If cytokinesis were to occur too soon, chromatids still lingering in the spindle midzone might get cut. Cytokinesis fails in animal cells with spindle midzone defects, and Barral's group now shows that the same is true for budding yeast. Although the actomyosin ring contracted normally, final membrane resolution (called abscission) was delayed in several mutants with altered midzone structure.
The midzone appears to be important for abscission perhaps by acting as a sink to inactivate the yeast Aurora kinase Ipl1. Although Ipl1 remained in the nucleus, it activated the export during mitosis of two anillin-like proteins called Boi1 and Boi2 that shuttled to the bud neck. There, they delayed abscission, as loss of the Boi proteins hastened cytokinesis, even in wild-type cells, thus causing chromosomal breaks.
At anaphase onset, Ipl1 is on chromosomes and is probably active. At telophase, Ipl1 accumulates at the midzone, where DNA is lacking, allowing its inactivation. Boi1 and Boi2 thus return to the nucleus.
Unlike other checkpoints, the delay did not stall mitotic progression; everything except cytokinesis went forward. “A yeast cell can survive if cytokinesis fails once,” says Mendoza. “The worst that can happen is it becomes diploid or forms a chain of multiple nuclei in a common cytoplasm. It's not tragic. What would be tragic is if chromosomes were cut.”
Mammalian Aurora B promotes initial furrow ingression and is thus required for cytokinesis. To determine whether it also prevents premature abscission, it will be necessary to knock out the kinase activity only after ingression.