The Evolutionary History of Yeast

The recent publication of genome sequences and phenotypes of 1,011 S. cerevisiae strains has provided a great resource for understanding the evolutionary history of this species (Peter et al. 2018). Dissection of genetic diversity in the analyzed strains has revealed that several segregating genes in the global yeast population are not present in the S288c lab reference strain. Specifically, as described by Anne Friedrich, while 2,908 genes are variably present across different S. cerevisiae strains, 1,712 genes are absent in the reference genome. This genetic variability in gene content is also present when looking at the “pangenome” of 6 other Saccharomycotina species which also have both a core and accessory genome. While this poses challenges to genotype-phenotype mapping, it suggests that yeast has an open pangenome possibly due to extensive interspecific hybridization between closely related species. The idea that genetic variation could result from hybridization events was also put forward by Anne Clark who described the history of introgressed sequences from S. paradoxus in 93 S. cerevisiae strains isolated from different ecological niches. The hypothesized 3,000 introgressed regions are distributed on all chromosomes. While the amount of introgression varies considerably between different strains, it accounts for ∼8% of all polymorphisms. Another path to variation comes from whole genome duplication, an ancient event that shaped the evolutionary history of yeast. Christian Landry described investigation of how gene duplication contributes to genetic robustness by studying the effect of gene duplicates on protein-protein interactions. Surprisingly, rather than providing functional redundancy, duplicates that function as heterodimers become functionally dependent on each other (Diss et al. 2017). As we learn more about the past of yeast, the future evolutionary history of yeast looks increasingly unnatural. Jef Boeke provided an update on the continuing work to modify the yeast genome. While efforts to reduce, relocate and modify the yeast genome to achieve inducible genetic flexibility are well on the way as part of the Sc2.0 project, he presented recent data on karyotype engineering to concatenate yeast chromosomes. Remarkably all genetic information in yeast can be faithfully transmitted in one (Shao et al. 2018) or two (Luo et al. 2018) chromosomes with little apparent fitness cost for the organism. On top of being a remarkable example of eukaryotic genome restructuring, this work suggests that karyotype evolution could have occurred in an ancient single-linear-chromosome organism by several rounds of genome duplications followed by gene loss. As described by Aimée Dudley, the utility of engineering yeast strains also has other applications. Genetic complementation assays in which human disease genes replace their yeast orthologs can be used to study the functional consequences of human genetic variation. Large-scale analysis of allelic variation, so called deep mutational scanning, has emerged as a powerful approach for structure-function analysis and has dramatically expanded the utility of yeast for predicting the effect of human variants of uncertain significance (VUS), by far the largest class of variants detected in clinical sequencing.

Although evolutionary biology benefits from generalizability of principles, sometimes, the exceptions are informative. Ken Wolfe described a fascinating story about the evolutionary origins of the mating type switching endonuclease gene, HO, which has no orthologs except in S. cerevisiae and is absent from almost all pre-whole genome duplication yeast species. The discovery of WHO genes (Weird HO) in Torulaspora delbrueckii points toward a model in which HO originated from the only intein-containing gene in yeast, VMA1. Its fusion to a zinc finger domain provides the protein sequence specificity and ability to integrate into empty alleles. Another exception was described by Anne-Ruxandra Carvunis, who summarized recent progress in understanding de novo gene birth, a process that generates proto-genes without ancestral gene modification. This process was initially thought to be extremely rare but has become a subject of greater interest following the identification of some 2,000 proto-genes by ribosome profiling data (Carvunis et al. 2012). Interestingly, 5% of 342 studied protogenes confers beneficial phenotypes upon overexpression. These beneficial proto-genes are enriched in transmembrane domains and have T-rich regions, likely encoding hydrophobic amino acids. In one case, alignment of the proto-gene sequence to other yeasts suggests that the transmembrane domain predated the coding potential.

Yeast as a model for species hybridization

The tractability of yeast as an experimental system has made it the preeminent tool for studying species hybridization. In an experimental tour de force, Jun-Yi Leu reported the construction of chromosomal replacement strains in which each S. cerevisiae chromosome has been replaced by the corresponding S. paradoxus chromosome. Multiple lines of evidence trace the decreased fitness in these “hybrids” to proteotoxic stress, derived from the inability of proteins encoded from the two different genomes to correctly assemble into complexes. Conversely, hybrids can also display higher fitness than the parental lines, as described by Rike Stelkens, who showed that F1 and F2 strains generated from S. cerevisiae/S. paradoxus hybrids have increased survival probabilities in stressful conditions. In order to rapidly map species-specific phenotypic differences, Rachel Brem described a recently published (Weiss et al. 2018) approach using genome-wide reciprocal hemizygosity testing coupled with random transposon mutagenesis in hybrids. This clever approach is also likely to have utility for identifying causative genes in traditional QTL mapping using F1 hybrids. Large-scale random insertional mutagenesis is clearly the way forward for functional genomics in diverse strains and species. Application of this method to C. albicans haploid clinical isolates by Judith Berman led to the definition of the “essentialome” in this difficult to manipulate human pathogen (Segal et al. 2018) offering new targets for therapeutic avenues.

Molecular mechanisms in an evolutionary context

Evolutionary consequences provide an explanation for why certain phenomena are observed, but understanding the molecular bases of regulatory processes that are critical in the life-history of the organism is required for understanding the details of how evolution works. How organisms respond to stress is of critical importance for understanding how organisms can survive in the wild. Naama Barkai reported new insights into how regulation of expression and noise by MSN2 and MSN4 balances the cell’s capacity to respond to stress, while maintaining full growth potential. MSN2 and MSN4 activate the same set of stress responsive targets. Whereas MSN2 is constitutively expressed at a low level, MSN4 is expressed noisily and is strongly inducible. This redundant two-factor regulatory system provides a mechanistic solution to the challenge of rapidly responding to transient stress while maintaining the capacity to proliferate in the presence of persistent stress. Non-genetic heterogeneity within a population of genetically identical cells may provide additional mechanisms for survival. Mark Siegal reported the role of MSN2 and MSN4 in generating cell-to-cell heterogeneity, which in turn contributes to the ability of the heterogeneous population to survive an acute stress (Li, Giardina, and Siegal 2018).

Studying the molecular mechanisms of regulation requires the generation of new tools. Naomi Ziv described the development of estradiol synthetic inducible systems (McIsaac et al. 2013) to recapitulate the regulatory mechanisms underlying the bistable stochastic switching between white and opaque phenotypes in C. albicans. Jens Nielsen and Kiran Patil described the use of multi-omics approaches to track the contribution of epigenetic changes to adaptive phenotypes and to highlight the role of metabolite exchange in complex microbial ecosystems, respectively.

Christian Brion described a clever single-cell approach to simultaneously mapping QTL that underlies variation in protein and mRNA abundance of a gene using an extension of the x-QTL based mapping approach (Albert et al. 2014). Chris Jakobson summarized his recent QTL study that provides insight into the relationship between genetic variation and quantitative traits using crosses between minimally diverged strains (She and Jarosz 2018). Interestingly, more than half the causative variants at QTL are synonymous or regulatory, suggesting a pervasive role for natural variation in tuning both transcription and translation. However, Ian Ehrenreich provided a sobering reminder of the complexity of genotype-phenotype relationships in his dissection of the multigenic and non-additive genetic basis of colony morphology phenotypes (Lee et al. 2018). The generation of heritable variation may also occur in unexpected and nonrandom ways, as Jon Houseley described the formation of circular DNA encoding the CUP1 gene when cells age in the presence of copper, an interesting extension of his recent report that CUP1 CNVs are stimulated by the presence of environmental copper (Hull et al. 2017).

Host-parasite interactions

Perhaps the most exciting and novel topic of the meeting was the emerging utility of using fungal systems for studying interactions between species and between hosts and their parasites. Harmit Malik told us about natural variation in the capacity for different yeast strains to maintain the 2 micron plasmid, a selfish genetic element that confers no benefit to the host. QTL mapping between a restrictive and permissive strain identified a single locus on chromosome V that appears to underlie this variation. The elucidation of the causative gene should provide new insight into this poorly understood class of genetic variation. Selfish genetic elements also likely contribute to some of the results that have been observed in experimental evolution populations. Sean Buskirk reported new studies of the role of the killer virus in experimental evolution. While present in the ancestor, killer ability is commonly lost from evolving populations, owing to mutations in the killer virus genome. These populations can subsequently develop sensitivity to killer toxin, leading to a rock-paper-scissor scenario in which evolved populations lose in competition assays with the ancestral genotype. A fascinating talk about the nematode eating fungus, Arthrobotrys oligospoara, by Yen-Ping Hsueh highlighted the amazing diversity of fungal lifestyles and their ability to trap nematodes. This fungal species is ubiquitous in the environment and the wild isolates are highly polymorphic in their prey-sensing ability. G-protein signaling plays a major role in sensing ascarosides secreted by nematodes, which induces trap formation by the fungus. Kirsten Nielsen told us that the human pathogen Cryptoccus neoformans tries to evade host immune cells and adapt to host lungs by generating titan cells that are up to 100μm in diameter through endoreduplication. These polyploid cells generate daughter cells that are genetically distinct and may be beneficial in the presence of various stresses, including antifungal drugs (Gerstein et al., 2015).