HLA-B67: a member of the HLA-B16 family that expresses the ME1 epitope

A new strategy for systematically classifying HLA alleles into serological specificities

HLA serological specificities were defined by the reactivity of HLA molecules with sets of sera and monoclonal antibodies. Many recently identified alleles defined by molecular typing lack their serotype assignment. We surveyed the literature describing the correlation of the reactivity of serologic reagents with AA residues. 20 - 25 AA residues determining epitopes (DEP) that correlated with 82 WHO serologic specificities were identified for HLA class I loci. Thirteen DEP each located in the beta-1 domains that correlated with 24 WHO serologic specificities were identified for HLA-DRB1 and -DQB1 loci. The designation of possible HLA-DPB1, -DQA1, -DPA1, and additional serological specificities that result from epitopes defined by residues located at both -DQA1 and -DQB1 subunits were also examined. HATS software was developed for automated serotype assignments to HLA alleles in one of the three hierarchical matching criteria: 1) all DEP (FULL); 2) selected DEP specific to each serological specificities (SEROTYPE); 3) one AA mismatch with one or more SEROTYPES (INCOMPLETE). Results were validated by evaluating the alleles whose serotypes do not correspond to the first field of the allele name listed in the HLA dictionary. Additional 85 and 21 DEP patterns that do not correspond to any WHO serologic specificities for common HLA class I and DRB1 alleles were identified, respectively. A comprehensive antibody identification panel would allow for accurate unacceptable antigen listing and compatibility predictions in solid organ transplantations. We propose that antibody-screening panels should include all serologic specificities identified in this study. This article is protected by copyright. All rights reserved.

HLAMatchmaker-Based Analysis of Human Monoclonal Antibody Reactivity Demonstrates the Importance of an Additional Contact Site for Specific Recognition of Triplet-Defined Epitopes

Five HLA-A3 reactive human monoclonal antibodies (mAbs) originating from a parous woman were screened against HLA-typed panels by means of complement-dependent lymphocytotoxicity, high-definition ELISA, and flow cytometry with single antigen beads. Antibody reactivity profiles were compared with triplet amino acid sequence polymorphisms identified by HLAMatchmaker, and a three-dimensional structural modeling program (Cn3D of the National Center for Biotechnology Information) was used to determine the topography of epitopes recognized by each mAb. These mAbs originated from a woman who during pregnancy developed antibodies to the paternal HLA-A3 antigen of her child. Each mAb was specific for one mismatched triplet on HLA-A3, and the reactivity patterns of these IgM-type mAbs were practically the same in lymphocytotoxicity and antigen-binding assays. One mAb was specific for 163dT, a unique triplet present only on A3. The other mAbs reacted with 62Qe, 142mI, or 144tKr; these triplets are present on different groups of HLA-A alleles, some of which, however, did not react. Topographic modeling of triplet-defined epitopes identified clusters of polymorphic surface residues that were shared between reactive alleles. These clusters may serve as primary contact sites for the specificity-determining complementarity-determining region (CDR) loops of antibody. The reactivity with these mAbs required also the presence of self-sequence elsewhere on the HLA molecular surface as a critical secondary contact site for antibody, likely through another CDR loop. For instance, the reactivity of the 62Qe-specific mAb required the presence of a glycine residue in position 56 and the reactivity of the 142mI-specific mAb required the presence of the GTLRG sequence in positions 79-83. Conversely, there were many other amino acid differences between the mAb-reactive alleles and HLA-A3 that did not prevent antibody binding. For instance, the 62Qe-specific mAb-reactive alleles had 35 and the 142mI-reactive alleles had 50 of such "permissive" residue differences. An HLAMatchmaker-based analysis of the reactivity of human mAbs will increase our understanding of the structural definition of HLA epitopes and their reactivity with alloantibodies.

The nature of introns 4-6 suggests reduced lineage specificity in HLA-B alleles

For most HLA-B alleles, coding sequences of the 3' part of the genes still need to be determined, and sequences of the 3' noncoding regions have yet to be studied systematically. In this study, we have determined the sequences of introns 4-6 in all HLA-B allelic groups, and computed nucleotide substitution rates and phylogenetic relationships. These sequences demonstrated an inconsistent pattern of intralineage specificity, intralineage diversity, and interlineage diversity that is best characterized by a patchwork pattern. Apart from phylogenetic studies about HLA diversity and diversification, the sequence data obtained in our study may prove valuable for haplotype-specific sequencing of the 3' part of HLA-B and for the explanation of recombination events in newly described HLA-B alleles.

HLAMatchmaker: a molecularly based algorithm for histocompatibility determination. I. Description of the algorithm

This report describes an algorithm for identifying acceptable HLA antigens for highly alloimmunized patients without the need for extensive serum screening. This algorithm is based on the concept that immunogenic epitopes are represented by amino acid triplets on exposed parts of protein sequences of human leukocyte antigen chains (HLA-A, HLA-B, and HLA-C) accessible to alloantibodies. A computer program (HLAMatchmaker) has been developed to determine class I HLA compatibility at the molecular level. It makes intralocus and interlocus comparisons of polymorphic triplets in sequence positions to determine the spectrum of non-shared triplets on donor HLA antigens. In most cases is it possible to identify certain mismatched HLA antigens that share all their polymorphic triplets with the patient's HLA antigens and could therefore, be considered fully compatible. HLAMatchmaker permits also the identification of additional mismatches that are acceptable as determined from the triplet information on HLA-typed panel cells that do not react with patient's serum.HLAMatchmaker provides an assessment of donor-recipient HLA compatibility at the structural level and this algorithm is different from conventional methods based on the mere counting of numbers of mismatched HLA antigens or CREGs. This donor selection strategy is suitable especially for allosensitized patients in need of a compatible transplant or platelet transfusion.

Novel HLA-B alleles formed by an inter-locus recombination with HLA-C, HLA-B*0713 and B*6702

This paper describes two novel HLA-B locus alleles. B*0713 appears to have resulted from an interlocus recombinational event which inserted a sequence from an HLA-C allele into B*0702. This event altered a significant portion of the alpha-1 domain alpha helix. B*6702 shares much of its exon 2 sequence with B*0713 but is identical in exon 3 to several HLA-B locus alleles including B*67012.

Error rate for HLA-B antigen assignment by serology: implications for proficiency testing and utilization of DNA-based typing methods

Until recently, the majority of HLA class I typing has been performed by serology. Expensive commercial typing trays are frequently used for testing non-Caucasian subjects and new strategies using DNA-based methods have been adopted for improving clinical histocompatibility testing results and adapted as supplements in proficiency testing. A double-blind comparison of the typing of HLA-B specificities in 40 samples was carried out between serology and two polymerase chain reaction (PCR) methods, PCR amplification with sequence-specific primers (PCR-SSP) and PCR amplification and subsequent hybridization with sequence-specific oligonucleotide probes (PCR-SSOP). The results demonstrated 22.5% misassignments of HLA-B antigens by serology. There was complete concordance between the results obtained with the two PCR based typing methods. A second panel of 20 donor samples with incomplete or ambiguous serologic results was analyzed by PCR-SSP and SSOP Both PCR methods identified correctly the HLA-B antigens. Our results suggest that more accurate typing results can be achieved by complementing serologic testing with DNA-based typing techniques. The level of resolution for HLA-B antigen assignment can be obtained by this combination of serology and limited DNA-based typing is equivalent to the HLA-B specificities defined by the WHO-HLA Committee. This level of resolution cannot routinely be achieved in clinical histocompatibility testing or in proficiency testing using serologic reagents only.

Interactions of HLA-B*4801 with peptide and CD8

Functional properties of the B*4801 allotype were investigated using HLA class I-deficient 221 cells transfected with B*4801 cDNA. From pool sequence analysis of endogenously bound peptides, B*4801 was shown to select for nonamer peptides having glutamine or lysine at position 2 and leucine at the carboxyl-terminus. In an in vitro cell-cell binding assay, B*4801 binds CD8 alpha homodimers weakly due to the presence of a threonine residue at position 245 in the alpha 3 domain. A mutant B*4801 molecule in which alanine replaces threonine 245, binds CD8 alpha homodimers at levels comparable to those of other HLA class I allotypes. Despite the low affinity of B*4801 for CD8 alpha, alloreactive T-cells that recognize B*4801 molecules expressed by the 221 transfectant are inhibited by anti-CD8 monoclonal antibodies. Analysis of 25 B*48-expressing individuals from various populations showed threonine 245 was encoded by every B*48 allele.

The novel HLA-Cw*1802 allele is associated with B*5703 in the Bubi population from Equatorial Guinea

The HLA-Cw*1801 specificity, a Cw7/Cw4 hybrid allele, has recently been described in association with B*8101 (formerly B"DT"). In this study, the new Cw*1802 variant, differing from Cw*1801 at exon 5, is found associated with B*5703 in Bubi individuals from Equatorial Guinea. Confirmatory complete coding regions of B*5703 and B*3910 are also reported.

A novel recombinant HLA-B*39 allele (B*3910) in a South African Zulu

The sequence of a new B*39 allele has been identified in a South African Zulu individual. This allele designated B*3910, differs at two nucleotide positions (246 and 272) from B*39011. The difference at position 246 is silent, while that at position 272 results in an amino acid change from cysteine (B*39011) to tyrosine (B*3910). As these same differences are found in other HLA-B alleles, they were probably introduced into the B*3910 sequence by a short gene conversion event with another allele. This finding provides further evidence for the diversification of HLA-B allelic sequences via recombination.

Characterization of the peptide-binding specificity of HLA-B*7301

Previous studies showed the human MHC class I heavy chain HLA-B*7301 has a sequence very divergent from other class I alleles. Despite the unusual sequence, we predicted B*7301 would retain the peptide-binding function typical of other HLA-A, B and C glycoproteins, and sequence similarity to B*2705 in a region of the peptide-binding site known as the B pocket suggested B*7301 would bind peptides with Arg at position 2. To test this hypothesis, the peptide-binding specificity of B*7301 was investigated. Sequence analysis of peptides bound endogenously by B*7301 indeed found selectivity for nonamer peptides possessing Arg at position 2 and a preference for small nonpolar residues such as Pro or Ala at the C terminus was also revealed. B*7301 therefore possesses the potential to function as a conventional antigen presenting class I glycoprotein. Functional similarities between B*7301 and B*2705 are discussed in the context of the association of B*27 subtypes with susceptibility to ankylosing sponylitis and arthritic diseases.

Novel HLA-B alleles, B*8201, B*3515 and B*5106, add to the complexity of serologic identification of HLA types

Three class I alleles, B*8201, B*3515 and B*5106, have been described using DNA and cDNA sequencing. The B*8201 allele is most structurally related to B*5602, differing from it by 14 nucleotide substitutions resulting in 5 amino acid differences. The other two alleles, B*3515 and B*5106, differ from their most closely related HLA-B alleles by 2-3 nucleotide substitutions resulting in 1-2 amino acid substitutions, respectively. The majority of nucleotide substitutions marking these new alleles are observed in other HLA-B alleles suggesting that gene conversion and/or reciprocal recombination have created this diversity. All of the amino acid substitutions are predicted to alter the antigen binding site of the HLA-B molecule. The newly defined HLA-B allelic products were originally defined by their unusual serologic reactivity patterns. The B*8201 allelic product is serologically typed as a B "blank" or as a variant of B22 or B45. These patterns and the serologic reactivity of the other newly described allelic products are consistent with the protein sequence homology among specific HLA-B molecules. While serology remains a powerful tool for detecting HLA diversity, alleles generated by events resulting in the sharing of HLA sequence polymorphisms among alleles at a locus will continue to create complexity in the interpretation of typing results.

Modification of an HLA-B PCR-SSOP typing system leading to improved allele determination

Modifications have been introduced to a previously reported HLA-B PCR-SSOP typing system. This has enabled further definition of alleles, determination of the probe pattern of some alleles not previously examined and identification of patterns of possible new alleles. However there are still some alleles that cannot be differentiated and there are several alleles which when present as a homozygote have the same pattern as in combination with another allele. When the method was applied to the typing of 66 consecutive cadaveric donors there were three donors whose type differed from the serological type.

Population diversity of B-locus alleles observed by high-resolution DNA typing

HLA B-locus typing by group-specific PCR and hybridization with SSOP was performed in 81 10th IHWS B cell lines and 334 selected subjects of our local panel, from four ethnic groups. Most of the B-locus serological specificities were well defined. However, some antigens like B41, B58, B56, the splits of B14, and some subtypes of B5, were not accurately assigned by serology. In the panel studied, we found 17 hybridization patterns that corresponded to probable new alleles. New patterns occurred in the four ethnic groups examined. Multiple subtypes of B35, B5, B15, B41, B44, B57, B58, B70, B14, B40, B22 were found in subjects of the same ethnic group. In view of the poor serological definition of some alleles, and the occurrence of multiple subtypes in the same ethnic population, it appears that high resolution B-locus typing may be an important addition for detection of potentially relevant HLA incompatibilities in transplantation. It should also be valuable for population studies and for the investigation of HLA associations with diseases.

Overlap in the repertoires of peptides bound in vivo by a group of related class I HLA-B allotypes

BACKGROUND

Polymorphism among class I molecules of the major histocompatibility complex (MHC) confers allotypic specificity on the peptides that these molecules bind and present to cytotoxic T lymphocytes. Evolution of new human HLA class I alleles usually involves gene recombination events that replace a segment of one allele with the homologous region of another. In this study, the impact of these evolutionary changes has been assessed by comparison of the peptide-binding specificities of six related HLA-B allotypes.

RESULTS

Endogenous peptides bound by HLA-B*5401, HLA-B*5501, HLA-B*5502, HLA-B*5601, HLA-B*6701 and HLA-B*0702 were characterized. Despite differing by 1-9 of the amino-acid residues comprising their peptide-binding sites, all these allotypes share a dominant preference for peptides that have proline at position 2. Polymorphism results in differing selection of carboxy-terminal and secondary anchor residues, but the peptide-binding specificities are sufficiently similar that there is overlap in the repertoires of peptides bound by these allotypes. Complete sequence determination of individual peptides revealed four that could be isolated from two or more allotypes. Members of the closely related HLA-B22 family--HLA-B*5401, HLA-B*5501, HLA-B*5502 and HLA-B*5601--show only minor differences in their peptide-binding specificities. This marked similarity is reflected at the functional level, as alloreactive cytotoxic T lymphocytes generated against HLA-B*5401 and HLA-B*5501 exhibited cross-reactive recognition.

CONCLUSION

The isolation of identical endogenously bound peptides from six HLA-B allotypes demonstrates overlap in the repertoires of peptides bound in vivo by different allotypes. We speculate that the shared preference for binding peptides with proline at position 2 reflects a selective pressure to retain this specificity, which may be based upon peptide availability in vivo. Characterization of the overlap between the repertoires of peptides bound by HLA-B allotypes could simplify the development of peptide-based vaccines that are targeted to cytotoxic T cells, as single peptides would be effective for humans of different HLA types.

Alloreactive cytotoxic T-lymphocyte-defined HLA-B7 subtypes differ in peptide antigen presentation

We investigated T-cell-defined HLA-B7 subtypes using cDNA sequencing, analysis of bound peptides, and reactivity with a panel of alloreactive cytotoxic T-lymphocyte (CTL) clones. Three subtypes (HLA-B*0702, HLA-B*0703, and HLA-B*0705) differ in nucleotide and predicted amino acid sequence. CTL reactivity and pooled peptide sequencing show that these three HLA-B7 subtypes bind distinct but overlapping sets of peptides. In particular B*0702 expresses D pocket residue Asp 114 and binds peptides with P3 Arg, whereas B*0705 expresses D pocket residue Asn 114 and binds peptides with P3 Ala, Leu, and Met. Consistent with different peptide-binding specificities, three alloreactive CTL differentiate between cells expressing B*0702, B*0703, and B*0705 by detecting specific peptide/HLA-B7 complexes. In contrast, three other T-cell-defined HLA-B7 subtypes are identical to HLA-B*0702. The B*0702-expressing cell lines are differentiated by two of ten CTL clones. One CTL clone differentiates B*0702-expressing cells by their ability to present peptide antigen. Thus differences in peptide presentation can explain differential CTL recognition of cell lines expressing structurally identical and variant HLA-B7.

HLA-B16 antigens: sequence of the ST-16 antigen, further definition of two B38 subtypes and evidence for convergent evolution of B*3902

The ST-16 antigenic specificity of the HLA-B locus is defined as a B39 variant of Mexican-Americans. Nucleotide sequencing of cDNA shows the ST-16 allele (B*3905) differs from B*39011 by a single substitution that substitutes tyrosine for aspartic acid at position 74 of the mature class I heavy chain. The complete coding region sequence for the common caucasoid allele encoding the B38 antigen has been determined. This B*3801 allele differs from B*3802 at two nucleotide substitutions within the Bw4 sequence motif. B*3801 and B*3802 may have been derived independently from B*39011 by conversion events with B alleles donating distinctive Bw4 motifs. A novel allele B*39022 derived from a Colombian Indian differs from the B*39021 allele of Japanese origin at two widely separated silent substitutions. Comparison of sequences for the known B16 alleles suggest that B*39021 and B*39022 were independently derived by recombination from B*39013 and B*39011 respectively.

A small test of a sequence-based typing method: definition of the B*1520 allele

Santamaria et al. (Human Immunology 1993 37: 39-50) describe a method of sequence-based typing (SBT) for HLA-A, B and C alleles said to give "unambiguous typing of any sample, heterozygous or homozygous, without requiring additional typing information". From SBT analysis, which involves determination of partial sequences of mixed alleles, these investigators reported that cell lines KT17 (HLA-B35,62) and OLGA (HLA-B62) from the reference panel of the 10th International Histocompatibility Workshop express novel variants of HLA-B15 (B1501-MN6) and HLA-B35 (B3501-MN7) respectively. To study further the novel alleles, we cloned and sequenced full-length HLA-B cDNA clones isolated from the KT17 and OLGA cell lines. We find that KT17 expresses B*3501, as assigned by SBT, and B*1501, the common allele encoding the B62 antigen. We were unable to confirm that KT17 expresses the novel B1501-MN6 variant identified by SBT. For OLGA our analysis confirms the partial sequences obtained by SBT. Thus OLGA expresses B*1501 and a novel HLA-B allele. The complete sequence of the latter shows it is a hybrid having exons 1 and 2 in common with B*1501 and other B15 subtypes and exons 3-7 in common with B*3501 and related molecules including B*5301 and B*5801. The novel allele has been designated B*1520 because of its sequence similarity with the B15 group; furthermore, serological analysis shows that the B*1520 product does not express epitopes in common with either B35, B53 or B58. The B*1520 heavy chain has a similar isoelectric point to A*3101; B*1520 was undetected by previous applications of isoelectric focusing because B*1520 and A31 are both expressed by OLGA. In conclusion, HLA-B typing of two cell lines by cDNA cloning and sequencing gives concordant results with SBT for three of the four alleles. The cause of the discrepancy for the fourth allele is unknown, however, this finding indicates that the novel HLA-A, B and C sequences emerging from SBT studies need independent verification.

Low-resolution DNA typing for HLA-B using sequence-specific primers in allele- or group-specific ARMS/PCR

The products of the human major histocompatibility complex (HLA Class I and II) have historically been detected using serological or cellular assays. With the availability of DNA sequence information for alleles of the HLA system, and with the development of molecular biological techniques it has become possible to tissue type for allelic differences in the HLA genes themselves. We describe here a polymerase chain reaction (PCR) system, based on the principle of the amplification refractory mutation system (ARMS), for low-resolution DNA typing of the HLA-B gene. The technique involves a one-step PCR from genomic DNA using sequence-specific primers in particular combinations that determine the specificity of each reaction. A low-resolution primer panel has been designed, based on published HLA-B gene nucleotide sequences, consisting of 34 sequence-specific primers (SSP) in 24 PCR reactions which cover all known HLA-B alleles, to give allele-specific or group-specific amplification of DNA fragments of defined size (344-784bp). Advantages of the system are that it can be performed in under 4 hours including DNA extraction, results are easy to interpret and it does not require viable cells.

The HLA-B73 antigen has a most unusual structure that defines a second lineage of HLA-B alleles

The nucleotide sequence of cDNA encoding the HLA-B73 antigen was determined; it is unusually divergent, differing from other HLA-B alleles by 44-77 nucleotide substitutions. Features that distinguish the B*7301 heavy chain from other HLA-B heavy chains include multiple substitutions in the alpha 3 domain and a duplication-deletion within the transmembrane region that increases the length of B*7301 compared to other HLA-B heavy chains. The duplication-deletion is shared with subsets of B alleles from the homologous gorilla (Gogo-B) and chimpanzee (Patr-B) loci. Other unusual features of B*7301 are individually shared with certain alleles of the HLA-A, HLA-C, HLA-F, Gogo-B and Patr-B loci. The B*7301 molecules has sequence elements in common with members of the B7 crossreacting group in the alpha 1 domain and is shown to possess the ME1 epitope, which is held in common with the B7, B22, B27, B42 and B67 antigens. B*7301 has a unique cysteine at position 270 of the alpha 3 domain which appears accessible but probably does not form disulphide-bonded B*7301 dimers in cell membranes. B*7301 represents a newly discovered but ancient lineage of HLA-B alleles that appears poorly represented in the modern human population.