Deciphering the fine nucleotide diversity of full HLA class I and class II genes in a well‐documented population from sub‐Saharan Africa

Abstract

With the aim to understand how next‐generation sequencing (NGS) improves both our assessment of genetic variation within populations and our knowledge on HLA molecular evolution, we sequenced and analysed 8 HLA loci in a well‐documented population from sub‐Saharan Africa (Mandenka). The results of full‐gene NGS‐MiSeq sequencing compared with those obtained by traditional typing techniques or limited sequencing strategies showed that segregating sites located outside exon 2 are crucial to describe not only class I but also class II population diversity. A comprehensive analysis of exons 2, 3, 4 and 5 nucleotide diversity at the 8 HLA loci revealed remarkable differences among these gene regions, notably a greater variation concentrated in the antigen recognition sites of class I exons 3 and some class II exons 2, likely associated with their peptide‐presentation function, a lower diversity of HLA‐C exon 3, possibly related to its role as a KIR ligand, and a peculiar molecular diversity of HLA‐A exon 2, revealing demographic signals. Based on full‐length HLA sequences, we also propose that the most frequent DRB1 allele in the studied population, DRB1*13:04, emerged from an allelic conversion involving 3 potential alleles as donors and DRB1*11:02:01 as recipient. Finally, our analysis revealed a high occurrence of the DRB1*13:04‐DQA1*05:05:01‐DQB1*03:19 haplotype, possibly resulting from a selective sweep due to protection to Onchorcerca volvulus, a prevalent pathogen in West Africa. This study unveils highly relevant information on the molecular evolution of HLA genes in relation to their immune function, calling for similar analyses in other populations living in contrasting environments.

HLA‐C genetic diversity and evolutionary insights in two samples from Brazil and Benin

“…It is not clear whether these results reflect a hitchhiking effect due to the high LD throughout HLA‐C or direct selection on these regions. Moreover, Goeury and colleagues also reported positive Tajima′s D values for all exons except exon 6 in another sub‐Saharan sample, with the highest value located at exon 2, as detected here in our Brazilian and Beninese samples 58 . Balancing selection was described for classical HLA coding regions at exons 2 and 3, enhancing antigen presentation capabilities, and hardly any demographic or genetic factors can explain the high degree of polymorphism and excess of nonsynonymous variants at these genes (reviewed by 60 ).…”

“…For the Brazilian sample, the CDS (and genomic) allele frequencies are similar to the ones reported in previous studies using different typing techniques 56,57 (http://www.allelefrequencies.net/). For Beninese, the CDS frequencies are similar to the ones reported in another sub‐Saharan African population sample from Senegal, West Africa, named Mandenka, as reported elsewhere 58 . This is the first survey addressing Beninese HLA‐C genetic diversity.…”

“…No evidence was detected when specific codons within exon 1 are evaluated using FUBAR. Similar results have been described in Mandenka, in which all exon 1 variants configure nonsynonymous mutations 58 . This is intriguing because exon 1 encodes the HLA‐C leader peptide and it is primarily involved in targeting HLA‐C to the cellular membrane, and it is not part of the mature HLA‐C molecule.…”

Mandenka from Senegal

“…. Actually, the results indicate the existence of one very frequent 3-locus haplotype, DRB1*13:04~DQA1*05:05:01~DQB1*03:19, that may have increased in frequency through selective sweep 5 . Whereas no relevant information is currently available for the 2 DQ alleles, DRB1*13:04 has formerly been reported as frequent in Gambia.…”

“…Accordingly, whereas no class I haplotypes reach a frequency higher than 7%, several 2‐locus class II haplotypes (in positive linkage disequilibrium) are observed at a very high frequency (up to 43% for DQA1*05:01:01 ~ DQB1*03:19 ). Actually, the results indicate the existence of one very frequent 3‐locus haplotype, DRB1*13:04 ~ DQA1*05:05:01 ~ DQB1*03:19 , that may have increased in frequency through selective sweep. Whereas no relevant information is currently available for the 2 DQ alleles, DRB1*13:04 has formerly been reported as frequent in Gambia .…”