The HLA profile and genetic affinities of three primitive Tamil-speaking endogamous groups: Kallars of Thanjavur, Piramalai Kallar and Vanniyar

Background: The present study was aimed to study the frequencies of HLA-DRB1/-DQB1 alleles and haplotypes of three endogamous groups of Tamil Nadu state, South India. PCR-SSP typing of HLA-DRB1 and -DQB1 alleles were performed on 111 Kallars of Thanjavur, 80 Piramalai Kallar of Madurai and 119 Vanniyar. Genetic distances, neighbor-joining phylogenetic dendrograms and correspondence analysis have been performed. Results: The HLA class II alleles, DRB1*07 (25.2%), DRB1*15 (15.7%), DRB1*14 (11.7%) and DRB1*12 (9.90%) among Kallars of Thanjavur; DRB1*15 (28.7

thought to be an important reservoir of genetic diversity and the source of at least three major waves of migration leading into Europe, the Americas, and India [1]. We recently established that two Tamil-speaking populations of the South India, the Mukkuvar, an endogamous group from coastal regions of South Tamil Nadu state, more closely related to ancient Hispanic and Guanche populations and the Valayar, a group inhabiting predominantly inland hilly regions with forest cover, are related to Austronesian and Micronesian populations [2].
The human leukocyte antigen (HLA) polymorphism in human populations has been studied for long to investigate immunogenetics, human genetic relationships and in reconstructing past migrational histories of populations. The extensive allelic variation among the HLA class I and class II genes distinguishes these as the most polymorphic coding sequence loci in the human genome [3]. The analysis of these genes has been a valuable tool in unraveling the historical relationships between ethnic groups and has greatly increased the knowledge of the ancestry and migration patterns of many populations including various endogamous groups in South India. South Indian populations are of interest for the study of genetic polymorphisms, because of the relatively long isolation of these populations from other ethnic groups. As Dobzhansky put it: 'South Indian caste systems are the biggest experiment ever done on Homo sapiens. ' However, in a recent South Indian population base study, an increased frequency of class II alleles DRB1*07, DRB1*15, DQB1*02, DQB1*05 and DQB1*06 have been reported. On the other hand, DRB1*01, DRB1*09 and DRB1*16 have been identified with low frequency, particularly among North Indian populations [4].
The purpose of this study was to analyze the alleles of the HLA class II (DRB1, DQB1) loci in three primitive Tamil-speaking Dravidian endogamous population groups such as Kallars of Thanjavur, Piramalai Kallar (of Madurai) and Vanniyar. Furthermore, the DNAs of most common two-locus haplotypes, such as DRB1*15-DQB1*06 and DRB1*07-DQB1*02 were typed for HLA-A, -B and -C alleles to identify the possible 5-locus extended haplotypes (EH) in these population groups. The frequencies of HLA-DRB1 and DQB1 alleles and haplotypes and phylogenetic relationships of these populations with nineteen other global populations are also constructed and discussed.

Study population
The Kallars of Thanjavur samples (n, 111) were collected from Thanjavur, Piramalai Kallar samples (n, 80) were collected from Madurai and Theni and Vanniyar samples (n, 119) were collected from Salem and Tiruvannamalai in the state of Tamil Nadu. Peripheral blood was collected via venipuncture from randomly chosen healthy individuals. Institutional ethical clearance was obtained from Madurai Kamaraj University (Ethical and Review Board Committee). Ethnicity was determined based upon current residence, place of birth, and family language history. Demographic data obtained included sex and age. Genomic DNA was extracted using salting out method as described previously [5].

Ethnographic notes Kallars of Thanjavur
Kallars are a wide-spread, ancient population living in southern parts of Tamil Nadu. Traditionally, they were described as semi agriculturists and semi warriors. Kallars are known to be the oldest immigrants of Neolithic period with Mediterranean racial elements [6]. The Chola country or Tanjore was their original abode of the Kallars, and from there they migrated to the Pandya Kingdom following its subjugation by the Cholas around eleventh century A.D [7]. The Cullaries are said to be in general a brave people, expert in the use of the lance and in throwing the curved stick called 'vullaree taddee' (Kallar Taddee or Valari) [8]. Portions of the Madurai and Thanjavur districts are divided into areas known as nadus, a name which is specially applicable to Kallar tracts [9]. In each "nadu" a certain caste, called the Nattan, is the predominant factor in the settlement of social questions which arise among the various castes living within the nadu (Pinnaiyur Nadu, Kasavala Nadu, Keel Vengai Nadu, Konoor Nadu, etc.). The Kallars of Thanjavur comprise an endogamous subgroup of the Kallar community exhibiting exogamy at the clan level while maintaining strict endogamy at the subgroup level (Kandiyar,Vandayar, etc.).

Piramalai Kallar
An endogamous subset of Kallars, called 'Piramalai Kallars' , live west of Madurai city (78.1E, 9.6N). In 1975, 95% of the population of Usilampatti and Thirumangalam taluks and 60% of Periyakulam taluks of Madurai district and 60% of Uthamapalayam taluk of Theni district were Piramalai Kallars, with a population size of approximately 350,000 [10]. In some old records they are referred to as Anaiyur Kallars which place was the stronghold of the Piramalai Kallars during the early years of their settlement [7].

Vanniyar
The Vanniyans, Pallis or Padaiyachis, are found in all the Tamil districts. The bulk of them are still labourers, but many now farm their own lands, while others are engaged in trade [11]. The Vanniyans are at the present time a small and obscure agricultural caste, but there is reason to believe that they are descendants of ancestors who, in former times, held a good position among the tribes of South India [9].
Vanniyas are mentioned in Ceylon archives. Wanni (Wavuniya, in the recent times) is the name of a district in Ceylon. It is situated towards Trincomalee in the northeast quarter [12]. The name "Vanniyan", seems to have been introduced by the Brahmans, possibly to gratify the desire of the Pallis for genealogical distinction. Padaiyachi means a soldier, and is also of late origin. After the fall of the Pallava dynasty, the Pallis became agricultural servants, and it is only since the advent of British rule that they have begun to assert their claims to a higher position [13].

Statistical analysis
Frequency of haplotypes of two and five loci was obtained by direct counting and expressed as indicated in the corresponding tables. The frequency data were converted into genetic distances in Arlequin (v3. 1). Dendrogram was constructed using Molecular and Evolutionary Genetics Analysis (MEGAv3.1) [15]. The genetic distances were also used for Principal Component Analysis (PCA) using GenAlEx 6 [16].

Hardy-Weinberg equilibrium examination
Hardy-Weinberg exact tests were performed on two HLA loci (DRB1 and DQB1). The observed, expected homozygosities, Chi-square (X 2 ) and the statistical P value are given in Table 2. Significant P-values were categorized by their levels (P < 0.0001, P < 0.0003) for convenience of discussion and whether the heterozygotes were in deficit or in excess. In order to make biologically meaningful comparisons, only heterozygotes which have a difference larger than two between the 'observed' and 'expected' counts were presented [17].

Haplotype frequencies
The two-locus haplotypes such as DRB1*07-DQB1*02    (Table 3). Based on the frequencies, the dominant two-locus haplotypes in these three populations were selected and typed for HLA-A, -B and -C locus alleles to find out the five-locus ancestral/ extended haplotype (AEH). The AEH represented more than five numbers was only presented for further analysis. The predominant five locus-extended haplotypes are presented in Table 4.  (Table 4).

Phylogenetic affinities
In order to assess the relationship between the populations, genetic distance (DA) was calculated and presented as neighbor-joining (NJ) tree and Principal Component Analysis. The neighbor-joining of the population relationships is presented in Fig. 2. The constructed NJ tree showed population relationships as a series of bifurcations, which are commonly interpreted as population splits.
The present study has been carried out to explore the utility of frequencies of alleles of HLA-DRB1 loci in inferring phylogenetic relationships. The NJ dendrogram revealed a unique genetic background of the study populations (Kallars of Thanjavur, Piramalai Kallar and Vanniyar). Based on the analysis, the N6 group may be genetically related to both N4 (Micronesian, Australian Aborigine and Austronesian) and N5 nodes (Melanesian) or they might reveal an admixture of Polynesian, N4 and N5 nodes. The Mediterranean populations are divided from N1 node and form a different cluster and cluster together with Black population (Fig. 1).
The PCA performed for HLA-DRB1 frequencies revealed the position of each population in two dimensions. The plot shows that the Kallars of Thanjavur belong to the same cluster as Austronesian, Australian Aborigine, indigenous groups and shared genetic similarities with the Mediterranean and Blacks. Similarly, Piramalai Kallar and Vanniyar belong to the same cluster as Austronesian and Australian Aborigine indigenous groups, Micronesian and Oriental indigenous populations. The PCA notably separated the admixed populations of Amerindian (who are Native Indians of the Americas), Mestizo (Spain and Latin America which originally meant to denote a person of combined European and Amerindian descent) and West Asia populations (Persian, Mulatto, Jew and Hispanic) from our study populations (Fig. 2).

Discussion
Southern India is one of the oldest geophysical regions of the world [18]. The majority of the people of southern India speak languages belonging to the Dravidian language family. The original name for the Dravidian family was 'Tamulic' , but the term 'Dravidian' was substituted by Bishop Caldwell, in order that the designation 'Tamil' might be reserved for the language of that name [19]. Studies on HLA polymorphism allow the delineation of their origin, geographic distribution patterns of these gene variants (alleles) and phylogenetic relatedness with world populations. In this study, the immunogenetic polymorphism of HLA-A/-B/-C/-DRB1*/-DQB1* alleles were genotyped by PCR using sequence specific primers (SSP). The present study on Kallars of Thanjavur, Piramalai Kallar and Vanniyar populations, three ancient and sympatrically isolated endogamous caste groups of Tamil Nadu, have revealed their unique HLA-DRB1/DQB1 allele and haplotype profiles. The HLA class II alleles DRB1*04, DRB1*07, DRB1*10, DRB1*12, DRB1*14 and DRB1*15, were found with higher frequencies in study populations of south India. However, these alleles were completely absent in some ethnic populations: such as the allele DRB1*04, was completely absent in Berber and Guanche; allele DRB1*07, was completely absent in Australian Aborigine,   (Table 4). Thus, the number of AEH is more in Kallars of Thanjavur than Piramalai Kallar and Vanniyar: this could be attributed to a number of socio-political and biological attributes. The Kallars of Thajavur were strategically located in a region of rich agricultural practice in the fertile plains of river Cauvery river water-fed area bordered on its eastern side by the Bay of Bengal coastal region. This region is more prone for biological / genetic amalgamation from outsiders, new immigrants into their region through coastal routes and also from plains of all direction (due to its fertile agriculture) that resulted in population admixture and genetic amalgamation. Also, Kallars of Thanjavur are less rigid in terms of their mate selection by having more number of marital clans (1300 clans), whereas the other two populations (Piramalai Kallar and Vanniyar) were relatively more rigid for genetic amalgamation (including miscegenation by socio-biological amalgamation) by way  of mate selection outside their group. Also, the marital clans in these later groups are comparatively less than the former. The alleles DRB1*01, DRB1*08, DRB1*09, DRB1*11 and DRB1*16, were found in low frequency in all the three study populations. Interestingly, these alleles have been reported with higher frequencies in different ethnic populations. The allele DRB1*01, was highly frequent in Berber (25.0%), Mulatto (25.0%), Caucasoid (23.45%), Persian (10.0%), Jew (10.0%) and Mestizo (9.37). The allele DRB1*08 showed higher frequencies in Polynesian (33.33%), Amerindian (32.81%), Melanesian We have analyzed and compared the DRB1* allele frequency results of 3 endogamous groups with 19 ethnic populations using neighbor-joining (NJ) method and Principal Component Analysis (PCA) (Figs. 1 and  2). The dendrogram analysis based on HLA-DRB1 profile, showed a unique genetic architecture of the south Indian study populations. These three populations groups have been connected to both N4 (Micronesian, Australian Aborigine and Austronesian) and N5 (Melanesian) nodes in one side, and joined with Polynesian cluster on the other side. The Australian continent holds some of the earliest archaeological evidence for the expansion of modern humans out of Africa, with the initial occupation ~ 40,000 years ago [21]. Austronesian are populations living mainly on islands in the north and east of New Guinea and coastal patches, and also in places as far away as Taiwan, Easter Islands, New Zealand and Madagascar, and in many densely populated areas such as Malaysia, Indonesia and the Philippines [22]. The origins of the Polynesians remain an enigma. Linguistic reconstructions of proto-Austronesian languages suggest a shared origin for Polynesians. Micronesians are from northern Borneo and Sulawesi [23] and Melanesians are from the Papua New Guinean coast (Madang), islands (Rabaul) and highlands (Goroka), and from New Caledonia and Fiji [24]. We recently reported that the phylogenetic affinities of South Indian populations with the South East Asia, most particularly with Austronesians possibly by forward migration of Indian population into South East Asia [2]. Perhaps, it was established that, the ancestral population from Africa left a genetic trail in Indian sub-continent en route Australia, in the distant past. The Mediterranean populations are divided from N1 node and form a different branch and clustered with Black. We have reported earlier that the Mukkuvar of Tamil Nadu (predominantly the eastern coastal inhabitants of South India) and Hispanic and Guanche populations share a very similar HLA genetic pool not withstanding with the fact of their distant geographic inhabitation [2]. The Guanches confirmed a North African origin and that they were genetically most similar to modern North African Berber people of the nearby North African mainland [25]. Some Guanches were known to be agriculturists and fish hunting communities in Spain. The Mukkuvar of south India were primarily fishers of east coast region of the country. These published evidences corroborate the HLA genetic affinities of Guanche, North African and South Indian Population.
The extended haplotype (EH), A*03-B*35-C*12-DRB1*07-DQB1*02 showed a higher frequency in the Kallars of Thanjavur (HF:0.1041) and Piramalai Kallar (HF:0.1000). Interestingly, this extended haplotype was not reported in any of the world population so far. However, A*24-B*35-C*04-DRB1*15-DQB1*06 haplotype observed in Kallars of Thanjavur (HF:0.1600) and Piramalai Kallar (HF:0.1166), already reported in Asian, Hispanic and Caucasoid populations (Table 4) Table 4). Such a complicated patterns of presence of different extended haplotypes in South India, compared to extended haplotypes from global populations throws interesting insights on the population dynamics and turn-over ostensibly influenced by frequent migrations and invasions during prehistoric and historic times.
A substantial amount of research has been conducted on the association of HLA polymorphisms with TB in different populations. HLA-DR2 is most consistently associated with TB in a diverse ethnic populations, including south Indian population [26,27], Polish [28], Thai [29], Indonesian and Russian [30]. Inter-population variations in HLA-TB associations have been established. DR14 was found at a significantly higher frequency among Iranian TB patients than controls [31]. HLA-DQB1*05:03 was found to influence TB progression in the Cambodian population [32], and DQB1*06:01 was associated with TB susceptibility in the South Indian, Thai and Uganda populations [26,29,33]. These alleles are very common in South Indian Population. Such an analysis for other highly frequent HLA alleles and haplotypes of south Indian population is worth exploring further.
The current level of diversity and the variation observed in allelic distributions for different populations could probably result from evolutionary forces that have changed as human populations have encountered new environments in their spread around the globe [34]. Maintenance of high levels of MHC polymorphism is crucial to counteract novel pathogenic challenges and to ensure long-term survival of organisms. Further, HLA alleles found in tropical countries tend to vary a lot from those in temperate parts of the globe, because the pathogens found there are different and highly divergent.

Conclusions
In conclusion, our results show a close relatedness among the study populations and other ethnic group of Oceania such as Polynesian, Melanesian, Micronesian, Australian Aborigine and Austronesian. The Pacific Islands are separate, isolated environments which may have hosted different pathogens from time immemorial and hence mounted differential evolutionary pressure on HLA genes. During the ancient past and at different times, when travel between the islands was not as frequent as it is at present, physical separation and limited inter population contacts may have contributed for the observed differences due to founder effects. But the Mediterranean populations when comparing with other ethnic populations are centrally located and in close vicinity with the Caucasian on one side of Southeast Europe and North African populations on the other side. Although, the Hispanic and Guanche populations were formed a unique ancestral node from the root of the dendrogram, represented that they did not settled in one place, often migrating along the coastal areas of different continents and most probably admixed with other local populations, amalgamated genes from neighboring populations and hence share a common HLA gene pool.