Genetic and genomic evolution of sexual reproduction: echoes from LECA to the fungal kingdom

Abstract

Sexual reproduction is vastly diverse and yet highly conserved across the eukaryotic domain. This ubiquity suggests that the last eukaryotic common ancestor (LECA) was sexual. It is hypothesized that several critical processes in sexual reproduction, including cell fusion and meiosis, were acquired during the evolution from the first eukaryotic common ancestor (FECA) to the sexual LECA. However, it is challenging to delineate the exact origin and evolution of sexual reproduction given that both FECA and LECA are extinct. Studies of diverse eukaryotes have helped to shed light on this sexual evolutionary trajectory, revealing that a primordial sexual ploidy cycle likely involved endoreplication followed by concerted chromosome loss and that cell-cell fusion, meiosis, and sex determination later arose to shape modern sexual reproduction. Despite the general conservation of sexual reproduction processes throughout eukaryotes, modern sexual cycles are immensely diverse and complex. This diversity and complexity has become readily apparent in the fungal kingdom with the recent rapid expansion of whole-genome sequencing. This abundance of data, the variety of genetic tools available to manipulate and characterize fungi, and the thorough characterization of many fungal sexual cycles make the fungal kingdom an excellent forum in which to study the conservation and diversification of sexual reproduction.

Introduction

Sexual reproduction is ubiquitous in eukaryotes and responsible for driving genetic diversity through a conserved ploidy reduction process known as meiosis. Though the origins of sexual reproduction remain unknown, there is a wide consensus that the last eukaryotic common ancestor (LECA) was sexual [1]. The sexual LECA is thought to have evolved from the first eukaryotic common ancestor (FECA), of archaeal origin, through massive gene gain and cellular membrane compartmentalization via both bacterial endosymbiosis and viral infection [2–4]. The earliest form of sex likely involved a ploidy increase via endoreplication followed by a rudimentary parasexual cycle with concerted chromosome loss to return to the haploid state. Subsequently, three major innovations – cell-cell fusion, meiosis, and the evolution of distinct sexes – yielded the modern versions of sexual reproduction (Figure 1) [5]. Modern sexual reproduction exists in myriad forms, defined by unique sexual identities and mating processes, but unified by the admixing of parental genomes and generation of recombinant offspring [1].

Figure 1. Evolution of sexual reproduction.

Evolution of sexual reproduction from the asexual first eukaryotic common ancestor (FECA) to the sexual last eukaryotic common ancestor (LECA) includes three major innovations: cell-cell fusion, meiosis, and sex. Depending on when cell fusion and sex were acquired, LECA can be either sexual (model A) or unisexual (model B). Evolution of fusogens, meiotic cycles, meiotic drivers, karyotypes, genomes, and epigenetic regulations drive the diversification of modern sexual reproduction in eukaryotes.

It is difficult to determine the complex evolutionary trajectory that sexual reproduction has undertaken due to the extinction of both FECA and LECA. It is therefore unclear how LECA achieved increased ploidy, how genetic circuits evolved to orchestrate diverse meiotic processes, and how species boundaries formed. Next-generation sequencing has enabled the expansion of genomic archives and helped to improve the genetic toolkits applied to investigate these evolutionary conundrums. In this review, we summarize current progress towards understanding the evolution of sexual reproduction and illustrate opportunities in the fungal kingdom to address evolutionary ambiguities.

From one to two

It is hypothesized that sexual reproduction first evolved as a form of selfing involving two identical genomes [1]. There are three hypothetical mechanisms that LECA could have utilized to increase ploidy: endoreplication, mitotic division without cytokinesis followed by nuclear fusion, or cell-cell fusion followed by nuclear fusion. Phylogenetic studies on cell fusion and karyogamy proteins favor the hypothesis that cell fusion and nuclear fusion were ancestral traits and therefore present in LECA [6]. Two recent studies characterized a class II viral fusogen HAP2 and found that it was highly conserved across eukaryotes, which suggests an early evolutionary acquisition of a viral protein by LECA [7,8]•. HAP2 membrane insertion motifs have further diversified to drive species-specific gamete fusion [9]. Despite its high conservation and important function in sexual reproduction, HAP2 has been lost several times during evolution, notably in the fungal kingdom [10]. Although a HAP2 ortholog has yet to be identified in fungi, the constantly increasing number of sequenced fungal genomes, specifically of basal fungal lineages, may soon identify a fungal HAP2 ortholog. Interestingly, a fungal-specific fusogen, Prm1, has been identified and functions in both gamete and somatic hyphal fusion [11]. It is unlikely that Prm1 is the product of convergent evolution, but more plausible that a somatic membrane fusion protein took up this role, enabling gamete fusion following HAP2 loss in fungi.

The evolution from reproduction via a single parent to sexual reproduction involving two parents could have provided opportunities in genomic diversification to increase the efficacy of selection and restrict inheritance of cytoplasmic elements. However, recent studies have revealed that meiosis involving a single parent occurs within the fungal kingdom, suggesting cell fusion may be an intermediate evolutionary step rather than ancestral to sexual reproduction. For example, the human fungal pathogens Candida albicans and Cryptococcus deneoformans are capable of both bisexual reproduction between cells of opposite mating types and unisexual reproduction in the absence of partners of the opposite mating type [12–14]. Deletion of either Prm1 or the broadly conserved nuclear fusion protein Kar5 does not block unisexual reproduction in C. deneoformans, and cells diploidized (likely through an endoreplication pathway) and subsequently undergo meiosis [15]. It will be of great interest to characterize endoreplication in this pathogen and explore its conservation across eukaryotes given that this may elucidate a path by which LECA underwent a primordial sexual cycle.

Genetic diversity dictates meiotic cycles

Meiosis is a critical step of sexual reproduction, during which, genome-wide homologous pairing, recombination, and ploidy reduction occur. Conservation and diversification of meiosis across eukaryotes has been extensively reviewed [16]. Broad conservation of meiotic genes throughout the eukaryotic domain hints that meiosis is an ancestral feature of sexual reproduction. It has been proposed that the presence of a core set of meiotic genes (the “meiotic toolkit”) can be used to identify sexual organisms [17]. These genes include SPO11, encoding an archaeal topoisomerase VI-derived enzyme that generates double-strand breaks (DSBs); RAD50 and MRE11 which process DSB DNA ends; DMC1/RAD51, HOP2, and MND1, which form a recombination complex; and MSH4 and MSH5, which promote the formation of Holliday junctions [18,19]. However, application of the meiotic toolkit criteria is not always successful, illustrating the diversity of meiotic machinery. For instance, meiosis and the subsequent production of recombinant progeny occur in social amoebae (dictyostelids) without Spo11, suggesting that an unknown enzyme generates DSBs in dictyostelids [20]. Identifying this unknown enzyme could reveal a missing evolutionary link and potentially expand the core meiotic gene set. In the ciliate Tetrahymena, a type II topoisomerase and Spo11 both function in post-meiotic DNA DSB generation, implying that type II topoisomerases may perform similar functions to Spo11 during meiosis in dictyostelids [21]•. Interestingly, in a recently sequenced ascomycete Trichoderma reesei, a type II topoisomerase functions in parallel with Spo11 in generating crossovers, which promote meiotic recombination [22]•.

The ascomycetous Candida species, C. lusitaniae and C. guilliermondii, represent another example of the limitations of the “meiotic toolkit” criteria, as these species retain a complete meiotic cycle, yet they lack DMC1, MND1, HOP2, MSH4, MSH5, and almost all genes encoding synaptonemal complex proteins, which are crucial for chromosomal pairing during meiosis. In contrast, the C. albicans genome encodes all of the above components, but lacks a complete meiotic cycle and instead undergoes a parasexual cycle with concerted chromosome loss to achieve ploidy reduction [12,23,24]. Recombination during the C. albicans parasexual cycle requires Spo11, indicating that either Spo11 has evolved to participate in this parasexual cycle or that this cycle is a degenerate form of meiosis. Characterization of meiotic genes involved in these Candida species will expand the known repertoire of meiotic genes.

Identifying a complete set of meiosis-specific genes will not only help to decipher the possible ancestral meiotic cycle of LECA, but more importantly will facilitate studies of sexual reproduction in unicellular organisms and help to identify genomes that have undergone reductive evolution, rendering organisms asexual.

Meiotic drive and counteracting forces

Despite the versatile meiotic regulatory circuits across eukaryotes, meiotic cycles play a conserved role in generating diversity for natural selection to act upon. However, certain DNA sequences, genes, or even chromosomes can enhance their own chances of inheritance in progeny without conferring any fitness benefits; these materials are known as selfish genetic elements [25]. The distorted inheritance of these elements can cause rapid population transitions, driving species extinction, speciation, and genomic landscape variation [26]. These selfish elements often impair the fitness of populations that carry them and are therefore generally considered deleterious. However, selfish elements can also have profound industrial or therapeutic potential, like the meiotic drivers that are capable of regulating the Anopheles gambiae mosquito population responsible for transmitting malaria [27,28]•.

Many selfish genetic elements, including killer meiotic drivers, transposable elements, selfish mitochondria, and supernumerary chromosomes, have been identified in fungi. Given their ability to spread quickly in a population, discovery and characterization of selfish elements relies on the identification of natural fertile hybrid populations where drivers have not yet saturated the population [29]. Recent whole-genome sequencing efforts across broad fungal lineages provide robust resources for meiotic driver studies. This new wealth of data has enabled the elucidation of both the evolution and molecular mechanisms of several selfish genetic elements.

The first recognized fungal meiotic driver, Spore killer (Sk) in Neurospora, contains two genetically distinct elements, Spore killer-2 (sk-2) and Spore killer-3 (sk-3), which are present in a 30-cM region surrounding a centromere where recombination is suppressed. Transposable elements enriched in the Sk region form tandem inversions, which further suppress recombination. Recombination suppression is hypothesized to contribute to linkage formation between killer and resistance genes for each Sk driver [30]. Studies have further characterized the rfk-1 gene in the Sk-2 region, which undergoes RNA editing during sexual development to yield an alternative transcript that likely confers resistance to spore killing [31]. The rfk-1 gene is located in a region that escapes meiotic silencing by unpaired DNA, suggesting that silencing of selfish elements could counteract meiotic drive in an evolutionary arms race [31]•. The poison-antidote wtf gene in fission yeast is another meiotic driver that manipulates RNA, utilizing different transcriptional and translational start sites to produce both a poison and an antidote, which protects spores that inherit the wtf gene from killing [32]. Interestingly, spore killing mediated by one wtf gene can be suppressed by another wtf gene that produces a mimic of the driver’s antidote, depicting a unique counter-meiotic drive scenario [33]•. The filamentous fungus Podosopora anserina harbors two different spore killer genes: het-s and Spok. Spread of these meiotic drivers in different fungal lineages is likely through horizontal gene transfer [34–36]. It is worth noting that Sk, wtf, and Spok meiotic drivers confer both killer and resistance functions within a single allele/locus, rendering meiotic recombination and chromosomal reassortment during the meiotic cycle incapable of disrupting the driver effect. To win the evolutionary arms race against meiotic drivers, the nuclear genome must silence the meiotic driver or evolve a suppressor.

Supernumerary chromosomes, uniparental mitochondrial inheritance, and transposable elements have also been described in several fungal species. In the filamentous fungus Zymoseptoria tritici, female supernumerary chromosomes were found to be meiotic drivers due to their inheritance in all meiotic progeny [37]. In the basidiomycete C. neoformans, mitochondria are typically inherited from the MATa parent; however, leaky biparental mitochondrial inheritance occurs in mutants lacking sex-specific transcription factors or when an aneuploid parent is involved, suggesting a mito-nuclear genome interaction dictates this meiotic driver behavior [38,39]. Additionally, the RNAi silencing pathway has been shown to suppress transposable elements in C. neoformans, similar to the finding in Drosophila that the Piwi-piRNA pathway functions as an adaptive defense mechanism in counteracting the selfish nature of transposons [40–42].

As more molecular mechanisms of selfish genetic elements are characterized, fungal organisms become attractive systems to study how meiotic drive shapes genomic landscapes during evolution due to several advantages: (i) with CRISPR-Cas9 technique advances, many single-gene-based meiotic drive systems can now be genetically “domesticated” and characterized through well-controlled experiments; (ii) short meiotic cycles and abundant production of meiotic progeny in many fungi allow detailed analysis of evolutionary trajectories; and (iii) the relatively small genomes of fungi allow efficient whole-genome sequencing of large numbers of meiotic progeny to observe changes in the genomic landscape in relatively large sample sizes.

Genomic landscapes govern sexual compatibility

Due to the high conservation of meiotic processes across eukaryotes, it is interesting to consider how genomic incompatibilities and other mechanisms enforcing species boundaries may have arisen. Although it is impossible to trace the specific events during evolution when sexual conflicts arose and promoted speciation, experimental analyses may shed light on this question.

Species hybridization represents a unique scenario of sexual conflict, and hybrids are largely sterile due to several genetic and epigenetic mechanisms. Genomic sequence divergence between two parental genomes leads to unpaired DNA bases during meiosis, which activates mismatch repair and anti-recombination mechanisms that can cause hybrid sterility. Impairment of the mismatch repair system can partially rescue hybrid sterility in both yeast and mouse models [43–45]. In Saccharomyces species, monitoring meiotic progeny in tetrads using a spore-autonomous fluorescent protein discovered that chromosomal nondisjunction during meiosis I is the main cause of hybrid fertility defects during inter- and intra-species crosses [46].

Besides sequence divergence, several other mechanisms contribute to hybrid incompatibility, including: (i) deleterious epistatic interactions between hybrid genomes, known as Bateson-Dobzhansky-Muller incompatibility; (ii) cytonuclear incompatibility; (iii) epigenetic incompatibility that causes lethal transcriptional activation or silencing in hybrid progeny; and (iv) transcriptomic incompatibility due to divergence of cis- and trans- regulatory elements [47–49]. These mechanisms suggest even small amounts of sequence divergence can directly contribute to postzygotic reproductive isolation.

Chromosomal rearrangements (CRs), such as reciprocal translocations and inversions, can also contribute to reproductive isolation among populations by reducing viability or fertility of progeny and/or via suppressed recombination that protects blocks of linked genes from recombination [50]. There is, however, an intrinsic difficulty in separating CR effects from the genome-wide genetic incompatibilities described above. Using yeast as a model to clarify the contributions of these phenomena, a S. cerevisiae genome was engineered to be collinear with S. mikatae, which structurally differs by two reciprocal translocations. This collinearity promoted hybrid spore viability, thereby supporting the hypothesis that CRs also contribute to reproductive isolation [51]. Another study applied the CRISPR-Cas9-mediated genome editing technique to engineer S. cerevisiae strains with increasingly reduced chromosome numbers and showed that tetrad formation and sporulation were markedly reduced in genetic crosses between the wild-type strain with 16 chromosomes and engineered strains with two chromosomes, indicating that karyotype contributes to driving reproductive isolation [52]••.

The growing number of chromosome-level genome assemblies provide increasing evidence that CRs play a major role in reproductive isolation. A holistic approach integrating genomic, epigenetic, and cytogenetic data will be crucial to advance our understanding of how genomic architecture potentially drives incipient divergence and ultimately speciation. While current studies on chromosome speciation focus on genic (Bateson-Dobzhansky-Muller incompatibilities) and non-genic (CRs) factors, the putative role of meiotic drive, genomic conflict, and disruption of epigenetic programming should not be overlooked in future research on chromosomal speciation.

Conclusion

The growing number of sequenced genomes will allow additional phylogenetic analyses of fusogen, karyogamy, and meiotic genes, will provide insights into the sexual nature of LECA, and will unravel evolutionary trajectories that have generated the astounding ecological diversity of sexual reproduction in modern life. There is also increasing evidence that missing evolutionary links of sexual reproduction lie in non-model organisms, especially in early-diverging lineages and newly described species. With the development of advanced molecular genetic techniques and the vast diversity of the fungal kingdom, we can strive to decipher the past evolutionary trajectories of sexual reproduction.