Change in mt DNA Genotype via Selection
An alternative hypothesis to both neutral drift and demographic bottlenecks for the generation of mt DNA barcode gaps between species is directional selection on mt genotypes. Natural selection has the potential to shape the mt genome in response to two distinct environments: the external environment (both biotic and abiotic) and the internal genomic environment created by the N genome (Rand et al. 2004; Zhu et al. 2014; Barreto et al. 2018; Sloan et al. 2018; Hill 2019a). There is now overwhelming evidence that the mt genome of at least some animal lineages—and likely all animal lineages—are subject to periods of directional selection as adaptive responses to the external environment (Dowling et al. 2008; Ballard and Pichaud 2014; Kazancioǧlu and Arnqvist 2014). In particular, thermal and chemical environments, oxygen pressure, diet, salinity, and UV exposure can all exert natural selection on the mt genome and lead to adaptive changes in protein-coding genes (Ballard and Pichaud 2014; Hill 2019a). The adaptive evolution of mt genomes in response to external environments is now a major research topic in evolutionary biology (Sunnucks et al. 2017; Hill et al. 2019), and such changes to the nucleotide sequence of mitochondria in response to directional selection pose a serious challenge to core arguments for why mt DNA sequences will often fail as a tool for diagnosing species (Hickerson et al. 2006). Adaptive divergence of mt genotype in response to external environment is a key reason why mt DNA is predicted to rapidly diverge between allopatric populations (Gershoni et al. 2009; Tobler et al. 2019).
Perhaps even more important, and certainly more pervasive, than changes to mt DNA gene sequence in response to external environment is the potential for perpetual evolutionary change in the mt DNA in response to changes in the internal genomic environment (Chou and Leu 2010; Burton and Barreto 2012; Barreto et al. 2018; Sloan et al. 2018; Hill 2020). The coadaptation of gene complexes is a foundational concept in evolutionary biology (Dobzhansky 1937; Wright 1942). In a discussion of the evolution of mt genomes, however, it is essential to grasp that there are unique features to the co-evolution and coadaptation between mt gene products and the products of a small list of N genes that code for products that function in intimate interaction with mt gene products (N-mt genes) (Hill 2019a; Shtolz and Mishmar 2019). First, the system that depends on coadaptation of mt and N-mt genes—the electron transport system—is the most critical biochemical system in the bodies of eukaryotes that depend on energy from aerobic respiration (Wallace 2010; Lane 2014). Second, because of the complexity of the ETS in controlling the flow of electrons and pumping of protons, very small changes to interacting components can have huge fitness effects (Lane 2011; Sloan et al. 2018; Hill 2019a; Hill et al. 2019). Third, mitonuclear coadaptation involves two genomes that can potentially undergo independent evolution (Rand et al. 2004; Gershoni et al. 2014; Wolff et al. 2014). Fourth and finally, the mt genome of animals does not generally engage in recombination (Barr et al. 2005) and so mitochondrial genes form one linkage group such that selection on one mt gene can affect the frequencies of other mt genes (Meiklejohn et al. 2007; Oliveira et al. 2008). Functional divergence in mt DNA will be particularly effective in creating Dobzhansky-Muller incompatibilities in hybrid offspring and hence in establishing barriers to gene flow because the mt DNA must maintain tight coadaptation with the N genome (Burton and Barreto 2012; Hill 2017).
If changes in mt genotype between species were entirely neutral, then matching the N genes of one species with the mt genes of a closely related species—either through hybridization or in cell culture by directly manipulating genomes—should result in no change in mitochondrial function in the resulting cells or organisms. Indeed, this logical extension of the neutral theory of mitochondrial evolution led to a failed research program to propagate endangered species by pairing mitochondria of donor species to the N genome of the species to be saved (Lanza et al. 2000). Observations from cybrid and hybrid studies, however, clearly established that, once sets of mt and N-mt genes diverge in nucleotide sequences to the extent seen in sister species, incompatibilities in non-coadapted gene sets cause a reduction in mitochondrial function when they are forced to work together (reviewed in Hill (2019a)). Mitonuclear incompatibilities in cybrid cells and hybrid organisms is strong evidence that the evolution of mt genotypes is not neutral with respect to the genomic environment (Barrientos et al. 1998; Ellison and Burton 2008b; Lee et al. 2008; Garvin et al. 2011; Latorre-Pellicer et al. 2016).
The evolution of uniquely coadapted mt and N-mt genotypes is a critical concept because it potentially explains both how the mt genotypes of sister species rapidly diverge and why there is so little introgression of divergent mt genotypes between species within most clades of bilaterian animals (Burton and Barreto 2012; Hill 2016). The evolution of a clean mt DNA barcode gap requires that the propagation of population-specific mitochondrial genotypes are constrained to remain within species boundaries across generations (Hebert et al. 2003b). Even a small amount of introgressive flow of mitochondrial genotypes, which would be inevitable under neutral models of mitochondrial evolution if species lived in sympatry, would add unacceptable ambiguity into barcoding efforts (Papadopoulou et al. 2008). In the rare cases in which mitochondria do introgress across species boundaries, the introgression tends to be rampant, with complete replacement of one mitochondrial genotype by another (Hill 2019b). All of these patterns are consistent with a process whereby coevolution of mt and N-mt genotypes leads to loss of fitness (at the level of the individual organism) when mt genotypes are paired to N-mt genes to which they are not coadapted. The barcode gap is more than an arbitrary marker of species boundaries—it is the functional boundary that reinforces the uniqueness of a species’ mitonuclear genotype (Lane 2009a; Chou and Leu 2010; Burton and Barreto 2012; Hill 2016, 2017).