Cell division is finely controlled by various molecules including small G proteins and kinases/phosphatases. Among these, Aurora B, RhoA, and the GAP MgcRacGAP have been implicated in cytokinesis, but their underlying mechanisms of action have remained unclear. Here, we show that MgcRacGAP colocalizes with Aurora B and RhoA, but not Rac1/Cdc42, at the midbody. We also report that Aurora B phosphorylates MgcRacGAP on serine residues and that this modification induces latent GAP activity toward RhoA in vitro. Expression of a kinase-defective mutant of Aurora B disrupts cytokinesis and inhibits phosphorylation of MgcRacGAP at Ser387, but not its localization to the midbody. Overexpression of a phosphorylation-deficient MgcRacGAP-S387A mutant, but not phosphorylation-mimic MgcRacGAP-S387D mutant, arrests cytokinesis at a late stage and induces polyploidy. Together, these findings indicate that during cytokinesis, MgcRacGAP, previously known as a GAP for Rac/Cdc42, is functionally converted to a RhoGAP through phosphorylation by Aurora B.

We have recently cloned a cDNA for a full-length form of MgcRacGAP. Here we show using anti-MgcRacGAP antibodies that, unlike other known GAPs for Rho family, MgcRacGAP localized to the nucleus in interphase, accumulated to the mitotic spindle in metaphase, and was condensed in the midbody during cytokinesis. Overexpression of an N-terminal deletion mutant resulted in the production of multinucleated cells in HeLa cells. This mutant lost the ability to localize in the mitotic spindle and midbody. MgcRacGAP was also found to bind ␣-, ␤-, and ␥-tubulins through its N-terminal myosin-like domain. These results indicate that MgcRacGAP dynamically moves during cell cycle progression probably through binding to tubulins and plays critical roles in cytokinesis. Furthermore, using a GAP-inactive mutant, we have shown that the GAP activity of MgcRac-GAP is required for cytokinesis, suggesting that inactivation of the Rho family of GTPases may be required for normal progression of cytokinesis.

Abstract. Brain dynein is a microtubule-activatedATPase considered to be a candidate to function as a molecular motor to transport membranous organdies retrogradely in the axon. To determine whether brain dynein really binds to retrogradely transported organelles in vivo and how it is transported to the nerve terminals, we studied the localization of brain dynein in axons after the ligation of peripheral nerves by light and electron microscopic immunocytochemistry using affinity-purified anti-brain dynein antibodies. Different classes of organelles preferentially accumulated at the regions proximal and distal to the ligated part. Interestingly, brain dynein accumulated both at the regions proximal and distal to the ligation sites and localized not only on retrogradely transported membranous organelles but also on anterogradely transported ones. This is the first evidence to show that brain dynein associates with retrogradely transported organelles in vivo and that brain dynein is transported to the nerve terminal by fast flow. This also suggests that there may be some mechanism that activates brain dynein only for retrograde transport.