Genomic evidence for a complete sexual cycle in Candida albicans

Abstract

Candida albicans is a diploid fungus that has become a medically important opportunistic pathogen in immunocompromised individuals. We have sequenced the C. albicans genome to 10.4-fold coverage and performed a comparative genomic analysis between C. albicans and Saccharomyces cerevisiae with the objective of assessing whether Candida possesses a genetic repertoire that could support a complete sexual cycle. Analyzing over 500 genes important for sexual differentiation in S. cerevisiae, we find many homologues of genes that are implicated in the initiation of meiosis, chromosome recombination, and the formation of synaptonemal complexes. However, others are striking in their absence. C. albicans seems to have homologues of all of the elements of a functional pheromone response pathway involved in mating in S. cerevisiae but lacks many homologues of S. cerevisiae genes for meiosis. Other meiotic gene homologues in organisms ranging from filamentous fungi to Drosophila melanogaster and Caenorhabditis elegans were also found in the C. albicans genome, suggesting potential alternative mechanisms of genetic exchange.

Meiosis represents a specialized cell division that is essential for sexual reproduction; it generates haploid germ cells from diploid parental cells (1). Because a sexual phase for Candida albicans has historically not been detected, it is classified among the Fungi imperfecti (2). However, the identification of a mating-type-like (MTL) locus and genes such as CPH1, CAG1, DLH1, NDT80, and HST6 in C. albicans (see C. albicans genome project information at http://alces.med.umn.edu/Candida.html and http://www-sequence.stanford.edu/group/candida), which participate in meiotic differentiation in S. cerevisiae, suggests that the classification of this diploid fungus belies the existence of a sexual cycle. Recently, in fact, genetic manipulation of the MTL locus resulted in the demonstration that C. albicans strains can mate to produce triploid or tetraploid progeny at very low frequency either in culture or in experimental animals (3, 4). It thus appears that Candida can undergo cell fusion, depending on mating type. However, completion of a sexual cycle, i.e., meiosis and sporulation, remains to be demonstrated.

C. albicans is an opportunistic pathogen that can cause disease in patients immunocompromised as a result of HIV infection, organ transplantation, and cancer chemotherapy (5). It is also a morphologically complex organism capable of proliferating either as a budding yeast or by the formation of pseudohyphae or filamentous hyphae. The inability to demonstrate a sexual cycle has significantly impeded conventional genetic analysis. Therefore the potential for its existence has both intrinsic and technical consequences.

Shotgun sequencing of the diploid C. albicans genome, undertaken by the Stanford Genome Technology Center, is complete with the sequencing of 10.4 haploid genome equivalents, which is sufficient to ensure identification of all of the genes in this organism. A web page and database have been made available over the World Wide Web (http://sequence-www.stanford.edu/group/candida). Given the evolutionary proximity between C. albicans and S. cerevisiae (6) and the differences in their virulence and habitat, genomic comparisons between these fungi are likely to illuminate aspects of the unique cell biology of both organisms. In this report, we address the potential for meiosis and sexual recombination in C. albicans.

Materials and Methods

Sequencing Library Construction, Shotgun Sequencing, Assembly, and Analysis.

C. albicans strain SC5314 was generously provided by Bristol-Meyers Squibb for use without restriction. DNA from this strain was used for all M13 and plasmid sequencing library constructions. Electrocompetent Escherichia coli DH12S and DH10B (GIBCO/BRL Life Technologies) were used for transformation of M13 and plasmid libraries, respectively. C. albicans cultures were grown in yeast/peptone/dextrose broth with shaking at 30°C. Spheroplasts were made, treated with SDS/proteinase K, and the DNA purified on a sucrose gradient. The purified C. albicans SC5314 genomic DNA was sheared by a point-sink shearing device (7) to a fragment size of 3–6 kb for cloning into the plasmid vector pUC119 and to 1–2 kb for cloning into M13mp18. Individual plasmid and M13 clones were picked and grown and template DNA prepared by using the automated instrumentation developed at Stanford University (8). Sequencing reactions were performed on plasmid and M13 template DNA by using BigDye Primer and BigDye Terminator kits from Perkin–Elmer Applied Biosystems according to the manufacturer's specifications, with slight modifications. Sequencing analysis was performed on ABI377-XL Automated Sequencers at 96 lanes/gel. Sequence data were derived by using the basecaller phred (9, 10), assembled with phrap (Phil Green, University of Washington), viewed with consed (11).

Identification of Meiotic Machinery in C. albicans.

We have used several approaches to identify meiotic homologues in C. albicans, focusing mainly on genes critical for mating and meiosis in budding yeast. We also searched for homologues of genes that are regulators of cellular reproduction in other organisms. The S. cerevisiae and other sequences were used as query sequences in basic blast searches (12); by using the blastp program, we searched for similar proteins in the C. albicans genome database Ver. 6 assembly. The blast output was sorted and top hits ranked by blast scores. The E-value cutoff used to assign homologues was 1e⁻⁶. However, each sequence required specific evaluation as there were exceptional instances where biological data indicated the presence of functional homologues, although these lacked significant sequence homology. In Table 1 (which is published as supplemental data on the PNAS web site, www.pnas.org) the functional groups used in the categories roughly correspond to: (i) mating differentiation; (ii) nutritional control; (iii) cell type control; (iv) initiation of meiosis; (v) checkpoint control and progression through meiosis; (vi) recombination and the formation of synaptonemal complexes; and (vii) spore wall morphogenesis and ascus formation.

Results and Discussion

Using comprehensive genomic comparisons, we have assessed the repertoire of gene homologues in C. albicans that in S. cerevisiae are required in pathways leading to sexual differentiation. In a genome-wide transcription analysis of sporulation in S. cerevisiae (13), nearly 500 genes were expressed; those of particular interest are genes that are either meiosis-specific or have shown meiotic-mutant phenotypes. By using these 500 genes as a reference point, C. albicans homologues were identified on the basis of their sequence similarity with S. cerevisiae counterparts from 10.4× sequencing data (see Materials and Methods). Genes analyzed in these comparisons are listed in Table 1 (www.pnas.org) and categorized on the basis of their presumptive function in meiosis and sporulation. In this study, homologues of S. cerevisiae genes involved in chromosome recombination and the formation of synaptonemal complexes (SC) were conspicuously absent from C. albicans, although other groups of genes important for meiosis, mating, and sporulation contained many candidates in the C. albicans genome.

Mating Differentiation.

In S. cerevisiae, mating between haploid cells is signaled by binding of pheromones to a cell-type-specific receptor on cells of the opposite mating type (STE2 expressed in a cells recognized by α-factor, and STE3 expressed in α cells recognized by a-factor). The signal is transmitted by interaction of a heterotrimeric G protein complex composed of Gα(Gpa1), Gβ(Ste4), and Gγ(Ste18) through a downstream mitogen-activated protein (MAP) kinase cascade encoded by STE20, STE11, STE7, and FUS3. The resulting activation of the transcription factor Ste12p is required for expression of mating type-specific genes, cell cycle arrest, fusion of mating partners, and karyogamy (14). In S. cerevisiae, pseudohyphal growth and invasive growth, respectively, are initiated in diploid and haploid cells on nutrient deprivation and signaled by genes shared with this MAP kinase pathway (15). As indicated in Table 1 (www.pnas.org), C. albicans homologues for all of these genes in the MAP kinase cascade have been identified, as well as homologues of the pheromone receptor genes STE2 and STE3. Genes involved in pheromone processing in S. cerevisiae such as STE14, AXL1, STE23, RAM1, RAM2, STE24, RCE1, KEX2, KEX1, and STE13 (16) have homologues in the Candida genome, although there is no independent evidence that C. albicans can produce or respond to pheromones. We also have identified several pheromone-induced genes such as FIG1, FIG3, and FIG4 (Table 1, www.pnas.org), which appear to be important for different steps of mating cell differentiation in S. cerevisiae (17). Together, these data suggest that C. albicans may have preserved the ability to produce and respond to mating pheromones. Although extensive blast analysis failed to identify any mating-factor homologues, computer programs that take into account pheromone gene structure have provided us with several candidates for pheromone genes (unpublished results).

In S. cerevisiae, α-agglutinin (SAG1) provides tight cell–cell adhesion during mating. It has been postulated that in C. albicans, the adhesiveness of the homologous ALS gene family contributes to its pathogenesis (18). The presence of FUS1 might suggest the ability of cell fusion for C. albicans. A KAR1-like sequence, which in S. cerevisiae is critical for nuclear fusion and spindle pole body formation (19), was not found in C. albicans.

One of the major outcomes of CPH1/STE12 activation through the MAP kinase pathway in C. albicans is to induce hyphal morphogenesis (20). Virulence of cst20/ste20 and cph1/ste12 disruption mutants is attenuated in the mouse model of systemic candidasis (21, 22), thereby establishing a potential link between pheromone signaling, filamentous growth, and virulence, as found in the pathogenic fungi Cryptococcus neoformans and Ustilago maydis (23, 24). In the evolution of pathogenesis, perhaps C. albicans has used the MAP kinase pathway to strictly control expression of its hyphal phenotype in response to changes in the host environment. The identification of potential ligands and possible environmental cues that are either recognized by the homologues of STE2- and STE3-like receptors or that stimulate the MAP kinase pathway through alternate receptors could help us understand hyphal induction and pathogenesis in C. albicans.

Nutritional Control.

In S. cerevisiae, meiosis is initiated only by diploid cells deprived of glucose and nitrogen and grown in the presence of a nonfermentable carbon source, whereas other fungi have different and complex nutrient requirements for this process. For instance, nitrogen starvation is required for mating and meiosis in Schizosaccharomyces pombe (25). The plant pathogen U. maydis enters meiosis only during growth in its host, Zea mays (26). C. albicans appears to contain homologues of S. cerevisiae genes involved in glucose repression and nitrogen metabolism: MIG1, GAT1, and UME6, as well as RAS/cAMP, SNF1, and MCK1, which are involved in the nutrient sensing pathway (27).

The complexity and cross talk between nutritional and meiotic pathways suggest that, although similar genes may be present in both organisms, their participation in these pathways may have different biological consequences. For example, SNF1 is essential for the viability of C. albicans (28) but is not essential in S. cerevisiae, where it coordinates glucose and acetate regulation of the early and late meiotic program (29). Another homologue found in C. albicans, MCK1, encodes a serine-threonine-tyrosine kinase, which functions as a positive regulator of meiotic gene expression in S. cerevisiae and is essential for ascus maturation; it governs centromere behavior in mitosis (27). Comparative genomic analyses also reveal the metabolic diversity of C. albicans (S.S., unpublished observations) and suggest that the conditions traditionally used in the laboratory for its culture may in part be responsible for the failure to detect a sexual cycle in this organism.

Cell Type Control.

In S. cerevisiae, only diploid cells that are heterozygous at the MAT locus can initiate meiosis and sporulation on nutritional starvation. An important feature of a/α cells is the presence of a transcriptional repressor, a1-α2, which is a heterodimeric homeodomain protein (30). RME1, which encodes a negative regulator of meiosis, is one of the genes turned off by a1-α2 (31). The product of another gene, IME4, mediates both cell type and nutritional activation of IME1 (see below) (32). No homologue of IME1 was found in C. albicans, although putative counterparts of RME1 and IME4 have been identified.

Mating type loci in yeasts are master regulators of cell fate specification and sexual morphogenesis (33). In C. neoformans, there is an association between mating type, hyphal phase, and infectivity (34). Recently, a mating-type-like (MTL) locus in C. albicans with homology to the MAT locus of S. cerevisiae was identified (35). The C. albicans MTL locus is large (approximately 8.8 kb), single copy, and without silent cassettes such as HML or HMR. Whether there is a correspondence between mating type and virulence in C. albicans has not been established. However, most clinical isolates of C. albicans are heterozygous at MTL locus (P. Magee, personal communication) and by analogy would be equivalent to a/α cells in S. cerevisiae.

Initiation of Meiosis.

Initiation of meiosis and sporulation in diploid S. cerevisiae by nutritional limitation occurs through a transcriptional cascade with sequentially expressed distinct classes of meiosis-specific genes (36). UME6 encodes a Zn₂Cys₆ DNA-binding protein that functions as a developmental switch for mitotic repression and meiotic activation of early meiotic genes (37). Interaction between Ume6p and Ime1p, a transcriptional activator, is required for induction of early meiotic gene expression. In S. cerevisiae, RIM11 and RIM15 kinases are required for Ume6p-Ime1p interaction (38). RIM101 defines a signaling pathway that activates IME1. As shown in Table 1 (www.pnas.org), homologues of UME6, RIM11, RIM15, and genes in the RIM101 pathway have been identified in C. albicans. Experiments have shown that the Candida RIM101 homologue participates in hyphal growth (39). In S. cerevisiae, Ime1p activates expression of IME2, a serine–threonine kinase, essential for premeiotic DNA replication (40). Despite extensive blast analysis, we have not found counterparts of IME1 in C. albicans. Similarly, efforts to functionally complement an ime1 null mutant of S. cerevisiae with a C. albicans genomic library have failed (41), lending further support to the apparent absence of this gene in C. albicans. However, we have found downstream targets of Ime1p such as IME2. In S. cerevisiae, the Ime1p/Ume6p complex and the URS1 consensus-binding site for early meiotic genes play a pivotal role in the cell's decision to enter meiosis. The absence of IME1 suggests that the switch machinery used in S. cerevisiae to effect commitment to the meiotic pathway is missing. It remains possible that a functional analogue of IME1 is present in C. albicans.

Checkpoint Control and Progression Through Meiosis.

Mitosis and meiosis, the two fundamental modes of cellular reproduction, have overlapping functions including DNA replication and chromosome segregation, as well as similar but distinct mechanisms to survey the progression of cell cycle events. We have identified C. albicans homologues of the mitotic DNA damage checkpoint genes MEC1, RAD17, RAD24, which are also required for meiotic progression (42), as well as meiosis-specific MEK1 and the chromatin-silencing factors SIR2 and DOT1, which are implicated in pachytene checkpoint control (43). We have also identified a set of gene homologues whose products in S. cerevisiae participate in DNA replication and chromosome segregation: CDC5, CDC7, CDC14, CDC20, CDC25, CDC28, IPL1, CDH1 (44), several components of the anaphase-promoting complex (APC1, APC2, APC11, CDC16, CDC23, CDC27), and its meiotic activator AMA1 (SPO70) (45). Overall, homologues for most of the genes that in S. cerevisiae participate both in mitosis and meiosis have been identified in C. albicans.

Meiosis I (MI) is a reductional division, whereas meiosis II, like mitosis, is equational. The separation of sister chromatids must be tightly regulated during meiosis; homologues of PDS5, SCC2, SMC1, SMC2, SMC3, which in S. cerevisiae function in sister chromatid cohesion, were also found in C. albicans (46). A homologue of YPR007 (REC8), proposed to encode a protein mediating meiotic sister chromatid cohesion (47), was identified, as was NDT80, a gene that encodes a meiosis-specific transcription factor. In S. cerevisiae, Ndt80p is required for expression of middle meiotic genes, meiotic division, and spore formation (48). Transcription of NDT80 depends on IME1, which remains unidentified in the C. albicans genomic sequence. Another important gene apparently absent in C. albicans is a homologue of SPO13. spo13 null mutants in S. cerevisiae eliminate the reductional division (MI) during meiosis to produce diploid spores (49). The absence of SPO13 might indicate that the C. albicans sexual cycle is one where cells undergo single-division meiosis.

Recombination and the Formation of Synaptonemal Complexes.

Meiosis is fundamentally different from mitosis in the occurrence of high-frequency recombination. Meiotic recombination, which is thought to proceed through a double-strand-break (DSB) repair mechanism (50), follows premeiotic DNA replication. More than 10 genes are required to produce DSBs in meiotic cells in S. cerevisiae; strikingly, several of these are absent from C. albicans: MER1, MER2, REC102, REC104, REC114, and MEI4. The absence of homologues of approximately half of the genes required for the initiation of recombination in S. cerevisiae suggests that C. albicans may be greatly impaired in this initial step. However, we have identified SPO11, a type II topoisomerase implicated in nicking DNA, to generate DSBs (51). Homologues of SPO11 have also been identified in Caenorhabditis elegans, Drosophila melanogaster, S. pombe, mouse, and human (52).

Homologues of most of the genes involved in strand invasion and Holliday junction formation have been identified as well: RAD51, RAD52, RAD54, RAD57, and DMC1 (53). The C. albicans DLH1 gene complements DMC1 null mutants, its meiosis-specific homologue in S. cerevisiae (54). Homologues of the meiosis-specific genes SAE2 and SAE3 are also absent from C. albicans, whereas those of genes that participate in both mitotic and meiotic mismatch repair MSH2, MSH6, MSH3, MLH1, and PMS1, are present (55).

Many unique characteristics of meiotic recombination are ensured through the formation of synaptonemal complexes (SC) (56). These structures not only exclude recombination between sister chromatids but also control the frequency of crossovers. Homologues of HOP1 and ZIP1, molecular components of the SC, were found in C. albicans, whereas those for other proteins, such as ZIP2 and HOP2, involved in synapsis of chromosomes and formation of SC, are absent (57, 58). Homologues of meiosis-specific genes MSH4 and MSH5, required to promote recombination between homologous chromosomes and Holliday junction resolution, were also identified. Together, these findings suggest that C. albicans is not competent to make mature SC. However, C. albicans may resemble other fungi such as S. pombe and Aspergillus nidulans, which exhibit meiotic recombination but do not form SC (59).

Spore Wall Morphogenesis and Ascus Formation.

In S. cerevisiae, the spore wall is formed de novo through a complex morphogenetic program depending on the initiation of meiosis. Although there are no detectable meiotic ascospores in C. albicans, homologues of genes involved in the sporulation pathway such as SPS1, a STE20-like protein kinase, and SMK1, a MAP kinase, were identified. We also found homologues of a set of mid/late sporulation-specific genes involved in spore wall maturation: DIT1, DIT2, SPR3, and YDR104. DIT1 and DIT2 are important for the formation of the outer layer of the spore wall. YDR104 is essential for spore wall formation in S. cerevisiae (13). The presence of these genes may reflect a possible MAP kinase pathway that can transmit environmental signals to modify cell wall structures in C. albicans. The extent to which homologues of these genes function in yeast or hyphal wall synthesis is not known. The C. albicans cell wall contains variable amounts of dityrosine, and its basic structural features are similar but not identical to those of S. cerevisiae (60).

Genomic Comparison Between C. albicans and Other Organisms.

Given that gametogenesis in metazoans and sporulation in yeasts are highly conserved evolutionary processes, it is not surprising that a survey of available databases revealed several genes that function in meiosis in multicellular organisms and that have homologues in C. albicans. Some of the genes analyzed are listed in Table 2 (which is published as supplemental data on the PNAS web site, www.pnas.org). For example, we have found homologues of pelota and DES-1, which are required for spermatogenesis in Drosophila. DOM34 is the S. cerevisiae homologue of pelota and is important for meiosis, pseudohyphal growth, and translation (61). DES-1, a transmembrane protein, is required for initiation of meiosis in Drosophila spermatogenesis. A DES-1 homologue has been identified in mouse, S. pombe, and Arabidopsis thaliana (62) but not in S. cerevisiae. The Drosophila protein is proposed to mediate the interaction between somatic cells and germ cells during development (63), raising the possibility that DES-1 homologue might be required for communication between C. albicans and its environment.

We have also found several S. pombe gene homologues (in C. albicans) such as RCD1 (YNL288), NRD1 (YPL184), STE20 (YER093), and MEI4 (FKH1), which are involved in sexual differentiation in the fission yeast (25). Their counterparts in S. cerevisiae have been identified, but their functional role in meiosis has not been established. For example, S. cerevisiae FKH1 regulates mitotic cell cycle progression and pseudohyphal growth (64) and is expressed early during sporulation (13).

The apparent dissimilarity in meiotic machinery between S. cerevisiae and C. albicans may reflect their relative phylogenetic distance (6). To investigate this further, we sought potential meiotic homologues in C. albicans that may resemble key meiotic regulators or cognates of genes that function in the sexual cycle in other fungi such as C. neoformans, A. nidulans, Neurospora crassa, and Podospora anserina. From this analysis, we have identified several meiotic homologues, including SPO14, MCK1, MEK1, and a pH regulatory system with components of palI, pacC, palA, palF, and palB shared by C. albicans, S. cerevisiae, and A. nidulans. With respect to pacC, a homologue of RIM101, this gene is conserved in all three of these fungal species, as are other elements of the RIM101 pathway. In S. cerevisiae, the RIM101 pathway functions both in meiosis and invasive growth. In C. albicans it is required for pathogenesis (65).

Asexual sporulation is common among diverse groups of fungi. C. albicans can produce structures known as chlamydospores under special conditions (5); their biological role is unknown. We have identified a set of C. albicans genes that are homologues of those that participate in asexual sporulation in A. nidulans (66). These include C. albicans genes designated SST2 (a homologue of flbA), YPR013 (a homologue of flbC), MYB1 (a homologue of flbD), TEC1 (a homologue of abaA), DOP1 (a homologue of dopA), and EFG1 (a homologue of stuA). Several of these were also found in S. cerevisiae. Characterization of these gene homologues may help to define the role of chlamydospores in C. albicans cell biology.

In considering alternative reproductive pathways, the identification of HET-C and HET-E-1 homologues in C. albicans is particularly interesting. The HET-C gene, which encodes a glycolipid transfer protein, is proposed to function in vegetative incompatibility and ascospore formation in Podospora anserina (67). HET-E-1 encodes a β-transducin-like protein (68). No HET-C gene homologue was found in S. cerevisiae. Homologues of MOD-D and MOD-E, modifiers of the het locus, were also identified in C. albicans. MOD-D is a homologue of Gα subunit, and MOD-E is an HSP90 homologue (69, 70). As the het loci appear to regulate self/nonself recognition during vegetative growth in filamentous fungi, heterokaryon formation could provide an opportunity for genetic exchange in imperfect fungi (71).

Conclusion

C. albicans can produce tetraploid progeny in animals and in culture when parental types homozygous at MTL locus are artificially created (3, 4). However, there is no evidence of reductional division from these matings, and it remains to be shown that this fungus has a meaningful sexual cycle. Most clinical isolates of Candida are heterozygous at the MTL locus, and most proliferation of naturally occurring populations appears to be clonal (72), suggesting that mating is infrequent. Our comprehensive comparison and analysis of the C. albicans genome, focusing on the question of sexuality in C. albicans, suggests that, whereas this organism suffers a multiple gene defect in a S. cerevisiae-like meiotic machinery, it possesses a repertoire of genes composed of homologues found in sexual pathways of S. cerevisiae, filamentous fungi, and metazoans, which in aggregate indicate that C. albicans has a sexual cycle in nature.

One of the more interesting results of our analysis is the finding that, whereas C. albicans appears to lack a homologue of IME1, in S. cerevisiae the master switch for entry into the meiotic pathway, it does possess a number of homologues of the downstream target genes of the Ime1p/Ume6p regulatory complex in S. cerevisiae. As IME1 is the integration point of genetic and nutritional signals for meiosis in S. cerevisiae (73), the regulation of initiation of meiosis might be achieved in Candida through an analogue of IME1, which integrates a different set of signals to effect commitment to meiosis. We have not been able to identify IME1 homologues in other organisms despite extensive blast analysis, indicating that perhaps the commitment to meiosis effected by the interaction between Ime1p and Ume6p is unique to S. cerevisiae. Similarly absent in C. albicans and other organisms as well is SPO13, which in S. cerevisiae is essential for proper execution of meiosis I. Within the meiotic pathway itself, C. albicans is missing 6 of 10 homologues of genes that in S. cerevisiae are necessary for the initiation of double-strand breaks and others that are involved in chromosome recombination and the formation of SC. Most of the missing genes are related to recombination and SC formation, although in each of these processes, C. albicans still possesses a number of homologues that may participate at various stages in the progression of meiosis.

The study of fungal developmental signaling pathways has revealed an intimate relationship between mating and filamentous growth (74) and suggests that different cell types can use shared signaling components to specify cell fate. C. albicans is able to proliferate in many forms—as yeast, pseudohyphae, and filamentous hyphae. It is also able to form chlamydospores and express different switch phenotypes in response to a variety of environmental cues. Any one of these forms may be capable of cell fusion, a possibility highlighted by the identification of several components of the het system as well as gene homologues that play crucial roles in cellular reproduction in various organisms. Our studies suggest the possibility for hyphal fusion and dikaryon formation in C. albicans and allude to more complex patterns of both asexual and sexual differentiation similar to those found in some filamentous ascomycetes and basidomycetes (75, 76).

Abbreviations

MTL locus

mating-type-like locus

MAP kinase

mitogen-activated protein kinase

SC

synaptonemal complexes