Polymorphism in the upstream regulatory region of DQA1 gene in the Italian population

Polymorphism in the Upstream Regulatory Region of DQA1 Genes and DRB1, QAP, DQA1, and DQB1 Haplotypes in the German Population

Johannes Peter Haas, Akinori Kimura, Adriane Andreas, Martina Hochberger, Elisabeth Keller, Gfinther Brfinnler, Maria de la Paz Bettinotti, Claudia Nevinny-Stickel, Bernhard Hildebrandt, Gabriele Sierp, Takehiko Sasazuki, and Ekkehard Albert

ABSTRACT: Polymorphism in the URR of the MHC class II DQA1 gene defines ten different alleles named QAP. Oligotyping for the alleles of DRB 1, QAP, DQA 1, and DQB 1 have been performed in 210 unrelated healthy controls from Germany. Moreover, 83 HTCs from the Tenth IHWS have been tested. Four point loci haplotypes (DRB1, QAP, DQA1, and DQB1) have been analyzed in the unrelated healthy population sample. Computer analysis of the linkage disequilibria leads to the conclusion that QAP alleles are in strong linkage disequilibrium with alleles either the DQA 1 or the DRB 1 locus. One typical ("common") haplotype was found to be associated with each DRB1 allele in the majority (86%) of the tested persons. Apart from that, 25 other less frequent ("un-

ABBREVIATIONS HLA human leukocyte antigen HTC homozygous typing cell IHWS International Histocompatibility Workshop MHC major histocompatibility complex PCR polymerase chain reaction

I N T R O D U C T I O N

The expression of major histocompatibility complex (MHC) class II genes is controlled by the upstream regu-

From the Immunogenetics Laboratory (].P.H., A.A., M.H., G.B., M.-P.B., C.N.-S., B.H., E.A.), Ludwig-Maximilians University Chil- dren's Polyclinic, Munich; and IGEL (G.S.), Augsburg, Germany; and the Department of Genetics (A.K., T.S.), Medical lnstitute of Bioregulation, Kyushu University, Fukuoka, Japan.

Address reprint requests to (present address) Dr.J.P. Haas, Universitiits- klinik fiir Kinder und Jugendliche, Loeschgestrafle 15, 91054 Erlangen, Germany.

Received (E) June 1, 1993; accepted August 4, 1993.

usual") haplotypes, with an overall frequency of 14% have been defined. Some of these "unusual" MHC class II haplotypes were found to differ only in the regulatory alleles of DQA1 (QAP alleles) while they are identical for the alleles coding for structural elements (DRB1, DQA1, and DQB1). Most of the "unusual" haplotypes were found to carry HLA-DQ6. Assuming that "unusual" (= rare) haplotypes have arisen from "common" (= fre- quent) haplotypes by point mutation and recombination, we propose the existence of three recombination sites in the MHC DR-DQ region: one between DRB1 and QAP, the second between QAP and DQA1, and the third be- tween DQA1 and DQB1. Human Immunology 39, 31-40 (1994)

QAP RFLP

SSP URR

DQA1 promoter restriction fragment-length

polymorphism sequence-specific primer upstream regulatory region

latory regions (URRs) in the 5'-flanking regions of HLA- D genes [1, 2]. Transcriptional regulation of M H C class II genes corresponds with interactions between cis-act- ing elements, located in the 5'-flanking region (URR) of the coding regions of HLA-D genes and trans-acting factors, which either activate or repress transcription [3, 4]. The cis-acting elements consist of the TATA; CCAAT; Y, X, W, and Z boxes; and others with either p romote r or enhancer functions. The sequences of URR elements are highly conserved within M H C class II

Human Immunology 39, 31-40 (1994) 31 © American Society for Histocompatibility and Immunogenetics, 1994 0198-8859/94/$7.00

32 J.P. Haas et al.

genes. Nevertheless, polymorphism in the URRs of HLA-D genes was found in DRA [5], DRB1 [6, 7], DQA1 [8-10], and DQB1 [11].

In the URR of DQA 1 polymorphism is concentrated in the hypervariable region between 280 and 240 base pairs (bp) upstream of exon 1, defining ten different alleles referred to as QAP (DQA 1 promoter) alleles 1.1, 1.2, 1.3, 1.4, 1.5, 2.1, 3.1, 3.2, 4.1, and 4.2 by Kimura and Sasazuki [9]. These QAP alleles were found to be associated with certain DQA 1 alleles.

MHC class II molecules are only expressed by a lim- ited number of cells; i.e., B lymphocytes and macro- phages, but their expression in other cell types such as fibroblasts and endothelial cells may be stimulated via lymphokines such as INF-5/[ 12]. MHC class-II-positive cells express different HLA-D molecules with certain levels of concentration on their cell surface. Generally, HLA-DR molecules have the strongest expression, fol- lowed by HLA-DP and -DQ molecules [13]. Differences in the URRs of MHC class II genes are supposed to be responsible for these effects; i.e., the lower concentra- tion of DQ molecules on the cell surface might be caused by the Y box in the URR of DQA 1, which differs from the consensus sequence (bp - 123 A instead of G [ 14]) found in all other Y boxes of class II URRs [3]. In addition, two of the QAP alleles, QAP 4.1 and 4.2, have been shown to carry a second substitution in their Y box (bp - 119 A instead of G), which causes a significantly decreased transcription of the DQA1 alleles (DQA 1"0401, *0501, *0601) carrying these QAP alleles [9, 15].

There is a strong linkage disequilibrium between the alleles of the different loci in the HLA-DR, DQ region. Thus, various HLA DR-DQ haplotypes have been ob- served in different populations [16]. It is evident that in spite of the strong linkage disequilibrium, recombina- tions might have happened in the past between DRB 1 and DQA1, as well as between DQA1 and DQB1 [17]. The existence of QAP alleles as a polymorphic marker located between DRB 1 and DQA 1 enables a more de- tailed analysis of HLA-DR/DQ linkage disequilibria and the locations of possible sites of recombination.

MATERIALS A N D M E T H O D S

MHC class II polymorphism was investigated in 22 healthy families (129 individuals) and 210 unrelated healthy individuals from Germany. In addition, 83 ho- mozygous typing cells (HTCs) from the Tenth Interna- tional Histocompatibility Workshop (IHWS) have been tested [18]. Oligotyping was performed for DRB1, QAP, DQA 1, and DQB 1. As cross-reference, we used restriction fragment length polymorphism (RFLP) typ- ings for DRB (Eco RI/Taq I), DQA 1 (Taq I), and DQB 1

(Eco RI/TaqI). Detailed information about RFLPs are given by Bidwell [19], Andreas et al [17], and Haas et al. [20].

The polymerase chain reaction (PCR) for DRB1, DQA1, and DQB1 was done by conventional means using the exon-specific primers from the 1 lth IHWS. Oligolabeling and detection were performed using the DIG-11-ddUTP-AMPPD system described by Nevinny-Stickel et al. [21, 22]. Typing for DRB1 in- cluded two steps. The first was a general typing with 15 oligos, followed by specific amplifications for the sub- groups of DR 1; 2; 4; and 3, 5, 6, 8. For DQA1 19 oligos and for DQB1 20 oligos have been used. All oligos are derived from the l l th IHWS except two new DQA1 oligos created according to the published MHC class II sequences [23] to detect position 46-51 in DQA 1"0201 (DQA 4702: TGG CAA TTG CCT CTG TTC) and DQAI*0401, "0501, and "0601 (DQA 4701: TGG TGT TTG CCT GTT CTC). For the determination of the allele DQAI*0104 [24], sequence-specific primer (SSP) typing was performed as described by Olerup et al. [25]. Primers for SSP typing of DQAI*0104 were kindly provided by Dr. O. Olerup. The QAP alleles were determined with 12 oligos [9]. For the PCR, the 5' primer QAP-amp 5 ( - 3 0 6 to -288 ) and the 3' primer QAP-amp3N ( - 11 to + 12 exon 1) were used, thus amplifying a 318-bp fragment. The PCR was per- formed in 30 cycles with an annealing temperature of 56°C. Oligos and primers for QAP typing are listed in Table 1.

The QAP typing system was established with the HTCs (references listed in Table 3) and the segregation of QAP alleles in 22 families (data not shown). HTC and family data have been excluded from haplotype analysis, which had been done on the basis of the 210 random controls. Allele frequencies and linkage disequilibria for DRB 1, QAP, DQA 1, and DQB 1 in the unrelated con- trol population were estimated by a computer program based on procedures described by Arnold and Albert [26] and Baur and Danilovs [27]. This procedure makes it possible to analyze the allele combinations between the four loci tested by calculating allele combinations of two or three different loci. The pairwise analysis of DQA1 and QAP alleles is presented in Table 2.

The calculation of genotypes on the basis of pheno- types is strongly influenced by the treatment of possible "blanks." In the population tested, an unrelated panel of 210 individuals, we have, with all methods available today apart from sequencing (serology, RFLP, SSO typ- ing, and SSP typing), not found a single DRB 1 "blank" allele. The investigation of the Hardy-Weinberg equi- librium (data not shown) did not reveal any excess of homozygotes, which would be expected if a significant number of "blank" alleles remains undetected. In addi- tion, the deduction of four point haplotypes (see below)

DRB1, QAP, DQA1, and DQB1 MHC Class II Haplotypes 33

TABLE 1 List of primers and oligos for typing of QAP alleles in the URR of DQA1

Name Sequence Twash (°C) Specificity

QAP-AMP5 5' CAG ACA TGC ACA CAC CAG 3' 5' primer U R R - D Q A 1 QAP-AMP3 5' CTT TAG GAT CAT CTT CT'F CCC AA 3' 3' primer U R R - D Q A 1 QAP 01 5' ATT GAG TGT CCT CCC TTT 3' 58 QAP 1.1 HR - 2 7 6 / - 2 5 9 QAP 02 5' ATT TAA TGT CCT CCC TTT 3' 58 QAP 1.2 HR - 2 7 6 / - 259 QAP 03 5' T I T AGT GTC CTC CCT TTA T 3' 60 QAP 1.3 HR - 2 7 5 / - 256 QAP 04 5' TAA AGG GAG GAC ACG AAA 3' 58 QAP 1.4 HR - 2 5 7 / - 2 7 5 QAP 05 5' GAA GAG AGG GAA GAC AC 3' 54 QAP 2.1 HR - 2 5 5 / - 2 7 0 QAP 06 5' TCT TCA GCT AGT AAC TG 3' 56 QAP 3.1 HR - 1 6 5 / - 149 QAP 08 5' GGA GGA CAC GAA ATC GG 3' 56 QAP 4.1 HR - 2 6 3 / - 279 QAP 10 5' GTG TCC TCC CTC TAT TC 3' 56 QAP 4.1 + 4,2 HR - 2 7 1 / - 2 5 5 QAP 12 5' TCC TCC CTC TGT TCA TAG 3' 58 QAP 3.1 HR - 2 6 8 / - 251 QAP 13 5' TCC TGC CTC TGT TCA TAG 3' 58 QAP 3.2 HR - 2 6 8 / - 251 QAP 14 5' TTG TGA CGC AGC CTC CA 3' 56 QAP 4.2 HR - 2 1 1 / - 225 QAP 15 5' ATT CCA ACT GAG TGT CCT 3' 56 QAP 1.5 HR - 2 8 4 / - 266

HR means hypervariable region and gives the positions of nucleotides in DQA 1-URR against which the SSO has been designed.

renders it very unlikely that more than some very occa- sional "blanks" exist in the tested population sample. Thus, we conclude that the low probability for a "blank" allows the assumption that essentially all alleles are ac- counted for.

Linkage disequilibrium was derived in allele combina- tions where a positive d value has been calculated. As our computerized analysis of linkage disequilibria is lim- ited to a maximum of three loci, the four point haplo- types (DRB1, QAP, DQA1, and DQB1; see Table 4) were combined logically. Consequently, there are no d values given for the four point haplotypes resulting from the described procedure. Finally, four point haplotypes have been verified by analyzing the phenotype of each individual tested.

For example, the phenotype of one individual was the following: HLA-DRB 1"0301, *0401/QAP4.1, 3.1/ DQAI*0501, *0301/DQB1*0201, *0302. From the two point haplotypes, it can be seen that the combina- tions QAP4.1-DQA 1 *0501 and QAP 3.1-DQA 1"0301 have strong linkage disequilibria, indicated by a relative d of 1.0. In other words, all QAP4.1 carry DQAI*0501, etc. (see Table 2). Likewise it has been determined that all DRBI*0301 carry QAP4.1, that all DRBI*0301 carry DQAI*0501, and that all DRBI*0301 carry DQBI*0201. The logical combination of this informa- tion leads to the four-point haplotype: DRBI*0301 - QAP4.1 - DQAI*0501 - DQBI*0201. In the same way, the second haplotype could be constructed from pair-wise analysis. Thus, the complete genotype, deduced from phenotype data on the basis of known linkage disequilibria, is DRBI*0301 - QAP4.1 - DQA 1"0501 - DQB 1"0201 and DRB 1"0401 - QAP3.1 - DQAI*0301 - DQBI*0302. The probability of this de- duction being erroneous is less than one in 210 (-- sample size). In cases where only one haplotype can

be deduced unequivocally from the phenotype data the second haplotype can be determined by subtraction. For example, individuum 1 in Table 4 has a phenotype including the haplotype DRB1*1501 - QAP1.2 - DQAI*0102 - DQBI*0602, which can be derived un- equivocally. The remaining alleles of this individuum must therefore be combined on the second haplotype: DRB 1 * 1401 - QAP4.1 - DQA 1 *0501 - DQB 1 *0201. While the combination of DR15 with DQ6 is a "com- mon" haplotype with a frequency of 12%, the second one of DR14 with DQ2 is an "unusual" haplotype found only in one individual. Any other possible genotype of this individual, however, would include two "unusual" haplotypes and therefore would be far less likely. In cases where the typing results were leading to unusual haplotype combinations the complete typing for DRB 1, QAP, DQA1, and DQB1 was repeated and confirmed.

RESULTS

The DNA typing for QAP alleles in 83 homozygous B- lymphoblastoid cell lines, derived from the 10th IHWS is presented in Table 3. The complete DNA typing re- suits of these cell lines, excluding the QAP alleles and the allele DQAI*0104, are published by Kimura et al. [18].

Segregation data from the 22 families tested are not shown. The haplotypes of all 44 parents deduced from the phenotype data on the basis of linkage disequilibria were found to be concordant with those determined by family segregation. Thus, the parents of these families have been included in the population sample, while their children have been excluded. Analyzing the HTC data, as well as those from our German population sample of unrelated controls, the linkage disequilibrium of the QAP alleles with certain DQA1 alleles was proven to be very strong. Complete positive association was found

34 J.P. Haas et al.

TABLE 2 DQA1-QAP haplotypes in unrelated controls 210 (n = 210)

Allele QAPI.1 QAP1.2 QAP1.3 QAPI.4 QAPl .5 QAP2.1 QAP3.1 QAP3.2 QAP4.1 QAP4.2 SUM

DQAI*0101 0 1005 982 ~ 878 b

0.97' 202.9d.~

DQAI*0102 23 - 1 9 5

- 0.90 5.4'

DQAI*0103 0

DQAI*0104 23 - 1 7

- 0 . 4 3 0.2

DQAI*0201 0

DQA 1"0301 0

DQAI*0302

DQA 1"0401

0

0

0

0

23 - 2 2 9

- 0 . 9 1 6.6'

0

0

1164 1026

1.00 214.2 g

0

0

0

DQAI*0501 0

0

0

0

0

1027

23 - 154

- 0 . 8 7 4.0 ~

1687 1314

0.95 156.0g

48 - 109

- 0.69 2.2

0

0

0

0

0

0

0

1758

0

92 - 160

- 0 . 6 4 3.2

753 647

0.82 108.1~

342 296

0.87 49.3 g

0

0

0

0

0

0

1187

0

276 193

0.63 13.1 g

89 55

0.16 2.2

23 8 0.02

- 1.0

0

0

0

0

0

0

388

DQA 1"0601 0

0 0

0

0

0

0

0

0

0 Sum 1187

0

0

0

0

0

1393 1189

1.00 207.1g

0

0

0

0

1393

68 58

1.00 - 1 . 0

0

o

o

68

0

23 - 5 4 0

- 0 . 9 6 19.6 g

0

0

0

0

0

0

2626 1930

1.00 216.4 g

o

2648

0

0 2123

0 890

0 388

0 1164

0 1461

0 0

274 274 265

1.00 - 1 . 0

0 2626

68 68 66

1.00 - 1 . 0

342 0

Thirty iterations have been performed. Delta values have been calculated for 10.000 haplotypes.

a First line: number per 10.000.

b Second line: absolute d value.

c Third line: relative d v.~due.

d Fourth line: chi-squared value, calculating the significance on the basis of the absolute number of haplotypes (significance levels: ' <0.05, / < 0 . 0 l, and g <0.001 ).

for QAP2.1 - D Q A 1 *0201, QAP3.1-DQA 1 *0301, and QAP4.1-DQAI*0501. All other QAP alleles appear with a certain DQA1 allele as a rule, but exceptions were found to occur. The QAP alleles were found to have an association with DQA1 as well as DRB1 alleles (see Table 4).

The haplotype combinations of DRB 1, QAP, DQA 1, and DQB 1 observed in the German population sample

are listed in Table 4. A total number of 55 different four point haplotypes could be defined in the german population sample. Another eight haplotypes were found only in HTCs from the 10th IHWS. Obviously these HTCs are selected and derive from different ethnic groups, such as the haplotype DRBI*0302-QAP4.2- DQAI*0401-DQBI*0402, which is known to occur in blacks [16].

DRB1, QAP, D Q A 1 , and DQB1 M H C Class II Haplotypes 35

TABLE 3 Typing results for DRB1, QAP, DQA1, and DQB1 in the HTCs from the 10th IHWS

ID code Name DRB 1 QAP DQA 1 DQB 1

WS9002 MZ070782 0102 0102 1.1 1.1 0101 WS9003 KAS 116 0101 0101 1.1 1.1 0101 WS 9004 JESTHOM 0101 0101 1.1 1.1 0101 WS9005 H O M 2 0101 0101 1.1 1.1 0101 WS9006 WT 100BIS 0101 0101 1.1 1.1 0101 WS 9007 DEM 0401 1602 3.1 1.2 0301 WS9008 D0208915 1501 1501 1.2 1.2 0102 WS9009 KAS011 1601 1601 1.2 1.2 0102 WS9010 AMAL 1501 1501 1.2 1.2 0102 WS9011 E4181324 1502 1502 1.2 1.2 0103 WS9013 SCHU 1501 1501 1.2 1.2 (1102 WS9014 MGAR 1501 1501 1.2 1.2 0102 WS 9016 RML 1602 1602 4.1 4.1 0501 WS9017 W T 8 1501 1501 1.2 1.2 0102 WS9018 L0081785 0301 0301 4.1 4.1 0501 WS9019 DUCAF 0301 0301 4.1 4.1 0501 WS9020 QBL 0301 0301 4.1 4.1 0501 WS 9021 RSH 0302 0302 4.2 4.2 0401 WS9022 COX 0301 0301 4.1 4.1 0501 WS9023 VAVY 0301 0301 4.1 4.1 0501 WS9025 DEU 0401 0401 3.1 3.1 0301 WS 9026 YAR 0402 0402 3.1 3.1 0301 WS 9028 PE 117 0404 0404 3.1 3.1 0301 WS9029 WT51 0401 0401 3.1 3.1 0301 WS 9030 JHAF 0407 0407 3.1 3.1 0301 WS 9031 BOLETH 0401 0401 3.1 3.1 0301 WS9032 BSM 0401 0401 3.1 3.1 0301 WS9033 BM 14 0401 0401 3.1 3.1 0301 WS9034 SAVC 0401 0401 3.1 3.1 0301 WS9036 SPO010 1101 1101 1.2 1.2 0102 WS9037 SWEIG007 1501 1501 1.2 1.2 0102 WS9038 BM 16 1201 1201 4.1 4.1 0501 WS9039 JVM 1102 1102 4.1 4.1 0501 WS9040 BM 15 1102 1102 4.1 4.1 0501 WS9042 TISI 1103 1103 4.1 4.1 0501 WS9043 BM21 1101 1101 4.1 4.1 0501 WS9045 TUBO 1104 1201 4.1 4.1 0501 WS 9047 PLH 07 07 2.1 2.1 0201 WS 9048 LBUF 07 07 2.1 2.1 0201 WS9050 MOU 07 07 2.1 2.1 0201 WS 9051 PITOUT 0701 0701 2.1 2.1 0201 WS9052 DBB 07 07 2.1 2.1 0201 WS9054 EK 1401 1401 1.3 1.3 0104 WS9055 HO301 1302 1302 1.4 1.4 0102 WS 9056 KOSE 1401 1302 1.3 1.4 0104 WS 9057 TEM 1401 1401 1.3 1.3 0104 WS9058 OMW 1301 1301 1.3 1.3 0103 WS 9059 SLE005 1302 1302 1.4 1.4 0102 WS 9060 CB 6B 1301 1301 1.3 1.3 0103 WS9061 31227ABO 1401 1401 1.3 1.3 0104 WS 9062 WDV 1301 1301 1.3 1.3 0103 WS9063 WT47 1302 1302 1.4 1.4 0102 WS 9064 AMALA 1402 1402 4.1 4.1 0501 WS 9065 HHKB 1301 1301 1.3 1.3 0103 WS 9066 TAB 089 0803 0803 1.4 1.4 0103 WS 9067 BTB 0801 0801 4.2 4.2 0401 WS9068 B M 9 0801 0801 1.3 4.2 0401 WS 9069 MADURA 0801 0801 4.2 4.2 0401 WS 9070 LUY 0803 0803 4.2 4.2 0601 WS 9071 OLGA 0802 0802 4.2 4.2 0401

0101 0501 0501 0101 0501 0501 0101 0501 0501 0101 0501 0501 0101 0501 0501 0102 0302 0502 0102 0602 0603 0102 0502 0502 0102 0602 0602 0103 0601 0601 0102 0602 0602 0102 0602 0602 0501 0301 0301 0102 0602 0602 0501 0201 0201 0501 0201 0201 0501 0201 0201 0401 0402 0402 0501 0201 0201 0501 0201 0201 0301 0301 0301 0301 0302 0302 0301 0302 0302 0301 0302 0302 0301 0301 0301 0301 0302 0302 0301 0302 0302 0301 0302 0302 0301 0302 0302 0102 0502 0502 0102 0602 0602 0501 0301 0301 0501 0301 0301 0501 0301 0301 0501 0301 0301 0501 0301 0301 0501 0301 0301 0201 0201 0201 0201 0201 0201 0201 0201 0201 0201 0201 0201 0201 0303 0303 0104 0503 0503 0102 0605 0605 0102 0503 0604 0104 0503 0503 0103 0603 0603 0102 0604 0604 0103 0603 0603 0104 0503 0503 0103 0603 0603 0102 0604 0604 0501 0301 0301 0103 0603 0603 0103 0601 0601 0401 0402 0402 0401 0402 0402 0401 0402 0402 0601 0301 0301 0401 0402 0402

36 J.P. Haas et al.

TABLE 3 (Continued) ID code Name DRB 1 QAP DQA 1 DQB 1

WS 9072 SPACHEO 0802 0802 4.2 4.2 0401 0401 0402 0402 WS 9075 DKB 0901 0901 3.2 3.2 0302 0302 0303 0303 WS9076 T7526 0901 0901 3.2 3.2 0301 0301 0303 0303 WS9078 PMG075 0102 0102 1.1 1.5 0101 0101 0501 0501 WS9079 LWAGS 0102 0102 1.5 1.5 0101 0101 0501 0501 WS9080 EHM 0101 0101 1.1 1.1 0101 0101 0501 0501 Ws9081 EA 1501 1501 1.2 1.2 0102 0102 0602 0602 WS 9082 HO 104 1501 1501 1.2 1.2 0102 0102 0602 0602 WS9083 LD2B 1501 1501 1.2 1.2 0102 0102 0602 0602 WS9086 L0541265 0301 0301 4.1 4.1 0501 0501 0201 0201 WS9087 STEINLIN 0301 0301 4.1 4.1 0501 0501 0201 0201 WS9088 PF04015 0301 0301 4.1 4.1 0501 0501 020l 0201 WS9089 BOB 1104 1104 4.1 4.1 0501 0501 0301 0301 WS9091 MLF(HCF) 0401 0401 3.1 3.1 0301 0301 0301 0301 WS 9092 BM 92 0404 0404 3.1 3.1 0301 0301 0302 0302 WS9093 BER 07 07 2.1 2.1 0201 0201 0201 0201 WS9094 CF996 07 07 2.1 2.1 0201 0201 020l 0201 WS9095 WIN 07 07 2.1 2.1 0201 0201 0201 0303 WS 9097 EMJ 1302 1302 1.4 1.4 0102 0102 0604 0604 WS9104 DHIF 07 07 2.1 2.1 0201 0201 0201 0201 WS9105 FPAF 1301 1301 1.3 1.3 0103 0103 0603 0603 WS9106 MANIKA 07 07 2.1 2.1 0201 0201 0201 0303 WS9107 LKT3 0405 0405 3.1 3.1 0301 0301 0401 0401

Note that DQAI typing has been corrected in HTCs carrying DQAI*0104. See also reference 18.

As expected from the linkage disequilibria in the HLA-DR/DQ region, one typical haplotype for each DRB 1 allele was found. These haplotypes (n = 30) are the most frequent ones and they will be subsequently named "common haplotypes" with respect to their fre- quency in the German population sample. For most of the DRB 1 alleles one or more unusual combinations or variation of the "common haplotype" was observed (see Table 4). Interestingly, the 30 "common haplotypes" correspond to 86% (n = 361) of all haplotypes (n = 420) tested in our population sample. Consequently, the 25 unusual haplotype combinations amount to 14% (n = 59) of the population sample.

Some DRB1 alleles were found only in association with one haplotype in the tested population sample of 420 independent haplotypes. These were the alleles DRBI*0101, 0103, 1601, 1602, 0301, 0402, 0403, 0407, 1101, 1102, 1103, 1104, 1201, 1303,0801,0802, and 1001. Other alleles of DRB1, namely, the DR6 group and DR15, showed a surprising multiplicity of haplotype variations.

Looking for a simple explanation for this haplotype distribution, all unusual haplotype combinations were investigated for possible mechanisms that would enable us to derive their evolution from a "common haplotype." Considering only the influence of two simple genetic mechanisms, namely recombination and point mutation,

"unusual" haplotypes might derive from "common hap- lotypes." This reasoning does not take into account, for example, a significant contribution of foreign ethnic ad- mixture to the observed haplotype variability. Of course, some of the "unusual" haplotypes could originate from different or even older populations. Thus, the question of the original donor haplotypes remains unclear and the explanations given in the last column of Table 4 are not exclusive. In our opinion they represent only the most likely solution.

The procedure is briefly discussed in the following example. In the German population sample, ten differ- ent haplotype combinations have been found to carry DRB1*1302. The haplotype DRB1*1302-QAP1.4- DQAI*0102-DQBI*0604, which was observed 11 times in the population sample tested and which is well defined by HTC 9063, was assumed to be the "common haplotype." The haplotypes DRB1*1302- QAP1.4#DQAI*0103-DQBI*0603 (n = 2) and DRB1*1302#QAP1.3 - DQAI*0103 - DQBI*0603 (n -- 2) seem both to be the results of recombinations with the DQ alleles of another "common haplotype" DRBl*1301 - QAP1.3 - DQAI*0103 - DQB 1"0603. Interestingly, the point of recombination indicated by a # seems to be located between QAP and the second exon of DQA 1 in the first case. This is indicated by the allele QAP 1.4, which is usually found in association with

DRB1, QAP, DQA1, and DQB1 MHC Class II Haplotypes 37

TABLE 4 Haplotype combinations and frequencies in 210 unrelated healthy individuals from the tested population sample (no. of haplotypes, 420)

D R B I - Q A P - D Q A I - D Q B 1 N o . F r e q u e n c y P o s s i b l e m e c h a n i s m

D R 1

0 1 0 1 -- 1.1 - 0 1 0 1 - 0 5 0 1 35 8 . 9 7

0 1 0 2 - 1.1 - 0 1 0 1 - 0 5 0 1 2 0 . 5 0

0 1 0 2 # 1.5 # 0 1 0 l - 0 5 0 1 -

0 1 0 3 - 1.1 - 0 1 0 1 - 0 5 0 1 3 0 . 7 5

D R 2

1 5 0 1 -- 1.2 - 0 1 0 2 - 0 6 0 2 4 8 12.1

1501 -- 1.2 - 0 1 0 2 # 0 5 0 2 ~ 1 0 . 2 5

1 5 0 1 -- 1.2 - 0 1 0 2 # 0 6 0 3 1 0 . 2 5

1 5 0 1 - 1.2 # 0 1 0 1 ~ # 0 6 0 2 1 0 . 2 5

1 5 0 2 - 1.2 - 0 1 0 2 - 0 6 0 2 1 0 . 2 5

1 5 0 2 - 1.2 # 0 1 0 3 ~ - 0 6 0 V 2 0 . 5 0

1 6 0 1 1.2 - 0 1 0 2 - 0 5 0 2 ~ 6 1 .50

1 6 0 2 - 1.2 - 0 1 0 2 - 0 5 0 2 ~ 1 0 . 2 5

D R 3

0 3 0 1 - 4 . I - 0 5 0 1 - 0 2 0 1 4 0 10 .0

0 3 0 2 - 4 . 2 - 0 4 0 1 ~ - 0 4 0 2 ~ - -

D R 4

0 4 0 1 - 3.1 - 0 3 0 1 - 0 3 0 2 23 5 , 8 7

0 4 0 1 3.1 - 0 3 0 1 # 0 3 0 V 10 2 . 5 5

0 4 0 2 - 3.1 - 0 3 0 1 - 0 3 0 2 4 1 .00

0 4 0 3 - 3 . l - 0 3 0 1 - 0 3 0 2 4 1 .00

0 4 0 4 - 3.1 - 0 3 0 1 - 0 3 0 2 9 2 . 3 0

0 4 0 4 - 3.1 - 0 3 0 1 # 0 4 0 2 1 0 . 2 5

0 4 0 5 3.1 - 0 3 0 1 - 0 3 0 2 1 0 . 2 5

0 4 0 5 - 3.1 - 0 3 0 1 # 0 4 0 1 - -

0 4 0 7 - 3.1 - 0 3 0 1 - 0 3 0 1 a 4 1 . 00

0 4 0 8 3.1 - 0 3 0 1 - 0 3 0 2 1 0 . 2 5

0 4 0 8 - 3.1 - 0 3 0 1 - 0 3 0 1 ~ 1 0 . 2 5

0 4 1 0 3.1 - 0 3 0 1 # 0 4 0 2 1 0 . 2 5

D R 5

1 1 0 1 4. I - 0 5 0 1 - 0 3 0 1 34 8 . 5 4

1 1 0 2 - 4 .1 0 5 0 1 - 0 3 0 I 3 0 . 7 5

1 1 0 3 4 .1 - 0 5 0 1 - 0 3 0 1 2 0 . 5 0

1 1 0 4 - 4 . l 0 5 0 1 - 0 3 0 1 13 3 . 2 7

1 2 0 1 - 4 .1 0 5 0 1 - 0 3 0 1 3 0 . 7 5

D R 6

1 3 0 1 - 1.3 - 0 1 0 3 - 0 6 0 3 31 7 . 7 9

1 3 0 1 - 1.3 0 1 0 3 # 0 6 0 2 ~ 2 0 . 5 0

1 3 0 1 # 1.2 - 0 1 0 2 ~ 0 6 0 2 ~ 1 0 . 2 5

1 3 0 1 - 1.3 # 0 1 0 2 ~ 0 6 0 4 ~ 1 0 . 2 5

1 3 0 1 # 1.4 - 0 1 0 2 ~ 0 6 0 4 ~ 2 0 . 5 0

1 3 0 1 # 1.4 # 0 1 0 3 ~ # 0 6 0 4 ~ 2 0 . 5 0

1 3 0 2 - 1.4 - 0 1 0 2 0 6 0 4 11 2 . 7 6

1 3 0 2 - 1.4 - 0 1 0 2 # 0 6 0 5 2 0 . 5 0

1 3 0 2 - 1.4 - 0 1 0 2 # 0 6 0 3 * 2 0 . 5 0

1 3 0 2 - 1.4 # 0 1 0 3 ~ - 0 6 0 3 ° 2 0 . 5 0

1 3 0 2 # 1.3 - 0 1 0 3 ~ - 0 6 0 3 ° 2 0 . 5 0

1 3 0 2 # 1.3 # 0 1 0 2 - 0 6 0 4 3 0 . 7 5

1 3 0 2 # 1.2 # 0 1 0 2 - 0 6 0 4 3 0 . 7 5

1 3 0 2 # 1.1 # 0 1 0 2 - 0 6 0 4 1 0 . 2 5

1 3 0 2 # 3.1 - 0 3 0 1 a - 0 3 0 2 ~ 1 0 . 2 5

P M : U R R - D Q A 1 - 2 2 8 T / C

R C : D Q w i t h D R B l * 1 6 H T

RC: D Q B w i t h D R B I * 1 3 H T

P M : D Q A I * 0 1 0 2 c o d o n 34 G / C

R C : D Q w i t h a r a r e D R B I * 0 8 0 3 H T

e x c l u d i n g Q A P v a r i a n t

R C : D Q B w i t h D R 1 3 1 " 0 4 0 7 H T

R C : D Q B w i t h D R B I * 0 8 H T

RC: D Q B w i t h D R B I * 0 8 H T + P M :

D Q B I * 0 4 0 2 c o d o n 23 G / T

R C : D Q B w i t h D R B I * 0 4 0 7 H T

RC: D Q B w i t h D R B I * 0 8 H T

R C : D Q B w i t h D R B l * 1 5 H T

R C : D Q w i t h D R B 1 * 1 5 0 1 H T incl .

Q A P v a r i a n t

R C : D Q w i t h D R B I * 1 3 0 2 H T exc l .

Q A P v a r i a n t

R C : D Q w i t h D R B 1 * 1 3 0 2 H T incl .

Q A P i n v a r i a n t

R C : D Q w i t h D R B l * 1 3 0 2 H T exc l .

D Q A 1

P M : D Q B I * 0 6 0 4 c o d o n 14 A / C a n d

c o d o n 30 C / T

R C : D Q B w i t h D R B I * I 3 0 1 H T

R C : D Q w i t h D R B I * 1 3 0 1 H T excl .

Q A P v a r i a n t

R C : D Q w i t h D R B I * 1 3 0 L H T incl .

Q A P i n v a r i a n t

P M : U R R - D Q A 1 - 2 2 5 C / A

P M : U R R - D Q A 1 - 2 2 5 C / A a n d

- 2 2 4 G / A

P M : U R R - D Q A 1 - 2 2 5 C / A a n d

- 2 2 6 T / G

R C : D Q w i t h D R B I * 0 4 H T incl . Q A P

v a r i a n t

38 J.P. Haas et al.

TABLE 4 (Continued)

D R B 1 - Q A P - D Q A 1 - D Q B 1 No. Frequency Possible mechanism

1302 # 4.1 - 0501 ~ - 0301 ~ 1 0 .25

1303 - 4.1 - 0501 - 0301 6 1.51

1401 - 1.3 - 0 1 0 4 0503 13 3.27

1401 # 1.4 # 0104 - 0503 I 0.25

1401 # 1.1 - 0 1 0 4 # 0502 a 1 0.25

1401 # 4.1 - 0501 - 0201 1 0.25

1402 - 4.1 .- 0501 - 0301 - -

DR7

0701 - 2.1 - 0201 - 0201 35 8 .79

0701 - 2.1 - 0201 # 0303 ~ 10 2.55

D R 8

0801 - 4.2 - 0401 - 0402 8 2 .06

0802 - 4.2 - 0401 - 0402 - -

0803 - 4.2 - 0601 - 0301 2 0 .50

0803 # 1.4 # 0103 ~ # 06014 - -

D R 9

0901 - 3.2 - 0301 - 0303 2 0 .50

0901 - 3.2 - 0302 - 0303 - -

D R 1 0

1001 - 1.3 - 0 1 0 4 - 0501 2 0 .50

RC: D Q with D R B I * I I, 12, 1303 H T s

incl. Q A P variant

D R B l * 1 3 0 3 der iv ing f rom D R B l * 1 1

PM: U R R - D Q A I - 2 2 5 A / C

Mul t ip le PMs and RCs o f unknown

origin

RC: D Q with D R B I * 0 3 0 1 H T incl.

Q A P

RC: D Q B with D R B I * 0 9 0 1 H T

PM: D Q A I * 0 4 0 1 codon 25 A / T ÷ RC:

D Q B with D R B 1 * 1 1 , 12, 1303

RC: D Q with a rare D R B l * 1 5 0 2 H T

excl. Q A P variant

PM: D Q A I * 0 3 0 1 codon 7 T / C and

codon G / C

Seven haplotypes, marked with a dash for f requency and number have been found only in H T C s from the 11th IHWS.

#, Presumed point of recombinat ion.

Dif ferences in haplotype composi t ion confirmed with RFLP patterns.

H T , haplotype; PM, point mutat ion (exchanges are indicated with the "original" nucleot ide in the first and the exchange in the second position); and RC, recombinat ion.

DQAI*0102 and not with DQAI*0103. In the second case, the recombination has obviously occurred between the DRB 1 locus and the QAP allele, as indicated by the allele QAP 1.3 in the presence of DQAI*0103. In both cases, oligotyping was repeated for all four loci and the DQA 1 and DQB 1 alleles have been confirmed by RFLP patterns. There is little evidence for point mutations in exon 2 in both cases, because the alleles DQAI*0102 and "0103 differ in three bases located at codons 7, 25, and 41, while DQBI*0603 and ~0604 differ in six bases located at codons 48, 57, 70, 86, 87, and 91. Thus, it would at least require nine point mutations to explain the observed variation if recombination is excluded as the major mechanism in this example.

However, point mutations may be the reason for other haplotypic variations. The haplotype DRBI* 1302#QAP1.2#DQA1*0102-DQB1*0604 (n -- 3) is another variation of the "common haplotype" carrying DRB 1" 1302. Obviously, the DRB 1 allele corresponds with both the DQA 1 and the DQB 1 alleles, while there is a change in the QAP allele. This might be the result of two recombinations on either side of the QAP allele,

but it seems more likely that two base exchanges in the positions -272 (C to A) and -271 (G to A) of the QAP1.4 allele are responsible for the change to the QAP1.2 allele. Individual typings of all persons showing aberrant QAP typings have been confirmed by re- peated typing.

Finally, four haplotypes have been found that are almost certainly the results of recombinations causing unusual DR-DQ combinations (see Table 5). These haplotype variations are even detectable by means of serology. They have been typed as DR15-DQ5, DR14-DQ2, DR1302-DQ8, and DR1302-DQ7. Not unexpected was the fact that the QAP allele segregates with the DQA1 fragment in all cases. The DR15-DQ5 seems to be the result of a DQB1 recombination be- tween the "common haplotypes" of DRB 1" 1501 and "1601. All the others are most likely the results of re- combination separating the DRB1 allele from QAP, DQA1, and DQB1. The haplotype DR14-DQ2 is probably due to a recombination of DRB1*1401 with the haplotype DRBI*0301-QAP4.1, DQAI*0501- DQBI*0201. The DR1302-DQ8 haplotype has a simi-

DRB1, QAP, DQA1, and DQB1 MHC Cl.ass II Haplotypes 39

TABLE 5 Phenotypes of four individuals with new haplotype combinations

Ind. no. H a p l o t y p e D R B * I Q A P D Q A * I D Q B * I Sero logy

Ind. 1 A 1401 # 4.1 - 0501 - 0201 ~ D R 1 4 - D Q 2

B 1501 - 1.2 - 0102 - 0602 D R 1 5 - D Q 6

Ind. 2 A 1501 - 1.2 - 0102 # 0502 a D R 1 5 - D Q 5

B 1501 - 1.2 - 0102 - 0602 D R 1 5 - D Q 6

Ind. 3 A 1302 # 4.1 - 0501 - 0301 D R 1 3 - D Q 7

B 1301 - 1.3 - 0103 - 0603 D R 1 3 - D Q 6

Ind. 4 A 1302 # 3.1 - 0301 a - 0302 a D R 1 3 - D Q 8

B 1301 - 1.3 - 0103 - 0603 D R 1 3 - D Q 6

#, Presumed point of recombination. Differences in haplotype composition confirmed with RFLP patterns.

lar background, combining DRB 1" 1302 with the QAP and DQ alleles of a DRBI*04 haplotype. The last variation, DR1302-DQ7, is unusual because DRB1*1302 is the allele present in this haplotype, and not DRBl*1303, as one might expect. The complete typings of the four individuals are listed in Table 5. Oligotyping results have been repeated and confirmed by RFLP data.

DISCUSSION

Our data confirm the strong linkage disequilibrium ex- isting between the QAP alleles encoded in the URR of DQA1 on one side and DRB1 or DQA1/DQB1 on the other side. However, this linkage disequilibrium, although very strong, is not complete.

In the German population sample, we found the ma- jority of haplotypes (86%) corresponding with one typi- cal combination for each DRB 1 allele. With respect to their frequency, they have been named "common haplo- types." Variations (unusual or rare haplotypes) were only observed in 14% of the haplotypes and they have pre- sumably arisen from "common haplotypes." The surpris- ing heterogeneity of DRB1, QAP, DQA1, and DQB1 haplotypes in the tested population (Table 4) is in some contrast to the situation found in the HTCs from the 10th IHWS (Table 3). Keeping in mind that these HTCs have been largely selected for HTC typing, it is obvious that in the HTCs "common haplotypes" are favored. Thus our study demonstrates the necessity of investiga- tions in sufficiently large population samples for compre- hensive description of the existing haplotypes. In this context, it has to be admitted that there is no family where a new haplotype arisen from D R - D Q recombina- tion has been documented so far. "Unusual" haplotypes must therefore be considered to be the result of recom- bination or point mutation in the past. This assumption is not more unlikely than the commonly accepted tenet that the subtypes of DRBI*04 have arisen by historic mutations.

It appears to be an important finding that in the investigated population sample most of the DRB 1-QAP-DQA1-DQB 1 haplotypes showed no variability at all, while some others were quite heteroge- neous, especially in the URR of DQA 1 where the QAP alleles are encoded. The same QAP allele was found in cis position with different DQA1 alleles (i.e., DR15 haplotypes) as one DQA 1 allele was observed with dif- ferent QAP alleles (i.e., DRBl*1302 haplotypes). This variability is, however, restricted to QAP alleles associ- ated with the DQ6 group. A possible explanation for this finding could be that all DQ6-associated QAP alleles have evolved by point mutations from a single primordial QAP sequence. Alternatively, assuming recombination as the mechanism for haplotype diversification, one would have to postulate that recombinations occur pre- dominantly within the DQ6 group. The fact that a QAP allele is in stronger association with the DRB 1 than with the DQA1 allele in some cases (i.e., DRB1*1501- QAP1.2), suggests that recombination must have oc- curred even in the very limited stretch of DNA between QAP and the second DQA1 exon.

Analyzing haplotype variability including the DQB 1 locus, we observed variations in the combination of DQA1 and DQB1 alleles also in the DQ6 group. This led us to propose three different points of recombination in DQ6 haplotypes: one between DRB1 and QAP, the second one between QAP and the second DQA1 exon and, finally, the third one between DQA1 and DQB1.

An important fact is the existence of haplotypes with identical structural alleles differing in their QAP alleles. The haplotype DRBl*1302-DQA1*0102-DQBl* 0604, for example, was found with the QAP alleles QAPI.1, 1.2, 1.3, and 1.4 (see Table 4). Because we must assume an influence of the polymorphism in the URR ofDQA1 (QAP alleles) on the expression of HLA- DQ molecules, it is likely that structurally identical hap- lotypes differ according to regulatory functions. It is therefore important to include functional characteristics

40 J.P. Haas et al.

of HLA haplotypes, such as the promoter polymor- phisms of MHC class II genes, in transplantation and disease studies.

ACKNOWLEDGMENTS Primers for SSP typing of DQA 1"0104 have kindly been pro- vided by Dr. O. Olerup. This work was supported by SFB 217.

REFERENCES

I. Sillivan KE, Peterlin BM: Transcriptional enhancers in the HLA-DQ subregion. Mol Cell Biol 7:3315, 1987.

2. Lee JS: Regulation of HLA class II gene expression. In Dupont B (ed): Immunbiology of HLA, vol 2. Springer- Verlag, 1989, p 49.

3. Benoist C, Mathis D. Regulation ofmajor histocompatibil- ity complex class II genes: X, Y and other letters of the alphabet. Annu Rev Immunol 8:681, 1990.

4. Glimcher LH, Kara CJ: Sequences and factors: a guide to MHC-class II transcription. Annu Rev Immunol 10:13, 1992.

5. Sherman PA, Basta PV, Ting JPY: Upstream DNA se- quences required for tissue specific expression of the HLA-DR~ gene. Proc Natl Acad Sci USA 84:4254, 1987.

6. Singal DP, Qui X, D'Souza M, Sood SK: Polymorphism in the upstream regulatory regions of HLA-DRB genes. Immunogenetics 37:143, 1993.

7. Perfetto C, Zacheis M, McDaid D, et al.: Polymorphism in the promoter region of HLA-DRB genes. Hum Immunol 36:27, 1993.

8. Del Pozzo G, Perfetto C, Ombra MN, et al.: DNA poly- morphism in the 5'-flanking region of the HLA-DQAI gene. lmmunogenetics 35:176, 1992.

9. Kimura A, Sasazuki T: Polymorphism in the 5'-flanking region of the DQA1 gene and its relation to DR-DQ haplotype. In Tsuji K, Aizawa M, Sasazuki T (eds): vol 2. New York, Oxford University Press, 1993, p 382.

10. Morzycka-Wroblewscka E, Harwood JI, Smith J, et aL: Structure and evolution of the promoter regions of the DQA genes. Immunogenetics 37:364, 1993.

11. Andersen LC, Beaty JS, Nettles JW, et al.: A11elic poly- morphism in transcriptional regulatory regions of HLA- DQB genes. J Exp Med 173:181, 1991.

12. Radka SF, Charron DF, Brodsky FM: Class II molecules of the major histocompatibility complex considered as differentiation markers. Hum Immunol 16:390, 1986.

13. Fakenburg JHF, Fibbe WE, Goselink HM, et al.: Human progenitor cells in long term cultures express HLA-DR antigens and lack HLA-DQ antigens. J Exp Med 162:1359, 1985.

14. Auffray C, Lillie JW, Korman AJ, et al.: Structure and expression of HLA-DQa and -DXoe genes: interallelic

alternate splicing of the HLA-DQc~ gene and functional splicing of the HLA-DXc~ gene using a retroviral vector. Immunogenetics 26:63, 1987.

15. Kimura A, Sasazuki T: HLA-DQA gene is differently regulated from other HLA class II genes. In Yoshida TO, Wilson JM (eds): Molecular Approaches to the Study and Treatment of Human Diseases. New York, Elsevier, 1992, p 97.

16. Stasmy P, Xiaojiang G, Morares ME, et al.: Class II haplo- types defined by oligonucleotide typing in 11 different ethnic groups. In Tsuji K, Aizawa M, Sasazuki T (eds): HLA 1991, vol 2. New York, Oxford University Press, 1993, p 233.

17. Andreas A, Chandanayingyong W, Attatippaholkum W, et al.: Unusual HLA-DR/DQ haplotypes: two different breakpoints in two different DR2-DQw3 haplotypes. Im- munogenetics 30:141, 1989.

18. Kimura A, Dong R-P, Harada H, et al.: DNA typing of HLA class II genes in B-lymphoblastoid cell lines homozy- gous for HLA. Tissue Antigens 40:5, 1992.

19. Bidwell J: DNA-RFLP analysis and genotyping of HLA- DR and DQ antigens. Immunol Today 9:18, 1988.

20. Haas JP, Andreas A, Rutkowsky B, et al: A model for the role of HLA-DQ molecules in the pathogenesis of juvenile chronic arthritis. Rheumatol Int 11:191, 1991.

21. Nevinny-Stickel C, Bettinotti MdlP, Andreas A, et al.: Nonradioactive HLA class II typing using polymerase chain reaction and digoxigenin-11-2'-3'-didesoxyuridi- netriphosphate labeled oligonucleotide probes. Hum Im- munol 31: 7, 1991.

22. Nevinny-Stickel C, Hinzpeter M, Andreas A, Albert ED: Nonradioactive oligotyping for HLA-DR1-DRw I0 using polymerase chain reaction, digoxigenin-labeled oligonu- cleotides and chemiluminiscence detection. Eur J Immu- nogenetics 18:323, 1991.

23. Marsh SGE, Bodmer JG: HLA class lI nucleotide se- quences 1991. Immunogenetics 33:321, 1991.

24. Lee KW, Johnson AH, Hurley CK, et al.: New DQwl diversity identified within DRw12 and DRw14 haplo- types. Tissue Antigens 38:231, 1991.

25. Olerup O, Aldener A, Fogdell A: HLA-DQB1 and -DQA 1 typing by PCR amplification with sequence-spe- cific primers (PCR-SSP) in 2 hours. Tissue Antigens 41:119, 1993.

26. Arnold A, Albert ED: Computerprogramm zur Berech- nung yon Haplotypenfrequenzen des HLA-Systems. Proc 7th ISBN 407, 1978.

27. Baur MP, Danilovs J: Reference tables of two and three locus haplotype frequencies for HLA-A, B, C, DR, Bf and GLO. In Terasaki PI (ed): Histocompatibility Testing 1980. Los Angeles, UCLA TTL, 1980, p 994.