A human immunoglobulinIGHG3 allele (Gmb0, b1, c3, c5, u) with anIGHG4 converted region and three hinge exons

The five humanIGHG genes consist of three constant domain exons plus one of or four hinge exon(s), the quadruplicated hinge region being characteristic of theIGHG3 gene. Besides this structural difference, theIGHG genes are polymorphic, as demonstrated by the restriction fragment length polymorphism and, at the protein level, by the Gm allotypic antigenic determinants. In this paper, we report the sequence of theG3m(b0, b1, c3, c5, u) IGHG3 allele, typical of the Black African populations and of populations with Negroid admixture, found in a homozygous Tunisian designated as LAT. We demonstrate that thisG3 allele contains only three hinge exons instead of four (the probable result of an unequal crossing over) and thatIGHG3 genes with triplicated hinge exons (and therefore encoding shorter γ 3 chains) are present in healthy individuals from different populations. Moreover, we show that the LAT G3m (b0, b1, c3, c5, u) coding sequence results from the conversion, in the CH3 exon, of theG3m (b0, b1, b3, b4, b5, u, v) allele, the most frequentIGHG3 gene in the Negroid populations, by the homologous region of aIGHG4 gene. The structural features of theLAT IGHG3 allele, which are the lack of one hinge exon and its conversion by theIGHG4 gene, demonstrate that both crossing-over and gene conversion events occur in the evolution of the humanIGHG genes. The nucleotide sequence data reported in this paper have been submitted to the GenBank nucleotide sequence database and have been assigned the accession number X16110.     

Duplications and deletions of the human IGHC locus: evolutionary implications

 A limited number of deletions and duplications within the human immunoglobulin heavy chain constant locus (IGHC) has previously been reported. We studied the IGHC locus in about 500 individuals representing three major races of human, Negroid (Gambian), Mongoloid (Japanese and Chinese), and Caucasoid (Iranian and Swedish). The haplotype frequency of duplications is highest in the Mongoloid population (22%), followed by the Caucasian (10%) and Negroid (5%) populations. The corresponding frequency of deletions are 2, 1.5, and 3.5%, respectively. New types of multiple duplications were found in this study on different genetic (H haplotype and racial) backgrounds. The most common duplication, found in all populations studied, encompasses the IGHA1-IGHE genes. The only deletion common to all racial groups is an isolated deletion of the IGHG4 gene. Our data are consistent with the hypothesis that the Caucasoid-Mongoloid group diverged from the hominoid ancestor after development of the Negroid populations, with subsequent evolution within the respective groups thereafter. Received: 14 June 1996/Revised: 2 August 1996     

Familial clustering of IGHC deletions and duplications: functional and molecular analysis

The human immunoglobulin heavy chain constant region locus (IGHC) comprises nine genes and two pseudogenes clustered in a 350 kilobase (kb) region on chromosome 14q32. Several IGHC haplotypes with single or multiple gene deletions and duplications have been characterized. The most likely mechanism accounting for these unusual haplotypes is the unequal crossing-over between homologous regions within the locus. Here we report the analysis of an unusual case of familial clustering of deletions/duplications. In the two branches of the BON family, three duplicated and two deleted haplotypes, all probably independent in origin, have been characterized. The structure of the haplotypes, one of which is described here for the first time, supports the hypothesis of homologous unequal crossing-over as the origin of recombinant haplotypes. The analysis of serological markers in a subject carrying one deleted and one duplicated haplotype allowed us the first direct inferences concerning the functions of the duplicated IGHC haplotypes.     

CYP21/C4 gene organisation in Italian 21-hydroxylase deficiency families

Summary A total of 33 Italian 21-hydroxylase (21-OH) deficiency families were investigated using a combination of short and long range restriction mapping of the CYP21/C4 gene cluster. The analyses revealed that large-scale length polymorphism in this gene cluster strictly conformed to a compound variable number of tandem repeats (VNTR) plus insertion system with between one and four CYP21 + C4 units and seven BssHII restriction fragment length polymorphisms (RFLPs) (75kb, 80kb, 105kb, 110kb, 135kb, 140kb and 180kb). A total of 9/66 disease haplotypes, but only 1/61 nondisease haplotypes, showed evidence of gene addition by exhibiting three or more CYP21 + C4 repeat units. Of these, two were identified in one 21-OH deficiency patient who has a total of eight CYP21 + C4 units, being homozygous for the HLA haplotype DR2 DQ2 B5 A28. This haplotype carries four CYP21 + C4 units, three of which contain CYP21A-like genes and one of which contains a CYP21B-like gene that presumably carries a pathological point mutation. Of the other gene addition haplotypes associated with 21-OH deficiency, four show three CYP21 + C4 units flanked by HLA-DR1 and HLA-B14 markers. Although such haplotypes have commonly been associated with non-classical 21-OH deficiency, three examples in the present study are unexpectedly found in two salt-wasting patients, who are respectively homozygous or heterozygous for this haplotype. Only 7/66 disease haplotypes showed evidence of a CYP21B gene deletion.     

Evolutionary dynamics of the immunoglobulin heavy chain variable region genes in vertebrates

Immunoglobulin heavy chains are polypeptides encoded by four genes: variable (IGHV), joining (IGHJ), diversity (IGHD), and constant (IGHC) region genes. The number of IGHV genes varies from species to species. To understand the evolution of the IGHV multigene family, we identified and analyzed the IGHV sequences from 16 vertebrate species. The results show that the numbers of functional and nonfunctional IGHV genes among different species are positively correlated. The number of IGHV genes is relatively stable in teleosts, but the intragenomic sequence variation is generally higher in teleosts than in tetrapods. The IGHV genes in tetrapods can be classified into three phylogenetic clans (I, II, and III). The clan III and/or II genes are relatively abundant, whereas clan I genes exist in small numbers or are absent in most species. The genomic organization of clan I, II, and III IGHV genes varies considerably among species, but the entire IGHV locus seems to be conserved in the subtelomeric or near-centromeric region of chromosome. The presence or absence of specific IGHV clan members and the lineage-specific expansion and contraction of IGHV genes indicate that the IGHV locus continues to evolve in a species-specific manner. Our results suggest that the evolution of IGHV multigene family is more complex than previously thought and that several factors may act synergistically for the development of antibody repertoire. Electronic supplementary materials The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

The “Sardinian”HLA-A30,B18,DR3,DQw2 haplotype constantly lacks the21-OHA andC4B genes. Is it an ancestral haplotype without duplication?

TheC4 and21-OH loci of the class III HLA have been studied by specific DNA probes and the restriction enzymeTaq I in 24 unrelated Sardinian individuals selected from completely HLA-typed families. All 24 individuals had theHLA extended haplotypeA30,Cw5,B18, BfF1,DR3,DRw52,DQw2, named “Sardinian” in the present paper because of its frquency of 15% in the Sardinian population. Eighteen of these were homozygous for the entire haplotype, and six were heterozygous at theA locus and blank (or homozygous) at all the other loci. In all completely homozygous cells and in four heterozygous cells at theA locus, the restriction fragments of the21-OHA (3.2 kb) andC4B (5.8 kb or 5.4 kb) genes were absent, and the fragments of theC4A (7.0 kb) and21-OHB (3.7 kb) genes were present. It is suggested that the “Sardinian” haplotype is an ancestral haplotype without duplication of theC4 and21-OH genes, practically always identical in its structure, also in unrelated individuals. The diversity of this haplotype in the class III region (about 30 kb less) may be at least partially responsible for its misalignment with most haplotypes, which have duplicatedC4 and21-OH genes, and therefore also for its decreased probability to recombine. This can help explain its high stability and frequency in the Sardinian population. The same conclusion can be suggested for the Caucasian extended haplotypeA1,B8,DR3 that always seems to lack theC4A and21-OHA genes.     

A linkage map of the mouse immunoglobulin lambda light chain locus

The mouse immunoglobulin lambda light chain locus has been linked using field inversion gel electrophoresis. The lambda light chain locus classically contains two V and four J-C gene segments in inbred mouse strains, and was physically mapped in the BALB/c cell line Wehi-3 which contains unrearranged lambda light chain gene segments. The locus is relatively small and spans 300 kb, as defined by a variety of single and double digests using methylation-sensitive restriction enzymes. The order of the lambda gene segments is V2-J2C2J4C4-V1-J3C3J1C1, as was originally proposed. No evidence for nonmethylated CpG rich areas (HTF islands) within the region was found. Fine mapping using the 1, 3 rearranged cell line J558 mapped the gap between the V and J-C gene segments in the lambda 1 gene cluster (VI-J3C3JIC1) to approximately 70 kb. The similar distance (60–100 kb) found in the lambda 2 gene cluster (V2-J2C2J4C4) is further evidence that duplication of an ancestral locus occurred.     

Electron microscopical analysis of Drosophila polytene chromosomes

An electron microscopic (EM) analysis was performed on regions of Drosophila melanogaster polytene chromosomes that contain inserted DNA segments of 19 and 8 kb. These segments had been inserted by P-elementmediated transformation. The 19 kb segment includes both the Drosophila hsp70 gene fused to the Escherichia coli -galactosidase gene and the rosy gene (Lis et al. 1983). This insert generates a new moderate-size band at the 9D4-9E1-2 region in polytene chromosomes. Upon heat shock, a puff originates from a portion of the new band. The 8 kb segment includes the Sgs7 and Sgs3 genes (Richards et al. 1983). This insert generates very diffuse thin bands that decondense at the stage of activation of the Sgs genes to produce wide interbands or small puffs. In all of the above cases, the insertion appears to occur at interband regions, and the genetically complex DNA segments that are inserted generate only a single detectable band.     

Polymorphism detection and sequence analysis of human T-cell receptor Vα-chain-encoding gene segments

The T-cell receptor (Tcr) provides specificity for antigen recognition by its variable domain, primarily consisting of two germline encoded variable (V) region gene segments. Thus it has been suggested that inherited polymorphisms in the TCRV gene segments could contribute to differential immune responsiveness (e.g., autoimmunity) in human populations. In the present study, we have sought potentially functional polymorphisms in the germline TCRAV gene segments. Using denaturing gradient gel electrophoresis on polymerase chain reaction (PCR)-amplified products from the pooled DNA of many individuals, we identified polymorphisms in the TCRAV2S1, AV4S1, AV7S1, and AV8S1 gene segments. A complete DNA sequence analysis of these PCR products identified polymorphisms that affected amino acids in the predicted antigen-binding regions of the Tcr chain, as well as polymorphisms in the introns. Genotype analysis of all nine DNA point mutations showed a 5%–50% range (averaging 35%) of minor allele frequencies, often resulting in individuals homozygous for the alternate allele forms. All possible haplotype combinations of the amino acid-affecting polymorphisms were found, indicating that in human populations there are a large number of different germline haplotypes encoding V gene segment alleles. These TCRAV coding region polymorphisms provide the rationale for, and allow the direct testing of, hypotheses concerning inherited polymorphisms within the T-cell receptor genes that may contribute to autoimmune susceptibility.The nucleotide sequence data reported in this paper have been submitted to the Genbank nucleotide sequence database and have been assigned the accession numbers L11159–L11162.     

Sharing of MHC haplotypes among apparently unrelated patients with congenital adrenal hyperplasia due to 21-hydroxylase deficiency

From the study of HLA, complement, and glyoxalase I alleles in 82 Venezuelan individuals belonging to 19 families of mixed ethnic origin having 20 affected newborns with salt-wasting congenital adrenal hyperplasia due to 21-hydroxylase (21-OH) deficiency, a total of 38 disease haplotypes and 53 nondisease haplotypes were found. Of the pathological haplotypes 47 % were found to share the HLA-B39 or -Bw62 specificities, 55 % of them in combination with the BFS, C2C, C4A4, C4B2 (SC42) complotype. The frequencies of HLA-B39 and -Bw62 among the affected haplotypes were 29 and 18% as compared with 6 and 0 % among the nondisease haplotypes of the same families. Statistical associations (P < 0.01) with salt-wasting adrenal hyperplasia were found with the SC42 complotype and with the combination SC42, HLA-B39. These results are markedly different from those reported in the literature which show an association at the population level among many Caucasoid samples of HLA-Bw47 and the extended haplotype (HLA-Bw47, DR7, FC91, 0) with the salt-wasting form of the disease. Furthermore, four of the unrelated patients reported here were homozygous for all the major histocompatibility complex loci tested, while three others were homozygous for at least two HLA loci. Analysis of the geographical origin of the grandparents indicated clustering of the deficiency carrier HLA haplotypes. This observation, together with the fact that there is an excess of homozygotes among the patients in Venezuela, strongly suggests that salt-wasting 21-OH deficiency congenital adrenal hyperplasia is mostly the result of a founder effect of relatively few independent mutations and, thus, of identity by descent of a few abnormal alleles at the 21-OHB locus in most cases. The mutation marked by HLA-Bw47 was not observed in this population.     

Order of class III genes relative to HLA genes determined by the haplotype method

The B18 C4A3 C4BQ0 BfF1 DR3 haplotype was found to be ideal for determining the order of C4 and Bf relative to HLA-B and DR by the haplotype method. All the copies of this haplotype are assumed to be derived from a single ancestral haplotype. Sixteen of the twenty-six BfFl-containing haplotypes carried all of the alleles from this ancestral haplotype. Most of the other BfFl-containing haplotypes could be derived from the ancestral haplotype by a single crossover event for one of the two possible gene orders. This suggests that B18 C4A3 C4BQ0 BfFl DR3 is the sole source of the BfFl allele. The uncommon C4 type on B18 C4A3 C4BQ0 BfFl DR3 facilitates recognition of the BfFl-containing products of recombination between Bf and C4. One such recombinant haplotype was found which shows that the orientation of the class III genes is as follows: C4 is closest to HLA-B and Bf is closest to HLA-DR. This gene order is supported by all the earlier unequivocal results obtained using the haplotype method (Olaisen et al. 1983, Marshall et al. 1984a). Combining these results with the information on class III genes obtained from overlapping cosmid clones (Carroll et al. 1984) and earlier mapping studies (Robson and Lamm 1984) shows that HLA-B is telomeric to 21B. C4B, 21A, C4A, Bf and C2 then follow 21B in that order covering 120 kb. HLA-DR is located further toward the centromere.     

Heterogeneity of human C4 gene size

In this article we present a study showing that the human C4 genes differ in length because of the presence or absence of a 6.5 kb intron near the 5 end of the gene. DNA from individuals of known HLA, factor B, and C4 haplotypes was analyzed for restriction fragment length polymorphism (RFLP) by Southern blot analysis with C4-specific cDNA probes. The RFLP patterns obtained showed that the C4 genes are either 22.5 kb or 16 kb in length. They are referred to as long and short C4 genes, respectively. A population study was carried out to examine the distribution of the gene size according to C4 allotypes and haplotypes. Long C4 genes included all C4A genes studied and also some C4B allotypes, e. g., B1 on most C4 A3B1 haplotypes. Similarly, C4B null genes were found to be of the long form. Other C4B allotypes tested were found to be coded for by short C4 genes, including B2, B1 in C4 A6B1 and C4 AQOB1 (with a single C4B gene haplotype).Abbreviations used in this paper C4 fourth component of complement - C2 second component of complement - BF factor B - MHC major histocompatibility complex - RFLP restriction fragment length polymorphism - EDTA ethylenediaminetetraacetic acid - SDS lauryl sulfate, sodium salt     

Physical map of the 3'' region of the human immunoglobulin heavy chain locus: clustering of autoantibody-related variable segments in one haplotype.

We have constructed the physical map of the 3' region of the human immunoglobulin heavy chain variable region (VH) genes. DNA segments extending to 200 kb upstream of the JH segment were isolated in two YAC clones. Five VH segments were identified in this region in the 5' to 3' order, V(II-5), V(IV-4), V(I-3), V(I-2), and V(VI-1) segments which were all structurally normal and orientated in the same direction as the JH segments. From DNA of a different cell line we have isolated a cosmid contig containing the same DNA region which has extraordinary polymorphism. The YAC and cosmid DNAs were called haplotypes A and B, respectively. Haplotype B contained an additional VH-I segment (V(I-4.1b)) between the V(II-5) and V(IV-4) segments. V(I-4.1b) segment is almost identical to a previously published VH sequence encoding a rheumatoid factor. Another VH segment in the B haplotype (V(I-3b)) corresponding to the V(I-3) segment also showed 99.7% nucleotide sequence homology with an anti-DNA autoantibody VH sequence. However, none of the VH sequences in haplotype A showed such strong homology with autoantibody VH sequences. The results suggest that VH haplotypes may have linkage with autoantibody production.     

Linkage between C2 deficiency and theHLA-A10,B18,Dw2/BfS haplotype in a French family

A linkage between C2 deficiency and the HLA-A10,B18/BfS antigens has been found in a French family from the Strasbourg area. The propositus, suffering from a chronic glomerulonephritis, is homozygous forHLA-A10,B18/BfS and totally C2-deficient. The parents and the brother are heterozygous for C2 deficiency and share theHLA-A10,B18/BfS haplotype. MLC tests and HLA-D typing revealed that the homozygous C2-deficient patient is also homozygous at theHLA-D locus for the w2 specificity. Evidence was obtained for a heterogeneity of the HLA-Dw2 specificity. This observation confirms the remarkable association between C2 deficiency and theHLA-A10,B18,Dw2 haplotype.     

Strong associations between RFLP and protein polymorphisms for CD46

Human CD46 (membrane cofactor protein) is a cell surface glycoprotein with cofactor activity for the factor I mediated cleavage of components C3b and C4b. Using a CD46 cDNA clone, three restriction enzymes give simple two allele restriction fragment length polymorphisms (RFLPs) in samples of over 300 Caucasians. For Pvu II, P1 with a 16.5 kilobase (kb) fragment and P2 with 14.8 kb + 1.9 kb fragments have frequencies of .40 and .60. For Hin dIII, H1 with a 4.3 kb fragment and H2 with a 2.3 kb fragment have similar frequencies. For Bgl. II, B1 with a 10 kb fragment and B2 with 8.3 kb + 1.8 kb fragments have frequencies of 0.08 and 0.92. There is strong linkage disequilibrium between these polymorphic sites. Designating haplotypes by Hin dIII, Pvu II, Bgl II alleles, there are two common haplotypes P2, H2, B2 and P1, H1, B2, expected at frequencies of .6 and .32, one less common haplotype P1, H1, B1 expected at a frequency .08. The two major protein isoforms of CD46, as detected on peripheral blood lymphocytes by western blot, of M _r 66 000 () and 56 000 () are determined by differential splicing in production of the mRNA. A strong association between protein isoform and RFLP haplotypes in 30 unrelated subjects suggests that the splicing preference site is in linkage disequilibrium with the RFLPs. The results are consistent with haplotypes P2, H2, B2 and P1, H1, B1 producing predominantly ; P1, H1, B2 producing predominantly in about 72% of cases and in 28% of cases. Address correspondence and offprint requests to: A. Wilton, at the present address.     

Search for MHC-associated genes in human: five new genes centromeric to HLA-DP with yet unknown functions

Taking advantage of five mouse genomic or cDNA probes [KE5(probe 14), KE4 (probe11), KE3 (probe7), KE2 (probe5), and SET] mapped on the H-2K region in mouse, we have identified and localized homologues of these five genes in the human major histocompatibility complex region (HKE5, HKE4, HKE3, HKE2, and HSET, respectively). Cosmid cloning and pulsed field gel electrophoresis analyses indicated that a human homologous gene, HKE5, is located 10 kilobases (kb) centromeric of the 2(XI) collagen (COL11A2) gene followed by HKE4. HKE3, closely linked to HKE2, is located 170 kb centromeric of HKE4. Furthermore, HSET is located 50 kb centromeric of HKE2. This gene organization outside the DP subregion is completely identical to that of the mouse H-2K region centromeric of I-Pb 3, a mouse homologue of the DPB gene, except the lack of genes corresponding to the H-2K and -K2 genes in human.     

Beta‐globin gene evolution in the ruminants: evidence for an ancient origin of sheep haplotype B

Domestic sheep (Ovis aries) can be divided into two groups with significantly different responses to hypoxic environments, determined by two allelic beta‐globin haplotypes. Haplotype A is very similar to the goat beta‐globin locus, whereas haplotype B has a deletion spanning four globin genes, including beta‐C globin, which encodes a globin with high oxygen affinity. We surveyed the beta‐globin locus using resequencing data from 70 domestic sheep from 42 worldwide breeds and three Ovis canadensis and two Ovis dalli individuals. Haplotype B has an allele frequency of 71.4% in O. aries and was homozygous (BB) in all five wild sheep. This shared ancestry indicates haplotype B is at least 2–3 million years old. Approximately 40 kb of the sequence flanking the ~37‐kb haplotype B deletion had unexpectedly low identity between haplotypes A and B. Phylogenetic analysis showed that the divergent region of sheep haplotype B is remarkably distinct from the beta‐globin loci in goat and cattle but still groups with the Ruminantia. We hypothesize that this divergent ~40‐kb region in haplotype B may be from an unknown ancestral ruminant and was maintained in the lineage to O. aries, but not other Bovidae, evolving independently of haplotype A. Alternatively, the ~40‐kb sequence in haplotype B was more recently acquired by an ancestor of sheep from an unknown non‐Bovidae ruminant, replacing part of haplotype A. Haplotype B has a lower nucleotide diversity than does haplotype A, suggesting a recent bottleneck, whereas the higher frequency of haplotype B suggests a subsequent spread through the global population of O. aries.     

Recombinational hotspot specific to female meiosis in the mouse major histocompatibility complex

Thewm7 haplotype of the major histocompatibility complex (MHC), derived from the Japanese wild mouseMus musculus molossinus, enhances recombination specific to female meiosis in theK/A interval of the MHC. We have mapped crossover points of fifteen independent recombinants from genetic crosses of thewm7 and laboratory haplotypes. Most of them were confined to a short segment of approximately 1 kilobase (kb) of DNA between theA 3 andA 2 genes, indicating the presence of a female-specific recombinational hotspot. Its location overlaps with a sex-independent hotspot previously identified in theMus musculus castaneus CAS3 haplotype. We have cloned and sequenced DNA fragments surrounding the hotspot from thewm7 haplotype and the corresponding regions from the hotspot-negative B10.A and C57BL/10 strains. There is no significant difference between the sequences of these three strains, or between these and the published sequences of the CAS3 and C57BL/6 strains. However, a comparison of this A3/A2 hotspot with a previously characterized hotspot in theE gene revealed that they have a very similar molecular organization. Each hotspot consists of two elements, the consensus sequence of the mouse middle repetitive MT family and the tetrameric repeated sequences, which are separated by 1 kb of DNA.The nucleotide sequence data reported in this paper have been submitted to the DNA Data Bank of Japan nucleotide sequence database and have been assigned the accession numbers d90007-9. Offprint requests to: T. Shiroishi.     

Deletion mapping of Ig VH gene segments expressed in human CD5 B cell lines. JH proximity is not the sole determinant of the restricted fetal VH gene repertoire.

VH gene segments expressed in a panel of monoclonal human CD5 B cell lines have been positioned on the IgH locus by deletion mapping. The analysis yielded a relative order of VH fragments of the VH2, VH4, VH5, and VH6 gene families that was consistent with, and provided a further refinement of existing maps of the human IgH locus. We demonstrate that four of six VH gene segments expressed in the CD5 B cell lines map > 500 kb from the cluster of JH segments. Two of the gene segments, positioned at approximately 850 kb (58p2) and approximately 500 kb (1-9III) from the JH segments, respectively, belong to the previously identified small cohort of second trimester fetal VH gene segments. The data show that JH proximity is not the sole determinant of restricted VH gene utilization in early human ontogeny.     

The evolutionary origin of the HLA-DR3 haplotype

The human HLA-DR3 haplotype consists of two functional genes (DRB1*03 and DRB3*01) and one pseudogene (DRB2), arranged in the order DRB1... DRB2... DRB3 on the chromosome. To shed light on the origin of the haplotype, we sequenced 1480 nucleotides of the HLA-DRB2 gene and aong stretches of two other genes, Gogo-DRB2 from a gorilla, Sylvia and Patr-DRB2 from a chimpanzee, Hugo. All three sequences (HLA-DRB2, Gogo-DRB2, Patr-DRB2) are pseudogenes. The HLA-DRB2 and Gogo-DRB2 pseudogenes lack exon 2 and contain a twenty-nucleotide deletion in exon 3, which destroys the correct translational reading frame and obliterates the highly conserved cysteine residue at position 173. The Patr-DRB2 pseudogene lacks exons 1 and 2; it does not contain the twenty-nucleotide deletion, but does contain a characteristic duplication of that part of exon 6 which codes for the last four amino acid residues of the cytoplasmic region. When the nucleotide sequences of these three genes are compared to those of all other known DRB genes, the HLA-DRB2 is seen as most closely related to Gogo-DRB2, indicating orthologous relationship between the two sequences. The Patr-DRB2 gene is more distantly related to these two DRB2 genes and whether it is orthologous to them is uncertain. The three genes are in turn most closely related to HLA-DRBVI (the pseudogene of the DR2 haplotype) and Patr-DRB6 (another pseudogene of the Hugo haplotype), followed by HLA-DRB4 (the functional but nonpolymorphic gene of the DR4 haplotype). These relationships suggest that these six genes evolved from a common ancestor which existed before the separation of the human, gorilla, and chimpanzee lineages. The DRB2 and DRB6 have apparently been pseudogenes for at least six million years (myr). In the human and the gorilla haplotype, the DRB2 pseudogene is flanked on each side by what appear to be related genes. Apparently, the DR3 haplotype has existed in its present form for more than six myr.The nucleotide sequence data reported in this paper have been submitted to the GenBank nucleotide sequence database and have been assigned the accession number M86691–94.