Paleoanthropological materials

The main area of the Tagar culture covers the steppe and forest-steppe zone in the basin of the middle Yenisei River and its tributaries (mainly in modern Khakassia Republic, Russia). Paleoanthropological materials from 95 individuals from 13 burial grounds of the Tagar culture located on the territory of the Minusinsk basin in the central part of the Tagar culture area were obtained (see details on Fig 1, S1 File). Materials from funeral complexes related to the main stages of the development of the Tagar culture, which replaced each other and were distinguished on the basis of archaeological data, are represented, with a predominance of materials from the Early and Middle stages: 46 samples (individuals) belonged to the Early Tagar period (Bainovo, Podgornovo, and Bidjinski types, according to the classification of M.P. Gryaznov) dated 9–6 centuries BC, 24 samples (individuals) belonged to the Middle Tagar period (Saragash type) dated 5–3 centuries BC, and 9 samples (individuals) belonged to the Late Tagar (Tes`type or Tes`culture) dated 2–1 centuries BC (see S1 File for detailed description of samples).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Location of Tagar archaeological sites from which samples for this study were obtained.

Burial grounds: 1—Novaya Chernaya-1; 2—Podgornoe Ozero, Barsuchiha-1, Barsuchiha-6, Barsuchiha-7; 3—Perevozinskiy; 4—Ulug-Kyuzyur, Kichik-Kyuzyur, Sovetskaya Khakassiya; 5—Tepsey-3, Tepsey-8, Tepsey-9; 6—Dolgiy Kurgan.

https://doi.org/10.1371/journal.pone.0204062.g001

Most of paleoanthropological materials from the Tagar population studied in this work were obtained during excavations conducted in the Minusinsk Basin by the Krasnoyarsk Archaeological Expedition of the Leningrad Division of the Institute of Archeology of the USSR Academy of Sciences in 1955–70 (head of the expedition—M.P. Gryaznov) [8]).

All specimens analyzed in this study were obtained from Repository of Paleoanthropological Collections of the Institute of Archaeology and Ethnography, Siberian Branch, Russian Academy of Sciences (IAET SB RAS Novosibirsk, Russia). All materials are available for further studies according to the full description (including the name of archaeological site (burial ground name), number of burial (and skeleton number for double or collective burials)) given in S1 File. All specimens are available for investigation under informal permission from IAET SB RAS. No additional permits were required for the described study. No permits were required for the described study, which complied with all relevant regulations.

Due to favorable climatic conditions of the Minusinsk basin, all the skeletal materials included in the study are characterized by a very high degree of macroscopic preservation. Fragments of long bones with a large layer of compact bone were included in the analysis.

All the ancient DNA experiments were carried out in the Laboratory of molecular paleogenetics of the Institute of Cytology and Genetics SB RAS (Novosibirsk, Russia).

DNA extraction

DNA was extracted as previously described [11, 12]. In brief, surfaces of bones were removed mechanically, and then treated with a 7% bleach solution and irradiated by UV light (for at least 1 hour on each side of each sample). Powder was drilled from the internal compact tissue of bones. DNA was extracted from bone powder by means of incubation in a 5M guanidinium thiocyanate (GuSCN) buffer (pH 8.0) for 48 h at 65°C, followed by phenol/chloroform extraction and precipitation with isopropanol. 2–4 extractions were performed for each individual under study.

Mitochondrial DNA analysis

The mtDNA analysis consisted of two steps. In the first step, mtDNA hyper-variable region (HVR) I was analyzed using two PCR strategies: 1) four overlapping fragments of mtDNA HVRI were amplified [13] and 2) a single long fragment was amplified by nested PCR [14] (S2 File). The HVRI fragments were sequenced directly using an ABI Prism BigDye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Foster City, CA, USA) (versions 1.1. and 3.1 for short and long fragment, respectively). Sequence data were analyzed using an ABI Prism 3130XL Genetic Analyzer (Applied Biosystems, USA) at the SB RAS Genomics Core Facility (Novosibirsk, Russia, www.sequest.niboch.nsc.ru).

Sequences were aligned with the revised Cambridge Reference Sequence of human mtDNA (rCRS) [15] by using BioEdit software version 7.0.5. Preliminary phylogenetic interpretation of sequences (based on HVRI sequences) was performed according to the current classification of mtDNA variability (mtDNA tree Build 17, http://www.phylotree.org) [16] by using the software tool HaploGrep (http://haplogrep.uibk.ac.at/) [17].

After the preliminary phylogenetic analysis of the HVRI data, informative positions in the mtDNA coding regions were analyzed, and haplogroups and sub-haplogroups were identified. Primer pairs from [18] were used to amplify mtDNA coding regions (S2 File). These coding regions were analyzed following the same methods used for HVRI. This combined two-step approach was used to clarify the phylogenetic relationships among samples and to verify the HVRI data.

Phylogenetic networks among the Tagar mtDNA haplotypes were constructed by using the Network program v.4.5.0.0. (www.fluxus-enginering.com). Phylogeographic analysis was conducted by using a database of mtDNA variability in modern Eurasian populations (S1 Table) collected from published sources.

Interpopulation differences of the Tagar population with other ancient and modern populations of Eurasia were estimated by F_ST distances [19] using the program Arlequin v.3.5.1.2 [20]. The significtableance level of Fst distances was assessed using the Monte-Carlo method, the number of interchanges was 100, and the level of significance of P = 0.05. The data of mtDNA haplogroups (and subgroups) frequencies (for interpopulation differences of the Tagar and modern Eurasian populations; S2 Table) and mtDNA HVRI sequences (for interpopulation differences of the Tagar and other ancient populations; S3 and S4 Tables) were used for this purpose. Multidimensional scaling, based on a matrix of pairwise F_ST differences, was carried out using the XLStat program (www.addinsoft.com).

Haplotype sharing analysis between Tagar population and other ancient groups was carry out as in the article [21].

Sex determination and autosomal STR analysis

Nine autosomal STRs and the sex-determining marker amelogenin were simultaneously co-amplified by using the AmpFlSTR Profiler Plus Kit (Applied Biosystems, USA) according to the manufacturer`s instructions. Results (see S7 File) were analyzed on an ABI Prism 3130XL Genetic Analyzer (Applied Biosystems, USA) at the SB RAS Genomics Core Facility (Novosibirsk, Russia, www.sequest.niboch.nsc.ru).

Precautions against contamination

Experimental work with the ancient material was carried out in specially equipped, isolated clean rooms using special clothes. All work surfaces and instruments were routinely cleaned with a 5% solution of bleach and irradiated by UV light. Blank controls were run in parallel with samples in all extraction and amplification procedures, to identify possible contamination. PCR with positive amplification in blanc control were discarded. MtDNA HVRI sequences were determined for all staff working in ancient DNA facility.

Representativity of the obtained results

We analyzed a number of mtDNA samples (N = 79) from representatives of the Tagar culture from the Minusinsk basin (the middle course of the Yenisei River, Sayan Mountains, Southern Siberia). The Tagar culture existed in the Minusinsk basin for a long time, at least from the 8^th to 1^st century BC, i.e., during the pre-Scythian, Scythian, and Early Xiongnu-Sarmatian periods. Only a small number of samples from the Tagar culture (belonging to one chronological period [4] or without indication of the Tagar cultural stage [10]) have been published previously. Our series includes samples from all the main chronological stages of the Tagar culture: Early Tagar, Middle Tagar, and Late Tagar periods (see S1 and S4 Files). These new results place the Tagar culture within the most well-studied (in terms of mtDNA diversity) early nomadic groups from Siberia.

Our work is a good example of the correlation between the number of mtDNA samples and the quality, generalizability and representativity of mtDNA diversity data for an ancient population. More than 74% of HVRI haplotypes and several haplogroups (covering almost 60% of individuals analyzed in this study) found in our Tagar series were absent from previously published data (N = 16) [4, 10].

For a mtDNA series including N samples, under a simple binomial approximation, we cannot rule out the possibility that some undetected mtDNA clusters existed in the population at a frequency p with (1–p)^N = 0.05. For N = 16 (data from [4, 10]), p = 0.171. Thus, previously published Tagar mtDNA data could occasionally not include clusters that were among the most highly represented in the gene pool of this ancient population. In particular, we observed haplogroups T1, U4, and A8, which were absent from previously published data, in our more representative series at frequencies of 13.9%, 10.1%, and 10.1%, respectively. For our series (N = 79), p = 0.037. With a high probability, we identified all of the main clusters of mtDNA present in the gene pool of the Tagar culture population in the Minusinsk basin. Thus, we obtained substantially more representative mtDNA diversity data for the Tagar population than were previously available.

Mitochondrial DNA diversity and genetic relationships of the Tagar population

According to our results, the Tagar populations show a similar mtDNA pool structure to those of other Iron Age populations representing the “Scythian World” (Figs 3 and 4; S3 and S5 Tables). These results are consistent with those of previous results [4] (note that the Tagar group was under-represented in this previous study). When analyzing the whole Tagar series, the greatest similarity was detected between the Tagar and Classic Scythians from the North Pontic region (Figs 3 and 4; S3 and S5 Tables). Interestingly, the close similarity in mtDNA pool structure is in agreement with the results of a craniometric study of these populations; both female and male series from the Tagar population show high craniofacial similarity with several local groups of Classic Scythians from the North Pontic region [63, 64].

It has been suggested that the genetic similarity between the western and eastern early nomadic Scythian-like groups reflects the multiregional (local) origin of these groups with similar genetic components [4]. According to this scenario, West Eurasian components of the mtDNA pool, which are largely responsible for the similarity in genetic composition (with respect to mtDNA) between eastern and western early nomads, were most likely introduced into Southern Siberia by migratory flows from the western part of Eurasia that occurred during the Bronze Age and are well documented based on archaeological data. There were two main suggested migration waves to the Minusinsk basin during the Bronze Age: (1) the migration of populations from Eastern Europe (possibly related to the Yamnaya culture), resulting in the rise of the Afanasievo culture in the Altai and Sayan Mountains during the Early Bronze Age (3^rd millennium BC) [9], and (2) the migration of the Andronovo (Fedorovo) culture population from adjacent regions of the Eurasian steppe (from modern Kazakhstan territory) during the 1^st half of the 2^nd millennium BC (Middle Bronze Age) [9, 65].

Our results are not inconsistent with the assumption of a probable role of gene flow due to the migration from Western Eurasia to the Minusinsk basin in the Bronze Age in the formation of the genetic composition of the Tagar population. Particularly, we detected many mtDNA lineages/clusters with probable West Eurasian origin that were dominant in modern populations of different parts of Europe, Caucasus, and the Near East (such as K and HV6) in our Tagar series based on a phylogeographic analysis.

To assess the potential influence of Bronze Age populations in the Minusinsk basin (and the related migration wave from west of the Eurasian steppe) on the formation of the genetic composition of the Tagar population we compared our Tagar series with several chronologically preceded ancient populations: 1) Early Bronze Age (pre-Andronovo) Afanasievo and Okunevo culture populations from Minusinsk basin and Ealy Bronze Age Yamnaya culture populations from Eastern Europe (since numerous data from archaeology, physical anthropology, and genomic analyses support the hypothesis that the Yamnaya populations were the ancestors of the Afanasievo group); 2) Middle Bronze Age populations, related with Andronovo migration wave during the 1^st half of the 2^nd millennium BC—Andronovo culture population from the Minusinsk basin and Andronovo-time population from the adjacent Baraba forest-steppe region.

We detected relatively low genetic distances between our Tagar population and two Bronze Age populations from the Minusinsk basin—the Okunevo culture population (pre-Andronovo Bronze Age) and Andronovo culture population, followed by Afanasievo population from the Minusinsk Basin and Middle Bronze Age population from the Mongolian Altai Mountains (the region adjacent to the Minusinsk basin) (Figs 3 and 6; S3 and S5 Tables). Among West Eurasian part of our Tagar series we also observed haplogroups/sub-haplogroups and haplotypes shared with Early and Middle Bronze Age populations from Minusinsk Basin and western part of Eurasian steppe belt (Fig 4; S5 Table). Thus, our results suggested a potentially significant role of the genetic components, introduced by migrants from Western Eurasia during the Bronze Age, in the formation of the genetic composition of the Tagar population. It is necessary to note the relatively small size of available mtDNA samples from the Bronze Age populations of Minusinsk basin; accordingly, additional mtDNA data for these populations are required to further confirm our inference.

However, based on our mtDNA data, we cannot exclude the possibility that some West Eurasian clusters of mtDNA were introduced (or reintrodused) to Southern Siberian Iron Age populations more recently, independent of the Bronze Age migration waves discussed above. In this regard, it is interesting to note some specific West Eurasian mtDNA lineages shared between eastern and western populations of the Scythian World. For example, a lineage with rCRS HVRI haplotype belonged to haplogroup U (haplotype number 39 in Table 1; Fig 2) was identified both in the Tagar population of Minusinsk basin (five individuals from two independent studies—this work; [10]) and in Classic Scythians from the North Pontic region [3]. This lineage is absent from the published mtDNA data for Bronze Age populations, both from the territory of Southern Siberia (including the Minusinsk basin) and the central and western regions of the Eurasian steppe belt. The source of this Western Eurasian component in the Tagar population gene pool is unclear.

Another substantial part of the mtDNA pool of the Tagar and other eastern populations of the Scythian World is typical of populations in Southern Siberia and adjacent regions of Central Asia (autochthonous Central Asian mtDNA clusters). Most of these components belong to the East Eurasian cluster of mtDNA haplogroups. Moreover, the role of each of these components in the formation of the genetic composition of subsequent (to the present) populations in South Siberia and Central Asia could be very different. In this regard, cluster C4a2a (and its subcluster C4a2a1), and haplogroup A8 are of particular interest.

Lineages in the C4a2a (and C4a2a1) cluster represent a significant portion of the haplogroup C diversity in our Tagar series (three out of five lineages). Interestingly, in the Early Iron Age, the mtDNA cluster C4a2a (including C4a2a1) did not appear to be widely distributed in Southern Siberia and Central Asia: these lineages were not detected in other populations of Scythian and Xiongnu-Sarmatian times, both in the territory of the Altai-Sayan mountain system and beyond. Only a single mtDNA line belonging to the related cluster C4a2 was detected in the Mongolian Xiongnu [36]. This differs substantially from modern populations: the C4a2a lineages (including mainly C4a2a1) are now a typical component of the mtDNA pool of the modern Turkic-speaking and Mongolian-speaking populations of Central Asia (including Southern Siberia) (S5 File). Therefore, a major increase in the role of this lineage in the genetic composition of Central Asian populations occurred over the past 2000 years. Descendants of the Tagar population (or genetically related populations) could potentially play a substantial role in this process (see more details in S5 File).

Unlike C4a2a, the role of A8 in the gene pool of the population of Southern Siberia has reduced substantially over the past ~2000 years. This cluster was widely represented in Scythian-Siberian groups, including the Tagar, Pazyryk and Aldy Bel culture populations (this study and [4]) (Fig 8; S6 File). In modern Eurasian population haplogroup A8 is a rare component of the mtDNA gene pool and is distributed with low frequencies mainly in Eastern (including the south of Siberia) and Western Central Asia (and it is very rare in Western Eurasia [66]) (Fig 8; S6 File). Interestingly, most variants of A8 found in the ancient Scythian-Siberian populations are characterized by a common haplotype with a C16278T substitution and obviously belong to a subcluster of haplogroup A8 that has not been annotated to date. Conversely, A8 lineages lacking С16278T are dominant in modern populations of Central Asia (see details in S6 File).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Phylogenetic tree of mtDNA haplogroup A8 lineages from modern and ancient populations of Eurasia.

Ascription to the populations showed by colors: grey—modern populations of Eurasia [43, 44, 47, 49, 50, 56, 57, 59, 66–70] (see S6 File for details); red—Tagar culture (this study); yellow—Pazyryk culture [4]; pink—Aldy-Bel culture [4]; purple—Okunevo culture [27, 28].

https://doi.org/10.1371/journal.pone.0204062.g008

These components, e.g., haplogroups A8 and C4a2a (C4a2a1), in the gene pool of populations in the Early Iron Age and more recent periods in the western regions of the Eurasian steppe belt are potential markers of the influence of the eastern populations during the Scythian period on western populations. According to the results of Unterlander et al. [4], East Eurasian mtDNA components in the Western Eurasian steppe belt increased during the Early Iron Age. However, their earlier penetration into the western regions of the Eurasian steppe belt cannot be excluded. For example, lineages of the C4a2 (but not the C4a2a or C4a2a1) haplogroup have been identified in the Neolithic populations of the North Pontic region in Ukraine [29].

Genetic features of successive Tagar groups

We compared successive Tagar groups (Early, Middle, and Late Tagar) with each other and with other Iron Age nomadic populations to evaluate changes in the mtDNA pool structure. Despite the genetic similarity between the Early and Middle Tagar series and Scythian World nomadic groups (Figs 5 and 6; S4 and S6 Tables), there were some peculiarities. For example, the Early Tagar series was more similar to North Pontic Classic Scythians, while the Middle Tagar samples were more similar to the Southern Siberian populations of the Scythian period (i.e., completely synchronous populations of regions neighboring the Minusinsk basin, such as the Pazyryk population from the Altay Mountains and Aldy-Bel population from Tuva).

We observed differences in the mtDNA pool structure between the Early and the Middle chronological stages of the Tagar culture population, as evidenced by the change in the ratio of Western to Eastern Eurasian mtDNA components. The contribution of Eastern Eurasian lineages increased from about one-third (34.8%) in the Early Tagar group to almost one-half (45.8%) in the Middle Tagar group.

At the level of mtDNA haplogroups, we detected a decrease in the diversity of phylogenetic clusters during the transition from the Early Tagar to the Middle Tagar. This decline in diversity equally affected the West Eurasian and East Eurasian components of the Tagar mtDNA pool. It should be noted that this decrease can be partially explained by the smaller number of Middle Tagar than Early Tagar samples. Under a simple binomial approximation the mtDNA clusters, observed at frequencies of 6.3% and 11.7%, could be lost by chance in our Early (N = 46) and Middle (N = 24) Tagar samples, respectively. However, the simultaneous lack of several such clusters, with a total frequency in the gene pool of the Early group of 34.8%, is unlikely.

Using a haplotype sharing analyses, we obtained similar results (Fig 6; S6 Table). Exactly half of the mtDNA lineages in the Middle Tagar series are identical to Early Tagar lineages. Moreover, almost all of the unique (non-identical) mtDNA lineages from the Middle Tagar series were phylogenetically closely related to the Early Tagar lineages and belonged to the same mtDNA phylogenetic cluster (except for one, lineage 16158T-16311C, HV*). On the other hand, 23 out of 29 mtDNA haplotypes, representing 72% of the Early Tagar series, were missing from the Middle series.

Thus, our results confirm the genetic continuity between the Early and Middle Tagar groups. Additionally, we detected a reduction in mtDNA diversity in the Tagar gene pool at the level of haplogroups and haplotypes. We did not find evidence for a migration wave or a change in the direction of interpopulation genetic interactions during this period.

The observed reduction in the genetic distance between the Middle Tagar population and other Scythian-like populations of Southern Siberia(Fig 5; S4 Table), in our opinion, is primarily associated with an increase in the role of East Eurasian mtDNA lineages in the gene pool (up to nearly half of the gene pool) and a substantial increase in the joint frequency of haplogroups C and D (from 8.7% in the Early Tagar series to 37.5% in the Middle Tagar series). These features are characteristic of many ancient and modern populations of Southern Siberia and adjacent regions of Central Asia, including the Pazyryk population of the Altai Mountains. We did not obtain strong evidence for an intensification of genetic contact between the population of the Minusinsk basin and the Altai Mountains in the Middle Tagar period compared with the Early Tagar period. Although, several archaeologists have found evidence for the intensification of contact at the level of material culture, namely, a cultural influence of the population of the Altai Mountains (represented by the Pazyryk population) on the population of the Minusinsk basin (the Saragash Tagar group) [6, 71, 72].

Another important issue is the change in the genetic structure of the Tagar population during the transition from the Middle (Saragash) to the Late (Tes`) stage. The Late Tagar stage refers to the Xiongnu period. Many archaeologists suggest that the formation of the Tes`stage involved the direct cultural influence of the Xiongnu and/or related groups of nomads from more eastern regions of Central Asia [71, 73]. Some archaeologists have even suggested renaming the Tes`stage in the Tes`culture [71], emphasizing the role of new eastern cultural elements. If this influence also existed at the genetic level, then we would expect to observe new genetic elements in the Tes`gene pool, particularly those of East Eurasian origin.

In this study, we only investigated a small series of Tes`mtDNA samples. We do not have a complete understanding of the genetic composition of the Tes`population. However, our results are of interest in the context of the possible influence of eastern nomads during this period. The Late Tagar group (Tes`stage) was most distantly related to Xiongnu and other eastern nomadic groups (see MDS plot at Fig 5; S4 Table). Moreover, eight of the nine Tes`mtDNA samples belonged to the West Eurasian mtDNA cluster. Furthermore, all of the mtDNA lineages detected in our small Late Tagar series were also present in the series from earlier Tagar groups (Figs 2 and 6; Table 1; S6 Table). Thus, we do not have any evidence for an influence of a genetically divergent population from more eastern regions of Central Asia on the population of the Minusinsk basin in this period. To assess the level of genetic continuity and identify any external genetic influence, a significant increase in the Tes`series is necessary.

Based on our results, we can preliminarily conclude that there was genetic continuity, at least partially, between the Early, the Middle and the Late Tagar populations. We did not find evidence of extensive gene exchange between the Tagar population and any genetically distinct (with respect to the mtDNA pool) human groups.