To grow or to shrink…

Aurora B puts chromosomes in their place mall molecule inhibitors of Aurora B activity, characterized by Hauf et al. (page 281) and Ditchfield et al. (page 267), reveal that the mammalian kinase and its budding yeast counterpart, Ipl1, have similar functions. Without Aurora B, mistakes in kinetochore–chromosome interactions go uncorrected. Early evidence of a function for the Aurora family in correcting syntelic attachments, those in which both chromatids are attached to the same spindle pole, was provided by the ipl1 mutant. But visualizing spindle– kinetochore attachments in yeast is difficult. The two articles in this issue examine attachments directly, by inhibiting Aurora B in mammalian cells. The groups used different compounds, but in both cases the Aurora B inhibitors left chromosomes misaligned and compromised the spindle checkpoint, thus causing division S To grow or to shrink… ooks can be deceiving. According to two articles in this issue, proteins that look like microtubule stabilizing proteins at times do just the opposite, revealing activities that can both build and destroy microtubules. Originally described as a Xenopus microtubule stabilizing protein, XMAP215 is a defining member of a large family of microtubule-associated proteins. Depletion of XMAP215 or its homologues leads to decreased spindle microtubule length in several systems, including fly, yeast, and worm. On page 349, however, Shirasu-Hiza et al. find that XMAP215 also promotes depolymerization of microtubules stabilized with a nonhydrolyzable GTP analogue (GMPCPP). This destabilizing activity, like its stabilizing activity, is specific to microtubule plus ends. The new work recalls a 10-year-old report demonstrating that XMAP215 has both activities in vitro. Sirasu-Hiza et al. used EM analysis to reveal a structure that supports a peeling-like mechanism of XMAP215, similar to that of KinI kinesin. Previously, the plus ends of microtubules stabilized by GMPCPP have been thought to resemble a " GTP cap, " a structure postulated to exist at the ends of growing microtubules. Here, the authors suggest instead that GMPCPP-stabilized structures may mimic a " paused state " —a hypothetical third state in microtubule dynamics, intermediate between the growing and shrinking states. They propose that XMAP215 destabilizes this paused state and increases either polymerization L or depolymerization rates depending on cellular conditions, thus explaining its dual activities. On page 359, van Breugel et al. find another XMAP215 family member with destabilizing activity— the budding yeast homo-logue Stu2p. In vitro, Stu2p depolymerized microtubules by binding directly to plus ends, probably hindering tubulin …