Extensive population genetic structure in the giraffe

Abstract

Background: A central question in the evolutionary diversification of large, widespread, mobile mammals is how substantial differentiation can arise, particularly in the absence of topographic or habitat barriers to dispersal. All extant giraffes (Giraffa camelopardalis) are currently considered to represent a single species classified into multiple subspecies. However, geographic variation in traits such as pelage pattern is clearly evident across the range in sub-Saharan Africa and abrupt transition zones between different pelage types are typically not associated with extrinsic barriers to gene flow, suggesting reproductive isolation.

Results: By analyzing mitochondrial DNA sequences and nuclear microsatellite loci, we show that there are at least six genealogically distinct lineages of giraffe in Africa, with little evidence of interbreeding between them. Some of these lineages appear to be maintained in the absence of contemporary barriers to gene flow, possibly by differences in reproductive timing or pelage-based assortative mating, suggesting that populations usually recognized as subspecies have a long history of reproductive isolation. Further, five of the six putative lineages also contain genetically discrete populations, yielding at least 11 genetically distinct populations.

Conclusion: Such extreme genetic subdivision within a large vertebrate with high dispersal capabilities is unprecedented and exceeds that of any other large African mammal. Our results have significant implications for giraffe conservation, and imply separate in situ and ex situ management, not only of pelage morphs, but also of local populations.

Figure 1

Genetic subdivision in the giraffe based on mitochondrial DNA sequences. (A) Approximate geographic ranges, pelage patterns, and phylogenetic relationships between giraffe subspecies based on mtDNA sequences. Colored dots on the map represent sampling localities (see Additional files 1 and 10 for detailed locality information). The phylogenetic tree is a maximum-likelihood phylogram based on 1707 nucleotides of mtDNA sequence (1143 nt of cytochrome b, 429 nt control region and 135 nt of tRNA) from 266 giraffes. Asterisks along branches correspond to node-support values of > 90% bootstrap support. Stars at branch tips identify paraphyletic haplotypes found in Masai and reticulated giraffes. (B) Minimum-spanning network of control region haplotypes using the molecular-variance parsimony algorithm (see Additional file 8), where circles represent haplotypes, numbers within them correspond to haplotype designations, and circle sizes are proportional to the haplotype's frequency in the population. Branches represent a single nucleotide change and black squares represent multiple changes (indicated by adjacent numbers). Colors are coded as in Figure 1A.

Figure 2

Genetic subdivision in the giraffe based on microsatellites alleles. Neighbor-joining network of allele-sharing distances (D_s) based on 14 microsatellite loci typed in 381 giraffes. Colors are coded as in Figure 1A.

Figure 3

Genetic subdivision among giraffe groups and populations based on Bayesian cluster analysis [23] of 14 microsatellite loci from 381 individuals. Shown are the proportions of individual multilocus genotypes attributable to clusters (K) indicated by different colors. Sample group designations and sampling locations are denoted. We varied K from 2–16 and at least six groups corresponding to currently defined subspecies and 11 geographic clusters are resolved as indicated.