Population structure, connectivity, and demographic history of an apex marine predator, the bull shark Carcharhinus leucas

Abstract

Knowledge of population structure, connectivity, and effective population size remains limited for many marine apex predators, including the bull shark Carcharhinus leucas. This large-bodied coastal shark is distributed worldwide in warm temperate and tropical waters, and uses estuaries and rivers as nurseries. As an apex predator, the bull shark likely plays a vital ecological role within marine food webs, but is at risk due to inshore habitat degradation and various fishing pressures. We investigated the bull shark's global population structure and demographic history by analyzing the genetic diversity of 370 individuals from 11 different locations using 25 microsatellite loci and three mitochondrial genes (CR, nd4, and cytb). Both types of markers revealed clustering between sharks from the Western Atlantic and those from the Western Pacific and the Western Indian Ocean, with no contemporary gene flow. Microsatellite data suggested low differentiation between the Western Indian Ocean and the Western Pacific, but substantial differentiation was found using mitochondrial DNA. Integrating information from both types of markers and using Bayesian computation with a random forest procedure (ABC-RF), this discordance was found to be due to a complete lack of contemporary gene flow. High genetic connectivity was found both within the Western Indian Ocean and within the Western Pacific. In conclusion, these results suggest important structuring of bull shark populations globally with important gene flow occurring along coastlines, highlighting the need for management and conservation plans on regional scales rather than oceanic basin scale.

Keywords: ABC‐RF; microsatellite DNA; mitochondrial DNA; mito‐nuclear discordance; population genetics.

Figure 1

Map of bull shark (Carcharhinus leucas) sampling locations (ZAN, Zanzibar; SEY, Seychelles; MOZ, Mozambique; SAF, South Africa; MAD, Madagascar; RUN, Reunion Island; ROD, Rodrigues Island; AUS1, Clarence River, Australia; AUS2, Sydney Harbour, Australia; NCA, New Caledonia; FLO, Florida). Sample sizes are in brackets. Boxes indicate ocean basins and dotted lines delineate regions

Figure 2

Graphical representations of the four scenarios depicting possible divergence histories for each pair of bull shark populations: FLO‐RUN, FLO‐AUS1, and RUN‐AUS1. The time was measured backward in generations before present. In black, is represented the ancestral population of effective population size N _anc; in dark gray, population 1 of effective population size N ₁ and in light gray, population 2 of effective population size N ₂. Double arrows represent bidirectional migration events. t ₂, time of divergence; t ₁, start and end of the isolation period for Scenario 2 and Scenario 4, respectively

Figure 3

Maximum clade credibility tree of the mitochondrial concatenated sequence CR‐nd4‐cytb for the bull shark. Only the different haplotypes are represented. Boxes delineate lineages discussed in the text. Below branches, are indicated node supports above 0.90; above branches, are indicated the mean divergence dates (in millions years ago; Mya) retrieved using either the time of divergence between Carcharhinus and Sphyrna genera (38 Mya; left) or the closure of the Isthmus of Panama separating Atlantic and Pacific populations (3.1–3.5 Mya; right)

Figure 4

TCS statistical parsimony network of 36 bull sharks mitochondrial concatenated sequence CR‐nd4‐cytb haplotypes. Each circle represents a haplotype and each segment, a mutation. Boxes and the dotted line separating the Western Indian Ocean in two groups demarcate lineages discussed in the text (WIO1/WIO2). Circle size is proportional to the number of individuals harboring each haplotype, and colors correspond to sampling locations (WIO1: ZAN, Zanzibar; SEY, Seychelles; MOZ, Mozambique; SAF, South Africa; MAD, Madagascar; WIO2: RUN, Reunion Island; ROD, Rodrigues Island; AUS1, Clarence River, Australia; AUS2, Sydney Harbour, Australia; NCA, New Caledonia; FLO, Florida)

Figure 5

Average probability of membership (y‐axis) of bull shark individuals (N = 357, x‐axis) using 25 microsatellites, assuming correlated allele frequencies and admixture as performed by Structure with the LOCPRIOR model. Only major modes for K varying from two to three are presented. ZAN, Zanzibar; SEY, Seychelles; MOZ, Mozambique; SAF, South Africa; MAD, Madagascar; RUN, Reunion Island; ROD, Rodrigues Island; AUS1, Clarence River, Australia; AUS2, Sydney Harbour, Australia; NCA, New Caledonia; FLO, Florida

Figure 6

Bull shark scatter plot output from a DAPC from microsatellites using the first and second components (a) all sampling locations kept and (b) removing FLO. Dots represent individuals with sampling locations in colors (ZAN, Zanzibar; SEY, Seychelles; MOZ, Mozambique; SAF, South Africa; MAD, Madagascar; RUN, Reunion Island; ROD, Rodrigues Island; AUS1, Clarence River, Australia; AUS2, Sydney Harbour, Australia; NCA, New Caledonia; FLO, Florida)