Single gene control of a complex phenotype hangs in the balance

Complex developmental events in eukaryotes are often controlled by complex regulatory pathways. From yeast to humans, the number of intricate transcriptional regulatory circuits is astounding (1). It is therefore striking that in the human fungal pathogen Candida albicans, two groups have independently discovered that a complex cellular switch and subsequent epigenetic behaviors are controlled almost exclusively by a single gene product (2, 3).

Two studies in this issue of PNAS by Zordan et al. (2) and Huang et al. (3) describe the discovery of a key transcriptional regulator (Wor1) and propose remarkably simple models to explain the mechanisms by which it controls the process of both establishing and maintaining alternative growth states (“white” and “opaque”). The ability to change between these two growth forms has recently been discovered to play key roles in both mating and pathogenesis of C. albicans (4). This morphological change, known as white–opaque switching, was discovered nearly 20 years ago (5), but its significance in the biology of the organism has only recently come to the fore.

White–opaque switching is a rare, stochastic switch between two growth forms. In the white form, cells grow as round, smooth, budding yeast and form white, domed colonies on solid agar. In the opaque form, cells grow as elongated, stippled cells and form flat, gray colonies (6). These growth forms have distinct growth properties and radically different behaviors. These differences have been noted in two critical areas of Candida biology. First, white and opaque cells have different pathogenic and growth properties: white cells are found more often in systemic infections, and opaque cells are associated with cutaneous infections (7). The opaque phase is unstable at higher temperatures and reverts to the white phase, whereas at lower temperatures, it is much more stable (5). Second, the white and opaque phases have distinct mating properties. C. albicans was long-thought to have lost its sexual cycle, but work over the last several years has revealed a cryptic mating reaction with fascinating features (8, 9). Unlike other fungi that mate with one another in their “regular” growth forms, C. albicans must undergo a morphological transition before the cells are competent to mate with one another (10). Opaque cells mate 1 × 10⁶ times better than their white counterparts, and mating is extremely efficient in a mouse model of skin infection, suggesting a relationship between pathogenesis and mating (11).

The processes of mating and switching are also intertwined by the fact that they are controlled by common regulatory proteins. In C. albicans, cell type is maintained by the transcriptional regulators a1 and α2 (homologs of the Saccharomyces cerevisiae mating type proteins with the same names) (8, 12). These homeodomain transcription factors are the key factors from the MTL (mating type-like) locus that controls mating behavior (13). In the presence of a1 and α2 together, cells cannot undergo mating; however, in the absence of either factor, the cells adopt an a or α cell identity. To mate efficiently, the a and α cells must then switch to the opaque form. The white–opaque switch is also inhibited by the presence of both a1 and α2 (10). For reasons that are not yet apparent, C. albicans has evolved to place white–opaque switching under the control of MTL locus proteins.

To gain insights into the white–opaque switching mechanism, Zordan et al. (2) investigated genes that were known from previous studies to be regulated directly by a1–α2 (14). They created deletion strains for each of the directly regulated genes and evaluated their ability to carry out the white–opaque switch. Of the six deletion strains tested, only one lost the ability to switch from white to opaque. The deleted gene, designated WOR1 for White–Opaque Regulator 1, was absolutely required for switching from the white form to the opaque form, and cells lacking WOR1 were locked in the white phase. Huang et al. (3) also identified WOR1, but their approach was quite different. They carried out a screen in S. cerevisiae to identify C. albicans genes that could restore the ability of specific mutants in invasive growth to adhere to solid agar. WOR1 allowed a bypass of key invasion genes by transcriptionally up-regulating downstream targets in S. cerevisiae and restored the ability of the mutant strain to invade the agar surface. To investigate its function in C. albicans, Huang et al. created a wor1Δ strain and also found a defect in white–opaque switching. Both sets of authors (2, 3) investigated the mechanism of WOR1 regulation using a series of molecular genetic experiments. In each case, the authors ectopically expressed WOR1 in white cells and observed a whole-scale switch to opaque cells. They also observed that the endogenous copies of WOR1 were induced, suggesting that Wor1 controls its own expression. In addition, both groups showed that ectopic expression of WOR1 can override a1–α2 repression of the WOR1 promoter.

In nonoverlapping experiments, Zordan et al. (2) carried out chromatin immunoprecipitation assays and found that Wor1 is in fact associated with its own promoter in four distinct locations. These data suggest that Wor1 is a DNA-binding protein (although Wor1 contains no known DNA-binding motif) or that Wor1 resides in a complex with proteins that bind DNA. Finally, Zordan et al. show that a pulse of Wor1 is sufficient to both initiate and sustain a switch from white to opaque, solidifying the idea that Wor1 acts in a continuous feedback loop, regulating its own promoter to induce WOR1. Huang et al. (3) addressed patterns of WOR1 expression in the cell population using a WOR1-GFP expression construct to visualize single cells. In this experiment, they observed that opaque cells displayed strong Gfp fluorescence and white cells were not detectably fluorescent, demonstrating that WOR1 expression is all-or-none, consistent with a bistable regulatory system (15).

These articles address two particularly important concepts. First, Zordan et al. (2) discuss the idea that epigenetic events can be controlled in multiple ways. Changes in chromatin structure have been identified as a mode of action for many epigenetic events (16); however, the model of a simple feedback loop in which an activator acts at its own promoter can achieve the same epigenetic effects (as seen here). Second, Huang et al. (3) frame the stochastic switching mechanism in terms of bistability (15). The principle concept of bistability is that a regulatory system controls the switch between two states, but the system cannot rest in intermediate states. In addition, bistable regulation occurs in a stochastic manner in a cell population. That is, the choice of which cells have altered gene expression is random. These properties have been observed in other systems such as Bacillus subtilis (competence) (17) and Xenopus laevis (oocyte maturation) (18). Both the complete switch between the white and opaque phases and the stochastic nature of the switch observed by Zordan et al. (2) and Huang et al. (3) support a model of bistable regulation of white–opaque switching in C. albicans.

Fig. 1[adapted from Zordan et al. (2)] shows a speculative model of how WOR1 could function as a bistable switch with epigenetic properties. In this model, white cells contain only low levels of the Wor1 protein. By an unknown mechanism, levels of the protein are held in check, perhaps by keeping occupancy of direct binding sites in the WOR1 promoter low. With only a small amount of the activator present, occupancy of the WOR1 promoter does not reach the threshold levels necessary for activation of high-level expression. At some low, random frequency, protein levels become high enough to allow greater occupancy of the WOR1 promoter and thus promote high-level expression of WOR1. Once this occurs, there is a switch to the opaque phase. [This switch is known to occur approximately once in 10,000 generations (2).] The opaque phase persists in a heritable manner until at some low frequency in the population, protein levels drop below the threshold required to maintain high levels of WOR1 expression, and the cell then switches to the white phase.

Fig. 1.

Integrated model of WOR1-controlled switching. In white a or α cells there is little Wor1 protein (blue circles) produced, causing the cells to remain in the white phase. In a rare, random event, the protein levels increase, activate more WOR1 expression (big green arrows) in an autoregulatory feedback loop, and cause a switch to the opaque phase (approximately once per 10,000 generations). Subsequent generations are primarily in the opaque phase because enough protein is inherited to sustain high-level expression of WOR1. However, rare, random events in which the Wor1 protein levels drop below a critical threshold lead to low levels of WOR1 expression (small gray arrows) and the production of white cells in the population. Changing concentrations of Wor1 tip the balance between two states in a bistable mechanism of control. Adapted from Zordan et al. (2).

By tipping the balance of Wor1 levels one way or the other, the cells have a switch-like behavior that could govern the capacity of the population to respond to changing environmental conditions. In a system such as this, individual cells within a population can exhibit distinct expression patterns. If a small portion of the population is already in the switched state, some cells are primed to respond to changing environmental conditions, such as a shift in location in the mammalian host (e.g., from bloodstream to skin where mating could occur). Obviously, there are many interesting questions that remain to be explored. For example, how is the switch mechanism regulated? Is there a mechanism to prevent runaway expression of WOR1 in white cells (such as the promoter occupancy model shown in Fig. 1) or is there an active mechanism to prevent autoregulation by Wor1 in white cells that has a low frequency of failure? Questions also remain with regard to how the system responds to other environmental signals, such as regulation by a1–α2 or changes in temperature that lead to switching. Perhaps the stability of the Wor1 protein is compromised at higher temperatures, leading to less functional protein and more switching to the white phase. Future studies into the kinetics of Wor1 protein stability will certainly lead to interesting findings regarding eukaryotic responses to different environmental conditions.

Although bistability has been studied in several prokaryotic systems (15), examples in eukaryotes are less common. The white–opaque switching system in C. albicans represents one of the most tractable eukaryotic systems for studying bistability in a biologically relevant context. The studies discussed here present clear insights into the mechanisms of a bistable switching system and open the door to further understanding and exploration of novel eukaryotic regulatory systems.