Scientific Report 2005

Cell Biology

Structure and Action of Molecular Machines

R.A. Milligan, J. Chappie, P. Chowdhury, T. Dang, A. Efimov, A. Mulder, G. Orca, M. Reedy,* M.K. Reedy,* C. Reyes, B. Sheehan, K. Thompson, A.B. Ward, E.M. Wilson-Kubalek, C. Yoshioka

* Duke University Medical Center, Durham, North Carolina

Macromolecular assemblies may be composed of from 2 to perhaps scores of proteins and are the functional units—the molecular machines—of the cell. We use electron cryomicroscopy and image analysis to study the structure and mechanism of action of several of these machines. We combine the 3-dimensional maps calculated from electron images of the machines with biochemical data and high-resolution x-ray structures of the individual components to provide insight into the operation of the machines. During the past year, we continued our work on members of the myosin and kinesin superfamilies, microtubule-stabilizing proteins, and membrane proteins.

Although the mechanism of plus end–directed, processive motion by conventional kinesins is now well understood, the mechanism by which members of the kinesin 14 class move toward the minus ends of microtubules is not. Likewise, in the myosin superfamily, how nucleotide-mediated conformational changes in the motor domain of class VI myosins result in “backward” motility is not known. We are elucidating the molecular mechanisms of these more unusual members of the myosin and kinesin superfamilies. (Movies showing the motions of conventional myosin and kinesin can be viewed at www.scripps.edu/milligan/projects.html.)

The kinesin Ncd belongs to the kinesin 14 class of motor proteins. Compared with the situation with plus end–directed kinesins, the nature and timing of the structural changes that underlie the motility of kinesin 14 motors are poorly understood. We used electron cryomicroscopy and image analysis to calculate 3-dimensional maps of Ncd bound to microtubules in various stages in its mechanochemical cycle. The maps revealed a minus end–directed rotation of approximately 70º of a coiled coil mechanical element of microtubule-bound Ncd upon ATP binding. In parallel with these structural studies, our collaborators, N. Endres and R. Vale at the University of California, San Francisco, showed that extending or shortening this mechanical element respectively increases or decreases movement velocity without affecting ATPase activity. These results indicate that as with other kinesins, the force-producing conformational change of Ncd occurs upon ATP binding but, unlike the situation with other kinesins, involves the swing of a rigid, lever arm–like mechanical element similar to that described for myosins.

Whereas most kinesins move along intact microtubules, members of the kinesin 13 class destabilize and depolymerize microtubules and do not appear to have motile properties. We found that a KinI fragment consisting of only the conserved motor core is necessary and sufficient for ATP-dependent depolymerization. The motor core binds along microtubules in all nucleotide states, but in the presence of a nonhydrolyzable ATP analog, depolymerization also occurs. Structural characterization of the analog-induced depolymerization products provided a snapshot of the disassembly machine at the microtubule ends.

Our data indicate that whereas conventional kinesins use the energy of ATP binding to execute a power stroke that results in unidirectional motion along the microtubule surface, KinIs at the ends of microtubules use the energy to bend the underlying protofilament, thereby destabilizing the microtubule lattice and leading to microtubule depolymerization. Furthermore, when the motor core is associated with the microtubule wall, the core is stalled in a weakly bound, nucleotide-free state. Progression to the strongly bound, ATP-containing state is possible only when the KinI encounters a microtubule end, where it can catalyze deformation of protofilaments and disassembly of microtubules. The unusual mechanochemical coupling of this kinesin provides an elegant mechanistic basis for its microtubule-depolymerizing activity.

The protein doublecortin is expressed in migrating and differentiating neurons. In humans, mutations in this protein disrupt brain development, causing lissencephaly. Although doublecortin is associated with and stabilizes the microtubule cytoskeleton, it has no homology with other microtubule-binding proteins such as MAP2 or tau. We found that doublecortin preferentially nucleates and binds to 13-protofilament microtubules. This specificity was explained when we discovered that the protein binds in the valleys between the protofilaments of the microtubule wall. This binding site is unique and appears to be ideally located for microtubule stabilization. In this location, doublecortin most likely contributes to both the longitudinal and the lateral interactions that stabilize the microtubule wall.

In collaboration with G. Chang, Department of Molecular Biology, we have grown well-ordered arrays of several membrane proteins that are involved in multidrug resistance. These arrays, helical tubes and 2-dimensional crystals of membrane-embedded proteins, are suitable for structural studies via electron microscopy. In one instance, we trapped a drug transporter in various stages of its mechanistic cycle and with substrates bound. We anticipate that 3-dimensional electron microscopy maps of membrane-embedded transporters in various states, together with the high-resolution x-ray structures of the detergent-solubilized protein, will provide insights into the mechanisms used to transport metabolites and drugs across membranes.

In other studies, we developed a general method for helical crystallization of proteins on lipid tubules that we are using to study the virulence factor PFO from Clostridium perfringens. PFO is a cytolysin, an important class of proteins that oligomerize and embed within membranes as part of their lytic function. We obtained helical crystals of wild-type and several mutant forms of PFO on nickel-lipid tubules. Three-dimensional maps of these proteins derived from images of the helical crystals will be used to complement our studies of PFO pore formation on lipid layers. These studies will provide a better understanding of the pathogenic function of cytolysins. Additional studies involving tubular crystallization of membrane proteins and other bacterial toxins are opening up promising new areas for future research.

Publications

Dang, T.X., Farah, S.J., Gast, A., Robertson, C., Carragher, B., Egelman, E., Wilson-Kubalek, E.M. Helical crystallization on lipid nanotubes: streptavidin as a model protein. J. Struct. Biol. 150:90, 2005.

Dang, T.X., Hotze, E.M., Rouiller, I., Tweten, R.K., Wilson-Kubalek, E.M. Prepore to pore transition of a cholesterol-dependent cytolysin visualized by electron microscopy. J. Struct. Biol. 150:100, 2005.

Neuman, B., Adair, B.D., Burns, J.W., Milligan, R.A., Buchmeier, M.J., Yeager, M. Complementarity in the supramolecular design of arenaviruses and retroviruses revealed by electron cryomicroscopy and image analysis. J. Virol. 79:3822, 2005.

O’Keefe, M.A., Turner, J.H., Musante, J.A., Hetherington, C.J.D., Cullis, A.G., Carragher, B., Junkins, R., Milgrim, J., Milligan, R.A., Potter, C.S., Allard, L.F., Blom, D.A., Degenhardt, L., Sides, W.H.. Laboratory design for high-performance electron microscopy. Microsc. Today 12:8, 2004.