Rho family small GTPases (Rac, RhoA, and Cdc42) function at the core of cytokinesis, the physical division of one cell into two. In this issue, Bastos et al. (2012. J. Cell Biol. http://dx.doi.org/10.1083/jcb.201204107) identify a new role for Rac inhibition: to release cell adhesion at the division plane and allow efficient constriction of the contractile ring. They show that the GTPase-activating protein, CYK4, suppresses equatorial cell substrate adhesion by inhibiting Rac and therefore its effectors ARFGEF7 and PAK1/2.
To ensure that cell division occurs at the correct time and place, spatially regulated signaling mediated by the mitotic spindle directs the assembly and constriction of an actomyosin contractile ring at the site of cytokinesis. Critical for cytokinesis is centralspindlin, a conserved tetramer consisting of a dimeric Rho family GTPase-activating protein (GAP), CYK4 (also called MgcRacGAP in mammals, CYK-4 in Caenorhabditis elegans, and RacGAP50C in Drosophila melanogaster), constitutively bound to a kinesin-6 dimer (Mklp1 in mammals, ZEN-4 in C. elegans, and Pavarotti in Drosophila). Centralspindlin localizes to the spindle midzone (or central spindle) in anaphase and to the midbody (or Fleming body) late in cytokinesis. At the midzone, centralspindlin contributes to the organization of antiparallel microtubule bundles and the recruitment of other players (Green et al., 2012). One controversial aspect of centralspindlin function is the role of CYK4 GAP activity during cytokinesis, as different studies have concluded that it inhibits RhoA (Jantsch-Plunger et al., 2000; Miller and Bement, 2009) or Rac signaling (Canman et al., 2008), activates RhoA signaling indirectly (Loria et al., 2012), or is dispensable for cell division altogether (Goldstein et al., 2005; Yamada et al., 2006).
Much of the controversy regarding CYK4 function has arisen because the main targets of its GAP activity in vitro, Rac and Cdc42, do not play an essential positive role during cytokinesis. When CYK4 was first identified in human male germ cells, it was shown biochemically to specifically stimulate the GTPase activity of Rac and, to a lesser extent, Cdc42, but not RhoA, and was thus termed MgcRacGAP (Touré et al., 1998). This increased activity toward Rac and Cdc42 relative to RhoA was also later reported for recombinant worm and mammalian CYK4 (Jantsch-Plunger et al., 2000; Kawashima et al., 2000). However, as RhoA is the only Rho family GTPase essential for cytokinesis during mitosis, it was assumed that CYK4 acts on RhoA in vivo (Jantsch-Plunger et al., 2000). It has also been postulated that Aurora B phosphorylation of S387 in the human CYK4 GAP domain active site can functionally convert the specificity from Rac/Cdc42 to RhoA during cytokinesis (Minoshima et al., 2003). This model has remained contentious, though, because of the lack of conservation of S387 in CYK4 across species.
Bastos et al. (in this issue) now address this controversy by demonstrating that, consistent with work performed in Drosophila and C. elegans, mammalian Rac plays a negative role in cytokinesis, and CYK4 is required to “release this brake” for cytokinesis to proceed. They first began their study with a comprehensive in vitro analysis of CYK4 GAP activity. By testing a recombinant human CYK4 GAP domain against a panel of small GTPases, they confirmed that the GAP activity is specific for Rac and Cdc42 but shows very little detectable activity for RhoA. They then tested native protein by assaying the activity of centralspindlin immunoprecipitated at different stages of the HeLa cell cycle and showed that endogenous centralspindlin also acts as a GAP for Rac and Cdc42 but not RhoA (at least in vitro). Using mass spectrometric analysis, they were unable to identify any phosphorylation sites within the CYK4 GAP domain, although previously identified phosphorylation sites outside this domain were found. They also showed that the CYK4 GAP specificity was independent of the mitotic kinases Cdk1, Aurora B, and Plk1. Lastly, the CYK4S387D mutation previously suggested to mimic phosphorylation by Aurora B did not up-regulate RhoA GTPase activity but rather inhibited all in vitro CYK4 GAP activity, refuting the model that Aurora B phosphorylation alters CYK4 specificity (Minoshima et al., 2003).
The authors then focused on identifying potential targets of these GTPases via pull-downs of recombinantly expressed Rac, RhoA, and Cdc42. They found that each GTPase likely regulates a functionally distinct set of proteins: proteins pulled down with Cdc42 are known to act in cell motility, those binding RhoA contribute to contractile ring formation and ingression, and those binding Rac act in cell adhesion. As CYK4 negatively regulates Rac, they hypothesized that limiting cell adhesion may be important for division. To support this model, they showed that depletion of Rac binding partners implicated in cell adhesion (ARHGEF7, PAK1, and PAK2) suppresses the cytokinesis failure induced by expression of either constitutively active Rac or GAP-dead CYK4 (Fig. 1). Furthermore, they found that cells expressing GAP-dead CYK4 had increased cell substrate adhesion during cytokinesis as well as aberrant equatorial adhesion fibers, indicating that CYK4 is important for inhibiting Rac-dependent adhesion at the division plane (Fig. 2).
An inhibitory role for CYK4 as a Rac GAP during cytokinesis has also been demonstrated genetically in both Drosophila (D’Avino et al., 2004) and C. elegans (Canman et al., 2008). Depleting or mutating Rac in the fly eye dramatically suppresses the rough eye phenotype of CYK4 GAP dead–expressing mutants, presumably caused by suppressing cell division failure. In C. elegans, depletion of Rac or Rac effectors that activate the Arp2/3 complex (a nucleator of branched actin filaments) also suppresses the cytokinesis defect of GAP-dead CYK4 mutant embryos (Canman et al., 2008). Thus, Rac inhibition at the cell equator may suppress multiple downstream pathways within the same cell or may control different pathways across species.
The inability of CYK4 to affect RhoA GTP hydrolysis under any condition tested seems to challenge models in which CYK4 acts directly on RhoA but certainly does not preclude an indirect effect. Cross talk among Rho GTPases is well documented in other cellular contexts (Burridge, 1999; Machacek et al., 2009). In motile cells, Rac activity can indirectly affect RhoA activity (Sander et al., 1999), and this could also occur during cytokinesis. Indeed, expression of GAP-dead CYK4 in dividing Xenopus laevis embryos defocuses the zone of active RhoA in the division plane (Miller and Bement, 2009). Therefore, CYK4 down-regulation of Rac may be indirectly shaping the zone of RhoA activity at the cell equator in vivo.
Nevertheless, this newly identified role for CYK4 as a Rac GAP controlling equatorial adhesion disassembly lends further support to a model in which Rac inactivation in the division plane is critical for cytokinesis. The use of an in vivo fluorescence resonance energy transfer activity sensor revealed that active Rac is lower in the division plane than in the polar regions of dividing cells (Yoshizaki et al., 2003). This reduction in equatorial Rac activity was not observed in cells expressing GAP-dead CYK4 (Yoshizaki et al., 2004). Together, these data suggest that CYK4 functions to inhibit Rac activity at the cell equator.
In summary, the Barr laboratory has helped to clarify some of the contentious issues surrounding CYK4 in Rho family GTPase regulation during cytokinesis. Their model, in which equatorial Rac inactivation reduces adhesion at the cell equator (Fig. 2), reveals fresh links among known cytokinetic regulators and effectors and implicates a novel role for ARFGEF7 and PAK1/2 inhibition in cell division (Fig. 1).