Cell cycle regulation of septins: implications for cytokinesis and the spindle position checkpoint

In most eukaryotic cells, cytokinesis is driven by a contractile actomyosin ring, which forms at the site of cell division and drives furrow ingression. In metazoans and fungi cytokinesis requires also other cytoskeletal proteins called septins. Septins are evolutionarily conserved proteins that can hydrolyse GTP and form polymers that assemble into higher order structures, such as filaments and rings. They can interact both with the actin and microtubule cytoskeleton and with membranes. Besides being involved in cytokinesis, septins are involved in many other functions, such as polarized growth, vescicle trafficking, cellular morphology and the creation of diffusion barriers. In budding yeast the first step towards cytokinesis is the assembly of a rigid septin ring composed of five different septins (Cdc3, Cdc10, Cdc11, Cdc12 and Shs1) at the bud neck, the constriction between the mother cell and the bud where cytokinesis takes place. The septin ring, which is formed in G1 at the site of future bud emergence, expands into a broader hourglass structure as cells bud and enter S phase, and it splits into two separate rings just before cytokinesis. These structural changes are accompanied by dynamic transitions. In fact, septins are static (during the “frozen” state) throughout most of the cell cycle and dynamic (during the “fluid” state) in late G1 and prior to cytokinesis. The transition from fluid to frozen state is promoted by protein kinases, such as Gin4 and the PAK kinase Cla4, which localize at the bud neck and directly phosphorylate septins. Conversely, the Tem1 GTPase and the PP2A protein phosphatase likely regulate the reverse transition. The yeast septin ring is also involved in proper spindle positioning, which is in turn crucial for balanced chromosome partitioning. Whenever budding yeast cells experience spindle positioning defects, they undergo anaphase within the mother cell, and then hold on in telophase with elongated spindles and high levels of mitotic CDKs. This cell cycle delay is imposed by the spindle position checkpoint (SPOC), which prevents mitotic exit and cytokinesis until errors are corrected, thus avoiding the generation of anucleate and binucleate cells. The SPOC target is the Tem1 GTPase, whose active GTP-bound form promotes a signal transduction cascade called Mitotic Exit Network (MEN) that ultimately drives cells out of mitosis by leading to inactivation of mitotic CDKs. The dimeric GTPase-activating protein (GAP) Bub2/Bfa1 keeps Tem1 inhibited until the spindle is properly aligned, thus coupling mitotic exit with nuclear division. The Kin4 protein kinase is also involved in the SPOC, by keeping Bfa1 active and by regulating the dynamics of Bub2/Bfa1 at spindle poles. In addition, the Elm1 kinase and the PP2A phosphatase contribute to the SPOC at least partly through Kin4 activation. Tem1 and several downstream MEN components are found at SPBs in a cell cycle-regulated manner and are thought to promote mitotic exit from this location. The Bub2/Bfa1 complex is found predominantly on the SPB that is pulled towards the bud, while it is present on both SPBs of misaligned spindles. Conversely, the Lte1 protein, which positively regulates Tem1, is confined in the bud from the G1/S transition to telophase, when it spreads throughout the cytoplasm of both mother cell and bud. Septins participate in the SPOC by preventing the unscheduled diffusion of Lte1 into the mother cell. Our laboratory had previously implicated the two functionally redundant ubiquitin-ligases Dma1 and Dma2 in proper septin ring positioning, cytokinesis and SPOC regulation. During my PhD, I have been trying to gain insights into the molecular mechanisms by which Dma proteins regulate these processes. Lack of both Dma1 and Dma2 compromises septin ring assembly and has additive effects with the deletion of the PAK kinase CLA4. Indeed, we found that Dma proteins are essential together with Cla4 for septin ring stabilization throughout the cell cycle. Consistently, FRAP (Fluorescence Recovery After Photobleaching) analyses revealed that concomitant deletion of DMA1 and DMA2 increases septin turnover at the bud neck, thus destabilizing the septin ring. Conversely, overexpression of DMA2 stabilizes the septin ring and delays its disassembly at the end of mitosis. Genetic analyses showed that lack of both Dma proteins is lethal when combined with the deletion or mutation in septin genes or genes involved in septin ring assembly and/or stabilization, underlying the importance of these two proteins in the control of septin ring dynamics. We also showed that the role of Dma1 and Dma2 in the SPOC is not due to Lte1 spreading in the mother cell and seems to be also independent of Bub2/Bfa1. Rather, Dma1 and Dma2 appear to control localization of the Elm1 kinase at the bud neck, thus providing a mechanistic explanation for their role in both septin dynamics and SPOC. Being Dma proteins ubiquitin ligases, we hypothesised that they could target for ubiquitylation a regulator of cytokinesis. Therefore, we carried out a genetic screen for extragenic suppressors of the lethality of dma1∆ dma2∆ cla4Δ cells. We got 44 mutants that could potentially identify targets of Dma1/2. We initially focused our attention on the dominant mutations, because some of them suppressed very efficiently the cytokinesis and septin deposition defects of dma1∆ dma2∆ cla4∆ cells. Remarkably, cloning and sequencing revealed that one of the suppressors corresponds to RHO1, encoding for the yeast counterpart of metazoan RhoA. Rho1/RhoA is a conserved GTPase that is required for the assembly and the contraction of the actomyosin ring. The RHO1 mutant allele that we have isolated in our screen is novel (RHO1-D72N) and, strikingly, three additional suppressors carried exactly the same RHO1 mutation. The RHO1-D72N allele is likely hyperactive, since we could also suppress the lethality of dma1∆ dma2∆ cla4∆ cells using known RHO1 hyperactive alleles, such as RHO1-G19V, but not by dominant-negative alleles, such as rho1-D125A. Since protein kinase C (PKC) is a major target of yeast Rho1, we tested whether a hyperactive PKC1 allele (PKC1-R398P) could also suppress the lethality of the dma1∆ dma2∆ cla4∆ triple mutant and found that this was indeed the case. In addition, both RHO1-D72N and PKC1-R398P alleles were able to suppress partially the temperature-sensitivity of cdc12 mutants, which undergo septin ring disassembly at high temperature. These data strongly argue that Rho1/Pkc1 hyperactivation stabilizes the septin ring. Indeed, genetic and FRAP analyses showed that the septin ring is more stable in RHO1 and PKC1 hyperactive mutants, while lack of these proteins destabilized it. Finally we also found that the deletion of RTS1, which promotes septin ring disassembly at the end of the cell cycle, suppresses the lethality and septin ring defects of dma1Δ dma2Δ cla4Δ cells. Altogether, our data suggest that Dma1/2 might promote septin stabilization by activating Elm1, the Rho1/Pkc1 pathway and/or inhibiting PP2ARts1, directly or indirectly. Whether these proteins are direct ubiquitylation targets of Dma1 and Dma2 is an important issue to be addressed in the future.