Complete Columbian mammoth mitogenome suggests interbreeding with woolly mammoths

Abstract

Background: Late Pleistocene North America hosted at least two divergent and ecologically distinct species of mammoth: the periglacial woolly mammoth (Mammuthus primigenius) and the subglacial Columbian mammoth (Mammuthus columbi). To date, mammoth genetic research has been entirely restricted to woolly mammoths, rendering their genetic evolution difficult to contextualize within broader Pleistocene paleoecology and biogeography. Here, we take an interspecific approach to clarifying mammoth phylogeny by targeting Columbian mammoth remains for mitogenomic sequencing.

Results: We sequenced the first complete mitochondrial genome of a classic Columbian mammoth, as well as the first complete mitochondrial genome of a North American woolly mammoth. Somewhat contrary to conventional paleontological models, which posit that the two species were highly divergent, the M. columbi mitogenome we obtained falls securely within a subclade of endemic North American M. primigenius.

Conclusions: Though limited, our data suggest that the two species interbred at some point in their evolutionary histories. One potential explanation is that woolly mammoth haplotypes entered Columbian mammoth populations via introgression at subglacial ecotones, a scenario with compelling parallels in extant elephants and consistent with certain regional paleontological observations. This highlights the need for multi-genomic data to sufficiently characterize mammoth evolutionary history. Our results demonstrate that the use of next-generation sequencing technologies holds promise in obtaining such data, even from non-cave, non-permafrost Pleistocene depositional contexts.

Figure 1

Mammoth mitochondrial DNA cladograms. (a) WM lineages (blue) are summarized from previous studies [9-11] with clades indicated and haplogroups labeled at the tips. Hypothetical CM lineage positions (green) are expected positions derived from strict interpretations of paleontological models that posit the two species were separate since the early Pleistocene. The multiple node positions reflect the general uncertainty surrounding the chronology and identity of the WM lineage common ancestor. The position of WM haplogroup B is poorly resolved, exhibiting deep common ancestry with the other haplogroups. Haplogroups A and C are endemic to Eurasia and North America, respectively; haplogroups B, D, and E occur on both continents. Radiocarbon chronologies indicate that haplogroup A went extinct approximately 35,000 ¹⁴Cya, and clade I by approximately 3,200 ¹⁴Cya. Calculated MRCA ages for all nodes yield wide confidence intervals. (b) Our estimated mtDNA cladograms of haplogroup C are depicted using two datasets: the black cladogram and associated scale and posterior probabilities (parameter set 1b, Figure S4 in Additional file 3) are estimated from 743 bp for which several dozen mammoths have been sequenced, whereas the red cladogram and associated scale and posterior probabilities (parameter set 4b, Figure S8 in Additional file 3) are estimated from full mitochondrial genomes, for which only one other haplogroup C mammoth has been sequenced. Each tip in the black cladogram represents a haplotype. M. columbi (haplotype C32) as represented by the Huntington Mammoth is indicated with a yellow star. Scale units are substitutions per site.

Figure 2

Schematic representation of elephantid mtDNA phylogenies under introgression scenarios. (a,b) Hypothetical mammoth (b) (this study) and observed African elephant (a) [38] cladograms, with male body size comparisons and predominant geographic ranges of the species indicated. Solid lines represent observed data; dashed lines represent predicted but presently unobserved lineages under an M. primigenius-M. columbi introgression hypothesis.