Introduction

Northwestern South America is a key territory for population studies, as it constitutes a passage for both North-South and South-North migrations in the new continent. There have been recent attempts to identify how early settlements in the Colombian territory took place based on biological samples and the analysis of different molecular markers (^{Casas-Vargas, et al., 2017}; ^{Díaz-Matallana, et al., 2016}; ^{Noguera-Santamaría, et al., 2015}; ^{Noguera, et al., 2014}; ^{Casas-Vargas, et al., 2011}; ^{Jara, et al., 2010}; ^{Uricoechea, et al., 2010}; ^{Keyeux & Usaquén, 2008}; Briceño, 2003; ^{Keyeux, et al., 2002}; ^{Noguera & Bernal, 2010}).

Studies on pre-Hispanic populations (^{Casas-Vargas, et al., 2011}; ^{Silva, et al., 2008}) have allowed an initial approach to human genetic variability in Colombia. However, although various archeological contexts have been identified in the Colombian north coast (^{Angulo-Valdés, 1951}; ¹⁹⁸¹; ¹⁹⁸³; ¹⁹⁸⁸; ^{Ramos & Archila, 2008}; ^{Rivera-Sandoval, 2018}; ^{Rojas-Sepúlveda & Martín, 2015}), little is known about their genetic variability within the pre-Hispanic Caribbean region.

One of the most thoroughly studied areas in the Colombian Caribbean coast is the archaeological site of Malambo, near the Magdalena River. Its inhabitants settled and adapted to environmental determinants in Lower Magdalena around 3070±200 years B.P. Their crops and the use of water resources such as rivers and lakes can be deducted from the remains of typical archaeofauna in these ecosystems. This type of sustenance is associated with well-known ceramic structures in western Venezuela and the Orinoco River valley (Angulo, 1981, ^{Langebaek, 1987}, ^{Meneses & Gordonés, 2005}). However, there are some questions regarding the Malambo timeline and its associations with other archaeological sites in the region, some of which seem to differ from the periods proposed by Angulo (^{Langebaek, et al., 1998}, ^{Langebaek & Dever, 2000}). In fact, it has been suggested that Malambo's excavations belong to different occupations, some of which were interpreted as a single establishment (^{Langebaek & Dever, 2000}), as verified by the detailed analysis of ceramics reported on the site since the incised modeling style known as the Malambo tradition is predominantly located in the deepest strata while the ceramics located in the superficial layers show similar characteristics to those of the indigenous settlements in the late period linked to the Malibú tradition between the 13^th and 16^th centuries (^{Rivera-Sandoval, 2018}).

One aspect to consider about Malambo, besides its deep history, is its relationship with other contemporary communities established in distant regions and with those occupying the region at different times. Genetic analysis can be very helpful in order to answer these questions. Previous works have defined specific American lineages (A2, B2, C1b, 21, C1c, C1c, C1d, D1, D4h3) described as founding lineages (Figure 1), as well as the sub-haplogroups found in North, Central, and South America (A2a, A2b, C4c, D2A, D3, D4E1, X2a, X2g, A2ad, A2af1 B2i2, B2J, B2K, B2L C1b13, DIG, D1J) (^{Tamm, et al., 2007}; ^{Perego, et al., 2009}; De Saint Pierre, et al., 2012; ^{Gómez-Carballa, et al., 2012}; ^{Romero-Murillo, 2016}). In this study, we selected a sample of four individuals from the Malambo site excavated by Angulo (1981): two middle-aged adults, one young adult, and one subadult. We analyzed and compared mitochondrial DNA with previous results in America (^{Tamm, et al., 2007}; Achilli, et al., 2008; Fagundes, et al., 2008; ^{Perego, et al., 2009}) to reconstruct population dynamics of ancient and contemporary communities.

Figure 1 Phylogenetic network of haplogroup B (network v5.0). Pre-Hispanic individuals from Malambo are shown in pink 

Materials and methods

Data analysis

For the alignment of sequences, we used the CLC Main Workbench (Qiagen). The hyper-variable region (HVR-I) nucleotides were aligned with a consensus sequence (Cambridge Reference Sequence -CRS-). We found the G16361A mutation recently described and confirmed in haplogroup B and sub-haplogroup B2j (^{Gómez-Carballa, et al., 2012}). We confirmed these results using restriction fragment length polymorphism (RFLP) analysis with the HaeIII enzyme to identify the intergenic 9-bp deletion in the COII/tRNA(Lys) as previously reported for this haplogroup. Fragments were visualized in polyacrylamide gel at 45 volts for 3 hours (Figure 2).

Figure 2 Mitochondrial DNA restriction fragment length polymorphism (RFLP) for haplogroup B in Malambo individuals 3018, 3019, 3020, and 3021 

We calculated the frequency of the haplogroup obtained by direct count. The genetic diversity was calculated based on the variant frequency: Ae= n/(n-1)= 1-Σpi2, where pi is the frequency of each allele (haplogroup) and n the size of the sample. Using the Arlequin 3.5.2 software, we analyzed the samples at intra- and inter-population levels by calculating their corresponding F _st (^{Excoffier, et al., 2005}) to estimate the genetic distances. We analyzed the phylogenetic networks of pre-Colombian mtDNA sequences obtained from the literature using a median-joining algorithm (NJ).

Intrapopulation analysis. For the assignment of the mitochondrial haplogroups, we compared the DNA sequences with the CRS reference sequence and determined the specific polymorphisms for each American-Indian haplogroup according to Torroni, et al. (1992).

Interpopulation analysis. For this analysis, we compared genetic data obtained from present and antique populations both in Colombia and in America. Then, a phylogenetic reconstruction was performed based in the genetic distances between haplogroups applying the Tamura and Nei Method (1993). Reduced Median Networks were established to determine the genetic relationship between haplogroups found in this study and other populations, as calculated and distributed by the Network 5.0 program. (Figure 1).

Results

Georeferencing

We used Google Maps to draw a map using the longitude and latitude references for each location where the samples were reported (Figure 3).

Figure 3 Geographical location of pre-Columbian groups referred in this study 

Discussion

Lower Magdalena is a geographically and culturally diverse region. To determine whether such diversity could be associated with the biological composition of the ancient human groups that settled in this region of the Caribbean, we carried out a genetic analysis of available bone and teeth remains. Based on our preliminary results with ancient DNA (aDNA), we found no evidence of population variability among the indigenous groups that settled in this region compared to other contemporary communities. Using bioarcheological and genetic results, we conducted a multi-variant analysis integrating the different data to correlate each result. We used the SPSS version 1.8 software for the analysis and generated a genetic distance matrix based on the Fst data; we also performing a principal component analysis (PCA) using haplogroup frequencies obtained upon the comparison of data from the present study with previously reported archeological populations of Colombia. For comparative purposes, we chose samples from Checua, Guane, and Valle de Santiago sites: the Checua are one of the oldest populations of Colombia while the Guane correspond to an indigenous group who inhabited the municipalities of Los Santos, Jordán, and Cabrera in the Northwestern department of Santander and the Valle de Santiago, a native population of the Lower Magdalena region (Figure 4,Tables 3 and 4).

Figure 4 Principal Component Analysis (PCA) of Malambo, Valle de Santiago, Guane, and Checua communities. Individual MC corresponds to the experimental researcher of the present work. 

Human groups established in Malambo around 3120 years B.P. started a new model of settlement in the region and promoted horticulture including tuber cultigens such as cassava (Angulo, 1981; ^{Groot, 1989}). From an archeological perspective, this subsistence pattern, as well as the pottery style identified in Malambo, have been closely related to populations in the Venezuelan Orinoquia, specifically the Barrancoid series (Angulo, 1981; ^{Langebaek, 1987}; ^{Langebaek & Dever, 2000}; ^{Meneses & Godonés, 2005}; ^{Sanoja, 1978}). The results regarding the genetic distances at an interpopulation level seem to confirm the relationship proposed from the archeological perspective, specifically with the identification of sub-haplogroup B2j, which has been reported among present Venezuelan populations and indicates that this sub-haplogroup could be traced back in Northern South America up to approximately 3900 years B.P. (^{Gómez-Carballa, et al., 2012}). This date approaches the earliest occupation of the archeological site of Malambo, i.e., 3070±200 years B.P. (Angulo, 1981).

A predominance of the founding American mitochondrial haplogroups (A-D) has been reported in Colombia (^{Silva, et al., 2008}; ^{Casas-Vargas, et al., 2011}; Noguera, et al., 2015; ^{Díaz-Matallana, et al., 2016}; ^{Casas-Vargas, et al., 2017}). In contemporary populations, the most frequent haplogroup is A (48.3%), one of the most common in Mesoamerica, which supports those models proposing early Chibchan linguistic group migrations from Mesoamerica to South America. Haplogroup C (28.4%) has also been found in previous studies and, less frequently, haplogroups B (16%) and D (6%) (Noguera, et al., 2015). Population dynamics among the Paleo-Colombian populations has been somewhat different: in the Cundi-Boyacense high plains, a predominance of haplogroups A and B has been reported (^{Casas-Vargas, et al., 2017}; ^{Silva, et al., 2008}), as well as a higher frequency of haplogroup B in different Colombian regions with the continued presence of diverse haplogroups such as the B2 and the C1 in the Andes region (^{Díaz-Matallana, et al., 2016}), Norte de Santander (^{Romero-Murillo, 2016}), and Santander (^{Casas-Vargas, et al., 2011}).

We identified haplogroup B2 in all four individuals of the Malambo population. This haplogroup is widely distributed in the American continent, which has been interpreted as evidence of the genetic continuity that endures since the first settlers up to contemporary mestizo populations (^{Díaz-Matallana, et al., 2016}). Thus, haplogroup B2 has been associated with the founding genetic lineages in the continent along with haplogroups A, C, and D, which would be supported by the integrated migration model of the early populating of northwestern South America as evidenced by genetic, geological, archeological, and anthropological findings. This model proposes that the first settlers of northern South America migrated across the coastal route of the Pacific Ocean. It is assumed that these groups moved sequential movements through the Andes in a West-East direction across the Cauca and Magdalena Rivers valleys where they found the advantage of easy feeding resources and were also used as fluvial migratory routes (^{Díaz-Matallana, et al, 2016}). This influenced adaptation strategies and settlements in these ecosystems directly.

A specific sub-haplogroup B2j was identified in individuals from the archaeological Malambo site and also reported across the Venezuelan Caribbean northern coastline with no reference to the maternal ancestry (^{Gómez-Carballa, et al., 2012}). The G16361A mutation present in all four individuals has been confirmed as a diagnostic mutation for this sub-haplogroup. Based on this coincidence, even with the inherent limitations of a preliminary analysis of only four pre-Columbian individuals and their comparison with only one individual in contemporary Venezuela, we can postulate both a maternal affiliation and a migration process to or from the Orinoco River valley down or up the Lower Magdalena territories. This relationship amongst pre-Hispanic populations of the Colombian Caribbean and Lower Orinoco may also be confirmed by comparing the ceramic styles identified by archeologists in Malambo (Colombia) and Barrancas (Venezuela). These sites share some stylistic designs that seem to belong to the same cultural tradition established in the Caribbean from 3070±200 years B.P. up to 1270±150 years B.P. (Angulo, 1981; ^{Sanoja, 1978}).

Sub-haplogroup B2j is characterized by T131C, A183G, C5270T, A15824T, A16166G, A16183C, G16361A, and T16519G mutations. Its coalescence age was estimated at 3.9 ky (95% C.I.: 0-7.8) and, as such, it has been considered a recent clade within the phylogenetic tree of haplogroup B (^{Gómez-Carballa, et al., 2012}). This coalescence age coincides with the earliest occupational period reported for Malambo. Even so, the population of Malambo is differentiated from the Guane and the Checua in that it presents the specific G16361A mutation associated with the B2j sub-haplogroup identifying the Malambo as a whole separate group.

On the other hand, continuity in these sites of occupation might have been much wider than what has been registered up to this moment with an incidence in genetic relationships of the Lower Magdalena populations. Even if specific cultural tradition and lifestyle have been identified for the Malambo group, recent radiocarbon analysis in bone samples from a particular individual of this study (individual 3018) dated this individual between 1405 and 1445 A.D. [573-613 B.P.] (^{Rivera-Sandoval, 2018}). This chronology is linked to the late period groups characterized by a very similar ceramic style as compared with the most superficial layers in Malambo (Angulo, 1981). This has also been reported in other sectors of Lower Magdalena (Angulo, 1955; 1983; 1988; ^{Ramos & Archila, 2008}; ^{Rivera-Sandoval, 2018}). These ceramics have also been associated with Malibú populations (^{Falchetti, 1993}) differing from Malambo's incised modeling decoration but with similar technological features in the paste composition and clay temper.

Although the coincidence of the G16361A mutation corresponding to a B2j sub-haplogroup could result from an independent mutational event, the geographical proximity in northern South America of the corresponding individuals within +500 years suggests a matrilineal genetic continuity among Lower Magdalena populations. This eventual biological continuity might not necessarily be reflected on the cultural traditions represented in the type of cultural material present in the archeological archives nor in the sustenance ecosystems and resource exploitation, which during the late period were oriented to the exploitation of corn and other plants, hunting, and fishing, as well as fauna recollections associated to rivers and lakes. Genetics would thus offer new data for bioarchaeological studies in the New World, including the eventual maternal filiation and continuity of northern South American populations with the presence of genetically related individuals who underwent cultural mutations as demonstrated by their ceramic vessels.

Conclusions

Our data support the importance of aDNA contribution to archeological studies as they are complemented with further evidence to reconstruct mobility patterns and relationships between ancient pre-Hispanic groups in this region of South America. We implemented non-disruptive DNA cleansing showing a high reproducibility of results for every DNA extraction, which allowed a confident mitochondrial DNA analysis of several bone fragments.

Out of the four mitochondrial haplogroups (A, B, C, and D) previously assigned to the first settlers in America, i.e., an exclusive presence of B lineage, was determined in all four samples of the Malambo archeological site showing a strong association with a modern contemporary individual in Venezuela who shared the B2j sub-haplogroup. This preliminary finding in five individuals (four pre-Columbian Malambo and one 20^th century mestizo from Venezuela) can be associated with possible migratory routes, contacts, and exchanges that could have occurred in the last 3000 years as, from a molecular genetic perspective, they adjust to the coalescence age of the corresponding mutations. However, it is necessary to integrate the genetic data with additional genetic markers in these four pre-Columbian individuals and further contextualize this information with archeological data.