|Home | About | Journals | Submit | Contact Us | Français|
Meiosis is an integral part of sexual reproduction in eukaryotic species. It performs the dual functions of halving the genetic content in the cell, as well as increasing genetic diversity by promoting recombination between chromosome homologs. Despite extensive studies of meiosis in model yeast, it is now apparent that both the regulation of meiosis and the machinery mediating recombination has significantly diverged, even between closely related species. To highlight this, we discuss new studies on sex in Candida species, a diverse collection of hemiascomycetes that are related to S. cerevisiae and are important human pathogens. These provide new insights into the most conserved, as well as the most plastic, aspects of meiosis, meiotic recombination, and related parasexual processes.
Sexual reproduction is a common attribute of eukaryotic species where it provides an adaptive advantage over species that are strictly asexual. During sex, fusion of gametes results in formation of cells with higher ploidy. Completion of sexual reproduction therefore requires a reductive DNA division, typically mediated by meiosis, in which diploid cells regenerate haploid gametes. Sexual reproduction evolved once very early in evolution in the eukaryotic lineage, and thus all asexual eukaryotes presumably derived from sexual ancestors .
Studies in yeast have played a central role in elucidating the regulation, machinery, and significance of sexual reproduction. In particular, experiments on the model ascomycetes, Saccharomyces cerevisiae and Schizosaccharomyces pombe, have detailed the molecular steps during mating and meiosis. However, despite the wealth of information obtained from these systems, it is now recognized that a surprising plasticity exists in sexual mechanisms from different species. We will review the meiotic programs of S. cerevisiae and S. pombe, and compare meiosis in these organisms with that recently described in Candida species. The latter are a diverse collection of hemiascomyces yeast related to S. cerevisiae, and are the most common cause of opportunistic fungal infections in humans. We emphasize differences in sexual differentiation between model and pathogenic yeast, and predict that additional surprises will emerge as genomic and functional studies are completed on these fungi that have important implications for human health.
S. cerevisiae and S. pombe show distinct meiotic programs, both in terms of their regulation and the physical structures underpinning nuclear division and recombination. Several reviews have discussed the regulation of meiosis [2–6], the role of cohesins [7,8], and the mechanism of meiotic recombination [9–12], and we will therefore provide a simplified overview.
Mating in S. cerevisiae occurs between haploid a and α cells producing a/α diploids. The latter are unable to mate, but are competent to undergo meiosis and sporulation in response to environmental cues. A key transcriptional regulator mediating this cell-type specificity is the a1/α2 heterodimer that represses expression of mating genes [13,14]. In addition, a1/α2 inhibits expression of RME1, which itself encodes a negative regulator of meiosis (Table 1). RME1 plays an important role in preventing haploid a or α cells from initiating meiosis, a potentially catastrophic event . The direct target of RME1 is IME1, which encodes a transcription factor that, in concert with Ume6, regulates downstream meiotic genes  (Figure 1). Multiple pathways converge on IME1 that are sensitive to cell type and nutritional signals. In particular, the IME4 gene is induced under starvation conditions and encodes an N6-methyladenosine activity that activates IME1, and potentially other targets, through modification of their mRNA [17,18]. More recently, it was shown that IME4 expression is also cell type regulated by a mechanism of transcriptional interference. Thus, haploid a or α cells produce an antisense RNA against IME4, whereas in a/α diploids expression of the antisense RNA is blocked by a1/α2, thereby enhancing expression of the sense RNA . This demonstrates a novel mechanism by which cell type specificity in yeast is mediated by a noncoding RNA. Immediately downstream of Ime1 is Ime2, a serine-threonine kinase that positively regulates subsequent events in meiosis [2,20]. Ime2 also negatively regulates Ime1, targeting it for degradation, thereby restricting Ime1 to a narrow window of activity . Profiling of synchronous meioses have revealed a cascade of sequential gene expression involving more than 1000 genes divided between seven temporal groups .
Mating in S. pombe occurs under conditions of nutritional stress and generates diploid h+/h− zygotes. Co-operation of h+ and h− mating-type proteins (Mc and Pm) induces expression of the mei3 gene, initiating entry into meiosis . Mc and Pm therefore provide cell-type specific regulation of meiosis analogous to that of a1/α2 in S. cerevisiae (Figure 1). Downstream of Mei3 there is a key interaction between Pat1, a serine/threonine kinase, and Mei2, an RNA-binding protein. Mei2 is a key factor for induction of meiosis, but is normally inhibited due to phosphorylation by Pat1 . Inhibition is relieved upon expression of Mei3, which binds to Pat1 thereby inactivating it , although a Mei3-independent pathway may also exist for activation of meiosis . Mei2 is essential for the initiation of meiosis and prevents the degradation of key meiosis-specific transcripts . Transcriptional waves subsequently accompany progression through sporulation, and transcription factors have been identified that regulate specific stages including Rep1 (early genes), Mei4 (middle genes), and Atf21/Atf31 (late genes) [32,33]. Comparison of S. pombe and S. cerevisiae transcriptomes reveals relatively little overlap between species, although shared components include those of the anaphase-promoting complex and recombination/cohesin genes, such as REC8 and DMC1 .
The first DNA division in meiosis is reductional, with homologous chromosomes segregating from one another (meiosis I), while the second meiotic division is equational, with separation of sister chromatids (meiosis II). A conserved complex of cohesins holds sister chromatids together during meiosis I, possibly via a ring structure that encircles the chromatids . In many eukaryotes, including S. cerevisiae and S. pombe, Rec8 provides meiosis-specific cohesin activity, and cleavage of this protein is necessary prior to separation of sister chromatids during meiosis II [35,36].
Meiotic recombination also plays an active role in pairing homologous chromosomes in yeast. Central to meiotic recombination is the formation and repair of DNA double-strand breaks (DSBs) introduced by the highly conserved Spo11 protein (Table 2) . DSBs are subsequently processed by the MRX complex (Mre11-Rad50-Xrs2) into single-stranded DNA regions in preparation for strand exchange by the RecA homologs, Rad51 and Dmc1 [9,37]. Dmc1, in particular, is a meiosis-specific activity important for homologous recombination in many, but not all, eukaryotes . Accessory factors act to promote DSB formation and strand exchange, although these factors tend to be poorly conserved or even species-specific [9,39].
Higher order structures mediate meiotic chromosome segregation in both S. cerevisiae and S. pombe. In S. cerevisiae, as in most other eukaryotes, synaptonemal complexes (SC) are formed during meiosis and these proteinaceous structures facilitate pairing of homologous chromosomes . A number of proteins are implicated in SC formation including the ZMM (Zip1-Zip4, Msh4/Msh5 and Mer3) family of proteins [11,40], which play a structural role in the SC (particularly Zip1) while also directing the program of genetic recombination. Thus, ZMM proteins promote the processing of meiotic DSBs into crossover events (where there is an exchange of flanking markers) rather than noncrossover events [41,42]. In contrast, meiosis in S. pombe does not involve SC formation, although minimal structures (linear elements) support pairing of homologous chromosomes. These structures share some components with the SC (Table 2), but are not essential for meiotic recombination [8,43].
Species belonging to the Candida genus were historically classified as budding yeast that were asexual and formed either pseudohyphae or true hyphae . Many of these species are human pathogens and cluster in a single clade within the hemiascomycetes (Figure 2)[45,46]. Sexual lifestyles have now been confirmed for several Candida species and these reveal a surprising diversity in their programs. Thus, C. glabrata, C. parapsilosis, and C. tropicalis have many of the genes necessary for mating and meiosis, and yet sex has never been observed in these species . In contrast, C. lusitaniae and C. guilliermondii have established sexual cycles including sporulation and ascospore formation . C. albicans is somewhere between these two extremes; efficient mating occurs but a conventional meiosis has yet to be demonstrated [47–50]. Molecular studies of sexual differentiation in C. lusitaniae and C. albicans have now been completed, and these will be contrasted with those in model yeast.
C. lusitaniae exhibits a surprisingly efficient program of meiosis and sporulation given that this species is lacking many ‘key’ meiotic components . Missing factors include the recombinase Dmc1, synaptonemal complex proteins Zip1-Zip4, as well as activities that promote crossover formation (Msh4/Msh5 and Mer3). In fact, C. lusitaniae is missing most of the top candidate genes identified as hallmarks of meiosis and sexual reproduction in eukaryotic species (the ‘meiotic toolkit’ genes, Table 2). Despite the loss of these factors, C. lusitaniae undergoes Spo11-mediated meiotic recombination at frequencies similar to that of other sexual fungi . The lack of conserved SC components suggests that meiosis may involve minimal structures for chromosome pairing similar to S. pombe. However, while a homolog of Rec8 is present in C. lusitaniae, other factors such as Hop1 and Mek1 that contribute to linear elements in S. pombe are missing [8,43,45]. It therefore appears that very rudimentary structures mediate chromosome pairing and recombination during C. lusitaniae meiosis.
Sexual differentiation in C. lusitaniae is also unusual in that, unlike the related hemiascomycete S. cerevisiae, it is no longer under control of the a1/α2 complex. C. lusitaniae has retained the a1 gene, and this gene is necessary for meiosis and sporulation, but the α2 gene is missing from the sequenced MTLα locus (Figure 1) [45,51]. Presumably, another α-specific gene is required for determining cell type as only a/α (not a/a or α/α diploids sporulate. One candidate is the α1 gene, although how a1 cooperates with α1 to specify cell type and promote meiosis is an intriguing question for the future. Furthermore, while C. lusitaniae contains homologs of many regulatory factors involved in meiosis (including IME2, IME4, RME1, and NDT80) there is no homolog of IME1, the master regulator in S. cerevisiae (Table 2) [45,51]. The transcriptional circuitry regulating meiosis has therefore diverged since S. cerevisiae and C. lusitaniae last shared a common ancestor. Similar rewiring of the mating type circuitry has also occurred between S. cerevisiae and the Candida species [53–55], again illustrating the highly plastic nature of sexual processes.
Sexual reproduction C. lusitaniae involves a haploid-diploid-haploid cycle similar to that in model yeast. Surprisingly, however, while the products of meiosis are mainly haploid, about one-third of meiotic products are diploid or aneuploid . The latter may be due to inefficient chromosome segregation and meiotic nondisjunction, particularly given the limited repertoire of meiotic components present in the C. lusitaniae genome. Interestingly, similar levels of aneuploidy (7–35%) also occur during human oogenesis and can lead to miscarriage and birth defects , although it is not clear what consequences aneuploidy may have for C. lusitaniae biology. In addition, C. lusitaniae asci typically contain two ascospores (dyads) rather than four ascospores (tetrads) as often formed during meiosis in S. cerevisiae and S. pombe. Chromosome missegregation during C. lusitaniae meiosis could generate many non-viable forms, and reducing the number of spores may help limit investment in dead-end products . A parallel exists with S. cerevisiae, as meiosis performed in limiting acetate conditions produces dyads rather than tetrads, again preserving cell resources while allowing sporulation to proceed [57–59]. In this case, unpackaged nuclei get degraded during ascus maturation leaving two intact, haploid ascospores.
In contrast to C. lusitaniae, which is rarely encountered in the clinic, C. albicans is a prevalent human pathogen that causes both mucosal and systemic infections . Despite its clinical importance, the sexual lifestyle of C. albicans has remained an enigma. Long thought to be an obligate asexual organism, mating of diploid strains has been demonstrated both in vitro and in vivo, and is regulated by a unique mechanism of phenotypic switching [48–50,61]. MTLa and MTLα cells can switch between white and opaque forms, but only cells in the opaque state are competent for mating . The key regulator of the opaque state is Wor1, as high levels of this protein are sufficient to drive cells to the opaque form, although a complex network of positive and negative feedback loops regulates heritable Wor1 expression . White-opaque switching is inhibited in a/α cells as the a1/α2 complex represses expression of the WOR1 gene, locking cells in the white state (Figure 1). The white-opaque switch therefore adds an additional layer of sophistication to C. albicans mating, presumably to regulate sexual reproduction in response to environmental cues from the mammalian host or other, yet to be discovered, niches.
Despite efficient mating between opaque a and α cells, and the presence of many ‘meiosis-specific’ genes in the sequenced genome (Table 2), a conventional meiosis has yet to be uncovered in C. albicans [45,62]. In its place, a parasexual mechanism has been described involving efficient and cooperative chromosome loss . Thus, culture of tetraploids on certain laboratory media causes genomic instability and loss of chromosomes from the tetraploid, resulting in the formation of diploid (and numerous aneuploid) forms of C. albicans [63,64]. One potential advantage of the parasexual program over a traditional meiosis is that it does not culminate in the production of spores. This may be significant given that spores are often highly antigenic and could therefore minimize detection of mating products by the host immune system .
Despite the apparent absence of a conventional meiosis in C. albicans, several meiosis-specific genes have been shown to encode functional products. For example, a homolog of Dmc1 is present in C. albicans and can successfully substitute for S. cerevisiae Dmc1 function in meiosis . In addition, Spo11 is required for the initiation of meiotic recombination in many eukaryotes, and was recently implicated in mediating recombination during the parasexual cycle of C. albicans . It therefore remains to be seen whether other meiosis-specific genes have also been reprogrammed to function in the parasexual program. If so, the parasexual cycle may represent a bona fide alternative to a conventional meiosis. Alternatively, the parasexual cycle may simply signify a particularly inefficient meiosis, especially given the high rates of aneuploidy now reported during meiosis in C. lusitaniae . Finally, an exciting possibility is that a cryptic meiosis still exists for C. albicans, and that the appropriate conditions have not been identified to initiate this program. In this regard, it is interesting to note that a novel mode of same-sex mating has recently been demonstrated for C. albicans . While the products of same-sex mating can undergo the parasexual pathway, further experimentation will determine if a conventional meiosis/sporulation can be performed by these cells.
An emerging theme is that programs of sexual reproduction exhibit a remarkable degree of plasticity. This is particularly true of meiosis, where recent genomic and functional studies have revealed marked differences even between closely related species. Thus, while much has been learned about sexual reproduction from studies in model yeast, it is now apparent that differences between species are likely to be almost as striking as the similarities. A striking example of this sexual diversity is provided by C. lusitaniae, which is able to undergo efficient sporulation and homologous recombination yet is apparently lacking much of the molecular machinery associated with meiosis in other organisms. Additional studies will be necessary to reveal how this is achieved with such a limited repertoire of meiotic components.
Finally, while we have focused on recent experiments in Candida, exciting developments have been made concerning sex and meiosis in other pathogenic fungi. In particular, a complete sexual cycle has now been demonstrated for Aspergillus neoformans , the most prevalent airbourne fungal pathogen, and novel modes of same-sex mating and meiosis have been shown to occur during ‘haploid fruiting’ in Cryptococcus neoformans , a causative agent of meningoencephalitis. As additional genomes are sequenced, the ability to analyze the role of putative mating and meiosis genes will accelerate, and will undoubtedly lead to a clearer picture of the evolution and consequences of sex in fungi and higher eukaryotes.
The authors would like to thank Kevin Alby for discussions and reading of the manuscript. We apologize to authors whose work could not be cited due to space restrictions. RS was supported by an F31 predoctoral fellowship (#F31 AI075607-02) from the NIH. RJB holds an investigator in the Pathogenesis of Infectious Disease Award from the Burroughs Wellcome Fund. This work was also supported by the NIH (R21 AI081560).