Introduction

During meiosis, the two alleles at a gene locus are separated into gametes, and each gamete has an equal chance of receiving either allele. This fundamental principle of inheritance, known as Mendel’s law of segregation (Abbott and Fairbanks, 2016), holds across most genetic loci in most sexual species. However, meiotic drivers, a class of selfish genetic elements, subvert fair segregation during gametogenesis and are transmitted to more than one-half (even to all) of the functional gametes produced by a heterozygote (Sandler and Novitski, 1957; Lyttle, 1991; Hurst and Werren, 2002; Bravo Nunez et al., 2018b). Meiotic drivers can spread in a population even when they impose fitness costs on their hosts (Crow, 1991; Lindholm et al., 2016). However, the spread of a meiotic driver can be thwarted by the costs imposed on its carriers or by its genetic suppressors (Lindholm et al., 2016).

The fission yeast wtf (with Tf Long Terminal Repeats) gene family provides an excellent model to study how meiotic drivers act and evolve (Hu et al., 2017; Nuckolls et al., 2017). Many wtf genes are autonomous one-gene poison-antidote meiotic drivers that encode both a spore-killing poison (short isoform) and an antidote to the poison (long isoform) using alternative transcriptional initiation (Hu et al., 2017; Nuckolls et al., 2017; Nuckolls et al., 2022). To archieve meiotic drive, all spores are exposed to the poison, whereas only those that inherit wtf express the antidote and are rescued (Hu et al., 2017; Nuckolls et al., 2017; Nuckolls et al., 2022). Some other wtf genes can act as drive suppressors (Bravo Nunez et al., 2018a; Bravo Nunez et al., 2020a). The poison and the antidote differ only in their N-terminal cytosolic tails containing PY (Leu/Pro-Pro-X-Tyr) motifs. PY motif-dependent ubiquitination promotes the transport of the antidote and the poison (physically interacted with the antidote) from the trans-Golgi network to the endosome, thereby preventing toxicity (Zheng et al., 2023).

The wtf gene family is experiencing rapid diversification: the Schizosaccharomyces pombe reference genome encodes 25 wtf genes, some of which are pseudogenes. The copy numbers of wtf genes vary greatly among different S. pombe strains, and frequent nonallelic gene conversion occurs between wtf genes (Hu et al., 2017; Nuckolls et al., 2017; Eickbush et al., 2019). However, these finding are based on a limited number of strains, and the pattern and extent of recombination in the wtf genes remain to be fully explored. Moreover, wtf driver genes are present in the last common ancestor (LCA) of the fission yeasts S. pombe, S. octosporus, S. osmophilus, and S. cryophilus, indicating that wtf genes have likely maintained the capacity to drive for more than 100 million years (De Carvalho et al., 2022). These fission yeast species carry varying numbers of wtf genes, ranging from 5 to 83 (De Carvalho et al., 2022). Yet, it remains perplexing how wtf genes achieved such diversification. On one hand, gene duplication can give birth to new wtf gene copies. On the other hand, a newly duplicated wtf gene might not drive because the poison produced by the newly duplicated wtf gene can be detoxified by the original wtf gene. Like newly duplicated genes, the most probable fate of a new wtf duplicate is pseudogenization (Lynch, 2007; Innan and Kondrashov, 2010). Thus, the vast majority of new wtf duplicates experience an early exit from the population, most probably never reaching fixation (Lynch, 2007; Innan and Kondrashov, 2010). To become a new driver, the new wtf copy should evolve coupling new poison and new antidote to the new poison through mutations. But the fate-changing mutations are likely to be rare. It follows that gene duplication might be insufficient to drive the diversification of wtf genes.

In this study, we analyzed the diversity and evolution of wtf genes in the genomes of 21 strains of S. pombe that were sequenced using long-read sequencing approaches (Tusso et al., 2022). Through knocking out each of 25 wtf genes in S. pombe laboratory strain 972h-, no significant attenuated growth was observed, indicating wtf genes might be not deleterious in the asexual life cycle. We found that recurrent recombination occurred among wtf genes. We generated a novel meiotic driver through artificial recombination between wtf drivers, and its encoded poison cannot be detoxified by the antidotes encoded by their parental wtf genes but can be detoxified by its own antidote. Therefore, we propose that recombination can generate wtf driver with new poisons and might shape the diversification of wtf genes.

Results

Diversity and evolution of wtf genes in fission yeasts

First, we analyzed the diversity and distribution of wtf genes in fission yeasts. The S. pombe reference genome (strain 972h-) encodes a total of 25 wtf genes. For these 25 wtf genes, the number of exons varies from 3 to 6 (Figure 1A) (Bowen et al., 2003). To investigate the relationship among exons from different wtf genes, we grouped these wtf exons into clusters based on nucleotide identity of 0.50 (Figure 1B and C). The wtf exons were grouped into 10 clusters with >2 members, and 12 exons exist as singletons in the similarity network (Figure 1B and C; Supplementary file 1c). Only exon 1 of all the 25 wtf genes group together in a cluster, indicating that the first exons are well conserved among the wtf genes. No other exon is conserved among all the 25 wtf genes. Therefore, the evolution of the wtf gene structures appear to be highly dynamic.

Figure 1.

We next identified wtf genes in 21 strains of S. pombe that were sequenced using long-read sequencing approaches (Supplementary file 1b ) (Tusso et al., 2022). The copy number of wtf genes varies among different S. pombe strains, ranging from 24 (strain JB879) to 37 (strain JB1206) (Figure 1D). Synteny analyses show that the wtf genes are present in 20 genetic loci (Figure 1D). Multiple wtf genes were present in 13 wtf loci. Within 20 wtf loci, at least one wtf gene is present in all of or nearly all of the 21 S. pombe strains, suggesting that these 20 wtf loci might have originated before the LCA of the 21 S. pombe strains. Putative wtf pseudogenes are prevalent in many wtf loci among the 21 S. pombe strains, indicating frequent pseudogenization occurred in the wtf genes. These results indicate that wtf copy number variation is prevalent among S. pombe strains.

We performed phylogenetic analyses of the wtf genes from 21 S. pombe strains and three other fission yeast species (S. octosporus, S. cryophilus, and S. osmophilus) (Figure 1E; Supplementary file 1d). The wtf genes of S. pombe form a monophyletic group. Orthologs of wtf14, wtf7, wtf11, and wtf15 form monophyletic groups, whereas orthologs of other wtf genes show complex phylogenetic mixing, indicating complex recombination might have occurred among these wtf genes (Figure 1E) (Eickbush et al., 2019). Moreover, the wtf genes with 6 exons (including the known functional drivers wtf4 and wtf23) cluster together, and exhibit a ladder-like phylogeny, which might be generated by continual selection driven by antidotes (like the ladder-like phylogeny of influenza A viruses H1N1 and H3N2, which is shaped by continual immune selection (Grenfell et al., 2004; Bedford et al., 2011)). Based on phylogenetic relationship, we divided the wtf genes of 21 S. pombe strains into eight groups, namely groups 1 to 8, among which groups 5 to 8 include orthologs of wtf14, wtf7, wtf11, and wtf15, respectively (Figure 1E). Exon 2 ATG codons (exon-2-ATG) and in-frame ATG within intron 1 and near the start of exon 2 (intron-1-ATG) of wtf genes can encode the start of poison protein isoforms (Hu et al., 2017). We found that most of exon-2-ATG and intron-1-ATG are present within group 1 wtf genes (Figure 1E). A majority of the wtf genes are flanked by solo-LTRs (Bowen et al., 2003) (Figure 1E). However, the solo-LTRs flanking the wtf genes do not cluster together but form many distinct groups, suggesting that solo-LTRs were inserted nearby the wtf genes multiple times (Supplementary file 1a). Together, our results reveal the rapid diversification and turnover of wtf genes in a single fission yeast species.

No attenuated growth of fission yeast without wtf genes

To explore the effect of wtf genes on the fitness of fission yeast, we knocked out each of the 25 wtf genes in the S. pombe laboratory strain 972h-using a method based on homologous recombination (Figure 2A). A total of 25 wtf knockout strains (Δwtf1 to Δwtf25) were generated. We used spot assay to evaluate the effect of wtf gene knockout on the yeast growth, and no growth defect was observed for all the 25 wtf knockout strains (Figure 2B). Therefore, our experiment suggests that the wtf genes might be largely neutral to the fitness of their carriers in the asexual life cycle at least in normal growth condition.

Figure 2.

Recurrent recombination in wtf genes

Given complex phylogenetic mixing observed among wtf genes (Figure 1E), we tested whether recombination occurred. We detected signals of recombination in the 25 wtf genes of the S. pombe reference genome (p = 0) and in the wtf genes of the 21 S. pombe strains (p = 0) using pairwise homoplasy index (HPI) test. Split network analysis also supports the frequent occurrence of recombination in the 25 wtf genes of the S. pombe reference genome (Figure 3A) and in the wtf genes of the 21 strains of S. pombe (Figure 3B). In contrast, no recombination signal was tested for groups 5 to 8 using HPI test (p = 1 for group 5, p = 1 for group 6, p = 0.53 for group 7, and p = 1 for group 8). We estimated recombination rates of the full-length wtf sequences, the first exons, and the wtf sequences without the first exons for wtf groups 1 to 4. We found that the recombination rate of group 1 wtf was highest among the four wtf groups (Figure 3C). For group 1, breakpoints are dispersed across the wtf sequences (Figure 3D). These lines of evidence suggest that wtf genes underwent recurrent and intricate recombination.

Figure 3.

Generation of a new driver gene through artificial recombination

Given gene duplication alone might be insufficient to shape the diversification of wtf genes, we hypothesize that recombination between wtf genes can generate new meiotic drives. To test this, we constructed four chimeric wtf genes through recombination among known functional meiotic drivers (wtf23) and an artificially generated meiotic driver (wtf18).

We used a proved Saccharomyces cerevisiae system to test the activity of poison and antidote proteins encoded by wtf genes (Nuckolls et al., 2020). As expected, the expression of the poison proteins (Wtf23^poison) encoded by wtf23 genes caused the yeast growth arrest (Figure 4A). The attenuated growth was alleviated, when the corresponding antidote proteins (Wtf23^antidote) were expressed (Figure 4A). We also experimentally analyzed wtf18 gene, which was known to encode only long (antidote-like) transcripts and probably act as a suppressor (Bravo Nunez et al., 2018a). We artificially introduced an in-frame ATG codon right before the start of exon 2, generating wtf18^poison/-0M. The expression of wtf18^poison/-0Mresulted in the yeast growth arrest, suggesting its product, Wtf18^poison/-0M, is indeed a poison protein (Figure 4B). When co-expressing wtf18^antidote and wtf18^poison/-0M, the attenuated yeast growth was rescued (Figure 4B), indicating that Wtf18^antidote can ameliorate the toxicity of Wtf18^poison/-0M. Therefore, the introduction of an ATG codon to wtf18 leads to the generation of a putative meiotic driver.

Figure 4.

We then constructed four chimeric wtf genes through artificial recombination between wtf23 and wtf18, including wtfC1 (possessing exons 1-2 of wtf23 and exons 3-6 of wtf18), wtfC2 (possessing exons 1-3 of wtf23 and exons 4-6 of wtf18), wtfC3 (possessing exons 1-4 of wtf23 and exons 5-6 of wtf18), and wtfC4 (possessing exons 1-5 of wtf23 and exon 6 of wtf18). The expression of the short isoforms of wtfC1, wtfC2, wtfC3, and wtfC4 resulted in yeast growth arrest, revealing their toxicity (Figure 4C-F). However, the antidote of wtfC1 and wtfC3 cannot detoxify the corresponding chimeric toxins (Figure 4C and E). Interestingly, we generated a putative novel meiotic driver, namely wtfC4. Our results show that wtfC4 encodes a functional poison (WtfC4^poison) (Figure 4F). The poison can be detoxified by its own long isoforms (dubbed as wtfC4^antidote), but cannot be detoxified by the antidote proteins of their parental genes (Figure 4F). Taken together, we generated a new meiotic driver through artificial recombination between pre-existing wtf genes.

Discussion

In this study, we analyzed the diversity and evolution of wtf genes in fission yeasts. The copy number of the wtf gene varies among different S. pombe strains, revealing rapid diversification and turnover of the wtf genes within a single fission yeast species. We detected signals of recurrent and intricate recombination among wtf genes as previously reports with limited genomes (Hu et al., 2017; Nuckolls et al., 2017; Eickbush et al., 2019; De Carvalho et al., 2022). We hypothesize that recombination between wtf genes can produce new wtf genes with new poisons and the antidotes to new poisons. These new wtf genes can then driver through populations. As proof-of-principle, we generated a chimeric wtf gene that represents a new meiotic driver. The encoded poison of the newly generated meiotic driver can be detoxified by its own long isoforms, but cannot be detoxified by the antidote proteins of their parental genes. However, the other three chimeric wtf genes did not show this property, possibly because the last exon might be crucial for antidote function. Indeed, our recombination breakpoint analyses (Figure 3D) reveal substantial recombination might have occurred in the last exon. Together, our results indicate that recombination is likely to drive the rapid diversification of wtf gene in fission yeasts.

Most of the known meiotic drivers impose costs on their carriers due to direct effects of the driver on survival or fertility, production of a biased sex ratio, or via deleterious mutations linked to the driver (Price and Wedell, 2008; Larracuente et al., 2012; Sutter and Lindholm, 2015; Fishman et al., 2015; Lindholm et al., 2016; Zanders and Unckless, 2019). In outcrossing between individuals from distinct yeast lineages, wtf drivers can provide a selective advantage to atypical spores, such as aneuploids and diploids (Bravo Nunez et al., 2020b). In this study, we assessed the effects of the wtf genes on the growth of fission yeast during the asexual life cycle through knocking out each of the 25 wtf genes in S. pombe laboratory strain 972h-separately. We did not observe obvious attenuated growth for these wtf knockout strains, indicating wtf genes are largely neutral to the fitness of their carriers during the asexual life cycle at least in the normal growth setting. It should be noted that the spot assay used in this study detect large differences in fitness between wide type and wtf knockout strains. Nevertheless, it is likely that wtf genes evolve mainly in a neutral manner during the asexual life cycle, which explains the presence of high proportion of pseudogenes in wtf gene repertoire. Moreover, asexual reproduction is much more frequent than sexual reproduction for yeasts (Tsai et al., 2008). Therefore, even if fate-changing mutations that simultaneously produce new poison and the antidote to new poison occur, the most probable fate of a new wtf gene generated by gene duplication is pseudogenization and removed from the population.

Gene duplication gives rise to new wtf genes. However, the newly generated wtf gene can be detoxified by the original wtf gene and thus cannot drive through its host population, when the original wtf is fixed in the population (Figure 5). Therefore, most, if not all, of wtf gene duplicates experience early exist from the host population. When recombination occurs between two pre-existing wtf genes, chimeric wtf gene with new poison and the antidote to new poison can be generated as this study shows. Then, the wtf gene with new driver property can spread in its host population, even reaching fixation (Figure 5). During asexual life cycle, wtf genes evolve mainly under genetic drift, and thus can accumulate disruptive mutations, leading to their pseudogenization. Taken together, our study highlights the significance of recombination in shaping the diversification of wtf genes.

Figure 5.

Methods

Identification of the wtf genes

We used the blastn algorithm to identify wtf genes within 21 S. pombe strains with 25 wtf genes from S. pombe reference genome as the queries and an e-cutoff value of 10^-5. The identified wtf genes were annotated based on the wtf genes of reference genome. The sequence identity among the exons of wtf genes was calculated using BioAider version 1.334 (Zhou et al., 2020). Exons was then clustered based on the nucleotide identity using igraph package version 2.0.1.1 (Csardi et al., 2006; Csardi et al., 2024). We extended 1000 bp flanking each wtf gene to establish their syntenic relationships.

Phylogenetic analysis

The coding sequences of wtf genes of 21 S. pombe strains and three other fission yeast species (S. octosporus, S. cryophilus, and S. osmophilus) were aligned using MAFFT version 7 (Katoh and Standley, 2013). To clarify the relationship of 21 S. pombe strains, 30 genes were randomly selected and concatenated using Phylosuite version 1.2.1 (Zhang et al., 2020). All the phylogenetic analyses in this study were performed using the maximum likelihood (ML) method implemented in IQ-TREE version 2 (Minh et al., 2020). The best-fit substitution model was selected using the ModelFinder algorithm (Kalyaanamoorthy et al., 2017). Node supports were assessed using the ultrafast bootstrap approximation (UFBoot) method with 1,000 replicates (Hoang et al., 2018). Solo-LTRs were identified using the blast algorithm and aligned using MAFFT version 7 (Katoh and Standley, 2013). Phylogenetic analysis was performed using the approximate maximum likelihood method implemented in FastTree version 2.1.1 (Price et al., 2010).

Recombination analysis

Split networks of wtf genes were generated using the neighborhood network analysis implemented in Splittree4 (Huson et al., 2006). Pairwise homoplasy index (PHI) test was performed using SplitsTree4 (Huson et al., 2006). Potential breakpoints were detected using 3SEQ (Lam et al., 2018). The recombination rate was estimated using FastEPRR package version 2.0 (Gao et al., 2016).

Generation of wtf knockout strains

The wtf knockout (Δwtf) strains generated in this study were derived from S. pombe strain 972h-. We constructed substitution cassettes for each of the 25 wtf genes of the S. pombe reference genome. Substitution cassettes contain a kanMX resistance marker and two homologous sequences flanking the target wtf genes (Moreno et al., 1991; García-Ríos et al., 2014). Substitution cassettes were transformation into S. pombe (strain 972h-) through the lithium acetate-based method (Moreno et al., 1991; García-Ríos et al., 2014). wtf gene knockout strains were selected for kanMX resistance and were verified by PCR.

Plasmid construction

Total RNA of fission yeast was extracted and reverse transcribed into cDNA. Coding sequences of wtf23^antidote, wtf23^poison, and wtf18^antidote were amplified using the corresponding primers (Supplementary file 1e). wtf18^poison/-M0 was generated using wtf18^antidote as the template and the primer with an artificially introduced ATG (Supplementary file 1e). We generated wtfC1 through recombining exons 1-2 of wtf23 and exons 3-6 of wtf18, generated wtfC2 through recombining exons 1-3 of wtf23 and exons 4-6 of wtf18, generated wtfC3 through recombining exons 1-4 of wtf23 and exons 5-6 of wtf18, and generated wtfC4 through recombining exons 1-5 of wtf23 and exon 6 of wtf18. These wtf and wtfC genes were then cloned into the GAL1/10 dual expression plasmid Gal_HF. Plasmids were first transformed into Escherichia coli and verified by PCR and sequencing. Plasmids were then transformed into S. cerevisiae (strain S288C) using the lithium acetate-based method. Yeast transformants were selected for kanMX resistance and were verified by PCR.

Spot assay

The yeast strains were cultured in YPD liquid medium at 30℃ with shaking at 200 rpm. The overnight cultures were transferred to fresh YPD liquid medium and grown to an OD₆₀₀ value of ∼3. Cells were collected by centrifugation, and the OD₆₀₀ was adjusted to 3. Subsequently, the strains were continuously diluted by a 10-fold gradient to 10^-5, generating five stock suspensions. The five stock suspensions were used for spot assays. We plated 1.5μL of each stock suspension on the surface of YPD and YPG solid media. The plates were incubated at 30℃, and the growth of colonies were observed.

Acknowledgements

This work was supported by National Natural Science Foundation of China (32270684 to G.- Z.H. and 32300511 to Z.G.).

Data Availability

No new data were generated in support of this research.