SummaryThe spindle assembly checkpoint (SAC) is a signalling network that delays anaphase onset until all the chromosomes are attached to the mitotic spindle through their kinetochores. The downstream target of the spindle checkpoint is the anaphase-promoting complex/ cyclosome (APC/C), an E3 ubiquitin ligase that targets several anaphase inhibitors for proteolysis, including securin and cyclin B1. In the presence of unattached kinetochores, the APC/C is inhibited by the mitotic checkpoint complex (MCC), a tetrameric complex composed of three SAC components, namely BubR1, Bub3 and Mad2, and the APC/C co-activator Cdc20. The molecular mechanisms underlying exactly how unattached kinetochores catalyse MCC formation and how the MCC then inhibits the APC/C remain obscure. Here, using RNAi complementation and in vitro ubiquitylation assays, we investigate the domains in BubR1 required for APC/C inhibition. We observe that kinetochore localisation of BubR1 is required for efficient MCC assembly and SAC response. Furthermore, in contrast to previous studies, we show that the N-terminal domain of BubR1 is the only domain involved in binding to Cdc20-Mad2 and the APC/C. Within this region, an N-terminal KEN box (KEN1) is essential for these interactions. By contrast, mutation of the second KEN box (KEN2) of BubR1 does not interfere with MCC assembly or APC/C binding. However, both in cells and in vitro, the KEN2 box is required for inhibition of APC/C when activated by Cdc20 (APC/C Cdc20 ). Indeed, we show that this second KEN box promotes SAC function by blocking the recruitment of substrates to the APC/C. Thus, we propose a model in which the BubR1 KEN boxes play two very different roles, the first to promote MCC assembly and the second to block substrate recruitment to APC/C Cdc20 .

Microtubule-binding drugs such as taxol are frontline treatments for a variety of cancers but exactly how they yield patient benefit is unclear. In cell culture, inhibiting microtubule dynamics prevents spindle assembly, leading to mitotic arrest followed by either apoptosis in mitosis or slippage, whereby a cell returns to interphase without dividing. Myeloid cell leukaemia-1 (Mcl-1), a pro-survival member of the Bcl-2 family central to the intrinsic apoptosis pathway, is degraded during a prolonged mitotic arrest and may therefore act as a mitotic death timer. Consistently, we show that blocking proteasome-mediated degradation inhibits taxol-induced mitotic apoptosis in a Mcl-1-dependent manner. However, this degradation does not require the activity of either APC/C-Cdc20, FBW7 or MULE, three separate E3 ubiquitin ligases implicated in targeting Mcl-1 for degradation. This therefore challenges the notion that Mcl-1 undergoes regulated degradation during mitosis. We also show that Mcl-1 is continuously synthesized during mitosis and that blocking protein synthesis accelerates taxol induced death-in-mitosis. Modulating Mcl-1 levels also influences slippage; overexpressing Mcl-1 extends the time from mitotic entry to mitotic exit in the presence of taxol, while inhibiting Mcl-1 accelerates it. We suggest that Mcl-1 competes with Cyclin B1 for binding to components of the proteolysis machinery, thereby slowing down the slow degradation of Cyclin B1 responsible for slippage. Thus, modulating Mcl-1 dynamics influences both death-in-mitosis and slippage. However, because mitotic degradation of Mcl-1 appears not to be under the control of an E3 ligase, we suggest that the notion of network crosstalk is used with caution.

The spindle checkpoint restrains anaphase onset and mitotic exit until all chromosomes are stably attached to the mitotic spindle via their kinetochores. The Tao1 protein kinase was recently reported as a novel spindle checkpoint component. When an siRNA was used to repress Tao1, the essential spindle checkpoint component Mad2 failed to localise to kinetochores, and cells rapidly exited mitosis. Tao1 was also shown to interact with BubR1, another essential checkpoint component, and be rapidly degraded after mitosis, a feature typical of many mitotic regulators. Here, we identify four different siRNAs that repress Tao1 protein levels as efficiently as the previously reported siRNA. However, these siRNAs do not override the spindle checkpoint. We also present data indicating that Tao1 does not interact with BubR1 and that it is not rapidly degraded after mitosis. We show that the previously reported siRNA not only represses Tao1 but also dramatically reduces Mad2 protein levels. Crucially, expression of exogenous Mad2, but not Tao1, rescued the spindle checkpoint phenotype induced by this siRNA. Thus, the key functional data implicating Tao1 in the spindle checkpoint can be explained by an off-target siRNA phenomenon that results in Mad2 inhibition. Taken together, our data do not support the notion that Tao1 is a component of the spindle checkpoint.