Sequence analysis of HLA-Bw53, a common West African allele, suggests an origin by gene conversion of HLA-B35

Like Wings of a Bird: Functional Divergence and Complementarity between HLA-A and HLA-B Molecules

Human leukocyte antigen (HLA) genes are among the most polymorphic of our genome, as a likely consequence of balancing selection related to their central role in adaptive immunity. HLA-A and HLA-B genes were recently suggested to evolve through a model of joint divergent asymmetric selection conferring all human populations, including those with severe loss of diversity, an equivalent immune potential. However, the mechanisms by which these two genes might undergo joint evolution while displaying very distinct allelic profiles in populations are still unknown. To address this issue, we carried out extensive data analyses (among which factorial correspondence analysis and linear modeling) on 2,909 common and rare HLA-A, HLA-B, and HLA-C alleles and 200,000 simulated pathogenic peptides by taking into account sequence variation, predicted peptide-binding affinity and HLA allele frequencies in 123 populations worldwide. Our results show that HLA-A and HLA-B (but not HLA-C) molecules maintain considerable functional divergence in almost all populations, which likely plays an instrumental role in their immune defense. We also provide robust evidence of functional complementarity between HLA-A and HLA-B molecules, which display asymmetric relationships in terms of amino acid diversity at both inter- and intraprotein levels and in terms of promiscuous or fastidious peptide-binding specificities. Like two wings of a flying bird, the functional complementarity of HLA-A and HLA-B is a perfect example, in our genome, of duplicated genes sharing their capacity of assuming common vital functions while being submitted to complex and sometimes distinct environmental pressures.

HLA B53 is associated with a poor outcome in black COVID-19 patients

A disproportionate incidence of death has occurred in African Americans (Blacks) in the United States due to COVID-19. The reason for this disparity is likely to be multi-factorial and may involve genetic predisposition. The association of human leukocyte antigens (HLA) with severe COVID-19 was examined in a hospitalized population (89% Black, n = 36) and compared to HLA typed non-hospitalized individuals (20% Black, n = 40) who had recovered from mild disease. For additional comparison, HLA typing data was available from kidney transplant recipients and deceased donors. Hospitalized patients were followed for 45 days after admission to our medical center with death as the primary end-point. One HLA allele, B53, appeared to be more prevalent in the hospitalized COVID-19 patients (percent of positive subjects, 30.5) compared to national data in US Black populations (percent of positive subjects, 24.5). The percent B53 positive in non-hospitalized COVID-19 patients was 2.6, significantly less than the percent positive in the hospitalized COVID-19 patients (p = 0.001, Fisher’s exact test) and less than the 8 percent positive listed in national data bases for US Caucasian populations. Significantly greater deaths (73 percent) were observed in HLA B53 positive hospitalized COVID-19 patients compared to hospitalized COVID-19 patients who were B53 negative (40 percent). Multi-variate analysis indicated that HLA B53 positive Black hospitalized COVID-19 patients were at a 7.4 fold greater risk of death than Black COVID-19 patients who were B53 negative. Consideration for accelerated vaccination and treatment should be given to HLA B53 positive Black COVID19 patients.

Insights into malaria susceptibility using genome-wide data on 17,000 individuals from Africa, Asia and Oceania

The human genetic factors that affect resistance to infectious disease are poorly understood. Here we report a genome-wide association study in 17,000 severe malaria cases and population controls from 11 countries, informed by sequencing of family trios and by direct typing of candidate loci in an additional 15,000 samples. We identify five replicable associations with genome-wide levels of evidence including a newly implicated variant on chromosome 6. Jointly, these variants account for around one-tenth of the heritability of severe malaria, which we estimate as ~23% using genome-wide genotypes. We interrogate available functional data and discover an erythroid-specific transcription start site underlying the known association in ATP2B4, but are unable to identify a likely causal mechanism at the chromosome 6 locus.  Previously reported HLA associations do not replicate in these samples. This large dataset will provide a foundation for further research on thegenetic determinants of malaria resistance in diverse populations.

Co-evolution of human leukocyte antigen (HLA) class I ligands with killer-cell immunoglobulin-like receptors (KIR) in a genetically diverse population of sub-Saharan Africans

Interactions between HLA class I molecules and killer-cell immunoglobulin-like receptors (KIR) control natural killer cell (NK) functions in immunity and reproduction. Encoded by genes on different chromosomes, these polymorphic ligands and receptors correlate highly with disease resistance and susceptibility. Although studied at low-resolution in many populations, high-resolution analysis of combinatorial diversity of HLA class I and KIR is limited to Asian and Amerindian populations with low genetic diversity. At the other end of the spectrum is the West African population investigated here: we studied 235 individuals, including 104 mother-child pairs, from the Ga-Adangbe of Ghana. This population has a rich diversity of 175 KIR variants forming 208 KIR haplotypes, and 81 HLA-A, -B and -C variants forming 190 HLA class I haplotypes. Each individual we studied has a unique compound genotype of HLA class I and KIR, forming 1-14 functional ligand-receptor interactions. Maintaining this exceptionally high polymorphism is balancing selection. The centromeric region of the KIR locus, encoding HLA-C receptors, is highly diverse whereas the telomeric region encoding Bw4-specific KIR3DL1, lacks diversity in Africans. Present in the Ga-Adangbe are high frequencies of Bw4-bearing HLA-B*53:01 and Bw4-lacking HLA-B*35:01, which otherwise are identical. Balancing selection at key residues maintains numerous HLA-B allotypes having and lacking Bw4, and also those of stronger and weaker interaction with LILRB1, a KIR-related receptor. Correspondingly, there is a balance at key residues of KIR3DL1 that modulate its level of cell-surface expression. Thus, capacity to interact with NK cells synergizes with peptide binding diversity to drive HLA-B allele frequency distribution. These features of KIR and HLA are consistent with ongoing co-evolution and selection imposed by a pathogen endemic to West Africa. Because of the prevalence of malaria in the Ga-Adangbe and previous associations of cerebral malaria with HLA-B*53:01 and KIR, Plasmodium falciparum is a candidate pathogen.

De novo assembly and genotyping of variants using colored de Bruijn graphs

Detecting genetic variants that are highly divergent from a reference sequence remains a major challenge in genome sequencing. We introduce de novo assembly algorithms using colored de Bruijn graphs for detecting and genotyping simple and complex genetic variants in an individual or population. We provide an efficient software implementation, Cortex, the first de novo assembler capable of assembling multiple eukaryotic genomes simultaneously. Four applications of Cortex are presented. First, we detect and validate both simple and complex structural variations in a high-coverage human genome. Second, we identify more than 3 Mb of sequence absent from the human reference genome, in pooled low-coverage population sequence data from the 1000 Genomes Project. Third, we show how population information from ten chimpanzees enables accurate variant calls without a reference sequence. Last, we estimate classical human leukocyte antigen (HLA) genotypes at HLA-B, the most variable gene in the human genome.

Population genetics of malaria resistance in humans

The high mortality and widespread impact of malaria have resulted in this disease being the strongest evolutionary selective force in recent human history, and genes that confer resistance to malaria provide some of the best-known case studies of strong positive selection in modern humans. I begin by reviewing JBS Haldane's initial contribution to the potential of malaria genetic resistance in humans. Further, I discuss the population genetics aspects of many of the variants, including globin, G6PD deficiency, Duffy, ovalocytosis, ABO and human leukocyte antigen variants. Many of the variants conferring resistance to malaria are 'loss-of-function' mutants and appear to be recent polymorphisms from the last 5000-10 000 years or less. I discuss estimation of selection coefficients from case-control data and make predictions about the change for S, C and G6PD-deficiency variants. In addition, I consider the predicted joint changes when the two β-globin alleles S and C are both variable in the same population and when there is a variation for α-thalassemia and S, two unlinked, but epistatic variants. As more becomes known about genes conferring genetic resistance to malaria in humans, population genetics approaches can contribute both to investigating past selection and predicting the consequences in future generations for these variants.

Associations between KIR epitope combinations expressed by HLA-B/-C haplotypes found in an HIV-1 infected study population may influence NK mediated immune responses

The Killer Ig-like receptors (KIRs) on NK cells regulate NK activity via the recognition of specific HLA class I products on the surface of target cells. To investigate the level at which these genetically polymorphic receptors and ligands influence HIV-1 disease progression, we examined the effect of KIR and HLA genotype on HIV outcome. We observed a significant association between particular combinations of KIR epitopes expressed by HLA-B/-C haplotypes and propose that the repertoire of KIR epitopes expressed by HLA-B and HLA-C alleles within the context of particular haplotypes may be an important component of the NK mediated immune response to HIV and/or other infectious pathogens.

The flexibility of the TCR allows recognition of a large set of naturally occurring epitope variants by HIV-specific cytotoxic T lymphocytes

Pathogens attempt to evade immune recognition by expressing mutated antigens. The present study shows that two mechanisms happen in vivo during the course of HIV infection to limit the escape of antigenic variants from cytotoxic T lymphocyte (CTL) recognition: recognition of several epitope variants by the same TCR and generation of several CTL populations specific for a single epitope but recognizing different variant sequences. We have studied two CTL populations directed towards the HIV-p24gag amino acids 176--184 QASQEVKNW epitope, presented by HLA-B5301. Both CTL populations were derived from a long-term asymptomatic HIV-infected child and they express different TCR. Each of the two CTL recognizes five of the 10 naturally occurring variants. These variants are distinct for both CTL and thus a total of eight variants are recognized. Thus, polyclonality of CTL specific for the same epitope but differing in variant sequences recognized may improve the control of variant viruses' replication in vivo. In addition to cross-recognition of several variant epitopes, promiscuous recognition of exogenous peptides complexed to allogeneic HLA-B molecules occurs, showing that the TCR can tolerate amino acid changes on both the peptide and the MHC molecule. This flexibility of the TCR is probably of great importance for control of viruses with high genetic variability, such as HIV.

HLA-B allele frequencies in Côte d'Ivoire defined by direct DNA sequencing: identification of HLA-B*1405, B*4410, and B*5302

Direct automated DNA sequencing was used to analyze exons 2 and 3 of HLA-B alleles present in forty-four unrelated individuals residing in the village of Adiopodoume, Côte d'Ivoire (Ivory Coast). Of the 23 HLA-B alleles observed, the most frequently detected allele was HLA-B*5301 (22.7%), which is believed to confer resistance to severe Plasmodium falciparum malaria. B*4501 (9.1%), B*1503 (8.0%), B*0705 (5.7%), B*1510 (5.7%) and B*3501 (5.7%) occurred frequently in the population. A second allele of B53 was identified; B*5302 contains a single amino acid variation at residue 171 (Y-->H). Two additional novel alleles, B* 1405 (a single amino acid variant of B*1402) and B*4410 (a five amino acid variant of B*4403) were characterized.

Characteristics of HLA class I and class II polymorphisms in Rwandan women

OBJECTIVE

To define HLA class I and class II polymorphisms in Rwandans.

METHODS

PCR-based HLA genotyping techniques were used to resolve variants of HLA-A, B, and C to their 2- or 4-digit allelic specificities, and those of DRB1 and DQB1 to their 4- or 5-digit alleles.

RESULTS

Frequencies of 14 A, 8 C, and 14 B specificities and of 13 DRB1 and 8 DQB1 alleles were >/=0.02 in a group of 280 Rwandan women. These major HLA factors produced 6 haplotypes extending across the class I and class II regions: A*01-Cw*04-B* 4501-DRB1*1503-DQB1*0602 (A1-Cw4-B12- DR15 - DQ6), A * 01 - Cw * 04 - B * 4901 -DRB1 * 1302-DQB1*0604 (A1-Cw4-B21-DR13-DQ6), A*30 - Cw*04 - B*15 - DRB1*1101 - DQB1*0301 (A19-Cw4-B15-DR11-DQ7), A*68-Cw*07-B* 4901-DRB1*1302-DQB1*0604(A28-Cw7-B21- DR13 - DQ6), A*30 - Cw*07 - B*5703 - DRB1* 1303-DQB1*0301(A19 - Cw7 - B17 - DR13 - DQ7), and A*74-Cw*07-B*4901-DRB1*1302-DQB1* 0604 (A19-Cw7-B21-DR13-DQ6), respectively. Collectively, these extended haplotypes accounted for about 19% of the total. Other apparent class I-class II haplotypes (e.g., Cw*17-B*42-DRB1*0302-DQB1*0402, Cw*06- B*58-DRB1*1102-DQB1*0301, and Cw*03- B*15-DRB1*03011-DQB1*0201) did not extend to the telomeric HLA-A locus, and other 3-locus class I haplotypes (e.g., A*68-Cw*04-B*15, A*74-Cw*04-B*15, and A*23-Cw*07-B*4901) completely or partially failed to link with any specific class II alleles.

DISCUSSION

Frequent recombinations appeared to occur between the three evolutionarily conserved HLA blocks carrying the class I and class II loci. The HLA class I profile seen in Rwandans was not directly comparable with those known in the literature, although the class II profile appeared to resemble those in several African populations. These data provide additional evidence for the extensive genetic diversity in Africans.

Diversity is demonstrated in class I HLA-A and HLA-B alleles in Cameroon, Africa: description of HLA-A*03012, *2612, *3006 and HLA-B*1403, *4016, *4703

To examine the genetic diversity in west Africa, class I HLA-A and HLA-B alleles of 92 unrelated individuals from two areas in the Cameroon, the capital Yaoundé and the village of Etoa, were identified by direct automated DNA sequencing of exons 2 and 3 of the HLA-B locus alleles and sequence-specific oligonucleotide probe (SSOP) and/or sequencing of the HLA-A locus alleles. HLA-A*2301 (18.7%), A*2902 (10.4%), B*5301 (10.9%), and B*5802 (10.9%) were the most frequently detected alleles, present in at least 10% of the population. A total of 30 HLA-A locus and 33 HLA-B locus alleles, including six novel alleles, were detected. The novel alleles were HLA-A*03012, A*2612, A*3006 and HLA-B*1403, B*4016, and B*4703. HLA-B*4703 contains a novel amino acid sequence that is a combination of the first 5 amino acids of the Bw6 epitope and the last 2 residues of the Bw4 epitope. The addition of 6 alleles to the ever-expanding number of known class I HLA alleles supports our hypothesis that extensive genetic diversity, including previously undescribed alleles, would be observed in this African population. In the Yaoundé population, the allele frequency distribution at the HLA-A locus is consistent with distributions indicative of balancing selection. Extensive HLA-A-B haplotypes were observed in this population suggesting that only a fraction of the Cameroon HLA-A-B haplotype diversity has been observed.

Sequence-Based Typing Provides a New Look at HLA-C Diversity

Although extensive HLA-A and HLA-B polymorphism is evident, the true diversity of HLA-C has remained hidden due to poor resolution of HLA-C Ags. To better understand the polymorphic nature of HLA-C molecules, 1823 samples from the National Marrow Donor Program research repository in North America have been typed by DNA sequencing and interpreted in terms of HLA-C diversification. Results show that HLA-Cw*0701 was the most common allele with a frequency of 16%, whereas 28% of the alleles typed as Cw12-18 (serologic blanks). The frequency of homozygotes was 9.8% as compared with previous studies of 18% for sequence-specific primers and 50% for serology. Most startling was the frequency at which new alleles were detected; 19 new HLA-C alleles were detected, representing a rate of ∼1 in 100 samples typed. These new HLA-C alleles result from 29 nucleotide substitutions of which 4 are silent, such that coding substitutions concentrated about the Ag-binding groove predominate. Polymorphism at the HLA-C locus therefore resembles that at the HLA-A and HLA-B loci more than previously believed, indicating that antigenic stress is driving HLA-C evolution. However, sequence conservation in the α-helix of the first domain and a clustering of unique amino acids around the B pocket indicate that HLA-C alleles respond to antigenic pressures differently than HLA-A and HLA-B. Finally, because the samples characterized were predominantly from Caucasians, we hypothesize that HLA-C polymorphism will equal or exceed that of the HLA-A and -B loci as DNA sequence-based typing is extended to include more non-Caucasian individuals.

Molecular analysis of HLA-B in the Malaysian aborigines

We have examined 56 unrelated individuals from Malaysian aborigines for their DNA polymorphism of the HLA-B gene by sequence specific oligonucleotide probe (SSO) method. Using the SSO hybridization, we found that one specific DNA allele with a B*1513 like pattern of epitope combination (ECB1513) was dominant among the Melayu Asli (Af = 41.9%) and the Senoi (Af = 24%). To determine the nucleotide sequences of ECB1513, a DNA fragment spanning from the beginning of exon 1 to the middle of exon 4 of the HLA-B gene was amplified by polymerase chain reaction (PCR) from two ECB1513 positive individuals, and the PCR products were cloned and sequenced. This sequencing analysis confirmed that ECB1513 was identical to HLA-B*1513 in exon 1, 2, 3, and 4. Amino acid sequence of this major allele, HLA-B*1513, in the aborigines especially around the peptide binding groove (B and F pockets), was compared with that of African B*5301 that had been suggested to confer resistance to malaria infection in Africa. The amino acid residues composing of the F pocket were completely identical in B*1513 and B*5301. These observations suggest that a common environmental factor, the malaria infection, might have independently enhanced the selection of functional change in the polymorphic portion of HLA-B gene in Africa and in South-East Asia.

An altered position of the alpha 2 helix of MHC class I is revealed by the crystal structure of HLA-B*3501

The crystal structure of the human major histocompatibility complex class I B allele HLA B*3501 complexed with the 8-mer peptide epitope HIV1 Nef 75-82 (VPLRPMTY) has been determined at 2.0 angstrom resolution. Comparison with the crystal structure of the closely related allele HLA B*5301 reveals the structural basis for the tyrosine specificity of the B*3501 F pocket. The structure also reveals a novel conformation of the 8-mer peptide within the binding groove. The positions of the peptide N and C termini are nonstandard, but the classic pattern of hydrogen bonding to nonpolymorphic MHC class I residues is maintained, at the N terminus by addition of a water molecule, and at the C terminus by a substantial shift in the alpha 2 helix.

Bound water structure and polymorphic amino acids act together to allow the binding of different peptides to MHC class I HLA-B53

The structure of the human MHC class I molecule HLA-B53 complexed to two nonameric peptide epitopes (from the malaria parasite P. falciparum and the HIV2 gag protein) has been determined by X-ray crystallography at 2.3 angstrom resolution. The structures reveal the architecture of a Pro-specific B pocket common to many HLA-B alleles. Relative to other alleles, the B53 peptide-binding groove is widened by a significant (up to 1.25 angstrom) shift in the position of the alpha 1 helix. Within this groove, bound water molecules, acting in concert with the side chains of polymorphic residues, provide the functional malleability of the MHC, which enables the high affinity/low specificity binding of multiple peptide epitopes.

Variation in the TNF-alpha promoter region associated with susceptibility to cerebral malaria

Tumour-necrosis factor-alpha (TNF-alpha) is believed to have an important role in the pathogenesis of severe infectious disease and fatal cerebral malaria is associated with high circulating levels of this cytokine. In a large case-control study in Gambian children we find that homozygotes for the TNF2 allele, a variant of the TNF-alpha gene promoter region, have a relative risk of 7 for death or severe neurological sequelae due to cerebral malaria. Although the TNF2 allele is in linkage disequilibrium with several neighbouring HLA alleles, we show that this disease association is independent of HLA class I and class II variation. These data suggest that regulatory polymorphisms of cytokine genes can affect the outcome of severe infection. The maintenance of the TNF2 allele at a gene frequency of 0.16 in The Gambia implies that the increased risk of cerebral malaria in homozygotes is counterbalanced by some biological advantage.

HLA-B15: a widespread and diverse family of HLA-B alleles

HLA-B15 embraces a multiplicity of antigenic specificities which vary in their distribution amongst human populations. To correlate B15 molecular structure with the serological picture we have sequenced alleles encoding the various subspecificities of the B15 antigen: B62, B63, B75, B76 and B77, and a number of "variants" of these antigens including the 8w66 split of B63. HLA-B63 (B*1517) and 8w66 (B*1516) heavy chains have sequence identity to B17 in the alpha 1 helix correlating with the antigenic crossreactivity of these molecules. HLA-B77(B*1513) and B75 (B*1502) heavy chains differ solely in segments determining the Bw4 and Bw6 public epitopes, consistent with the serological description of the B77 and B75 antigens. One allele encoding the B76 antigen (B*1512) appears to be the product of gene conversion between the HLA-A and -B loci and differs from B*1501 in codons 166 and 167. In contrast, a second allele encoding the B76 antigen (B*1514) differs from B*1501 by an unrelated substitution in codon 167 which confers similarily with B45, an antigen crossreactive with B76. A third allele encoding B76, B*1519, differs from B*1512 by a unique point substitution in exon 4. Three alleles encoding variant B15 and B62 antigens (B*1508, B*1511 and B*1515) differ from B*1501 by localized clusters of substitutions that probably result from interallelic conversion. The B15 sequences described in this paper, in combination with those previously determined, define a family of 22 alleles, including those encoding the B46 and B70 antigens. Within this family the patterns of allelic substitution are analogous to those of other HLA-A and -B families, in that pairwise differences almost always involve functional positions of the antigen recognition site and recombination is the major agent of diversification.

A method to quantify binding of unlabeled peptides to class I MHC molecules and detect their allele specificity

A general method has been developed for measuring the stabilization of class I MHC molecules in extracts of the mutant cell lines .174/T2 and RMA-S. 35S-Met-labeled class I molecules which have been stabilized by peptides in vitro are immunoprecipitated with conformation dependent monoclonal antibodies and electrophoresed on polyacrylamide gels. The heavy and light chains are excised from the dried gel and quantified on a flat bed scintillation counter. The stabilizing effect of peptides on class I molecules in vitro correlates well with peptide binding measured by direct methods and can be therefore used to assess peptide binding affinity. We show that a peptide from HIV-1 gag (which has a high affinity for Db) is a CTL epitope restricted through Db, and also use the assay to analyse the effects of amino acid substitution on peptide affinity. In addition, the effect of a given peptide on a class I molecule within a mixture of human class I molecules can be distinguished by immunoprecipitation with the monomorphic antibody W6/32 and separation by 1-D isoelectric focussing. The technique therefore requires neither labeled peptide ligands nor allele-specific antibodies. It can be used to identify the peptide ligand of any human class I molecule, and gives a measure of peptide binding affinity. The technique should be of value in identifying epitopes recognized by CTL since we have found that these tend to bind with the highest affinities.

Different rates of HLA class I molecule assembly which are determined by amino acid sequence in the alpha 2 domain

Assembly of HLA class I molecules was studied using pulse-chase labeling of B-lymphoblastoid cell lines with 35S-methionine, immunoprecipitation with antibodies detecting free or beta 2-microglobulin-associated heavy chain and isoelectric focusing. Marked differences between the products of different class I alleles were noted. HLA-B51 assembled very inefficiently, with considerable free heavy chain still detected in an unsialated form after a four hour chase. The closely related molecule HLA-B35 was in contrast rapidly assembled, all newly synthesized heavy chain being detected in a beta 2m-associated sialated form within 30 minutes. Analysis of naturally occurring variants related to HLA-B35 and HLA-B51 localized the region determining assembly efficiency to the alpha 2 domain, in which these molecules differ at eight amino acid residues. The effect was not due to a linked dominant gene, as both patterns of assembly were observed in a single cell line.

Molecular analysis of the association of HLA-B53 and resistance to severe malaria

The protective association between the human leukocyte antigen HLA-B53 and severe malaria was investigated by sequencing of peptides eluted from this molecule followed by screening of candidate epitopes from pre-erythrocytic-stage antigens of Plasmodium falciparum in biochemical and cellular assays. Among malaria-immune Africans, HLA-B53-restricted cytotoxic T lymphocytes recognized a conserved nonamer peptide from liver-stage-specific antigen-1 (LSA-1), but no HLA-B53-restricted epitopes were identified in other antigens. These findings indicate a possible molecular basis for this HLA-disease association and support the candidacy of liver-stage-specific antigen-1 as a malaria vaccine component.

Common west African HLA antigens are associated with protection from severe malaria

A large case-control study of malaria in West African children shows that a human leucocyte class I antigen (HLA-Bw53) and an HLA class II haplotype (DRB1*1302-DQB1*0501), common in West Africans but rare in other racial groups, are independently associated with protection from severe malaria. In this population they account for as great a reduction in disease incidence as the sickle-cell haemoglobin variant. These data support the hypothesis that the extraordinary polymorphism of major histocompatibility complex genes has evolved primarily through natural selection by infectious pathogens.

HLA class I typing by PCR: HLA-B27 and an African B27 subtype

We describe a rapid method of HLA class I typing using the polymerase chain reaction and oligonucleotide hybridisation that eliminates the requirements for viable lymphocytes and allows subtypes to be defined. We have used this to demonstrate that the predominant subtype of HLA-B27 in the Gambia, West Africa, is HLA-B*2703, which is very rare or absent in other racial groups. This subtype differs from the common Caucasian HLA-B27 subtypes in its recognition by cytotoxic T cells. We propose that HLA*B-2703, unlike other HLA-B27 subtypes, may not be associated with ankylosing spondylitis, thus accounting in part for the rarity of this condition in black populations.