A "silent" nucleotide substitution in exon 4 is responsible for the "alternative expression" of HLA-A*01:01:38L through aberrant splicing.

A Welsh Bone Marrow Donor Registry donor was serologically typed, using both alloantisera and monoclonal antibodies, as human leukocyte antigen (HLA)-A2, A-, but typed by polymerase chain reaction sequence-specific priming as HLA-A*01, A*02. Full gene sequencing of the A*01 separated allele indicated an apparently normal A*01:01:01:01 apart from a silent change at nucleotide 705 in exon 4, codon 211 (alanine: normally GCG but GCA in this donor). Sequence analysis of the amplified A*01 allele in cDNA synthesized from RNA indicated that exons 1, 2, 3, and 5 had typical A*01:01 sequences. However, exon 4 was truncated in this allele (87 nucleotides shorter), beginning just after the single nucleotide polymorphism (SNP) identified in genomic DNA sequencing. The nucleotide sequence up to, and 1 nucleotide after, the SNP is homologous with the 3' end of human leukocyte antigen (HLA)-A intron 3 and thus resembles a splice site. However, a small amount of "normal" HLA-A1 was detected on the surface of cells from an Epstein-Barr virus transformed B-cell line (BCL), but not on peripheral blood mononuclear cells, by flow cytometry. Additionally, a trace amount of "normal sized" A*01 was amplified from cDNA. We suggest that in this A*01 variant allele (A*01:01:38L) intron 3 is largely spliced out with a part of exon 4; exon 4 is still in-frame but the protein is smaller than the wild type. This is likely to affect folding and assembly of the "wild type" mature protein on the cell surface, thus explaining the apparent null phenotype when assayed by conventional serology. However, a small amount of A1 protein is made from correctly spliced A*01 mRNA and is detectable on BCLs using flow cytometry.

One-tube HLA-B27/B2708 typing by flow cytometry using two "Anti-HLA-B27" monoclonal antibody reagents.

BACKGROUND: Flow cytometry-based methods are widely used to detect the ankylosing spondylitis-associated HLA-B27/B2708 antigens. However, the generally used "HLA-B27" monoclonal antibodies (moabs) cross-react with many HLA specificities, including the common HLA-B7 antigen. Thus, using two "B27" moabs is highly recommended. METHODS: The assay used two "HLA-B27" reagents, FITC and PE conjugated, respectively and a PE-Cy5 anti-CD3 antibody. Assay verification used 51 reference subjects possessing B*2705, B*2702, and B*2708 and a range of cross-reactive HLA antigens. A total of 1,006 consecutive patients' samples, referred for "HLA-B27 typing", were assayed alongside our standard flow cytometry method. A further 12 low frequency HLA-B*27 specificities were tested. Samples reacting with one "B27" moab only were B*27 allele typed by PCR using sequence-specific primers. RESULTS: All patient B27/B2708 positives (28.3%) were identified by our one-tube method which detected B*2705, B*2702, B*2708, and 8/12 other B*27 specificities. It was unaffected by HLA-B7 and other cross-reactive antigens but required a minor adjustment, a reduction in the volume of one of the "B27" moabs used, to avoid detecting a minority of HLA-B57 subjects. CONCLUSIONS: Our one-tube B27/B2708 assay is simple, robust, uses two "B27" moabs for typing precision and security, does not suffer from interference by HLA-B7 or other cross-reactive antigens and has the obvious advantage of using a single tube per typing.