Integrating Different Biological Evidence Around Some Microevolutionary Processes: Bottlenecks and Asian-American Arctic Gene Flow in the New World Settlement

Uniparental Genetic Markers

Most of the recent advances in the genetic reconstruction of the peopling of the Americas are based on the analysis of uniparental genetic systems mainly represented by mitochondrial DNA (mtDNA, particularly the first hypervariable region or HVS-I), which has an exclusively maternal inheritance, and the Y-chromosome with exclusively paternal inheritance. The intergenerational passage of unmodified mtDNA genomes and the non-recombinant portion (NRY) of the Y-chromosome over generations means that all lineages present in modern populations should have descended from a single strain that was part of a specific ancestral gene pool.

Early analyses of the HVS-I reveal that the majority of contemporary Native Americans had five distinct mitochondrial lineages (or haplogroups), named A, B, C, D, and X (Schurr and Sherry 2004). Each haplogroup has a sequence of specific mutations that identify it (Table 1). Above and beyond the support to the theory of Asian origin of the first Americans that these data provide, the presence of A, B, C, D, and X was interpreted by some scholars as supporting a hypothesis of multiple migrations to the New World (Horai et al. 1993; Torroni et al. 1992). However, this view was gradually disregarded because of new investigations based on larger samples and more sophisticated approaches, which gave more support to a single-migration scenario (Bonatto and Salzano 1997a,b; Merriwether et al. 1995; Stone and Stoneking 1998).

Mutations which define sub-lineages of A, B, C, and D haplogroups are found outside the HVS-I region

Studies involving markers in the non-recombinant Y-chromosome portion make a perfect combination with those based on mtDNA, since they reveal demographic and evolutionary histories mediated through a male perspective. The phylogenetic tree of Y-chromosome haplogroups, or paternal lineages, is constituted by the combination of binary alleles recognized as unique-event polymorphisms or single-nucleotide polymorphisms (SNPs). However, since the beginning of the mid-1990s, progress of knowledge in this field has been slow due to difficulties in finding good Y-chromosome population—or at least group-specific—markers (Jobling and Tyler-Smith 1995). This problem is being circumvented thanks to a combination of efforts from various research groups. In the last paper of the Y-Chromosome Consortium team (Karafet et al. 2008), reveals the most recent level of resolution achieved. The tree now uses a hierarchical naming system adopted in 2002 in order to standardize the nomenclature (The Y-Chromosome Consortium 2002). Additionally, the combination of SNP information along with microsatellite data, whose mutation rate is much faster, offers valuable insights into human population evolutionary and demographic histories. But how has the Y-chromosome data helped unravel questions related to settlement of the Americas? The discovery of a single founder Y-chromosome (Pena et al. 1995; Underhill et al. 1996) first suggested that migrants would have come in a single wave of migration. A C→T mutation at the M3 (DYS199) locus defines the autochthonous Native American founder haplogroup now named Q1a3a (Karafet et al. 2008). It is present in high frequencies in Native Americans only and in some few Siberian populations, probably due to reverse gene flow from Beringia to western Siberia (Bortolini et al. 2003; Karafet et al. 1997a; Lell et al. 1997; Lell et al. 2002; Santos et al. 2007; Schurr and Sherry 2004; Underhill et al. 1996). Subsequently, the existence of other founder lineages was postulated (Bortolini et al. 2003; Karafet et al. 1999; Lell et al. 2002; Ruiz-Linares et al. 1999). For instance, Y chromosomes bearing a specific C→T mutation at locus M242, named as Q, were observed initially in natives of America, Siberia, and Mongolia, an arrangement identified as the ancestral chromosome of Q1a3a (Bortolini et al. 2003; Karafet et al. 2008; Seielstad et al. 2003). Besides that, Siberian and Central Asian undifferentiated Q chromosomes have also been observed at low frequencies in Turkey, India, Pakistan, Korea, Japan, and Oceania (see revisions in Zhong et al. 2010a, b).

Studies with microsatellites on Q1a3a chromosomes clearly show the accumulation of new variants associated with the process of radiation across the American continent and its probable Beringian origin (Bortolini et al. 2003). More recently, other Q1a3a sub-lineages, resulting from in situ differentiation and with a population-/tribe-/region-specific distribution, have been identified (e.g., Q1a3a1, Q1a3a2, Q1a3a3, and Q1a3a4; Fig. 1). Moreover, Fig. 1 also shows that only two C3 clade lineages outside the Q clade should be considered as Native American (C3* and C3b; Bortolini et al. 2003; Geppert et al. 2010; Karafet et al. 2008; Zegura et al. 2004).

Our recent studies with 20 Native American populations tested for 18 SNPs that can identify all currently known lineages of the Y-chromosome Q clade (Pinhasi R 2009) indicate a higher number of distinct Asian and/or Beringian lineages in North America than in South America. Besides haplogroup Q lineages, Native Americans also present a low frequency of C3b, found only in North America (Karafet et al. 2008; Zegura et al. 2004), while the more ancient C3* was found in Northwestern South America (Geppert et al. 2010). The presence of more ancient C3* in Northwestern South America may indicate its existence in North America nowadays (not detected so far) and/or in earlier times, since the peopling of South America from the Beringian source likely occurred by land through North America (Pinhasi 2009).

As was the case with mtDNA, studies based on the Y-chromosome have been used to infer the time of entry into America, mainly using Q1a3a and its sub-lineages. These estimates provide dates from 9,000 to 22,000 years before present (YBP) for the initial peopling of the continent (Bortolini et al. 2003; Santos et al. 2007).

Autosomal Genetic Markers

The potential of human autosomal or biparental markers for addressing the history of populations has been widely demonstrated, but few studies using these systems in Native American populations have been performed. However, despite the scarcity of data, some biparental marker-based findings have been crucial in outlining the peopling of the New World. For example, Schroeder et al. (2009) identified an allele (9AR) at autosomal locus D9S1120 that is highly frequent among all Native American populations but absent from other worldwide populations. According to the authors, this particular distribution is most consistent with the hypothesis that all modern Native Americans derive from a common founding population, in accordance with the scenario depicted by the more recent uniparental loci studies.

Up to now, the most extensive research using autosomal chromosome markers has been that performed by Wang et al. (2007), comprising 78 populations typed for 678 microsatellite loci, including 29 Native American groups. Recently, Ray et al. (2009) used this large data set and an approximate Bayesian computation framework to contrast three scenarios that summarize the main controversies concerning the colonization of the continent. The first scenario considers a single-wave (SW) model, which posits that all Native American diversity stems from a single migration event from Asia, without any subsequent gene flow between the two continents. The second scenario is a two-wave (2W) model that allows for a second migration from Asia that would have occurred more recently. Finally, the third scenario is identical to the SW model but allows for asymmetric and recurrent gene flow (RGF) between Asia and the Americas after the initial colonization (Ray et al. 2009). The authors strongly reject the settlement of the Americas by a single, discrete, colonization wave from Asia (SW model), as well as the 2W model, and underline the importance of gene flow between Asia and America during the evolution of Native American populations. Thus, RGF was considered the most parsimonious model to explain the results. Ray and colleagues also suggest that the initial settlement very likely occurred between 13,200 and 15,250 YBP, in agreement with the results based on Y-chromosome and some mtDNA analyses.

In conclusion, the recent genetic analysis using mtDNA, Y-chromosome, and autosomal markers supports an origin in Beringia for the common autochthonous Native American genetic lineages. Many other lineages evolved in situ within America and can be considered to derive from the former ones. Concomitantly, many typical Asian lineages/alleles were lost during the Beringia colonization due to bottlenecks and other drift events (see below). Additionally, the genetic data also indicate that there is a higher number of distinct Asian and/or Beringian Y-chromosome (Pinhasi 2009) and mtDNA lineages (Perego et al. 2010) in North America than in South America. As a result, South Americans are genetically related to North Americans but less than North Americans are to Asians (especially Siberians). Basically, two major causes, not mutually exclusive, must be involved to explain these findings: (a) a pronounced founder effect at the beginning of South America’s colonization via the Panama Isthmus and/or (b) continuous gene flow between North American and Asian populations via the Circum-Arctic route after the Beringian collapse (see below).

Craniofacial Traits

Human populations in general and Native Americans in particular display high levels of variation in cranial morphology. It is highly debated whether this diversification originated through genetic drift or other neutral processes or through adaptation by natural selection to climatic and other environmental factors. Since non-neutral markers can potentially provide useful information about biological relationships and affinities among ancient populations, the debate around the relative importance of stochastic versus non-stochastic evolutionary agents is of crucial importance (González-José et al. 2004; Harvati and Weaver 2006; Relethford 1994; Roseman and Weaver 2007). Regarding settlement of the New World, the putative set of adaptations that arose in response to global climate change corresponding to the Last Glacial Maximum and the potentially enormous amount of local adaptations that occurred while the first settlers occupied the virgin environments of the continent are both events that influenced Asian and American populations, respectively.

The exploration of large databases using the theoretical and methodological tools of population and quantitative genetics has received much attention in recent years. Vast databases consisting of classical inter-landmark measurements made by Williams W. Howells (Howells 1973; 1989), Tsunehiko Hanihara (Hanihara 1996), and Héctor Pucciarelli (Pucciarelli et al. 2003, 2006b, 2008) provide much of the input for several recent analyses. Alternatively, data consisting of landmark coordinates analyzed using geometric–morphometric methods are in widespread and increasing use in the recent literature on human variation (e.g., (Ackermann 2005; González-José et al. 2008a,b; Harvati et al. 2004; Harvati and Weaver 2006; Perez et al. 2007, 2009)). These methods provide some advantages over traditional morphometrics because they preserve the geometry of the object studied better than traditional measurements and thus allow for a better analysis of shape; readily account for size correction; enable the identification of landmarks where shape differences occur and the relative levels of difference at each landmark; enable visualization of the shape changes between specimens in specimen space; and, perhaps most importantly, enable the quantification of some traits that are difficult to measure with conventional measurements (Adams et al. 2004; Mitteroecker and Gunz 2009; Rohlf and Marcus 1993).

The aim of most of the studies made on the classical databases is to determine if the patterns of within- and between-group variability are compatible with evolution mediated by stochastic factors or by natural selection. As expected, depending on the type of analysis performed, results indicate a variable influence of stochastic versus non-stochastic factors. For instance, the computation of levels of differentiation (Fst) indicates that craniometric traits behave much as neutral markers: around 10–15% of variation is accumulated in differences among populations (Relethford 2001, 2002). The Americas in general (Relethford 2002) and South America in particular (González-José et al. 2001; Sardi et al. 2005) display higher levels of variation than expected, considering their relatively recent occupation.

Analysis of intertrait covariation (rather than absolute phenotypic differences) indicates a pattern of stability in the intertrait correlation and covariance structure, probably as the effect of morphological integration of traits linked by a common function and/or developmental chronological and spatial connectivity (González-José et al. 2004). These levels of morphological integration guarantee a conservative and coordinated response to environmental pressures. Maybe, this is why many studies report a high coincidence among morphological and genetic between-group distances (González-José et al. 2004; Manica et al. 2007; Roseman 2004). Recently, a group of papers debated the relative importance of ancient demography and climate in determining worldwide patterns of within-population cranial diversity. Overall, these studies suggest that among-population differences in extant H. sapiens cranial morphology are proportional to among-population differences in neutral molecular characteristics (Betti et al. 2009a,b; Manica et al. 2007). For most of the populations studied, cranial morphology varies among regions in a manner consistent with neutral expectations. Some authors (Harvati and Weaver 2006; Roseman and Weaver 2004) state that there is an excess of differentiation involving Circum-Arctic populations, thus suggesting that some aspects of cranial morphology might have evolved as adaptations by natural selection to coldness of climate. However, this statement was recently challenged by Betti et al. (2009a), who use an explicit theoretical and methodological framework to measure the relative influence of climate and distance from Africa. The authors conclude that distance from Sub-Saharan Africa is the sole determinant of human within-phenotypic diversity, showing that neutral processes have been much more important than climate in shaping the human cranium.

In summary, evidence collected from trait and intertrait variations at the within- and between-population levels on classical databases indicates that neutral processes (genetic drift) have been much more important than natural selection driven by climate in shaping patterns of craniofacial variation. Some papers have also demonstrated that much of the selective signal comes from groups inhabiting extremely Circum-Arctic regions.

Quantitative genetics theory demonstrates that within-group variation in the ancestral population may be an important source of between-group diversity in descendant groups (Lande 1979). In this context, the analysis of early within-population variability is of particular importance in the light of the South America settlement, since the most productive ancient site in terms of complete skulls is located in Lagoa Santa, central Brazil. In other words, the most confident estimations of ancient biological variability come from South America, rather than from North America. The Lagoa Santa series was extensively studied recently by Neves and coworkers (see a review in Neves and Hubbe 2005). The resulting studies congruently claim that the Lagoa Santa (Neves and Hubbe 2005) and other South American (González-José et al. 2005a; Mena et al. 2003; Neves et al. 1999) ancient samples tend to have low affinities with modern east Asians and most modern Amerindians. In a recent paper (González-José et al. 2008a,b), we suggest that the set of craniofacial characters defining the range of New World phenotypic variation is already present in its putative ancestor, the late Pleistocene early members of our species, as well as in their two immediate sister groups—modern Australians and northeastern Asians. In this context, the precursor of the single ancestral population, whose existence is suggested by mtDNA or Y-chromosome markers (see above), could well be a late Pleistocene Asian population inhabiting someplace in Northeast Asia and showing high degrees of craniofacial diversity. If craniofacial variation in the ancestral population was high enough to accommodate specimens with a mosaic of characteristics, then the demographic expansion of this population from Northeast Asia is enough to explain without serious contradictions both molecular and craniofacial data in prehistoric America.

To sum up, it seems that several independent analyses of classical, massive databases certify the use of craniofacial phenotypes as a good proxy to infer biological affinities among past populations, in combination with the advantages of molecular analyses. However, more research is still needed in order to further refine the information extracted from skull phenotypes.