No description available.

Determining the genetic basis of reproductive isolation is a fundamental goal in evolutionary biology. Intrinsic reproductive isolation often arises due to epistasis between divergent interacting genes. The rapid evolution of hybrid male sterility is known to have several causes, including the exposure of recessive X-linked incompatibilities in males and the rapid evolution of male reproductive traits. Despite these insights, little is known about the genetics of reproductive isolation during the early stages of speciation. This deficiency inspired parallel studies on the molecular evolution of male reproduction in house mice and the genetic basis of hybrid male sterility between two mouse species, Mus domesticus and M. musculus. Evolutionary analysis of 946 genes showed that the intensity of positive selection varies across sperm development and acts primarily on phenotypes that develop late in spermatogenesis (Appendix A). Several reciprocal crosses between wild-derived strains of M. musculus and M. domesticus were used to examine F1 hybrid male sterility (Appendix B). These crosses revealed hybrid male sterility linked to the M. musculus X chromosome and a novel sterility polymorphism within M. musculus. A large introgression experiment was used to further dissect the genetic basis of X-linked incompatibilities between M. musculus and M. domesticus (Appendix C). Introgression of the M. musculus X chromosome into a M. domesticus genetic background produced male sterility and involved a minimum of four factors. No sterility factors were uncovered on the M. domesticus X chromosome. These data demonstrate the complex genetic basis of hybrid sterility in mice and provide numerous X-linked candidate sterility genes. The molecular evolution of five rapidly evolving candidate genes was examined using population and phylogenetic sampling in Mus (Appendix D). Four of these loci showed evidence of positive natural selection. One locus, 4933436I01Rik, showed divergent protein evolution between M. domesticus and M. musculus and was one of a handful of testis-expressed genes within a narrow interval involved in hybrid male sterility. In summary, these data demonstrate that hybrid male sterility has a complex genetic basis between two closely related species of house mice and provide a foundation for the identification of specific mutations that isolate these species.

Knowledge of the distribution of fitness effects of new mutations (DFE) can enable us to quantify the amount of genetic change between species that is driven by natural selection and contributes to adaptive evolution. The primary focus of this thesis is the study of methods to infer the DFE and the study of adaptive evolution in the house mouse subspecies Mus musculus castaneus. Firstly, I extended previous methodology to model the DFE based on polymorphism data. Methods that have previously been used to infer the DFE from polymorphism data have relied on the assumption of a unimodal distribution. I developed new models that can be used to fit DFEs of arbitrary complexity, and found that multimodality can be detected by these models given enough data. I used these new models to analyse polymorphism data from Drosophila melanogaster and M. m. castaneus, and found evidence for a unimodal DFE for D. melanogaster and a bimodal DFE for M. m. castaneus. Secondly, I investigated the contribution of change in coding and non-coding DNA to evolutionary adaptation. I used a polymorphism dataset of ~80 loci from M. m. castaneus sequenced in 15 individuals to investigate selection in protein-coding genes and putatively regulatory DNA close to these genes. I found that, although protein-coding genes are much more selectively constrained than non-coding DNA, they experience similar rates of adaptive substitution. These results suggest that change in functional non-coding DNA sequences might be as important as protein-coding genes to evolutionary adaptation. Thirdly, I used whole genome data from 10 M. m. castaneus individuals to compare the rate of adaptive substitution in autosomal and X-linked genes. I found that, on average, X-linked genes have a 1.8 times faster rate of adaptive substitution than autosomal genes. I also found that faster-X evolution is more pronounced for male-specific genes. I used previously developed theory to show that these observations can be explained if new advantageous mutations are recessive, with an average dominance coefficient less than or equal to 0.25. These results can help to explain the long-studied phenomenon of the large effect of the X chromosome in speciation.

Whole-genome studies of rates of protein evolution show that genes underlying reproduction and immunity tend to evolve faster than other genes, consistent with the frequent action of positive selection. The evolution of immunity has been well-studied at the interspecific level, but much remains unknown about the population-level dynamics of immunity. This project described genetic variation at immunity and non-immunity loci as well as variation among levels of infection for diverse pathogens in a natural population of mice from Tucson. Analysis of autosomal and X-linked loci in the native range of Mus domesticus, the species from which Tucson mice are primarily descended, revealed low levels of variation consistent with a recent population expansion, resulting in a slight excess of rare alleles across the genome. Genetic variation among a set of classical inbred strains represented a small fraction of wild variation. An overlapping set of genes sequenced in mice from Tucson revealed that there is significant introgression from Mus castaneus. After controlling for gene flow, Tucson mice showed evidence of a mild bottleneck that produced a slight excess of intermediate frequency alleles, but did not result in a dramatic loss of genetic variability. Most of the 15 pathogens and parasites studied in Tucson were found at low to intermediate frequency, and most mice had one to three infections, suggesting that there are many opportunities for host-pathogen coevolution, and a possible role for coinfection. A study of Fv-4, which confers resistance to murine leukemia viruses, confirmed that the resistance allele originated in M. castaneus and is now found at intermediate frequency in Tucson after introduction through gene flow. Finally, a study of the recently duplicated Ceacam1 and Ceacam2 genes, previously shown to be involved in resistance to mouse hepatitis virus (MHV), revealed that a gene conversion event moved a suite of mutations from Ceacam2 to Ceacam1. An elevated rate of protein evolution showed that Ceacam2 had experienced positive selection after duplication. Interestingly, there was no association between MHV antibody presence and Ceacam1 genotype in Tucson. This project showed that gene flow and gene conversion mediated resistance to infections in wild mice.

I asked whether there is female sterility in hybrid offspring of Mus musculus domesticus and M. m. musculus using two wild derived inbred strains representing each subspecies. I evaluated F1 hybrid female fertility by crossing F1 females to a tester male and comparing multiple reproductive parameters between intra-subspecific controls and inter-subspecific hybrids. Hybrid females had smaller litter size, reduced embryo survival, fewer ovulations, and fewer small follicles relative to controls. Significant variation in reproductive parameters was seen among different genotypes. Genes contributing to hybrid female infertility are polymorphic within subspecies. Differences in reproductive phenotypes in F1's of reciprocal crosses suggest that female subfertility may be due to either cyto-nuclear incompatibility or to imprinting. These findings suggest a greater complexity in hybrid sterility than has been previously appreciated and highlight the potential importance of hybrid female sterility in the early stages of speciation.