Major mitochondrial DNA haplotype heterogeneity in highland and lowland Amerindian populations from Bolivia
Abstract This study provides the frequencies of four mitochondrial DNA (mtDNA) haplogroups of 233 native South Amerindians in eight populations living in the Beni Department of Bolivia, including six populations not previously studied. Linguistically, these populations belong to the three principal South Amerindian language stocks, Andean, Equatorial-Tucanoan, and Ge-Pano-Carib. Frequency analyses under geographic, historic, linguistic, and genetic configurations using the theta statistic of Weir (Weir 1990) and analysis of molecular variance (AMOVA) show similar results. Results are also similar when phenetic cluster is used. Aymara belongs almost exclusively to haplogroup B, Quechua- and Moseten-speaking tribes belong to haplogroups A and B, but the first tribe presents high frequencies of haplogroup B. Yuracare, Trinitario, and Ignaciano exhibit high frequencies of A, B, and C haplogroups, and the Movima present a large proportion of haplogroup C. There is some correspondence between mtDNA haplogroup frequencies and language affiliation and historical connections, but less so with geographic aspects. The present study provides a context for understanding the relationship between different Amerindian populations living in a multiethnic area of Bolivia.
From the earliest studies on mitochondrial DNA (mtDNA) diversity among native Americans, research has been ongoing in an effort to determine the origin of the indigenous peoples of the Americas (see Bonatto and Salzano  for a review of the literature). Little agreement has been reached, though, concerning the relationship between the distribution of the four founding lineage haplotypes in native Americans and the number of waves of migration to the New World (Merriwether and Ferrell 1996). Few detailed studies have been made of the genetic haplogroup variability in reduced geographical areas of South America. Some studies have been done in small regions of Brazil (Ward et al. 1996), Panama (Kolman and Bermingham 1997; Batista et al. 1995), and Costa Rica (Santos et al. 1994), but not in Bolivia.
The highlands of Bolivia constitute a natural path for migration from northern areas into the coastal South and into the lowlands in Amazonia. No genetic studies of indigenous peoples living in the Bolivian lowlands have been carried out, however. The “Llanos de Mojos” is a region of the Bolivian lowlands located in the Beni Department. This area is particularly interesting from many points of view: the great multiethnicity of the human populations that live there, their geographic marginality, and their unique demographic history (Denevan 1966; Block 1980). Before the arrival of the first missionaries, the estimated native population size was 200,000, but after one hundred years this number declined dramatically.
What is known of the history of the “Llanos de Mojos” region was written by the Jesuits and Franciscans. These missionaries founded large villages or settlements, called “reductions,” for the religious and cultural education of the Indians. Several tribes have mixed there since 1682. In 1746 the Jesuits where expelled from the Llanos de Mojos, and the reductions were transformed into the present villages of San Ignacio and Trinidad, where Ignaciano and Trinitario populations live. From Jesuit ethnographic records, Saignes (1985) showed that the Inca had occupied the oriental Andean range and the upper and middle valleys of the Beni river, were the Chimane and Moseten Indians live at present. However, no further information is available about the importance of the interaction between the Inca and the local populations in the Llanos de Mojos.
The present study provides data on the frequencies of the four major mtDNA haplogroups in eight Amerindian populations, six of them not previously studied. Their genetic affinities are studied, taking into account the different hypotheses on the colonization of the South American continent. This study will be useful not only for understanding the complex history of the peopling of Bolivia, but also for extending our knowledge on the distribution of mtDNA variation among native South Americans.
Materials and Methods
Samples. All the populations studied live in Llanos de Mojos. The sample includes eight populations: Aymara, Quechua, Chimane, Moseten, Yuracare, Ignaciano, Trinitario, and Movima (estimated sizes of these populations are shown in Table 1). Other populations living in the Beni Department include Itonama, Canichana, Baure, More, Cavineno, Esse-Ejja, Chacobo, Pacahuara, and Cayuvaba, who speak different languages, enhancing the greatly varied ethnicity of this region. The Aymara and Quechua may be considered colonizers in the area, since they originally occupied the Andean highlands. Chimane, Moseten, and Yuracare live in the Andean foothills, whereas the Ignaciano, Trinitario, and Movima occupy the seasonally overflowing savanna. However, this distribution is merely tentative, since, in fact, their geographic areas of dispersion overlap (Figure 1). A total of 233 individuals with known ethnic affiliation were studied. Ethnicity was assigned to individuals based on self-identification as recorded from interview data. Only unrelated individuals were studied. Family relationships were determined in situ by a demographic survey done at the same time that genetic samples were obtained. Most of the sample sources were bulb hairs, but some blood stains were also obtained. Pulled hairs were stored in sterile tubes until analysis, and blood stains were obtained on 5-mm filter paper (Rubilabor 589^sup3^ blue ribbon) and air-dried for several minutes before storage in sterile tubes.
Polymorphism Typing. DNA extraction was performed using ChelexTM 100, as described by Walsh et al. (1991). All samples were typed using specific mtDNA restriction sites for diagnosis of haplogroups A, B, C, and D. These haplogroups are defined by restriction size or length polymorphism: (A) HaeIII site gain at nucleotide position (np) 663, (B) 9-base-pair deletion in the COIl-tRNA intergenic region (region V), (C) Alul site gain at nucleotide position (np) 13262 and HincH loss at np 13259, and (D) AM site loss at no 5176.
The primers used were previously described in Wrischnik et al. (1987). For each haplogroup these primers are: (A) H708 (5′-TGAACTCACT GGAACG-3′) and L630 (5′-ATGTTTAGACGGGCTC-3′); (B) H8366 (5’TTTCACTGTAAAGAGGTGTTGG) and L 8150 (5′-CCGGGGGTATACT ACGG-3′); (C) H13305 (5′-GTGCAGGAATGCTAGG-3′) and L13204 (5′-ACTCTGTTCGCAGCAG-3′); (D) H5237 (5′-CAAAAAGCCGGTTA GC-3′) and L5147 (5′-AAACTCCAG-CC-ACG-3′)
Polymerase chain reaction amplifications were done in a total volume of 25 (mu)L, containing 2.5 gL of lOX reaction buffer, 0.2 mM of each dNTP, 1 (mu)L W-1 (BRL), 1.5 mM Cl^sub2^Mg, 1 (mu)L of each primer (approximately 0.2 mM stock), 0.25 units of Taq polymerase (BRL), and 5 (mu)L of template DNA extracted from bulb hairs or 1 (mu)L of DNA extracted from blood stains. Initial denaturation was at 94degC for 4 min, followed by 30 cycles of 94degC denaturation for 1 min, 55degC annealing for 1 min, 72degC extension for 1 min, and a final extension cycle of 72degC for 5 min. From the reaction tubes, 8 (mu)L of the amplified products were resolved by electrophoresis on 2% agarose gel containing ethidium bromide. The remaining 17 (mu)L were digested with haplogroup-specific restriction enzymes for 1 h 30 min at 37degC. The restriction products were electrophoresed in a 4% agarose gel, stained with ethidium bromide, and observed under 260 nm UV light to determine if the amplified fragment had been restricted by the enzyme.
These statistics were initially calculated for the eight populations considered, as well as for the whole Amerindian continent (Table 2). Clusters of populations were made using geographic, genetic distance, linguistic, and historic criteria, and these statistics were computed for each one of the five sets of clusters obtained (Table 3).
1. Geographic criteria (Andean versus Amazonian populations). Two groups were formed according to their original territory. Bolivian populations are clearly divided into two geographical areas divided by the Andean Mountains. Even though all populations studied live in the Amazonian region, the Aymara and Quechua may be considered colonizers in the area, since they originally occupied the Andean highlands and the principal stock of these populations remains there.
2. Genetic distance criteria. Four groups were derived from the unweighted pair group method with arithmetic mean (UPGMA) tree, according to the four principal branches of this tree.
3. Linguistic criteria. According to the Amerindian language classification (Ruhlen 1991), there are four major groups of Amerindian languages in South America, three of which are spoken in the Beni Department by the studied populations. We have used those three major groups to classify our populations. At one level, those major stocks are divided into many related languages that have been used for a more specific analysis.
4. Historic criteria. Five groups were formed based on two historical events. The first one takes into account the expansion of the Incan Empire from the Andean region to the Chimane and Moseten habitats. The Quechua are considered descendants of the Incas. The second event refers to the more recent Jesuit reductions in Llanos de Mojos, where the Ignaciano and Trinitario populations were joined during the 18th century.
The comparison of the theta values for each clustering of populations with the initial value obtained for the whole sample makes it possible to determine if the heterogeneity among the clusters increases, as is expected when the clustering of samples is based on valid assumptions (Lorenz and Smith 1996). The AMOVA also analyzes the heterogeneity among clusters using the percentage of variance resulting from diversity within populations, among groups of populations, or among populations within a group. A permutation resampling procedure was used to confirm the statistical significance of the analyses (Excoffier 1992).
The genetic affinities among populations were computed using DNA distances derived from the haplogroup frequencies obtained. The genetic distances by Nei (1972), Reynolds et al. (1983), and Cavalli-Sforza and Edwards (1967) were calculated. Although they all gave essentially the same results, the chord distance by Cavalli-Sforza and Edwards (1967) was chosen to build the matrix distance between populations. It is in better accordance with the human demographic history in South America than the others, it allows for differences in population sizes among groups, and it does not require constant population sizes through time (Cavalli-Sforza and Edwards 1967). UPGMA trees were constructed using the PHYLIP genetic package (Felsenstein 1993).
A total of 233 samples were typed for all four major mtDNA haplogroups. The frequencies of each haplogroup by population are shown in Table
1. No compound haplotypes (positive restriction for more than one haplogroup) were found. Only six individuals (2.6%) did not belong to one of the four principal Native American haplogroups (A, B, C, or D) and were therefore classified as “others.” They may be the result of admixture with Caucasian populations (Torroni et al. 1993), or belong to some rare Native American haplogroup (Bailliet et al. 1994). These samples were removed from their respective groups, and the frequencies of the four principal haplogroups were re-scaled to unity in these groups for further analyses. Figure 2 shows the cumulative haplogroup frequencies by populations. Haplogroup B is the most common in the studied area (51.9%), and haplogroup D is the least common (3%). Haplogroups A (21.8%) and C (20.6%) show similar values.
A clear highland-lowland gradient in the frequencies of B and C haplogroups for the different populations can be observed (Figure 2). Haplogroup B shows high frequencies in the Andean region but decreases in the lowlands, whereas for haplogroup C the highest frequencies are observed in the lowlands, with a decrease in the highlands. In the Aymara, haplogroup B is by far the most frequent, as has also been described in other Aymara samples (Merriwether et al. 1995a). In the Movima sample the prevailing haplogroup is C, whereas in the rest of populations at least two haplogroups share similar importance. In the Chimane and Moseten, haplogroups A and B are the most frequent, whereas in the Ignaciano and Trinitario, B and C are preponderant. The Yuracare have moderate frequencies of haplogroups A, B, and C. Finally, the Quechua show a higher frequency of haplogroup B, but haplogroups A and C are also present, a finding that contrasts with those of other studies of Quechua populations (Merriwether et al. 1995a), which show higher frequencies of haplogroups A and D.
MtDNA Diversity. The diversity indices (h) by Nei (1987) measuring mtDNA diversity for all the groups considered are shown in Table 2. The Aymara show the lowest index of diversity (h = 0.119), and the rest of the populations show values of h between 0.698 (Yuracare) and 0.514 (Moseten). The diversity index of the whole population from the Beni area is h = 0.630 (n = 233), a value similar to that obtained for the joint sample of all Amerindians populations (h = 0.697). This value was calculated by us for 1877 individuals from 52 ethnic groups using published data (Bonatto and Salzano 1997; Horai et al. 1993; Lorenz and Smith 1996; Merriwether et al. 1995b; Schurr et al. 1990; Torroni et al. 1993 and 1994).
AMOVA Analysis. The analysis of molecular variance may be used at different hierarchical levels: considering whole populations in only one group; considering groups of populations under different arbitrary assumptions; and, finally, comparing each pair of populations to determine if there is significant interpopulation variance and if they are genetically distinguishable.
At the first level, the AMOVA (Table 3) shows a significant interpopulation variability between the eight Beni populations, with 80.1 % of the total variance due to within-population differences and only 19.9% due to among– population differences. All the between-group and within-group variability values obtained are similar to those obtained for the whole Amerindian continent. If the Aymara and Movima are excluded from the analysis, the withingroup variability increases from 80.1 % to 91.6%, and the among-populations variance decreases from 19.9% to 8.4%, indicating that these two groups are responsible for most of the variability observed.
The variance observed in the distribution of four principal haplogroup frequencies can be partitioned into three components: (1) variance within populations, (2) variance among groups of populations, and (3) variance among populations within a group (Excoffier et al. 1992). Most of the variance observed in the frequency distribution of the four principal haplogroups is due to variance at the intrapopulation level.
At the interpopulation level, higher values of variance among groups of populations, combined with lower values of variance among populations within a group, indicates the reliability of the clustering criteria. Under this assumption, the best values are obtained when six clusters of closely related linguistic groups (Ruhlen 1991) are considered. If only three linguistic clusters are considered (three subgroups of major Amerind languages), the heterogeneity levels decrease a little. The historical criteria and the UPGMA distance also fall under valid assumptions. Groups generated with geographic criteria do not have significant differences between them (p = 0.084), thus invalidating these criteria.
Finally, AMOVA was applied to each pair of populations. Most of the comparisons show significant differences, with some exceptions (Table 4). The populations that form a clear linguistic cluster do not differ significantly from each other. The significance of the interpopulation variance between Trinitario and Ignaciano (p = 0.962), and between Chimane and Moseten (p = 0.816) reveals that, at this level, both pairs of populations are genetically indistinguishable. The Yuracare, placed in an intermediate position, only show significant differences when they are compared to the Movima, Quechua, and Aymara. Neither were there significant differences between the Quechua and Moseten populations (p = 0.091). The Movima and Aymara show significant differences in haplogroup frequencies when compared with all the other groups.
0 Analysis. The overall theta value obtained for a single group, including all populations studied, was theta = 0.199, similar to that obtained for the whole continent (theta = 0.206). Intergroup heterogeneity varies depending on how the populations are clustered prior to estimating theta. If two geographic clusters, dividing Andean from Amazonian populations, are considered, the theta obtained was 0.191 (smaller than the initial value). The analysis based on six linguistic groups gives a value of theta = 0.217, which decreased to theta = 0.212 when the three major linguistic subfamilies (Andean, Ge-Pano-Carib, and Equatorial-Tucanoan) were considered (Ruhlen 1991). The historical clustering criteria yield a theta value of 0.220, indicating the historic relationship between Quechua- and Moseten-speaking tribes (Saignes 1985). The value obtained for the genetic clustering is theta = 0.222 for four clusters, which indicates a fair agreement with the linguistic and historic criteria.
Genetic Distance. The distance matrix derived from the haplogroup frequencies, using the Cavalli-Sforza and Edwards (1967) genetic distance for the populations, is shown in Table 5. The UPGMA tree obtained from these distances is shown in Figure 3. The most distinct groups in the plot are the Movima and Aymara, both having one preponderant haplogroup (C and B, respectively). The cluster formed by Chimane, Moseten, and Quechua share haplogroups A and B, and the cluster formed by Ignaciano and Trinitario share haplogroups B and C.
The populations considered here show great genetic and linguistic heterogeneity despite currently living in neighboring areas of the same latitude. Actually, the multiethnicity of the Beni area is reflected by the diversity values (h, Nei 1987) for this area, which are similar to those observed for the whole of the Amerindian populations. The Trinitario-Ignaciano and Chimane– Moseten groups cluster in the UPGMA genetic tree, suggesting that they come from the same genetic source, which is consistent with their related languages. However, the recent history of these populations indicates a not-so-recent ancestry. The Trinitario and Ignaciano belong to the Macro-Arawak equatorial language stock; their past can be traced back only as far as when they were confined and protected in reductions by the Jesuits in the 17th century. Little is known about the origin of the Chimane and Moseten populations, which speak the Pano language from the Ge-Pano stock, except that they were partially influenced by the Inca culture. The Aymara and Quechua are Andean populations, as evidenced by their geography and their languages, but the relationships between them remain unclear. The Movima lived in the northern savanna, near Santa Ana de Yacuma, until they moved southward, in pre-Columbian time, and now overlap with the Arawak people, who arrived recently in the Llanos de Moios.
The AMOVA presented here confirm the aforementioned high genetic variability. As for intergroup genetic relationships, when Aymara and Movima are removed from the analyses, the percentage of between-group variance decreases drastically, and the rest of the samples-Quechua, Moseten, Chimane, Yuracare, Ignaciano, and Trinitario-cluster together. The haplogroup frequencies in Figure 2 show that the Aymara and Movima are probably the most isolated groups in this study. On the other hand, the similarities observed among the rest of the groups might be the result of intense genetic flow. The AMOVA and theta analyses are useful for genetic analysis when geographic, linguistic, and historic factors interact (Excoffier et al. 1992). From our data, it appears that geography was not a principal factor in genetic differentiation, despite the mountains being an important barrier. In fact, the genetic similarities observed here between the two Andean populations, Aymara and Quechua, may well be due to significant genetic flow between them. Other studies, though, show greater similarities between Aymara and Quechua than those observed here, perhaps because of the genetic influence of the Chimane and Moseten upon the Quechua sample studied here.
Linguistic and historic components have been important in the recent demographic dynamics of the Beni area. They satisfactorily explain the genetic affinities between the populations studied. However, the three major South Amerindian linguistic stocks represented in the area (Ruhlen 1991) are less correlated with the genetic variability observed than when the six linguistic groups (Aymaran, Quechuan, Mosetenean, Movima, Yuracare, and Mojo) are considered (Tablet). The high level of linguistic and genetic variability observed in the reduced area of the Beni Department suggests that the peopling of the Bolivian lowlands was a complex process. The differentiation between highland and lowland populations detected in this study might reflect separate evolutionary histories. However, important levels of genetic flow between indigenous populations from the three different ecohabitats (Andean, pre-Andean hills, and savanna lowlands) cannot be discarded, as indicated by the similarities found among the Trinitario, Ignaciano, Yuracare, Chimane, Moseten, and Quechua, despite their different habitats and languages. The differences in the mitochondrial haplogroups found in the Beni area are in agreement with the linguistic and ethnic diversity of the populations studied.
Acknowledgments This study was funded by the Spanish Ministerio de Educaci6n y Ciencia through DGICYT Projects PB-93-0021 and PB-96-1485. We are grateful to J. Castro of San Ignacio Hospital, Greet Diltiens, and Roser Montagut for their cooperation in the collection of samples. We are also grateful to J. Cerda of Laboratorio de Biologia “San Calixto” from La Paz, Bolivia, for his help.
Bailliet, G., F. Rothhammer, F.R. Carnese et al. 1994. Founder mitochondrial haplotypes in Amerindian populations. Am. J. Hum. Genet. 55(1):27-33.
Batista, O., C.J. Kolman, E. Bermingham. 1995. Mitochondrial DNA diversity in the Kuna Amerinds of Panama. Hum. Mol. Genet. 4(5):921-929.
Block, D. 1980. In search of El Dorado: Spanish entry into Moxos, A Tropical Frontier, 15501767. Ph.D. diss., The University of Texas, Austin, TX.
Bonatto, S. L., and F.M. Salzano. 1997. A single and early migration for the peopling of the Americas supported by mitochondrial DNA sequence data. Proc. Natl. Acad. Sci. USA 94(5):1866-1871.
Bonatto, S.L., and F.M. Salzano. 1997. Diversity and age of the four major mtDNA haplogroups, and their implications for the peopling of the New World. Am. J. Hum. Genet. 61:14131423.
Cavalli-Sforza, L.L., and A.W.F. Edwards. 1967. Phylogenetic analysis: Models and estimation procedures. Evolution 32:550-570.
Denevan, W. 1966. The Aboriginal Cultural Geography of the Llanos de Mojos of Bolivia. Berkeley: University of California Press.
Diez Astete, A., and Riester J. 1996. Etnias y territorios indigenas. In Comunidades, territorios indigenas y biodiversidad en Bolivia. CIMAR Universidad Aut6noma Gabriel Rene Moreno. Santa Cruz, Bolivia: Miothek K. Eds.
Excoffier, L., P.E. Smouse, and J.M. Quattro. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes: Application to human mitochondrial DNA restriction data. Genetics 1992:479-491.
Felsenstein, J. 1993. PHYLIP: Phylogeny inference package. Version 3.2. Distributed by the author. Department of Genetics, University of Washington, Seattle, WA.
Horai, S., R. Kondo, H.Y. Nakagawa et al. 1993. Peopling of the Americas, founded by four major lineages of mitochondrial DNA. Mol. Biol. Evol. 10(1):23-47.
Kolman, CT, and E. Bermingham. 1997. Mitochondrial and nuclear DNA diversity in the Choco and Chibcha Amerinds of Panama. Genetics 147(3):1289-1302.
Lorenz, J.G., and D.G. Smith. 1996. Distribution of four founding mtDNA haplogroups among Native North Americans. Am. J. Phys. Anthropol. 101(3):307-323.
Merriwether, D.A., and R.E. Ferrell. 1996. The four founding lineage hypothesis for the New World: A critical reevaluation. Mol. Phylogenet. and Evol. 5(1):241-243.
Merriwether, D.A., R.E. Ferrell, and F. Rothhammer. 1995a. mtDNA D-loop 6-bp deletion found in the Chilean Aymara: Not a unique marker for Chibcha-speaking Amerindians Better]. Am. J. Hum. Genet. 56(3):812-813.
Merriwether, D.A., F. Rothhammer, and R.E. Ferrell. 1995b. Distribution of the four founding lineage haplotypes in Native Americans suggests a single wave of migration for the New World. Am. J. Phys. Anthropol. 98(4):411-430.
Nei, M. 1972. Genetic distance between populations. American Naturalist 106:283-292. Nei, M. 1987. Molecular Evolutionary Genetics. New York: Columbia University Press. Reynolds, J.B., B.S. Weir, C.C. Cockerham. 1983. Estimation of the coancestry coefficient:
Basis for a short-term genetic distance. Genetics 105:767-779.
Ruhlen, M. 1991. A Guide to the World’s Languages. Stanford, CA: Stanford University Press. Saignes, T. 1985. Caciques, tribute and migration in the Southern Andes. Indian society at the
17th. century colonial order (Audiencia de Charcas). London University: Institute of Latin American Studies.
Santos, M., R.H. Ward, and R. Barrantes. 1994. MtDNA variation in the Chibcha Amerindian Huetar from Costa Rica. Hum. Biol. 66(6):963-977,
Schurr, T.G., S.W. Ballinger, Y.Y. Gan et al. 1990. Amerindian mitochondrial DNAs have rare Asian mutations at high frequencies suggesting they derived for four primary maternal linea es. Am. J. Hum. Genet. 46:613-623.
Torroni, A., J.V. Neel, R. Barrantes et al. 1994. Mitochondrial DNA “clock” for the Amerinds and its implications for timing entry into North America. Proc. Natl. Acad. Sci. USA 91:1158-1162.
Torroni, A., T.G. Schurr, M.F. Cabell et al. 1993. Asian affinities and continental radiation of the four founding Native American mtDNAs. Am. J. Hum. Genet. 53:563-590.
Walsh, P.S., D.A. Metzger, and R. Higuchi. 1991. Chelex 100 as a medium for simple extraction of DNA for PCR-based typing from forensic material. Biotechniques 10(4):506-513. Ward, R.H., F.M. Salzano, S.L. Bonatto et al. 1996. Mitochondrial DNA polymorphism in three
Brazilian Indian tribes. Am. J. Hum. Biol. 8(3):317-323.
Weir, B.S. 1990. Intraspecific Differentiation. Sunderland, MA: Sinauer Associates. Wrischnik, L.A., M. Higuchi, M. Stoneking et. al. 1987. Length mutations in human mitochondrial DNA: Direct sequencing of enzymatically amplified DNA. Nucleic Acids Res. 529-542.
FRANCESC BERT,1 ALFONS CORELLA, MANEL GENE,2 ALEJANDRO PEREZ– PEREZ-,1 AND DANIEL TURBON1
ISeccio Antropologia, Departament de Biologia Animal, Facultat de Biologia, Universitat de Barcelona, Avinguda Diagonal 645, 08028 Barcelona, Spain.
2 Departament de Medicina Legal, Laboratori de Gen&ica Forense, Facultat de Medicina, Universitat de Barcelona, Barcelona, Spain.
Human Biology, February 2001, v. 73, no. 1, pp. 1-16. Copyright 2001 Wayne State University Press, Detroit, Michigan 48201-1309
Received 16 April 1999; revision received 27 June 2000.
Copyright Wayne State University Press Feb 2001
Provided by ProQuest Information and Learning Company. All rights Reserved