Patterns of adaptive and neutral diversity identify the Xiaoxiangling mountains as a refuge for the giant panda

Abstract

Genetic variation plays a significant role in maintaining the evolutionary potential of a species. Comparing the patterns of adaptive and neutral diversity in extant populations is useful for understanding the local adaptations of a species. In this study, we determined the fine-scale genetic structure of 6 extant populations of the giant panda (Ailuropoda melanoleuca) using mtDNA and DNA fingerprints, and then overlaid adaptive variations in 6 functional Aime-MHC class II genes (DRA, DRB3, DQA1, DQA2, DQB1, and DQB2) on this framework. We found that: (1) analysis of the mtDNA and DNA fingerprint-based networks of the 6 populations identified the independent evolutionary histories of the 2 panda subspecies; (2) the basal (ancestral) branches of the fingerprint-based Sichuan-derived network all originated from the smallest Xiaoxiangling (XXL) population, suggesting the status of a glacial refuge in XXL; (3) the MHC variations among the tested populations showed that the XXL population exhibited extraordinary high levels of MHC diversity in allelic richness, which is consistent with the diversity characteristics of a glacial refuge; (4) the phylogenetic tree showed that the basal clades of giant panda DQB sequences were all occupied by XXL-specific sequences, providing evidence for the ancestor-resembling traits of XXL. Finally, we found that the giant panda had many more DQ alleles than DR alleles (33∶13), contrary to other mammals, and that the XXL refuge showed special characteristics in the DQB loci, with 7 DQB members of 9 XXL-unique alleles. Thus, this study identified XXL as a glacial refuge, specifically harboring the most number of primitive DQB alleles.

Figure 1. Distribution of giant panda populations and network relationships among panda mtDNA haplotypes.

Current distribution of extant giant panda populations (A) and network relationships among the panda mtDNA haplotypes (B). The 6 isolated populations are indicated in dark green, according to the most recent survey . Population-scale networks are shown in a and b (QLI, red; MSH, blue; QLA, yellow; DXL, purple; XXL, sky-blue; and LSH, green). The solid circles represent each unique haplotype, with their sizes proportional to their frequency. Empty circles indicate the undetected haplotypes that are necessary to link all observed haplotypes to the network.

Figure 2. MtDNA-based mismatch distributions and Bayesian skyline plot.

MtDNA-based mismatch distributions (A) and Bayesian Skyline Plot (B) for the 2 subspecies and the species as a whole.

Figure 3. DNA fingerprint-based median-joining network.

DNA fingerprint-based median-joining network relationships of the 6 panda populations. The populations from which the individuals were collected are indicated in the same color scheme given in Figure 1B).

Figure 4. Maximum likelihood phylogenetic trees.

Maximum likelihood (ML) phylogenetic relationships of the Aime-MHC class II alpha (A) and beta (B) alleles. Bootstrap values less than 50 (50%) are not shown.