Cooperation between Mitotic Kinesins Controls the Late Stages of Cytokinesis

Current Biology 16:301 (2006/02)
Francis Barr @ Max-Planck-Institute of Biochemistry

There are two possible Aurora B phosphorylation sites in MKLP1, S821 and S911. Due to the failure in generating phospho-S821-specific antibody, only S911 was addressed in this paper. Authors showed that Aurora B phosphorylates MKLP1 S911 during anaphase. The S911A mutant did not cause any defect in early stages of cytokinesis, including contractile ring build-up and furrow ingression. However, this phosphorylation on S911 is required for MKLP1 stay at the midzone/midbody during late cytokinesis. A bipartite NLS in MKLP1 was disrupted by S911 phosphorylation so MKLP1 will not be recruited back to nucleus after nuclear membranes reformed.

Plk1, Cdk1, and Aurora B were all tested and only Aurora B could phosphorylate MKLP1 S911 both in vitro and in vivo. Besides, MKLP2 is required for this phosphorylation since it regulates the translocation of Aurora B in cytokinesis. Unfortunately authors didn’t address how the mislocalization of MKLP1 causes cytokinesis failure during late stages. The only evidence authors had is a increase of binucleate cell percetage. It is just not solid enough.

Unlike vertebrates and Drosophila, C. elegans appears to lack MKLP2. Besides, S911-containing NLS region in C-terminus of mammalian MKLP1 is absent from the C. elegans homolog ZEN-4. It suggests this regulatory mechanism is specific for vertebrates and either does not occur in C. elegans or involves a different mechanism. The second site, S812 in the tail region downstream of the actin binding domain, is conserved in ZEN-4, where it is required for regulation of the late stages of cytokinesis. Whether MKLP2 is also required for S812 phosphorylation in human cells and for other Aurora B substrates such as Cyk-4 (MgcRacGAP), is still unknown.