The regulatory mechanism of the ABO gene is not clear, especially in the hematopoietic compartment. Whilst much is known about the cause of weak A or B antigen expression on red blood cells in inherited weak ABO subgroups, less is understood about weakening of ABO during hematological disease and antigenic development during erythropoiesis. An enhancer element located 3.8 kbp upstream of exon 1 in the ABO gene contains either one (A1, O2) or four (A2, O1, O1v, B) 43-bp repetitive elements [
Total RNA was prepared from peripheral blood from apparently healthy donors [n=16] genotyped as homozygous for five common ABO alleles: A1, A2, O1, O1v and B. The O2 genotype was only available as heterozygotes (A1O2 and O1O2). Two TaqMan assays detecting ABO transcripts with probes located over the boundaries between exons 3–4 and 6–7 were used to quantify the level of transcripts. Analysis of a housekeeping gene assured the quality of the RNA preparations investigated. Surprisingly, no transcripts from the A1 or A2 alleles could be detected in the samples whereas B, O1, O1v and O2 transcripts were readily detectable. Since this finding did not correlate with the number of repeats in the enhancer element, DNA fragments encompassing 4,050 bp upstream of exon 1 and 2,000 bp downstream of the stop codon in exon 7 were sequenced to search for other polymorphisms that may explain the differences. In the 5′-region, 22 polymorphisms were detected of which 19 were not previously published. A single-bp deletion (at nt. −4036) in the B allele was the only difference between A2 and B alleles in this area. In the downstream region, which proved to be highly repetitive, a total of 45 new polymorphic sites were found of which 25 were specific for the O1v allele. A four-bp deletion 757–760 bp downstream of the stop codon was the only polymorphic site unique and common to A1 and A2 alleles, and hence potentially involved in the absence of A1 and A2 transcripts in mature blood cells.
In this study, we showed that transcripts from A alleles are lacking in peripheral blood from normal donors. None of the polymorphic sites in the upstream regulatory region appeared to be correlated with expression levels, suggesting that alterations of transcription factor binding sites does not explain the finding. The importance of the newly-found downstream deletion common to A1 and A2 alleles deserves further investigation as does measurement of ABO transcription levels during various stages of erythroid differentiation. In general, transcripts containing nonsense mutations are known to be more rapidly eliminated. In our study, O transcripts with the 261delG mutation were present in blood whilst "consensus" A1 transcripts were undetectable. The reason for this contra-dogmatic finding remains to be clarified.