An explanation for why mitochondrial barcoding fails
DNA barcoding using sequences from mt-encoded proteins or rRNA works
very well for bilaterian animals, but it is much less effective in
delimiting species boundaries of some other eukaryotic taxa,
particularly plants (Chase et al. 2005; Kress et al. 2005) and fungi (Xu
2016) but also Porifera (sponges) and Anthozoa (corals and sea
anemones)(Huang et al. 2008). Recombination of mt genomes, which is rare
or non-existent in bilaterian animals, slows down or stops selective
sweeps because beneficial alleles can be fixed in a mt genotype
independent of the frequencies of other genes on the mt chromosome
(Charlesworth et al. 1993; Rokas et al. 2003; White et al. 2008). The
hypothesis for the evolution of barcode gaps that I outline in this
paper, therefore, provides testable hypotheses for why mt DNA barcoding
might fail for some taxa. If the efficacy of barcoding is dependent on
selective sweeps, which in turn is dependent on lack of recombination of
mt genomes, then it follows that taxa with recombination of mt genes
will have a poor mt DNA barcode signal. Intriguingly, the mt genomes of
Porifera and Anthozoa, for which mt DNA barcoding also works poorly,
include introns, have very low mutation rates, and likely engage in
recombination (Gissi et al. 2008; Huang et al. 2008; Brockman and
McFadden 2012). Recombination of mitochondrial DNA has been documented
in some plants and fungi (Barr et al. 2005), but the scope of
recombination across these eukaryotic groups remains poorly known. For
plants, the extent of recombination and the potential for selection
sweeps is likely irrelevant to a failure of an effective mt DNA
barcode—the rates of nucleotide substitution in plants (with some
exceptions) is far lower than in other eukaryotic taxa, leaving little
opportunity for the evolution of species-specific mt genotypes (Cowan et
al. 2006). A broad-scale comparison of the efficacy of mt DNA barcoding
in relation to rates of recombination and nucleotide substitution of mt
DNA could be very illuminating.
Rampant introgression of mt genomes, wherein the mitochondrial genotype
of one species replaces the mt genotype of another species with little
change to N genotypes, will also erase a barcode signal (Toews and
Brelsford 2012; Hill 2019b). Such mt introgression is hypothesized to
occur when (1) the fitness gain from a better adapted heterospecific
mitochondrion compensate for fitness losses from mitonuclear
incompatibilities, (2) escape from mutational erosion and loss of mt
function compensate for loss of mitonuclear incompatibilities, or (3)
when a maternally transmitted parasite like Wolbachia infects a
new host species and, because it is co-transmitted with mitochondria,
causes the spread the mt genotype of the original host species in the
new host species (Sloan et al. 2017; Hill 2019b). The effects of
endosymbionts may be particularly problematic for the persistence of a
mt DNA barcode gap because endosymbionts can drag mitochondria across a
species boundary and could be an explanation for why phenotypically
distinct populations of animals like blowflies (Diptera: Calliphoridae)
which have high rates of infection by endosymbionts often share a
mitochondrial genotype (Whitworth et al. 2007). Loss of a uniquely
coadapted mitonuclear genotype could be viewed as loss of species
identity such that a lack of a DNA barcode gap in cases of rampant mt
introgression is correctly failing to diagnose a collapsed species
(Vonlanthen et al. 2012). Such an argument carries a risk of
circularity, but the congruence between mt DNA barcode gaps and both
conventional species boundaries (Hebert et al. 2003b; Tavares and Baker
2008) and the transitions in ornamentation used during mate choice for
species recognition (Hill 2018) establishes a clear link between
transitions in mitochondrial genotype and transitions between
populations that taxonomists have recognized as species. The cases of
rampant introgression of mt genomes then become rare exceptions that can
be explained within the context of the mitonuclear compatibility species
concept (Hill 2019b).