The selective neutrality of mitochondrial DNA variation is increasingly refuted based on data from natural populations and model organisms. Instead, mitochondrial and associated nuclear (mitonuclear) variation emerge as drivers of adaptive evolution and divergence. Species with discordant mitonuclear geographic patterns inexplicable by commonly advocated selectively-neutral causes are powerful systems for studying mitonuclear fitness, adaptation and speciation.
The Eastern Yellow Robin is a common, widespread Australian bird. Populations carrying mtDNA of either an inland or a coastal haplogroup occupy different climates, show interspecific-level mitochondrial DNA difference, yet exchange many neutral nuclear genes. To examine mitonuclear evolution in this species, we sequenced whole mitogenomes, ~1000 nuclear sequence markers, and screened >60,000 genomewide, short-read sequence markers. We tested for selection, derived coalescent estimates of divergence times, gene flow and effective population sizes, conducted genome-scale analyses of population differentiation and linkage disequilibrium, and mapped the short-read markers to a reference genome to ascribe genome position and putative gene functions.
Reconstruction of population history indicated allopatric separation ~2,000,000 years ago, then secondary contact ~100,000 years ago. Current mitonuclear discordance can be explained by two separate adaptive introgressions of mitochondrial DNA. Inland and coastal populations have generally similar nuclear genomes but substantial difference in mitochondrial DNA and a subset of nuclear genes. Remarkably, most of these nuclear genes map to two main locations in the reference genome: a 16 megabase region of autosome 1A, and a smaller region of the Z sex chromosome. The inferred genes in 1A disproportionately have mitonuclear functions and experience very low recombination. Thus 1A acts as a ‘supergene’ of co-inherited, functionally-related genes.
We hypothesize that variants of the 1A mitonuclear supergene are co-adapted to their local environments, possibly via metabolic adaptation. The supergene genomic architecture can explain how strong mitonuclear adaptation could survive substantial gene flow.