Historical biogeography of the leopard (Panthera pardus) and its extinct Eurasian populations

Abstract

Background: Resolving the historical biogeography of the leopard (Panthera pardus) is a complex issue, because patterns inferred from fossils and from molecular data lack congruence. Fossil evidence supports an African origin, and suggests that leopards were already present in Eurasia during the Early Pleistocene. Analysis of DNA sequences however, suggests a more recent, Middle Pleistocene shared ancestry of Asian and African leopards. These contrasting patterns led researchers to propose a two-stage hypothesis of leopard dispersal out of Africa: an initial Early Pleistocene colonisation of Asia and a subsequent replacement by a second colonisation wave during the Middle Pleistocene. The status of Late Pleistocene European leopards within this scenario is unclear: were these populations remnants of the first dispersal, or do the last surviving European leopards share more recent ancestry with their African counterparts?

Results: In this study, we generate and analyse mitogenome sequences from historical samples that span the entire modern leopard distribution, as well as from Late Pleistocene remains. We find a deep bifurcation between African and Eurasian mitochondrial lineages (~ 710 Ka), with the European ancient samples as sister to all Asian lineages (~ 483 Ka). The modern and historical mainland Asian lineages share a relatively recent common ancestor (~ 122 Ka), and we find one Javan sample nested within these.

Conclusions: The phylogenetic placement of the ancient European leopard as sister group to Asian leopards suggests that these populations originate from the same out-of-Africa dispersal which founded the Asian lineages. The coalescence time found for the mitochondrial lineages aligns well with the earliest undisputed fossils in Eurasia, and thus encourages a re-evaluation of the identification of the much older putative leopard fossils from the region. The relatively recent ancestry of all mainland Asian leopard lineages suggests that these populations underwent a severe population bottleneck during the Pleistocene. Finally, although only based on a single sample, the unexpected phylogenetic placement of the Javan leopard could be interpreted as evidence for exchange of mitochondrial lineages between Java and mainland Asia, calling for further investigation into the evolutionary history of this subspecies.

Keywords: Ancient DNA; Hybridisation capture; Leopards; Mitochondrial genomes; Mitogenomes; Palaeogenetics; Panthera pardus; mtDNA.

Fig. 1

Map indicating the location of the samples included in our study, with the approximate historical and ancient (Pleistocene) distribution of the leopard (adapted from Uphyrkina et al., [19]; Diedrich [6]). The current distribution of leopards is severely reduced compared to the historical range, and highly fragmented [7]. Sample PP3 is not displayed due to its ambiguous provenance (“East Indies”)

Fig. 2

Calibrated mitogenomic phylogeny of 25 leopard mitogenome sequences. Node support is indicated by Bayesian Posterior Probabilities, blue node bars indicate the 95% credibility interval of divergence times. The lower axis shows the estimated coalescence times in thousands of years. Colours indicate the locality of the samples; the unexpected placement of the Javan leopard is highlighted dark yellow. The three-letter code corresponds to the putative subspecies for each individual, following Miththapala et al. [25]; Uphyrkina et al., [19]; Diedrich [6]. Asterisks indicate the Late Pleistocene samples. The RAxML maximum likelihood phylogeny can be found in Additional file 1: Figure S1

Fig. 3

Proposed phylogeographical history for the leopard (Panthera pardus), comparing the genetic data with the fossil record. Shaded areas indicate the age of the fossil record for P. pardus (grey) for each region, and the disputed Asian leopard-like Panthera fossils (red). The solid lines indicate the relationships between the mitogenome lineages recovered in this study