Patterns of molecular evolution and epistasis on a genomic and genic scale.

摘要

Epistasis describes non-additive interactions which affect gene expression and phenotype. It can happen on multiple levels, including on a genomic level with interactions between genes or even chromosomes affecting global patterns of gene expression. It can also happen within a gene itself, with epistatic interactions between amino acids affecting gene expression and resultant phenotypes. I present three studies in two organisms to study this phenomenon on a global-genomic scale, and also on a local-genic scale.;First, I present evidence that epistatic interactions between Y-linked regulatory polymorphisms and genetic background affect global gene expression in Drosophila melanogaster. The Y chromosome is a heterochromatic, degenerate chromosome and thought to have little evolutionary consequence. I studied Y chromosomes from two populations of D. melanogaster that are known to have major effects on the thermal tolerance of spermatogenesis. I show that these Y chromosomes differentially modify the expression of hundreds of autosomal and X-linked genes, but the effect depends on the genetic background the Y finds itself in. Second, I present novel evidence suggesting that the mechanism for Y-regulatory variation (YRV) is heterochromatin-based. Imprinting (due to parent-of-origin inheritance) in Drosophila has been documented mainly in heterochromatic regions, in particular the Y chromosome. I show that sex-specific transmission of the Y can lead to polymorphic imprinting and can change the magnitude and scope of YRV, perhaps through differential titration of chromatin proteins. In particular, genes responding to this polymorphic imprint were more likely to be male-specific, testis-specific, and involved in rDNA transcript levels. This is particularly intriguing as rDNA processing is known to be affected by heterochromatin formation.;Finally, I study how mutational interactions within one gene can constrain evolutionary trajectories. The human malaria parasite, Plasmodium vivax, varies at several positions in the gene dihydrofolate reductase (DHFR), reflecting the mark of selection for drug resistance. Variation at four amino acid sites allows us to reconstruct a complete fitness landscape using all possible combinations of mutational variants within the gene. The results suggest that sign epistasis, where one mutation does well on some backgrounds but poorly on others, is common within P. vivax DHFR. In addition, drug concentration and effective population size can have a strong effect on whether the most resistant, quadruple-mutant, allele will fix in a population. I propose this may explain why the most resistant allele is missing from common polymorphic natural isolates.