The biological machines that generate cell migration and intracellular movements consist of a protein motor that hydrolyzes ATP and a filamentous polymer (either actin filaments or microtubules) that provides a track for the unidirectional motion of the motor. We have been studying movement by the molecular motor kinesin since the mid-1980s. Together with Robert Fletterick's laboratory (University of California, San Francisco), we determined the atomic resolution structure of the kinesin motor domain and discovered unexpectedly that it is similar in structure to myosin, an actin-based motor. We also use custom-built microscopes that can measure the force, steps, processivity, and velocity produced by single kinesin molecules.
Using a combination of single-molecule spectroscopy, cryoelectron microscopy (with Ron Milligan's group, Scripps Research Institute), pre-steady-state kinetics, and mutagenesis techniques, we identified a critical mechanical element in kinesin (called the neck linker) and showed that it undergoes nucleotide- and microtubule-dependent conformational changes. This allowed us to construct a structural model that explains the known features of kinesin movement. We also have shown recently that intramolecular tension conveyed through the neck linkers allows the two motor domains in the kinesin dimer to communicate with one another and efficiently couple ATP hydrolysis into forward stepping.
The key features of the model are highlighted in this movie from the 2000 Science review by Vale and Milligan (see below). Kinesin is a dimeric motor protein that travels processively towards the microtubule plus end by taking 8 nm steps, which corresponds to the distance between adjacent alpha/beta tubulin binding sites. The coiled coil dimerization domain is shown in grey (the attached cargo would be at the end of the coiled coil, which is much longer than shown here). The Rice et al. paper proposed that a small peptide called the neck linker docks to the catalytic core in “ATP” states and undocks in “ADP” or nucleotide-free states. (The docked state is Yellow in the movie and undocked state is Red). When the rear head detaches (after phosphate release), then the neck linker docking in the front head pulls the detached partner head from a rear to a forward position. After a Brownian search (illustrated by the bouncing motion of the head), it binds to the forward tubulin binding site and this interaction causes ADP to be released. This microtubule binding event completes the 8 nm step and generates force. This cycle can then repeat as the kinesin takes many steps along the microtubule and the rear head can pass on either side of front head so as not to build up twist in the coiled coil. This asymmetric hand-over-hand motion is supported by studies from Yildiz et al. (see pdf above) in conjunction with work from the Block and Hirose/Higuchi labs. The timing of events in this movie is not accurate, but intended to illustrate structural states in the cycle. Normally, the “step” (rearward detachment, translation past the partner head, diffusional search, and docking) would occur extremely rapidly and occupy a very small fraction of the ATPase cycle (see also studies by Carter and Cross).
However, many questions remain, such as whether kinesin waits in between steps as a two-head bound intermediate (as shown in this movie) or as a one-head-bound intermediate, which has been favored by other investigators. The mechanism of how the two kinesin motor domains communicate so that their ATPase cycles remain coordinated and out of phase also remains a very active topic of investigation in the field. The structural states of the neck linker during the ATPase cycle also require further investigation. The ATP docked state is well supported, but the ADP “undocked” state might have a specific conformation and not completely disordered as shown in the movie.
More recently, we have turned our attention to understanding the mechanism of dynein, the least understood of the cytoskeletal motor proteins. Dynein is a large molecule: the minimal motor domain is approximately 10-fold greater in size than kinesin's motor domain. Studying dynein has required strategies to express and mutagenize its large motor domain and to develop single-molecule assays to study its activity. We have developed budding yeast as a system for expressing and purifying the dynein motor. Using in vitro motility assays, we have observed single-molecule motility of cytoplasmic dynein and find that it is more processive than kinesin and generates a comparable amount of force. We also find that processivity (long-distance movement along a microtubule without detachment) requires two motor domains working in a coordinated manner. By attaching bright fluorescent markers (quantum dots), we also have clearly resolved individual steps taken by dynein motor domains. We have been able to coax dynein to take such steps by pulling on it with an optical trap without providing chemical energy (ATP). This result has led to a new model for dynein stepping. We are currently studying the structural changes that drive dynein motility, and we have recently obtained the crystal structure of dynein's microtubule-binding domain, in collaboration with Ian Gibbons (University of California, Berkeley).
Click on links below to learn more about kinesin and dynein.