Reproductive isolation due to divergent ecological selection is accompanied by vast genomic instability in experimentally evolved yeast populations

Abstract

Populations evolving independently in divergent environments accumulate genetic differences and potentially evolve reproductive isolation as a by-product of divergence. The speed and mechanisms underlying this process are difficult to investigate because we rarely get the opportunity to witness them in natural settings, and histories of selection and gene flow between populations are often unknown. Here, we experimentally evolved yeast for 1000 generations of evolution in both divergent and parallel environments. At regular time points during experimental evolution, we made crosses between parallel- and divergent-evolving populations to measure postzygotic reproductive isolation (gamete viability). We used whole genome population sequencing to determine the mutational load, the number and types of structural variation, and other genomic features of the parent, F1, and F2 intraspecific hybrids. We found evidence for large scale phenotypic and genome-wide differentiation in response to divergent laboratory selection. Divergent-selected populations produced hybrids with reduced gamete viability - a classic signature of postzygotic reproductive isolation in the form of hybrid breakdown. Parallel-selected populations on the other hand remained more reproductively compatible (with exceptions). We found that F2 hybrid genomes contained vast genomic instability, i.e., new structural variants (especially insertions, deletions, and interchromosomal translocations) that were not observed in parent and F1 genomes, which is likely a result of chromosome missegregation and recombination errors in hybrid meiosis. Our results provide phenotypic and genomic evidence that partial reproductive isolation evolved due to adaptation to divergent environments, consistent with predictions of ecological speciation theory.