To the Editor:The Rh blood group locus comprises two closely linked genes, designated RHCE andRHD, encoding integral membrane proteins that carry the Cc/Ee and D antigens, respectively. The D− phenotype is usually due to the complete deletion of the RHD gene from the Rh locus.1 The D antigen is extremely immunogenic and is associated with hemolytic disease of the newborn (HDN). HDN occurs when antibodies from D− women, who have been sensitized to the D antigen, cross the placenta and react with antigens on fetal red blood cells. Polymerase chain reaction (PCR) assays to detect theRHD gene have been developed for determining fetal RHDgene type2-4 and paternal RHD gene dosage3 with potential clinical value in the management of pregnancies at risk for HDN. Such assays assume the RHD gene will be completely absent in D− serotypes. This appears valid generally for D− ce haplotypes of white origin (frequency [f] = .39) who account for the vast majority of D− phenotypes. However, exceptions have been reported from whites with the less frequent Ce and cE haplotypes (f = .0098 and .0119, respectively) and amongst nonwhites.5-10
We described two D− CCee+ white blood donors where the RHD gene was present in some form.5 One lacks RHD gene exons between 2 and 9 and would be correctly identified using multiplex PCR assays (unpublished observations, April 1996). We report here that the other donor, designated B1,5 carries a four-nucleotide deletion at a splice junction along an otherwise normal RHD gene that would prevent expression of the D antigen.
Total RNA extracted from whole blood buffy coat preparations was reverse transcribed into cDNA. This was used as template in PCR reactions to amplify four overlapping products, spanning the entireRHD gene, from the 5′ untranslated region (nucleotide −19) to the 3′ untranslated region (nucleotide 1536). Sequencing cDNA-derived PCR products showed a 4-base deletion between nucleotide positions 487 and 492 compared with two previously published RHD gene sequences11,12 (GenBank accession no. AF 037626). This corresponds to the intron 3/exon 4 boundary. Where the sequence ACAGACT was expected commencing at the 5′ region of exon 4 the sequence ACT was observed. Although the gene is still transcribed into a full-length mRNA transcript and the remainder of the sequence is normal, the reading frame is altered from nucleotide 488 and a premature stop codon introduced at positions 496-498.
To confirm that the four-nucleotide deletion was not a splicing error and was indeed present at the genomic level, genomic DNA from B1 was used to amplify across intron 3 into exon 4. The sense primer was common to both the RHCE and RHD genes (5′-TGC TGG TGG AGG TGA CAG-3′) and antisense primer (5′-GAA CAC GTA GAT GTG CAT CAT-3′) specific to the RHD gene. PCR products included a band at approximately 670 bp not present in D− controls. Sequencing with the antisense primer showed this band spanned the intron 3/exon 4 junction and confirmed that B1’s sequence lacked the four nucleotides corresponding to positions 488 to 491. It also provided an additional intron 3 sequence from which a new sense primer (5′-CAC CTC CTA AGT GAA GCT CTG-3′) was designed and used with the above antisense primer to amplify across the intron 3/exon 4 junction region in genomic DNA studies of family members. PCR products of 150 or 146 bp were expected for the normal and abnormal RHD genes, respectively.
Three distinct banding patterns were observed. A single PCR product was obtained for B1 and his two brothers who were also serologically D− and Ce+ with genotypes (D)Ce/Ce [where (D) is the phenotypically silent D gene]. This product migrated slightly faster than the single PCR product observed from D+ family members, including two with DCe/Ce genotypes and two genetically unrelated members with DCe/DcE and DCe/DCe genotypes. Unexpectedly, a third pattern was observed for the two heterozygote family members, who were the daughter of B1 and his brother, respectively. Both were D+ with the DCe/(D)Ce genotype. These showed a doublet corresponding to the expected 150 and 146 bp fragments, and a third slower migrating band shown to be a heteroduplex comprising the combined 150- and 146-bp bands. The genotype of each band was confirmed by sequencing.
Avent et al6 recently described a CCee individual with a single nucleotide substitution in the RHD gene at nucleotide 121 in exon 1, which results in an in frame stop codon. As for the mutation we describe above, the RHD gene was associated with the Ce haplotype and in both cases would have typed as RHD gene positive by current multiplex PCR assays potentially available for determining the risk of HDN. However, these rare occurrences would have little clinical significance. A fetus carrying these deletions would be typed as RHD gene positive and presumably the full precautions for monitoring potential antibody-induced red blood cell destruction would be used. However, these findings do add to the small but diverse array of genetic variations, other than complete D gene deletion, which can generate the D− trait.
Supported by the National Health and Medical Research Council of Australia, the Alexander Steele Young Memorial Lions Foundation, and the Brisbane North Regional Authority Liver Transplant Unit.