Genetic Analysis of a Cryptic Contact Zone between Mitochondrial Clades of the Eastern Red-Backed Salamander, Plethodon cinereus
When evolutionarily divergent lineages adjoin their geographic ranges after a period of isolation, myriad outcomes can occur, from population anastomosis to the evolution of reproductive isolation by way of reinforcement. Hybrid zones represent natural experiments that may indicate whether lineages will maintain their evolutionary independence. Here, we report on a hybrid zone in the Eastern Red-Backed Salamander, Plethodon cinereus, a highly abundant and wide-ranging terrestrial salamander found in the northeastern United States and in southeastern Canada. An earlier study identified six distinct mitochondrial clades across the range of P. cinereus. Populations of two of these clades were as close as 9.6 km apart in Lorain County, Ohio, USA. To investigate the nature of this contact zone, we sampled 316 individuals from 16 sites along a 53-km transect, and analyzed 10 microsatellite loci and one mitochondrial locus. We found a clinal transition for mtDNA haplotypes. In contrast, most studies of terrestrial plethodontid salamanders commonly exhibit sharp boundaries between mtDNA clades. Microsatellite markers, however, revealed little differentiation and weak population structure, suggesting the nuclear cline, if it exists, lies outside of our sampling region. Explanations for the discordance between the mitochondrial DNA and our microsatellite data include lineage sorting, male-biased dispersal, or historical introgression of mtDNA, among other possibilities. We compare our results with other studies of introgression in terrestrial salamanders, and discuss the causes of mitonuclear discordance.Abstract
(A) Map of the eastern United States and Canada with geographic range of Plethodon cinereus shaded in light gray (IUCN et al., 2015). The asterisk denotes the sampling region of the contact zone in northern Ohio. (B) Sampling region including 16 localities. Pies represent the proportion of individuals of the OH (white) and PA (black) mtDNA clades discovered at each site, and pie size corresponds to sample size. Notable riverine barriers are labeled; locality numbers are positioned to indicate the side of rivers that sites are located. Sites 10 and 12 in bold were sampled for mtDNA by Radomski (2017), and were designated as PA and OH populations, respectively. (C) STRUCTURE results (K = 2) for microsatellite DNA using the admixture model, without incorporating the LOCprior. (D) STRUCTURE results (K = 2) using the admixture model and the LOCprior (LOCISPOP). (E) NewHybrids results using uniform priors. Each individual is a vertical bar, and the y-axis indicates an individual's posterior probability of assignment to each genotype class (summing to 1). Black portions of the bar represent the assignment probability to one of two parent populations. White portions represent backcrosses to this first parent population. Probabilities of assignment to all other genotype classes (parent population 2, F1, F2, backcross to parent population 2) are combined and represented by gray, although these combined posterior probabilities are low (<0.04) and generally not visible.
(A) Discriminant analysis of principle components (DAPC) of microsatellite data when assuming six clusters and retaining 10 PCs and five discriminant functions. BIC values did not demonstrate a clear number of clusters (Supplemental Fig. S4), but other values from K = 3–10 did not provide more (or less) useful summaries of the data. The x- and y-axes represent the first and second discriminant axes (DA), respectively. Ellipses are set to approximate a 95% confidence interval (ellipse = 2.5). Inset displays cumulated variance (%) explained against the number of retained PCs. (B) Scatterplot of the same data as (A), but individuals are labeled as OH (blue) and PA (red) mtDNA clade. Note that mtDNA clade membership was not used in clustering analysis; these results simply display mtDNA clade onto the DAPC of K = 6.
Geographic cline for mtDNA haplotype frequencies. The x-axis represents distance, χ (km), from our westernmost site along the single-coordinate cline axis; the y-axis represents the proportion of individuals at each site assigned to the OH clade. The solid line is the best-fit, right-tailed model, and the surrounding grey region represents the 95% credible cline region. The dashed line represents the simple, no-tailed model. Although the right-tailed model had the lowest AICc score, the 95% credible region overlapped with the simple cline model, and two log-likelihood estimates showed wide variation in parameter estimates for the right-tailed model.
Contributor Notes
