Homothallic and heterothallic mating in the opportunistic pathogen Candida albicans

Abstract

Candida albicans is the most common fungal pathogen in humans, causing both debilitating mucosal infections and potentially life-threatening systemic infections^1,2. Until recently, C. albicans was thought to be strictly asexual, existing only as an obligate diploid. A cryptic mating cycle has since been uncovered in which diploid a and α cells undergo efficient cell and nuclear fusion, resulting in tetraploid a/α mating products (refs. ^3–6). Whereas mating between a and α cells has been established (heterothallism), we report here two pathways for same-sex mating (homothallism) in C. albicans. First, unisexual populations of a cells were found to undergo autocrine pheromone signalling and same-sex mating in the absence of the Bar1 protease. In both C. albicans and Saccharomyces cerevisiae, Bar1 is produced by a cells and inactivates mating pheromone α, typically secreted by α cells^7–10. C. albicans Δbar1 a cells were shown to secrete both a and α mating pheromones; α-pheromone activated self-mating in these cells in a process dependent on Ste2, the receptor for α-pheromone. In addition, pheromone production by α cells was found to promote same-sex mating between wild-type a cells. These results establish that homothallic mating can occur in C. albicans, revealing the potential for genetic exchange even within unisexual populations of the organism. Furthermore, Bar1 protease has an unexpected but pivotal role in determining whether sexual reproduction can potentially be homothallic or is exclusively heterothallic. These findings also have implications for the mode of sexual reproduction in related species that propagate unisexually, and indicate a role for specialized sexual cycles in the survival and adaptation of pathogenic fungi.

The hemiascomycetous fungi are a diverse range of species that include the prototypical model yeast, Saccharomyces cerevisiae, as well as the prevalent human pathogen, Candida albicans. Sexual reproduction in both fungi is driven by the secretion of sex-specific pheromones that promote mating between partners of the opposite sex^11–13. Pheromone gradients are sensed by G-protein-coupled cell-surface receptors that initiate the mating response^14,15. C. albicans mating is thought to be exclusively heterothallic, whereas isolates of S. cerevisiae that express HO endonuclease are homothallic because of mating-type switching¹⁶. The discovery that unisexual populations of C. albicans can express both a and α mating pheromones ¹⁷ led us to test whether autocrine pheromone signalling and homothallism can occur in this species.

C. albicans a cells derived from the laboratory strain SC5314 were treated with synthetic α-pheromone and the quantitative polymerase chain reaction used to compare the induction of both a- and α-pheromone genes (MFa, and MFα, respectively). Opaque-phase cells were used for these experiments, as these are the mating-competent form of C. albicans ⁶. Both MFa (300-fold) and MFα (373-fold) were highly induced in response to α-pheromone, in contrast to S. cerevisiae where only a-pheromone is upregulated¹⁸. In both C. albicans and S. cerevisiae, an extracellular barrier activity acts to antagonize the function of α-pheromone. The activity, encoded by BAR1, is an aspartyl protease that cleaves α-pheromone, inactivating it^7–10. We therefore analysed a C. albicans Δbar1 mutant to determine whether α-pheromone production has a functional role in these cells. C. albicans Δbar1 cells exhibited distinctive colony morphologies when cultured on solid media. Mutant colonies appeared increasingly wrinkled during growth on yeast peptone dextrose (YPD) medium (Fig. 1a). Cells retrieved from these colonies were highly elongated and showed similarities to cells that had undergone extended exposure to pheromone¹². To confirm that elongated cells were undergoing a mating-type response and were not hyphal (filamentous) cells, green fluorescent protein (GFP) reporters were introduced under the control of pheromone-specific promoters. Both FIG1 and FUS1 genes are highly expressed in response to pheromone¹⁷, and pFIG1-GFP and pFUS1-GFP constructs were therefore used to monitor activation of the mating program. Significantly, both FIG1 and FUS1 expression was induced, confirming that cells were undergoing a mating-like response, even in the absence of exogenous pheromone or a mating partner of the opposite sex (Fig. 1b). To confirm that this phenomenon was not limited to strains derived from SC5314, Δbar1 mutants were also generated in P37005, a natural a/a isolate of C. albicans, and found to exhibit similar cell and colony phenotypes (Supplementary Fig. 1).

Figure 1. Deletion of BAR1 results in an autocrine mating response.

a, Patches of wild-type and Δbar1 a cells after 4 days growth on YPD medium. b, Mating response of wild-type and Δbar1 a cells with and without α-pheromone in Spider medium for 36 h. Differential interference contrast (DIC, left) and GFP fluorescence (GFP, right) images are shown. Induction of the mating response was monitored using GFP reporters under the control of the mating-response genes FIG1 or FUS1. All C. albicans cells used in this study were in the opaque phase as this is the mating-competent form of the organism⁶,³⁰. Scale bars, 16 µm.

Two models were considered for how autocrine pheromone signalling could lead to activation of a mating response. First, a-pheromone secretion could signal to Ste3, the receptor for a-pheromone (Fig. 2a, model I). Alternatively, α-pheromone could induce mating via binding to Ste2, the receptor for α-pheromone (model II). Although not mutually exclusive, these two models were tested by construction of strains lacking different combinations of pheromone or pheromone receptor genes. Significantly, loss of either α-pheromone (MFα) or its receptor (Ste2) completely abolished the wrinkled colony appearance of Δbar1 a strains (Fig. 2b). In contrast, loss of STE3 (encoding the receptor for a pheromone) had no effect on colony phenotype (Fig. 2b). Thus, autocrine signalling in a cells is driven by α-pheromone production and requires the presence of Ste2, the cognate receptor for α-pheromone (model II).

Figure 2. Autocrine signalling in C. albicans a cells.

a, Model of conventional paracrine signalling between a and α cells and two models for autocrine signalling. Note that ‘autocrine’ signalling is used here to refer to signalling within populations of a cells. b, C. albicans a cells after 4 days growth on YPD medium. Whereas Δbar1 strains showed a highly wrinkled morphology, reconstituted strains in which the BAR1 gene had been reintegrated did not show wrinkling (not shown). c, Demonstration of autocrine signalling among different populations of a cells after 4 days growth on YPD medium. Wrinkling at the edge of a test colony is indicative of autocrine signalling.

Autocrine pheromone signalling was also observed between neighbouring colonies of cells, as Δbar1 a cells induced a response in adjacent colonies of a cells. Cells from wrinkled colonies again exhibited polarized growth indicative of a mating response (data not shown). Neighbouring cells failed to respond in the absence of STE2, indicating that signalling required the receptor for α-pheromone (Fig. 2c, row 2). Furthermore, Δbar1 cells were unable to induce a response in adjacent cells if either MFα or STE2 was also deleted (Fig. 2c, rows 3 and 4). The dependence of signalling on STE2 is demonstrative of a positive feedback loop in these cells; α-pheromone secreted by Δbar1 cells binds to Ste2 receptors on these cells and results in further augmentation of pheromone production. Only when this feedback loop is in effect does pheromone secretion reach the levels necessary for autocrine signalling. STE3, encoding the receptor for a-pheromone, was not involved in this signalling (Fig. 2c, column 4 and row 5). Production of α-pheromone by Δbar1 a cells was also evident in halo assays and was shown to increase biofilm formation by these cell types (see Supplementary Information).

The existence of a robust autocrine mating pathway led us to test whether C. albicans is capable of homothallic mating. Two different auxotrophic Δbar1 strains were mixed and co-incubated on Spider medium for 3 days. Cells were recovered from the mating mixture and plated onto selective medium to detect potential a–a mating products. A number of prototrophic mating products were obtained and analysis by PCR and flow cytometry revealed that these were stable tetraploid cells of mating type a (Fig. 3a, b). To compare the frequency of homothallic and heterothallic mating in C. albicans, several quantitative mating crosses were performed (Fig. 3c). Mating between wild-type a and α cells occurred efficiently (94%), whereas no mating was observed between two wild-type a strains (<0.00002%). Significant levels of unisexual mating were obtained, however, using two Δbar1 a strains (0.037%), an increase of more than 1,850-fold over wild-type a–a mating. The efficiency of unisexual mating was further increased in crosses between Δbar1 and wild-type a strains, with 0.88% of cells undergoing mating, an increase of more than 40,000-fold over that between wild-type a strains (Fig. 3c). In fact, Δbar1 a cells mated as efficiently with other a cells as with α cells (0.80%), indicating a loss in sexual preference for one mating-type partner over the other. These experiments demonstrate that mating can take place even within unisexual populations of C. albicans cells, and that Bar1 acts to limit these types of self-fusion events. Notably, mating between wild-type a cells also occurred at an appreciable frequency (0.22%) if co-cultured with α cells. In these ménage à trois matings, α cells presumably provided levels of α-pheromone that were sufficient to overcome Bar1 activity and drive a–a mating. Levels of α-pheromone may be augmented by autocrine pheromone signalling in a cells, as deletion of MFα in a cells led to a slight reduction in the frequency of ménage à trois mating (0.12%). In addition, α cells did not need to be in contact with a cells to drive same-sex mating (Supplementary Fig. 4). Mating between wild-type α cells (0.043%) was similarly observed when co-cultured with a cells (see Fig. 3).

Figure 3. Homothallic mating in C. albicans.

a, Flow cytometric analysis indicating the tetraploid content of progeny isolated from a–a matings. The x-axis of each graph (Sytox) represents a linear scale of fluorescence and the y-axis (Cell counts) a linear scale of cell number. For analysis of additional same-sex mating products see Supplementary Fig. 6. b, PCR confirming the mating type of a–a mating products using primers directed at OBPa and OBPα. Odd-numbered lanes are OBPa reactions, whereas even-numbered lanes are OBPα reactions. Lanes 1, 2 and 3, 4 are control reactions for a and α cells, respectively. c, Table of mating frequencies for different genetic crosses. Values are listed as mean frequency ± s.d.

If self-mating in unisexual populations is a direct consequence of autocrine pheromone signalling, then mutants deficient in signalling should be unable to self-mate. Both Δbar1/Δste2 and Δbar1/Δmfα a strains did not exhibit autocrine signalling and also showed no detectable mating with wild-type a strains (<0.0002%; Fig. 3c). In contrast, Δbar1/Δste3 a strains were competent for both autocrine signalling and unisexual mating (1.2%). The potential for unisexual mating is therefore dependent on the ability of cells to undergo autocrine pheromone signalling. To examine the efficiency of inbreeding and outbreeding by different C. albicans strains, SC5314 Δbar1 a strains were also mated with different clinical isolates (Fig. 4a). Again, Δbar1 strains mated with both a and α isolates at a similar frequency (average of 9.8% and 17.6%, respectively). In fact, mating was generally more efficient between SC5314 and clinical isolates (up to 26% a–a mating) than that between SC5314 strains alone. Mating zygotes could be detected in unisexual crosses and fluorescent labelling of the nuclei confirmed that karyogamy had taken place (Fig. 4b and Supplementary Fig. 5). The tetraploid products of unisexual mating were generally stable, but could be induced to undergo chromosome loss to produce diploid cells, thereby completing a parasexual mating cycle (Supplementary Fig. 7).

Figure 4. Homothallic mating across clades in C. albicans.

a, Mating frequencies of an SC5314 Δbar1 a cell crossed with naturally occurring isolates from different clades of C. albicans. 1–8 are crosses between Δbar1 a cells and clinical α cells. 9–18 are crosses between Δbar1 a cells and clinical a cells. For a complete list of strains see Methods. b, Mating zygote formed between a Δbar1 a cell derived from SC5314 (DSY906, expressing HTB–RFP; histone H2B gene–red fluorescent protein) and a clinical a cell derived from IHEM16614 (DSY908, expressing HTB–YFP). Arrows indicate fused nuclei expressing both fluorescent proteins. DIC (left), YFP fluorescence (observed using the GFP channel, middle) and RFP fluorescence (right) images are shown. Scale bar, 15 µm.

Our findings have implications for sexual reproduction in C. albicans and related species, as they reveal a mechanism for same-sex mating in hemiascomycetes. In C. albicans, heterothallic mating between a and α cells can occur, yet a predominantly clonal population structure suggests that outbreeding is rare¹⁹. The discovery of self-mating indicates that homothallism could act to create a burst of genetic diversity by increasing recombination within unisexual populations. Completion of the parasexual cycle could promote allele shuffling or increase karyotypic variation within the diploid genome. Furthermore, isolates of some related species are thought to be exclusively homothallic. For example, isolates of Lodderomyces elongisporus are able to form asci from identical cells, yet the genome apparently encodes only one pheromone gene (MFα) and one receptor gene (STE2)²⁰. Discovery of autocrine signalling and self-mating reveals a mechanism by which such species could reproduce sexually despite the absence of mating partners of the opposite sex.

Homothallism has been well characterized in S. cerevisiae where it is a direct result of mating-type switching. C. albicans and related species are unable to undergo mating-type switching, yet we demonstrate two potential pathways by which self-mating can occur. First, inhibition of Bar1 activity promotes autocrine signalling and mating within unisexual populations of C. albicans. We hypothesize that certain niches in the mammalian host will favour self-fertilization by downregulating Bar1 activity. Second, we show that same-sex mating can occur because of pheromone signalling in conventional a–α mixes. In this case, pheromone secretion by even a limiting number of α cells can promote a–a mating (Supplementary Table 3). In addition, certain combinations of C. albicans a and α strains are incompatible for heterothallic mating as nuclei fail to undergo karyogamy²¹. Our results show that pheromones produced during contact between two such ‘infertile’ strains could still result in self-mating within sexes.

The program of same-sex mating in C. albicans shows unexpected similarities to that in other human fungal pathogens. In the basidio-mycetous fungus Cryptococcus neoformans mating between a and α isolates can occur, yet clinical and environmental isolates are over-whelmingly of the α mating type. Same-sex mating between α isolates of C. neoformans has been observed both in the laboratory and in the wild, and is thought to be the predominant mode of sexual reproduction in this organism^22–25. Pneumocystis carinii is another fungal pathogen predicted to undergo a modified sexual cycle involving cells of the same mating type²⁶. It thus seems that autocrine pheromone signalling and homothallism may be traits shared by a diverse group of medically relevant fungi. These pathogens have retained the ability to undergo sexual reproduction and yet have evolved modifications to these programs that support their survival and propagation. Elucidating the respective mechanisms of homothallism and heterothallism will therefore be critical to the understanding of these species as human pathogens.

METHODS SUMMARY

C. albicans strains were derived from SC5314 except where otherwise noted. Mutant strains were created using either the PCR fusion or SAT1 flipper methods described previously^27,28. For a complete list of strains and primers see Supplementary Tables 1 and 2. Media were made as described previously¹⁰. All C. albicans cells used in this study were in the opaque phase as this is the mating-competent form of the organism⁶. Quantitative PCR was performed on a cells that had been treated with 10 µg ml⁻¹ α-pheromone for 4 h in Spider medium, as described previously^10,29. For analysis of colony morphology, patches of C. albicans cells (~0.5 cm²) were spaced 0.75 cm apart on YPD medium and imaged after 4 days growth at 30 °C using a Zeiss Stemi 2000-C stereoscope with an Infinity 1 camera. Cells were examined using a Zeiss Axioplan 2 microscope equipped with a Hamamatsu-ORCA camera. Mating frequencies were determined by mixing 2 × 10⁷ cells of differently marked auxotrophic strains on Spider medium and incubating at room temperature for 3 days. Cells were then resuspended in water and plated onto selective media to determine frequency of mating products¹⁰. For analysis of a–a mating in the presence of α cells, 2 × 10⁷ cells of each of three strains were mixed and incubated on Spider medium (two differently marked a strains and an α strain negative for both markers). Cells were recovered and analysed as above. The same strategy (two differently marked α strains and an a strain negative for both markers) was applied to detect α–α mating. Flow cytometry to determine cell ploidy was performed by staining DNA with Sytox Green and analysis on a FACSCalibur⁴. PCR to determine mating type was performed using primers directed at OBPa/α as described previously³⁰.