Deep intra-island divergence of a montane forest endemic: phylogeography of the Puerto Rican frog Eleutherodactylus portoricensis (Anura: Eleutherodactylidae)

Abstract

Aim Hypotheses proposed for lineage diversification of tropical montane species have rarely been tested within oceanic islands. Our goal was to understand how basin barriers and Pleistocene climatic fluctuations shaped the distribution of diversity in Eleutherodactylus portoricensis (Eleutherodactylidae), a frog endemic to the montane rain forests of Puerto Rico.Location The north-eastern (Luquillo) and south-eastern (Cayey) mountains of Puerto Rico.Methods We generated mitochondrial DNA (mtDNA) control region sequences (c. 565 bp) from 144 individuals of E. portoricensis representing 16 localities, and sequenced 646 bp of cytochrome b and 596 bp of nuclear DNA (nDNA) rhodopsin exon and intron 1 from a subset of individuals. We conducted a phylogenetic analysis on the mtDNA sequence data and explored population substructure with maximum parsimony networks, a spatial analysis of molecular variance, and pairwise F ST analysis. Coalescent simulations were performed to test alternative models of population divergence in response to late Pleistocene interglacial periods. Historical demography was assessed through coalescent analyses and Bayesian skyline plots. ResultsWe found: (1) two highly divergent groups associated with the disjunct Luquillo and Cayey Mountains, respectively; (2) a shallow mtDNA genetic discontinuity across the La Plata Basin within the Cayey Mountains; (3) phylogeographic congruence between nDNA and mtDNA markers; (4) divergence dates for both mtDNA and nDNA pre-dating the Holocene interglacial (c. 10 ka), and nDNA suggesting divergence in the penultimate interglacial (c. 245 ka); and (5) historical demographic stability in both lineages. Main conclusionsThe low-elevation Caguas Basin is a long-term barrier to gene flow between the two montane frog populations. Measures of genetic diversity for mtDNA were similar in both lineages, but lower nDNA diversity in the Luquillo Mountains lineage suggests infrequent dispersal between the two mountain ranges and colonization by a low-diversity founder population. Population divergence began prior to the Holocene interglacial. Stable population sizes over time indicate a lack of demonstrable demographic response to climatic changes during the last glacial period. This study highlights the importance of topographic complexity in promoting within-island vicariant speciation in the Greater Antilles, and indicates long-term persistence and lineage diversification despite late Pleistocene climatic oscillations.

Sea level, topography and island diversity: phylogeography of the Puerto Rican Red‐eyed Coquí, Eleutherodactylus antillensis

“…; Barker et al . ), and likely shaped genetic diversity in E. antillensis by providing mesic refugia during particularly arid glacial periods. Most well‐supported clades in the E. antillensis nuDNA phylogenies are comprised of individuals from Puerto Rico (Fig.…”

“…We adjusted the upper boundary for the prior distribution for effective population size parameters to the geographic size of eastern Puerto Rico and the Eastern Islands relative to western Puerto Rico because genetic variation may be lower in smaller populations (Frankham ), and the values are reasonable with respect to the estimated density of adult male E. antillensis per hectare on Guana Island (Ovaska ) and the large effective population sizes attained by other eleutherodactyline frogs (Crawford ; Barker et al . ). The divergence model for the Ancient Eastern Islands Isolation Hypothesis depicts the Eastern Islands group diverging from western and eastern Puerto Rican groups between the start of the penultimate interglacial and the early middle Pleistocene ( t 1 , prior distribution: 250 000–500 000 generations), and the western and eastern Puerto Rican groups diverging between the Holocene interglacial and the early middle Pleistocene ( t 2 , prior distribution: 1–500 000 generations; Fig.…”

“…This estimate is refined from an earlier study of genetic diversification in E. portoricensis (Barker et al . ). In two E. antillensis individuals for which both CR and cyt b sequences were available, CR had approximately three times as much variation as cyt b , and therefore we believe that a rate of 4.1 × 10 −8 sub/s/y for CR is appropriate for this species.…”

“…Because this rate was estimated from coding regions of mtDNA, which may evolve more slowly than CR, we estimated an upper boundary for CR by inferring its rate relative to a coding mtDNA locus, cytochrome b (cyt b ), using sequences collected from E. portoricensis (Barker et al . ) (cyt b accession numbers—–; CR accession numbers—, , , , , , , , , , , , , , , , , , , , , , , , , , ‐, , , , , , , ). Mutation rates of a particular marker can vary across species (Nabholz et al .…”

“…; Barker et al . ), with sea level fluctuations promoting speciation in the Eastern Islands (Brandley & de Queiroz ; Oneal et al . ).…”