The RHD gene encoding the Rhesus blood group antigen D is flanked by two highly homologous DNA segments of about 9,000 bp, the upstream and downstream Rhesus boxes. In haplotypes with an RHD deletion, the fusion of the two Rhesus boxes within the 1,463 bp identity region generates the single hybrid Rhesus box. Its detection has been applied to determine RHD zygosity. For instance, testing for RHD heterozygosity is of clinical interest, because all children of an RHD homozygous father will be D positive and at risk of hemolytic disease of the newborn, if the mother carries an anti-D. Aberrant Rhesus boxes can confound this application and appear to be frequent among Africans. A comprehensive, systematic analysis was lacking. We sequenced about 5,850 bp of the upstream and downstream Rhesus boxes in 18 samples that were representative for all four D clusters and of the hybrid Rhesus box in 4 samples that were mistyped in assays for the hybrid Rhesus box. The known differences between upstream and downstream Rhesus boxes were in part restricted to subsets of RHD alleles. We detected 46 additional polymorphism, which represented single nucleotide substitutions, short insertions or deletions. Gene conversions were found in the upstream Rhesus boxes of RHDψ, DAU-1 and DAU-3 and in the downstream Rhesus boxes of Ccdes, weak D type 4.1, type 4.2 and DAU-0. All 6 different gene conversions started or ended in or near the identity region, which may have been instrumental for generating these gene conversions. Recombinations between haplotypes were likely in several alleles like DIII type 4. Four non-standard hybrid Rhesus boxes were suggestive of multiple independent RHD deletion events.
We found considerable variation of Rhesus box sequences, which explained the limitations for RHD heterozygosity testing that were previously observed in several laboratories. The general topology of the genomic Rhesus locus was however similar in all cases and consistent with our initial description. There was evidence for 5 probably independent RHD deletion events. Testing RHD heterozygosity in Africans is best based on quantitative PCR or amplification of the full length hybrid Rhesus box. Because in Europeans who harbored RHD alleles belonging to African D clusters the Rhesus box variation was also substantial, use of two complementary methods for hybrid Rhesus box detection may even be advisable in whites. We conclude that a long range PCR spanning the whole hybrid Rhesus box, while technically demanding, is not influenced by the kinds of aberrations observed in this study and may hence remain the standard for research purposes and quality assurance.