Genotyping TAP2 Variants in North American Caucasians, Brazilians, and Africans

Genes and Immunity (2001) 2, 32–40  2001 Nature Publishing Group All rights reserved 1466-4879/01 $15.00 www.nature.com/gene Genotyping TAP2 variants in North American Caucasians, Brazilians, and Africans

J Tang1,2, DO Freedman2, S Allen3, E Karita4, R Musonda5, C Braga6, BD Jamieson7, L Louie8 and RA Kaslow1,2,3 1Program in Epidemiology of Infection and Immunity, School of Public Health, 2Division of Geographic Medicine, Department of Medicine; 3Department of Epidemiology and International Health, University of Alabama at Birmingham, Birmingham, AL 35294, USA; 4Tropical Disease Research Center, Ndona, Zambia; 5National AIDS Control Program, BP 780, Kigali, Rwanda; 6Centro des Pesquisas Aggeu Magalhaes, Recife, Brazil; 7Department of Medicine, UCLA School of Medicine, Los Angeles, CA 90095, USA; 8Children’s Hospital Oakland Research Institute, Oakland, CA 94609, USA

The protein forms of transporter associated with antigen processing, subunit 2 (TAP2), differ either by amino acid substitutions (Thr374Ala, Ile379Val, Ile467Val, Thr565Ala, Val577Met, Cys651Arg, and Ala665Thr) or by a truncation (Gln687Stop) of 17 amino acid residues at the C-terminus. Nonsynonymous single nucleotide polymorphisms (N-SNPs) causing these amino acid variations except 577Val were detected in genomic DNA samples from North American Caucasians (n = 76), Brazilians (n = 148), Rwandans (n = 285), and Zambians (n = 117). Exclusive (100%) and nearly exclusive (Ͼ95%) linkage disequilibrium was seen with a number of N-SNPs. The average heterozygosity at any given dimorphic site ranged from 7.3% to 44.6%, and at least four N-SNPs showed clear population specificity. N-SNP combinations alone led to the identification of 16 relatively common alleles, which appeared to form at least three lineages. Further analyses of 101 cDNA samples from Brazilians detected nine expressed TAP2 alleles, four of which matched the official assignments. Genetic complexity at the TAP2 locus was further enhanced by two out of five synonymous SNPs (S-SNPs), especially the GGT386GGG (Gly) that had similar heterozygosity rates in Caucasians (28.9%), Rwandans (33.3%), and Zambians (33.3%). Overall, distribution of both synonymous and nonsynonymous SNPs in the various ethnic groups examined here conformed well to the Hardy-Weinberg equilibrium, and between 57.9% and 77.0% of subjects in each ethnic group were heterozygous with two TAP2 alleles predicted to differ by at least one amino acid residue. Such complexity of TAP2 polymorphisms, in the form of SNPs as well as alleles, is likely to complicate the analyses of disease associations and haplotype structures in the HLA class II region. Genes and Immunity (2001) 2, 32–40.

Keywords: allele; genetics; polymorphism; population; SNP; TAP2

Introduction plex have also been carefully mapped.5–12 Homodimers of TAP1, TAP2, and oligomers of TAP1-TAP2 complex As important chaperones and peptide transporters in the are also potentially functional.13 For example, induction HLA class I antigen presentation pathway, the trans- or transfection of TAP1 alone enhances or restores the porters associated with antigen processing (TAP) have expression of class I molecules otherwise under- been well studied in relation to immunological speci- expressed on tumor cells.14,15 1–3 ficity. In general, two subunits, TAP1 and TAP2, func- Description of single nucleotide polymorphisms (SNPs) tion as a heterodimer, with each subunit containing seven at the human TAP2 locus (also known as PSF2, RING11, to eight N-terminal segments spanning the membrane of Tap2,orTAP-2)16–20 and documentation of functional dis- endoplasmic reticulum, along with a C-terminal domain tinctions for TAP variants in rats and Syrian hamsters 21–23 2–7 that binds ATP in the cytosol. The multiple domains have led to the investigation of their potential roles in mul- for binding peptides and those interacting with calnexin, tiple human diseases,24–31 transplantation,32,33 and vaccine ␤ tapasin, and the class I/beta 2-microglobulin ( 2m) com- response.34 The findings are largely inconsistent, especially when distinct ethnic groups are compared.35 More recently, TAP2 variants in combination with HLA class I Correspondence: Dr Jianming Tang, Program in Epidemiology of Infection alleles have been shown to mediate the rates of HIV-1 and Immunity, 610 Ryals Public Health Building, 1665 University Blvd., disease progression.36,37 The tight linkage disequilibrium University of Alabama at Birmingham, Birmingham, AL 35294-0022, between TAP2 alleles and those in the neighboring class Ȱ USA. E-mail: jtang uab.edu II loci including DQB1 and DRB118,38–40 is often used to Grant Support: This work has been supported jointly by Center for imply that TAP2 effects observed in epidemiological AIDS Research and the Health Services Foundation, University of Alabama at Birmingham. Additional support was provided from studies may just signify the involvement of additional grants AI-31552, AI-36661, AI-41951, AI-41530, and AI-40591 from molecules or their combinations on the extended haplo- NIAID. types.41–43 However, there is evidence that TAP2 variants Characterization of human TAP2 polymorphisms J Tang et al 33 alone can play primary or independent roles in human Consistent with our previous analyses of Rwandan and diseases.44 Zambian samples,39 several N-SNPs and S-SNPs were At the amino acid sequence level, TAP2 alleles can be always found on the same chromosome, making no defined by eight nonsynonymous SNPs (N-SNPs) at independent contribution to the allelic diversities. For codon positions 374 (Ala↔Thr in exon 5), 379 (Val↔Ile example, exclusive (100%) linkage disequilibrium was in exon 5), 467 (Val↔Ile in exon 7), 565 (Ala↔Thr in exon seen for 467Ile with four other sites: 374Thr, 458Thr 9), 577 (Met↔Val in exon 9), 651 (Arg↔Cys in exon 11), (ACA), and 466Gly (GGA); for 651Cys with two 665 (Thr↔Ala in exon 11), and 687 (Stop↔Gln in exon additional sites: 665Thr, 687Stop; for 379Ile with 374Ala; 11). The actual number of alleles involving these N-SNPs for 436Asn (AAT) with 665Ala-687Gln-697Val (GTT), and has not been sufficiently examined. To date, four TAP2 for 436AAT with 467Val. The 374Ala was more frequent alleles have been officially assigned by the WHO than 379Ile, and a proportion of 374Ala-related sequences Nomenclature Committee for factors in the HLA sys- did not carry 379Ile. On the other hand, nearly mutually tem,45 while 10 TAP2 alleles are anticipated by the 12th exclusive (Ͼ95%) linkage disequilibrium was observed IHWC TAP Workshop. Unofficial designations (local between 665Thr-687Stop and 665Ala-687Gln. names) are often used to describe alleles sharing specific The N-SNPs were differentially distributed among the N-SNPs, creating great difficulty in comparing TAP2- four ethnic groups: Caucasians lacked the 374Thr, 467Ile, associated experimental or epidemiological effects. In this and 565Thr; Rwandans, 565Thr; and Zambians, 651Cys study, we examined eight TAP2 N-SNPs and five synony- (Table 2). The 565Thr (TAP 2D) was also absent in Nigeri- mous SNPs (S-SNPs) in four ethnic groups in order to ans.46 In contrast, all seven N-SNPs were found in Brazili- define the common alleles, and to identify potentially ans at a higher or intermediate frequency compared with informative SNPs and alleles for future studies. Caucasians and Africans. The average heterozygosity at each N-SNP site ranged from 7.3% to 44.6% (Table 2), while the overall hetero- Results zygosity defined by the 7 N-SNP sites in TAP2 was 57.9% in Caucasians, 77.0% in Brazilians, 71.6% in Rwandans, TAP2 SNPs and allele structure and 63.2% in Zambians, all of which closely fit Hardy– Between 1994 and 2000, 626 genomic DNA samples Weinberg equilibrium. Greater TAP2 heterozygosity in derived from North American Caucasians, Brazilians, Brazilians compared with other groups probably Rwandans, and Zambians were typed for TAP2 polymor- reflected frequent racial admixture among local Caucasi- phisms using a combination of PCR-based techniques ans, Blacks, and Hispanics. Heterozygosity at N-SNPs in including single-strand conformation polymorphism about 20–25% of samples in each ethnic group did not (SSCP), restriction fragment length polymorphism allow the reliable assignment of allele pairs or portrayal (RFLP), nucleotide sequencing, and PCR with sequence of allelic distributions. (ie, SNP)-specific primers (SSP). The TAP2 coding region TAP2 alleles defined by seven N-SNPs in the four eth- was found to carry seven nonsynonymous single nucleo- nic groups could be split further by S-SNPs at codons tide polymorphisms (N-SNPs) that collectively produced 386 and 436. In particular, complete equilibrium between 16 alleles (Figure 1). Six of the TAP2 alleles (indicated by GGT386GGG, I379V, A665T, and Q687Stop further separ- arrow heads) were assigned unambiguously on the basis ated a number of protein-coding alleles into pairs of of repeated observations of homozygous genotypes, nucleotide sequences. In contrast, AAC436AAT only split while others could be established in heterozygous indi- the alleles carrying 665Ala-687Gln-697Val (GTT) because viduals following SSCP and PCR-SSP that linked specific of tight linkage disequilibria. SNP sites in the same or different exons. Four alleles (Figure 1, underlined) matched those recognized by the Analyses of TAP2 cDNA sequences WHO Nomenclature Committee for factors in the HLA PCR-SSP analyses of cDNA samples from 101 Brazilians system.45 Temporary designations were given to other also identified 7 N-SNP sites (Table 3). Previous geno- alleles, with considerations being given first to the adop- typing of a subset of 22 genomic DNA samples yielded tion of nomenclature proposed in previous work (Table identical results. Linking SNP sequences by RT-PCR gre- 1) or by the 12th International Histocompatibility Work- atly simplified the procedure for allele assignment, but shops (IHWC). However, neither local names nor IHW the alleles in 20 (19.8%) heterozygous genotypes (arrows) nomenclatures fully covered the N-SNP sites or their still could not be resolved. For the rest of samples, 13 combinations defined in our genotyping scheme. As a different genotypes could be attributed to nine different result, our temporary designations (Figure 1) could not alleles. Seven additional alleles (*01G, *01H, *02B, *02F, * be reliably related to any official or other consensus 03A, and *04A) detected in genomic DNA (Figure 1) did assignments available. not arise from these samples. The homozygosity and het- Five synonymous SNPs (S-SNPs), apart from codon erozygosity rates at each polymorphic site did not differ 163 (Val-GTC to Val-GTT, not targeted), were also exam- from the expected frequencies, suggesting again that ined: 386GGG to GGT (Gly), 436AAC to AAT (Asn), their distributions conformed to Hardy–Weinberg equi- 458GGG to GGA (Gly), 466GGG to GGA (Gly), and librium. 697GTG to GTT (Val) (at each site the predominant sequence is listed first) (Figure 1). These sites were not Attempts to fully assemble TAP2 allelic sequences included in the allele nomenclature. A rare variant, Detection of six homozygous TAP2 genotypes in various 513Ala (GCT) to Ser (TCT) in exon 8, was detected in one ethnic groups provided an opportunity to define two of the 76 Caucasian samples. The same individual would new alleles by direct sequencing without cloning; one of be carrying *0101 and *0102 if the 513Ala/Ser site were such alleles was also seen in TAP2 typing of homozygous excluded. cell lines (L Louie et al, unpublished). Sequencing of

Genes and Immunity Characterization of human TAP2 polymorphisms J Tang et al 34

Figure 1 Characterization of major TAP2 alleles defined by nonsynonymous single nucleotide polymorphisms (N-SNPs). The TAP2 gene consists of 11 exons (boxed and drawn to approximate scale). A recombination hot spot has been mapped to intron 2 region.55,84 The ATG translation initiation codon starts at nucleotide position 40428 on the extended genomic sequence (GenBank accession number X66401). Depending on the nucleotide sequence at codon 687, the stop codon in exon 11 can be either 687TAG (49753 to 49755) or 704TGA (49804 to 49806). The seven N-SNP sites (shown in seven-digit numbers) were commonly encountered in four ethnic groups. The predicted amino acid residues (in single-letter codes) are also shown, with codon positions defined by three-digit numbers. At each position the less common nucleotide (for S-SNPs) and amino acid sequence (for N-SNPs) is underlined. Open bidirectional arrows indicate nonconservative amino acid substitutions. Boxed N-SNPs are exclusively linked, with rare exceptions (by arrows). In addition, five synonymous SNPs (S-SNPs, in parentheses) are also defined. Meanwhile, the 577Val (Bky2) variant found earlier in Japanese58 could not be detected here.

genomic and cDNA was performed for the cell line adequate informative sites among the various human AWELLS, and genomic sequencing only was done for a TAP2 alleles, both maximum likelihood and neighbor- single Rwandan individual bearing homozygous 374Thr joining trees identified three allelic lineages, with the *01 and 467Ile variants. Alignment of exon 5 to exon 6 group including three officially assigned alleles (*0101, sequences further revealed polymorphisms in intron 5 *0102, and *0103), and the *02 group with a single official (GenBank accession numbers AF100416 and AY007425). allele (*0201). Allele *03A was placed much deeper in the Thus, identical TAP2 mRNA or cDNA sequences could tree than any other human alleles, while the placement of be encoded by very different genomic sequences. the allele *04A differed between the two methods used. Designations of various TAP2 alleles without official Evolutionary relationships of TAP2 alleles assignments were solely based on these preliminary Complete TAP2 sequences have been described for the phylogenetic analyses, which offered generally weak mouse (Mus musculus),47 rat (Rattus norvegicus),48 quail support for the various clades and their internal top- (Coturnix japonica),49 chicken (Gallus gallus),50 Atlantic sal- ologies. Overall, allele *03A appeared to be the most mon (Salmo salar),51 rainbow trout (Oncorhynchus ancestral human TAP2 allele, but the relationships among mykiss),52 frog (Xenopus laevis),53 as well as gorilla (Gorilla other alleles within each group were poorly resolved. gorilla).54 Partial sequences are further available for the 55 common chimpanzee (Pan troglodytes), cattle (Bos Discussion taurus),56 and zebrafish (Danio rerio).57 Overall, all species carry TAP2 alleles у 699–703 amino acids in length, while TAP2 typing in four ethnic groups yielded 16 alleles pre- some humans carry alleles with only 686 amino acids. dicted to carry different combinations of amino acid Alignment of TAP2 codon sequences that differ within sequences defined by seven dimorphic sites alone. While and between humans and gorillas were used to produce some alleles (eg, *0101, *0201) were common regardless phylogenetic trees (data not shown). In the absence of of ethnic background, others were present in low fre-

Genes and Immunity Characterization of human TAP2 polymorphisms J Tang et al 35 Table 1 Ofﬁcial and local designations of TAP2 variants deﬁned by amino acid sequences

374a 379 467a 565 651 665 687 References

*0101 A V V A R T Stop 1 *0102 A V V T R T Stop 19 *0103 A V V A C T Stop 80 *0201 A V V A R AQ 19 2.1 I 18 2.2 V 18 2.3 A 18 2.4 T18 2.5 Q 18 2.6 Stop 18 2A V T Stop 18 2B V AQ 18 2C I T Stop 18 2D IAQ18 2Ab(=*0101) V A T Stop 19, 81 2Bb (=*0201) V A A Q 19, 81 2C I A T Stop 19 2D ITT Stop 19 2E (=*0102) V T T Stop 19, 81 2F (=*0103) V A C T Stop 81, 82 2G V A R T Q 82 A R T Stop 83 B R AQ 83 C C T Stop 83 *0202 I ARAQ 82 +++c T Stop 69 −−−c AQ 69 +−+c T Q 69 Nw1 or Nw2d T V I A R T Stop 39 Nw3 or Nw4d T V V A R T Stop 39 Nw5d T VVARAQ 39 aPositions not included in the 12th IHWC TAP workshop. bCan also match PSF2A, PSF2B, RING11A, and RING11B16,17. cAccording to nucleotide positions 1993, 2059, and 2091 in the TAP2 cDNA sequence. dSeveral synonymous sites were also included in these allele designations. quency only in some ethnic groups. Such distinct popu- TAP2-like sequences have been repeatedly detected in lation specificity may further apply to the TAP2-577Val humans (eg, GenBank sequence AF097669),65 rats,66 and variant commonly found in Japanese.58 A systematic Xenopus.53 Indeed, salmon and rainbow trout carry two documentation of TAP2 polymorphisms should benefit separate TAP2 loci (TAP2A and TAP2B).51,52 Both TAP2 direct comparison of experimental and epidemiological genes appeared to be polymorphic and transcribed, findings based on different ethnic groups. However, the although the locus linked to MHC was expressed at a complexity of TAP2 polymorphisms, as reflected by the much higher level than the homologue not linked to degree of heterozygosity in the coding and non-coding MHC.52 Recombination is often associated with CpG-rich sequences, poses great difficulties in such efforts. A more regions67 and further requires signal sequence motifs;68 selective genotyping of major TAP2 N-SNPs based on some of the candidates have been mapped within the their frequencies may become an effective approach in TAP2 locus, and probably explain the recombination hot the early stages of studies. spot in intron 2.55,68 More definitive data may lie in TAP2 Mechanisms responsible for the generation of such intronic sequences.69 complex TAP2 polymorphisms are probably similar to The relationships between TAP2 polymorphisms and those recognized for classic HLA genes: point mutation function of the TAP1/TAP2 heterodimer are debatable. and gene conversion coupled with balancing selec- Initial experimental data have largely rejected the likeli- tion.59,60 For example, the 467Ile variant in human TAP2 hood of functional distinctions caused by human TAP2 has shown exclusive linkage disequilibria with two polymorphisms.70 Examination of various ethnic groups additional synonymous nucleotide substitutions at further suggested that the distribution of TAP2 variants codons 458 and 466 in the same exon, but no other inter- does not appear to the result of vigorous selection.71 mediate alleles were detected. Such phenomenon is However, the evidence has been limited to: (1) analyses strongly indicative of gene conversion/recombination, based on few TAP2 proteins and their TAP1 partners which has been recognized as a driving force for HLA commonly found in Caucasians; (2) in vitro studies of cell allele and haplotype diversities,61–64 with pseudogenes lines with restricted/modified functions; and (3) only a and probably paralogous genes serving as the potential subset of the possible TAP2 SNPs. The complexity of source of donor sequences.65 TAP2 pseudogene and TAP2 polymorphisms, as implied in our identification of

Genes and Immunity Characterization of human TAP2 polymorphisms J Tang et al 36 Table 2 Frequency and heterozygosity of TAP2 single nucleotide polymorphisms (SNPs) observed in four ethnic groups

SNP sequencesa T374A I379V I467V T565A C651R A665T Q687Stop GGT386GGG AAT436AAC

Caucasians (n = 76) SNP frequencies: 1.000 0.816 1.000 0.993 0.947 0.849 0.849 0.814 0.908 0.000 0.184 0.000 0.007 0.053 0.151 0.151 0.184 0.092 Heterozygosity: Observed 0 31.6% 0 1.3% 10.5% 25.0% 25.0% 28.9% 18.4% Expected 0 30.0% 0 1.4% 10.0% 25.7% 25.7% 30.0% 16.7% Brazilians (n = 148) SNP frequencies: 0.980 0.868 0.983 0.899 0.960 0.676 0.676 nd nd 0.020 0.132 0.017 0.101 0.040 0.324 0.324 nd nd Heterozygosity: Observed 4.0% 25.0% 3.4% 20.3% 8.1% 54.7%b 54.7%b nd nd Expected 3.9% 22.9% 3.3% 18.2% 7.7% 43.8% 43.8% nd nd Rwandans (n = 285) SNP frequencies: 0.933 0.882 0.947 1.000 0.900 0.698 0.694 0.806 0.993 0.067 0.118 0.053 0.000 0.100 0.302 0.306 0.194 0.007 Heterozygosity: Observed 13.0% 22.5% 10.2% 0 18.6% 43.1% 45.3% 33.3% 1.4% Expected 12.5% 20.8% 10.0% 0 18.0% 42.2% 42.4% 31.3% 11.4% Zambians (n = 117) SNP frequencies: 0.932 0.898 0.949 0.817 1.000 0.765 0.752 0.778 0 0.068 0.102 0.051 0.183 0 0.235 0.248 0.222 0 Heterozygosity: Observed 13.7% 18.8% 10.3% 35.0% 0 40.2% 42.7% 33.3% 0 Expected 12.7% 18.3% 9.7% 29.9% 0 36.0% 37.3% 34.5% 0 Average 9.4% 23.5% 7.3% 11.5% 11.7% 43.1% 44.6% 32.6% 5.1% heterozygosity

aAt each codon position the less common amino acid or nucleotide sequence is listed ﬁrst and/or underlined as deﬁned in Figure 1. bP = 0.185 compared with the expected frequency. For all other sites P Ͼ 0.559 based on Mantel-Haenszel ␹2 tests.

Table 3 TAP2 genotypes observed in 101 cDNA samples from Brazilians

Exon 5 to exon 7 Exon 9 to exon 11 Genotypeb n 374-379-467a 565-651-665-687a

A-V-V A-R-T-Stop *0101 homozygous 20 A-V-V A-R-A-Q *0201 homozygous 7 A-V-V A-R-T-Stop, A-R-A-Q *0101, 0201 31 A-V-V A-R-T-Stop, T-R-T-Stop *0101, 0102 7 A-V-V A-R-T-Stop, A-C-T-Stop *0101, 0103 6 A-V-V T-R-T-Stop, A-R-A-Q *0102, 0201 1 A-V-V, A-I-V A-R-T-Stop *0101, 01E 2 A-I-V A-R-T-Stop *01E, homozygous 1 A-V-V, T-V-I A-R-T-Stop *0101, 01F 1 A-I-V A-R-T-Stop, T-R-T-Stop *01D, 01E 1 A-V-V, A-I-V A-R-A-Q *0201, 02C 2 A-V-V, T-V-I A-R-A-Q *0201, 02D 1 A-V-V T-R-T-Stop, A-R-A-Q *0102, 0201 1 A-V-V, A-I-V A-R-T-Stop, T-R-T-Stop *0101, 01D/*0102, 01E 7Ȥ A-V-V, A-I-V A-R-T-Stop, A-R-A-Q *0101, 02C/*0201, 01E 7Ȥ A-V-V, A-I-V T-R-T-Stop, A-R-A-Q *0102, 02C/*0201, 01D 4Ȥ A-V-V, A-I-V A-R-T-Stop, T-R-A-Q *0101, 02F/01E, 02E 1Ȥ A-I-V, T-V-IT-R-T-Stop, A-R-A-Q *01D, 02D/*02C, 01x? 1Ȥ

aLinked by PCR using SNP-speciﬁc primers. bSeven alleles (*01G, *01H, *02B, *02E, *02F, *03A, and *04A) could not be detected unambiguously among these samples.

multiple TAP2 alleles other than those bearing ofﬁcial only in humans and not in any other vertebrates. TAP2.3 assignments, questions any assumption in the non- (=665Ala), the variant rarely found in the truncated TAP2 Caucasian groups. For example, mutagenesis of human molecules, has been shown in three cohort studies as one TAP2 sequences to mimic polymorphisms observed in of the factors contributing to delayed disease progression rats has revealed a critical role for the 374 position in among HIV-1-infected Caucasians.36,37 The same epide- mediating peptide transport speciﬁcity.72 On that basis, miological studies also highlight the importance of HLA the naturally occurring Thr374Ala in Africans and Brazil- class I and TAP combinations and/or interactions, which ians deserves particular attention. It is also interesting are coherent with the antigen presentation pathway but that the truncation of TAP2 protein at codon 687 occurs have largely escaped the attention of most investigators.

Genes and Immunity Characterization of human TAP2 polymorphisms J Tang et al 37 Table 4 Oligonucleotides used in PCR-SSP and sequencing analyses of human TAP2 variants

Oligo name Speciﬁcity 5Ј⇒ 3Ј sequences Annealing sites: Annealing sites: (underlined = polymorphic) genomic DNAa cDNAa

TAP2-5/1 374Ala ggA gAg ACC Tgg AAC gCg Exon 5, 121783 ⇒ 121800 1179 ⇒ 1196 TAP2-5/2 374Thr Cgg AgA gAC CTg gAA CgC A Exon 5, 121782 ⇒ 121800 1178 ⇒ 1196 TAP2-5/3 379Val CgC gCC TTg TAC CTg CTC g Exon 5, 121797 ⇒ 121815 1193 ⇒ 1211 TAP2-5/4 379Ile CgC gCC TTg TAC CTg CTC A Exon 5, 121797 ⇒ 121815 1193 ⇒ 1211 TAP2-3/1 TAP2, general Agg CTT ggC TTC TCg CTC AC Intron 5, 122118 ⇐ 122137 NA TAP2-3/2 467Val ACg TCT Tgg AAT TTC ACA ACC Exon 7, 123769 ⇐ 123789 1474 ⇐ 1494 TAP2-3/3 467Ile gAC gTC TTg gAA TTT CAC AATT Exon 7, 123769 ⇐ 123790 1474 ⇐ 1495 TAP2-5/5 TAP2, general TTg gAA CAC Tgg ggT ATT gg Intron 6, 123459 ⇒ 123478 NA TAP2-5/6 565Ala ggT TCT gTg Agg AAC AAC ATT g Exon 9, 124397 ⇒ 124418 1748 ⇒ 1769 TAP2-5/7 565Thr ggT TCT gTg Agg AAC AAC ATT A Exon 9, 124397 ⇒ 124418 1748 ⇒ 1769 TAP2-5/8 TAP2, general CCT CAg ATg TAg ggg AgA Ag Intron 10, 124897 ⇒ 124927 NA TAP2-5/9 577Met gCT gCg AAg ATg ATA Agg TgA Exon 9, 124434 ⇒ 124454 1785 ⇒ 1805 TAP2-5/10B 577Val CTg CgA AgA TgA TAA ggT gg Exon 9, 124435 ⇒ 124454 1786 ⇒ 1805 TAP2-3/4 651Arg CAC TgT gCg ATC CCC ACg Exon 11, 125434 ⇐ 125451 2037 ⇐ 2044 TAP2-3/5 651Cys CAC TgT gCg ATC CCC ACA Exon 11, 125434 ⇐ 125451 2037 ⇐ 2044 TAP2-3/6b 665Thr gTg ggC gCg CTg AAC TgT Exon 11, 125476 ⇐ 125493 2069 ⇐ 2086 TAP2-3/7 665Ala ggT ggg CgC gCT gAA CTg C Exon 11, 125476 ⇐ 125494 2069 ⇐ 2087 TAP2-3/8c 687Stop ATA gAg gTC CTg TCC CTC CTA Exon 11, 125542 ⇐ 125562 2234 ⇐ 2255 TAP2-3/9c 687Gln ATA gAg gTC CTg TCC CTC CTg Exon 11, 125542 ⇐ 125562 2234 ⇐ 2255 TAP2-3/10 TAP2, general ACT gTC CCC TcG CCT CTC Intron 10, 125062 ⇐ 125079 NA DRB1-C5 DRB1, general TgC CAA gTg gAg CAC CCA A Codon 173 ⇒ 179 in exon 3 NA DRB1-C3 DRB1, general gCA TCT TgC TCT gTg CAg AT Codon 200 ⇐ 193 in exon 4 NA TAP2-c5a TAP2, general gCT gAA gCA gAA gTC CCC gg NA 28 ⇒ 37 TAP2-c3a TAP2, general gCT CAg CCg TgA gTT CAg CTC Exon 4, 119113 ⇐ 119093 818 ⇐ 838 aThe numbering of nucleotides in TAP2 genomic and cDNA sequences is based on GenBank accession numbers X87344 and Z22936 (=TAP2* 0102), respectively. Subjects and methods polymorphism (SSCP) analyses and sequencing;39 (2) 24 additional oligos for specific amplification of TAP2 Subjects sequences carrying individual SNPs (Table 4). The same A total of 626 subjects were pooled from studies of 24 oligos along with several other internal ones (not diverse disease manifestations of HIV/AIDS or lym- shown) also served as the sequencing primers. In phatic Wuchereria bancrofti infection in four geographi- addition, when cDNA served as the templates, PCR with cally different areas. For studies on HIV/AIDS, criteria combinations of the SNP-specific primers was able to link for selecting subjects included: (1) HIV-1 infection and T374A, I379V with I467V, as well as T565A, C651R with risk for infection that determined the eligibility of indi- A665T. Haplotyping using the SNP-specific primers was viduals; (2) availability of genotyping materials and also possible, but the PCR conditions outlined below clinical data that allowed analyses of outcomes; required further modifications (available upon request). (3) preliminary epidemiological data that suggested the associations between TAP2 polymorphisms and HIV-1 disease progression. Patients and control subjects in the Characterization of TAP2 variants W. bancrofti study were selected on a more random basis. Segments (146–313 bp in size) of TAP2 exon sequences The ethnic backgrounds represented were pure native amplified by PCR were first analyzed by single-strand African (Rwanda73–75 and Zambia76,77), largely North conformation polymorphism (SSCP), followed by selec- American Caucasian (United States, Multicenter AIDS tive sequencing, and restriction digestion with appropri- 39 Cohort Study (MACS) (http://statepi.sph.jhu.edu/ ate endonuclease as described earlier. PCR-based classi- macs/macs.html)), and mixed Caucasian/Hispanic/ fication of SNPs and their combinations further relied on 78 the use of control and SNP-specific primers. Briefly, each African (Recife, Brazil ). A homozygous typing cell line ␮ × (AWELLS) was used to define the complete coding PCR mix (10 l) consisted of 1 buffer C (60 mm Tris- sequence of a common TAP2 allele. HCI, pH 8.5, 15 mm (NH4)2SO4, 2.5 mm MgCl2), 50–70 ng of genomic DNA, 0.3 units of AmpliTaq polymerase DNA samples (Fisher Scientific, Norcross, GA, USA), 120 nm of each Genomic DNA was prepared from buffy coats, purified control primer, 250 nm each of specific primer, 0.4 mm peripheral mononuclear cells, or immortalized B cell lines each of dGTP, dCTP, dTTP and dATP, 10% (v/v) gly- using the standard salting out technique and the QIAamp cerol, and 0.02% cresol red. PCR cycling began with 10 blood kit (QIAGEN Inc., Chatsworth, CA, USA). All higher-stringency cycles of denaturing at 95°C for 25 s, DNA samples were diluted to 150–200 ng/␮l and stored annealing at 65°C for 45 s, and extension at 72°C for 45 s; at 4°C in TE buffer (10 mm Tris-HCl, pH 8.0, 2 mm EDTA) it was followed by 22 more lower-stringency cycles of before use. denaturing at 95°C for 25 s, annealing at 61°C for 40 s, and extension at 72°C for 40 s. Half of each PCR reaction PCR primers product was loaded directly to 1.5% agarose gel for The oligonucleotides used in this study included: (1) 12 electrophoresis, and the banding patterns were recorded pairs to amplify 11 exons for single-strand conformation on photographs of ethidium bromide-stained gels.

Genes and Immunity Characterization of human TAP2 polymorphisms J Tang et al 38 Analyses of TAP2 allelic transcripts 10 Nijenhuis M, Hammerling GJ. Multiple regions of the trans- A subset of 101 complementary DNA (cDNA) samples porter associated with antigen processing (TAP) contribute to derived from Brazilians and another cDNA from its peptide binding site. J Immunol 1996; 157: 5467–5477. AWELLS homozygous cell line were analyzed to define 11 Suh WK, Mitchell EK, Yang Y et al. MHC class I molecules form the structure of transcribed TAP2 alleles. Procedures for ternary complexes with calnexin and TAP and undergo peptide- regulated interaction with TAP via their extracellular domains. mRNA extraction and reverse transcription (RT) have 78 J Exp Med 1996; 184: 337–348. been described elsewhere. RT-PCR amplification of 12 Vos JC, Spee P, Momburg F, Neefjes J. Membrane topology and cDNA sequences followed protocols similar to those used dimerization of the two subunits of the transporter associated for amplifying genomic DNA, except that AmpliTaq was with antigen processing reveal a three-domain structure. J replaced by Platinum Taq (GIBCO BRL Life Techno- Immunol 1999; 163: 6679–6685. logies, Grand Island, NY, USA) to facilitate the hot-start 13 Lapinski PE, Miller GG, Tampe R, Raghavan M. Pairing of the PCR, and that the lower-stringency cycle number was nucleotide binding domains of the transporter associated with increased from 22 to 27. antigen processing. J Biol Chem 2000; 275: 6831–6840. 14 Zhu K, Wang J, Zhu J et al. p53 induces TAP1 and enhances the Phylogenetic analyses transport of MHC class I peptides. Oncogene 1999; 18: 7740–7747. The phylogenetic relationships among human TAP2 15 Singal DP, Ye M, Bienzle D. Transfection of TAP1 gene restores HLA class I expression in human small-cell lung carcinoma. Int alleles detected in various ethnic groups were analyzed 79 J Cancer 1998; 75: 112–116. using subsets of software in the PHYLIP packages. Two 16 Colonna M, Bresnahan M, Bahram S, Strominger JL, Spies T. 54 TAP2 sequences from gorillas (Gorilla gorilla) served as Allelic variants of the human putative peptide transporter the outgroup. The final sequence alignment included involved in antigen processing. Proc Natl Acad Sci USA 1992; 89: only 11 codon sequences carrying N-SNPs and two 3932–3936. codons (458, 697) with S-SNPs that were exclusively 17 Powis SH, Mockridge I, Kelly A et al. Polymorphisms in a linked to one or more N-SNPs within and across species. second ABC transporter gene located within the class II region Concordance in topology between maximum likelihood of the human major histocompatibility complex. Proc Natl Acad and neighbor-joining trees was treated as indicative of Sci USA 1992; 89: 1463–1437. the most likely evolutionary relationships. 18 Carrington M, Colonna M, Spies T, Stephens JC, Mann DL. Haplotypic variation of the transporter associated with antigen processing (TAP) genes and their extension of HLA class II Acknowledgements region haplotypes. Immunogenetics 1993; 37: 266–273. 19 Powis SH, Tonks S, Mockridge I et al. Alleles and haplotypes of We thank staff and study participants in the various stud- the MHC-encoded ABC transporters TAP1 and TAP2 [pub- ies in the US, Brazil, Rwanda, and Zambia. We also thank lished erratum appeared in Immunogenetics, 1993 Vol 37, page Adam Plier, Charles R Rivers, Caroline Costello, Yuting 480]. Immunogenetics 1993; 37: 373–380. Zhang, Pingping Xia, Xiaochuan Bai, Jareen Meizen-Derr, 20 Aoki Y, Isselbacher KJ, Pillai S. Polymorphisms involving the Ulgen Fideli, and Kathy Schultz for technical assistance transmembrane domains of human TAP2. Immunogenetics 1993; and data management. 38: 382. 21 Powis SJ, Young LL, Joly E et al. The rat cim effect: TAP allele- dependent changes in a class I MHC anchor motif and evidence References against C-terminal trimming of peptides in the ER. Immunity 1996; 4: 159–165. 1 Kelly A, Powis SH, Kerr LA et al. Assembly and function of 22 Wang P, Gyllner G, Kvist S. Selection and binding of peptides the 2 ABC transporter proteins encoded in the human major to human transporters associated with antigen processing and histocompatibility complex. Nature 1992; 355: 641–644. rat cim-a and -b. J Immunol 1996; 157: 213–220. 2 Androlewicz MJ, Ortmann B, van Endert PM, Spies T, Cresswell 23 Lobigs M, Mullbacher A, Blanden RV, Hammerling GJ, Mom- P. Characteristics of peptide and major histocompatibility com- burg F. Antigen presentation in syrian hamster cells: substrate plex class I/beta 2-microglobulin binding to the transporters selectivity of TAP controlled by polymorphic residues in TAP1 associated with antigen processing (TAP1 and TAP2). Proc Natl and differential requirements for loading of H2 class I mol- Acad Sci USA 1994; 91: 12716–12720. ecules. Immunogenetics 1999; 49: 931–941. 3 Momburg F, Hammerling GJ. Generation and TAP-mediated 24 van Endert PM, Lopez MT, Patel SD, Monaco JJ, McDevitt HO. transport of peptides for major histocompatibility complex class Genomic polymorphisms, recombination, and linkage disequi- I molecules. Adv Immunol 1998; 68: 191–256. librium in human major histocompaitbility complex-encoded 4 Spies T, Bresnahan M, Baram S et al. A gene in the human major antigen processing genes. Proc Natl Acad Sci USA 1992; 89: histocompatibility complex class II region controlling the class 11594–11597. I antigen presentation pathway. Nature 1990; 348: 744–747. 25 Powis SH, Rosenberg WM, Hall M et al. TAP1 and TAP2 poly- 5 Suh W, Cohen-Doyle MF, Fruh K et al. Interaction of MHC class Immunogenetics 38 I molecules with the transporter associated with antigen pro- morphism in celiac disease. 1993; : 345–350. cessing. Science 1994; 264: 1322–1326. 26 Donato L, de la Salle H, Hanau H et al. Association of HLA 6 Heemels MT, Ploegh H. Generation, translocation, and presen- class I antigen deficiency related to a TAP2 gene mutation with tation of MHC class I-restricted peptides. Annu Rev Biochem familial bronchiectasis. J Pediatr 1995; 127: 895–900. 1995; 64: 463–491. 27 Moins-Teissernc H, Bobrynina V, Loiseau P, Charron D. New 7 Suh WK, Derby MA, Cohen-Doyle MF et al. Interaction of polymorphisms within the human TAP1 and TAP2 coding murine MHC class I molecules with tapasin and TAP enhances regions. Immunogenetics 1995; 42: 308. peptide loading and involves the heavy chain alpha3 domain. 28 Takeuchi F, Nakano K, Nabeta H et al. Polymorphisms of the J Immunol 1999; 162: 1530–1540. TAP1 and TAP2 transporter genes in Japanese SLE. Ann Rheum 8 Heemels MT, Ploegh HL. Substrate specificity of allelic variants Dis 1996; 55: 924–926. of the TAP peptide transporter. Immunity 1994; 1: 775–784. 29 Rau H, Nicolay A, Usadel KH et al. Polymorphisms of TAP1 9 Momburg F, Armandola EA, Post M, Hammerling GJ. Residues and TAP2 genes in Graves’ disease. Tissue Antigens 1997; 49: in TAP2 peptide transporters controlling substrate specificity. 16–22. J Immunol 1996; 156: 1756–1763. 30 Takeuchi F, Nakano K, Matsuta K et al. Polymorphism of TAP1

Genes and Immunity Characterization of human TAP2 polymorphisms J Tang et al 39 and TAP2 in Japanese patients with rheumatoid arthritis. Tissue Pasquier L. Expression, linkage, and polymorphism of MHC- Antigens 1997; 49: 280–282. related genes in rainbow trout, Oncorhynchus mykiss. J Immunol 31 Foley PJ, Lympany PA, Puscinska E et al. Analysis of MHC enco- 1999; 163: 774–786. ded antigen-processing genes TAP1 and TAP2 polymorphisms 53 Ohta Y, Powis SJ, Coadwell WJ et al. Identification and genetic in sarcoidosis. Am J Respir Crit Care Med 1999; 160: 1009–1014. mapping of Xenopus TAP2 genes. Immunogenetics 1999; 49: 32 Kobayashi T, Yokoyama I, Hayashi S et al. TAP1, TAP2, LMP2, 171–182. DMA, and DMB genetic polymorphisms in renal transplan- 54 Loflin PT, Laud PR, Watkins DI, Lawlor DA. Identification of tation. Transplant Proc 1998; 30: 29–30. new TAP2 alleles in gorilla: evolution of the locus within homi- 33 Kobayashi T, Yokoyama I, Inoko H et al. Significance of trans- noids. Immunogenetics 1996; 44: 161–9. porter associated with antigen processing gene polymorphism 55 Jeffreys AJ, Ritchie A, Neumann R. High-resolution analysis of in living related renal transplantation. Hum Immunol 2000; 61: haplotype diversity and meiotic crossover in the human TAP2 670–674. recombination hotspot. Hum Mol Genet 2000; 9: 725–733. 34 Hayney MS, Poland GA, Dimanlig P et al. Polymorphisms of 56 Hess M, Goldammer T, Gelhaus A et al. Physical assignment of the TAP2 gene may influence antibody response to live measles the bovine MHC class IIa and class IIb genes. Cytogenet Cell vaccine virus. Vaccine 1997; 15: 3–6. Genet 1999; 85: 244–247. 35 Ocal L, Russell K, Beynon H et al. Genetic analysis of TAP2 in 57 Takami K, Zaleska-Rutczynska Z, Figueroa F, Klein J. Linkage systemic lupus erythematosus patients from two ethnic groups. of LMP, TAP, and RING3 with Mhc class I rather than class II Br J Rheumatol 1996; 35: 529–533. genes in the zebrafish. J Immunol 1997; 159: 6052–6060. 36 Kaslow RA, Carrington M, Apple R et al. Influence of combi- 58 Kumagai S, Kanagawa S, Morinobu A et al. Associations of a nations of human major histocompatibility complex genes on new allele of the TAP2 gene, TAP2*Bky2 (Val577), with suscepti- the course of HIV-1 infection. Nat Med 1996; 2: 405–411. bility to Sjogren’s syndrome. Arthritis Rheum 1997; 40: 1685– 37 Keet IPM, Tang J, Klein MR et al. Consistent associations of HLA 1692. class I and class II and transporter gene products with pro- 59 Hughes AL, Hughes MK. Natural selection on the peptide-bind- gression of human immunodeficiency virus-1 infection in homo- ing regions of major histocompatibility complex molecules. sexual men. J Infect Dis 1999; 180: 299–309. Immunogenetics 1995; 42: 233–243. 38 Djilali-Saiah I, Benini V, Daniel S et al. Linkage disequilibrium 60 Hughes AL, Yeager M, Carrington M. Peptide binding function between HLA class II (DR, DQ, DP) and antigen processing and the paradox of HLA disease associations. Immunol Cell Biol (LMP, TAP, DM) genes of the major histocompatibility complex. 1996; 74: 444–448. Tissue Antigens 1996; 48: 87–92. 61 Hughes AL, Hughes MK, Watkins DI. Contrasting roles of inter- 39 Tang J, Allen S, Karita E, Musonda R, Kaslow RA. New TAP2 allelic recombination at the HLA-A and HLA-B loci. Genetics polymorphisms in Africans. Tissue Antigens 1998; 51: 556–562. 1993; 133: 669–680. 40 Mora B, Battat S, Grillo R et al. Strong linkage disequilibrium of 62 McAdam SN, Boyson JE, Liu X et al. A uniquely high level of TAP1*0301 and TAP2D alleles with the HLA A1-B35-DRB1* recombination at the HLA-B locus. Proc Natl Acad Sci USA 1994; 1104-DQA1*0103-DQB1*0603 extended haplotype in Ashkenazi 91: 5893–5897. Jews. Eur J Immunogenet 1999; 26: 331–335. 63 Zangenberg G, Huang M-M, Arnheim N, Erlich H. New HLA- 41 Ronningen KS, Undlien DE, Ploski R et al. Linkage disequilib- DPB1 alleles generated by intraallelic gene conversion detected rium between TAP2 variants and HLA class II alleles; no pri- by analysis of sperm. Nat Genet 1995; 10: 407–414. mary association between TAP2 variants and insulin-dependent 64 Hogstrand K, Bohme J. Gene conversion can create new MHC diabetes mellitus. J Immunol 1993; 23: 1050–1056. alleles. Immunol Rev 1999; 167: 305–317. 42 Caillat-Zucman S, Daniel S, Djilali-Saiah I et al. Family study of 65 Kasahara M. Genome dynamics of the major histocompatibility linkage disequilibrium between TAP2 transporter and HLA complex: insights from genome paralogy. Immunogenetics 1999; class II genes. Absence of TAP2 contribution to association with 50: 134–145. insulin-dependent diabetes mellitus. Hum Immunol 1995; 44: 66 Yamaguchi Y, Kasano M, Terada T, Sato R, Maeda M. An ABC 80–87. transporter homologous to TAP proteins. FEBS Lett 1999; 457: 43 Hilarby MC, Davies EJ, Donn RP, Grennan DM, Ollier WE. TAP 231–236. 2D is associated with HLA-B44 and DR4 and may contribute to 67 Hogstrand K, Bohme J. Gene conversion of major histocompat- rheumatoid arthritis and Felty’s syndrome. Clin Exp Rheumatol ibility complex genes is associated with CpG-rich regions. 1996; 14: 67–70. Immunogenetics 1999; 49: 446–455. 44 Keicho N, Tokunaga K, Nakata K et al. Contribution of TAP 68 Cullen M, Nobel J, Erlich H et al. Characterization of recombi- genes to genetic predisposition for diffuse panbronchiolitis. nation in the HLA class II region. Am J Hum Genet 1997; 60: Tissue Antigens 1999; 53: 366–373. 397–407. 45 Bodmer JG, Marsh SGE, Albert ED et al. Nomenclature for fac- 69 Carrington M, Stephens JC, Ma W-Y et al. Identification of a tors of the HLA system, 1998. Tissue Antigens 1999; 53: 407–446. novel TAP2 allele in a Colombian black population: gene con- 46 Awomoyi AAO, Donn RP, Davies EJ et al. Absence of TAP 2D version, ancestral intermediate, or convergent change? Mol Bio in Yoruba Nigerians. Eur J Immunogen 1995; 22: 187–190. Evol 1997; 14: 892–894. 47 Marusina K, Iyer M, Monaco JJ. Allelic variation in the mouse 70 Obst R, Armandola EA, Nijenhuis M, Momburg F, Ham- Tap-1 and Tap-2 transporter genes. J Immunol 1997; 158: 5251– merling GJ. TAP polymorphism does not influence transport 5256. of peptide variants in mice and humans. Eur J Immunol 1995; 48 Joly E, Deverson EV, Coadwell WL et al. The distribution of 25: 2170–2176. TAP2 alleles amongst laboratory RAT RT1 haplotypes. unpub- 71 Faucz FR, Probst CM, Petzl-Erler ML. Polymorphism of LMP2, lished, GenBank accession number X75307 1993. TAP1, LMP7 and TAP2 in Brazilian Amerindians and Cauca- 49 Shiina T, Shimizu C, Oka A et al. Gene organization of the quail soids: implications for the evolution of allelic and haplotypic major histocompatibility complex (MhcCoji) class I gene region. diversity. Eur J Immunogenet 2000; 27: 5–16. Immunogenetics 1999; 49: 384–394. 72 Armandola EA, Momburg F, Nijenhuis M et al. A point 50 Guillemot F, Billault A, Pourquie O et al. A molecular map of mutation in the human transporter associated with antigen pro- the chicken major histocompatibility complex: the class II beta cessing (TAP2) alters the peptide transport specificity. Eur J genes are closely linked to the class I genes and the nucleolar Immunol 1996; 26: 1748–1755. organizers. EMBO J 1988; 7: 2775–2785. 73 Lifson AR, Allen S, Wolf W et al. Classification of HIV infection 51 Grimholt U. Transport-associated proteins in Atlantic salmon and disease in women from Rwanda. Ann Intern Med 1995; 122: (Salmo salar). Immunogenetics 1997; 46: 213–221. 262–270. 52 Hansen JD, Strassburger P, Thorgaard GH, Young WP, Du 74 Lindan CP, Allen S, Serufilira A et al. Predictors of mortality

Genes and Immunity Characterization of human TAP2 polymorphisms J Tang et al 40 among HIV-infected women in Kigali, Rwanda. Ann Intern Med 79 Felsenstein J. PHYLIP (Phylogeny Inference Package). 1992; 116: 320–328. Department of Genetics, University of Washington: Seattle, 75 Tang J, Costello C, Keet IPM et al. HLA class I homozygosity WA, 1993. accelerates disease progression in human immunodeficiency 80 Cano P, Baxter-Lowe LA. Novel human TAP2*103 allele shows virus type 1 infection. AIDS Res Hum Retroviruses 1999; 15: further polymorphism in the ATP-binding domain. Tissue Anti- 317–324. gens 1995; 45: 139–142. 76 McKenna SL, Muyinda GK, Roth D et al. Rapid HIV testing and 81 Ishihara M, Ohno S, Ishida T et al. Analysis of allelic variation of counseling for voluntary testing centers in Africa. AIDS 1997; the TAP2 gene in sarcoidosis. Tissue Antigens 1997; 49: 107–110. 11 (Suppl 1): S103-S110. 82 Cesari MM, Dulay SJ, Caillens H et al. A new human transporter 77 Meinzen-Derr J, Kautzman M, Haworth A et al. Self-report of associated with antigen processing allele encodes a large C- condom use among cohabiting heterosexual discordant couples terminal protein domain. Immunogenetics 1997; 45: 280–281. in Lusaka, Zambia: validation with biological markers. Am J Pub 83 Szafer F, Oksenberg JR, Steinman L. New allelic polymorphisms Health 2001; (in press). in TAP genes. Immunogenetics 1994; 39: 374. 78 Freedman DO, Plier DA, de Almeida A et al. Biased TCR reper- 84 Cullen M, Erlich H, Klitz W, Carrington M. Molecular mapping toire in infiltrating lesional T cells in human Bancroftian filari- of a recombination hotspot located in the second intron of the asis. J Immunol 1999; 162: 1756–1764. human TAP2 locus. Am J Hum Genet 1995; 56: 1350–1358.

Genes and Immunity