EXAMINATION OF HLA-DP INTRON/EXON VARIATION IDENTIFIES TWO DPB1 EVOLUTIONARY GROUPS INCLUDING INTRON 2 STR VARIANTS THAT MAY REGULATE DPB1 EXPRESSION LEVELS

Recent studies showed the existence of biologically relevant DPB1 expression variants that associate with two 3 ′ UTR SNPs of DPB1. The knowledge of sequence variation in DPA1 and DPB1 alleles is limited to a few exons. Only a few alleles are fully sequenced resulting in limitations for DP typing by NGS. We developed robust and efficient methods for long range amplification and molecular cloning that allowed us to examine and characterize numerous alleles. Clones were sequenced by NGS and consensus was generated by a novel de novo assembly algorithm. The fragments generated span exon 1–4 and 2–4 of DPA1 and DPB1. We characterized 16 DPA1 and 40 DPB1 alleles defined but their unique sequences in 63 samples. We obtained novel sequence information in 11 DPA1 alleles including 7 extended sequences, 3 novel intron variants and 1 novel exon variant. In DPB1 we extended the coverage in 20 allleles and identified 12 novel intron variants and 2 novel exon variants. In addition to the well known variations in exons, the long DPA1 intron1 (3556–3590 nt) showed 201 SNPs and 17 indels. DPA1 intron 2 (340 nt) and intron 3 (214 nt) are short, not variable in size with low polymorphism (12 and 9 SNP, respectively). In DPB1 intron 3 is conserved in length (547 nt) and presents only 17 SNPs. Intron 2 of DPB1 (3952–4025 nt) has the largest variability with 176 SNPs and 9 indels; in addition this intron has one variable STR region located at 44 nt before the beginning of exon 3 with 4–14 repeats of the tetramer AAGG. Two groups of alleles that have only 4 repeats (short STR) and 8–14 repeats (long STR) were identified. Also, phylogenetic trees define two distinct groups at each, DPA1 and DPB1 loci. The DPB1 clusters are tighter when exon 2 is excluded from the analysis and correlate well with the intron 2 STR short and long variants. We observed that the 3 ′ UTR variants of DPB1 that correlate with high and low expression, also associate strictly in a biunivocal fashion with the short and long intron 2 STR variants, respectively. Our studies provide useful information for mapping reads in NGS typing. In addition our findings identify potential regulatory elements that may affect gene expression possibly by altering the rate of mRNA splicing. In transplantation, the expression levels of the mismatched allele appear to determine both immunogenicity and outcome.