Several models have been suggested above, describing possible modes of spindle checkpoint action: 1. Cdc20 sequestration (by Mad2-Cdc20 and/or MCC). 2. Stable MCC-APC/C association. 3. Cdc20 turnover (in budding yeast). 4. Cdc20-APC/C modification (by Mps1, Bub1, MAPK, Aurora B or BubR1 kinases). Several of these mechanisms could affect APC/C activity by modifying, competing for, and/or blocking the binding site(s) for its substrates. Alternatively, they could reduce the processivity of ubiquitination of substrates, or prevent the release of substrates and thereby reduce substrate turnover. Indeed, the processivity of ubiquitination can determine the order of destruction of APC/C substrates (Rape et al., 2006). Most substrates require multiple APC/C binding events in order to build polyubiquitin chains, and only polyubiquitinated substrates are recognised by the 26S proteasome for destruction. Thus, if the processivity of ubiquitination or the turnover of APC/C substrates were impaired in mitosis, the degradation of securin and cyclin would no longer take place, which would result in mitotic arrest. Our results have highlighted the importance of Mad3 as an anaphase inhibitor, and suggest that it usually acts in concert with Mad2 to efficiently inhibit Cdc20-APC/C. Further experiments are necessary to fully understand their mechanism of action, and this will require a wide range of approaches including dynamic studies of the 'flux' of Mad2 and BubR1 through signalling scaffolds, further structural insights, the identification of important phosphorylation sites on both the checkpoint proteins and Cdc20-APC/C, and an in vitro reconstitution of MCC inhibition of the APC/C. We look forward to seeing the complex regulation of mitotic progression being described over the coming years.