I was reading through an excellent review of the recent literature about mtDNA and selection, from Damian Dowling and colleagues (2008). The review focuses on the patterning of evidence for selection in ecological and phylogenetic terms, and to some extent upon the function of mtDNA or the mito-nuclear complex of proteins involved in oxidative metabolism. It includes a long passage covering the significant mismatch between mtDNA variation and effective population sizes across animals (but not mammals). A short section discusses the possibility of adaptive polymorphism maintained by mito-nuclear interactions:
Knowing that deleterious mutations in mtDNA can accumulate within populations because of genetic drift , there certainly seems to be scope for mito-nuclear co-evolution to proceed via a compensatory model. Under this model, deleterious mutations accumulate in the mitochondrial genome, with selection then favouring an adaptive response in the nuclear genome to restore any compromised metabolic function . In effect, mtDNA mutations will act as the drivers of adaptive evolution in nuclear genes. This scenario is not unlikely, given that more than 1000 nuclear-encoded proteins, which are essential for metabolism, are transported into the mitochondrion .
Additionally, given that at least some mtDNA polymorphism might have been shaped via positive selection  and , scope might also exist for mito-nuclear co-evolution to proceed via a model in which adaptive mutations in one genome select for a response in the other.
There has been recent interest in the coinheritance of sex chromosomes and mtDNA. Because the sex-determining chromosome is opposite in birds from mammals, a number of natural experiments may be available to examine the role of coevolution for the mtDNA and co-inherited sex chromosomes. Further, a number of studies have identified a substantial cytoplasmic contribution to fitness and lifespan variance in Drosophila, suggesting that adaptive variation in mtDNA may be segregating within populations.
The review discusses the possible importance of the adaptive perspective for aspects of biology ranging from life history and aging to speciation (where fast-evolving mtDNA genes may induce hybrid incompatibilities). And sperm are a surprising focus of research -- mtDNA mutations affect motility, fertility, and the outcome of sperm competition. On that topic, more later.
Dowling DK, Friberg U, Lindell J. 2008. Evolutionary implications of non-neutral mitochondrial genetic variation. Trends Ecol Evol 23:546-554. doi:10.1016/j.tree.2008.05.011