Evolution of yeast mating systems: all roads lead to selfing

Fungi sexually reproduce in heterothallic outcrossing and homothallic selfing modes, and transitions between the two are common throughout the fungal tree of life. A new study in this issue of Current Biology by Krassowski et al [1] reports from the palette of 332 whole genomes for yeast species that the transition from heterothallism to homothallism is common and has repeatedly punctuated the evolutionary trajectory across one of the major lineages of the fungal kingdom. This provides further evidence for pervasive evolutionary forces that promote transitions to selfing modes of sexual reproduction.

Sex is conserved and pervasive throughout eukaryotes. The elaborate and costly process of sexual reproduction typically involves cell fusion (plasmogamy), nuclear fusion (karyogamy), and reductive division (meiosis). Although the tool kit for meiosis is highly conserved in major eukaryotic lineages [2], the mechanisms of sex determination that dictate plasmogamy and karyogamy are highly plastic and diverse. In fungi, sex determination systems are classified into heterothallism and homothallism. Homothallic fungi are self-fertile: they can sexually reproduce in a culture derived from a single spore or cell. Heterothallic fungi are self-sterile: they require another compatible individual for sexual reproduction. It is widely accepted that the last eukaryotic common ancestor (LECA) was sexual. Whether LECA was homothallic or heterothallic, however, is still a matter of debate and experimentation. Transitions between heterothallism and homothallism are observed, even among closely related species. For evolutionary biologists, deciphering the ancestral state of sex-determination and the evolutionary forces that drive these transitions have become a Holy Grail.

Transitions between heterothallism and homothallism are a choice between outcrossing and inbreeding, and these transitions are responsive to selective pressures that favor one or the other breeding strategy. By analyzing genome sequences of 332 yeast species, Krassowski and colleagues demonstrate a surprising favoritism of homothallism over heterothallism among these fungi, despite the more challenging mechanisms to transition to homothallism through mating-type switching [1]

The mating type locus (MAT) defines the cell type in fungi, analogous to sex chromosomes that define sexuality of multicellular eukaryotes. In ascomycete yeasts, individuals of heterothallic species harbor only one of two stable mating types. Homothallic species harbor both mating types in a single genome (either fused or in different genomic locations) [3], or they switch between mating types (Figure 1). Mating-type switching gives rise to compatible cells within the same colony derived from a single cell, which effectively renders the species homothallic. The budding yeast Saccharomyces cerevisiae and the fission yeast Schizosaccharomyces pombe independently evolved mating-type switching systems that involve three cassettes (Figure 1). Such three-cassette systems have served as paradigms to understand other switching systems.

Figure 1. Mating systems in yeast species.

The ancestral mating system in yeast species is heterothallism, with MATa and MATα defined by alternative alleles at the MAT locus. Homothallism, including both primary and secondary homothallism, can evolve from heterothallism through chromosomal rearrangements. For primary homothallism, the two MAT alleles can be either physically linked or located on different chromosomes. For secondary homothallism, one of the two MAT alleles in the genome is silenced and the cell can undergo mating-type switching through either flip/flop or three-cassette mechanism. The transition from heterothallism to homothallism occurred at much higher frequency compared to the transition from homothallism to heterothallism in yeast species. It is likely that transition between the two modes of homothallism also occur, although the frequencies of such transitions are not clear.

The cassette model for mating-type switching in S. cerevisiae was first proposed by Hicks and his colleagues [4]. This budding yeast harbors an active MAT locus and two silent MAT cassettes, HML and HMR (Figure 1). The active MAT locus determines the cell type while HML and HMR loci are transcriptionally silent due to heterochromatin invoked by action of the histone deacetylase Sir protein complex [5]. Having only one mating type expressed is necessary for haploid strains to exhibit a single defined mating type competent for mating. A DNA double-strand break at the active MAT locus inflicted by the Ho endonuclease provokes a gene conversion event, generating a uni-directional mode of switching from HML or HMR to MAT [6]. The mating-type switching system in S. pombe shares similar features with the three-cassette system of S. cerevisiae. However, the enzymes that incite mating-type switching and the precise nature of silencing differs from the budding yeast and the two systems evolved independently.

A simpler version of mating-type switching, namely the flip/flop system, was identified in two Ascomycetous yeast species, Pichia pastoris and Hansenula polymorpha [7, 8]. In this case, only two physically linked mating-type cassettes are present in the genome and one cassette is silenced by virtue of its location in a heterochromatic chromosomal region (Figure 1). In H. polymorpha, the silenced cassette lies next to a centromere, while in P. pastoris, it is located close to a sub-telomeric region. Mating-type switching is achieved by inversion of the chromosomal region encompassing the two MAT cassettes, which is facilitated by repeat sequences flanking the MAT cassettes. Inversion switches the physical locations of the two MAT cassettes, thus activating the alternate MAT cassette. Interestingly, this model has been first proposed many years ago as an alternative to the three cassette switching model [9].

By analyzing the genomes of 332 yeast species, Krassowski et al. revealed extensive diversity in the configuration of the mating-type locus, even among closely related species [1]. They deduced the modes of sexual reproduction based on the organization of the MAT loci, and found that secondary homothallism, primarily by the flip/flop system, repeatedly and independently evolved from heterothallic ancestors. The dynamic nature of the MAT locus organization, and that transitions from heterothallism to homothallism dominate, suggest an evolutionary advantage driving homothallism.

Specifically, of the 332 yeast species analyzed, 192 harbor only one MAT locus in the genome that is indicative of heterothallism. 138 species harbor both the MATa and MATα alleles in the genome, indicative of homothallism. Interestingly, of the twelve lineages analyzed in this study, the classic three-cassette mating-type switching system was found only in budding yeasts of the family Saccharomycetaceae, pointing to a single evolutionary origin of this reproductive mode. On the other hand, the flip/flop switching mechanism was found in 31 species that belong to 4 different lineages, consistent with the hypothesis that the simpler flip/flop system might be the ancestral stage of the three-cassette switching system [10]. Additionally, 41 species have both MATa and MATα alleles without an apparent mechanism for mating-type switching, suggesting that these species might be primary homothallic. Finally, there are 2 species for which no MAT genes were detected in their genomes. These findings paint a picture of highly dynamic MAT locus configurations in budding yeast species, and highlight the power of whole genome sequencing and comparative genomic analyses in revealing the surprising diversity hidden among natural isolates.

How transitions between MAT configurations occurred is not clear. It is hypothesized that the homothallic species evolved via a diploid stage in which both MAT alleles co-existed in the genome, followed by ectopic recombination that established physical linkage between the two MAT alleles [10]. This could also be achieved by homologous recombination in the chromosomal region encompassing the MAT locus, if it is inverted between the opposite mating types. Not surprisingly, transitions from heterothallism to homothallism often occurred via introgression of MAT genes of the opposite mating type at a sub-telomeric or rDNA locus, consistent with previous studies in which these regions, as well as centromeres, are hotspots for chromosomal rearrangements [11]. The discovery of this extraordinary diversity in the configuration of the mating-type locus among yeast species is a necessary and important first step. Functionally characterizing the mating systems in these species is the next challenge. For example, which endonucleases initiate mating-type switching in these homothallic species? Is there evidence of novel domesticated transposable elements functioning as initiators of mating-type switching and promoting sexual reproduction, as shown in some yeast species [12–15]? What are the epigenetic mechanisms that silence one of the MAT loci? For the species that transitioned from heterothallism to homothallism, are there other regions of the genome that also underwent extensive chromosomal rearrangements? Are there conserved features in their centromeres?

One striking discovery by Krassowski et a. [1] is that transitions between heterothallism and homothallism among these yeast species are biased. Specifically, compared to 3 transitions from homothallism to heterothallism, the transition in the opposite direction occurred at least 31 times. Of these evolved homothallic fungi, at least 11 were due to independent transitions from ancestral heterothallism to homothallism by mating-type switching. Thus, there must be a significant selective advantage for homothallism in yeast species. The strong favoritism of homothallism could reflect a selection favoring sexual reproduction. For heterothallic species, the maximum mating opportunity is 50% when the two mating types are in equilibrium. The opportunities for an encounter with an opposite sex are even lower when the distribution of the mating types is uneven. By contrast, the mating opportunity can be 100% for homothallic species. In addition to breaking down genetic linkage, generating genetic diversity, and facilitating more efficient natural selection, for yeast species, mating can lead to diploid cells or spores, both of which vary in physiology compared to haploid cells and could be advantageous under certain environments.

Interestingly, even for heterothallic fungal species, homothallism can be maintained or evolved. For example, the human fungal pathogen Cryptococcus species complex has a bipolar mating system. Each cell inherits a stable mating type defined by a single MAT locus, and there is no mating-type switching. Nonetheless, multiple species from this species complex are capable of unisexual reproduction without the need for a compatible mating partner [16, 17]. It has been proposed that this mode of sexual reproduction might have evolved in response to and/or given rise to the sharply distorted distribution of mating types in the natural population (MATα:MATa >90%) with limited mating opportunities. Thus, the biased transitions from heterothallism to homothallism observed in yeast species could be due to a general favoritism of inbreeding forms of sexual reproduction that increases the opportunities for successful mating in many fungi. These general principles may also apply to other eukaryotic microbes, such as oomycetes (devastating pathogens of plants) in which transitions between heterothallism and homothallism are frequently observed [18]. Additionally, recent studies have revealed a transition from heterothallism to homothallism in Pneumocystis sp., obligate animal pulmonary pathogens that are related to fission yeasts, in which the P and M mating-type loci have been fused and genes involved in sexual reproduction are expressed during infection [19, 20]. This provides further evidence that these unusual, obligate pathogens of animals undergo their sexual cycle in the lungs of infected animals. Thus, the repeated and independently emergence of homothallism may reflect the origins of sex itself and could suggest that LECA underwent homothallic or unisexual reproduction. If so, then the repeated emergence of homothallism may reflect a return to an ancestral mode of reproduction.