Haplotype structures of the UGT1A gene complex in a Japanese population

Abstract

Genetic polymorphisms of UDP-glucuronosyltransferases (UGTs) are involved in individual and ethnic differences in drug metabolism. To reveal co-occurrence of the UGT1A polymorphisms, we first analyzed haplotype structures of the entire UGT1A gene complex using the polymorphisms from 196 Japanese subjects. Based on strong linkage disequilibrium between UGT1A8 and 1A10, among 1A9, 1A7, and 1A6, and between 1A3 and 1A1, the complex was divided into five blocks, Block 8/10, Block 9/6, Block 4, Block 3/1, and Block C, and the haplotypes for each block were subsequently determined/inferred. Second, using pyrosequencing or direct sequencing, additional 105 subjects were genotyped for 41 functionally tagged polymorphisms. The data from 301 subjects confirmed the robustness of block partitioning, but several linkages among the haplotypes with functional changes were found across the blocks. Thus, important haplotypes and their linkages were identified among the UGT1A gene blocks (and segments), which should be considered in pharmacogenetic studies.

Introduction

Glucuronidation, catalyzed by UDP-glucuronosyltransferases (UGTs), is one of the critical steps in the detoxification and elimination of various endogenous and exogenous compounds.1, 2 As for the genes coding UGTs, two subfamilies, UGT1 and UGT2, have been identified in humans. The human UGT1A gene complex spans approximately 200 kb, is located on chromosome 2q37, and consists of nine active and four inactive exon 1 segments (in the following segment order: UGT1A12P, 1A11P, 1A8, 1A10, 1A13P, 1A9, 1A7, 1A6, 1A5, 1A4, 1A3, 1A2P, and 1A1) and common exons 2–5 (Figure 1). One of the nine active exon-1's (namely, 1A1 and 1A31A10) can be used in conjunction with the common exons.2, 3 The UGT1A N-terminal domains (encoded by the exon-1's) determine the substrate-binding specificity and the C-terminal domain (encoded by exons 2–5) is important for binding to UDP-glucuronic acid.1 Thus, the exon 1 segments confer the substrate specificity of UGT1A isoforms,4 and the 5′-flanking region (and possibly the 3′-flanking region) of each exon 1 is acknowledged to independently regulate the expression of each isoform.3, 4

Figure 1
figure1

The organization of the human UGT1A gene complex and the polymorphisms used for genotyping. The human UGT1A gene complex (nine active and four inactive exon-1's, and common exons 2–5) spans approximately 200 kb. Four pseudogenes (1A2p, 1A11p, 1A12p, and 1A13p), and 1A5, which mRNA expression has not yet been detected in any tissue, were not analyzed in this study and are depicted by the small and gray bars. The complex was divided into five blocks, Block 8/10, Block 9/6, Block 4, Block 3/1, and Block C.

A number of genetic polymorphisms including single nucleotide polymorphisms (SNPs) in UGT1As have been identified and published on the UDP glucuronosyltransferase home page (http://som.flinders.edu.au/FUSA/ClinPharm/UGT/allele_table.html). Some of these polymorphisms are known to affect glucuronidation rates.5, 6, 7, 8, 9, 10, 11, 12, 13, 14 Regarding 1A1, a TATA box variant (−40_−39 insTA: *28 allele), increases the risk of irinotecan-induced toxicity via decrease in detoxicating glucuronidation of 7-ethyl-10-hydroxycamptothecin (SN-38), an active metabolite of irinotecan.5 Another 1A1 polymorphism, 211G>A (G71R: *6 allele), also shows reduced activity to SN-38.7, 15 In addition, it has been reported that the 1A7 alleles, *2 (387T>G (N129K), 391C>A and 392G>A (R131K)), *3 (N129K, R131K, and 622T>C (W208R)) and *4 (W208R) show reduced activities towards benzo(a)pyrene metabolites: for SN-38 glucuronidation, *3 and *4, but not *2, are less active.6, 7

Haplotypes, linked combinations of SNPs on a chromosome, have the advantage of providing more useful information on phenotype–genotype links than individual SNPs.16 Co-occurrence of the SNPs or segmental haplotypes with functional changes in the UGT1A complex could lead to a cooperative alteration in glucuronidation activity. Kohle et al.17 reported close linkages among 1A1*28, 1A6*2 (T181A/R184S), and 1A7*3 (N129K/R131K/W208R) in Caucasians and Egyptians. Moreover, a recent analysis has shown that the low-activity alleles, 1A7*2 and *3, were completely associated with the 1A9 −126_−118 T9 allele, whereas the high-activity 1A7*1 allele was linked with the −126_−118 T10 allele (1A9*22: high expression) in Americans.18 However, there is no haplotype analysis with a high SNP density for the entire UGT1A complex, especially for Asian populations, which includes the Japanese.

Previously, we have reported the segmental haplotype structures for 1A1, 1A4, and 1A6 exon-1's in Japanese subjects.19, 20, 21 In this study, additional first exons (1A3, 1A7, 1A8, 1A9, and 1A10) and their surrounding promoter or intronic regions were sequenced for the same 196 Japanese subjects as used for the analysis of 1A1, 1A4, and 1A6, and the haplotypes for the UGT1A complex were inferred in linkage disequilibrium (LD) blocks, 1A8-1A10, 1A9-1A7-1A6, and 1A3-1A1. Then, the tagged polymorphisms with functional changes were genotyped for additional 105 Japanese subjects. Finally, several linkages among the block haplotypes were inferred in a total of 301 subjects and compared with those of other ethnic groups.

Results

UGT1A8, 1A10, 1A9, 1A7, and 1A3 polymorphisms detected in a Japanese population

All the exon-1's and their flanking regions of UGT1A8, 1A10, 1A9, 1A7, and 1A3 were sequenced in 196 Japanese subjects (108 arrhythmic patients and 88 cancer patients). As for 1A6, 1A4, 1A1, and common exons 2–5, their SNPs and segmental haplotypes have already been reported.19, 20, 21 UGT1A5 was omitted from the current analysis because the expression of 1A5 mRNA has not been shown in any tissue.2 AF297093.1 (GenBank accession number) was used as the UGT1 reference sequence. All the allele frequencies were in Hardy–Weinberg equilibrium. No statistically significant differences in allelic frequencies of the detected SNPs were found between the subjects with the different disease types (P0.05 by χ2 test or Fisher's exact test). Thus, the data for all subjects were analyzed as one group.

In 1A8, seven SNPs, including two novel synonymous ones, were detected (Figure 2). The known nonsynonymous SNP, 518C>G (A173G, *2 allele), was found at a frequency of 0.594. As for 1A10, eight SNPs were detected. Among them, two polymorphisms, 4G>A (A2T) and 200A>G (E67G), were novel (Figure 2). Previously reported SNPs 177G>A (M59I, *2 allele) and 605C>T (T202I, *3 allele) were both found at a 0.010 frequency. Seven polymorphisms were detected in 1A9, and two of them, −126_−118 T9>T11 and 422C>G (S141C), were novel (Figure 3). The SNPs 726T>G (Y242X, *4 allele) and 766G>A (D256N, *5 allele) were found at frequencies of 0.003 and 0.010, respectively. In this study, the known insertion, −126_−118 T9>T10 (*22 allele), was also detected at a 0.666 frequency. A total of nine SNPs including two novel ones (−70G>A, and 726T>C (Y242Y)) were detected in 1A7 (Figure 3). The known nonsynonymous SNPs, 387T>G (N129K), 391C>A and 392G>A (R131K: the SNPs at 391 and 392 are completely associated), and 622T>C (W208R), were also detected at frequencies of 0.350, 0.350, and 0.219, respectively. In 1A3, 10 SNPs were detected, and only 46C>T (L16L) was novel. The known nonsynonymous SNPs, 17A>G (Q6R), 31T>C (W11R), 133C>T (R45W), and 140T>C (V47A), were detected at frequencies of 0.051, 0.265, 0.043, and 0.133, respectively (Figure 4).

Figure 2
figure2

SNPs and haplotypes of UGT1A8 and 1A10 (Block 8/10) in a Japanese population. aPosition on AF297093.1. bA of the initiation codon in each gene segment is numbered 1. For intron SNPs, their positions were numbered from the nearest exon. cNo amino-acid change. ddbSNP number in the National Center for Biotechnology Information. eThe SNPs included in the PharmGKB database (http://www.pharmgkb.org/). fThe haplotypes are described as numbers plus alphabetical letters. The prediction of rare haplotypes that were inferred from only one subject (frequency was 0.003) is sometimes inaccurate.

Figure 3
figure3

SNPs and haplotypes of UGT1A9, 1A7, and 1A6 (Block 9/6) in a Japanese population. aPosition on AF297093.1. bA of the initiation codon in each gene segment is numbered 1. For intron SNPs, their positions were numbered from the nearest exon. cNo amino-acid change. ddbSNP number in the National Center for Biotechnology Information. eThe haplotype numbering of UGT1A6 basically followed the numbering by Nagar et al.14 fThe haplotypes are described as numbers plus alphabetical letters. The prediction of rare haplotypes that were inferred from only one subject (frequency was 0.003) is sometimes inaccurate.

Figure 4
figure4

SNPs and haplotypes of UGT1A3 and 1A1 (Block 3/1) in a Japanese population. aPosition on AF297093.1. bA of the initiation codon in each gene segment is numbered 1. For intron SNPs, their positions were numbered from the nearest exon. cNo amino-acid change. ddbSNP number in the National Center for Biotechnology Information. eThe haplotypes are described as numbers plus alphabetical letters. The prediction of rare haplotypes that were inferred from only one subject (frequency was 0.003) is sometimes inaccurate.

LD analysis

Next, pairwise LD analysis was performed for the UGT1A genes using the polymorphisms detected in this study (1A8, 1A10, 1A9, 1A7, and 1A3) and previous studies (1A6, 1A4, 1A1, and common exons 2–5)19, 20, 21 from the same 196 subjects, and the values for rho square (r2), chi square (χ2), and D were obtained. Variations found in only one subject were excluded from the analysis. Since the data for r2 and χ2 values were almost equivalent, only the data for r2 is depicted in Figure 5. Several close linkages within each segment were seen between the variations of the 1A8, 1A10, 1A7, and 1A3 gene segments, as seen in the 1A6, 1A4, and 1A1 segments and common exons that were previously analyzed. Furthermore, strong linkages were also observed across the segments, especially between 1A8 and 1A10, among 1A9, 1A7, and 1A6, and between 1A3 and 1A1, where approximately 19, 35.5, 10, 11, 26, 10.5, and 29.5 kb separate 1A10 from 1A8, 1A9 from 1A10, 1A7 from 1A9, 1A6 from 1A7, 1A4 from 1A6, 1A3 from 1A4, and 1A1 from 1A3, respectively. Thus, the region from 1A8 to 1A1 was divided into four LD blocks: Block 8/10 (1A8 and 1A10), Block 9/6 (1A9, 1A7, and 1A6), Block 4 (1A4), and Block 3/1 (1A3 and 1A1). The data from the D values supported this block partitioning (data not shown). In addition, a few exceptional strong linkages (over 0.7 for r2) beyond the LD blocks were also observed. The perfect linkage (r2=1) was detected between 1A10 177G>A (M59I) and 1A9 766G>A (D256N), and between 1A4 IVS1+101G>T and 1A1 686C>A (P229Q). A strong LD (r2=0.95) was also shown between 1A8 711A>C (T237T) and 1A3 17A>G (Q6R). In addition, 1A7 756G>A (L252L) was strongly linked with 1A4 −219C>T, −163G>A, 142T>G (L48V), 448T>C (L150L), 804G>A (P268P), and IVS1+43C>T (r2=0.78 or greater). The other linkages were less than 0.67 for r2.

Figure 5
figure5

Linkage disequilibrium (LD) analysis of UGT1A SNPs. Pairwise LD is expressed as r2 (from 0 to 1) by a 10-graded blue color for the polymorphisms (n=2 or greater) in 1A8, 1A10, 1A9, 1A7, 1A6, 1A4, 1A3, 1A1, and common exons 2–5 located in the same order on the chromosome. The denser color represents the higher linkage. Block C: the same as Block 2 in the previous paper, and the nucleotide positions of polymorphisms in the exons 2–5 are numbered as in UGT1A1.19 UTR: untranslated region.

Haplotype analysis

Next, the haplotypes for Block 8/10, Block 9/6, and Block 3/1 were analyzed. The haplotypes of Block 4, consisting of the 1A4 segment,21 and Block C, covering common exons 2–5 (Block 2 in a previous paper19), have already been reported. The block haplotypes were tentatively named with Roman numerals plus small alphabetical letters. The haplotypes were also estimated for each gene segment (segment haplotypes), where a group of haplotypes without amino-acid changes was defined as *1.

As for Block 8/10, consisting of the two segments 1A8 and 1A10, six haplotypes were first unambiguously assigned by the presence of homozygous SNPs at all sites (*Ia, *Ib, *IIa, and *IIb) or a heterozygous SNP at only one site (*IIc and *IIIa). We separately estimated the diplotype configuration (a combination of haplotypes) for each subject by LDSUPPORT software. The diplotype configurations of 184 subjects were inferred with a probability greater than 0.96, and the remaining diplotypes (12 subjects) had a probability greater than 0.54. The haplotype inferred in the diplotypes with low probabilities were described with a question mark in Figure 2 (i.e., *IVa?). There were eight additionally inferred haplotypes. The block haplotypes were also described in the form of segment haplotype (1A8 haplotype-1A10 haplotype) combinations in Figure 2: in the 1A8 segment, the segment haplotype bearing the nonsynonymous A173G (*2 allele) was named *2; in 1A10, the haplotypes bearing M59I (*2 allele), T202I (*3 allele), A2T, and E67G were named *2, *3, *2T, and *67G, respectively. The most frequent block haplotype (segment haplotype combination in parenthesis) was *IIa (*2a*1a) (frequency: 0.492), followed by *Ia (*1b*1a) (0.311), *IIb (*2a*1b) (0.084), and *Ib (*1c*1a) (0.051). The frequencies of other block haplotypes were less than 0.05. It is noteworthy that the low-activity 1A10 haplotype *3 was completely linked with the 1A8*1 haplotype (Block 8/10 *IV).

Regarding Block 9/6 (1A9-1A7-1A6), 14 haplotypes were first unambiguously assigned by homozygous SNPs at all sites (*Ia, *Ib, *IIa, and *IIIa) or a heterozygous SNP at only one site (*Ic, *Id, *IIc, *IId, *IIIb, *IVa, *Va, *VIIIa, *XIa, and *XIIa). Additionally, eight haplotypes (*IIb, *IIe, *IIIc, *VIa, *VIb, *VIIa, *IXa, and *Xa?) were inferred, and diplotype configurations of 191 subjects were inferred with a 1.00 probability by the software. The haplotype inferred in the diplotype with a low probability was *Xa? (Figure 3). The block haplotypes were also described as combinations of segment haplotypes (1A9 haplotype-1A7 haplotype-1A6 haplotype) in Figure 3: in the 1A9 segment, the segment haplotype bearing Y242X (*4 allele), D256N (*5 allele), −126_−118 T9>T10 (*22 allele), −126_−118 T9>T11, or S141C were named *4,*5, *22, *T11, or *141C, respectively; in the 1A7 segment, the haplotype bearing N129K/R131K (*2 allele) was named *2, and the haplotype bearing N129K/R131K/W208R (*3 allele) was named *3; in 1A6, the haplotypes bearing S7A/T181A/R184S (*2 allele), S7A/R184S (*4 allele), S7A/S103X/T181A/R184S (*5 allele), and S7A/R90H/T181A/R184S (*6 allele) were named *2, *4, *5, and *6, respectively, as described previously.20 The most frequent haplotype of Block 9/6 was *Ia (*22a*1a*1a) (0.594), followed by *IIa (*1a*3a*2a) (0.184), and *IIIa (*1a*2a*1a) (0.074) (Figure 3). The frequencies of the other haplotypes were under 0.05. Notably, most (97.6%) of the high-activity segment haplotype 1A9*22 was linked with 1A7*1 and 1A6*1 (Block 9/6 *I). The 1A7 low-activity haplotype *3 was mostly linked (97.7%) with 1A6*2 haplotype (*II and *IVa in Figure 3).

Regarding Block 3/1 (1A3-1A1), six haplotypes were first unambiguously assigned by the presence of homozygous SNPs at all sites (*Ia, *IIa, *IIIa, and *Va) or a heterozygous SNP at only one site (*Ib and *VIa). The diplotype configurations of 188 subjects were inferred with a 1.00 probability. The additionally inferred haplotypes were *IIb, *IIc, *IIIb, *IIIc, *IVa*IVe?, and *VIIa. The haplotype *IVe? was inferred with a low probability (Figure 4). The combinations of segment haplotypes (1A3 haplotype-1A1 haplotype) were also described in Figure 4: in 1A3, the group bearing the nonsynonymous variations Q6R/W11R, W11R, R45W, and W11R/V47A were named the *6R11R, *11R, *45W, and *11R47A haplotypes, respectively;22 in 1A1, the haplotypes bearing G71R (*6 allele), −40_−39 insTA (*28 allele with or without *60 allele), and −3279T>G (*60 allele without *28 allele) were named the *6, *28, and *60 haplotype groups as described previously.19 The most frequent haplotype of Block 3/1 was *Ia (*1a*1a) (frequency: 0.564), followed by *IIa (*11R47A*28b) (0.122), *IIIa (*1a*6a) (0.102), *IVa (*11Ra*60a) (0.056), and *Va (*6R11R*60a) (0.051). The frequencies of the other block haplotypes were less than 0.05. It is noteworthy that the high-activity segment haplotype 1A3*11R47A was completely linked with the low-activity haplotype 1A1*28 (Block 3/1 *II). The low-activity haplotype 1A1*6 was mostly linked (71.3%) with the 1A3*1 haplotype (*III). The high-activity 1A3*11R haplotype was perfectly linked with the low-activity 1A1*60 haplotype (*IV).

Finally, no statistically significant differences in haplotype frequencies were found between the subjects with the different disease types in Block 8/10, Block 9/6, and Block3/1 (P0.05 by χ2 test or Fisher's exact test).

Genotyping and haplotype analysis across the LD blocks

A typing method was developed and additional 105 Japanese subjects (16 arrhythmic patients and 89 cancer patients) were genotyped, where direct sequencing (for nine polymorphisms in the 1A9 5′-flanking region and 1A4) and pyrosequencing (for the rest of the polymorphisms) were used for detection of 41 polymorphisms (see Materials and methods and the Table 1 legend) with (potentially) functional importance. The frequencies from 301 subjects in total are described in Table 1. Again, all the allele frequencies were in Hardy–Weinberg equilibrium, and statistically significant differences were not observed in any of the allelic frequencies between the two disease types (P0.05 by χ2 test or Fisher's exact test). Almost the same LD map as in Figure 5 was obtained between the 41 tagged variations (data not shown), indicating the robustness of our block partitioning. It is noteworthy that the known variation 1A1 1456T>G (Y486D, *7 allele) was newly found in one subject (frequency: 0.002).

Table 1 Frequencies of variations in the UGT1A gene complex detected in 301 Japanese subjects

Several reports have shown that some polymorphisms in 1A9, 1A7, 1A6, and 1A1 were closely linked,17, 18 and we also observed several weak linkages beyond the LD blocks (see Figure 5). Therefore, the block-haplotype combinations (whole complex haplotypes) were analyzed among Block 8/10, Block 9/6, Block 4, Block 3/1, and Block C (common exons 2–5) by LDSUPPORT software utilizing the polymorphisms. In 1A4 (Block 4), the haplotypes bearing L48V (*3 allele) and R11W (*4 allele) were named *3 and *4, respectively, as described previously.21 Polymorphisms found at a frequency less than 0.010, and subjects with these polymorphisms were excluded in this analysis. When Block 8/10 or Block C was included in the analysis, the whole-complex haplotypes were highly complicated (data not shown). However, if Block 8/10 and Block C were excluded, the diplotype configurations of 278 subjects were inferred with a probability greater than 0.91 (mostly>0.95) using the 18 tagged polymorphisms (see Figure 6 legend for polymorphisms). The haplotypes covering Block 9/6, Block 4, and Block 3/1 are summarized in Figure 6. Again, we did not find any statistically significant differences in frequencies of haplotypes covering Block 9/6, Block 4 and Block 3/1 between the subjects with the different disease types (P0.05 by χ2 test or Fisher's exact test). The region from 1A9 to 1A1 is approximately 90 kb length. Since the 18 variations were used for haplotyping, the number of inferred haplotype combinations (only 26) is unexpectedly small compared to the theoretical ones (Figure 6).

Figure 6
figure6

Combinations of Block 9/6, Block 4, and Block 3/1 haplotypes in a Japanese population. The used variations were UGT1A9 −126_−118 T9>T10, 1A7 387T>G, 391C>A, 392G>A and 622T>C, 1A6 19T>G, 541A>G and 552A>C, 1A4 31C>T and 142T>G, 1A3 17A>G, 31T>C, 133C>T and 140T>C, and 1A1 −3279T>G, −40_−39 insTA, 211G>A and 686C>A. In 1A4 (Block 4), the haplotypes bearing L48V (*3 allele), and R11W (*4 allele) were named *3 and *4, respectively.

Several functionally important linkages were found across the blocks. Block 9/6 *VI (1A9*1-1A7*2-1A6*4) and Block 3/1 *IIb (1A3*11R47A-1A1*28c containing the *60, *28, and *27 alleles) were perfectly linked (6/6 cases). Most of the 1A1*6-containing haplotypes (Block 3/1 *III and *VI) (69/85 cases) were associated with Block 4 (1A4) *1 and Block 9/6 *II (harboring 1A7*3 and 1A6*2). The 1A1*60-harboring haplotypes (Block 3/1 *IV and *V) were very closely linked with Block 9/6 *III (harboring 1A7*2) and Block 4 *3 (59/71 cases of 1A1*60-harboring haplotypes). Most of Block 3/1 *VI (1A3*45W-1A1*6) (25/26 cases) was associated with Block 9/6 *II (1A9*1-1A7*3-1A6*2), and Block 4 *4 was perfectly linked (4/4 cases) with both Block 3/1 *VI and Block 9/6 *II.

In addition, we found that Block 8/10 *IV (containing the low-activity allele 1A10*3 (T202I)) was strongly linked with Block 9/6 *III (1A9*1-1A7*2-1A6*1), 1A4*3, and Block 3/1 *IV (1A3*11R-1A1*60) (4/5 cases of Block 8/10 *IV, data not shown). Block 3/1 *V (harboring 1A3*6R11R and 1A1*60a) was perfectly linked with Block C *IB (25/25 cases of Block 3/1 *V, data not shown).

Discussion

Previously, we have reported the genetic variations of UGT1A6, 1A4, 1A1 segments and common exons 2–5 found in 196 Japanese subjects.19, 20, 21 In this study, we first directly sequenced 1A8, 1A10, 1A9, 1A7, and 1A3 using genomic DNA from the same Japanese subjects and detected 7, 8, 7, 9, and 10 genetic polymorphisms, respectively (Figures 2, 3 and 4). Two and one novel nonsynonymous SNPs were found in 1A10 (4G>A, A2T; 200A>G, E67G) and 1A9 (422C>G, S141C), respectively. As for 1A9 S141C, our preliminary results have shown that this amino-acid substitution reduces the enzymatic activity against 7-hydroxy-4-trifluoromethylcoumarin in vitro (Jinno et al., unpublished data). Since the guanine base at position +4 is important for translation initiation,23 1A10 4G>A might decrease the translation rate. Moreover, the luciferase-reporter activity of 1A9 −126_−118 T10 (1A9*22 allele) was reported to increase 2.6-fold as compared to that of 1A9 −126_−118 T9.24 Therefore, the novel variation 1A9 −126_−118 T9>T11 may also affect transcriptional activity. Further studies are needed to ascertain these possibilities. Recently, 1A7 −57G was reported to reduce the luciferase activity by 70% of the wild-type −57T.25 While this SNP is linked with either 1A7*3 (129K/131K/208R) or *4 (208R) in Germans, our study showed that −57G was completely linked with 1A7*3 due to the absence of 1A7*4 in Japanese.

For the 1A8 alleles, only *1 and *2 were detected. Our segment haplotypes *1a, *1b, and *2a correspond to alleles *1a, *1, and *2, respectively, in a previous study on Americans.8 The frequencies obtained in the United States,8 0.282, 0.551, and 0.145, for *1a, *1, and *2, respectively, are different from those obtained in this study, 0.023, 0.316, and 0.587 for *1a, *1b, and *2a, respectively. The allele frequency of 1A9 −126_−118 T9>T10 (*22 allele) in our data (0.666) was similar to that reported previously in Japanese (0.60), but higher than those in Caucasians (0.39) and African-Americans (0.44).24 For 1A7, the frequencies of *1, *2 (129K/131K), and *3 (129K/131K/208R) haplotypes were 0.651, 0.130, and 0.219, respectively. Our data are comparable to the previous data for a Japanese population,26 but not to those on Caucasians (0.355, 0.280, and 0.365) and Egyptians (0.420, 0.200, and 0.380).17 As for 1A3, the frequencies of the haplotypes *1, *11R, *6R11R, *11R47A, and *45W were 0.692, 0.082, 0.051, 0.133, and 0.043, respectively. These are similar to the previous data obtained from the Japanese.22

Recently, linkages among the SNPs in 1A9, 1A7, 1A6, and 1A1 have been reported in Americans.18 By our LD analysis, strong linkages were shown between the SNPs in 1A8 and 1A10, among those in 1A9, 1A7, and 1A6, and also between those in 1A3 and 1A1. Moreover, this is the first report on the haplotype analysis using high-density SNPs for the entire UGT1A complex. By block haplotyping, several close linkages between the segmental haplotypes were observed: between the 1A8*1 and 1A10*3 haplotypes in Block 8/10; between the 1A7*3 and 1A6*2 in Block 9/6; between the 1A3*11R47A and 1A1*28 and between the 1A3*11R and 1A1*60 haplotypes in Block 3/1. Carlini et al.18 reported that 1A7 low-activity alleles (1A7*2 and *3) were perfectly linked to 1A9*1 in Americans (including Caucasians (83%) and African-American (14%)). Also in this study, most (95.6%) of the 1A7*2 or *3 alleles were linked to the 1A9*1 allele in Japanese (Figure 3).

We conducted additional typing of 105 subjects by pyrosequencing and direct sequencing, and confirmed the presence of several functionally important haplotype combinations beyond the blocks (Figure 6). In Americans (including Caucasians (83%) and African-American (14%)),18 Caucasians, and Egyptians,17 75, 78, and 57%, respectively, of 1A7*3 were associated with the 1A1*28 allele, though only the *28 allele was genotyped in 1A1 in these analysis. In our more intensive analysis, most of the 1A7*3 haplotype was associated with either the 1A1*28b haplotype (having 1A1*60 and *28 alleles) (26.1% of the 1A7*3 haplotype) or the 1A1*6 haplotype (67.2% of UGT1A7*3). Thus, different profiles for the linkage of 1A7*3 with the 1A1 polymorphisms between the Caucasians and Japanese reflect the facts that the frequency of the 1A1*6 haplotype in the Asian populations was relatively high, and that the 1A1*28 and *6 alleles were mutually exclusive.19 In fact, linkage between 1A1*6 and 1A7*3 alleles was recently suggested in Taiwanese.27 Innocenti et al. reported the three most common 1A9-1A1 haplotype combinations were 1A9*22-1A1*1 (36.4%), 1A9*1-1A1*28b (28.0%), and 1A9*1-1A1*1 (18.6%) for Caucasians, and 1A9*22-1A1*1 (45.3%), 1A9*1-1A1*60 (22.3%), and 1A9*1-1A1*6 (12.7%) for Asians.28 In this study for Japanese, 1A9*22-1A1*1, 1A9*1-1A1*60, and 1A9*1-1A1*6 (58.5, 11.9, and 13.3%, respectively) were also the most common three combinations. Furthermore, we revealed that most (98.2%) of the 1A1*1 haplotype was linked with 1A9*22, and 87.1% of 1A1*6, 100% of 1A1*28c, and 93.0% of 1A1*60 were associated with 1A9*1. Collectively, haplotype combinations are suggested to be different between Caucasians and Asians. In addition, several interesting linkages were found between the segmental haplotypes as shown in the Results. For example, the segment haplotypes 1A6*4 (S7A/R184S) and 1A1*28c were strongly linked in Japanese subjects.

These linkages might be crucial for the metabolism of a certain drug for which two or more UGT1A isoforms significantly contribute to its metabolism. In fact, multiple UGT isoforms are involved in glucuronidation of several compounds, for example SN-38,7, 29 estrogens and their metabolites (estron, estradiol, 2-hydroxyestrone, and others),30, 31 and arachidonic acid and its metabolites.32, 33 UGT1A1, 1A9, and 1A7 play important roles in SN-38 glucuronidation.7, 29 The 1A1*60, *28b, and *6 haplotypes are associated with reduced UGT1A1 activity to SN-38.15, 19, 34, 35 Since the 1A9 high-activity (high transcription) haplotype *22 was dominant in Japanese (0.666), 1A9*1 can be considered (relative to *22) as a low-activity haplotype. The 1A7*3, but not *2, haplotype has a reduced glucuronidation activity (by 59%) to SN-38.7 A more recent report has shown that UGT1A10 is also responsible for SN-38 glucuronidation,36 and that the 1A10 *3 (T202I) is a low-activity allele.11 We found that the Block 8/10 *IV haplotype (harboring 1A10*3) was closely linked with Block 9/6 *III (harboring 1A9*1) and Block 3/1 *IV (harboring 1A1*60). Furthermore, most of Block 9/6 *II (harboring 1A9*1 and 1A7*3) were estimated to be linked with Block 3/1 *III or *VI (having 1A1*6), or Block 3/1 *IIa (having 1A1*28b). Though the functional significance of 1A10 T202I toward SN-38 is currently unknown, it is possible that the concurrently reduced activities of UGT1A10, 1A9, 1A7, and 1A1 may influence SN-38 glucuronidation.

Arachidonic acid and its metabolites prostaglandins were conjugated with UGT1A1, 1A3, 1A9, 1A10, and 2B7.33 UGT1A1, 1A3, and 1A4 also had catalytic activities toward a hydroxylated metabolite of arachidonic acid, 12- and 15-hydroxyeicosatetraenoic acid.32 Furthermore, glucuronidation of leucotriene B4, another arachidonic acid metabolite that mediates the inflammation process, can be catalyzed by UGT1A1, 1A3, 1A8, and 2B7.32 Thus, co-occurrence of the functionally less active haplotypes (such as Block 9/6 *II (including 1A9*1)-Block 3/1 *VI (harboring 1A3*45W and 1A1*6)), might cooperatively influence the metabolism of several important compounds in the arachidonic acid cascade.

Since plural UGT isoforms are often involved in the glucuronidation of ‘one’ compound, co-occurrence of the functionally less active haplotypes in the entire UGT1A gene complex needs to be carefully considered in studies on the association of genetic polymorphisms with pharmacokinetic parameters and clinical and epidemiologic data. Our findings would provide fundamental and useful information for genotyping or haplotyping of UGT1As in the Japanese and probably other Asian populations.

Materials and methods

Human genomic DNA samples

The 301 Japanese subjects consisted of 124 arrhythmic patients, who were administered β-blockers, and 177 cancer patients, who were administered irinotecan. Genomic DNA was extracted directly from blood leukocytes. The ethical review boards of the National Cancer Center, the National Cardiovascular Center, and the National Institute of Health Sciences approved this study. Written informed consent was obtained from all patients.

Polymerase chain reaction (PCR) conditions for DNA sequencing

First, the fragments were amplified from genomic DNA (150 ng) using 2.5 U of Z-Taq (Takara Bio Inc., Shiga, Japan) with 0.2 μ M primers (see ‘1st amplification’ in Table 2 for primer sequences). The exon-1's of UGT1A8 and 1A10 were simultaneously amplified by mixed primers for each gene, and those of 1A9 and 1A7 were amplified as one fragment. The primer sequences for the 1st amplification of 1A10 were described previously.37 The first PCR conditions consisted of 30 cycles of 98°C for 5 s, 55°C for 5 s, and 72°C for 190 s. Then, each exon 1 was amplified by Ex-Taq (0.625 U) (Takara Bio Inc.) using the first PCR products as templates with the 2nd amplification primers (0.2 μ M) that were designed in the introns (see ‘2nd amplification’ in Table 2 for primer sequences). The PCR primers for the 2nd amplification of 1A10 and 1A9 were described previously.37, 38 Exon 1 in 1A3 and the promoter region of 1A9 were first directly amplified from genomic DNA (100 ng) using Ex-Taq as in the 2nd round of PCR described below. The second round of PCR consisted of one cycle at 94°C for 5 min, followed by 30 cycles of 94°C for 30 s, 55°C for 1 min, and 72°C for 2 min, and then a final extension for 7 min at 72°C. These PCR products were then treated with a PCR Product Pre-Sequencing Kit (USB Co., Cleveland, OH, USA) and directly sequenced on both strands using an ABI BigDye Terminator Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA) and the primers listed in Table 2 (see ‘Sequencing’). The excess dye was removed by a DyeEx96 kit (Qiagen, Hilden, Germany), and the eluates were analyzed on an ABI Prism 3700 DNA Analyzer (Applied Biosystems). All the SNPs were confirmed by repeating the PCR on genomic DNA and sequencing the newly generated PCR products.

Table 2 Primers for amplification and sequencing of exon-1's in UGT1A8, 1A7, 1A3, and 5′-flanking region of 1A9

Genotyping (pyrosequencing)

Genotyping was performed by pyrosequencing, except for UGT1A4 31C>T (R11W), 127delA (43fsX22: frameshift from codon 43 resulting in the termination at the 22nd codon, codon 65), 142T>G (L48V), 175delG (59fsX6), 271C>T (R91C), 325A>G (R109G), and IVS+1G>T, and 1A9 −126_−118 T9>T10 or T11, which were genotyped by direct sequencing because these polymorphisms were not clearly determined by pyrosequencing. Fragments were directly amplified from genomic DNA (10–15 ng) by Ex-Taq (1 U) with amplification primer pairs (either primer was biotinylated) (Table 3). The PCR conditions consisted of 1 cycle at 94°C for 5 min, followed by 50 cycles of 94°C for 30 s, 55°C for 30 s (except for UGT1A 1598A>C (H533P), in which annealing was carried out at 58°C for 45 s), and 72°C for 30 s. Primers for 1A1 −3279T>G, −40_−39 insTA, 211G>A (G71R), 247T>C (F83L), 686C>A (P229Q), and 1456T>G (Y486D) in common exon 5 were described previously.39 Biotinylated single-stranded DNA fragments were generated as described previously.39 Briefly, PCR products were mixed with streptavidin beads for 10 min. The beads were transferred to a MultiScreen-HV Plate (Millipore Corporation, Billerica, MA, USA), and the buffer was removed by vacuum. DNA attached to the beads was denatured, washed twice, and then suspended in 20 mM Tris-acetate containing 2 mM Mg-acetate (pH 7.6). After transferring to a 96-well PSQ plate (Pyrosequencing AB, Uppsala, Sweden), 10 pmol of the sequencing primer (PAGE-purified grade) (Table 3) for SNP analysis was added to the single-stranded fragments. The mixture was incubated at 95°C for 2 min, and then cooled to room temperature for annealing. An automated pyrosequencing instrument, the PSQ™96MA (Pyrosequencing AB), and the PSQ 96 SNP reagent kit (Pyrosequencing AB) were used to perform the genotyping. To validate the typing methods, the results for 48 samples were confirmed to be identical to those obtained by direct sequencing (data not shown).

Table 3 Primers in pyrosequencing

LD and haplotype analysis

Hardy–Weinberg equilibrium analysis and LD analysis were performed using SNPAlyze software (version 3.2). (Dynacom Co. Ltd., Yokohama, Japan), and pairwise two-dimensional maps between SNPs were obtained for the D, χ2, and r2 values. Some of the haplotypes were unambiguously determined from the subjects with homozygous SNPs at all sites or a heterozygous SNP at only one site. Separately, the diplotype configurations (combination of haplotypes) were inferred by an expectation-maximization-based program, LDSUPPORT, which determines the posterior probability distribution of the diplotype configuration for each subject based on the estimated haplotype frequencies.40

The diplotype configurations of the subjects were inferred with probability (certainty) values over 0.96 for 184, 191, and 188 out of 196 subjects in the UGT1A8-1A10 block (Block 8/10), the 1A9-1A7-1A6 block (Block 9/6), and the 1A3-1A1 block (Block 3/1), respectively. The Block 4 (1A4) haplotypes were described previously.21 Note that the predictability for the extremely rare haplotypes inferred from only one subject is known to be low in some cases. Haplotype analysis was also performed among the representative SNPs in Block 9/6, Block 4, and Block 3/1 by LDSUPPORT software.

Accession codes

Accessions

GenBank/EMBL/DDBJ

Abbreviations

LD:

linkage disequilibrium

PCR:

polymerase chain reaction

SN-38:

7-ethyl-10-hydroxycamptothecin

SNP:

single nucleotide polymorphism

UGT:

UDP-glucuronosyltransferase

UTR:

untranslated region

References

  1. 1

    Radominska-Pandya A, Czernik PJ, Little JM . Structural and functional studies of UDP-glucuronosyltransferases. Drug Metab Rev 1999; 31: 817–899.

  2. 2

    Tukey RH, Strassburg CP . Genetic multiplicity of the human UDP-glucuronosyltransferases and regulation in the gastrointestinal tract. Mol Pharmacol 2001; 59: 405–414.

  3. 3

    Gong QH, Cho JW, Huang T, Potter C, Gholami N, Basu NK et al. Thirteen UDPglucuronosyltransferase genes are encoded at the human UGT1 gene complex locus. Pharmacogenetics 2001; 11: 357–368.

  4. 4

    Metz RP, Ritter JK . Transcriptional activation of the UDP-glucuronosyltransferase 1A7 gene in rat liver by aryl hydrocarbon receptor ligands and oltipraz. J Biol Chem 1998; 273: 5607–5614.

  5. 5

    Guillemette C . Pharmacogenomics of human UDP-glucuronosyltransferase enzymes. Pharmacogenomics J 2003; 3: 136–158.

  6. 6

    Guillemette C, Ritter JK, Auyeung DJ, Kessler FK, Housman DE . Structural heterogeneity at the UDP-glucuronosyltransferase 1 locus: functional consequences of three novel missense mutations in the human UGT1A7 gene. Pharmacogenetics 2000; 10: 629–644.

  7. 7

    Gagne JF, Montminy V, Belanger P, Journault K, Gaucher G, Guillemette C . Common human UGT1A polymorphisms and the altered metabolism of irinotecan active metabolite 7-ethyl-10-hydroxycamptothecin (SN-38). Mol Pharmacol 2002; 62: 608–617.

  8. 8

    Huang YH, Galijatovic A, Nguyen N, Geske D, Beaton D, Green J et al. Identification and functional characterization of UDP-glucuronosyltransferases UGT1A8*1, UGT1A8*2 and UGT1A8*3. Pharmacogenetics 2002; 12: 287–297.

  9. 9

    Elahi A, Bendaly J, Zheng Z, Muscat JE, Richie Jr JP, Schantz SP et al. Detection of UGT1A10 polymorphisms and their association with orolaryngeal carcinoma risk. Cancer 2003; 98: 872–880.

  10. 10

    Jinno H, Saeki M, Saito Y, Tanaka-Kagawa T, Hanioka N, Sai K et al. Functional characterization of human UDP-glucuronosyltransferase 1A9 variant, D256N, found in Japanese cancer patients. J Pharmacol Exp Ther 2003; 306: 688–693.

  11. 11

    Jinno H, Saeki M, Tanaka-Kagawa T, Hanioka N, Saito Y, Ozawa S et al. Functional characterization of wild-type and variant (T202I and M59I) human UDP-glucuronosyltransferase 1A10. Drug Metab Dispos 2003; 31: 528–532.

  12. 12

    Villeneuve L, Girard H, Fortier LC, Gagne JF, Guillemette C . Novel functional polymorphisms in the UGT1A7 and UGT1A9 glucuronidating enzymes in Caucasian and African-American subjects and their impact on the metabolism of 7-ethyl-10-hydroxycamptothecin and flavopiridol anticancer drugs. J Pharmacol Exp Ther 2003; 307: 117–128.

  13. 13

    Ehmer U, Vogel A, Schutte JK, Krone B, Manns MP, Strassburg CP . Variation of hepatic glucuronidation: novel functional polymorphisms of the UDP-glucuronosyltransferase UGT1A4. Hepatology 2004; 39: 970–977.

  14. 14

    Nagar S, Zalatoris JJ, Blanchard RL . Human UGT1A6 pharmacogenetics: identification of a novel SNP, characterization of allele frequencies and functional analysis of recombinant allozymes in human liver tissue and in cultured cells. Pharmacogenetics 2004; 14: 487–499.

  15. 15

    Jinno H, Tanaka-Kagawa T, Hanioka N, Saeki M, Ishida S, Nishimura T et al. Glucuronidation of 7-ethyl-10-hydroxycamptothecin (SN-38), an active metabolite of irinotecan (CPT-11), by human UGT1A1 variants, G71R, P229Q, and Y486D. Drug Metab Dispos 2003; 31: 108–113.

  16. 16

    Judson R, Stephens JC, Windemuth A . The predictive power of haplotypes in clinical response. Pharmacogenomics 2000; 1: 15–26.

  17. 17

    Kohle C, Mohrle B, Munzel PA, Schwab M, Wernet D, Badary OA et al. Frequent co-occurrence of the TATA box mutation associated with Gilbert's syndrome (UGT1A1*28) with other polymorphisms of the UDP-glucuronosyltransferase-1 locus (UGT1A6*2 and UGT1A7*3) in Caucasians and Egyptians. Biochem Pharmacol 2003; 65: 1521–1527.

  18. 18

    Carlini LE, Meropol NJ, Bever J, Andria ML, Hill T, Gold P et al. UGT1A7 and UGT1A9 polymorphisms predict response and toxicity in colorectal cancer patients treated with capecitabine/irinotecan. Clin Cancer Res 2005; 11: 1226–1236.

  19. 19

    Sai K, Saeki M, Saito Y, Ozawa S, Katori N, Jinno H et al. UGT1A1 Haplotypes associated with reduced glucuronidation and increased serum bilirubin in irinotecan-administered Japanese cancer patients. Clin Pharmacol Ther 2004; 75: 501–515.

  20. 20

    Saeki M, Saito Y, Jinno H, Sai K, Kaniwa N, Ozawa S et al. Genetic polymorphisms of UGT1A6 in a Japanese population. Drug Metab Pharmacokinet 2005; 20: 85–90.

  21. 21

    Saeki M, Saito Y, Jinno H, Sai K, Hachisuka A, Kaniwa N et al. Genetic variations and haplotypes of UGT1A4 in a Japanese population. Drug Metab Pharmacokinet 2005; 20: 144–151.

  22. 22

    Iwai M, Maruo Y, Ito M, Yamamoto K, Sato H, Takeuchi Y . Six novel UDP-glucuronosyltransferase (UGT1A3) polymorphisms with varying activity. J Hum Genet 2004; 49: 123–128.

  23. 23

    Kozak M . Recognition of AUG and alternative initiator codons is augmented by G in position +4 but is not generally affected by the nucleotides in positions +5 and +6. EMBO J 1997; 16: 2482–2492.

  24. 24

    Yamanaka H, Nakajima M, Katoh M, Hara Y, Tachibana O, Yamashita J et al. A novel polymorphism in the promoter region of human UGT1A9 gene (UGT1A9*22) and its effects on the transcriptional activity. Pharmacogenetics 2004; 14: 329–332.

  25. 25

    Lankisch TO, Vogel A, Eilermann S, Fiebeler A, Krone B, Barut A et al. Identification and characterization of a functional TATA box polymorphism of the UDP glucuronosyltransferase 1A7 gene. Mol Pharmacol 2005; 67: 1732–1739.

  26. 26

    Ando M, Ando Y, Sekido Y, Ando M, Shimokata K, Hasegawa Y . Genetic polymorphisms of the UDP-glucuronosyltransferase 1A7 gene and irinotecan toxicity in Japanese cancer patients. Jpn J Cancer Res 2002; 93: 591–597.

  27. 27

    Huang MJ, Yang SS, Lin MS, Huang CS . Polymorphisms of uridine-diphosphoglucuronosyltransferase 1A7 gene in Taiwan Chinese. World J Gastroenterol 2005; 11: 797–802.

  28. 28

    Innocenti F, Liu W, Chen P, Desai AA, Das S, Ratain MJ . Haplotypes of variants in the UDP-glucuronosyltransferase1A9 and 1A1 genes. Pharmacogenet Genomics 2005; 15: 295–301.

  29. 29

    Ciotti M, Basu N, Brangi M, Owens IS . Glucuronidation of 7-ethyl-10-hydroxycamptothecin (SN-38) by the human UDP-glucuronosyltransferases encoded at the UGT1 locus. Biochem Biophys Res Commun 1999; 260: 199–202.

  30. 30

    Lepine J, Bernard O, Plante M, Tetu B, Pelletier G, Labrie F et al. Specificity and regioselectivity of the conjugation of estradiol, estrone, and their catecholestrogen and methoxyestrogen metabolites by human uridine diphospho-glucuronosyltransferases expressed in endometrium. J Clin Endocrinol Metab 2004; 89: 5222–5232.

  31. 31

    Basu NK, Kubota S, Meselhy MR, Ciotti M, Chowdhury B, Hartori M et al. Gastrointestinally distributed UDP-glucuronosyltransferase 1A10, which metabolizes estrogens and nonsteroidal anti-inflammatory drugs, depends upon phosphorylation. J Biol Chem 2004; 279: 28320–28329.

  32. 32

    Turgeon D, Chouinard S, Belanger P, Picard S, Labbe JF, Borgeat P et al. Glucuronidation of arachidonic and linoleic acid metabolites by human UDP-glucuronosyltransferases. J Lipid Res 2003; 44: 1182–1191.

  33. 33

    Little JM, Kurkela M, Sonka J, Jantti S, Ketola R, Bratton S et al. Glucuronidation of oxidized fatty acids and prostaglandins B1 and E2 by human hepatic and recombinant UDP-glucuronosyltransferases. J Lipid Res 2004; 45: 1694–1703.

  34. 34

    Beutler E, Gelbart T, Demina A . Racial variability in the UDP-glucuronosyltransferase 1 (UGT1A1) promoter: a balanced polymorphism for regulation of bilirubin metabolism? Proc Natl Acad Sci USA 1998; 95: 8170–8174.

  35. 35

    Sugatani J, Yamakawa K, Yoshinari K, Machida T, Takagi H, Mori M et al. Identification of a defect in the UGT1A1 gene promoter and its association with hyperbilirubinemia. Biochem Biophys Res Commun 2002; 292: 492–497.

  36. 36

    Oguri T, Takahashi T, Miyazaki M, Isobe T, Kohno N, Mackenzie PI et al. UGT1A10 is responsible for SN-38 glucuronidation and its expression in human lung cancers. Anticancer Res 2004; 24: 2893–2896.

  37. 37

    Saeki M, Ozawa S, Saito Y, Jinno H, Hamaguchi T, Nokihara H et al. Three novel single nucleotide polymorphisms in UGT1A10. Drug Metab Pharmacokinet 2002; 17: 488–490.

  38. 38

    Saeki M, Saito Y, Jinno H, Sai K, Komamura K, Ueno K et al. Three novel single nucleotide polymorphisms in UGT1A9. Drug Metab Pharmacokinet 2003; 18: 146–149.

  39. 39

    Saeki M, Saito Y, Jinno H, Tohkin M, Kurose K, Kaniwa N et al. Comprehensive UGT1A1 genotyping in a Japanese population by pyrosequencing. Clin Chem 2003; 49: 1182–1185.

  40. 40

    Kitamura Y, Moriguchi M, Kaneko H, Morisaki H, Morisaki T, Toyama K et al. Determination of probability distribution of diplotype configuration (diplotype distribution) for each subject from genotypic data using the EM algorithm. Ann Hum Genet 2002; 66: 183–193.

Download references

Acknowledgements

This study was supported in part by the Program for Promotion of Fundamental Studies in Health Sciences and by the Health and Labour Sciences Research Grants from Ministry of Health, Labour and Welfare, and the grant from the Japan Health Sciences Foundation. We thank Ms Chie Sudo for her secretarial assistance.

Author information

Correspondence to Y Saito.

Additional information

DUALITY OF INTEREST

None declared.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Saeki, M., Saito, Y., Jinno, H. et al. Haplotype structures of the UGT1A gene complex in a Japanese population. Pharmacogenomics J 6, 63–75 (2006). https://doi.org/10.1038/sj.tpj.6500335

Download citation

Keywords

  • UGT1A
  • single nucleotide polymorphism
  • linkage disequilibrium
  • haplotype

Further reading