How spindles assemble has attracted a lot of attention. However, little is known about the mechanism of spindle disassembly, which takes places at the end of mitosis. Now though, reporting in Cell, Yixian Zheng and colleagues have identified the AAA-ATPase Cdc48/p97 and the Ufd1Npl4 adaptor complex as important regulators of this process.

Given its role in the dissociation of proteins from large cellular structures, such as the endoplasmic reticulum, the authors tested whether the Cdc48/p97–Ufd1–Npl4 complex could assist spindle disassembly. They found that, in Xenopus laevis egg extracts, a dominant-negative form of Xenopus p97 (p97QQ), which was defective in ATP hydrolysis, blocked spindle disassembly and the formation of interphase nuclei and microtubule arrays. And, the Ufd1–Npl4 adaptor complex was also required for spindle disassembly.

Zheng and co-workers found that the microtubule-binding proteins XMAP215 and TPX2, and the spindle-associated Polo-like kinase Plx1, were co-immunoprecipitated with p97–Ufd1–Npl4 from extracts that had been induced to exit mitosis. In addition, compared with p97 wild-type treatment, p97QQ treatment caused increased binding of both XMAP215 and TPX2 to microtubules after mitotic exit. So, p97–Ufd1–Npl4 seems to regulate the binding of certain spindle-assembly factors to microtubules at mitotic exit.

To study the function of p97 in vivo, the authors used the yeast homologue, Cdc48, and found that — at the restrictive temperature — the cells of a heat-sensitive mutant strain (cdc48-3) arrested in mitosis and the spindles first elongated and then collapsed into medium-length spindles that persisted. A temperature-sensitive mutant of Npl4 also had defects in spindle disassembly, so the authors concluded that Cdc48–Ufd1–Npl4 regulates spindle disassembly in yeast in vivo.

Inactivation of the Cdc2 kinase at the end of mitosis also has a role in regulating spindle disassembly. However, Cdc48/p97 is not required for Cdc2 inactivation in either Xenopus egg extracts or in yeast.

So, what's the mechanism that Cdc48 uses to mediate spindle disassembly? Zheng and colleagues hypothesized that Cdc48 might regulate spindle disassembly by binding to Cdc5 (the yeast homologue of Plx1) and Cdc5 and Ase1 and is required for their degradation after mitotic exit.

The authors conclude that, at the end of mitosis, both Cdc2 inactivation and the activity of Cdc48/p97–Ufd1–Npl4 are required for spindle disassembly. They further suggest that this complex binds certain spindle-assembly factors and prevents them from promoting spindle assembly.