however, is ruled out by diffraction patthat the moire repeat of 14-protofilament tubulin-GMPCPP microtubules (540 2 41 nm;n = 68) was 21% longer than that terns of cryo-electron microscopic images of microtubules, of tubulin-GTP microtubules (4502 32 nm; n = 51).This finding which show very little, if any, change in the spacing between that thetubulin-GMPCPP microtubule protofilaments are less tubulin subunits in the microtubule lattice. The difference in flexural rigidity between tubulin-GMPCPP and tubulin-GDP tightly supertwisted than those of tubulin-GDP microtubules suggests a difference in the lateral interactions between proto- microtubules also isunlikely to change thespeed of transport, since Young's modulus is sufficiently great topreclude any sigfilaments in thesetwo microtubule types. nificantchangein effective step size through microtubule DISCUSSION strain. Using EGop= 1162 pN/nm2 (33), an average inter-kineThe role(s) of GTP hydrolysis by tubulin hasbeen a topic of sin spacing of 100 nm along the microtubule, power stroke force of 4 pN, inner radius of microtubule = 11.48 nm, and outer considerable interest. Kinetic studies and experiments with nonhydrolyzable GTP analogues strongly implicate GTP hy- radius = 14.18 nm (371, one can calculatethat thereduction in drolysis as a mechanism for generating dynamic instability of step size at most would only be 7.9 x 10" nm. These valuesare microtubule (9). We now show that GTPhydrolysis by tubulin negligible compared to the normal step size of -8 nm (36). An influences the structural andmechanical properties of the mi- alternative idea to a change in step size is that the coupling crotubule, as well as the process of microtubule-based motility efficiency (hydrolysis of an ATP followed by completion of an 8-nm step) is higher with tubulin-GMPCPP microtubules. An driven by kinesin. By examining the curvatures of microtubules on kinesin- understanding of the physical basis of this phenomenon, howcoated surfaces and in solution, we conclude that tubulinever, must await more detailed information regarding the kiGMPCPP microtubules are stiffer than tubulin-GDP polymers. nesin-microtubule interaction and the power stroke.
b
Properties of lhbulin-GDP and lhbulin-GMPCPP Microtubules Although microtubules in cells are composed almost entirely of tubulin-GDP, the resultspresented here may be relevant for understanding the growing ends of microtubules, which are thought to be composed of tubulin-GTP. Such ends, whose size is still in debate, might be more rigid and have a different protofilament organization and monomer spacing than the remainder of the microtubule. This prediction could beexamined in subsequent studies. The finding that thenucleotide state of tubulin influences kinesin interactions also raises the possibility that other types of microtubule-associated proteins might interact differently with tubulin-GTP and tubulin-GDP. Proteins that preferentially interact with tubulin-GTP,forexample, might cap the plus-ends of microtubules and stabilize them to disassembly.Tubulin-GMPCPP microtubules might serve as useful reagents for identifying such proteins. Acknowledgments-We thank several studentsinthe Physiology Course a t the Marine Biological Laboratory (Matt Frerking, Susan McLeskey, Tehyen Chen, Keith Kozminski, and Kris Novak) for contributing to the workdescribed in this paper. We are also grateful to Carsten Timmerman for assistance with printing and densitometry of electron micrographs. Laura Romberg and TimMitchisonprovided helpful comments on the manuscript. REFERENCES 1. Cleveland, D. W. (1993)in Guidebook to the Cytoskeletal and Motor Proteins (Kreis, T., and Vale, R., eds) pp. 101-105, Oxford University Press, Oxford 2. Mitchison, T., and Kirschner, M.(1984)Nature 312,237-242 3. Horio, T.,and Hotani, H. (1986)Nature 321,605-607 4. Walker, R. A,, O'Brien, E. T., Pryer, N. K., Soboeiro, M. F., Voter, W. A., Erickson, H. P., and Salmon, E. D. (1988)J. Cell Biol. 107, 1437-1448 5. Cassimeris, L., Pryer,N. K., andSalmon, E. D. (1988)J. Cell Biol. 107, 2223-2231 6. Sammak, P. J., and Borisy, G . G . (1988)Cell Motil. Cytoskel. 20, 237-245 7. Schulze, E.,and Kirschner, M.W. (1988)Nature 334,356-359 8. Oosawa, F.,and Asakura,S. (1975)Thermodynamics of the Polymerization of Protein, Academic Press, New York 9. Erickson, H. P., and OBrien, E. T. (1992)Annu. Rev. Biophys. Biomol. Struct. 21, 145-166
23775
10. Walker, R. A,, Pryer, N. K., and Salmon, E. D. (1991)J. Cell Biol. 114, 73-81 11. Hyman, A. A., Salser, S., Drechsel, D., Unwin,N., and Mitchison, T. J. (1992) Mol. Biol. Cell 3, 1155-1167 12. Sandoval, I., Jameson, J. L..Niedel, J., MacDonald, E., and Cuatrecasas, P. (1978)Proc. Natl. Acad. Sci. U. S. A. 76, 3178-3182 13. Sandoval, I., MaeDonald, E., Jameson, J. L., and Cuatrecasas,P. (1977)Proc. Natl. Acad. Sci. U. S. A. 74, 4881-4885 14. Skoufias, D. A., and Scholey, J. (1993)Cum Opin. Cell Bid. 6,95-107 15. Vale, R.D. (1992)Dends Biochem. Sci. 17,300304 16. Hyman, A. A. (1991)J . Cell Sci. Suppl. 14,125-127 17. Schnapp, B. J., and Reese, T. S. (1989)Proc. Natl. Acud. Sci. U. S. A. 86, 1548-1552 18. Johnson, K A. (1986)Methods Enzymol. 134,305-317 19. Navone, F., Niclas, J., Horn-Booher, N., Sparks, L., Bernstein,H.D., McCafiey, G., and Vale, R. D. (1992)J . Cell Bioi. 117, 1263-1275 20. Howard, J., Hudspeth, A. J., and Vale, R. D. (1989)Nature 342, 154-158 21. Vale, R. D., Reese, T. S., and Sheetz, M. P. (1985)Cell 42, 3%50 22. Vale, R. D., and lbyoshima, Y. Y. (1989)J . Cell Biol. 108, 2327-2334 23. Catterall, W. A,, and Pederson, P. L.(1971)J. Biol. Chern. 246,49874994 24. Adrian, M., Dubochet, J., Lepault, J., and McDowall, A. (1984)Nature 308, 32-36 25. Milligan, R. A,, Brisson, A., and Unwin, P. N. T. (1984)Ultramicroscopy 13, 1-10 26. Stoops, J. K., Schroeter, J. P., Bretaudiere, J-P., Olson, N. H., Baker, T.S., and Strickland, D. S. (1991)J. Struct. Biol. 106, 172-178 27. Stewart, R. J., Farrell, K. W.,and Wilson, L. (1990)Biochemistry 29, 64896498 28. Hunt, A. J., and Howard, J. (1993)Proc. Natl. Acad. Sci. U.S. A. 90, 1165311657 29. Andrews, S. B., Gallant, P. E., Leapman, R. D., Schnapp, B. J., and Reese, T. S. (1993)Proc. Natl. Acad. U. S. A. 90, 6503-6507 30. Warshaw, D. M., Desrosiers, J. M., Work, S. S., andTryhus, K.M. (199O)J.Cell Biol. 111,453463 31. Amos, L.A,, and A m o s , W. B. (1991)J. Cell Sci. Suppl. 14, 95-101 32. Allen, R. D., Weiss, D. G., Hayden, J. H., Brown, D. T., Fujiwake, H., and Simpson, M. (1985)J. Cell Biol. 100, 1736-1752 33. Gittes, F., Mickey, B., Nettleton, J., and Howard, J. (1993)J . Cell Biol. 120, 923-934 34. Dye, R. B., Fink, S. P., and Williams, R. C., Jr. (1993)J. Biol. Chem. 268, 6847-6850 35. Mandelkow, E."., Mandelkow, E., and Milligan, R. A. (1991)J . Cell Biol. 114, 977-991 36. Svoboda, K., Schmidt, C. F., Schnapp, B. J., and Block, S . M. (1993)Nature 366,721-727 37. Chretien, D., and Wade, R. H. (1991)Biol. Cell 71, 161-174 38. Wade, R. H., Chretien, D., and Job, D. (1990)J. Mol. Biol. 212, 775-786
