Single molecule mechanics of Kif15

Kinesin-12 is a motor protein that has a role in the processes of mitotic spindle formation and maintenance. The human Kinesin-12, Kif15, has been shown to have some functional redundancy with Eg5, a Kinesin-5 that plays key roles in the formation of the bipolar spindle and is a potential target for anti-cancer drugs. Eg5 is thought to contribute to spindle formation by cross-linking and sliding microtubules, however little is known about the mechanism of Kif15.\ud \ud We have used laser tweezers to investigate the mechanical properties of Kif15 compared to those of kinesin-1. We have found that Kif15 is plus end directed and takes multiple steps along the microtubule without detaching. Full-length Kif15 walks faster and supports more load than full-length Eg5. Kif15 is less processive under load than kinesin-1, although it has a similar stall force. A second, diffusive, microtubule binding site in Kif15 supports processivity at zero load, and slows flyback following a detachment in the optical trap.\ud \ud The microtubule-associated protein, Tpx2, is necessary for the localisation of Kif15 to spindle microtubules. We find that Tpx2 binding arrests the motion of Kif15 and creates a stable binding state that resists both assisting and hindering loads. We also find evidence of a tail-mediated auto-inhibitory mechanism that creates a stable MT binding state and causes pausing during processive runs. C-terminal truncation of the Kif15 tail relieves this inhibition leading to faster overall stepping and abrogates the effects of Tpx2.\ud \ud We examined the detachment behaviour of Kif15 from microtubules, under assisting and hindering loads. We find that assisting loads cause single Kif15 and Kinesin-1 motors to detach from the microtubule more easily than hindering loads. Kif15 shows a much more asymmetric response to load in low levels of ATP than Kinesin-1, and both show more asymmetry than Eg5: previous work has shown that the behaviour of Eg5 does not change dramatically with differing loading directions. This has interesting implications for the roles of Kif15 and Eg5 motors in both parallel and anti-parallel microtubule bundles. Overall our data supports an in vivo mechanism for Kif15 that it distinct from that of Eg5.\ud \ud We investigated the load-dependent detachment of Kinesin-1 and Kif15 in millimolar concentrations of ADP, AMPPNP and micromolar concentrations of ATP. Kinesin-1 in ADP detached at low loads, and in AMPPNP at two different loads, both higher than in ADP. These two AMPPNP states of Kinesin1 likely corresponding to single and double headed microtubule binding, as proposed by Ishiwata and colleagues. Kif15 behaved broadly similarly. At micromolar ATP concentrations and hindering loads, both Kinesin-1 and Kif15 again showed two different high load detachment states. This is inconsistent with the model proposed by Ishiwata and possible modifications are discussed.