DQA1 and DQB1 Promoter diversity and linkage disequilibrium with class II haplotypes in Mexican Mestizo population

(2001) 2, 216 221 2001 Nature Publishing Group All rights reserved 1466-4879/01 $15.00 www.nature.com/gene DQA1 and DQB1 Promoter diversity and linkage disequilibrium with class II haplotypes in Mexican Mestizo population C Alaez, MN Vázquez-García and C Gorodezky Department of Immunogenetics. Instituto de Diagnóstico y Referencia Epidemiológicos, InDRE, SSA, Mexico The upstream sequences in the 5 flanking region of HLA class II genes, regulate their expression and contribute to the development of immunological diseases. We analyzed 105 healthy unrelated Mexican Mestizos for QAP and QBP polymorphism. DNA typing for DRB1, DQA1, DQB1, QAP1 and QBP1 was done using a standardized PCR-SSOP. Although all QAP alleles previously described were found in Mexicans, the distribution differed as compared to other populations. QAP-3.1, 4.1 and 4.2 were the most frequent alleles and were associated with DQA1*03, *0501 and *0402 respectively. The prevalent QBP alleles were 3.21, 3.1 and 4.1 found mainly associated with DQB1*0302, *0301 and *0501. Linkage disequilibria between the promoter and the corresponding DQA1 and DQB1 allele, are in general the same as described by others. A total of 61 different haplotypes were defined, only six of them with a frequency above 4%. The haplotypes DRB1*0407-QAP-3.1-DQA1*03-QBP-3.21-DQB1*0302 (HF = 14.37%) and DRB1*0802-QAP-4.2- DQA1*0401-QBP-4.1-DQB1*0402 (HF = 14.22%), which have an Amerindian ancestry, are the most frequent in Mexicans. Some rare combinations were detected such as DRB1*0405-QAP-1.3-DQA1*0101/4-QBP-5.11/5.12-DQB1*0501 and DRB1*0403-QAP-3.2-DQA1*03 QBP-3.21-DQB1*0302, probably due to ancient recombination events. This knowledge is relevant as a basis to evaluate functional implications and to explore the role of promoter diversity in disease expression. (2001) 2, 216 221. Keywords: class II promoters; DQA/DQB promoters; polymorphism of URR; QAP/QBP diversity Introduction The major histocompatibility complex (MHC) class II genes encode cell-surface glycoproteins essential for the development of the T cell repertoire at the thymus and for the presentation of processed antigen to T cells. 1 Thus, modulation of T cell activation depends on the strength of signal delivery coming from MHC class II proteins and from costimulator molecules. One of the factors influencing this process is the different density of class II molecules on the cell surface. Generally, HLA-DR molecules are strongly expressed followed by DQ and DP proteins; their different level of expression is probably determined by variations in the class II regulatory complex, located in the 5 upstream regulatory regions of class II genes. 2 These regions contain cis-acting sequences, such as the CCAAT, Y, X, W, Z boxes which have promoter or enhancer functions. 3 These sequences bind nuclear transcription proteins involved in the constitutive and/or Correspondence: Dr Clara Gorodezky, Head of the Department of Immunogenetics, Instituto de Diagnóstico y Referencia Epidemiológicos, InDRE, SSA, Carpio 470 1st floor, México. D.F. 11340. Mexico E-mail: cgorodea mailer.main.conacyt.mx The two first authors contributed equally to this work. This work was partly supported by CONACYT fellowships No 83542 for C. Alaez as PhD student of the Program of Biological and Chemical Sciences ENCB-IPN and No 115763 for MN Vázquez- García, student of the Biomedicine Molecular Program, CINVES- TAV-IPN. Received 9 February 2001; revised and accepted 16 April 2001 inducible expression as well as in the specific tissue distribution displayed by class II genes. A different level of polymorphism has been identified in the cis-acting sequences or in the nucleotide spacer between them. The changes may affect both, protein interaction and the degree of class II expression. 4 The analysis of the variability in the DQA1 URR sequence has revealed that polymorphism is concentrated in the region located between 280 and 240 base pairs upstream of exon 1. Ten alleles have been found named QAP: 1.1, 1.2, 1.3, 1.4, 1.5, 2.1, 3.1, 3.2, 4.1, 4.2, which show mutations in the X, Y and W boxes. 5 Ten alleles have also been described for URR DQB1 and have been called QBP: 2.1, 3.1, 3.21, 3.31, 4.1, 5.11, 5.12, 6.11, 6.2, 6.3. The mutation sites in these alleles are localized in the X 1, X 2 and W boxes. No variability has been detected for the Y box. 6,7 Population studies of DQA1 and DQB1 promoters diversity have shown that one DQA1 allele is found in linkage disequilibrium with different promoter genes within certain groups of DQA1 alleles. 8 However, for QBP1 polymorphism a tight linkage between the promoter region and exon two of DQB1 locus was demonstrated with the exception of what has been found for DQ5 and DQ6 haplotypes. For the latter, single DQB1 alleles are linked to different but closely related QBP1 alleles and vice versa. 9 The majority of these findings were originally described in Caucasian populations such as Germans and in certain Mediterranean groups, namely Italians. 8,10 The

exploration of QAP and QBP distribution in different ethnic groups, other than whites, will most probably show unknown polymorphisms or new associations between the promoter alleles and the class II haplotypes. This knowledge may be helpful in understanding the biological significance of haplotype conservation in those class II associated diseases in which DQ alleles play an important role such as type I diabetes (IDDM), celiac disease and others. In fact, some studies looking for QAP or QBP involvement have been carried out trying to further define disease susceptibility in IDDM 11,12 and in systemic lupus erythematosus 13 with negative results. The aim of this study was therefore to analyze the QAP and QBP promoter distribution in Mexican Mestizos and to determine their linkage with the different DRB1- DQA1-DQB1 haplotypes present in this population. Results QAP and QBP alleles were assigned in 105 apparently healthy Mexican individuals. Their distribution in Mexicans is shown in Table 1. All QAP alleles described previously were found in this population, but the most frequent ones were: QAP1 3.1 (AF = 31%), QAP1 4.1 (AF = 27.1%) and QAP1 4.2 (AF = 15.9%) which were always associated with DQA1*03, DQA1*0501 and DQA1*0402 respectively. Linkage disequilibria of QAP variants with certain DQA1 alleles was very strong and a complete association was found for QAP-2.1-DQA1*0201 (HF = 78/1000), QAP-3.1-DQA1*03 (HF = 345/1000, QAP- 4.1-DQA1*0501 (HF = 297/1000) and QAP-4.2-DQA1* 0401 (HF = 177/1000). However some others, such as QAP-1.1, 1.2, and 1.3 were combined with different DQA1*01 alleles (Table 2). The QBP distribution (Table 1) also shows that all described QBP alleles are present in Mexicans, with QBP- 3.21 (AF = 30.3%), QBP-3.1 (AF = 23.4%) and QBP-4.1 (AF = 15.9%) being the three most frequent ones. These were mainly associated with DQB1*0302, *0301 and *0501 respectively. In some individuals, it was impossible to distinguish between QPB-5.11 and 5.12 due to the particular combination of the promoters found in the subjects. DQB1 alleles were not as strongly associated with the QBP variants, as observed for QAP-DQA1 haplotypes. Several QBP-DQB1 combinations were very frequent: 3.21-*0302 (HF = 324/1000), 3.1-*0301 (HF = 235/1000); 4.1-*0402 (HF = 179/1000); 2.1-*0201 (HF = 121/1000) Table 1 Distribution of DQA1 and DQB1 promoter alleles in Mexican Mestizos QAP1 alleles n = 105 AF a % QBP1 alleles n = 105 AF a % 1.1 15 7.4 2.1 21 10.8 1.2 8 3.9 3.1 43 23.5 1.3 11 5.3 3.21 54 30.8 1.4 2 1 3.31 2 1 1.5 2 1 4.1 31 16.2 2.1 13 6.4 5.11 6 2.9 3.1 55 31 5.12 5 2.4 3.2 2 1 5.11 or 5.12 7 3.4 4.1 49 27 6.11 1 0.5 4.2 31 16 6.2 13 6.5 6.3 4 2 a AF, allele frequency. while others were found in a low proportion (HF between 5/1000 and 29/1000) (Table 2). The most probable haplotype combinations were determined in each case and the frequencies were calculated. The combinations of DRB1-QAP1-DQA1-QBP1- DQB1 were inferred from the data shown in the tables and from the class II most probable haplotypes found in Mexicans. These data are shown in Table 3. We were able to define a total of 61 different haplotypes, but only six of them had a frequency over 4%,: DRB1*0802-QAP4.2- DQA1*0401-QBP-4.1-DQB1*0402 (HF = 14.22%), DRB1* 0407-QAP-3.1-DQA1*03-QBP-3.21-DQB1*0302 (HF = 14.4%), DRB1*1602-QAP-4.1-DQA1*0501-QBP-3.1-DQB1* 0301 (HF = 5.89%), DRB1*0701-QAP-2.1-DQA1*0201- QBP-2.1-DQB1-*0201 (HF = 5.89%), DRB1*1406-QAP-4.1- DQA1*0501-QBP-3.1-DQB1*0301 (HF = 5.38%), DRB1*1402- QAP-4.1-DQA1*0501-QBP-3.1-DQB1*0301 (HF = 4.88%). The rest of the haplotypes were present in lower frequencies (0.48 3.39%) and they represent 50.33% of all haplotypes found. Discussion The distribution of QAP1 and QBP1 alleles and their linkage disequilibrium with DQA1 and DQB1 genes has been analyzed for the first time in the Mexican population. The history of admixture goes back to the landing of Cortés and this process has been the vertebral axis of the history of Mexicans during the last 500 years. It has been a complex phenomenon that gave rise to the Mestizos, a term coined by the Spaniards after the conquest; it remains as the anthropological definition of the admixture between Indian and Caucasian genes. 14 Part of the results shown here were included in the 12th International Histocompatibility Workshop (IHW) Joint Report. 5,7 Even if all QAP alleles described were found in the 105 individuals, the most frequent allele QAP1 3.1 was always associated with DQA1*03, which is the prevalent DQA1 allele in Mexicans and in Mestizos of all Latin American countries. 15 The high frequency of QAP1 3.1-DQA1*03 is the result of the presence of DR4 haplotypes that are very common here and in all populations with an Amerindian ancestry. Moreover, the allele frequency of DR4 haplotypes present in the different Indian groups from North, Central and South America, oscilates between 10% observed in Chiriguanos from Argentina, to 75% in the Yukpa from Venezuela. 15 Interestingly, only two DR4 haplotypes in Mexicans had a different QAP-DQA1 association: one of them was DRB1* 0405-QAP-1.3-DQA1*0101/4-QBP-5.11/5.12-DQB1*0501. It may have been the result of a recombination event between DRB1 and QAP1 loci. The other rare combination found was DRB1*0403-QAP-3.2-DQA1*03 QBP- 3.21-DQB1*0302 (Table 3), which was detected in only one individual; this haplotype has been found in a very low frequency in other groups as well, such as Germans; 8 therefore it is most likely of Caucasian origin. In Germans it is part of DRB1*0901-QAP-3.2-DQA1*0301-DQB1*0303 and does not appear with DR4 alleles. Complete linkage disequilibrium between QAP1 and DQA1 alleles was detected in combinations such as QAP- 2.1-DQA1*0201, QAP-4.1-DQA1*0501 and QAP-4.2- DQA1*0401. However as shown in other groups, DQA1* 01 alleles may be associated with QAP1 1.1, 1.2, 1.3, 1.4 and 1.5 8 (Table 2). The same variability was also found 217

218 Table 2 QAP1-DQA1 and QBP1-DQB1 haplotype frequencies in Mexican Mestizos (n = 105) QAP1 Alleles DQA1 Alleles HF a X1000 P Fisher QBP1 Alleles DQB1 Alleles HF a X1000 P Fisher 1.1 *0101/4 35 2.04E-03 2.1 *0201 121 8.84E-19 1.1 *0102 33 1.17E-02 3.1 *0301 235 1.27E-20 1.2 *0102 41 1.64E-05 3.21 *0302 324 5.37E-21 1.2 *03 20 NS 3.31 *0303 6 2.30E-02 1.3 *0103 23 1.33E-03 4.1 *0402 179 2.96E-22 1.3 *0101/4 22 2.69E-02 5.11 *0502 17 1.49E-04 1.3 *0102 13 NS 5.11 *0301 11 NS 1.4 *0101/4 5 2.45E-01 5.12 *0501 24 9.43E-05 1.4 *0103 6 1.33E-01 5.11/5.12 *0501 29 1.26E-04 1.5 *0101/4 1 1.85E-02 6.2 *0602 58 7.56E-10 1.5 *03 12 NS 6.2 *0604 10 9.46E-02 2.1 *0201 78 1.35E-15 6.3 *0603 11 9.17E-03 3.1 *03 345 2.30E-23 3.2 *03 1 3.16E-01 2.1 *0201 78 1.35E-15 4.1 *0501 297 1.33E-23 4.2 *0401 177 1.27E-21 a HF, haplotype frequency. Table 3 Distribution of the most probable haplotypes in Mexican Mestizos DRB1-QAP-DQA1-QBP-DQB1 HF a % DR1 *0101-1.1-*0101/4-5.11-*0501 2 0.96 *0101-1.1-*0101/4-5.12-*0501 2 0.96 *0102-1.1-*0101/4-5.11-*0501 1 0.48 *0102-1.1-*0101/4-5.12-*0501 3 1.44 *0102-1.1-*0101/4-511512-*0501 1 0.48 *0102-1.5-*0101/4-511512-*0501 2 0.96 DR2 *1501-1.1-*0102-6.2-*0602 3 1.48 *1501-1.2*0102-6.2-*0602 4 1.92 *1501-1.3-*0101/4-5.11-ND 1 0.48 *1501-1.3-*0102-6.2-*0602 1 0.48 *1501-1.3-*0103-ND-*0603 1 0.48 *1502-1.3-*0103-6.11-*0601 1 0.48 *1503-1.2-*0102-6.2-*0602 1 0.48 *1601-1.1-*0102-5.11-*0502 1 0.48 *1601-1.2-*0102-511/512-*0502 1 0.48 *1602-4.1-*0501-3.1-*0301 12 5.89 DR3 *0301-4.1-*0501-2.1-*0201 8 3.88 DR4 *0401-3.1-*03-3.21-*0302 1 0.48 *0402-3.1-*03-3.21-*0302 3 1.44 *0403-3.1-*03-3.1-*0301 1 0.48 *0403-3.1-*03-3.21-*0302 4 1.92 *0403-3.2-*03-3.21-*0302 1 0.48 *0404-3.1-*03-3.21-*0302 7 3.39 *0405-3.1-*0101/4-511512-*0501 1 0.48 *0405-3.1-*03-3.21-*0302 6 2.90 *0407-3.1-*03-2.1-*0201 1 0.48 *1407-3.1-*03-3.21-*0301 1 0.48 *0407-3.1-*03-3.21-*0302 28 14.40 *0411-3.1-*03-3.21-*0302 7 3.39 DR5 *11-1.1-*0102-5.11-*0502 1 0.48 Continued Table 3 Continued DRB1-QAP-DQA1-QBP-DQB1 HF a % *1101-1.1-*0102-6263-*0602 1 0.48 *1101-4.1-*0501-3.1-*0301 4 1.92 *1104-4.1-*0501-3.1-*0301 4 1.92 *1201-4.1-*0501-3.1-*0301 2 0.96 DR6 *06-4.1-*0501-3.1-*0301 1 0.48 *1301-1.1-*0103-6.3-*0603 1 0.48 *1301-1.3-*0102-ND-*0604 1 0.48 *1301-1.3-*0102-6.3-*0604 1 0.48 *1301-1.3-*0103-511/512-*0501 1 0.48 *1301-1.3-*0103-6.2-*0603 1 0.48 *1301-1.3-*0103-6.3-*0603 1 0.48 *1301-1.4-*0103-6.3-*0602 1 0.48 *1301-4.1-*0501-3.1-*0301 1 0.48 *1302-1.2-*0102-6.2-*0604 1 0.48 *1302-1.4-*0101/4-6.2-*0604 1 0.48 *1303-4.1-*0501-3.1-*0301 2 0.96 *1304-4.1-*0501-3.1-*0301 1 0.48 *1401-1.2-*0101/4-5.11-*05031 1 0.48 *1401-1.3-*0101/4-511/512-*05031 1 0.48 *1401-1.3-*0102/5.11-*0502 1 0.48 *1402-4.1-*0501-3.1-*0301 10 4.88 *1402-4.1-*0501-3.21-*0301 1 0.48 *1406-4.1-*0501-3.1-*0301 11 5.38 DR7 *1701-2.1-*0201-2.1-*0201 12 5.89 *0701-2.1-*0201-3.1-*0201 1 0.48 DR8 *0801-4.2-*0401-4.1-*0402 3 1.44 *1802-4.2-*0401-4.1-*0402 28 14.22 DR9 *1901-3.2-*03-3.31-*0303 1 0.48 DR10 *1001-1.1-*0101/4-5.12-*0501 1 0.48 a HF, haplotype frequency.

in the QAP alleles associated with the DQ6 group, as demonstrated in Germans. The explanation claimed by the authors 8 for this finding is that all DQ6 associated QAP alleles have evolved by point mutations from a single precursor QAP sequence. We did not find the strong association between DRB1*1501 and QAP-1.2 demonstrated in Germans, suggesting that a recombination event occurred in a limited stretch of DNA between these two genes in the DR2 haplotypes. Nine out of 11 DR1 carriers showed the combination QAP-1.1-DQA1*0101/4, suggesting a stronger linkage between these two genes (QAP and DQA1) than between QAP and DRB1 locus, since different DRB1*01 alleles were present in association with the same QAP allele. None of the QAP-DQA1 combinations found in this study differed from those reported by the 12th IHW in Han Chinese, Dai minority of Southern China, South American or Seri Indians (from Mexico), Caucasians and African-Americans. 5 The association analysis between DQB1 and the corresponding promoters (Table 2) demonstrated an almost absolute linkage disequilibrium between the promoter and the structural gene. These data are in accordance with the hypothesis of rare events of recombination occurring between DQB1 and QBP. 9 Several exceptions were however detected. As an example: DQB1*0501 was associated with QBP1 5.11 or 5.12. We were not able to assess the precise frequency for each combination because when QBP-4.1 or QBP-3.21 were present in one chromosome and 5.11 or 5.12 were detected in the other one, the discrimination was only possible with probe 1941. We used different conditions to achieve the greatest stringency to get a specific hybridization signal for this probe, but it was not possible. The reason is that QBP1 5.11 and 5.12 differ from each other only at position 181 to 191 bp and no alternative position is available for discrimination. The same cross-hybridization problem found here has been reported previously by others. 9 Another example of DQB1 and QBP association, which is not absolute, was the DQ6 group; QBP-6.2 was found more frequently in combination with DQB1*0602, but also with DQB1*0604 in some haplotypes. DRB1*1301- QAP1 1.3-DQA1*0103-QBP1 6.2-DQB1*0603; DRB1*1302- QAP1 1.2-DQA1*0102-QBP1 6.2-DQB1*0604; DRB1*1302- QAP-1.4-DQA1*0101/4-QBP1 6.2-DQB1*0604. However, the variability of QBP-DQB1 associations observed herein, was low compared with the QAP-DQA1 group. Of the 61 different haplotypes found in Mexicans, the six prevalent ones represent 50.63% of all the detected combinations. Five of them are of Amerindian origin: DRB1*0802-QAP-4.2-DQA1*0401-QBP-4.1-DQB1*0402 (HF = 14.22%), DRB1*0407-QAP-3.1-DQA1*03-QBP -3.21- DQB1*0302 (HF = 14.4%), DRB1*1602-QAP-4.1-DQA1* 0501-QBP-3.1-DQB1*0301 (HF = 5.89%), DRB1*1406-QAP- 4.1-DQA1*0501-QBP-3.1-DQB1*0301 (HF = 5.38%), DRB1 *1402-QAP-4.1-DQA1*0501-QBP-3.1-DQB1*0301 (HF = 4.88). These combinations are strongly represented in most American Indian groups. For example, the frequency of DRB1*0802 haplotype varies from 10% in Yukpa from Venezuela to 52.9% in Kaingag from Brazil. In Mexican Lacandon from Chiapas 16 this haplotype is absent, but in Seris from northeast Mexico, the frequency is 36.7%. 15 The same is true for the other combinations mentioned above. On the contrary, the haplotype frequency of this DR8 haplotype in Caucasians from Europe is only 2.6/10,000; of the DRB1*1602 is 0.3/10,000; *1402 and * 1406 combinations are poorly represented: 1.4 and 0.9/10,000 17 respectively. These data indicate the strong influence of Asian background to the genetics of the ancestral groups of Mexican Mestizos. DRB1*0701-QAP-2.1-DQA1*0201-QBP-2.1-DQB1*0201 (HF = 5.89%), DRB1*0301-QAP-4.1-DQA1 *0501-QBP-2.1- DQB1*0201 (HF = 3.88%), and DRB1*0402-QAP-3.1- DQA1*03-QBP-3.21-DQB1*0302 (HF = 1.44%) are mainly of Mediterranean background. 18 DRB1*0701 haplotype has a frequency of 768/10,000 in Caucasians from Europe and DRB1*0301 of 1280/10,000, while both are absent in Mexican Indians. 15 17 As shown, the contribution of Indian markers to the total gene pool of Mestizos, is clearly stronger than the one coming from white genes. The existence of haplotypes with identical DRB1- DQA1-DQB1 structural genes, differing only in the polymorphism of QAP and QBP regions, suggests that probably the same HLA molecule may have a different rate of expression in different individuals depending on the promoter variant which controls the synthesis. For this reason it will be important to perform functional studies to identify the importance of this fact in transplantation and in susceptibility to autoimmune or infectious diseases. There are some evidences suggesting that MHC class II haplotype-specific expression may affect cytokine levels in the mouse 19 and disease resistance in chickens. 20 Undoubtedly, this variability contributes to increase population diversity. Some groups have shown that allelic diversity in regulatory regions affect the rate of expression of a given DQA1 variant. Indeed, DQA1*0401, *0501 and *0601 are less expressed at the RNAm level than DQA1*0201. QAP1 4.1 or 4.2, that regulate DQA1*0401, *0501, *0601, have a substitution in the Y box that affect transcription, while QAP1 2.1 which controls DQA1*0201 has the consensus sequence in the Y box. 21 Similar results were shown for several DQB1 promoter alleles in which the X box polymorphism has been associated with differences in protein binding pattern and transcriptional activity. 22 For example, the presence of a dinucleotide TG at position 180/ 179 in DQB1*0302 correlates with a lower level of expression compared to the one of DQB1*0301. The functional significance of such regulatory polymorphism has been confirmed by quantitative RT-PCR of the allele specific mrnas, demonstrating that the interferon- induced transcription of *0301 occurs earlier and at higher levels than the transcription of DQB1*0302. 22 If a difference in allele expression exists, it is likely that the amount of peptides presented by the class II molecules will differ as well. The implications of this hypothesis in HLA linked susceptibility may be of great relevance for disease outcome. One of the factors influencing the shift of an immune response towards Th1 or Th2 phenotypes is the level in which an MHC-peptide complex is expressed on the surface of an antigen presenting cell. 2 In this regard, some studies have shown that the modulation of HLA DRB1 by INF- is transcriptionally regulated, and the variation in promoter sequence might contribute to such regulation at the promoter level. 23 Thus, studies analyzing diversity and function in different ethnic groups are needed to clarify the role of promoters variability in disease pathogenesis. 219

220 Materials and methods Population A random group of 105 apparently healthy Mexican Mestizos previously typed for DRB1, DQA1 and DQB1 genes according to the protocols recommended by the 12th IHW were included in this study. The subjects were adults born and living in the country, most of them from the central part and the highlands of Mexico; a small proportion were born in the north, along the coast and the southeast. The Mexican Mestizos are an admixed group, with a genetic background composed by Caucasian genes coming from the Spaniards, that arrived during the 16th century and Indian genes derived from the inhabiting local native tribes. Later, the African slaves brought by the Spanish conquerors introduced some black component. 24 DNA extraction Twenty ml of peripheral EDTA blood was drawn from every subject included in the study. Genomic DNA was obtained using the conventional proteinase K phenol/ chloroform extraction method. 25 Oligotyping of QAP and QBP loci Typing of the DQA1/DQB1 promoter alleles was performed using the primers and probes designed for the 12th IHW. 5,6 Protocols were provided to all of us, participants in the promoter component. The following primers were used for the amplification of DQA1 promoter: QAP- Amp5/1 and QAP-Amp3-N. Sixteen oligonucleotide probes were utilized for the hybridization step. QBP amplification was done with the primers: QBP1-Amp5/2 and QBP1-Amp3. To define the different QBP alleles, 25 probes were included in the study. To differentiate QBP1 6.2 alleles from the 6.3 subtype, an additional PCR reaction that identified the presence of the 6.3 allele was required. A set of amplified DNAs from DR/DQ homozygous cell lines derived from the 10th IHW were used as positive controls for each promoter allele. 5 The quality of the PCR products was assessed by electrophoresis. For the dot blot analysis, 5 ul of the PCR product was spotted onto positively charged nylon membranes and was covalently fixed with UV light. The hybridization was carried out with the corresponding oligonucleotide probes, labeled with Dig-ddUTP. The membranes were washed with 2 SSPE and 5 SSPE at the Tm of each probe; after this stringent washing, membranes were incubated with the Fab fragment of anti-dig antibody (Boheringer-Mannheim No. 1363905, Mannheim, Germany) for 30 min. The hybridization was visualized by means of chemiluminiscence detection using CSPD as the substrate (Boheringer-Mannheim No. 1655884). The membranes were exposed on an X-ray film for 5 15 min. The hybridization patterns were captured and the results were analyzed using the data analysis program developed by G. Brunnler for the 12 IWS. 5 Statistical analysis Gene frequencies (AF) were assessed according to Haldane. 26 QAP1-DQA1 and QBP1-DQB1 linkage disequilibrium ( ) and haplotype frequencies were calculated according to Mattiuz et al 27 using a computer program adapted in our lab by E. Infante. Since the program calculates only allele combinations of two loci, the five-point haplotype DRB1-QAP1-DQA1-QBP1-DQB1 were combined logically, using the above information. When only one common haplotype was present, the other one was assigned by subtraction according to the frequencies and linkage disequilibria found in the population. Acknowledgments We are truly grateful to E. Infante for his help with the computer analysis. References 1 Rammensee HG, Falk K, Rotzchke O. MHC molecules as peptide receptors. Curr Opin Immunol 1993; 5: 35 44. 2 Guardiola J, Maffei A, Lauster R, Mitchison NA, Acolla RS, Sartoris S. Functional significance of polymorphism among MHC class II gene promoters. Tissue Antigens 1996; 48: 615 625. 3 Glimcher LH, Kara CJ. Sequences and factors: a guide to MHC class II transcription. Ann Rev Immunol 1992; 10: 13 49. 4 Shewey LM, Beaty JS, Andersen LC, Nepom GT. Differential expression of related HLA class II DQB genes caused by nucleotide variation in transcriptional regulatory elements. J Immunol 1992; 148: 1265 1273. 5 Brunnler G, Haas JP, Fan LA, et al. DQA1 promoter polymorphism: 12th International Histocompatibility Workshop study. In: Charron D (ed). Genetic Diversity of HLA: Functional and medical implications. Vol. I. EDK: Paris, 1997, pp 171 175. 6 Reichstetter S, Krellner PH, Meenzen CM, Kalden JR, Wassmuth R. Comparative analysis of sequence variability in the upstream regulatory region of the HLA-DQB1 gene. Immunogenetics 1994; 39: 207 212. 7 Reichstetter S, Adorno D, Alaez C et al. DQB1 promoter polymorphism: 12th International Histocompatibility Workshop study. In: Charron D (ed). Genetic Diversity of HLA: Functional and Medical implications. Vol. I. EDK: Paris, 1997, pp 176 183. 8 Haas JP, Kimura A, Andreas A et al. Polymorphism in the upstream regulatory region of DQA1 genes and DRB1, QAP, DQA1, DQB1 haplotypes in the German population. Hum Immunol 1994; 39: 31 40. 9 Reichstetter S, Brunnler G, Meenzen CM, Kalden JR, Wassmuth R. DQB1 promoter sequence variability and linkage in Caucasoids. Hum Immunol 1996; 51: 73 80. 10 Petronzelli F, Kimura A, Ferrante P, Mazzilli Mc. Polymorphism in the upstream regulatory region of DQA1 gene in the Italian population. Tissue Antigens 1995; 45: 258 263. 11 Donner H, Rau H, Braun J, Herwing J, Usadel KH, Bandenhoop K. Highly polymorphic promoter regions of HLA DQA1 and DQB1 genes do not help to further define disease susceptibility in insulin-dependent diabetes mellitus. Tissue Antigens 1997; 50: 642 645. 12 Knight SW, Mijovic C, Barnett AH. HLA-DQB1 upstream regulatory region polymorphism and type I diabetes. Tissue Antigens 1996; 47: 231 236. 13 Yao Z, Kimura A, Hartung K et al. Polymorphism of the DQA1 promoter region (QAP) and DRB1, QAP, DQA1, DQB1 haplotypes in systemic lupus erythematosus. Immunogenetics 1993; 38: 421 429. 14 Serrano Sánchez C. Mestizaje e Historia de la población de México (con un esbozo antropológico de los lacandones de Chiapas). In: Municio MA and García Barreno P (eds). Plimorfismo Génico (HLA) en poblaciones hispanoamericanas. Real Academia de Ciencias Exactas, Física y Naturales: Madrid, 1996, pp 171 194. 15 Petzl-Erler ML, Gorodezky C, Layrisse Z et al. Anthropology report for region Latin America: Amerindian and admixed population. In: Charron D (ed). Genetic diversity of HLA. Functional and Medical implications. Vol. I. EDK: Paris, 1997, pp 337 344. 16 Olivo A, Alaez C, Debaz H et al. The influence of the MHC on

survival of human isolates. In: Charron D (ed). Functional and Medical implications. Vol. I. Paris: EDK, 1997: pp 277 279. 17 Gjertson DW, Lee Su-Hui. HLA-A/B and DRB1/DQB1 allele level haplotype frequencies. In: Gjertson DW, Terasaki P (eds). HLA 1999. Lenexa: Kansas, ASHI 1998, pp 365 450. 18 Appendix B. HLA allele and haplotype frequencies by DNA typing: local assignment In: Charron D (ed). Genetic Diversity of HLA: Functional and Medical implications. Vol. I EDK: Paris, 1997, pp 712 775. 19 Dieli F, Sireci G, Lio D, Bonnano CT, Salermo A. Major histocompatibility complex regulation of interleukin-5 production in the mouse. Eur J Immunol 1993; 23: 2897 2902. 20 Kaufman J, Vainio O, Wiles MV, Wallny HJ. MHC determination of resistance to pathogens. Basel Inst Immunol Ann Rep 1994; 74 75. 21 Indovina P, Megiorni F, Ferrante P, Petronzelli F, Mazzilli MC. Different binding of NF-Y transcriptional factor to DQA1 promoter variants. Hum Immunol 1998; 59: 751 757. 22 Beaty JS, West KA, Nepom GT. Functional effects of a natural polymorphism in the transcriptional regulatory sequence of HLA DQB1. Mol Cel Biol 1995; 15: 4771 4782. 23 Lee JL, Kim YH, Kim JD, Kim SJ, Park JH. Molecular analysis of HLA DR gene expression induced by IFN-gamma in malignant melanoma cell lines. Immunogenetics 1999; 49: 287 294. 24 Gorodezky C. Genetic differences between European and Indians: tissue and blood types. Allergy Proc 1992; 13: 243 250. 25 Bignon JD, Fernandez-Viña MA, Cheneau ML et al. HLA DNA class II typing by PCR-SSOP: 12th International Worshop Experience. In: Charron D (ed). Genetic Diversity of HLA. Functional and Medical Implication. Vol. II. EDK: Paris, 1997, pp 21 25. 26 Haldane JBS. The estimation and significance of the logarithm of a ratio of frequencies. Ann Hum Genet 1956; 20: 309 311. 27 Mattiuz PL, Ihde D, Piazza A, Cepellini, Bodmer WF. New approaches to the population genetic segregation analysis of the HLA system. In: Terasaki P (ed). Histocompatibility Testing. Munksgaard: Copenhagen, 1970, pp 193 205. 221