The Roles of Mutation, Selection, and Expression in Determining Relative Rates of Evolution in Mitochondrial versus Nuclear Genomes

Abstract

Eukaryotes rely on proteins encoded by the nuclear and mitochondrial (mt) genomes, which interact within multisubunit complexes such as oxidative-phosphorylation enzymes. Although selection is thought to be less efficient on the asexual mt genome, in bilaterian animals the ratio of nonsynonymous to synonymous substitutions (x) is lower in mt-compared with nuclear-encoded OXPHOS subunits, suggesting stronger effects of purifying selection in the mt genome. Because high levels of gene expression constrain protein sequence evolution, one proposed resolution to this paradox is that mt genes are expressed more highly than nuclear genes. To test this hypothesis, we investigated expression and sequence evolution of mt and nuclear genes from 84 diverse eukaryotes that vary in mt gene content and mutation rate. We found that the relationship between mt and nuclear x values varied dramatically across eukaryotes. In contrast, transcript abundance is consistently higher for mt genes than nuclear genes, regardless of which genes happen to be in the mt genome. Consequently, expression levels cannot be responsible for the differences in x. Rather, 84% of the variance in the ratio of x values between mt and nuclear genes could be explained by differences in mutation rate between the two genomes. We relate these findings to the hypothesis that high rates of mt mutation select for compensatory changes in the nuclear genome. We also propose an explanation for why mt transcripts consistently outnumber their nuclear counterparts, with implications for mitonuclear protein imbalance and aging.

The dynamics of mito-nuclear coevolution: A perspective from bivalve species with two different mechanisms of mitochondrial inheritance

“…The transcription level of mtDNA‐encoded OXPHOS subunits is much higher than that of nuclear‐encoded subunits in both the species (Figure , Supporting Information Table ). This pattern was already reported in a wide range of eukaryotes, where the transcript abundance of mitochondrial subunits is, on average, 20‐fold higher than that of nuclear subunits in animals, 18‐fold higher in plants, and 6‐fold higher in fungi (Havird & Sloan, ; Nabholz et al., ). Here, we report a high abundance of mitochondrial transcripts in both males and females of R. decussatus , in females of R. philippinarum, and a particularly high abundance in males of R. philippinarum —respectively, 43‐fold, 34‐fold, 12‐fold, and 100‐fold higher than the transcription of nuclear subunits.…”

“…First, dN/dS becomes ineffective when dS reaches saturation—which is a likely event when comparing evolutionarily distant genes (Anisimova & Liberles, ). Second, as shown by Havird and Sloan (), the relationship between mitochondrial and nuclear dN/dS varies dramatically across eukaryotes mainly because of differences in dS. Unfortunately, more sensitive and robust methods (e.g., McDonald–Kreitman test, methods based on linkage disequilibrium) require data not yet available for most organisms, so dN/dS is often the only possible choice.…”

“…In the present work, we investigated the relationship between transcripts abundance and dN/dS in OXPHOS genes of two bivalve species. This is an interesting analysis, since R. philippinarum has a particularly high transcription level of mitochondrial subunits compared to nuclear subunits: up to 100‐fold higher for M‐type, a difference about fivefold higher than in other animals (Havird & Sloan, ; Nabholz et al., ). Therefore, according to the E‐R correlation hypothesis, a much lower dN/dS should characterize mitochondrial subunits of R. philippinarum .…”

“…Second, because mitochondrial genomes in animals tend to accumulate mutations from nine to 25 times faster than the nuclear genome (Lynch, Koskella, & Schaack, ), positive selection would act on nuclear OXPHOS subunits to compensate the insurgence of mutations in mitochondrial genes, ensuring a proper structural/functional match among the subunits of OXPHOS complexes. This theory, called the “nuclear compensation hypothesis”, was adopted by several authors (Aanen, Spelbrink, & Beekman, ; Burton & Barreto, ; Burton, Ellison, & Harrison, ; Dowling, Friberg, & Lindell, ; Havird & Sloan, ; Osada & Akashi, ). Third, Nabholz et al.…”

Targeted capture of complete coding regions across divergent species

“…2014; LoVerso and Cui 2015; Yang et al. 2015; Havird and Sloan 2016; Phillips et al. 2016; Wu et al.…”