More on Amerindian mtDNA haplogroups X2, B2 and C4: Evidence from Siberian Tubalars, Tuvans, Evens and Ulchi

American Journal of Physical Anthropology Vol 148, No. 1, pages 123-138, May 2012 DOI: 10.1002/ajpa.22050

Mitochondrial Genome Diversity in the Tubalar, Even and Ulchi: Contribution to Prehistory of Native Siberians and Their Affinities to Native Americans

Sukernik, Rem I., Natalia V. Volodko, Ilya O. Mazunin, Nikolai P. Eltsov, Stanislav V. Dryomov, and Elena B. Starikovskaya.

To fill remaining gaps in mitochondrial DNA diversity in the least surveyed eastern and western flanks of Siberia, 391 mtDNA samples (144 Tubalar from Altai, 87 Even from northeastern Siberia, and 160 Ulchi from the Russian Far East) were characterized via high-resolution restriction fragment length polymorphism/single nucleotide polymorphisms analysis. The subhaplogroup structure was extended through complete sequencing of 67 mtDNA samples selected from these and other related native Siberians. Specifically, we have focused on the evolutionary histories of the derivatives of M and N haplogroups, putatively reflecting different phases of settling Siberia by early modern humans. Population history and phylogeography of the resulting mtDNA genomes, combined with those from previously published data sets, revealed a wide range of tribal- and region-specific mtDNA haplotypes that emerged or diversified in Siberia before or after the last glacial maximum, ∼18 kya. Spatial distribution and ages of the “east” and “west” Eurasian mtDNA haploclusters suggest that anatomically modern humans that originally colonized Altai derived from macrohaplogroup N and came from Southwest Asia around 38,000 years ago. The derivatives of macrohaplogroup M, which largely emerged or diversified within the Russian Far East, came along with subsequent migrations to West Siberia millennia later. The last glacial maximum played a critical role in the timing and character of the settlement of the Siberian subcontinent.


This paper is another mtDNA survey of a couple of western and eastern Siberian populations. A section of it is devoted to the “peopling of the Americas.” There are a couple of data points that deserve a mention. First, Sukernik et al confirm that hg X detected in South Siberia represents a recent backflow from the west. They write (p. 132),

“The haplogroup X sequence we revealed in a sole Tubalar and one Russian mtDNA samples seem to be attributed to the European X2e2 sub-branch. From the Druze and Georgian X2e2, defined by 3948 and 12084 mutations (Reidla et al., 2003; Shlush et al., 2008), the Tubalar differed at 13327. The age of Siberian X2e2- 13327 lineage calculated on the basis of five similar mtDNA sequences, one Tubalar, one Teleut, two Altai-kizhi, and one Buryat (the latter four being attested by Derenko et al., 2007) is at most ~1.5 kya (Table 4). It is obvious that Siberian X2e2-13327 sub-cluster, separated from the Near Eastern X2e root by three mutational steps (3948-12084-13327), represents a portion of relatively recent gene flow toward Altai-Sayan.”

This means that the huge geographic gap between North American X2a and X2g and the Near Eastern versions of X2 remains unfilled. It’s noteworthy that Western Siberians show an appreciable portion of West Eurasian hg U (U2, U4, U5, U7), with U2e1 apparently of Upper Paleolithic origin, as U2 in Kostenki (~30,000) indicates, but no hg X of any considerable antiquity. U is a powerful West Eurasian clade within macrohaplogroup R. American Indians have representatives from East Eurasian R, namely B2, but no traces of West Eurasian U. On the other hand, American Indians have hg X with West Eurasian connections, which is part of macrohaplogroup N (traditionally, a parent of macrohaplogroup R). Non-U West Eurasian lineages are found in Siberia, e.g. Tuvans have 2.1% of W (highest frequencies in northern Pakistan) and Kets 2.6% of N2a (Sukernik et al. 2012, Table 2). These must be ancient relics associated with small-size foraging demes and, judging by their frequencies and the phylogenetic position, the closest Siberian parallels to American Indian X2 lineages.

Second, with respect to hg B, Sukernik et al. detected (p. 131-132) a sister-like relationship between American Indian hg B2 and Tubalar B4b1a:

“Of special interest are the Tubalar B4b1a samples (6.2%) that have 16086 variant, resulting in the 16136- 16189-16217-16519 motives, previously described in a few ancient and modern mtDNA samples from the Egyin Gol Valley (south of Lake Baikal in northern Mongolia) (Keyser-Tracqui et al., 2003; 2006). Compared at the complete sequence level, the Tubalar B4b1a shared 6023, 6413, and 16136 with only a part of the eastern Asian B4b1 full sequences albeit with the same ancestral node 499-4820-13590 shared with Native American B2 (Starikovskaya et al., 2005; Hill et al., 2007; this study).”

Finally, Sukernik et al. (p. 135) had something interesting to say about a minor American Indian hg C4c,

“Unceasing comparison of the complete mtDNA sequences from Siberia and the Americas has extended the number of founding haplotypes from Beringia significantly (Volodko et al., 2008; Perego et al., 2010; Kumar et al., 2011). In this regard, it is worth noting relic C4c mtDNA sequences defined in two geographically distant Native American populations. One comes from the Ijka-speaking tribe from Colombia (Tamm et al., 2007), and the other from a sole Shuswap-speaker from British Columbia, Canada (Malhi et al., 2010). However, on the other side of the Bering Strait, an extensive survey of Siberian/Asian samples failed to reveal C4c, whereas its sister lineages, C4a and C4b, were widely dispersed in Siberia (Table 2).”

This is another case when Siberia and the New World have a mtDNA gap, rather than a link.

In addition, they report (Table 2, p. 134) a few more versions of the puzzling hg D5 in Tubalars, Mansi, Tuvans, Evenkis, Evens, Yukagirs and Ulchis. Hg D5 was first reported among the Mansi by Derbeneva et al. “Traces of Early Eurasians in the Mansi of Northwest Siberia Revealed by Mitochondrial DNA Analysis,” Am J Hum Gen 70 (2002), 1010-1011. It lacks the classical 10394 DdeI+/10397 AluI- sites, which characterize macrohaplogroup M at the RFLP level. On these grounds alone, D5 is identical with macrohaplogroup N (see the (-/-) sign in Table 1)). On the sequence level it corresponds to a transition A10397G. In addition, it has a back-mutation T16189C, which is a very frequent back-mutation occurring on different branches of hgs D and M. Otherwise, hg D5 shows transition C10400T characteristic of macrohaplogroup M.

This kind of “noise” (see also here on hg B and 9bp deletion) casts doubts on the robustness of existing mtDNA phylogenies. One may pose a question whether macrohaplogroup M is in fact a subset of macrohaplogroup N (just like macrohaplogroup R is), with hg D5 being the first branch between N and M, with all other hg D lineages opposing all other hg M lineages below the D5 node. This possibility is consistent with the younger dates obtained by Behar et al. 2012 for M (49,590 YBP) vs. N (58,859 YBP) and even R (56,523 YBP). Geographically, M is already a subset of N, as N lineages are found on every continent, while M lineages seem to be missing from West Eurasia.

Update: 08.28.12. Gisele Horvat (pers.comm.) has put one D5 sequence (FJ147320 Tubalar, not in PhyloTree) through a couple of mutation-identifying programs (e.g., Ian Logan’s one) and reported that the two character states typical for macrohaplogroup M (10394 DdeI+/10397 AluI-) are not missing on this D5 sequence. However, Volodko et al. (“Mitochondrial Genome Diversity in Arctic Siberians, with Particular Reference to the Evolutionary History of Beringia and Pleistocenic Peopling of the Americas,” American Journal of Human Genetics 82, 2008, 1092) report the same lack of mhg restriction sites in their 5 Yukaghir lineages classified as hg D5a1. Transition A10397G (equals “-/-” at 10394 DdeI+/10397 AluI-) is recorded as part of the D5 definition in PhyloTree.

It is therefore possible that Sukernik et al. (2012) admitted a human error in their RFLP work. And it is true that peer review may not be enough of a fine filter to catch these errors. Thanks go to Gisele for bringing it up, but we will have to await further developments to confirm if hg D5 has indeed a non-M RFLP signature on it.