Cell：“分子发动机和刹车”作用机制的研究

e1p has no polarity preference, MTs become oriented randomly (Figure S5 and Movie S16) and, compared to model 1, the symmetry of the bundle-midzone is consequently decreased by almost a factor of 2 when the unbundled model is used as a baseline (Figure 5D). Previously, a parallel bias in crosslinking for the motors dynein and HSET was, based on computer simulations, proposed to aid the formation of monopolar asters out of randomly oriented MTs (Chakravarty et al., 2004). We can now imagine that in order to construct a complete bipolar spindle out of dispersedly nucleated MTs, cells use two different sets of bundlers and motors. One only interacts with parallel MTs and takes care of focusing MT minus ends into spindle poles, while the other interacts with antiparallel MTs and organizes the spindle midzone (see model, Figure 6B). Because of the difference in polarity bias, these two sorting systems may have very little crosstalk, e.g., the bundling of antiparallel MTs by ase1p during anaphase would not hinder the sorting of parallel MTs elsewhere. Ase1p can therefore act as a selective brake on motors that slide antiparallel MTs. In this light, it will be key to investigate the polarity preference of other mitotic motors and bundlers like Eg5 and NuMA.

MT End Tracking of Motors Creates Responsive MT Networks

How do cells regulate which MTs slide and which MTs are stably connected? Seemingly, when bundlers and motors accumulate between two MTs proportional to the amount of overlap, either sliding forces or bundling forces dominate, and MTs either slide or are immobile. Here, we observed ase1p and klp2p disproportionally along MTs since ase1p localized along the full length of bundled MTs while klp2p accumulated only at MT plus ends (see model, Figure 6A). The molecular tug-of-war between ase1p and klp2p therefore resulted in a net sliding force that decreased with MT length. Computer simulations indicated that the bundling strength of ase1p can be made strong enough to create a bundle-midzone that resists compression forces by MT polymerization yet is weak enough to enable MT sliding of short MTs. Additionally, klp2p localization to MT plus ends made sliding a position-dependent process, since MT sliding stopped when a MT plus end moved beyond the bundle midzone. Sliding of MTs within long MT bundles occurs in various cell types, including neurons (Baas et al., 2006), mature megakaryocytes (Patel et al., 2005), and Ustilago maydis (Straube et al., 2006). Interestingly, observations in neurons show that, similarly to S. pombe, only short MTs are motile whereas longer MTs are stationary. We can thus imagine that end tracking of motors is a more commonly used strategy that allows for dynamic sorting of short MTs on one hand and the establishment of stable MT contacts on the other hand.