Limited contemporary gene flow and high self-replenishment drives peripheral isolation in an endemic coral reef fish

Abstract

Extensive ongoing degradation of coral reef habitats worldwide has lead to declines in abundance of coral reef fishes and local extinction of some species. Those most vulnerable are ecological specialists and endemic species. Determining connectivity between locations is vital to understanding recovery and long-term persistence of these species following local extinction. This study explored population connectivity in the ecologically-specialized endemic three-striped butterflyfish (Chaetodon tricinctus) using mt and msatDNA (nuclear microsatellites) to distinguish evolutionary versus contemporary gene flow, estimate self-replenishment and measure genetic diversity among locations at the remote Australian offshore coral reefs of Middleton Reef (MR), Elizabeth Reef (ER), Lord Howe Island (LHI), and Norfolk Island (NI). Mt and msatDNA suggested genetic differentiation of the most peripheral location (NI) from the remaining three locations (MR, ER, LHI). Despite high levels of mtDNA gene flow, there is limited msatDNA gene flow with evidence of high levels of self-replenishment (≥76%) at all four locations. Taken together, this suggests prolonged population recovery times following population declines. The peripheral population (NI) is most vulnerable to local extinction due to its relative isolation, extreme levels of self-replenishment (95%), and low contemporary abundance.

Keywords: Chaetodon; Lord Howe Island; Norfolk Island; coral reefs; extinction risk; marine dispersal.

Figure 1

Location maps and focal species. (A) Google Earth image of eastern Australia showing Middleton Reef (MR), Elizabeth Reef (ER), Lord Howe Island (LHI), and Norfolk Island (NI) in the South West Pacific Ocean. (B) Chaetodon tricinctus swimming in the open (Photo courtesy of Justin Gilligan). Aerial photographs of MR (C); ER (D), LHI (E), and NI (F; the bay measures 1 km in length)

Figure 2

mt and msatDNA Chaetodon tricinctus analyses. (A) A phylogram of mtDNA (cyt b) sequences from 97 C. tricinctus individuals from Middleton Reef, Elizabeth Reef, Lord Howe Island, and Norfolk Island. This represents the best ML tree from 10 individual analyses. Numbers on branches indicate support for each clade, based on ML, MP, MB, and BEAST analyses. (B) Haplotype minimum spanning tree (MST) with number of substitutions between haplotypes indicated on connectors. Different fills represent each of the four locations as shown on the key to the figure, and (C) Scatterplots of the discriminant analysis of principal components of the microsatellite data for four C. tricinctus locations using geographic sample site as priors for genetic clusters. Individual genotypes appear as dots surrounded by 95% inertia ellipses. Eigenvalues show the amount of genetic information contained in each successive principal component with X and Y axes constituting the first two principle components, respectively. Boxes indicate haplotype (h), nucleotide (% π), and genetic diversity (gd) indices for C. tricinctus.

Figure 3

Posterior probability of assignment of each individual genotype to four Chaetodon tricinctus populations as indicated by DAPC. The names of the possible assignment populations are given on the x-axis. 108 genotypes are listed on the y-axis, along with the population from which they were sampled. Colored bars correspond to a 0.2–0.8 probability of assignment to a given population.

Figure 4

Migration rates among Chaetodon tricinctus locations. The thickness of the line is directionally proportional to the number of migrants (M) and the line colors indicate the predominant direction of gene flow. Population size (θ, within parentheses) is also shown for each location. (A) Migrate-n evolutionary gene flow (mtDNA), (B) Migrate-n contemporary gene flow (msatDNA), and (C) BAYESASS analysis of self-replenishment and recent migration rates (msatDNA) shown as a percentage.