Mutation and selection at silent and replacement sites in the evolution of animal mitochondrial DNA

Abstract

Two patterns are presented that illustrate the interaction of mutation and selection in the evolution of animal mtDNA: 1) variation among taxa in the ratio of polymorphism to divergence (r(pd)) at silent and replacement sites in protein-coding genes, and 2) strand-differences in polymorphism and divergence at 'silent' sites that suggest a mutation-selection balance in the evolution of codon usage. Cytochrome b data from GenBank show that about half of the species pairs tested have a significant excess of amino acid polymorphism, relative to divergence. The remaining half of species pairs do not depart from neutrality, but generally do show an excess of amino acid polymorphism. Sequences from Drosophila pseudoobscura displaying a signature of an expanding population show a slight, but non-significant, deficiency of amino acid polymorphism suggestive of recently intensified selection on mildly deleterious mutations. Genes whose reading frames lie on the major coding strand of Drosophila mtDNA show a preponderance of T-> C substitutions, while genes encoded on the minor strand experience more A-> G than T-> C substitutions between species at both silent and replacement sites. However, silent mutations at third codon positions are introduced into the population in proportions opposite to those observed as fixed differences between species (e.g., an excess of T-> C polymorphisms are found at the ND5 gene on the minor coding strand). The high A+T content of insect mtDNAs imposes strong codon usage bias favoring A-ending and T-ending codons resulting in a distinct mutation-selection balance for genes encoded on opposites strands. Thus, at both replacement and silent sites, mutations that appear to be constrained in terms of divergence between species are in excess within species. The data suggest that mildly deleterious mutations are common in mitochondrial genes. A test of this, and a competing, hypothesis is proposed that requires additional sequence surveys of polymorphism and divergence. An important challenge is to tease apart the impact of mutation and selection on levels of polymorphism versus divergence in a genome that does not generally recombine.