White-opaque switching in Candida albicans: cell biology, regulation, and function

SUMMARY

Candida albicans remains a major fungal pathogen colonizing humans and opportunistically invading tissue when conditions are predisposing. Part of the success of C. albicans was attributed to its capacity to form hyphae that facilitate tissue invasion. However, in 1987, a second developmental program was discovered, the “white-opaque transition,” a high-frequency reversible switching system that impacted most aspects of the physiology, cell architecture, virulence, and gene expression of C. albicans. For the 15 years following the discovery of white-opaque switching, its role in the biology of C. albicans remained elusive. Then in 2002, it was discovered that in order to mate, C. albicans had to switch from white to opaque, a unique step in a yeast mating program. In 2006, three laboratories simultaneously identified a putative master switch gene, which led to a major quest to elucidate the underlying mechanisms that regulate white-opaque switching. Here, the evolving discoveries related to this complicated phenotypic transition are reviewed in a quasi-chronological order not only to provide a historical perspective but also to highlight several unique characteristics of white-opaque switching, which are fascinating and may be important to the life history and virulence of this persistent pathogen. Many of these characteristics have not been fully investigated, in many cases, leaving intriguing questions unresolved. Some of these include the function of unique channeled pimples on the opaque cell wall, the capacity to form opaque cells in the absence of the master switch gene WOR1, the formation of separate “pathogenic” and “sexual” biofilms, and the possibility that a significant portion of natural strains colonizing the lower gastrointestinal tract may be in the opaque phase. This review addresses many of these characteristics with the intent of engendering interest in resolving questions that remain unanswered.

INTRODUCTION

Most microorganisms, both prokaryotic and eukaryotic, undergo changes in phenotype during development and, in many cases, in response to environmental signals. The regulation of these changes can be genetic, at the level of DNA reorganization, or epigenetic, at the level of gene regulation without DNA reorganization. Microorganisms have used a variety of strategies to effect phenotypic change, and the human pathogen Candida albicans is no exception. C. albicans replicates either as a budding, round to ellipsoidal yeast or as a multicompartmented filament (hypha) (1). The reversible transition between these two morphologies has been intensively investigated because of the valid perception that the filamentous form represents a major vehicle of tissue invasion, a key step in pathogenesis. And because of the rapid reversibility between the two growth forms and no proof to argue otherwise, it has been the general perception that the transition between the two morphologies is regulated epigenetically. Through the latter half of the 20th century, the bud-hypha transition remained the major focus of research related to the developmental biology of C. albicans. However, lingering in the background were the rare observations that low-density clonal cultures of C. albicans plated on nutrient agar gave rise to variant colony morphologies that differed in the spatial distribution of yeast cells, hyphae, and pseudohyphae (intermediate between yeast and hypha forms) within colony domains (2–6). However, in 1985, reversible, high-frequency switching between multiple-colony morphologies was demonstrated in laboratory strains of C. albicans (7, 8), and soon after, a variety of reversible high-frequency switching systems were revealed in clinical isolates (9). These observations proved to be a rediscovery of the previously reported phenomenon (2–6), which had received little attention over the years.

One of these newly observed switching systems, referred to as “the white-opaque transition,” or “white-opaque switching,” was first observed in blood and lung isolates obtained from a transplant patient with systemic candidiasis (10). Of the several distinct switching systems that had been identified by 1987, white-opaque switching was selected as a model for investigating in depth the underlying molecular mechanisms of switching for several reasons. First, it was a simple phase transition between only two alternative phenotypes (10), rather than between six phenotypes, as was the case for the reversible switching system in strain 3153A, first reported by Slutsky et al. in 1985 (8). Second, it was spontaneous and reversible through sequential rounds of switching and could occur without affecting the frequency of subsequential switching (10, 11). Third, it was identified in a significant portion of a collection of relatively fresh clinical isolates (9). And fourth, it involved a unique cell phenotype, the “opaque phenotype,” which was easily distinguished from the white cell phenotype at the cellular as well as colony level. During the 5 years that followed the discovery of white-opaque switching, it was demonstrated that cells of the opaque phase exhibited a wide variety of phenotypic characteristics that distinguished it from cells of the white phase. But in spite of the numerous and, in some cases, unique characteristics of the opaque phenotype, including the identification of the first genes differentially expressed in the alternative phenotypes (12–14), white-opaque switching remained an enigma for more than a decade following its discovery primarily because it was not clear what function it served in the life history and pathogenicity of C. albicans. But that changed by a seminal set of observations made soon after the identification of the C. albicans mating-type locus (MTL) in 1999 (15). In 2002, 15 years after the discovery of white-opaque switching and 3 years after the discovery of the mating-type locus, Miller and Johnson (16) reported first that the zygosity of the mating-type locus regulated the capacity to switch and, second that in order to mate, cells had to switch from white to opaque. Therefore, to mate, natural C. albicans strains which are primarily diploid and a/α at the mating-type locus, first had to undergo homozygosis at the mating-type locus, then switch from white to opaque. The question of what function white-opaque switching played therefore appeared to have been answered and attention then focused on the regulation of switching at the level of transcription. Networks of interacting transcription factors (TFs) were identified which regulated switching (17, 18), “sexual” white cell biofilms were discovered, which were induced by pheromone released by minority opaque cells and which facilitated mating (19), and mutations in genes repressing switching in a significant minority of clinical MTL-heterozygous (a/α) strains resulted in expression of the opaque phenotype at sites of colonization (20). In this review, selective details of this evolving story will be reviewed in somewhat chronological order when possible, with the intent of pointing out what we know as well as do not know about white-opaque switching. Hopefully, this review will engender interest in the ever-evolving story and result in experiments that further elucidate the molecular and cytological changes that regulate the transition. The review is semi-chronological with landmark observations listed by date in Table 1 to provide the reader with a historical perspective of the sequence of discoveries limited to or impacting white-opaque switching.

TABLE 1.

Chronology of landmark discoveries that impacted our understanding of the cell biology, regulation, and function of the white-opaque transition^a

Year	Discovery(ies)
1985	Rediscovery of high-frequency, reversible switching of colony morphology (7, 10).
1987	Discovery of white-opaque switching at the University of Iowa Hospitals and Clinics (8).
1987	Unique opaque cell morphology; the unique opaque cell wall pimple (11, 21).
1990	Antigenic differences between white and opaque cell surface (11, 21).
1992	Identification of first phase-specific genes (12–14).
1997	Difference in virulence between white and opaque in mouse model for systemic infections (22).
1999	Discovery that the transacting factor EFG1 represses white to opaque switching (23, 24).
1999	Opaque cells, but not white cells, colonize skin in a mouse cutaneous model (22).
2001	Acetylation of histones affects the frequency of switching (25–28).
2002	Switching involves the regulation of a large number of genes (29–31).
2002	Switching is a prerequisite for mating (16).
2002	MTL homozygosis is a prerequisite for mating (16, 32).
2002	Species related to C. albicans undergo white-opaque switching (33, 34).
2003	Skin facilitates opaque cell mating (35).
2004	Opaque cells do not release white blood cell chemoattractant and evade the cellular immune system (36–39).
2004	Switching can occur in a/α cells with a hemoglobin response gene mutation (40).
2006	Simultaneous identification of the “master switch gene” WOR1, a positive regulator, by three research groups (24, 41, 42).
2006	Identification of self-sustaining WOR1-positive feedback loop (41).
2006	Minority opaque cells stimulate a majority of white cells to form a sexual biofilm (19, 43).
2007	Development of a TF model of “interlocking feedback loops” regulating WOR1 expression and white-opaque switching (17, 44).
2008	Passage of white cells through a mouse gut induces switching (45).
2009	High CO₂ and GLcNAc support switching from white to opaque (45–48).
2010	Identification of numerous complex changes in the transcriptome accompanying switching (31).
2013	Discovery that a significant proportion of a/α strains can switch to opaque (39, 49).
2014	Discovery of a “gray” cell phenotype and role of EFG1 mutations (20, 49–51)
2015	pH regulates white-opaque switching (52).
2020	Double deletion mutants of EFG1 and WOR1 can switch en masse to opaque; hypothesized alternative opaque pathway (53).
2021	Three of the five TF-binding sites along the UIR of WOR1 are individually essential for opaque commitment in the white-to-opaque transition (54).

^a References begin with the first observation followed by subsequent observations expanding the discovery. References are numbered according to appearance in the text.

DISCOVERY OF WHITE-OPAQUE SWITCHING

The story of white-opaque switching began in 1979 when Bedell and Soll (55) described a frequent variant of C. albicans that, unlike its parent strain, formed hyphae at stationary phase in liquid medium and at the periphery of clonal colonies grown on nutrient agar containing a growth-limiting concentration of zinc. Five years later, in 1984, in a chapter on the role of zinc in C. albicans dimorphism (56), it was noted that the M10 phenotype appeared in populations of C. albicans strain 3153A after irradiation with low doses of UV, at frequencies far too high to be explained by point mutations. In this review article (56), it was proposed that because C. albicans was diploid, the frequent appearance of the M10 variants was very likely the result of mitotic recombination (57), resulting in homozygosis of a recessive gene, that gave rise to the M10 phenotype. Having read the review article, the author’s Ph.D. advisor called him and noted that it was simple enough to test the proposed hypothesis. If correct, the mass plating of M10 cells should give rise to the wild-type phenotype (i.e., colonies with no hyphal perimeter) at extremely low frequencies, consistent with the low frequency of mutation. When tested, the opposite was observed (8). A mixture of six variant colony phenotypes (“star,” “ring,” “irregular wrinkle,” “stipple,” “hat,” and “fuzzy”) arose at frequencies of approximately 10⁻⁴. Plating of clonal populations of the variants revealed that each switched interconvertably to the other five phenotypes as well as reversibly to the parental smooth colony phenotype at frequencies varying from 3 × 10⁻⁴ to 6 × 10⁻². This “3153A-type” switching system (8), named after the a/α strain employed, was therefore reversible, and the variants were interconvertible at high frequency. But this was not the first time such spontaneous colony variants had been observed. As early as 1932, Negroni (2) described a high-frequency “rough” colony variant. In 1968, Brown-Thomsen (4) identified 15 colony variants in cultures stored for up to 6 months. Studies by Brown-Thomsen (4) and Ireland and Sarachek (6) revealed physiological differences between parental and variant offsprings. In most of these early studies, reversibility of variants was usually not established, with exceptions such as the study by Ireland and Sarachek (6). Finally, in 1985, a study by Pomes et al. (7) demonstrated that a rough colony variant switched back to the parental smooth colony phenotype at extremely high frequencies and that this switching system, like the 3153A switching system described by Slutsky et al. (8), was stimulated by low doses of UV. Together, the studies by Slutsky et al. (8) and Pomes et al. (7) brought to the attention of the medical mycology community observations and concerns that there were instabilities of general phenotype other than the bud-hypha transition that could play a role in colonization, virulence, and pathogenesis.

In 1986, Soll and colleagues, realizing that the 3153A switching system was identified because of the growth-limiting zinc content of the employed nutrient agar (55, 56), set out using this agar to assess the variety of switching systems, switching frequencies, and cellular phenotypes for a large collection of clinical C. albicans isolates. Although this screen was never fully completed, a number of switching systems were identified, including variations of the 3153A switching system, a petite-normal colony transition, a smooth-heavy myceliated colony transition, the rough-smooth colony transition already described by Pomes et al. (7), and a white-opaque transition (8). Realizing that there may be multiple switching systems functioning in the same cells and that some switching systems like that of strain 3153A might include too many variant phenotypes for mechanistic studies, Soll and colleagues focused on the biphasic, reversible white-opaque switch observed in a number of clinical isolates, most notably in collections of vaginal and blood isolates. The transition occurred at high frequency and was reversible between a white, hemispheric colony (the “white phase”) and a gray, flat, slightly larger colony (the “opaque phase”), when assessed on supplemented Lee’s agar (55) at 25°C in air, after 5–7 days of incubation (10). The strain selected for further analysis and first observed to undergo the transition was WO-1 (“white-opaque one”), which was isolated from the blood and lungs of a transplant patient suffering from systemic candidiasis (10). WO-1 cells gave rise to a majority of colonies of the white phenotype and a minority of colonies of the alternative opaque phenotype. Cells also gave rise at far lower frequencies to irregular wrinkled and fuzzy colonies, the former due to hypha formation within colonies and the latter to aerial hyphae (10). In 1987, quite ignorant of the seminal role it played in mating (16, 58), white-opaque switching was fortuitously selected as the model system for investigating the molecular mechanisms regulating switching. Soon after its discovery, James Hicks, then at Cold Spring Harbor Laboratory, found that phloxine B, a vital dye, differentially stained opaque cells red (11) (Fig. 1A through D), which facilitated its use as an experimental system.

Fig 1.

Examples of white-opaque switching in clonal colonies formed on nutrient agar containing phloxine B, a vital dye which selectively stains opaque cells red. (A) White colony (W) formed by a white cell exhibits a late switch from white to opaque, generating a small single opaque sector (op). (B) White colony (W) with two white to opaque switches forming two opaque sectors (op). (C) White colony (W) with an early switch generating a large opaque sector. (D) Opaque colony (op) with an early white sector (wh) surrounded by faster spreading opaque cells. Reprinted from reference (11).

FREQUENCY OF WHITE-OPAQUE SWITCHING

The original frequencies of white-opaque switching were not assessed at the cellular level but rather at the colony level. Cells from white colonies were clonally plated on supplemented Lee’s nutrient agar, which contains glucose as the carbon source, and were then incubated for 5–7 days at 25°C in air. The number of white and opaque (switched) colonies were then counted. In the first studies with strain WO-1, the average frequency of homogeneous opaque colonies that formed was approximately 6 × 10⁻³, and the frequency of white colonies with opaque sectors was approximately 3 × 10⁻² (10, 11). The mutation frequency for a haploid yeast is approximately 10⁻⁶. It was therefore reasonable to conclude at the time that the frequency of switching from white to opaque was many orders of magnitude higher than that of mutation. In subsequent serial platings of cells from white colonies, the average frequency of switching was approximately 3 × 10⁻² (11). In parallel platings of cells from clonal opaque colonies, the average frequency of switching from opaque to white was approximately three times higher than that of white to opaque. In initial switching experiments (10), it was demonstrated that the transition was sequentially reversible. But there were three caveats to these initial results. First, in the original plating experiments by Slutsky et al. (10), the samples from the patient colonized with strain WO-1 were cultured in sequence on three different agars before clonal plating on supplemented Lee’s agar, which may explain the lower initial frequency obtained in the first plating beginning the sequences. This was the first observation that suggested the contents of nutrient media used to grow the initial cell populations may affect subsequent switching frequencies. Second, it was observed that multiplying opaque cells spread along an agar surface faster and therefore more extensively than white cells due to their long pseudohyphal-like shape and apical pseudohypha-like pattern of bud formation (11, 59). For this reason, most sectored colonies were primarily white with opaque sectors (Fig. 1A through C) rather than opaque with white sectors. On occasion, opaque colonies with white sectors were observed. Usually, these sectors were completely surrounded by opaque cells, the sectors appearing as white islands (Fig. 1D). Third, in one of two sequential plating lineages (11) performed a year after the report by Slutsky et al. (10), the frequency of switching decreased progressively, suggesting a heritable but progressive change, which may explain why some WO-1 samples sent to colleagues and some of the clones of WO-1 that had been transferred serially exhibited decreases in the frequency of switching (11).

Realizing that switching at the colony level did not accurately reflect frequencies at the cellular level, Rikkerink et al. (60) assessed switching frequencies by generating minicultures from single cells and calculating the zeroth term of the Poisson distribution, using the Luria-Debruck fluctuation formula, which was originally developed to test whether mutations appeared in the absence of selective pressure (i.e., spontaneously). The experiments gave estimates of 10⁻⁵ to 10⁻⁴ per cell generation for white to opaque switching and 5 × 10⁻⁴ per cell generation for opaque-to-white switching. These were frequencies far lower in both directions than those indicated by earlier experiments based upon colony frequencies on agar (10, 11). Bergen et al. (61) then performed continuous video analyses at high magnification under phase contract microscopy, beginning with single cells plated on films of supplemented Lee’s agar cast on the bottoms of Petri dishes or on glass slides. Switching events were identified by a change in cell morphology. All of the cells in microcolonies emanating from single white or opaque cells were followed to approximately 1,000 cells and the proportions of opaque and white cells assessed. Bergen et al. (61) found that the proportion of white-origin microcolonies with one or more opaque cells was 4.4 × 10⁻³, and the proportion of opaque-origin microcolonies with one or more white cells was 2.5 × 10⁻¹. A mathematical model was used to determine the probability of a switch in either direction and to estimate the number of generations preceding a switch (61). The probability that a white cell generated an opaque daughter cell was 1.7 × 10⁻⁵, and the probability that an opaque cell generated a white daughter cell was 1 × 10⁻¹. Bergen et al. (61) concluded that, under their conditions, the expected average number of generations before a white cell formed an opaque daughter cell was 15.8, and the expected number of generations before an opaque cell formed a white daughter cell was 3.4. While the two studies estimated similar rates of switching from white to opaque, 10⁻⁵ to 10⁻⁴ and 1.7 × 10⁻⁵, respectively, the rates of reversion from opaque back to white differed markedly (5 × 10⁻⁴ versus 10⁻¹, respectively).

Subsequent to these initial studies, it became increasingly clear that the stability of the opaque phenotype was highly sensitive to a number of environmental conditions, which may explain the difference between the results of Rikkerink et al. (60) and Bergen et al. (61). The first study of white-opaque switching (10) demonstrated that a shift from 24°C to 4°C induced one-third of opaque cells to switch to white, and a shift from 24°C to 37°C caused mass conversion (>90%) of opaque to white. Similarly, Rickkerink et al. (60) demonstrated that increasing the temperature from 24°C to 34°C for 15 hours caused over 90% of opaque cells to form white colonies when returned to 24°C. Since these initial temperature studies, it has been demonstrated that other conditions, including pH (52), CO₂ levels (46), low levels of UV (62), and the nature of the carbon source (47), affect the frequency of white to opaque and opaque to white switching. Therefore, it should be kept in mind when comparing data from different studies that frequency measurements are highly dependent upon and therefore may be specific to the conditions employed.

UNIQUE OPAQUE CELL ARCHITECTURE: THE PIMPLE

As noted, one of the major reasons that led to the selection of the white-opaque transition as a model system for investigating the molecular mechanisms regulating phenotypic switching was the unique morphology of the opaque cell. White cells of C. albicans strain WO-1 grown in liquid medium or on nutrient agar were round to slightly ellipsoidal (Fig. 2A) (10, 11, 21), similar in shape to yeast phase cells of standard laboratory and clinical strains that do not undergo the white-opaque transition. White cells of strain WO-1 contained a single nucleus, granular cytoplasm, multiple mitochondria, a smooth cell wall, and a small or indistinguishable vacuole (11, 21). In marked contrast, opaque cells were bean shaped, usually asymmetric, possessed a pimpled cell wall, were larger than white cells, and contained a large dominant vacuole (Fig. 2B through D) (10, 11, 21). In most other natural strains that undergo the transition, the shapes of white cells were similar to those of strain WO-1, but in a minority of strains, the opaque cells were rounder than WO-1 opaque cells. While mature white cells of strain WO-1 grown on supplemented Lee’s medium at 25°C in air exhibited an average volume of 33 µm³, mature opaque cells exhibited an average volume of 114 µm³, a threefold difference (11), and a mass on average twice that of white cells (10). However, the DNA content of white and opaque cells was similar (10) and based on the DNA content of Saccharomyces cerevisiae, consistent with a diploid genome.

Fig 2.

The opaque-specific pimples at the surface of opaque cells distinguished by scanning electron microscopy. (A) Mature white cell and attached daughter cell with smooth surfaces. (B) Opaque cell with pimpled surface and extracellular vesicles (EVs) protruding from pimples. (C) Opaque cell with EVs protruding from wall without distinct pimple mounds and an early daughter cell devoid of EVs in panel A. (D) Opaque cell with pimples, which has formed a hypha with smooth surface. Arrow head in panel A, bud scar. Arrows in panels B, C, and D, pimples or pimples and EVs. Reprinted from reference (21). Scale bars, 1 µm.

Perhaps the most fascinating and unique morphological characteristic of the opaque cell is the presence of cell wall pimples (Fig. 2B through D) (11). Scanning electron microscopy (SEM) revealed that while the walls of mature white cells were smooth, except for birth and bud scars (Fig. 2A), the walls of mature opaque cells were covered with pimples (Fig. 2B through D). The pimples appeared as mounds with or without small apical blebs, or vesicles. In some micrographs, the cell surface was covered by blebs without visible mounds (Fig. 2C), but the blebs exhibited distribution pattern similar to mounds (21). Pimples were not present on opaque daughter cells until they attained approximately two-thirds of their final mother cell volumes (Fig. 2C). The final average number of pimples per mature opaque cell ± standard deviation was 141 ± 39 (21). Thin sections of TEMs verified the presence of these unique pimples and apical vesicles (Fig. 3) and revealed that each pimple possessed a single central channel traversing it, from the cytoplasm to the outer edge of the wall. In some micrographs, vesicles were found embedded in the cytoplasm just under the pimple. In a few micrographs, a vesicle was observed associated with the bottom or apical end of a pimple channel, and in a few micrographs, a vesicle appeared to stretch through the pimple channel (21). The architecture of the cell wall pimple suggests it is involved in the extrusion of extracellular vesicles (EVs). This is an intriguing possibility in light of the possible role of extracellular vesicles in a number of diverse biological processes including infections caused by fungi. In C. albicans, intriguing aspects of EVs are their possible associations with C. albicans biofilm formation and matrix-dependent drug resistance (63, 64) and, as noted later in this review, mating.

Fig 3.

Transmission electron micrograph of a longitudinal section of an opaque cell. Cell wall pimples are indicated by arrow heads. The enlarged vacuole is filled with vesicles. Examples of vesicles in vacuole are enlarged in the upper right corner of the figure (A, B, C, D). Panel D shows smaller vesicles within a larger one. Scale bars, 1.0 µm for main micrograph and 0.1 µm for panels A, B, C, and D. Reprinted from reference (21).

Transmission electron microscopy (TEM) studies suggested that pimple-associated EVs may originate in the enlarged opaque cell vacuole. Both phase contrast microscopy and TEM revealed that mature opaque cells contain a giant vacuole, which in mature opaque cells presses the nucleus, granular cytoplasm, and mitochondria into a cortical cytoplasmic shell below the cell wall (Fig. 3) (11, 21). The dominating opaque vacuole is filled with vesicular structures (Fig. 3), which in some TEMs appear to have collapsed due to fixation (21). Many of these intravacuolar vesicles are of similar size to the vesicles associated with pimples (21) (Fig. 3, major micrograph and minor panels A, B, and C). The mean diameter ± standard deviation of vesicles in vacuoles is 46 ± 17 nm, approximately twice the diameter of vesicles observed at the pimple apices. However, TEM studies suggest that vesicles narrow as they traverse the pimple channel, and the blebs measured at the pimple apices may be only a portion of the full vesicle. These observations (21) support the possibility that the vesicles in the vacuole may be the source of the pimple vesicles.

The elongated nature of opaque cells dominated by a large vacuole (11, 21) is reminiscent of the architecture of hypha compartments. It is therefore logical to address whether opaque cells share other traits with the filamentous growth form. Studies performed soon after the discovery of the opaque phenotype suggested that there were both overlapping and unique opaque and hypha characteristics. First, when a white yeast phase cell forms a white phase bud, there is a constriction at the mother cell-bud junction, the site of septum formation (Fig. 4A). However, when a yeast phase cell forms a germ tube (incipient hypha), there is no constriction or septum formed at the mother cell-bud junction (65) (Fig. 4B). In the case of evagination of an opaque cell, there is a constriction (Fig. 4C), as is the case for white yeast phase cells (Fig. 4A). Second, whereas the maturing opaque daughter cell forms cell wall pimples (Fig. 2B through D), hyphae do not form pimples (Fig. 2D) (59). Third, F-actin granules in the growing opaque cell cortex (60) (Fig. 4C) distribute in a pattern intermediate between that in a growing yeast phase daughter cell (Fig. 4A) and an elongating hypha (Fig. 4B) (65). Fourth, mouse antiserum generated against heat-fixed opaque cells, when absorbed with nonpermeabilized hyphae, loses antibodies that commonly bind to the cell surfaces of both hyphae and opaque cells, but not the surface of white cells, indicating that opaque cells and hyphae share surface antigens not present on white cells (21). And fifth, although opaque phase cells form hyphae, they do so in response to signals different from those that stimulate hypha formation in white cells (59). Moreover, when opaque hyphae that form at pH 6.7 at 37°C are transferred to medium with a pH of 4.5 at 25°C, conditions that induce the transition from hypha to budding yeast, opaque hyphae form opaque-shaped daughter cells, not white-shaped daughter cells, indicating that hyphae formed by opaque cells, though morphologically indistinguishable from those formed by white cells, are still in the opaque phase (59). These observations, reported in the 3 years following the discovery of the white-opaque transition, distinguished it from the bud-hypha transition.

Fig 4.

A cytological comparison between white daughter cell formation (A), hypha formation and compartmentalization (B), and opaque daughter cell formation (C). Note, in particular, differences in cell shape, the distribution of actin granules, septum localization, and cell separation. Symbol key below represents cell formation panels.

OPAQUE-SPECIFIC CELL SURFACE ANTIGENS

In the 5 years following the discovery of the white-opaque transition in 1987 (66), it was demonstrated that white and opaque cells differed in (i) their capacity to adhere to buccal epithelial cells (59), (ii) hydrophobicity (67), (iii) sterol content (68), (iv) susceptibility to antifungal agents (69), (v) sugar assimilation pattern (70), (vi) release of acid protease (12), and (vii) cell wall and vacuole ultrastructure (11, 21). Given the extent of the phenotypic changes accompanying the white to opaque switch, it was assumed in 1987, the year of discovery, that it involved the differential expression of genes. The first indication that white and opaque cells expressed phase-specific genes was the identification of one or more opaque-specific antigens on the opaque cell surface (21, 59). A polyclonal rabbit antiserum generated against heat-fixed opaque cells stained both heat-fixed, nonpermeabilized white and opaque phase cells, and did so uniformly across the surfaces of both cell types, indicating wall antigens existed that were common to both white and opaque cells. Absorption of the serum with nonpermeabilized, heat-fixed white cells resulted in an antiserum that stained hyphae and opaque cells but not white cells. When absorbed with white cells forming hyphae to remove antibodies against both white cells and hyphae, the subtracted antiserum stained opaque cell surfaces but not white cells or hyphae, and did so in a punctate fashion, suggesting that the antiserum contained antibodies that targeted cell wall pimples containing opaque-specific antigens. To verify that opaque-specific antigens were located at the pimple, nonpermeabilized fixed opaque cells were stained with anti-opaque cell rabbit antiserum. The opaque cells treated with this antiserum were in turn treated with goat anti-rabbit antibody conjugated to colloidal gold (21). TEM revealed localization of gold at the apices of pimples (21). Western blot analyses of cell lysates labeled with the anti-opaque cell antiserum revealed three opaque-specific antigens at 31, 21, and 14.5 kd. When antiserum absorbed first with nonpermeabilized white cells and then absorbed with nonpermeabilized opaque cells was used to stain cell lysates in westerns, only the antibodies against the 14.5 kd antigen were absent, indicating that this opaque-specific antigen was located on the opaque cell surface (21). Subsequent preliminary studies suggested that the antigen might be a tyrosine kinase but that conclusion remains tentative (T. Srikantha and D. R. Soll, unpublished data) and deserves further exploration.

IDENTIFICATION OF THE FIRST PHASE-SPECIFIC GENES

Although the anti-opaque antiserum studies (11, 21) suggested that white-opaque switching involved differential gene expression, more direct evidence was needed. In a 1991 review article on switching, Soll et al. (69) published images of two-dimensional polyacrylamide gel electrophoresis autoradiograms of ³⁵S-methionine pulse-labeled proteins of white and opaque cells. Out of hundreds of distinguishable labeled polypeptides, only one was white specific and two opaque specific. This technique favors abundant proteins. The three proteins differentially expressed were never identified. It was not until 1992, 5 years after the discovery of the white-opaque transition, that the first opaque-specific gene was cloned (12). An opaque cell λgt10 DNA library generated from poly(A)⁺ RNA was hybridized with radiolabeled cDNA of either white or opaque cells and clones differentially hybridized to opaque but not white cDNA identified. One gene was identified which encoded PEPI, previously identified as a member of a secreted aspartyl proteinase gene family (“SAP gene family”) (71, 72). The gene was therefore subsequently renamed SAP1. Two years later, White and Agabian (73) demonstrated that of the three SAPs identified at that time, two (SAP2, SAP3) were preferentially expressed in the white phase and reaffirmed that SAP1 (PEP1) was selectively expressed in the opaque phase. Eight years after SAP1 was identified as opaque specific, Kvaal et al. (22) demonstrated that opaque cells, but not white cells, adhered to and multiplied on baby mouse skin. Misexpression of SAPI in white cells conferred this opaque-specific virulence trait upon white cells in a mouse model of cutaneous colonization. Recently Lohse et al. (74) demonstrated that opaque cells require excreted SAPs to utilize extracellular proteins as a nitrogen source and that, in mixtures of white and opaque cells, opaque-secreted proteinases facilitated white cell proliferation in a protein-based nutrient medium.

In 1993, 1 year after identifying SAP1(PEP1), Morrow et al. (13) used the same differential hybridization screen to identify a second opaque-specific gene, OP4. Southern blot analysis with 14 endonucleases revealed no DNA reorganization in the vicinity of OP4. The amino acid sequence suggested Op4 is a membrane protein, with an amino-terminal hydrophobic leader sequence. OP4 is also differentially up-regulated in the switch phenotypes irregular wrinkle, revertant smooth, and star of the laboratory strain 3153A but not the original smooth phenotype (75).

In 1993, Srikantha and Soll (14) identified the first white-specific gene, WH11. Although its function has not been established, Wh11 exhibits 47% homology of amino acids 1 through 59 with the S. cerevisiae glucose/lipid-regulated GLPI protein, HSP12 (14). Wh11 localizes in the cytoplasm of opaque cells but not hyphae. It is also expressed in budding yeast phase cells, but not in hyphae, of strain 3153A, which does not switch to opaque. It has been suggested that WH11 plays a role in generating the round morphology of budding yeast phase cells (14). Misexpression experiments suggest that it may also play a role in both opaque to white switching and virulence. Misexpressing WH11 did not affect the transition from the white to opaque phase but caused an increase in the frequency of switching from opaque to white and enhanced kidney colonization in a mouse model for systemic infections (76). Park et al. (77) subsequently deleted WH11 and found no effect on switching from white to opaque, verifying the results of Kvaal et al. (76).

By the year 2000, it had been demonstrated that white-opaque switching regulated the expression of a number of phase-specific genes, including SAP1 (12, 73), WH11 (14), SAP3 (73), OP4 (13), CDR3 (78), NIK1 (79), and EFG1 (23). Efg1 was shown by Sonnenborn et al. (23) to be a potent repressor of switching from white to opaque and will be discussed in detail in the section “EFG1 and WOR1” of this review. All seven of these genes were identified in studies that resulted in the identification of one or a few phase-specific genes. However, in 2002, the first transcription-profiling study comparing hundreds of genes between white and opaque cells was reported by Lan et al. (29), employing labeled double-stranded cDNA generated from total RNA of white and opaque cells, and a high-density oligonucleotide Gene Chip, manufactured by Affymetrix. This early and seminal microarray analysis (29) identified 221 open reading frames (ORFs) that underwent a twofold or more increase in expression at the level of transcription in the transition from white to opaque and 152 ORFs that underwent a twofold or more decrease. The identified genes were involved in a wide variety of biological functions. Twenty-five percent of the genes up-regulated in white cells and 20% of the genes up-regulated in opaque cells were not identified at the time of publication. One aspect of differential regulation that Lan et al. (29) found most interesting was the difference in expression of genes involved in aerobic and anaerobic metabolism. The patterns of up-regulation suggested an advantage for white cells to undergo fermentative (anaerobic) metabolism and an advantage for opaque cells to undergo oxidative (aerobic) metabolism. The authors suggested that this difference might be related to the areas of the host preferentially colonized by the alternative phenotypes (29). Differentially expressed genes were also revealed in the Lan et al. study (29) that played roles in mating, stress responses, drug resistance, and intra- and intercellular signaling. The study was the first to show that the white-opaque transition involved the regulation of a large number and broad variety of genes and, hence, that the white-opaque transition represented a complex differentiation, as previously suggested by the differences in cell architecture.

A PREREQUISITE FOR MATING

Over the 15 years following its discovery, white-opaque switching remained an enigma, primarily because no critical function had been established and second because a majority of clinical strains appeared unable to undergo the transition. However, in 2002, two major discoveries were reported by Miller and Johnson in a single publication (16) that dramatically stimulated interest, first that switching was regulated by the zygosity of the mating-type locus and thus was probably repressed by the a1-α2 co-repressor and second that switching was a prerequisite for mating. These discoveries came on the heels of the discovery in 1999 that C. albicans, previously believed to be asexual, harbored a mating-type locus similar to that of S. cerevisiae a/α cells. In 1999, Hull and Johnson (15) identified the mating-type locus (MAT) of C. albicans by chromosome walking (Fig. 5A), using probes that were based on sequences provided by the Stanford C. albicans Sequencing Project. In one copy of the MAT locus, the MATa1 gene, similar to S. cerevisiae MATa1, and an MATa2 gene were present, and in the second copy, MATα1 and MATα2, similar to S. cerevisiae MATα1 and MATa2, respectively, were present (Fig. 5A). Thus, the strain of C. albicans that Hull and Johnson (15) analyzed was a/α, and the MAT locus harbored the genes for the homeobox proteins a1 and α2, which together formed the corepressor a1-α2, which represses mating in related fungi. On the other hand, there were no indications of a silent mating-type cassette, which exits in S. cerevisiae and is involved in mating-type switching. The composition of the MAT locus suggested that C. albicans underwent mating, and within a year of the identification of the mating-type locus, evidence of mating was established. Two studies, one in vivo (80) and one in vitro (81), demonstrated that MTL-hemizygotes and MTL-homozygotes underwent mating. In both the in vivo (80) and in vitro (81) studies, it appeared that mating was a rare event. These results combined with those of studies on the population structure of C. albicans (82–84), indicated that mating did occur but that it was a rare event, that it occurred presumably between a/a and α/α cells, that it resulted in the formation of tetraploids, and that the resulting tetraploids returned to the diploid state either by a meiotic-like mechanism or random chromosome loss, the latter suggested in studies by Bennett and Johnson (85, 86) (Fig. 5B). Meiosis was not excluded, given that Tzung et al. (30), who analyzed the sequences of C. albicans genes homologous to genes involved in S. cerevisiae mating, concluded that while several genes involved in meiosis in S. cerevisiae lacked credible homologs in C. albicans, many did. The preceding results set the stage for the discovery by Miller and Johnson (16) that white-opaque switching played a unique and fundamental role in the mating process of C. albicans.

Fig 5.

The mating system of C. albicans includes homozygosis from a/α to a/a or α/α, a switch from white to opaque, fusion between a/a and α/α opaque cells, and homozygosis. (A) The mating-type locus of MTLa/MTLα cells, which includes mating-type genes MTLa1, MTLa2, MTLα1, and MTLα2, and nonmating-type genes PAP, OBP, and PIK. (B) Homozygosis, switching, and mating. (C) SEM example of conjugation tube extension. (D) SEM example of conjugation tube contact. (E) SEM example of tube fusion (conjugation bridge). (F) SEM example of first daughter cell formation. Arrows in D, E, and F, fusion point; PC, parental cell; DC, daughter cell. SEMs reprinted from Lachke et al. (35). Scale bars, 1 µm.

In 2002, Miller and Johnson (16) reported two seminal observations that transformed the white-opaque transition from a phenomenon of interest to a process central to the life history of C. albicans and related species. First, they demonstrated that yeast-phase cells of an a/α strain did not switch to opaque. In contrast, clonal plating of white cells of a1/- or -/α2 hemizygous derivatives formed opaque-like sectors at frequencies of 10.4% and 4.5%, respectively. Miller and Johnson (16) demonstrated that high temperatures caused mass conversion of the bean-shaped, opaque-like cells to the white phenotype, and opaque-like cells differentially expressed the opaque-specific gene OP4 but not the white-specific gene WHII. If, as these results indicated, the a1-α2 corepressor of mating also repressed switching from white to opaque, then the original white-opaque switching strain WO-1 might have lost either MTLa1 or MTLα2, or both. PCR analyses demonstrated that indeed WO-1 did not contain MTLa1 but did contain MTLα2 and, hence, could not form the a1-α2 corepressor (16). It was verified that the clinical isolate WO-1 was indeed α/α (32). Miller and Johnson (16) then presented evidence that opaque cells, but not white cells, of the a1/- and -/α2 hemizygous strains were mating competent, thus demonstrating that white-opaque switching played a major role in the basic life history of C. albicans. Miller and Johnson (16, 32) also observed that the complemented offsprings were primarily in the white phase, which was consistent with the re-establishment of a1-α2 repression of switching. Their results suggested that in order to mate, a/α strains, which make up over 95% of clinical isolates (32, 87, 88), first have to undergo homozygosis at the MTL locus, then switch from white to opaque (Fig. 6B). Soon after this discovery, Lockhart et al. (32) demonstrated that clinical strains that could switch from white to opaque were homozygous at the mating-type locus. It should be noted that in 2004, Pendrak et al. (40) demonstrated that deletion of the hemoglobin response gene 1 (HBRI) resulted in switching to opaque in a/α cells by suppressing α2 expression. In 2017, Scaduto et al. (89) demonstrated that overexpressing select MAP kinases resulted in mating of white cells locked in the white phase due to deletion of a positive regulator of switching. Surprisingly, those mating “white-locked” cells formed conjugation tubes and mated without assuming opaque cell architecture.

Fig 6.

Opaque cells, but not white cells, colonize and mate at high frequency on the skin of baby mice. Cells were incubated on skin under cotton and imaged by SEM. (A) White cells do not colonize infant mouse skin. (B) Opaque cells colonize infant mouse skin. (C, D) Example of high-frequency mating (arrows) of opaque cells in craters induced by opaque cells on infant mouse skin. Scale bars represent 20 µm for panel A, 33 µm for B, and 2 µm for C and D. Arrows in panels C and D denote examples of mating events. Panels A and B are reprinted from reference (22), and panels C and D are reprinted from reference (35).

THE CELL BIOLOGY OF MATING BY OPAQUE CELLS

Miller and Johnson (16) using differential interference contrast (DIC) microscopy first demonstrated that in mixed opaque MTLa and MTLα cultures, elongated cells with thick tubular extensions, similar in morphology to conjugation tubes in other fungi, are formed with no constriction or septum at the opaque-tube junctions. In mating mixtures of opaque MTLa and white MTLα cells, or white MTLa and opaque MTLα cells, these morphologies were not observed. Early evagination morphologies resembled the “shmoo” formed by haploid S. cerevisiae cells in response to pheromones of opposite mating type (90–93). The year after the Johnson and Miller report (16), Lockhart et al. (58), using a variety of light microscopic, computer reconstruction and fluorescent staining techniques, described in detail at the cellular level several steps of C. albicans mating, including shmooing, conjugation tube formation, conjugation tube fusion, and daughter cell formation, in mixtures of natural MTLa/MTLa and MTLα/MTLα strains (Fig. 5C through F), with the exception of karyogamy. First using DIC microscopy, mixtures of opaque cells of an MTLa/MTLa strain and MTLα/MTLα cells were first shown to shmoo and then form long conjugation tubes (Fig. 6C) at high frequency. Tubes did not form if either of the cells of mixed a/a and α/α strains were white, or if one strain was a/α. The conjugation tubes differed from hyphae in that there was no constriction or septation along the conjugation tubes (Fig. 5B and C). Although at the early stages of tube formation, the majority of evaginated cells were shmoo like, the projections extended apically with time to form tubes. The tubes could grow to lengths far greater than the diameter of their mother cells (Fig. 5C). Continuous video recordings and SEM of mating in perfusion chambers revealed contact (Fig. 5D) and fusion (Fig. 5E) only at the apices of opposing tubes. Computer-assisted 3D reconstructions of DIC optical sections over time using DIASemb software (94) were generated from the fusion of tubes (58). Three-dimensional reconstructions over time suggested that opposing conjugation tubes released gradients of opposing mating pheromones that directed growth at the apices of the conjugation tubes for accurate apical contact preceding fusion (Fig. 5D). To prove that only opaque MTLa/a and MTLα/α cells fused, Lockhart et al. (58) labeled a/a cell surfaces with the vital dye FITC-conA (green) and α/α cell surfaces with the vital dye rhodamine-ConA (red). Fusants were composed exclusively of FITc-conA (a/a) and rhodamines-conA (α/α) stained zygotes. In cells that formed very long conjugation tubes, the nucleus migrated from the mother cells far into the tube. After conjugation tube fusion, the a and α nuclei, which had entered the conjugation bridge, met at the position of subsequent daughter cell formation. Placement of the nuclei was usually acentric along the bridge. In the Lockhart et al. (58) study, the a and α nuclei failed to undergo karyogamy (nuclear fusion). However, in a study by Pujol et al. in 2004 (95), it was demonstrated at the cellular level that mating occurred between opposite mating types of C. albicans and the highly related species Candida dubliniensis. The interspecies mating events were identical to C. albicans intraspecies crosses, but in the interspecies crosses, karyogamy occurred. In 2005, Bennett et al. (96) demonstrated at the cellular level that the nuclei of their a and α C. albicans strains underwent karyogamy when mating, indicating that C. albicans mating led to a tetraploid cell with a single nucleus (85, 86). The failure to observe karyogamy in the study by Lockhart et al. (58) was never explained but most likely was the result of strain incompatibility. In the study by Lockhart et al. (58), when the daughter cell had expanded to roughly half its final volume, one nucleus resulting from nuclear division in the conjugation bridge entered the daughter cell. The studies by Pujol et al. (95) and Bennett et al. (96) also demonstrated that after karyogamy and nuclear division, one nucleus entered the daughter cell. A developing septum then formed at the junction of the conjugation bridge and daughter cell. In 2003, Daniels et al. (97) presented evidence that the eccentric position of daughter cell formation (Fig. 6F) was restricted to that portion of the conjugation bridge contributed by the opaque a/a parent cell. The conclusion was based upon the results of all possible crosses of four natural a/a and four natural α/α strains. The a/a portion of the conjugation bridge was also selectively stained by anti-Hwp1 rabbit polyclonal antiserum. Hwp1 is an adhesin selectively expressed on the surface of hyphae, which plays a role in adhesion to human epithelial cells (98) and in the binding of biofilms to venous catheters (99). HWP1 transcription is stimulated by pheromone in opaque a/a cells but not in opaque α/α cells (97). Together, these results indicate that a/a and α/α cells do not play equivalent roles in the mating process. The alternative mating types express different pheromone receptors, and as Daniels et al. (97) demonstrated, this results in a difference in HWP1 expression and adhesion properties, and localization of daughter cell formation and septation along the conjugation bridge. Finally, one must wonder if, in addition to the differences in the MTL genes and their roles in gene expression between the two mating types, differences may be at play between the a and α alleles of the three nonsex genes harbored in the alternative mating loci (Fig. 5A), which have been shown to have diverged more rapidly than average genes outside of the mating locus on chromosome 5 (100) in sequence comparisons. The three nonsex genes in the MTL locus, PIK (the essential phosphatidyl inositol kinase gene), PAP (the essential poly (A) + polymerase gene), and OBP (the nonessential oxylate-binding protein gene), were subsequently shown to play roles in biofilm formation by a/α cells (100).

CONDITIONS AFFECTING SWITCHING FREQUENCY: EARLY OBSERVATIONS

It was clear in the first comparisons between clinical isolates of C. albicans that the frequency of white-opaque switching varied between strains and, in some cases, decreased during sequential plating of single strains (11). In 1987, before the discovery of white-opaque switching was published, strain WO-1 was distributed by request and, in turn, redistributed secondarily by first recipients. At least two recipients complained that the clones of WO-1 that they received had stopped switching or did so at reduced frequency. The conditions used and the number of colonies assessed differed markedly between recipients, as well as the conditions used between recipients and those used in the original studies by Slutsky et al. (10) and Soll et al. (9) in 1987. As previously discussed in the original report of white-opaque switching (10), the switching affected by low temperature (4°C) and high temperature (37°C) was demonstrated.

By the late 1980s, the mechanisms of several phase transitions or switching systems in other microbes had been clearly elucidated at the molecular level and, in most cases, were found to involve DNA recombination (101–107), a process stimulated by UV (108). Switching in C. albicans strain 3153A (8) and strain 1001 (7) had both been demonstrated to be stimulated by low doses of UV in 1985. Morrow et al. (62), therefore, tested whether low doses of UV affected the frequency of white-opaque switching in strain WO-1. They found that a low dose of UV irradiation caused an increase in the frequency of opaque colony formation in two independent experiments. Morrow et al. (62) ruled out enrichment due to a shorter opaque cell generation time as a possible explanation for the observed results. They also found that UV-stimulated sectoring was a heritable effect through several generations. Morrow et al. (62) suggested two possible models for the regulation of white-opaque switching, the first one based on the reversible phase transition in Salmonella (105) and the second based on the reversible mating type switch in S. cerevisiae (107). Both of these transitions involved reversible recombination events, and the latter was induced by UV (108). No observations have supported either model to date.

CONDITIONS AFFECTING FREQUENCY: LATER OBSERVATIONS

In addition to the effects of temperature and UV, Soll et al. (69) noted in a 1991 review of switching that if a 5-day-old Petri dish culture containing white WO-1 colonies was wrapped air-tight in parafilm and incubated an additional 8 days, multiple opaque sectors formed at the edge of each colony. This indicated that a decrease in oxygen or a build-up of CO₂ induced white to opaque switching. Therefore, through the 1990s, evidence had slowly accumulated that the frequency of switching in both directions could be affected by changes in disparate environmental conditions. Although the discovery in 2002 by Miller and Johnson (16) that the white to opaque switch was a prerequisite to mating had renewed interest in the white-opaque transition, the effects of environmental conditions took a backseat to studies on the genetic mechanisms regulating the switch event and the role of the opaque phenotype in the mating process. Most in vitro studies continued to be performed under potentially nonphysiological conditions, including glucose as the carbon source, temperatures of 22°C–25°C, in air which is low in CO₂ (0.04%), and on a supporting agar substrate or in suspension under rotation or agitation to keep the cultures mixed and aerated. However, over time, studies were reported that underscored the profound effects environmental conditions could have on switching. In 2007, Dimitrou et al. (48) demonstrated that the opaque phenotype generated under aerobic conditions at 30°C was stabilized by transferring cultures to anaerobic conditions at 37°C. To achieve anerobic conditions, tubes containing growth medium were prebubbled with pure nitrogen for 30 minutes, which removed both O₂ and CO₂. This set of conditions supported opaque cell mating. In 2008, Ramirez-Zavala et al. (45) further explored the effects of anaerobic conditions and temperature on switching frequencies. Under anaerobic conditions, they obtained mass conversion of white to opaque after 48 hours at 24°C and after 24 hours at 37°C. These results appeared to demonstrate two points. First, anaerobic conditions induced switching to opaque. Second, switching could be induced at physiological temperature (37°C) under anaerobic conditions. The authors then inoculated mice intragastrically with white cells of strain WO-1, plated the feces of the mice each subsequent day on supplemented Lee’s agar, and incubated the plates for 7 days at 25°C in air. Switching to opaque by the passaged cells occurred at frequencies up to two orders of magnitude higher than unpassaged white control cultures. The authors concluded from their in vitro and in vivo results that the low O₂ levels in the GI tract induced switching to opaque. Ramirez-Zavala et al. (45) employed an in vitro technique to obtain anaerobic conditions that not only reduced O₂ levels but also increased CO₂ levels to approximately 18%, 450 times the concentration in air. Therefore, neither the in vitro nor the in vivo experiments distinguished between decreased O₂ and increased CO₂ as the inducing condition. In the GI tract, while O₂ levels are low, CO₂ levels range from 5% to over 20% (109, 110). Huang et al. (46) therefore tested the effects of high CO₂ by incubating clonally plated agar cultures of white cells in air containing different concentrations of CO₂. In air containing 5% CO₂, the frequency of opaque and sectored colonies increased from 4- to 16-fold over that in air alone. In air containing 20% CO₂, the frequency increased to more than 90%. Huang et al. (46) concluded that it was high CO₂, not hypoxia (low O₂), that induced the switch to opaque. In further support of this conclusion, the authors showed that high CO₂ induced switching similarly in the absence and presence of O₂ (46). Huang et al. (46) suggested that HCO₃ mediated the CO₂ effect, but recent studies by Zhang et al. (111) raise the possibility that PP2C2 phosphatases may function as direct CO₂-sensing systems.

The effects of specific sugars on switching also became a focus of attention, most notably N-acetylglucosamine (GlcNAc), a monosaccharide that is the subunit of chitin, a polysaccharide component of C. albicans and bacterial cell walls, of peptidoglycans in the wall of the gastrointestinal (GI) tract, and of hyaluronan, an extracellular matrix molecule of host connective tissue. Most likely GlcNAc released into the GI tract is primarily from bacteria of the microbiome (112). Over 50 years ago, GlcNAc had been demonstrated by Simonetti et al. (113) to induce hypha formation in C. albicans. Huang et al. (47) therefore tested whether GlcNAc also induced white to opaque switching by substituting GlcNAc for glucose in supplemented Lee’s agar medium (“supplemented Lee’s GlcNAc agar”). After 5 days of incubation on supplemented Lee’s GlcNAc agar, over 80% of plated white cells formed colonies that were opaque or white with opaque sectors. If white cells were clonally plated on supplemented Lee’s GlcNAc agar and incubated at 25°C in high CO₂ (1%), switching was further enhanced, suggesting additivity or synergy of high CO₂ and GlcNAc. GlcNAc appears to be sensed by an N-acetyltransferase that acetylates promoter histones (114).

Another major environmental factor which surprisingly was not examined early on for its effect on white-opaque switching was pH. In the case of the bud-hypha transition of C. albicans, Buffo et al. (115) had already established in 1984 that at low pH (pH 4.5) at 37°C, yeast cells continued to form budding yeast-phase cells, but at high pH (e.g., pH 6.5) at 37°C, yeast-phase cells formed hyphae en masse. It was not until 2015, almost 30 years after the discovery of white-opaque switching, Sun et al. (52) performed a study on the effects of pH on white-opaque switching. They found that low pH favored switching from white to opaque and high pH favored switching from opaque to white. The pH of the GI tract varies from very low in the stomach (pH 1.5–2.0), to slightly above neutral in the small intestine (mpH 7.3), to basic in the colon (7.9–8.5). In summary, high CO₂, GlcNAc as the major carbon source, and low pH are all conducive to the white to opaque switch at 37°C.

HYPHA FORMATION BY OPAQUE CELLS

In 1989, 2 years after the discovery of switching, Anderson et al. (59) reported that although opaque cells formed hyphae (Fig. 2D), the environmental conditions that were inducive differed from those for hypha formation in yeast phase a/α cells or white MTL-homozygous cells. The conditions of high pH (pH 6.5) at high temperature (37°C) in air, which were inducive for hypha formation by white cells (115), did not support the formation of hyphae by opaque cells (59). However, when opaque cells were anchored to a polylysine-coated glass wall of a Sykes-Moore perfusion chamber, a proportion was induced to form hyphae at frequencies quite variable between clones of strain WO-1 (59). The nature of the adhesive substrate that facilitated hypha formation did not seem to matter. Uncoated plastic, polylysine-coated plastic, and fibronectin-coated glass were similarly inducive. The most potent induction of hyphae formation, however, was obtained when opaque cells were incubated on glass cover slips coated with a monolayer of human epithelial cells (59). Together, these results indicated that environmental induction of hypha formation differed between white and opaque budding cells, and anchoring to epithelium or a variety of other unrelated substrates favored hypha formation by opaque cells. This latter characteristic may be related to the selective colonization of skin by opaque cells versus white cells (22, 35). In addition, mechanical sensing and thigmotropism have been demonstrated quite elegantly for hyphal growth by C. albicans (116, 117) and may be related to the nonspecific substrate response of opaque cells in hypha formation (59). Twenty-four years after the study by Anderson et al. (59), Si et al. (118) identified three additional environmental cues inducing opaque cell filamentation, sorbital as the carbon source, growth on minimal medium lacking amino acids, and GlcNAc as the carbon source, although in the last case, colony incubation was extended and filaments formed in the pseudohyphal phenotype. Si et al. (118) also demonstrated that opaque cells were capable of hypha formation without anchoring to a substrate (i.e., in suspension) in synthetic complete dextrose (SCD) medium with low phosphate and in SCD medium supplemented with 1 M sorbitol. What seems interesting from a cell biological point of view is the similarity between hypha-forming colonies on the different solid media in the Si et al. (118) study and the frizzy variant formed by strain WO-1 that was heritable and reversible, but which appeared spontaneously at very low frequency and therefore was not investigated after its initial description by Slutsky et al. (10). The fuzzy variant formed aerial hyphae profusely on supplemented Lee’s agar at 25°C in air (10), suggesting it was the result of a second switching system in strain WO-1 that altered environmental constraints on the bud-hypha transition.

ADHESION AND COLONIZATION

Immediately after the discovery of white-opaque switching, it was reported that clinical strains capable of undergoing the white-opaque transition could be isolated from different host niches. The results suggested that the alternative phenotypes, white and opaque, might selectively colonize different host niches and therefore may express different adhesive characteristics. The first study testing differences in adhesion was performed in 1987 by Kennedy et al. (67). White cells of three separate natural strains that underwent white-opaque switching adhered more frequently to human buccal epithelial cells grown in culture and were less hydrophobic than opaque cells. Ten years later, Kvaal et al. (22) demonstrated that opaque cells were far more adherent than white cells to skin, using a cutaneous model in which white or opaque cells were incubated under a nonwoven cotton patch on the hairless skin of newborn mice. While few white cells colonized the skin (Fig. 6A), opaque cells adhered and multiplied (Fig. 6B), sometimes forming dense multilayer opaque cell sheets in a matrix over the skin, suggestive of an opaque cell biofilm. Individual opaque cells in contact with skin caused indentations (cavitation), possibly the result of secreted proteinases (Fig. 6C and D) (22). One year after Miller and Johnson (16) demonstrated that opaque was the mating-competent phenotype, Lachke et al. (35) tested whether opaque cells would mate on skin, using the same cutaneous mouse model as Kvaal et al. (22). Mixed cultures of opaque a/a and α/α cells incubated under cotton gauze patches on newborn mouse skin underwent mating at frequencies up to five times that in suspension cultures (Fig. 6C and D). In a comparison of opaque cell mating on different substrates, Lachke et al. (35) quantified the proportion of adherent cells mating among applied mixtures (50:50) of opaque a/a and α/α cells. They observed the following fusion (mating) levels: on skin, 39%; on glass, 0%; on plastic, 0%; and on silicon elastomer, the material catheters are composed of 2%. In marked contrast, in experiments to assess virulence and mating in an oropharyngeal model of candidiasis, Solis et al. (119) found that white cells were far more successful than opaque cells in adhering to and colonizing the oropharyngeal mucosa. This observation explained in part the decrease in invasion of the oral tissue by opaque cells in comparison to white cells and the lack of mating between a/a and α/α cells (119). Perhaps the most intriguing aspect of differential adherence may be related to the formation of a sexual biofilm by white cells, in which minority opaque cells through release of mating pheromone signal majority white cells to form an adherent, complex biofilm that facilitates opaque cell mating. In the original study by Daniels et al. (19), it was demonstrated that α-pheromone stimulated a/a white cells and a-pheromone stimulated α/α white cells to adhere to plastic or silicon elastomer, the first step in sexual biofilm formation in vitro. The formation of a sexual biofilm and the role of pheromones are discussed in greater detail in the section “Mating, Intercellular Signaling, and Formation of a Sexual Biofilm” of this review.

WHITE AND OPAQUE CELLS IN SYSTEMIC AND GI TRACT MODELS

The first comparison of colonization between white versus opaque cells in a mouse model for systemic infection was performed in 1997 by Kvaal et al. (76). The authors found that when white cells were injected into the mouse tail vein, they colonized the kidneys and maintained their original white cell phenotype. In marked contrast, when opaque cells were injected, kidney colonization was reduced 10-fold, and the colonizing population consisted of over 90% white cells rather than opaque cells after 12 days. These results demonstrated that white cells were far more proficient than opaque cells in kidney colonization in the tail injection model and suggested that when opaque cells were injected, either the few white cells in the opaque cell population were preferentially selected for or the few original opaque cells that did adhere switched to the white phenotype during colonization. Twenty years later, in 2019, Takagi et al. (120) revisited the question of comparative colonization in the mouse model of systemic infection, verifying the results of Kvaal et al. (76) and expanding the study to other organs. Equal proportions of white and opaque cells, expressing m-Cherry and GFP, respectively, were coinjected into mice. Kidneys, liver, heart, spleen, and brain were harvested after 24 hours, and the level and phenotype of colonization in each organ were assessed by plating and counting colony phenotypes and colony color. Three categories were discriminated in the colonization assay: (i) white colonies emanating from injected white cells, (ii) opaque colonies emanating from injected opaque cells, and (iii) white colonies originally opaque that had switched to white. First, they found that colonization was most intense in the kidneys of four of the eight injected mice in the study. Unlike other organs, kidney colonization in mice was primarily by white cells, with significant minor portions consisting of opaque cells that had switched to white. Liver, spleen, and brain had similar proportions of white and opaque cell colonization. Brain contained white cells resulting from opaque to white switching. In marked contrast, heart was colonized predominantly by opaque cells that were originally opaque. Thus, while white cells preferentially colonized the kidney, the most heavily colonized organ, opaque cells preferentially colonized the heart, the least intensively colonized of the five organs. Because injected yeast cells are cleared from the blood within an hour, one might assume that colonizing opaque cells that switch to white most probably do so after initial organ colonization. Wu et al. (121) had previously found that in mixed infections performed in the systemic mouse model to assess competitiveness, natural a/α strains were more competitive than their a/- or -/α derivatives in colonizing the kidneys. In the experiments by Wu et al. (121), the entire MTL-locus was deleted in the MTL-hemizygous strains.

EVADING THE IMMUNE SYSTEM

In several of the most prominent switching systems studied in pathogenic microorganisms, one or more changes in antigenicity allow the infecting organism to evade the host immune system (101–103, 105, 106). Although Anderson et al. (59) demonstrated that antigenic changes at the cell surface accompany the white-opaque transition, it has never been demonstrated that these changes function in immune evasion. However, white and opaque cells do interact differently with white blood cells. In 1990, Kolotila and Diamond (122) found that opaque cells were more sensitive than white cells to killing by neutrophils and oxidants and that both neutrophils and oxidants stimulated white to opaque switching. In 2004, Geiger et al. (36) extended these studies of the interactions between leukocytes and the two switch phenotypes. Previously, Cutler (123) using an agar diffusion assay demonstrated that clinical strains of C. albicans, presumably a/α, released a polymorphonuclear leukocyte (PMN) chemoattractant that bound to the chemotactic receptor for fMLP, a potent PMN attractant. Since the assay used by Cutler (123) did not discriminate between chemotaxis and chemokinesis, Geiger et al. (36) employed a single-cell analysis chamber fashioned after the chamber developed by Zigmond (124), which distinguishes chemotaxis from chemokinesis by following single-cell behavior microscopically in a defined gradient of chemoattractant. First, Geiger et al. (36) found that C. albicans strain 3153A (a/α), C. dubliniensis, C. tropicalis, C. parapsilosis, and C. glabrata, but not S. cerevisiae, released one or more PMN chemoattractants as potent as fMLP. Second, in a comparison between white and opaque cells of C. albicans strains WO-1 (α/α), P37005 (a/a), 19 f (α/α), L26 (a/a), and 12C (a/a), white cells released a potent chemoattractant, but opaque cells did not. Hence, opaque cells may be invisible to leukocytes. Protein kinase A treatment voided chemoattractant activity. The molecular mass of the chemoattractant was estimated to be 10³ Da, which was similar to the molecular weight estimated by Edens et al. (125), who also presented evidence that the fMLP receptor mediated leukocyte chemotaxis in gradients of the C. albicans chemoattractant. Geiger et al. (36) proposed that opaque invisibility may represent a strategy to increase the success of mating in vivo. Lohse and Johnson (37) subsequently demonstrated that both cells from a cell line, S2, derived from Drosophila melanogaster hemocytes, and from a cell line, RAW 264.7, derived from mouse macrophages, preferentially phagocytosed C. albicans white cells at frequencies five times higher than that for opaque cells. One might conclude from the chemotaxis experiments of Geiger et al. (36) and the phagocytosis experiments of Lohse and Johnson (37) that the absence of chemoattractant explains the low levels of phagocytosis since attractants stimulate pseudopod-like projections that function as phagocytic cups that engulf bacterial and fungal pathogens. Lohse and Johnson (37) noted that although the adherence of white cells to RAW 264.7 cells was two- to threefold higher than adherence of opaque cells, the difference was too low to account alone for the much greater difference in phagocytosis. Sasse et al. (38) subsequently demonstrated that human PMNs, when mixed with white and opaque cells, preferentially phagocytosed white cells. In these experiments, WO-1 (α/α) white and opaque cells harboring GFP and RFP under control of the WHII promoter and OP4 promoter, respectively, were mixed together with human PMNs and analyzed by time lapse video. The frequency of phagocytosis of white versus opaque cells was 99–1. Sasse et al. (38) observed that the great majority of white cells had formed germ tubes (incipient hyphae) under the conditions employed on glass surfaces in solution. However, in a 3D type I collagen gel, phagocytosis of white and opaque cells was equivalent. The presented video frame suggested that the opaque cells in the collagen matrix were undergoing germ tube formation. Dendritic cells also appeared to phagocytose white and opaque cells similarly, but again the presented video frames suggested opaque cells were forming germ tubes or wider evaginations with shmoo-like morphologies. These interesting observations warrant further investigation.

a/α SWITCHING, EFG1, AND THE “GRAY” PHENOTYPE

For approximately 10 years after Miller and Johnson’s discoveries (16), it was the general perception that natural a/α strains could not switch. However, in 2013, Xie et al. (39) presented evidence that a significant proportion of a/α strains could be induced to switch to opaque. Using GlcNAc as the sugar source and high CO₂ (5%) at 25°C (opaque-inducing conditions), they found that approximately one-third of a collection of 96 clinical isolates from China, mainly a/α, could be induced to switch from white to opaque, albeit at low frequencies. They also found that 50% of a collection of 29 clinical a/α strains from different countries could also be induced to switch to opaque at low frequencies. The a/α opaque cells exhibited the unique characteristics of MTL-homozygous opaque cells but could not mate, presumably because they still expressed the a1-α2 corepressor. Xie et al. (39) further demonstrated that deletion of the genes encoding any one of the three transcription factors, Efg1, Rfg1, or Brg1, derepressed switching in a/α cells. In 2020, Park et al. (20) reported that a similar proportion of a collection of clinical isolates could be induced to switch from white to opaque. Park et al. (49) added to the list of repressors the transcription factor Sfl2. But the story became even more complicated in 2014 due to the discovery by Tao et al. (50) that select a/α strains that switched to opaque could also switch reversibly to a third phenotype, suggesting a possible triphasic switching system. Tao et al. (50) first observed that the a/α strain BJ1097, a vaginal isolate, switched between white and opaque as well as a third phenotype they designated “gray.” Gray cells were smaller than white or opaque cells but elongate like opaque cells. However, unlike opaque cells, they lacked cell surface pimples. The three cell types, white, opaque, and gray, differed in gene expression patterns, virulence in the mouse systemic infection model, adhesion to skin in the mouse cutaneous infection model, and mating efficiency. Tao et al. (50) also assessed whether the two major transcription factors that regulate white-opaque switching, Efg1 and Wor1, discussed in detail in the section “EFG1 and WOR1” of this review, played a role in the switch to gray. They concluded that they did not. In 2019, two studies verified and extended the basic observations by Tao et al. (50). First, Park et al. (49) verified that deletion of EFG1 resulted in triphasic switching in the a/α strain SC5314. Second, Liang et al. (51) found that clinical isolates that contained EFG1 mutations formed gray cells. Liang et al. (51) showed that natural a/α EFG1 heterozygous strains (EFG1/efg1⁻) transitioned to efg1⁻/efg1⁻ mutants at high frequency, leading to higher colonization in a mouse gastrointestinal model and argued that this transition was involved in adaptation to host colonization. To compare colonization in the mouse GI tract, Liang et al. (51) performed competition experiments between white EFG1/EFG1 cells or white EFG1/efg1⁻ cells, and gray efg1⁻/efg1⁻ cells, and between white EFG1/efg1⁻ and gray EFG1/efg1⁻ cells. All of the mutants were generated in the a/α strain SC5314. In all cases, plating of final samples revealed gray cell dominance after 4 days. Dominance of the gray phenotype was also suggested when pure white cell populations of three clinical a/α strains (P75063, 1619, BJ1097) that formed gray cells were tested in the mouse GI tract model. In fecal samples, a majority of the colonizing cells formed gray colonies. These studies (49, 51) verified triphasic switching and linked it to EFG1 mutations. In the following year, Park et al. (20) analyzed 27 randomly selected clinical a/α isolates for switching and sequenced their EFG1 open reading frames. The collection of clinical strains included 17 that were homozygous for wild-type EFG1 (EFG1/EFG1), 2 heterozygous mutants (EFG1/efg1⁻), and 8 homozygous mutants (efg1⁻/ efg1⁻). Four of the 17 natural EFG1/EFG1, one of the two natural EFG1/efg1⁻ strains, both constructed efg1Δ/efg1Δ mutant strains, and all of the eight natural efg1⁻/efg1⁻ strains could be induced to switch to opaque. Thirteen of the 17 EFG1/EFG1 strains and 1 of the EFG1/efg1⁻ strains could not be induced to switch to opaque. All of the homozygous EFG1 mutants (two efg1Δ/efg1Δ, eight efg1⁻/efg1⁻) could be induced to switch to opaque, and all but one of them formed gray cells. In contrast, only one of the four heterozygous or hemizygous strains could be induced to form opaque cells, and none of the natural efg1⁻/efg1⁻ strains could be induced to form gray cells. Park et al. (20) found that one copy of the wild-type EFG1 gene was sufficient to repress gray cell induction, but in five cases, four EFG1/EFG1 and one EFG1/efg1⁻ natural strains, cells could still be induced to form opaque cells, suggesting that the gray phenotype was not an essential intermediate in the switch from white to opaque. The four clinical EFG1/EFG1 and one clinical EFG1/efg1⁻ strains that could not form gray cells but could form opaque cells could not be complemented with a constitutively expressed copy of EFG1. Complementation in this case would have been the reinstatement of repression of white to opaque switching. These results (20) were consistent with the conclusion by Tao et al. (50) and Liang et al. (51) that the gray phenotype is distinct from opaque. Most likely the four EFG1/EFG1 and one EFG1/efg1⁻ strains in the Park et al. (20) study that could be induced to switch to opaque, but not gray, harbored mutations in genes other than EFG1 that also repress the white to opaque switch but not the white to gray switch. As was the case in the studies by Liang et al. (51), Park et al. (20) found that regardless of the injected phenotype of efg1⁻/efg1⁻ cells (white, gray, or opaque) in the GI tract model, the fecal samples eventually contained a majority of cells that formed gray colonies, assessed by plating fecal samples on GlcNAc agar at 25°C in 5% CO₂. However, whereas Liang et al. (51) observed that fecal samples contained cells that formed almost exclusively gray colonies, Park et al. (20) observed in competition experiments that between a third and a half of C. albicans cells in fecal samples formed opaque colonies after 15 days. Since Park et al. (20) had found that the gray phenotype was either not expressed or highly unstable at 37°C in vitro, Park et al. (20) questioned the validity of deducing in vivo cell phenotypes from colony phenotypes in fecal plating experiments. They were concerned that after plating, the colonizing cells could have switched phenotype. Therefore, Park et al. (20) examined microscopically cell phenotypes of C. albicans in fresh fecal samples of a mouse GI colonization model before cells could multiply, using an original injection combination in the model of 50% SC5314 EFG1/EFG1 white cells and 50% SC5314 efg1⁻/efg1⁻ white cells, the latter harboring the mCherry gene under the regulation of the opaque-specific OP4 promoter. Using phase contrast microscopy to assess cell morphology and fluorescence microscopy to identify opaque-specific mCherry expression, Park et al. (20) found that 90% of 100 C. albicans cells from the original efg1⁻/efg1⁻ cells in fecal samples exhibited the opaque cell morphology and fluoresced, indicating that efg1⁻/efg1⁻ opaque cells, not gray cells, dominated the cell populations in colonized feces. In other words, cells transitioned from opaque to gray under the in vitro colony assay conditions on agar, demonstrating that the plating assay was misleading. Together, these results are important and must be verified for several reasons. First, they indicate that a significant proportion of clinical isolates may harbor mutations in genes, including EFG1, that encode repressors of switching, and these variants may express the opaque phenotype at sites of colonization in a significant number of hosts. Second, the distinction between strains that can or cannot switch to opaque is rare, if ever made when studying drug resistance, virulence, pathogenesis, or the basic biochemistry and molecular genetics of Candida albicans and related species, resulting in generalizations that may not be representative of all strains at sites of colonization. Third, the proportion of a/α strains harboring loss-of-function EFG1 alleles in the Park et al. study (20) was far higher than the proportion reported by Liang et al. (51). It therefore is possible that the high frequency reported by Park et al. (20) may be due to culturing techniques that favor EFG1 mutations. This discrepancy requires resolution. Fourth, if the gray phenotype is not expressed or is not stable at 37°C, what is its function in commensalism and pathogenesis? And fifth, since efg1⁻/efg1⁻ mutants outcompete EFG1/EFG1 wild-type strains in the GI tract, why are not they the super majority of strains in nature (i.e., in human hosts)?

EFG1 AND WOR1

As is evident in the treatment of switching literature after 2006, it became difficult to discuss the regulation of switching at the level of gene regulation without alluding to the roles of two major regulators, the repressor Efg1 and the inducer Wor1. Therefore, their discoveries warrant a detailed chronological description. In the first comprehensive review of switching in C. albicans (66) published in 1992, it was hypothesized that a master switch gene, positioned adjacent to a heterochromatic region, switched spontaneously between a repressed (silent) and derepressed (expressed) state (126), which in turn dictated spontaneous switching. This simple hypothesis provided a conceptual framework for regulation but proved false. As noted, in 1999, Sonnenborn et al. (23) reported observations on the role of the transcription factor EFG1 (127, 128) in white-opaque switching. Previously, EFG1 had been demonstrated to be essential for the induction by serum of hypha formation. Because Sonneborn et al. (23) were unable at that time to delete both copies of EFG1 in strain WO-1, they employed an overexpression derivative of WO-1, efg1⁻/PCK1p-EFG1, in which glucose could be used to regulate the level of EFG1 expression in the opaque phase. They found that high glucose-treated opaque cells of the PCKp-EFG1 derivative switched back to the white phenotype. They showed by northern blot hybridization that decreased EFG1 levels correlated with an increased frequency of transition from white to opaque. Finally, they found that low levels of EFG1 caused cells to form a cell type very much like the gray phenotype described 14 years later by Tao et al. (50). One year after the report by Sonnenborn et al. (23), Srikantha et al. (24) successfully deleted both alleles of EFG1 in the α/α strain WO-I. They showed that EFG1 was transcribed as an abundant 3.2 kb mRNA in white cells and a far less abundant 2.2 kb mRNA in opaque cells, the result of different transcription start sites. They also demonstrated that the efg1⁻/efg1⁻ mutant formed opaque colonies at 25°C on supplemented Lee’s medium in air, and when these opaque cells were plated on the same nutrient agar in air at 42°C, which induces opaque cells to switch to white, the mutant cells reverted to a phenotype elongate-like opaque cells but smaller and lacking pimples, similar to the subsequently identified gray phenotype (50). When efg1⁻/efg1⁻ opaque cells were transferred to 42°C, expression of the opaque-specific genes SAP1 and OP4 immediately turned off. These results were similar to results reported by Sonneborn et al. (23) for cells expressing low amounts of EFG1. Together, these results suggested that Efg1, a transcription factor, was involved in repressing one or more genes involved in the switch to opaque, as well as activating one or more genes involved in producing the round shape of normal white cells. In 2020, Park et al. (49) compared switching by a/α efg1⁻/efg1⁻ cells under eight sets of environmental conditions, which included all permutations of three environmental parameters [glucose versus GlcNAc as the carbon source, 25°C versus 37°C, and air (0.04% CO₂) versus 5% CO₂]. Their results indicated that under the combined physiological conditions of GlcNAc as a sugar source, high-CO₂ (5%) and high-temperature (37°C) a/α white cells lacking Efg1 switched to opaque cells en masse. This may explain why Park et al. (20) found that a/α efg1⁻/efg1⁻ cells colonizing mouse feces expressed almost exclusively the opaque phenotype.

Although Efg1 had been established as a repressor of the white to opaque transition, it did not seem to satisfy the expected role of a master switch gene. The expectations were based on the assumptions that white is the default phenotype in the transition, and differentiation to opaque required up-regulation of a master switch gene that, in turn, activated and repressed a variety of genes involved in the complex combination of physiological, structural, and regulatory changes that had been shown to accompany the switch to opaque. It was therefore quite exciting in 2006 when three laboratories simultaneously published papers (41, 42, 129) identifying the same putative “master regulator” of the white-opaque transition, which was named in two of the studies (41, 42) “WOR1” (“white opaque regulator one”) and, in the third study (129), “TOS9,” the latter the name of an ortholog previously identified in Schizosaccharomyces pombe. The three research groups used different strategies to identify WOR1 (TOS9), but all were based on the assumption that a master switch gene controlling the transition to opaque would be down-regulated in the white phase and up-regulated in the opaque phase. In their discovery of the Wor1 regulator, Zordan et al. (41) selected six candidate genes that were shown in a previous microarray study (130) to be differentially up-regulated in the opaque phase. Homozygous mutants were generated for all six TFs in an a/a derivative of strain SC5314 and tested for switching. Of the six genes, only the deletion mutant of WOR1 resulted in the complete failure of cells to undergo white to opaque switching. Zordan et al. (41) demonstrated that ectopic expression of WOR1 in a white cell population caused mass conversion to opaque. Furthermore, Zordan et al. (41) demonstrated that the WOR1 protein bound to its own promoter in five regions, suggesting self-regulation. In their discovery of the WOR1 regulator, Huang et al. (42) performed a high copy screen for C. albicans genes that had been shown to suppress S. cerevisiae invasiveness in a flo8⁻ mutant background (131). Huang et al. (42) converted an a/α wor1⁻/wor1⁻ strain to an a/a wor1⁻/wor1⁻ strain to test the role of WOR1 in switching. The a/a wor1⁻/wor1⁻ strain did not switch from white to opaque. As in the study of Zordan et al. (41), Huang et al. (42) found that ectopic expression of WOR1 resulted in white to opaque switching in a/α cells. Huang et al. (42) also found that Wor1 induced self-expression at the native WOR1 locus. Srikantha et al. (129), hypothesizing that a master switch gene would harbor in its promoter region an a1-α2 corepressor-binding site, used a chromatin immunoprecipitation—microarray (ChIP-Chip) strategy to identify candidate genes that bound the a1-α2 repressor, then tested their expression patterns for any that were consistent with that expected of a positive regulator of switching. The expected pattern was simply no expression in a/α, a/a, or α/α white cells, and expression in a/a and α/α opaque cells. ChIP-chip analyses of Myc-tagged Mata1, the a1 component of the a1-α2 repressor, identified 52 candidate genes, including TOS9 (WOR1). Of the 25 genes with interpretable expression patterns, only WOR1 (TOS9) exhibited a pattern that was consistent with a master switch regulator. Srikantha et al. (129) demonstrated that WOR1 (TOS9) localized in the nucleus, that the WOR1(TOS9) homozygous deletion mutant, tos9Δ/tos9Δ, did not switch and that ectopic expression of WOR1 (TOS9) in white a/α cells induced the opaque phenotype. It may be a rare event when three studies simultaneously identify such a pivotal regulator and exhibit no contradictions between results. The initial studies of WOR1 and EFG1 could easily be combined into a simple two-component Yin-Yang regulatory model, in which WOR1 was self-inducing (Fig. 7A and B). When EFG1 was expressed and WOR1 repressed, the phenotype was white, and when EFG1 was repressed and WOR1 expressed, the phenotype was opaque. The analyses of WOR1 suggested that WOR1 was the master regulator of white-opaque switching and, by inference, that WOR1 was essential for switching to opaque. However, in 2020, Park et al. (53) demonstrated that the double-mutant efg1Δ/efg1Δ wor1Δ/wor1Δ generated in both MTL-heterozygous and MTL-homozygous strains could be induced to undergo the white to opaque transition. They found that the double mutants underwent mass conversion to opaque under optimum opaque-inducing conditions (GlcNAc agar, 37°C, 5% CO₂), which were also physiological. The opaque cells formed by these double mutants conformed to all aspects of the established opaque phenotype. To explain their results, Park et al. (53) argued that the basic Yin-Yang model of EFG1 and WOR1 regulation and the TF network regulating expression of WOR1 represented the dominant switch mechanism, but there also existed an alternative opaque pathway (AOP) that could bypass WOR1, which will be considered in the section “Regulatory TF Circuits” dealing with proposed models of TF circuits regulating switching.

Fig 7.

Evolution of TF models for the regulation of white and opaque. (A and B) Model based on the discoveries of Efg1 as a repressor in 1999 (23) and Wor1 as an activator in 2006 (41, 42, 129). The model does not include physical interactions between Efg1 and WOR1. It does not incorporate self-induction by Wor1. (C and D) Model developed by Zordan et al. (17) of “interlocking transcriptional feedback loops” involving WOR1, WOR2, EFG1, and CZF1 in 2007. It should be noted that in the 2007 models (17) developed by Zordan in panels C and D, the arrows are based on functional as well as binding characteristics. (E and F) Model in C, D expanded in 2013 to six TFs by Hernday et al. in 2013 (44). The 2013 models by Hernday et al. (44) in panels E and F are based solely on binding characteristics, as noted in the legend of Fig. 3, panel B and Fig. 7, panel B of reference (44). Note that the arrows in the models in panels E and F denote binding only.

PHASE REGULATION OF GENE EXPRESSION: EARLY OBSERVATIONS

But before considering proposed models based on the discovery of the central role of WOR1, it seems worthwhile to consider some of the attempts at elucidating the mechanisms of regulation prior to 2006. In 1995, Srikantha et al. (132) reported that efficient integration into the white-specific gene WH11 and the opaque-specific gene OP4 was phase specific. Integration into the native genes was higher when expressed and lower when unexpressed. If a phase-specific ORF was moved to a constitutively expressed locus, the phase specificity of integration was lost (132). The integration results suggested chromatin alterations of the genes between phases. In the following year, Srikhantha et al. (133) functionally characterized the up-stream intergenic region (UIR) of the white-specific gene WH11, generating and analyzing sequential deletion derivatives of the UIR fused to the highly sensitive reporter gene RLUC (134) at the native WH11 locus. Two strong white-specific activation sites and one weak one were identified. Gel retardation with the distal activation site and white or opaque protein extracts revealed a white-specific protein complex. Srikantha et al. (133) demonstrated that the same activation complex was absent during hypha formation, in which WH11 is also down-regulated. In 1998, Lockhart et al. (135) performed a similar functional analysis of the UIR of the opaque-specific gene OP4, generating sequential deletion derivatives of the UIR fused in frame to RLUC. A single strong transcription activation region which contained a MADS box consensus sequence was identified and found to bind to three complexes. Competition experiments with the MADS box consensus sequence of the promoter of SAP1, which is coordinately up-regulated with OP4 in the opaque phase, and a human MADS box consensus sequence demonstrated that one of the three OP4 promoter complexes was specific to the OP4 sequence (135). Although these early studies provided insights into the transcriptional control of phase-specific genes, they did not target genes one would expect to be involved in the switch event. In other words, they revealed regulation upstream or downstream of the switch event.

REGULATORY TF CIRCUITS

The two original genome-wide transcription profiling studies (29, 130) revealed that the transition from white to opaque was accompanied by the down-regulation and up-regulation of a large number of genes involved in a broad array of functions. In 2010, 7 years after the two original transcription profiling studies, Tuch et al. (31) used a “strand-specific massively parallel sequencing” analysis of cellular RNA to annotate and compare the transcriptomes of white and opaque cells. What appeared to disturb Tuch et al. (31) concerning the previous comparisons of the transcriptomes (29, 130) was that although they identified over 400 genes regulated by switching, Zordan et al. (17) found that the Wor1 regulator bound to only 58 of the 221 genes potentially regulated by Wor1. In their sequencing study, Tuch et al. (31) identified 1,306 transcripts differentially transcribed twofold or more between white and opaque cells. Approximately one-third (488) of the transcripts was novel and included antisense transcripts that encoded proteins as well as transcripts that did not encode proteins. Of the novel transcripts, 213 were encoded in a DNA sequence overlapping a transcribed region of the opposite strand, suggesting significant antisense regulation. Of the 274 new transcripts representing ORFs encoding proteins ≥40 amino acids, Tuch et al. (31) highlighted clusters of 24 new transcripts encoding short proteins differentially expressed in opaque, clustered in three genomic locations that were found only in C. albicans and its closest relation, C. dublienensis. It appears from the Candida literature that little has been done to explore the functions of these genes in switching. One take-home message from the micro-array and RNA seq analyses is that the transcriptome differences between the white and opaque phases are numerous and complex and provide a wealth of avenues for future exploration into the mechanisms that distinguish white from opaque.

The elucidation of transcriptome differences and the functional analyses of EFG1 and WOR1 represented the first steps in providing a contextual framework for studies that led to the development of models for the regulation of switching at the transcriptional level. As previously noted, a simple Yin-Yang model emerged for the regulation of switching, in which the white phenotype resulted from selective EFG1 expression and the opaque phenotype resulted from selective WOR1 expression (Fig. 8A and B). Hence, EFG1 was pivotal in orchestrating the white transcriptome, whereas WOR1 was pivotal in orchestrating the opaque transcriptome. Several initial observations were seminal in expanding this model. In 2006, in their discovery paper, Zordan et al. (41) reported that Wor1 bound to five sites along its own UIR, indicating self-regulation. Huang et al. (42) also presented evidence of a WOR1-positive feedback loop. Then in 2007, Vinces and Kumamoto (136) found that the TF Czf1 when over-expressed in white cells stimulated switching to opaque. They also found that it too bound to its own promoter. They concluded that CZF1 was an antagonist or repressor of EFG1. In the same year, Zordan et al. (17) also reported that Czf1 was a repressor of EFG1 and that an additional TF, Wor2, was a positive regulator of switching and was dependent on expression of WOR1. Deletion of CZF1 resulted in a 50-fold decrease in white to opaque switching and deletion of WOR2, the complete loss of white to opaque switching. Over-expression of CZF1 in white cells led to mass conversion to opaque, but over-expression of WOR2 had only a small stimulatory effect. The binding data of Wor1 to the upstream intergenic regions of itself, CZF1, WOR2, and EFG1 as well as mutation and expression studies led Zordan et al. (17) to propose the first complex model for the regulation of switching at the level of transcription (17) (Fig. 8C and D). In the opaque phase, Wor1 bound to a single site along the CZF1 UIR, to three sites along the WOR2 UIR, and four sites along the EFG1 UIR and five sites along its own UIR. Based on regulation in white versus opaque, relationships based on epistasis (i.e., the effects of one gene upon another assessed by mutations), binding studies, and inferred activation versus repression, a model of a genetic network of “interlocking transcriptional-feedback loops” was formulated by Zordan et al. (17) (Fig. 7C and D). The model was restricted to transcription factors and, true to the definition of a model, was the best description of the information at hand. In the model for MTL-homozygous strains, EFG1 still played a central role in establishing the white phenotype and WOR1, the opaque phenotype. When EFG1 was up-regulated and CZF1, WOR2, and WOR1 down-regulated, it resulted in the expression of the white phenotype, and when EFG1 was down-regulated and CZF1, WOR2, and WOR1 up-regulated, it resulted in the opaque phenotype. In the white phenotype, Efg1 repressed WOR2, which encoded an activator of WOR1. Because Wor1 activates WOR2, itself, and CZF1, downregulation of WOR1 releases EFG1 from repression. In the opaque phase, WOR1 was up-regulated through activation by WOR2 and by self-activation. The models for MTL-heterozygous (a/α) cells were similar, with the added role of a1-α2 as a repressor of WOR1 (17). It should be emphasized that this model (Fig. 7C and D) was based upon three types of information. For those with a theoretical bent, they will find an interesting treatment of this model in an article by Sriram et al. (137) reported in 2009, in which the white-opaque transition is regulated by interlocked positive feedback loops. In 2011, Wang et al. (138) identified the TF, Ahr1 (Zcf37), which repressed the frequency of white to opaque switching and increased the frequency of switching from opaque to white, as demonstrated by analyses of deletion mutants and ectopic expression derivatives. In 2013, 6 years after the four-component model (Fig. 4C and D) (17), Lohse et al. (139) identified a fifth TF, WOR3, which when overexpressed in white cells caused mass conversion to opaque, and when deleted, stabilized the opaque phenotype at 37°C. When overexpressed in an efg1Δ/efg1Δ background, it increased the frequency of white to opaque switching, suggesting it was a modulator of switching. Most notably, Wor3 bound to the WOR1 promoter and orf, as well as its own promoter and orf, suggesting it played a role in WOR1 expression (139). In the same year, Hernday et al. (44) developed more complex models incorporating ChIP-Chip and microfluidic DNA-binding data to identify binding interactions between the six TFs Wor1, Wor2, Wor3, CƶfI, Efg1, and Ahr1 (Fig. 7E and F). ChIP-chip analysis of white-phase cells revealed no differential binding of Wor1, Wor2, or Wor3 to any genomic location, consistent with negligible expression levels of the three in the white phase. However, Efg1, Czf1, and Ahr1 bound to UIRs controlling 73, 55, and 93 genes, respectively, in the white phase. In 23 cases, two of the three TFs bound to the same UIRs, and in five cases, all three TFs bound. More importantly, in the white phase, Efg1 bound to WOR1, WOR2, and itself, Cƶf1 bound to EFG1, WOR2, and itself, and Ahr1 bound to EFG1, WOR2, and itself. Hernday et al. (44) in their report formulated a network for the white phase based on the binding of Efg1, Cƶf1, and Ahr1 genome wide, i.e., binding to genes in the entire genome and a network for the opaque phase based solely on genome-wide binding of Wor1, Wor2, Wor3, Ahr1, Efg1, and Cƶf1 [see Fig. 2 of Hernday et al., reference (44)]. In the model, one or more of the six regulatory components in the network bound to the UIRs of 748 genes in the opaque phase. Wor1 alone bound to 68% of UIRs, and two or more of the six TFs bound to 45% of UIRs. Hernday et al. (44) then generated models based on their binding data as well as binding data from the studies of Zordan et al. (17) and Lohse et al. (139) on binding between core components of the regulatory circuitry (Wor1, Wor2, Wor3, Ahr1, Czf1, and Efg1)(Fig. 7E and F). Hernday et al. (44) also presented models of binding between core TFs as well as to other transcription regulatory genes alone or in combination. In the white phase, Efg1, Ahr1, and Czf1 bound to a total of 31 regulatory genes, and in the opaque phase, Wor1, Wor2, Wor3, Efg1, Czf1, and Ahr1 bound to approximately 67 regulatory genes. But at this point, it should be emphasized that binding of network TFs to UIR sites does not necessarily represent a regulatory interaction (140–157). In studies assessing TF-binding sites and functionality in archaebacteria (150) and S. cerevisiae (142), up to 50% of the identified binding sites proved to be potentially nonfunctional. For instance, De Vilbiss et al. (153), using ChIP-seq, identified three GATA-1/2 binding sites in the GATA2 promoter, but a deletion analysis revealed that none of them were functional.

Fig 8.

Functional analysis of EFG1 upstream intergenic region. Individual deletions of each of the nine core network TF-binding sites upstream of the white transcription start point of EFG1 have little effect on the up-regulation of expression of the RLUC reporter gene in the white phase or down-regulation in the opaque phase. Deletion derivatives of the EFG1 promoter were fused in frame to the RLUC gene and inserted in an EFG1 allele. (A) The binding of core network TFs to sites in the upstream intergenic region of EFG1. The EFG1 gene and upstream region are diagrammed, and binding of TFs noted (+) for the white and opaque phases, based on analyses synopsized by Hernday et al. (44, 158). (B) Diagrams of the deletion derivatives and expression of the RLUC reporter in the white and opaque phase. Activity is graded from ≥wild type (ⴕⴕⴕⴕ) to negligible (—). RLUC, Renilla reniformis luciferase gene; WhTSP, white transcription start point; OpTSP, opaque transcription start point. The 10 core binding sites are numbered 1 through 10. Deletions in panel B presented as multiple Xs. These diagrams were derived from reference (159).

One of the most effective methods to assess the function of TF binding is to generate targeted deletions of the binding sites and assess the effects on transcription of either the tagged orf or the activity of a reporter gene fused in frame to the gene or to the promoter deletion derivative. Pujol et al. (159) performed this type of analysis for the EFG1 UIR, generating individual targeted deletions of the 10 sites that bind one or more of the TFs Czf1, Efg1, Ahr1, Wor1, Wor2, and Wor3 (Fig. 8A and B). Pujol et al. (159) also analyzed six combinatorial deletion derivatives. The deletion derivatives were generated in the α/α strain WO-I, in which each deletion derivative of the EFG1-binding sites was fused in frame with the reporter gene RLUC (134). It should be noted that white and opaque cells were obtained from cultures grown on supplemented Lee’s medium at 30°C in air, and deletion derivatives assessed for RLUC expression. Only 4 of the 10 binding sites when individually deleted affected transcription of RLUC, but the effects were minor (Fig. 8B). Combinatorial deletion of all nine binding sites upstream of the white transcription start point resulted in negligible RLUC expression, verifying that the upstream region was necessary for EFG1 transcription. In the opaque phase, none of the single deletions or combinatorial deletions derepressed expression of RLUC (Fig. 9B), indicating that none of the binding sites functioned individually as cis-acting repressors of EFG1 expression (Fig. 8B). These results suggest that at the level of the EFG1 promoter, expression may be regulated solely by activation. However, as pointed out by Pujol et al. (159), there is still the possibility of additivity, redundancy, or functional specialization of the cis-acting binding sites under untested stimulatory conditions, which would not have been assessed in their study. Unfortunately, the 10 targeted deletion strains in the study by Pujol et al. (159) still contained one native EFG1 allele, and therefore, the heterozygous deletion derivatives could not be used to assess for the effects of the deletions on switching.

Fig 9.

Functional analysis of the WOR1 up-stream intergenic region. Individual deletions of the five core network TF-binding sites in the WOR1 locus have no effect on the white to opaque phenotypic switching frequency, but individual deletions of three (P2, P3, P4) block phenotypic commitment to the opaque phase and reduce WOR1 transcription levels. (A) The binding of core network TFs to sites in the WOR1 locus. The WOR1 gene and up-stream region are diagrammed, and binding TFs noted (+) in the opaque and white phase, based on analyses synopsized by Hernday et al. (44, 158) and Lohse et al. (160). (B) Diagrams of the deletion derivatives and functional analyses, including frequency of white to opaque switching, opaque commitment, and levels of WOR1 transcription for strain WO-1. All three functional values were graded from ≥wild type (ⴕⴕⴕⴕ) to negligible (—). np, not performed; Start and Stop, start and stop transcription points. Deletion regions presented as multiple Xs. Source of images and functional data, Conway et al. (54).

Conway et al. (54) subsequently performed a functional analysis of the binding sites of the network TFs for WOR1 but, in their study, generated targeted homozygous deletion derivatives of the 10 WOR1 sites that bound one or more of the six network core TFs, Efg1, Cƶf1, Ahr1, Wor1, Wor2, Wor3, and two additional TFs added to the network in 2016, Ssn6 (158) and Wor4 (160) (Fig. 9A). Conway et al. (54) assessed the effect each of the targeted homozygous deletions (ΔPI, ΔP2, ΔP3, ΔP4, ΔP6) and three combinatorial deletions (ΔPI-2, ΔPI-3, ΔPI-WOR1) had on transcription of WOR1 as well as the effects on the frequency of white to opaque switching (Fig. 9B). Individually, targeted deletion of any one of these five binding sites, four upstream and one downstream, of the WOR1 orf (P1, P2, P3, P4, P6), or combinatorial deletion of P1 and P2, had no effect on mass conversion (100%) from white to opaque on GlcNAc agar at 25°C in 5% CO₂, for the α/α strain WO-1 or the a/a strain P37005 (54) (Fig. 9B). Combinatorial deletion of P1, P2, and P3, however, blocked completely white to opaque switching. Individual deletion of one site, P3, caused a decrease in frequency of 18% in the α/α strain WO-1 and 90% in the a/a strain P37005. The results of the functional analysis of binding sites for WOR-1 indicated that individual deletion of anyone of the five tested network TF-binding sites or combinatorial deletion of P1 and P2 in the WOR1 UIR and, hence, their binding interactions with TFs do not affect switching from white to opaque on GlcNAc agar at 25°C in 5% CO₂. They may, however, function differently in different strains under suboptimal conditions for switching, which was shown to be the case in the a/a strain P37005 for P2 and P3 on GlcNAc agar in 5% CO₂ (54). Or they may play combinatorial functions under conditions not tested. But as will be discussed in the section “Chromatin Modification, Switching, and Commitment,” some of these binding sites appear to play individual and essential roles in the “commitment event,” simply defined as the time at which a white cell, transferred from opaque-inducing conditions to opaque-noninducing conditions, still proceeds to form an opaque cell, i.e., when a white cell commits to the formation of the opaque phenotype, even when switched to noninducing conditions. Commitment will be dealt with in the section “Commitment and the WOR1 Promoter” of this review.

In addition to the interactive TF circuit models that evolved from the original one formulated by Zordan et al. (17) in 2007, other models have been proposed, which, in most cases, were also based on the tenet that WOR1 is pivotal in the regulation of the white to opaque transition. In particular, the model of Alkafeef et al. (161) in 2018 highlighted the role of TUP1, a member of the Gro/TLE family of proteins that bind to TF complexes, not DNA, in the regulation of gene expression. In 2002, Zhao et al. (162) deleted TUP1 in the α/α strain WO-1 and found that the homozygous mutant formed a filamentous colony phenotype but did not undergo switching from white to opaque on supplemented Lee’s medium at 25°C in air. By placing TUP1 under the regulation of the inducible MET3 promoter, Zhao et al. (162) found that down-regulation of TUP1 in both white and opaque cells resulted in a transition from white to the filamentous growth form. Subsequent up-regulation of TUP1 resulted in the transition to the opaque phenotype. Prior to the study by Zhao et al. (162), Braun and Johnson (163) had demonstrated that deletion of TUP1 caused filamentous growth, and therefore, Tup1 was a repressor of filamentation. Zhao et al. (162) suggested that Tup1 was recruited to the site of the switch event and that the switch event involved chromatin modification. Alkafeef et al. (161) subsequently found by tandem affinity purification that Tup1 and associated proteins differentially bound to Wor1 in the white phase. They demonstrated that depletion of Tup1 bypasses Wor1 self-activation and that unlike WOR1, TUP1 does not regulate downstream opaque-regulated genes. In the model developed by Alkafeef et al. (161), white is induced by a glycolytic carbon source (e.g., glucose) and opaque by a nonglycolytic carbon source (e.g., glycerol) at 37°C. TupI blocks WOR1 activation in the white phase and is inhibited by Wor1 in the opaque phase, which releases WOR1 repression. Interestingly, Alkafeef et al. (161) presented data which suggest that Tup1, Ssn6, and Tcc1 are the prevalent proteins associated with Wor1 repression in the white phase. It should be noted here that Frazer et al. (164) demonstrated that the Tfs in a model based on binding among SSN6, AHR1, WOR1, EFG1, CZF1, WOR4, WOR1, and WOR3 form condensates mediated by prion-like sequences. Co-assembly would play a role in co-localization at the binding sites along the promoter regions. In considering alternative models, the one by Park et al. (53) explaining their efg1Δ/efg1Δ wor1Δ wor1Δ double-mutant results must be considered. As noted previously, in 2020, Park et al. (53) presented a simple model to explain how a white cell could switch to opaque in the simultaneous absence of Wor1 and Efg1. They argued that a basic Yin-Yang model for the regulation of white-opaque switching, in which Efg1 expression is central to the expression of the white phenotype and WOR1 expression is central to the expression of the opaque phenotype, was insufficient to explain the finding that white cells of the efg1Δ/efg1Δ wor1Δ/wor1Δ double mutant could be induced to form bona fide opaque cells (53). They concluded that simply adding to the basic model an alternative opaque pathway, which could bypass WOR1 and mediate the white to opaque switch in double mutants, rendered it sufficient. They hypothesized that for wild-type cells in the white state, Efg1 would repress WOR1 and an AOP-targeted transacting factor (Fig. 10A). For wild-type cells in the opaque state, EFG1 would be downregulated, and both WOR1 and the AOP TF derepressed, activating the white-opaque switch (Fig. 10B). In the absence of both Efg1 and Wor1 in the double-mutant efg1Δ/efg1Δ wor1Δ/wor1Δ, the AOP would be derepressed, and the unidentified AOP TF would assume Wor1 function (Fig. 10C). The model would benefit from having WOR1 also repress the AOP TF in the opaque phase (boxed in Fig. 10B). There are several caveats to this hypothesized scenario, but all can be addressed experimentally. First, what is the reason for redundancy? Could it be that the alternative opaque pathway transduces select signals that differ from those that target WOR1? In other words, are there upstream receptors and signal transduction pathways that are not shared? If this is the case, then the efg1Δ/efg1Δ WOR1/WOR1 and efg1Δ/efg1Δ wor1Δ/wor1Δ mutants might respond differently to one or more of the four sets of conditions on supplemented Lee’s GlcNAc agar since the former mutant is activated by Wor1 and the latter by the AOP TF. This seems indeed to be the case for MTL-hemizygous double mutants at 25°C in 5% CO₂ and at 37°C in air (53). Second, is it plausible that both WOR1 and the hypothetical AOP TF evolved to activate the same downstream genes involved in generating the complex opaque phenotype? Or is it plausible that both Wor1 and the AOP TF bind to and activate the same single downstream TF which in turn up-regulates and down-regulates the many downstream genes involved in the phenotypic transition? The large number and functional variety of opaque-specific characteristics suggest that a large number of opaque-specific genes are involved. Therefore, it seems more economical that both WOR1 and the AOP TF activate one or more common down-stream TFs which in turn regulate the same set of genes involved in the genesis of the same complex opaque phenotype. The hypothesized existence of an alternate opaque pathway, which explains the results of Park et al. (53), in no way detracts from the central function of WOR1 and the TF networks that have been proposed to regulate its expression. It may also not be the only model consistent with the results of Park et al. (53). It simply adds a layer of complexity that must be explored.

Fig 10.

Model to explain how double mutants that cannot express either EFG1 or WOR1 (efg1Δ/efg1Δ wor1Δ/wor1Δ) can undergo white to opaque switching at high frequency and form opaque cells that exhibit all of the characteristics of bona fide opaque cells, including the capacity to mate and the formation of cell wall pimples (53). In the model, an alternative opaque pathway other than the one involving EFG1 and WOR1 is hypothesized that activates a transcription factor (AOP TF) in double mutants. The AOP in the absence of EFG1 and WOR1 activates the same downstream genes targeted by Wor1 that are involved in the expression of the complex opaque phenotype. (A) In the white phase of wild-type cells, Efg1 represses both WOR1 and AOP TF expression. (B) In the opaque phase, EFG1 is down-regulated, resulting in up-regulation of WOR1 and the AOP TF gene, which in turn regulate the genes involved in generating the opaque phenotype. Alternatively (in brackets), in the opaque phase, down-regulation of EFG1 derepresses WOR1, which not only regulates the genes involved in generating the opaque phenotype but also represses AOP TF expression. (C) When both WOR1 and EFG1 are simultaneously deleted or repressed, the AOP TF is derepressed and activates the genes involved in generating the opaque phenotype. This model was developed simply to explain the formation of an opaque cell in the absence of EFG1 and WOR1, and in no way replaces or contradicts the TF network models for the regulation of WOR1 in wild-type (EFG1/EFG1 WOR1/WOR1) cells, Fig. 8C through F. Source of the model, Park et al. (53).

SIGNAL TRANSDUCTION PATHWAYS

Previously in this review, we considered early and late observations on conditions that affected the frequency of switching. Some of these conditions, like high CO₂ and GlcNAc, appeared stimulatory, while others, like glucose, inhibitory. One might expect these disparate signals are transduced by different receptors and/or signal transduction pathways. Inducing conditions include UV, temperature, CO₂, carbon source, and pH. In 2009, Huang et al. (46) provided evidence that CO₂ in the range of physiological concentrations (5%–20%) caused an increase in the frequency of white to opaque switching in five MTL-homozygous strains, 7- to 18-fold in 5% CO₂ and 10- to 100-fold in 20% CO₂. They presented evidence that CO₂ acted through conversion to HCO₃⁻ by carbonic anhydrase at low but not high CO₂ concentrations and showed that switching to opaque was enhanced by adenylate cyclase and Ras1 at low but not at high CO₂ concentrations. In a subsequent study, Du et al. (165) presented evidence that the TF, Flo8, was essential for CO₂-induced white to opaque switching and functions up-stream of Wor1 and Wor2. Furthermore, they found that an unidentified pathway in addition to the cAMP/PKA pathway transduced the CO₂ signal that activates white to opaque switching. Yan et al. (166) subsequently demonstrated that high CO₂ up-regulated WOS1 which encoded a SUMO E3 ligase that catalyzes Wor1 SUMO regulation, thus enhancing switching from white to opaque. In 2010, Huang et al. (47) reported that substituting GlcNAc for glucose in supplemented Lee’s agar resulted in more than a 20-fold increase in white to opaque switching after 10 days of incubation at 25°C in air. Huang et al. (47) demonstrated that the GlcNAc effect was enhanced at physiological temperature (37°C) and that the Ras1/cAMP pathway also mediated GIcNAc induction. However, they found that another minor unidentified pathway functioned in the absence of the Ras1/cAMP pathway, mediating GIcNAc induction. In a model of regulation, Huang et al. (47) indicated that all three signals they considered at the time inducive (glucose, GlcNAc, CO₂) were transduced through both the Ras1/cAMP pathway and one or more unidentified pathways, all targeting WOR1 expression. Du et al. (165) in 2015 demonstrated that overexpression of the zinc-finger TF Ofi1 differentially enhanced GlcNAc induction at 25°C in air, with no effect on the frequency of white to opaque switching on glucose agar. In 2014, Liang et al. (167) reported that HOG1 deletion mutants (hog1Δ/hog1Δ) of a/a and α/α strains, but not a/α strains, were induced to switch from white to opaque by GlcNAc and that induction in the MTL-homozygous hog1Δ/hog1Δ strains depended upon WOR1. Deletion mutants of PBS2 and SSK2, upstream components of the Hog1 MAPK pathway, reacted to GlcNAc like the HOG1 deletion mutant. They concluded that the HOG1 stress-activated protein kinase pathway, composed of Ssk2-Pbs2-Hog1, repressed WOR1-dependent switching from white to opaque in MTL-homozygous cells and may or may not do so in a/α cells, in which WOR1 is repressed by the a1-α2 corepressor. The next step in elucidating the role of this pathway in switching will be to assess the mutants for switching under all combinatorial permutations of temperature (25°C versus 37°C), sugar (glucose versus GlcNAc), CO₂ (air versus 5% CO₂), and pH (4.5 versus 6.7).

Sun et al. (52) tested the frequency of switching from white to opaque for two α/α and two a/a strains on GlcNAc agar in air or 5% CO₂ through a pH range of pH 5.0 to pH 8.0. They found that high pH favored hypha formation. Since the cAMP signaling pathway was implicated in the induction by CO₂ and GlcNAc of white to opaque switching, Sun et al. (52) tested whether deletion of the adenylate cyclase gene CYR1 influenced the effect of pH on switching. It did not. Deletion of RIM101, which encodes a transcription factor that has been shown to be involved in a pH sensing pathway, also had no effect on the decrease of switching to opaque in response to high pH. The signal transduction pathway for the pH effect on switching, therefore, remains unidentified. Finally, genotoxic and oxidative stress have been shown to affect the frequency of switching. As noted, Morrow et al. (62) demonstrated in 1989 that low doses of UV caused increases in the frequency of switching in both the white to opaque and in the opaque to white direction, presumably due to effects on chromatin structure. Twenty years later, Alby and Bennett (168) demonstrated that two other genotoxic agents, MMS, a DNA alkylating agent, and hydroxyurea (HU), an inhibitor of DNA replication, both caused increases in switching from white to opaque. Interestingly, the HU effect must be reconciled with the observation by Rikkerink et al. (60) that suspending white cells in water, which blocks cell multiplication, also blocks switching. In addition to genotoxic stress, Alby and Bennett (168) demonstrated that hydrogen peroxide, which causes oxidative stress, also causes an increase in white to opaque switching. Alby and Bennett (168) also assessed generation time both for agent-treated cells and two high-frequency switching strains and manipulated nutrient levels to affect generation time. They concluded that stress conditions that lengthen the division cycle (i.e., slowed growth rate) cause an increase in WOR1 expression, which in turn causes an increase in the frequency of switching. This conclusion contradicted the results of a study by Lohse et al. (169), in which natural strains that switched from white to opaque at different frequencies were analyzed for growth rate. They found no correlation between growth rate and switching frequency. If, in fact, the hypothesis of Alby and Bennett (168) proves to be correct, then one may consider the possibility that any environmental or nutrient condition that alters the rate of division would affect the level of Wor1 in a switching-competent cell and, in turn, the frequency of switching. The effect of growth rate on the frequency of switching, therefore, deserves further clarification. In summary, evidence of upstream signal transduction pathways affecting Wor1 levels and subsequent frequencies of switching have implicated the Ras/cAMP pathway, the Ssk2-Pbs2-Hog1-MAPK pathway, Wos1 segmentation of Wor1, and other pathways and processes affected by growth rate. The results so far support a model in which multiple signals and signal transduction pathways converge on Wor1 expression.

COMMITMENT AND THE WOR1 PROMOTER

Because switching from white to opaque or opaque to white in populations of MTL-homozygous cells can be induced to occur synchronously and en masse, it is possible to pinpoint the time at which cells transferred from noninducing to inducing conditions commit to the induced phenotype. Synchrony and mass conversion allow one to transfer temporal measurements of a population to single-cell measurements in the analysis of a phenotypic transition. In the case of the white to opaque transition, this is accomplished by transferring cells from opaque-inducing conditions back to noninducing conditions at time intervals and assessing at what time after cells are transferred to inducing conditions, they continue to form the opaque phenotype after they are transferred to noninducing conditions. Using synchronized populations, one can then compare the time of commitment with gene expression, biochemical and metabolic changes, and changes in cell morphology and intracellular architecture. In 1979, Mitchell and Soll (170) first used this procedure to identify the time at which pH-induced budding yeast cells of C. albicans commit to hypha formation. In reporting the identification of the first white-specific gene WH11 in 1993, Srikantha and Soll (14) employed a similar procedure to assess the time of commitment to the white phase in the opaque to white transition, using high temperature as the inducing condition. They found that when mid-log phase opaque cells grown in supplemented Lee’s medium at 25°C in air were transferred to fresh medium at 42°C, they synchronously committed to the white phase after 3.5 hours under inducing conditions, concomitant with the second cell doubling. The white-specific gene WH11 was abruptly up-regulated at 3.5 hours as well. In 2010, Lohse and Johnson (171) employed a similar protocol to assess the time during the transition from opaque to white at which the WOR1 transcript level decreased and the EFG1 transcript level increased in relation to commitment to white cell formation. They obtained kinetics of commitment similar to those of Srikantha and Soll (14), with a T₅₀ (time of 50% commitment) of 5 versus 3.5 hours. They found that the level of Wor1 began to decrease immediately after transfer to inducing conditions. When 50% of the population had committed to white, the level of Wor1 had decreased to 10% of its original level at time 0. Efg1 tagged with GFP, and the EFG1 transcript increased concomitantly with the increase in white-committed cells in the population. The increase in Efg1 at the time of commitment was similar to the time WH11 expression was observed to increase in the study by Srikantha and Soll (14). Single-cell measurements by Lohse and Johnson (171) of Wor1-GFP and Efg1-CFP fluorescence supported those population measurements, indicating that the synchronized population results reflected single-cell values.

The discovery of CO₂ (46) and GlcNAc (47) induction provided a method for inducing synchronous switching en masse from white to opaque and, hence, the analysis of commitment in the white to opaque transition. In 2021, Conway et al. (54) employed the combination of high CO₂ and GlcNAc as the inducing conditions, and the combination of glucose and air (0.04% CO₂) as the noninducing conditions, to analyze when a white cell transferred from noninducing to inducing conditions at 25°C committed to the opaque phenotype. Commitment time was compared to the time of evagination, chitin ring formation at the incipient septum, and WOR1 expression, in the α/α strain WO-1 and the a/a strain P37005. In their procedure (54), cells were grown under noninducing conditions to stationary phase, in which they accumulated as unbudded white cells. These unbudded white cells were then plated on glucose-based agar in air at 25°C (“noninducing conditions”) to continue to support synchronous formation of white daughter cells or on GLcNAc agar in 12% CO₂ at 25°C (“inducing conditions”) to induce synchronous formation of opaque daughter cells. At 0- and 2-hour time intervals thereafter, cells plated under inducing conditions were replated under noninducing conditions to assess the time of commitment. Under noninducing and inducing conditions, cells exhibited the same synchronous evagination kinetics, with evagination beginning at 4 hours and reaching a T₅₀ (time 50% evagination) at approximately 6 hours in both cases. The kinetics of chitin ring formation at the mother cell-bud junction were also assessed by calcofluor white staining and found to be similar to evagination kinetics under both noninducing and inducing conditions. The kinetics of opaque commitment under inducing conditions were similar to the kinetics of evagination and chitin ring formation. However, the increase in WOR1 transcription began right after cells were transferred to inducing conditions, reaching 50% maximum level at 2 hours, approximately 4 hours prior to opaque commitment by 50% of cells. These results demonstrated that induced white cells commit to the opaque phase at the same time they evaginate and initiate chitin ring formation (septation) but 4 hours after WOR1 transcription is up-regulated. Conway et al. (54) hypothesized that since WOR1 was essential to switch from white to opaque for cells harboring wild-type EFG1, the commitment event may involve a heritable change that maintains WOR1 expression under noninducing conditions and that such a change may occur at the level of the regulatory upstream intergenic region of the WOR1 ORF. As previously noted, they generated individually targeted deletion mutants of the four network TF-binding sites (41) (Fig. 10A) upstream of the WOR1 ORF, labeled P1, P2, P3, and P4, respectively, and the one binding site downstream of the WOR1 ORF, P6, as well as three combinatorial deletions, ΔP1-2, ΔP1-3, and ΔP1-wor1 (Fig. 9B). These deletion derivatives were then analyzed for switching, evagination, opaque commitment, and WOR1 expression under opaque-inducing conditions (Fig. 10B) (54). Mass conversion from white to opaque was induced in all of the single mutants (ΔP1, ΔP2, ΔP3, ΔP4, ΔP6) and in the double mutant ΔP1-2 but not in the triple mutant ΔP1-3 or the multiple mutant ΔP1-9 (Fig. 9B). Although all five individual mutants switched en masse to opaque, only the individual mutants ΔP1 and ΔP6 underwent opaque commitment like the wild-type parent strain (Fig. 9B). Hence, the individual mutants ΔP2, ΔP3, and ΔP4 derivatives switched cellular phenotype but did not commit (Fig. 9B), indicating that P2, P3, and P4 were individually essential for commitment but not for switching of cellular phenotype. Interestingly, the increase in WOR1 expression in the opaque-induced mutant derivatives ΔP1 and ΔP6 was similar to that in the wild-type WO-1 strain (Fig. 9B). WOR1 expression also increased in the mutants ΔP2, ΔP3, and ΔP4 when induced to switch, but expression was depressed by approximately half (Fig. 9B). The results of this analysis (54) revealed several aspects of WOR1 promoter function. First, induction of the switch from white to opaque by a combination of GlcNAc plus high CO₂ does not require P1, P2, P3, P4, or P6 individually, nor P1-P2 combinatorially. Second, while P1 and P6 are individually not essential for commitment to the opaque phenotype, P2, P3, and P4 are individually essential. Third, although WOR1 is essential for the switch to opaque, up-regulation occurs hours before commitment. Fourth, individual deletion of P2, P3, and P4, but not P1 and P6, causes a reduction in WOR1 transcript levels. Therefore, switching and commitment can be uncoupledat the level of the promoter. Thus, the decreases in WOR1 induction in the individual deletion mutants ΔP2, ΔP3, and ΔP4 correlate with the loss of opaque commitment but not a loss in mass conversion to the opaque phenotype. Fifth, it should be noted that in the opaque phase, sites P1, P2, P3, and P4 all bind the TFs Wor1, Wor2, Wor3, Wor4, efg1, Ahr1, and SSN6, and P2 and P3 also bind Czf1, but P1 and P4 do not. Therefore, except for Czf1 binding, the binding of the remaining TFs provides no correlative insight into function. If only induced switching is considered, individual deletion of any one of the six sites has no effect, but the combinatorial deletion of the first three sites P1, P2, and P3 (ΔPI-P3) blocks the induction of switching completely by GlcNAc and high CO₂, suggesting an additive or synergistic effect. On the other hand, P1 and P4 have the same TF-binding profile, but P1 is not essential for commitment, whereas P4 is essential. These results suggest that there are other TF components or cofactors, or a positional effect, that determine the differences in site-specific function. Fifth, the results of Conway et al. (54) also reveal differences in the role of binding sites between the α/α strain WO-1 and the a/a strain P37005. The differences are not notable for P1, P2, and P6. For CO₂ induction of mass conversion from white to opaque, deletion of P2 alone blocked GlcNAc-induced switching by a/a P37005 but not by α/α WO-1, and deletion of P1 or P6 blocked switching by P37005 but not WO-1. It should be noted that recently Ziv et al. (172) performed a single-cell analysis employing a microfluidic chamber and found that a switch to opaque involves two steps, the first predisposes a cell to switch, the second commits to the switch. This second step may in fact represent the commitment event mediated at regions P2, P3, and P4 of the WOR1 promoter (54). The relationship between these two steps and the commitment event involving P2, P3, and P4 warrants further investigation.

CHROMATIN MODIFICATION, SWITCHING, AND COMMITMENT

The roles of chromatin modification and protein complexes interacting with chromatin have been intensively studied not only in the establishment of cell fate in development and the differentiation of stem cells but also in the phenotypic changes associated with cancer (173–179). Most of the observations on the role of chromatin modification related to the white-opaque switch have been indirect. As previously noted, Srikantha et al. (132) demonstrated in 1995 that the frequency of integration in the phase-specific genes WH11 and SAP1 correlated with their level of phase-specific expression, indicating that changes in chromatin architecture accompany gene expression in white-opaque switching. Kvaal et al. (76) demonstrated that misexpression of WH11 in the opaque phase caused a 100-fold increase in switching from opaque to white but had no effect on the frequency of switching from white to opaque. However, since WH11 localizes in the cytoplasm (180) and Sap1 is secreted, it is unlikely that either plays a direct role at the genomic level in the switch event. In 2001, Grewal et al. hypothesized that his work on gene silencing in the Schizosaccharomyces pombe mating system and the role of histone deacetylases (181) might serve as a model for the regulation of switching, and therefore tested the effect of trichostatin-A (TSA), a histone deacetylase inhibitor. Klar et al. (25) demonstrated that TSA increased the frequency of white to opaque switching in C. albicans by over 100-fold. They further demonstrated that deletion of HDA1, encoding a histone deacetylase, a prime target of TSA, resulted in a 30-fold increase in switching from white to opaque. In a follow-up study, Srikantha et al. (26) demonstrated that deletion of RPD3, a second histone deacetylase, not only caused a 25- to 50-fold increase in switching from white to opaque but also an 83-fold and a 16-fold increase in the frequency of switching from opaque to white in repeat experiments. Northern blot analysis revealed that both HDA1 and RPD3 transcript levels in the respective mutants were down-regulated in the opaque phase, but expression of WH11 and EFG1 was still up-regulated in the white phase and OP4 and SAP1 up-regulated in the opaque phase. In 2009, Hnisz et al. (27) identified eight transcription regulators and chromatin modifiers that affected the frequency of switching. Hnisz et al. (27) found that the fold change in the frequency of white to opaque switching by the mutants set1Δ/Δ, hda1Δ/Δ, rpd3Δ/Δ, set3Δ/Δ, hos2Δ/Δ, hst2Δ/Δ, and nat4Δ/Δ was 1.7-, 2.7-, 2.8-, 0.2-, 0.13-, 0.04-, and 0.12-fold, respectively. No mutant exhibited mass conversion from white to opaque or was completely blocked in the white or opaque phenotype, as was also the case in the previous studies of HDA1(175) and RPD3 (26). Thus the role of these chromatin modifiers was either as modulators of frequency at the site of switching (i.e., WOR1) or indirect. In 2016, Xie et al. (28) demonstrated in an a/α strain capable of white-opaque switching that deletion of several histone modifiers increased or decreased the frequency of white to opaque switching several fold, and deletion of the paralogs Ppd3 and Rpd31 played opposite roles. In 2011, Stevenson and Liu (182) demonstrated that deletion of the histone H3K56 acetyltransferase, Rtt109, depressed the frequencies of spontaneous and induced switching from white to opaque. The frequency of switching from white to opaque by wild-type cells on glucose agar was approximately 3.5% whereas that of the deletion mutant rtt109/rtt109 was 0.5%, a sevenfold decrease. Switching by wild-type cells on GlcNAc agar was approximately 97% and by the rtt109/rtt109 mutant 10%, a 10-fold decrease. Acetylation of the histone H3 at lysine 56 (H3K56) causes disassembly of chromatin, which occurs in gene activation. The conclusions by Stevenson and Liu (182) were consistent with the observation that genes activated in white or opaque cells are more accessible to plasmid integration (132). However, as with most of the studies of chromatin modifiers and genes encoding proteins regulating chromatin architecture, mutants of these genes may also affect other aspects of cell physiology related to gene activation, for instance, the cell cycle in which HST3 and HST4 are differentially expressed in the G2 and M phases in S. cerevisiae (183, 184). These mutations may also affect the activation of genes downstream or upstream of WOR1, which are activated or silenced in the transition from white to opaque or opaque to white. One year later, Stevenson and Liu (185) presented evidence, indicating that the chaperone proteins Cac2 and Hir1 of the complexes CAF-1 and HIR, respectively, which function in nucleosome assembly, also played roles in the frequency of switching in a/a cells. Cac2 appeared to have a modulatory effect on white to opaque switching on glucose agar and no role in switching on GlcNAc agar, while HIR2 played no role on either sugar. Stevenson and Liu (185) did find that, on GlcNAc agar, the frequency of switching from opaque to white was over 90% for wild type, the cac2Δ/cac2Δ mutant and the hir1Δ/hir1Δ mutant, but it was reduced fivefold to 18% in the cac2Δ/cac2Δ hir1Δ/hir1Δ double mutant, suggesting they may perform synergistic functions. In a report by Lohse et al. (186) in 2016, 196 individual mutants, each containing a mutated gene for a transcription regulator, were analyzed for white-opaque switching. The study included results for 16 additional genes encoding proteins known or presumed to play roles in histone modification and chromatin remodeling. In 2021, Qasim et al. (187) reviewed the switching data of these 16 genes and 12 additional genes from other studies, for a total of 28 genes that included 8 histone acetyltransferases, 3 histone methyltransferases, 12 histone deacetylases, and 5 other genes (2 phosphatases, 2 chromatin remodelers, 1 histone chaperone). They concluded that 10 of the 28 mutants had no effect (SPT10, HPA2, SAS2, ELP3, SET2, DO1, HOS1, HO33, ORF19.4736, HIR1), 14 had a modest effect (≤10-fold) (RΠ109, NAT4, HAT1, SET1, HDA1, HDA2, HDA3, RPD31, HST1, HST3, HOS2, SET3, CAC2), and 4 had a substantial effect (>10-fold) (YNG2, RPD3, HST2, SWR1). Of the latter four, two, RPD2 and HST2, had already been identified as having substantial effects in previous studies (25–27). One of the most interesting studies, by Zhang et al. (188) in 2013, demonstrated that mutations in components of the Mediator complex had component-specific effects on the frequency of switching in both the white to opaque and opaque to white direction. Mediator complexes play roles in gene silencing in epigenetically regulated differentiations or responses to environmental perturbations by regulating chromatin architecture and polymerase II function (179, 189). Deletion of seven C. albicans Mediator subunits resulted in significant increases or decreases in the frequency of white to opaque switching and, in the case of two components, resulted in frequencies of <10⁻⁴ (188). In five cases, the mutation increased or decreased the frequency of opaque to white switching. In summary, the mutational studies of chromatin modifier genes and mediator genes indicate that some play modulatory roles in the frequency of the switch event. The next challenge is to differentiate between which of them do so at the site of the switch event and which affect switching indirectly.

MATING, INTERCELLULAR SIGNALING, AND FORMATION OF A “SEXUAL BIOFILM”

When C. albicans a/α cells are incubated on a silicone elastomer in RPMI 1640 medium, they form a basal yeast cell polylayer and an upper layer of hyphae vertically oriented and embedded in dense matrix (19, 190, 191). In 2006, Daniels et al. (19) found that MTL-homozygous (a/a and α/α) white cell populations formed biofilms with the same layered architecture as a/α biofilms but were approximately one-third as thick. In contrast to MTL-homozygous white cells, when MTL-homozygous opaque cells were similarly plated, they formed a thin loose layer of interspersed opaque cells and hyphae, easily disrupted by agitation. However, if a minority of opaque cells (1%–10%), half a/a and half α/α, were mixed with a majority of white cells, the resultant biofilm doubled in thickness and retained the layered complex architecture. These results suggested that under optimum conditions for biofilm formation in RPMI medium, a medium originally formulated for host cell viability, minority opaque cells dramatically enhance majority white cell biofilm formation. It had previously been demonstrated by Lockhart et al. (192) in 2003 that treatment of white a/a cells with α-pheromone did not block white cells in the G1 phase of the cell cycle, induce mating tube formation, up-regulate expression of FIG1, a membrane protein required for efficient mating, or up-regulate KAR4, required for efficient karyogamy, as it did in opaque cells (Fig. 11A). It did, however, up-regulate STE2, the α-pheromone receptor and STE4, a G-protein subunit in the mating response pathway, as it did in opaque cells (Fig. 11). Using biotinylated α-pheromone staining, Daniels et al. (19) further demonstrated that a/a white cells expressed α-pheromone receptors, evenly dispersed across their surface. Daniels et al. (19) postulated that the selective biofilm response of white cells to pheromone of opposite mating type played a role in facilitating mating of a/a and α/α cells in host niches (Fig. 11B and C). In natural niches, switching to opaque by MTL-homozygous white cells was a rare event in a developing white cell biofilm and pheromones released by minority opaque cells would reciprocally stabilize the phenotypes of opaque cells of opposite mating types by blocking them in GI. These rare opaque cells would also release gradients of pheromone that would direct chemotropism of a/a and α/α conjugation tubes so that tube fusion at the apices of the tubes was spatially accurate. Released pheromones would also stimulate the formation of a white cell biofilm with a matrix that stabilized pheromone gradients over long distances, while at the same time porous enough not to interfere with the directed elongation of the conjugation tubes in the process of chemotropism. Daniels et al. (19) postulated that pheromone stimulated MTL-homozygous white cells to form a “sexual biofilm” in contrast to a/α cells, which formed a “pathogenic biofilm” (Fig. 11B). In support of this hypothesis, several MTL-homozygous biofilm characteristics were identified (19). First, it was demonstrated that pheromone-induced formation of an adherent, coherent, compact polylayer of white cells on plastic, the first step in the formation of a complex C. albicans biofilm (19, 190, 191). Pheromones did not induce a similar coherent polylayer of opaque cells (Fig. 11A through C). Second, Daniels et al. (19) found that the addition of minority opaque cells (1% or 10%) to majority of the white cells resulted in a doubling of biofilm thickness. Third, Daniels et al. (19) demonstrated at the cellular level that in white cell biofilms, minority a/a and α/α opaque cells within 50 µm of each other formed conjugation tubes that extended in the direction of each other, indicating that they were undergoing chemotropism in gradients of pheromone of opposite mating type. Together, these observations suggested that majority-MTL-homozygous white cells, stimulated by pheromone released by minority opaque cells, formed a biofilm that facilitated mating between the minority opaque cells of opposite mating type.

Fig 11.

White cell pheromone response and opaque-white signaling in the formation of a white “sexual” biofilm. White MTL-homozygous cells form “sexual” biofilms in RPMI medium that support mating between minority opaque cells, while MTL-heterozygous cells form “pathogenic” biofilms that are resistant to antifungals and impenetrable by white blood cells but do not support mating (19). (A) Responses of MTL-homozygous opaque cells and MTL-homozygous white cells to pheromone of opposite mating type. (B) Characteristics of MTL-heterozygous (a/α) “pathogenic” biofilms and MTL-homozygous (a/a, α/α, a/a + α/α) “sexual” biofilms formed in RPMI 1640 medium. (C) Majority white cells switch at low frequency to minority opaque cells, which in turn signal majority white cells to form a white sexual biofilm by releasing pheromone of opposite mating type in a paracrine-type fashion (C, first panel). Addition of minority opaque cells increases the pheromone signal and enhances sexual biofilm formation (C, second panel). Note that opaque cells alone do not make biofilms under the in vitro conditions employed.

In further support of this model (19), Yi et al. (43, 193) demonstrated that although MTL a/α biofilms and both white MTL a/a and MTL α/α biofilms formed in RPMI 1640 medium exhibited similar complex architecture, a/a and α/α biofilms differed functionally from a/α biofilms. a/α biofilms were relatively impermeable to low molecular weight molecules, impenetrable by leukocytes, and resistant to fluconazole (Fig. 11B). In marked contrast, white a/a and α/α biofilms were permeable, penetrable, and susceptible to fluconazole (Fig. 11B). This suggests that while a/α strains, which dominate in nature, form a “pathogenic” biofilm that is impermeable to antifungals and impenetrable by white blood cells, a/a and α/α white cells form a “sexual” biofilm that supports mating. It was then demonstrated that the white cell pheromone response pathway shared the same upstream components with the upstream pheromone response pathway of opaque cells but targeted a different transcription factor (193, 194). The homology between the pheromone response pathway in C. albicans and that in S. cerevisiae suggested that they evolved from a common ancestral pathway and that the pheromone response pathway of C. albicans white cells, resulting in a sexual biofilm, may have evolved from the opaque pheromone response pathway, with the substitution of Tec1 for CpH1 as the target TF.

The next year, Yi et al. (195) delved into the observations that deletion of STE2, the receptor for α-pheromone, caused a reduction in white a/a biofilms in the absence of opaque cells, the source of α-pheromone. Alby et al. (196) had previously shown that a/a cells secreted α-pheromone, resulting in self-induction of low-frequency same-sex mating. Yi et al. (195) demonstrated that white a/a cells released α-factor that self-induced white a/a cell biofilm formation (Fig. 11C) and that α/α white cells released a-factor that self-induced white α/α cells to form white α/α biofilms. These biofilms could be further enhanced by adding a minority of opaque cells of opposite mating type presumably as a source of pheromone (Fig. 11C). The results of Yi et al. (195) demonstrated self-activation in MTL-homozygous same sex biofilm formation (197) (Fig. 11C) that required switching and was paracrine-based—i.e., involved extracellular signaling between two phenotypes of the same origin and sex. In the model for a/a white cell biofilm formation, white a/a cells switch to opaque at low frequency (Fig. 11C). These rare opaque a/a cells then release in a noncanonical fashion α-pheromone at low levels, which induces majority a/a white cells to form a thicker complex a/a white cell biofilm that facilitates opaque cell mating. In the model for α/α white cell biofilm formation, white α/α cells switch at low frequency to opaque α/α cells, which release at low levels a-pheromone that induces majority α/α white cells to form a thick complex α/α white cell biofilm. As with any evolving model, there are caveats. In 2013, Lin et al. (198) found that in contrast to the results of Sahni et al. (194), Cph1, not Tec1, was the targeted TF of the map kinase pathway, in the white a/a cell response to α-pheromone. Although the result of Lin et al. (198) would appear to contradict those of Sahni et al. (194), the conditions employed by Lin et al. (198) differed from those employed by Sahni et al. (194), raising questions of legitimate comparison. Daniels et al. (19, 191), Yi et al. (43, 193, 195, 199), and Sahni et al. (194) employed conditions that generated a highly organized biofilm on a silicon elastomer or a tissue culture plastic Petri dish surface. The conditions included RPMI 1640 medium, a temperature of 37°C , in air with gentle rocking. Lin et al. (198) employed quite different conditions for biofilm formation. The initial cells were grown in YPD medium, inoculated onto silicon squares, gently agitated for 90 minutes at 37°C in air, then washed in PBS, and finally incubated in Spider medium with or without pheromone. The differences in the resulting biofilms in the alternative studies were experimentally demonstrated in a study by Daniels et al. (191). Daniels et al. (191) demonstrated that the use of Spider medium does not provide an experimental platform for studying the unique interplay between MTL-homozygous white and opaque cells in the formation of a sexual biofilm. On the other hand, the biofilms formed in Spider medium may more accurately reflect the architecture of some biofilms formed in select natural niches, in particular catheters. The distinctions between “pathogenic” and “sexual” biofilms formed by MTL-heterozygous and MTL-homozygous white cells, respectively, represent a unique characteristic of C. albicans and one of the more interesting functions of white-opaque switching. It should be noted that in bacteria, which form complex biofilms , a distinction between a pathogenic and sexual biofilm has not been demonstrated, indicating that such a distinction is unique to C. albicans and the white-opaque transition. It therefore deserves further exploration.

CONCLUDING REMARKS

In this general and to some degree chronological review, the objective was to describe not only the variety and evolution of information that has been gleaned over the last 35 years since the discovery of the white-opaque transition but also to point out interesting questions that remain unanswered. Therefore, rather than simply summarize the results and conclusions presented in the preceding text in these concluding remarks, it would be more productive to emphasize some of the major questions which remain unanswered with the hope that readers find them interesting enough to pursue and ultimately resolve.

Opaque EVs and cargo: are they involved in mating?

The first EVs of C. albicans described in the literature were those formed at the pimple apices of the mature opaque cell. Renewed interest in EVs of C. albicans has emerged regarding their association with biofilm matrix formation and drug resistance. In the formation of a sexual biofilm, opaque cells signal white cells to form a biofilm by releasing pheromones in a noncanonical fashion. Could the cargo of EVs released by opaque cells contain mating pheromones? In addition, could cargo be related to the cavitation in mouse skin caused by opaque cells or the matrix formed by opaque cell populations formed on mouse skin?

Opaque cell colonization: is it for real?

One of the most basic assumptions in the study of host colonization of white and opaque cells is that the assay for the phenotype of the colonizing cells, namely plating of bodily fluids, organ macerates or feces on agar medium, and subsequent assessment of colony phenotypes 5–7 days later, provides an accurate measure of phenotypes at sites of colonization. The influence of in vitro assay conditions on switching in both the white to opaque and opaque to white direction after 5–7 days, on a variety of different agar media under nonphysiological conditions, could very well change the original phenotype, especially from opaque to white, and even from white or opaque to gray. The gray phenotype may not be expressed or may be unstable at 37°C in vitro and very likely in vivo as well. This has led to the recent conclusion that EFG1 mutants, which may represent a significant portion of natural strains, may express the opaque phenotype, not the gray phenotype, in gastrointestinal colonization. Yet, most studies of the physiology, developmental biology, drug susceptibility, and other aspects of C. albicans are performed in vitro on a/α cells without regard to their phenotypic potential. We may therefore be missing part of the story.

EFG1 mutants in clinical isolates: why so frequent?

Deleting EFG1 provides a competitive advantage for colonization in the mouse GI model. If this is the case, one might have expected the natural efg1⁻/efg1⁻ mutants to take over (i.e., become the dominant phenotype) if the GI tract is the main C. albicans reservoir. Is it in the process of doing so?

“Spontaneous” versus “inductive”: is it the wrong nomenclature?

For decades after the discovery of switching, the frequency of white to opaque switching on glucose-based agar at 25°C in air was far higher than mutational frequencies. It usually varied between 0.1% and 10%. However, studies of conditional effects on frequency revealed that at 37°C, in high CO₂ and GIcNAc as the sugar source, one obtained mass conversion (˃90%) from white to opaque. Then it was demonstrated that while GlcNAc induces switching, glucose added to GlcNAc medium actually inhibited switching. Referring to switching on glucose-based agar as “spontaneous” and switching on GlcNAc-based agar as “inductive” must therefore be reevaluated. The same question is applicable to studies in air rather than high CO₂. One could argue that the conditions of high CO₂ and GlcNAc as the carbon source are physiological and conditions of low CO₂ and glucose repressive. Unfortunately, most studies in vitro of switching and the opaque phenotype have been performed under nonphysiological conditions.

“Developmental transition” versus “terminal phenotypes”

Because the conditions for obtaining synchronous mass conversion from white to opaque were elucidated approximately 25 years after the discovery of white-opaque switching, most of the earlier research on switching involves comparisons of the two terminal, established phenotypes, “white” versus “opaque.” Only a few studies include temporal descriptions of the events accompanying the transition in time conversion from the white to opaque, which very likely involves transient events missed by simply comparing the terminal phenotypes. Synchronous mass conversion now provides the tool to investigate the time course of the events basic to the switching process at the genetic, biochemical, and cellular levels of analysis.

Models of TF regulation

The simple model of TF regulation that included EFG1 and WOR1 has evolved into “clouds” of genes that bind to a core set of TFs which are represented as hubs. With the increase in complexity inferred from binding studies, it has become challenging to develop regulatory TF models. Some of the newest and provocative models of TF regulation of switching based on binding include hundreds of connections between core TFs and genes, without proof of regulatory or functional relationships. The challenge in interpreting these complex models is threefold. First, it is imperative to elucidate which TF-binding events are involved in the phenotypic transition based upon functional analyses. It has been emphasized in this review that binding alone does not necessarily translate to regulatory function, and this point must be emphasized for the uninitiated reader. Second, investigations must be performed on the fascinating differences gleaned from the transcriptome data of Tuch et al. (31), which include Wor1-bound genes that are noncoding, antisense transcripts of genes that code proteins, transcripts of genes with unusually long UTRs, and the identification of 1,306 differentially expressed transcripts. Finally, temporal comparisons between white and opaque must be performed under multiple sets of environmental conditions with multiple strains.

Wor1 and an AOP: is there a common downstream yet-to-be-discovered “master switch gene”?

Demonstration that the double-mutant efg1Δ/efg1Δ wor1Δ/wor1Δ cells can switch to bona fide opaque cells must be further investigated. There is no question that Wor1 plays a central role in the regulation of the white-opaque transition. Therefore, it was no accident that three independent research groups using different strategies simultaneously identified Wor1 as a major positive TF regulator of the white to opaque transition. The initial characterizations of WOR1 mutants all indicated that it represented the essential “master switch gene” regulating the downstream events leading to the opaque phenotype. However, 15 years later, it was demonstrated that if WOR1 is deleted in an efg1Δ/efg1Δ mutant background, white cells of these double mutant can still be induced to switch to opaque en masse by high CO₂ and GlcNAc as the carbon source, at 37°C. Opaque cells of the induced double mutants exhibited all tested opaque-specific characteristics. Therefore, WOR1 and EFG1 are not essential for the switch from white to opaque when both are simultaneously deleted. These results could be explained by a model in which EFG1 represses an alternative WOR1-independent opaque pathway, the “AOP”, which regulates the same downstream genes as those regulated by Wor1 and may do so by targeting an additional TF also regulated by Wor1. The hypothesized AOP and common target TF must now be verified and identified. It has also been proposed that Wor1 and the AOP might target a common downstream regulator that may in fact prove to be the true “master switch gene.” In addition, we must find out what role the AOP plays—i.e., why might it exist?

What is upstream and what is downstream of Wor1 and the AOP TF?

No clear picture has emerged of the receptors and signal transduction pathways upstream of Wor1 and the hypothesized AOP TF that regulate activation and, hence, dictate switching frequencies. And even less attention has been paid to the events downstream of Wor1 and, now possibly AOP as well, that are directly involved in generating the unique combination of cellular characteristics and functions of the opaque phenotype. Identifying proteins involved in opaque-specific traits, such as the formation of the opaque-specific pimple, and elucidating how the genes encoding them are regulated may facilitate the elucidation of upstream regulatory events.

Is there really a “sexual biofilm”?

By using RPMI 1640 medium, which was developed to culture mammalian cells and which is optimal for supporting the formation of a complex, we organized C. albicans biofilm, a distinction can be made between a “pathogenic biofilm” formed by a/α cells and a “sexual biofilm” formed by MTL-homozygous cells. The pathogenic biofilm is relatively impermeable to low molecular weight molecules, impenetrable by human white cells, and resistant to antifungals, while the sexual biofilm is permeable, penetrable, and susceptible to antifungals and facilitates mating. Formation of a sexual biofilm by MTL-homozygous cells depends on a paracrine-like extracellular signaling mechanism in which minority opaque cells, formed by low-frequency switching, signal majority white cells by releasing in a noncanonical manner pheromone of opposite mating type at low concentrations, which stimulates white cells to form the complex sexual biofilm. The sophistication and complexity of this signaling system are quite remarkable and may provide possible insights into why C. albicans may have incorporated the white-opaque transition into the mating process. This complex paracrine signaling system must be verified.

Why is white-opaque switching restricted to C. albicans and highly related species

White-opaque switching appears to be restricted to three related species of Candida, C. albicans (8), C. dubliniensis (33), and C. tropicalis (34), all of which are pathogenic in humans. Less related pathogenic Candida species, such as C. parapsilosis (200) and C. glabrata (197), do not appear to undergo white-opaque switching. Might there be a specific aspect of virulence or pathogenicity shared solely among the three related switching species that is mediated by white-opaque switching?

The discovery of white-opaque switching was initially considered interesting, but an enigma lacking a function in the basic biology of C. albicans. However, a few scientists persisted in their characterization based on the belief that a phenotypic transition so complex must serve an important function. Their faith was born out over time as new discoveries revealed its uniqueness and complexity and, finally, its possible roles in mating and host colonization. The literature reviewed here suggests that the scientific community has just scratched the surface in understanding how white-opaque switching is regulated, its role in C. albicans mating, its role in sexual biofilm formation, and where and how switching affects host colonization. Clearly, the best is yet to come.

Biography

David R. Soll received his BS, MS, and PhD at the University of Wisconsin, was a postdoctoral fellow at Brandeis University and was the Carver/Emil Witschi Professor of Biology at the University of Iowa. Currently he is Professor Emeritus. He also served as Director of the Developmental Studies Hybridoma Bank, a national resource created by NIH. His research interests in cell motility and in the developmental biology of fungi began at Wisconsin in 1965. Having discovered white-opaque switching in 1987, he continued studying the phenomenon until his retirement in 2022.