During the cell cycle, the cyclin-dependent kinases (CDKs) at the center of the cell cycle clock trigger a diverse set of events, including remodeling of the cell's cytoskeleton. A number of internal surveillance pathways called checkpoint controls assess how key events are progressing and, if there is a hitch in some important process, they signal the clock to wait until the defect is corrected. In the past few years we have learned a lot about how the central clock works. However, several central questions remain concerning how the CDKs actually trigger many of the events, and how the checkpoint controls "know" whether things are proceeding according to plan.

We work with the tractable budding yeast as a model system, allowing us to make rapid progress on complex problems. One focus in the lab concerns a checkpoint pathway called the morphogenesis checkpoint, which monitors cytoskeletal polarization and bud formation, and inhibits G2 CDK activation if environmental stresses affect these processes. We are trying to understand how information about the cytoskeleton and cell shape is sensed and transmitted to the CDKs. A second focus concerns cell polarity, which is switched on by G1 CDKs and switched off by G2 CDKs in yeast. We would like to understand how global CDK activation makes cells develop (or dismantle) an asymmetric cytoskeleton. Because the genes and processes we study are highly conserved, we anticipate that learning the answers to fundamental questions in yeast will be relevant and informative for a wide range of organisms.

Some years ago, we identified a cell cycle checkpoint that we called the morphogenesis checkpoint in budding yeast. Our research on this pathway led us to realize that yeast cells "know" whether or not they have a bud, and whether or not the actin cytoskeleton is intact. If either budding or the actin cytoskeleton is defective, the cells arrest the cell cycle in G2 and delay mitosis until the defect has been corrected (reviewed in Lew, 2003). Although this problem is rarely encountered by yeast cells growing in the lab (where their environment is tightly controlled), it is more frequent in the wild, where changes in temperature, osmolarity, ethanol levels and other factors can stress the cells and (temporarily) stop the budding process. We discovered that failure to bud causes stabilization of an unstable mitosis-inhibitory protein called Swe1p, leading to Swe1p accumulation and arrest in G2 until the problem is corrected.

How does the cell know whether or not it has formed a bud? The Swe1p degradation machinery becomes strikingly concentrated at the neck between mother and bud, attached to a family of filaments called septins. Our current model is that the local cortical geometry of the cell changes upon budding, and that geometry causes a rearrangement of the septin filaments, leading to activation of a checkpoint kinase called Hsl1p that is part of the Swe1p degradation pathway.

Outside of yeast, the septins were largely unknown until recently, but in 2007 several advances brought the septins (which are conserved in all animal cells) into the limelight. Structural studies gave us the first inside look at septins and revealed an unusual septin polymerization mechanism. Disease researchers found increasing numbers of links (as yet quite mysterious) between septins and various diseases including cancer. And investigators working on the mammalian nervous system found that septins form rings at the "necks" of dendritic spines, which mediate synaptic transmission between neurons. Spine number, size, and shape all change in response to stimulation, and septins are well-positioned to detect the local cortical geometry (just as they do at yeast mother-bud necks). These findings raise the tantalizing possibility that the septin cortical sensing we discovered in yeast might have unexpected parallels in more complex cells, possibly even during learning and memory in the mammalian brain.

Current projects investigate how the checkpoint kinase Hsl1p is regulated by septins and other pathways. We would also like to:

• elucidate the signaling pathways linking Hsl1p to Swe1p degradation
• understand how other stresses that affect actin but not septins impact Swe1p
• discern how septin filaments respond to the local cortical geometry
• examine how the checkpoint pathways we discovered are deployed in other systems

Our ultimate goal is to understand how cells monitor their cytoskeleton and cell shape, and how they use that information both for cell cycle regulation and in other contexts.

Cell polarity is a nearly universal feature of eukaryotic cells. A polarized cell has a single, clear axis of asymmetry: a "front" and a "back". This general description encompasses an enormous variety of polarized morphologies, differing between cell types in a single organism as well as between organisms. Thus, it was not clear, a priori, whether regulation of "cell polarity" would entail diverse pathways linked to the diverse morphologies, or a single "master" pathway that would coordinate differing machineries in different cells. In the past several years it has become apparent that the highly conserved Rho-family GTPase Cdc42p, first discovered in yeast, is a component of such a master pathway, employed time and again to promote polarity in different contexts.

Most cells know which way to polarize. Concentration gradients of attractants, repellents, nutrients, or pheromones reveal the optimal directions for successful attack, escape, feeding, or mating. Some cells, including yeast, also carry internal landmarks inherited from their parents that guide polarization without environmental input. Polarization might therefore be viewed as a response to specific external or internal cues. However, it is now apparent that cells can and do polarize even when deprived of directional cues, choosing a random axis and committing to it as if they knew where they were going. This process, called "symmetry breaking", reflects the presence of a core internal polarity program whose direction can be influenced by appropriate cues.

Symmetry breaking is thought to reflect the action of positive feedback loops that reinforce inequalities in the local concentrations of polarity factors, so that stochastic fluctuations are amplified into a single dominating asymmetry. This idea was first suggested by the mathematician Alan Turing in 1952, but the molecular nature of the feedback loops involved in cell polarity has been difficult to pin down. We are addressing this issue in the tractable model eukaryote Saccharomyces cerevisiae (budding yeast), where we investigate how Cdc42p and its collaborators break symmetry.

We have shown that symmetry breaking requires a scaffold protein and involves cycles of GTP binding and hydrolysis by Cdc42p (Irazoqui et al. 2003, 2004, 2005). These cycles require a GEF (guanine nucleotide exchange factor) to catalyze release of GDP and binding of GTP by Cdc42p. Our work suggests a hypothesis in which the scaffold tethers an active GEF to a pre-existing molecule of GTP-Cdc42p. This tethering leads to localized Cdc42p GTP-loading and thus the amplification of a localized concentration of GTP-Cdc42p in a positive feedback loop. Current projects test key predictions of our model, delve into the mechanisms behind GEF activation, and probe the role of the cytoskeleton. We would also like to:

• investigate the roles of other Cdc42p regulators
• devise mathematical models of the process that make quantitative testable predictions
• determine how the cell cycle engine controls the timing of polarization
• ask whether similar mechanisms are at play in other systems

Our ultimate goal is to understand how this core polarity program works, and how it has been adapted to generate the amazing variety of polarized cell behaviors.