MHC diversity in Caucasians, investigated using highly heterogeneous noncoding sequence motifs at the DQB1 locus including a retroviral long terminal repeat element, and its comparison to nonhuman primate homologues

Long terminal repeats (LTRs) are common retrovirus-related sequences spread throughout the human genome. We previously reported the human-specific integration of one LTR (DQLTR3) located 15 kb upstream of HLA DQB1. To elucidate the contribution of retroviral sequences to the variability and phylogenetic background of HLA DQB1 we investigated another LTR (DQLTR13), located 1.3 kb upstream of HLA DQB1, in German families, great apes, and Old World monkeys. Within German families, DQLTR13 presence was strongly linked to HLA DQB1*0302, *0303, and *0402 haplotypes. All other haplotypes had a low frequency or were devoid of DQLTR13. Phylogenetic analysis of DQLTR13 and adjacent nucleotide sequences in humans and non-human primates revealed a high degree of similarity and recent origin of HLA DQB1*0302, *0303, and *0402. Nevertheless, two lineages leading to DQB1*0301 and *0302 were generated by an ancient split of a DQB1*0301, *0302 progenitor. A third lineage consisting of DQB1*05/*06-related sequences may have evolved from the DQB1*0302 lineage, and a DQB1*0201-related sequence shared common ancestry with DQB1*0301. Among the human haplotypes, HLA DQB1*0201 and *0301 are linked to two different DQA1 alleles. Based on the small genetic distance of DQLTR13 as well as the adjacent sequences on these haplotypes, we suggest that a recent recombination is responsible for these associations. In the analysis of nonhuman primate species, we detected DQLTR13 in two lowland gorillas, dating the integration at at least 8 million years ago. We therefore conclude that noncoding sequences up to 1.3 kb upstream of DQB1 provide novel insight into the generation of MHC gene diversity.