CYP2A7 Polymorphic Alleles Confound the Genotyping of CYP2A6*4A Allele

CYP2A7 polymorphic alleles confound the genotyping of CYP2A6*4A allele

T Fukami1, M Nakajima1, Human cytochrome P450 (CYP) 2A6 metabolizes nicotine to cotinine. 1 2 Genetic polymorphisms of CYP2A6 contribute to the interindividual H Sakai , HL McLeod and variability of nicotine metabolism. We encountered some subjects possessing 1 T Yokoi two copies of the CYP2A6 gene, although they were genotyped as

1 heterozygotes of the CYP2A6*4A allele (entire CYP2A6 gene deleted allele). Department of Drug Metabolism and From the subjects, we found CYP2A7 polymorphic alleles (CYP2A7*1B, Toxicology, Division of Pharmaceutical Sciences, 0 Graduate School of Medical Science, Kanazawa CYP2A7*1C, and CYP2A7*1D) in which the sequences in the 3 -flanking University, Kakuma-machi, Kanazawa, Japan region were converted to the corresponding CYP2A6 sequences, being and 2Department of Medicine, Washington confused with the CYP2A6*4A. These allele frequencies in European- University School of Medicine, St Louis, MO, USA Americans (n ¼ 187) were 1.3, 2.1, 0.3%, respectively, but these were very Correspondence: rare in African-Americans (n ¼ 176), Japanese (n ¼ 184), and Koreans Dr M Nakajima, Drug Metabolism and (n ¼ 209). By an improved genotyping method, the allele frequency of Toxicology, Division of Pharmaceutical CYP2A6*4A of 3.7% in European-Americans was corrected to 0%. The Sciences, Graduate School of Medical Science, comprehensible and reliable genotyping method developed in this study Kanazawa University, Kakuma-machi, Kanazawa 920-1192, Japan. would be useful to evaluate associations between the genotype and E-mail: [email protected] phenotype. The Pharmacogenomics Journal (2006) 6, 401–412. doi:10.1038/sj.tpj.6500390; published online 25 April 2006

Keywords: CYP2A6; CYP2A7; deletion allele; genetic polymorphism

Introduction

Cytochrome P450 (CYP) enzymes are involved in the metabolism of numerous exogenous and endogenous substances including drugs, environmental chemicals, steroid hormones and bile acids. The CYP2A gene subfamily comprises three genes, CYP2A6, CYP2A7, and CYP2A13, as well as the split pseudogene CYP2A18. CYP2A6 can metabolize nicotine,1 coumarin,2 tegafur,3 losigamone,4 letrozole,5 and nitrosamines such as 4-methylnitrosamino-1-(3-pyridyl)-1-butanone.6 CYP2A7 has been shown to not incorporate heme and to be functionally inactive.2,7 There are genetic polymorphisms in the CYP2A6 gene. In addition to a variety of single nucleotide polymorphism (SNP)s, several alleles that are produced by crossover with the CYP2A7 gene are known to exist. The CYP2A7 gene is located approximately 25 kb upstream of the CYP2A6 gene with a identity of 96.5% in the coding nucleotide sequence.2 The CYP2A6*1B allele has a gene conversion with CYP2A7 in the 30-untranslated region.8 In the CYP2A6*4 (CYP2A6*4A, CYP2A6*4B, and CYP2A6*4D) alleles, the entire CYP2A6 gene is deleted and the enzymatic activity is lacking.8–11 The CYP2A6*1X2 allele has a duplication of the CYP2A6 gene as the reciprocal product of the CYP2A6*4D allele after an unequal Received 3 November 2005; revised 17 12 February 2006; accepted 24 February 2006; crossover event. The CYP2A6*12 is a CYP2A7/CYP2A6 hybrid allele created by published online 25 April 2006 an unequal crossover in intron 2.13 Identification of novel CYP2A7 polymorphic alleles T Fukami et al 402

Recently, we developed a genotyping method to distin- the sequences at region 1 as well as region 2 are deleted. In guish the CYP2A6*4A, CYP2A6*4D, CYP2A6*1F, and CY- the other allele termed CYP2A7*1D, the sequences at regions P2A6*1G alleles.14 Using this method, we found that 14 out 1, 2, and 3 are deleted. This is the first report of the CYP2A7 of 187 European-American subjects heterozygously pos- polymorphic alleles. sessed the CYP2A6*4A allele, resulting in an allele frequency of 3.7%. Heterozygotes of the CYP2A6*4 allele have one Frequencies of CYP2A7 polymorphic alleles copy of the CYP2A6 gene. To our surprise, several subjects With the conventional genotyping method of CYP2A6*4,14 with the CYP2A6*4A allele were judged as possessing two the subject not possessing the CYP2A6*4 allele could not copies of the CYP2A6 gene, as they were genotyped as have the CYP2A7 polymorphic allele. Therefore, we per- heterozygotes for other SNPs with the wild type. From these formed the genotyping analyses of the CYP2A7 poly- subjects, we found novel CYP2A7 polymorphic alleles in morphic alleles (Figure 1b–e) for 14 European-Americans, which the sequences in the 30-flanking region have under- three African-Americans, 73 Japanese and 42 Koreans who gone conversion with the corresponding CYP2A6 sequences, were once genotyped as heterozygotes or homozygotes of leading to misgenotyping as the CYP2A6*4A allele. the CYP2A6*4 allele. Among the 14 European-Americans, five subjects were genotyped as CYP2A7*1A/CYP2A7*1B; seven subjects were genotyped as CYP2A7*1A/CYP2A7*1C; Results and two subjects were genotyped as CYP2A7*1A/CYP2A7*1D (Table 1). One African-American subject and one Japanese Sequence analysis of the 30-flanking region of the CYP2A7 gene subject were genotyped as CYP2A7*1A/CYP2A7*1D With polymerase chain reaction-restriction fragment length and CYP2A7*1A/CYP2A7*1B, respectively. In contrast, the polymorphism (PCR-RFLP) analysis using the primers CYP2A7 polymorphic alleles were not found in the Korean of 2Aint7F and 2A6R2,14 we found that 14 out of 187 subjects. The allele frequencies of the CYP2A7 polymorphic European-American subjects and three out of 176 African- allele are summarized in Table 1. American subjects were heterozygotes of the CYP2A6*4A allele, resulting in allele frequencies of 3.7 and 0.9%, respectively. However, all European-American and one out CYP2A6 copy number in the subjects possessing the CYP2A7 of three African-American heterozygotes of the CYP2A6*4A polymorphic alleles allele were genotyped as a heterozygote for SNPs of To investigate the copy number of the CYP2A6 gene in g.À48T4G (found in CYP2A6*9 allele) or g.51G4A (found the subjects with the CYP2A7 polymorphic alleles, we in CYP2A6*14, CYP2A6*18B, CYP2A6*20, CYP2A6*21, and performed quantitative analyses of the PCR-amplified frag- CYP2A6*22 alleles) with the wild type, indicating the ments. The genomic DNA samples from the Japanese or presence of two copies of the CYP2A6 gene. In contrast, Korean subjects who did not have the CYP2A7 polymorphic no contradiction in the genotype was observed for 73 alleles and were genotyped as CYP2A6*4A/CYP2A6*4A, Japanese and 42 Korean subjects heterozygously or homo- CYP2A6*1/CYP2A6*4A, CYP2A6*1/CYP2A6*1 or CYP2A6*1/ zygously possessing the CYP2A6*4A allele. CYP2A6*1X2 were used for the standard curve. The region As we previously reported,14 the primer 2Aint7F can from exon 3 to intron 3 in both the CYP2A6 and CYP2A7 anneal to both the CYP2A6 and CYP2A7 genes, but the genes was amplified (Figure 2). The ratios of the PCR primer 2A6R2 specifically anneals to the CYP2A6 gene. If the products corrected with the fragment lengths of CYP2A6/ primer 2A6R2 anneals to the CYP2A7 gene, the PCR-RFLP CYP2A7 for the standard samples were as follows: 0.07 and pattern would cause us to mistakenly genotype it as the 0.08 for CYP2A6*4A/CYP2A6*4A (n ¼ 2), 0.81–0.93 for CYP2A6*4 allele. To investigate the cause of the contra- CYP2A6*1/CYP2A6*4A (n ¼ 6), 1.69–2.06 for CYP2A6*1/ diction observed in these subjects, sequence analyses of the CYP2A6*1 (n ¼ 6), and 2.81 for CYP2A6*1/CYP2A6*1X2 30-flanking region of the CYP2A7 gene were performed. (n ¼ 1). With the standard curve, the ratios of the copy Consequently, three novel CYP2A7 alleles were found numbers of CYP2A6/CYP2A7 in the 14 European-American, (Figure 1a). In an allele termed CYP2A7*1B, the sequences one African-American and one Japanese subjects who at the region of the primer 2A6R2 are converted with the possessed the CYP2A7 polymorphic alleles were calculated corresponding sequences of the CYP2A6 gene, that is, a 6-bp as 0.88–1.12, indicating the presence of two copies of the deletion at region 2. In another allele termed CYP2A7*1C, CYP2A6 gene.

Figure 1 (a) Sequence alignments of the 30- flanking region of CYP2A7*1A, CYP2A7*1B, CYP2A7*1C, and CYP2A7*1D alleles, as well as the CYP2A6 gene. The accession number of the genomic DNA sequences of CYP2A7 and CYP2A6 is NG_000008.5. The nucleotide numbering refers to the TGA of the stop codon with the next nucleotide of A as 1 with the reference sequence of the CYP2A7 gene. Deletions are denoted by dashes. The solid boxes represent the regions that are different from those of CYP2A6 gene. The horizontal arrows indicate the location and direction of the primers for the polymerase chain reaction. The recognition sites of Tai I are indicated by vertical arrows with dashed boxes. (b) Schematic and (c) representative polymerase chain reaction-restriction fragment length polymorphism patterns for CYP2A7 alleles. (d) Schematic and (e) representative allele specific-PCR patterns for CYP2A7*1A/CYP2A7*1C or CYP2A7*1A/CYP2A7*1D. In the present study, there was no homozygote of the CYP2A7 polymorphic alleles.

The Pharmacogenomics Journal Identification of novel CYP2A7 polymorphic alleles T Fukami et al 403

The region from exon 5 to intron 5 in both the CYP2A6 and CYP2A7 forthestandardsampleswereasfollows:0.00for CYP2A7 genes was also amplified (Figure 3). The ratios of the CYP2A6*4A/CYP2A6*4A (n ¼ 2), 0.75–0.82 for CYP2A6*1/ PCR products corrected with the fragment lengths of CYP2A6/ CYP2A6*4A (n ¼ 6), 1.30–1.53 for CYP2A6*1/CYP2A6*1

The Pharmacogenomics Journal Identification of novel CYP2A7 polymorphic alleles T Fukami et al 404

Table 1 Allele frequencies of CYP2A7 alleles and CYP2A6*4A in four populations

Allele European-Americans African-Americans Japanese Koreans (number of allele) (n ¼ 374) (%) (n ¼ 352) (%) (n ¼ 368) (%) (n ¼ 418) (%)

CYP2A7*1B 1.3 (0.6–3.0) (n ¼ 5) 0.0 (0.0–0.8) 0.3 (0.1–1.6) (n ¼ 1) 0.0 (0.0–0.7) CYP2A7*1C 1.9 (0.9–3.8) (n ¼ 7) 0.0 (0.0–0.8) 0.0 (0.0–0.8) 0.0 (0.0–0.7) CYP2A7*1D 0.5 (0.1–1.9) (n ¼ 2) 0.3 (0.1–1.6) (n ¼ 1) 0.0 (0.0–0.8) 0.0 (0.0–0.7) CYP2A6*4A 0.0 (0.0–0.8) 0.6 (0.1–2.1) (n ¼ 2) 22.3 (18.3–26.8) (n ¼ 82) 11.0 (8.3–14.4) (n ¼ 46)

The values of 95% confidence interval are in parentheses.

Figure 2 Copy number assay for the CYP2A6 and CYP2A7 genes by polymerase chain reaction-restriction fragment length polymorphism (PCR- RFLP) targeting exon 3. (a) Schematic structures of the CYP2A6 and CYP2A7 genes. Boxes represent exon 3 and lines represent intron 3. Polymerase chain reaction amplification from exon 3 to intron 3 was performed with the primer pair indicated by horizontal arrows. The recognition site of EcoN I is indicated by a vertical arrow. (b) Representative photograph of PCR-RFLP patterns for different CYP2A6 genotypes. The intensities of the 391-bp (CYP2A6) and 273-bp (CYP2A7) fragments were quantified. The ratio of the intensities corrected by the fragment length was calculated as the ratio of the copy numbers of CYP2A6/CYP2A7. Homozygotes of CYP2A6*4A, heterozygotes of CYP2A6*1/CYP2A6*4A, homozygotes of CYP2A6*1, and heterozygotes of CYP2A6*1/CYP2A6*1X2 have 0, 1, 2 and 3 copies of the CYP2A6 gene, respectively. (c) Standard curve of the ratio of the copy numbers of CYP2A6/CYP2A7.(d) The ratios of the copy numbers of CYP2A6/CYP2A7 in the 14 European-American, one African-American and one Japanese subjects who possessed the CYP2A7 polymorphic alleles were calculated using the standard curve shown in (c).

(n ¼ 6), and 2.10 for CYP2A6*1/CYP2A6*1X2 (n ¼ 1). With the and one Japanese subjects who possessed the CYP2A7 standard curve, the ratios of the copy numbers of CYP2A6/ polymorphic alleles were calculated as 0.84–1.00, indicating CYP2A7 in the 14 European-American, one African-American, the presence of two copies of the CYP2A6 gene.

The Pharmacogenomics Journal Identification of novel CYP2A7 polymorphic alleles T Fukami et al 405

Figure 3 Copy number assay for CYP2A6 and CYP2A7 genes by polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) targeting exon 5. (a) Schematic structures of the CYP2A6 and CYP2A7 genes. Boxes represent exon 5 and lines represent intron 5. Polymerase chain reaction amplification was performed with the primer pair indicated by horizontal arrows. The restriction site of Taq I is indicated by a vertical arrow. (b) Representative photograph of PCR-RFLP patterns for different CYP2A6 genotypes. The intensities of the 330-bp (CYP2A6) and 189-bp (CYP2A7)fragments were quantified. The ratio of the intensities corrected by the fragment length was calculated as the ratio of the copy numbers of CYP2A6/CYP2A7.(c) Standard curve of the ratio of the copy numbers of CYP2A6/CYP2A7.(d) The ratios of the copy numbers of CYP2A6/CYP2A7 in the 14 European-American, one African-American, and one Japanese subjects who possessed the CYP2A7 polymorphic alleles were calculated using the standard curve shown in the (c).

Allele frequency of CYP2A6*4A mined. Among them, the data for 163 European-Americans The improved genotyping method for CYP2A6*4A which (Figure 5a) and 116 African-Americans (Figure 5b) whose can exclude the polymorphic CYP2A7 alleles was established genotypes were relevant to the CYP2A7 polymorphic alleles (Figure 4) and was applied for the subjects (14 European- or the CYP2A6*4A allele are shown. Among the European- Americans, three African-Americans, 73 Japanese, and 42 Americans, 12 subjects who had been mistakenly genotyped Koreans) who possessed the CYP2A6*4A allele or the CYP2A7 as CYP2A6*1/CYP2A6*4A owing to the polymorphic polymorphic alleles. The subjects possessing the CYP2A7 CYP2A7 alleles were correctly genotyped as CYP2A6*1/ polymorphic alleles were not genotyped as the CYP2A6*4A CYP2A6*1 (Figure 5a). The cotinine/nicotine ratios in these allele. Thus, no European-American subject possessed the subjects (6.473.5, 1.5–14.9, n ¼ 12) were in the range of the genuine CYP2A6*4A allele in the present study (n ¼ 187). The ratios in the homozygotes of CYP2A6*1 not possessing the allele frequencies of CYP2A6*4A in the African-Americans CYP2A7 polymorphic alleles (7.975.6, 0.6–36.5, n ¼ 115). (n ¼ 176) and Japanese (n ¼ 184) were correctly recalculated Two subjects who had been mistakenly genotyped as as 0.6 and 22.3%, respectively (Table 1). CYP2A6*4A/CYP2A6*9 or CYP2A6*4A/CYP2A6*14 owing to the polymorphic CYP2A7 alleles were correctly genotyped as In vivo nicotine metabolism in subjects possessing the CYP2A7 CYP2A6*1/CYP2A6*9 or CYP2A6*1/CYP2A6*14, respectively. polymorphic alleles The cotinine/nicotine ratios in the two subjects (4.3 and 3.7) In this study, the phenotype of nicotine metabolism in 176 were in the range of the ratios in the subjects with European-Americans and 160 African-Americans was deter- CYP2A6*1/CYP2A6*9 (5.973.2, 0.8–13.7, n ¼ 24) and the

The Pharmacogenomics Journal Identification of novel CYP2A7 polymorphic alleles T Fukami et al 406

Figure 4 Improved genotyping of CYP2A6*1A, CYP2A6*1B, CYP2A6*4A, and CYP2A6*4D alleles by the combination of allele specific-PCR (AS-PCR) and polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP). (a) Schematic structures of the CYP2A6 and CYP2A7 genes. Closed and open boxes represent the exons of CYP2A7 and CYP2A6, respectively and lines represent introns. The primers of 2A6int8F and 2A7int8F specifically anneal to the CYP2A6 and CYP2A7 genes, respectively. The amplified DNA was digested by Acc II. (b) Schematic and (c) representative AS-PCR and PCR-RFLP patterns for the different CYP2A6 genotypes. Three subjects with the CYP2A7 polymorphic allele were confirmed not to have the CYP2A6*4 allele. Subjects 1 and 2 who were previously genotyped as CYP2A6*1A/CYP2A6*4A have been correctly re-genotyped as CYP2A6*1A/ CYP2A6*1A and CYP2A6*1A/CYP2A6*1B, respectively. Subject 3 who was previously genotyped as CYP2A6*1B/CYP2A6*4A has been correctly re- genotyped as CYP2A6*1B/CYP2A6*1B. subjects with CYP2A6*1/CYP2A6*14 (6.774.5, 2.5–17.1, If the phenotype data of the subjects possessing the n ¼ 10), not possessing the CYP2A7 polymorphic alleles, polymorphic CYP2A7 alleles might be mistakenly inter- respectively. preted as the heterozygotes of the CYP2A6*4 allele, the data Among the African-Americans, a subject who had been analyses of the relationship between the genotype and mistakenly genotyped as CYP2A6*4A/CYP2A6*17 was cor- phenotype would be perverted. rectly genotyped as CYP2A6*1/CYP2A6*17 (Figure 5b). The cotinine/nicotine ratio in this subject (2.5) was in the range of the ratios in the subjects with CYP2A6*1/CYP2A6*17 not Discussion possessing the CYP2A7 polymorphic alleles (5.473.0, 1.4– 10.7, n ¼ 21). The cotinine/nicotine ratios in the subjects In the present study, we found novel CYP2A7 polymorphic with CYP2A6*1/CYP2A6*4A or CYP2A6*4A/CYP2A6*9 were alleles (CYP2A7*1B, CYP2A7*1C, and CYP2A7*1D) 4.3 and 0.9, respectively. The ratios in the homozygotes of that have a gene conversion with the CYP2A6 sequence CYP2A6*1 not possessing the CYP2A7 polymorphic alleles in the 30-flanking region and confound the genotyping were 8.075.1 (0.9–30.4, n ¼ 92). of the CYP2A6*4 allele. These alleles were found in

The Pharmacogenomics Journal Identification of novel CYP2A7 polymorphic alleles T Fukami et al 407

CYP2A6*4 was 4.0% in the French population (white subjects) by this genotyping method. The possibility cannot be excluded that the allele frequency of CYP2A6*4 might be overestimated. In the genotyping method by Oscarson et al.,8 the CYP2A6*4 allele was detected with a two-step PCR analysis. In the first PCR reaction (PCR I), the primer set of 2Aex7F and 2A6R1 was used (Figure 6b). The sense primer was designed to anneal to a common sequence located in exon 7 of both the CYP2A6 and the CYP2A7 genes, whereas the antisense primer was designed to anneal only to the sequence of the 30-flanking region in the CYP2A6 gene (Figures 1a and 6b). In the second step (PCR II), the allele specific (AS)-PCR was performed with the primer sets of 2A6ex8F or 2A7ex8F and 2A6R2. However, this genotyping method cannot distinguish between CYP2A6*4 and CY- P2A7*1D alleles, as the primer of 2A6R1 can anneal to the CYP2A7*1D allele (Figure 6b). Oscarson et al.8 reported that the allele frequencies of CYP2A6*4 in Finns (n ¼ 100) and Spaniards (n ¼ 100) were 1.0 and 0.5%, respectively. Owing to the presence of the CYP2A7*1D alleles, the allele frequencies of CYP2A6*4 in these populations might be actually lower than those values. The frequency of the CYP2A6*4 allele in Europeans remains to be determined by the improved genotyping method with a large number of subjects. In the genotyping method by Goodz and Tyndale,17 the CYP2A6*4 allele was detected by the two-step PCR analysis reported by Oscarson et al.8 with a modification of the primer. In the first PCR reaction, the primer 2A6R3 was used instead of 2A6R1 used in the method by Oscarson et al.8 The 2A6R3 primer was designed to anneal only to the sequence Figure 5 Cotinine/nicotine ratios in plasma 2 h after chewing one of the 30-flanking region in the CYP2A6 gene (Figures 1a and piece of nicotine gum. The cotinine/nicotine ratios in (a) 163 European- 6c). As the 2A6R3 primer cannot anneal to any polymorphic Americans and (b) 116 African-Americans who were genotyped for CYP2A7 alleles (Figure 6c), this genotyping method could CYP2A6 alleles. The closed circles indicate the subjects who were once identify the genuine CYP2A6*4A allele. Schoedel et al.18 genotyped as the heterozygotes of the CYP2A6*4 allele with the conventional genotyping method, but re-genotyped as not CYP2A6*4 reported that the frequency of the CYP2A6*4 allele in allele owing to the CYP2A7 polymorphic alleles. The open circles Canadian white subjects (n ¼ 1168) was 1.2% using this indicate the subjects not possessing the CYP2A7 polymorphic alleles. genotyping method. However, this genotyping method The numbers of subjects are shown in parentheses. could not distinguish between the CYP2A6*4A and CY- P2A6*4D alleles. In contrast, our improved genotyping method can distinguish the CYP2A6*4A, CYP2A6*4D, and European-Americans with a moderate frequency (0.5–1.9%), the polymorphic CYP2A7 alleles. whereas they were very rare in the African-Americans, The CYP2A6*4A allele completely lacks the enzymatic Japanese, and Koreans. By the discrimination of the CYP2A7 activity, as we previously reported with the in vivo phenotyp- alleles and genuine CYP2A6*4 allele, the previous allele ing of nicotine.19,20 The misgenotyping of the CYP2A6*4 frequencies of CYP2A6*4 of 3.7% (European-Americans, allele results in an inconsistency between the genotype and n ¼ 187), 0.9% (African-Americans, n ¼ 176), and 22.6% (Japa- phenotype. As some heterozygotes of the CYP2A6*4 allele nese, n ¼ 184) were corrected to 0, 0.6, and 22.3%, respectively. can metabolize nicotine at levels similar to homozygotes of Thus, the polymorphisms of CYP2A7 would lead to misgeno- CYP2A6*1A,19,20 the effects of the misgenotyping as hetero- typing and overestimation of the allele frequency of CY- zygotes of the CYP2A6*4A owing to the CYP2A7 polymorphic P2A6*4A, particularly with individuals of European origin. alleles on the phenotype were not dramatic (Figure 5). In the genotyping method by Ariyoshi et al.,15 the primer However, if subjects were misgenotyped as the CYP2A6*4A set of 2A6-B4 and 2A6UTR-AS1 was used (Figure 6a). The allele in combination with the other CYP2A6 alleles that are 2A6-B4 primer can anneal to exon 8 of both the CYP2A6 and known to be inactive, such as CYP2A6*2, CYP2A6*7, CYP2A7 genes. The 2A6UTR-AS1 primer is the same as the CYP2A6*10,orCYP2A6*20, it would clearly confound the 2A6R2 primer but is 2 bp longer. Therefore, the CYP2A7 relationship between the genotype and phenotype. polymorphic alleles would be also amplified (Figures 1a and Several studies have demonstrated an association between 6a). Loriot et al.16 reported that the allele frequency of CYP2A6 genetic polymorphisms and the risk of lung cancer

The Pharmacogenomics Journal Identification of novel CYP2A7 polymorphic alleles T Fukami et al 408

Figure 6 Summary of previous genotyping methods for the CYP2A6*4 allele. Closed and open boxes represent exons of the CYP2A7 and CYP2A6 genes, respectively and lines represent the introns. The horizontal arrows indicate the location and direction of the primers for the PCR. Dashed lines between the horizontal arrows represent the amplification of PCR. (a) Methods by Ariyoshi et al.15 and Nakajima et al.14 The primers of 2A6-B415 and 2Aint7F14 anneal to both the CYP2A6 and CYP2A7 genes. The primers of 2A6UTR-AS115 and 2A6R214 anneal to the CYP2A7 polymorphic alleles (CYP2A7*1B, CYP2A7*1C and CYP2A7*1D) as well as the CYP2A6 gene, resulting in the misgenotyping of CYP2A6*4.(b) Method by Oscarson et al.8 The primer of 2Aex7F anneals to both the CYP2A6 and CYP2A7 genes. The primer of 2A6R1 anneals to CYP2A7*1D allele as well as the CYP2A6 gene (PCR I). In PCR II, the primer of 2A7ex8F can anneal to the CYP2A7*1D allele as well as to the CYP2A6*4 allele. Therefore, this method also leads to misgenotyping of CYP2A6*4A.(c) Method by Goodz and Tyndale.17 The primer of 2A6R3 specifically anneals to the CYP2A6 gene, but not to the polymorphic CYP2A7 alleles. This method can distinguish between the CYP2A6*4A allele and the CYP2A7 polymorphic alleles. in the Japanese population.21–23 In contrast, no clear gene. These CYP2A7 alleles confound the genotyping association was observed in Europeans and Chinese popula- analysis of CYP2A6*4A. The CYP2A7 polymorphic alleles tions.16,24,25 In those studies, the misgenotyping of the were found in European-Americans with a moderate CYP2A6*4 might be one of the possible factors in the frequency but were very rare in African-Americans, contradiction. It is important to investigate the association Japanese, and Koreans. We must pay attention to the of the CYP2A6 genetic polymorphisms and the interindivi- CYP2A7 alleles in the genotyping of the CYP2A6 gene. dual differences in the enzymatic activity or the risk of cancer with an accurate CYP2A6 genotyping method. The comprehensive and reliable genotyping method for the Materials and methods CYP2A6*4 allele developed in the present study will be essential to improve the results of such population studies. Chemicals and reagents In conclusion, we found CYP2A7 polymorphic alleles in Taq DNA polymerase and Blend Taq DNA polymerase were which the sequences of the 30-flanking region were con- obtained from Greiner Japan (Tokyo, Japan) and TOYOBO verted with the corresponding sequences of the CYP2A6 (Osaka, Japan), respectively. Restriction enzymes were

The Pharmacogenomics Journal Identification of novel CYP2A7 polymorphic alleles T Fukami et al 409

Table 2 Primers used in the present study

Primer Sequence Location

2A6-51G 50-GCTGGTCTGCCTGACTGTG-30 Exon 1 2A6-51A 50-GCTGGTCTGCCTGACTGTA-30 Exon 1 2A6int1ASa 50-TCCTGTCTTTCTGATGCTGA-30 Intron 1 2Aex3F 50-TTCAGCAACGGGGAGCGCG-30 Exon 3 2A6/7int3AS 50-GATCGATATTGGCGCCTGCG-30 Intron 3–exon 4 2A6E5Fa 50-GATGTTCTCTTCGGTGATG-30 Exon 5 2A6/7int5AS 50-TGCTGGGATTACAGGCTGTT-30 Intron 5 2A6int8F 50-CAAGTGTACCTGGCAGGAAA-30 Intron 8 2A7int8F 50-CAAGTGTATCTGGCAAGAAG-30 Intron 8 2A6-delSb 50-TTCGCGGAAGAGGCGGGTA-30 30-UTR 2A6R1c 50-GCACTTATGTTTTGTGAGACATCAGAGACAA-30 30-Flanking region 2A7R1d 50-GCACTTATGTTTTGTGAGACATCAGATAGAG-30 30-Flanking region 2A6 FR-F 50-CTTAGAAAGTTGTCTCTGATG-30 30-Flanking region 2A7 FR-F 50-CTTAGAAAGTTGGCTCTATCT-30 30-Flanking region 2A7-REV 50-TTTACGTGAGGAATGGGTGG-3 30-Flanking region 2A6reverse 50-TAATTGGGTTGTTTTCTATTGAGT-30 30-Flanking region 2A7 FR-R 50-ATAGTGCTGTAATTACAGGTG-30 30-Flanking region aFukami et al.30 bNunoya et al.9 cOscarson et al.8 dOscarson et al.11 purchased from Fermentas (Hanover, MD, USA), New of 25 ml. After initial denaturation at 941C for 3 min, England BioLabs (Beverly, MA, USA) or Takara (Kyoto, amplification was performed by denaturation at 941C for Japan). Primers were commercially synthesized at Hokkaido 25 s, annealing at 571C for 25 s and extension at 721C for 25 s System Sciences (Sapporo, Japan). All other chemicals and for 28 cycles, followed by a final extension at 721C for 5 min. solvents were of the highest grade commercially available. The expected size of the PCR product was 285 bp.

Genomic DNA Sequence analyses of 30-flanking region of the CYP2A7 gene This study was approved by the Human Studies Committee The PCR product with the primers of 2A6-delS and 2A7 FR-R of Washington University School of Medicine (St Louis, MO, (Table 2, Figure 1a) was subcloned into pT7Blue T-vector USA) and the Ethics Committees of Kanazawa University (Novagen, Madison, WI, USA). The PCR reaction mixture (Kanazawa, Japan) and Soonchunhyang University Hospital was the same as described above except for primers. After (Chonan, Korea). We recruited 187 European-American, 176 initial denaturation at 941C for 3 min, amplification was African-American, 184 Japanese and 209 Korean subjects. performed by denaturation at 941C for 25 s, annealing at Written informed consent was obtained from all subjects. 561C for 25 s, and extension at 721C for 2 min for 35 cycles, Blood samples were collected from a cubital vein. Genomic followed by a final extension at 721C for 5 min. The plasmid DNA was extracted from peripheral lymphocytes by use of a DNA was submitted to DNA sequencing using a Long-Read Puregene DNA isolation kit (Gentra Systems, Minneapolis, Tower DNA sequencer (GE Healthcare Bio-Science, NJ, USA). MN, USA). Detection of CYP2A7 polymorphisms Genotyping of CYP2A6 alleles To detect the CYP2A7 polymorphic alleles (CYP2A7*1B, The genotyping of CYP2A6*1B,14 CYP2A6*1F,14 CYP2A7*1C and CYP2A7*1D), a PCR-RFLP method was CYP2A6*1G,14 CYP2A6*1X2,26 CYP2A6*2,27 CYP2A6*3,27 developed. The PCR reaction mixture was the same as CYP2A6*4A,14 CYP2A6*4D,14 CYP2A6*5,19 CYP2A6*6,19 described above except for the primers (2A6-delS and 2A7- CYP2A6*7,26 CYP2A6*8,26 CYP2A6*9,28 CYP2A6*10,26 REV). The primers 2A6-delS and 2A7-REV anneal to the CYP2A6*11,26 CYP2A6*12,14 CYP2A6*13,29 CYP2A6*14,29 CYP2A7 gene (Table 2 and Figure 1). After initial denatura- CYP2A6*15,29 CYP2A6*16,29 CYP2A6*17,30 CYP2A6*18,29 tion at 941C for 3 min, amplification was performed by CYP2A6*19,29 and CYP2A6*2031 was performed as described denaturation at 941C for 25 s, annealing at 581C for 25 s and previously. An AS-PCR method for the SNP of g.51G4A was extension at 721C for 30 s for 35 cycles, followed by a final developed in the present study. The primer sets were 2A6- extension at 721C for 5 min. The PCR product was digested 51G or 2A6-51A, and 2A6int1AS (Table 2). The reaction with Tai I restriction enzyme and electrophoresed on a 4% mixture contained the genomic DNA samples (100 ng), agarose gel. The wild-type of the CYP2A7 gene (CYP2A7*1A

1 Â PCR buffer, 1.5 mM MgCl2, 0.4 mM of each primer, 250 mM allele) yields 197-, 132- and 101-bp fragments; the of dNTPs and 1 U of Taq DNA polymerase in a final volume CYP2A7*1B allele yields 197-, 132- and 95-bp fragments;

The Pharmacogenomics Journal Identification of novel CYP2A7 polymorphic alleles T Fukami et al 410

the CYP2A7*1C allele yields 197-, 132- and 90-bp fragments; bp fragments by Taq I, whereas that from CYP2A6 gene was the CYP2A7*1D allele yields 197-, 128- and 90-bp fragments not digested (330 bp). These products were electrophoresed (Figure 1b). on a 2% agarose gel and visualized by ethidium bromide To distinguish between CYP2A7*1A/CYP2A7*1C and CY- staining (Figure 3b). Standard curve was made with the P2A7*1A/CYP2A7*1D, we established an AS-PCR method. genomic DNA samples described above (Figure 3c). The The PCR reaction mixture was the same as described above relative gene copy ratios of CYP2A6/CYP2A7 for European- except for the primer sets (2A7 FR-F and 2A7-REV or 2A6 FR- American subjects who were genotyped as heterozygotes of F and 2A7-REV). The 2A7 FR-F primer anneals to CYP2A7*1A the CYP2A6*4A allele were calculated using the standard and CYP2A7*1C, whereas the 2A6 FR-F primer anneals to curve. CYP2A7*1D (Table 2 and Figure 1). After initial denaturation Although there is no CYP2A7 gene in the CYP2A6*12 at 941C for 3 min, amplification was performed by denatura- allele, no subject possessed the CYP2A6*12 allele in our tion at 941C for 25 s, annealing at 561C for 25 s and study. In addition, as no duplicate or multiplicate alleles extension at 721C for 25 s for 35 cycles, followed by a final have been reported for CYP2A7 gene, the intensity of extension at 721C for 5 min. An aliquot (10 ml) of the PCR CYP2A7 fragment was used as control for two copies. As product was electrophoresed on a 4% agarose gel. The the sequences of CYP2A6 and CYP2A7 genes to which the CYP2A7*1A and CYP2A7*1C alleles were amplified with the primers anneal are completely identical, it was assumed that primer set of 2A7 FR-F and 2A7-REV (122 bp). The the PCR efficiencies for the two genes would be equal. Even CYP2A7*1D allele was amplified with the primer set of 2A6 if the PCR efficiencies might be different, it would appear in FR-F and 2A7-REV (118 bp) (Figure 1c). the slope of the standard curve. Thus, the relative gene copy ratios of CYP2A6/CYP2A7 were calculated using the stan- Determination of the relative gene copy ratio of CYP2A6/CYP2A7 dard curve. To investigate whether the subjects with the CYP2A7 polymorphic allele have two copies of the CYP2A6 gene, Improved genotyping method for the CYP2A6*4 alleles PCR analyses with the quantified genomic DNA (100 ng) To avoid the misgenotyping of the CYP2A6*4A allele owing were performed. The 2Aex3F and 2A6/7int3AS (Table 2) to the amplification of the polymorphic CYP2A7 alleles, an primers can anneal to both the CYP2A6 and CYP2A7 genes improved genotyping method was designed (Figure 4). The (Figure 2a). After an initial denaturation at 941C for 3 min, antisense primer 2A6reverse (Table 2) specifically anneals to amplification was performed by denaturation at 941C for the CYP2A6 gene and does not anneal to CYP2A7 poly- 25 s, annealing at 621C for 25 s, and extension at 721C morphic alleles. The primers of 2A6int8F and 2A7int8F for 30 s for 35 cycles, followed by a final extension at 721C (Table 2) specifically anneal to the CYP2A6 and CYP2A7 for 5 min. The PCR product derived from the CYP2A7 gene genes, respectively. The reaction mixture contained geno- was digested to 273- and 118-bp fragments by EcoNI, mic DNA (100 ng), 1 Â PCR buffer, 0.2 mM dNTPs, 0.4 mM whereas that from CYP2A6 gene was not digested (391 bp). each primer, and 0.5 U of Blend Taq DNA polymerase in a These products were electrophoresed on a 2% agarose gel final volume of 25 ml. After initial denaturation at 941C for and visualized by ethidium bromide staining (Figure 2b). 3 min, amplification was performed by denaturation at 941C The intensities of the 391-bp (CYP2A6) and 273-bp for 25 s, annealing at 541C for 25 s, and extension at 721C for (CYP2A7) fragments were quantified using ImageQuant TL 2 min for 35 cycles, followed by a final extension at 721C for (GE Healthcare Bio-Science). The ratio of the intensities 5 min. An aliquot (5 ml) of the PCR product was analyzed by corrected by the length of the fragment was calculated as the electrophoresis with 0.8% agarose gel. The CYP2A6*1A and relative gene copy ratio of CYP2A6/CYP2A7. Standard curve CYP2A6*1B alleles were amplified with the primer set of was made using genomic DNA samples from the subjects 2A6int8F and 2A6reverse (1937 and 1936 bp, respectively). who were genotyped as CYP2A6*4A/CYP2A6*4A (n ¼ 2, The CYP2A6*4A and CYP2A6*4D alleles were amplified with Japanese), CYP2A6*1/CYP2A6*4A (n ¼ 6, Japanese), the primer set of 2A7int8F and 2A6reverse (1935 and CYP2A6*1/CYP2A6*1 (n ¼ 6, Japanese), and CYP2A6*1/CY- 1936 bp, respectively). These PCR products were digested P2A6*1X2 (n ¼ 1, Korean) (Figure 2c). Using the standard with Acc II at 371C for 3 h and electrophoresed on a 2% curve, we quantified the relative gene copy ratios of agarose gel (Figure 4b and c). The PCR products derived from CYP2A6/CYP2A7 for European-American subjects who were the CYP2A6*1B and CYP2A6*4A alleles were digested by genotyped as heterozygotes of the CYP2A6*4A allele. Acc II, but those from CYP2A6*1A and CYP2A6*4D were not. To confirm the determination of the targeted exon 3, a similar analysis was performed at the different region (from In vivo phenotyping of nicotine metabolism exon 5 to intron 5). The 2A6E5F and 2A6/7int5AS (Table 1) Written informed consent was obtained from 187 healthy primers were used (Figure 3a). Both primers can anneal to European-American and 176 healthy African-American non- both the CYP2A6 and CYP2A7 genes. After an initial smokers. The phenotyping of in vivo nicotine metabolism denaturation at 941C for 3 min, amplification was per- was performed according to the method established in our formed by denaturation at 941C for 25 s, annealing at 551C previous study.19 Briefly, the subjects chewed one piece of for 25 s and extension at 721C for 30 s for 35 cycles, followed nicotine gum for 30 min, chewing for 10 s per 30 s. Blood by a final extension at 721C for 5 min. The PCR product samples were collected from a cubital vein just before and derived from the CYP2A7 gene was digested to 189- and 141- 2 h after the start of chewing. The concentrations of nicotine

The Pharmacogenomics Journal Identification of novel CYP2A7 polymorphic alleles T Fukami et al 411

and cotinine in the plasma samples were determined entire coding region of CYP2A6 in Japanese. Pharmacogenetics 2004; 14: by high-performance liquid chromatography as described 701–705. ´ previously.32 The cotinine/nicotine ratio of the plasma 11 Oscarson M, McLellan RA, Gullsten H, Yue QY, Lang MA, Bernal ML et al. Identification and characterization of novel polymorphisms in the concentration was calculated as an index of nicotine CYP2A locus: implications for nicotine metabolism. FEBS Lett 1999; 460: metabolism. As nicotine and cotinine were detected in the 321–327. plasma before chewing nicotine gum, 11 European- 12 Rao Y, Hoffmann E, Zia M, Bodin L, Zeman M, Sellers EM et al. Americans and 16 African-Americans were judged as Duplications and defects in the CYP2A6 gene: identification, genotyping, and in vivo effects on smoking. Mol Pharmacol 2000; 58: smokers. Therefore, the phenotype data from 176 European- 747–755. Americans and 160 African-Americans were analyzed along 13 Oscarson M, McLellan RA, Asp V, Ledesma M, Ruiz MLB, Sinues B et al. with the CYP2A6 genotype data. Characterization of a novel CYP2A7/CYP2A6 hybrid allele (CYP2A6*12) that causes reduced CYP2A6 activity. Hum Mutat 2002; 20: 275–283. 14 Nakajima M, Yoshida R, Fukami T, McLeod HL, Yokoi T. Novel human Acknowledgments CYP2A6 alleles confound gene deletion analysis. FEBS Lett 2004; 569: 75–81. 15 Ariyoshi N, Takahashi Y, Miyamoto M, Umetsu Y, Daigo S, Tateishi T This study was supported in part by a grant from the Japan Health et al. Structural characterization of a new variant of the CYP2A6 gene Sciences Foundation with the Research on Health Science Focusing (CYP2A6*1B) apparently diagnosed as heterozygotes of CYP2A6*1A and on Drug Innovation, by an SRF Grant for Biomedical Research in CYP2A6*4C. Pharmacogenetics 2000; 10: 687–693. Japan and by Philip Morris Incorporated. The enthusiasm and 16 Loriot MA, Rebuissou S, Oscarson M, Cene´e S, Miyamoto M, Ariyohi N research support of Tracy Jones, RN, Arnita Pitts, RN, Phyllis Klein, et al. Genetic polymorphisms of cytochrome P450 2A6 in a case–control RN, and Ladonna Gaines, Washington University Center for Clinical study on lung cancer in a French population. Pharmacogenetics 2001; Studies, and Margaret Ameyaw, MD are greatly appreciated. We 11: 39–44. 17 Goodz SD, Tyndale RF. Genotyping human CYP2A6 variants. Methods acknowledge Mr Brent Bell for reviewing the manuscript. Enzymol 2002; 357: 59–69. 18 Schoedel KA, Hoffmann EB, Rao Y, Seller EM, Tyndale RF. Ethnic variation in CYP2A6 and association of genetically slow nicotine DUALITY OF INTEREST metabolism and smoking in adult Caucasians. Pharmacogenetics 2004; 14: 615–626. None declared. 19 Nakajima M, Kwon J-T, Tanaka N, Zenta T, Yamamoto Y, Yamamoto H et al. Relationship between interindividual differences in nicotine metabolism and CYP2A6 genetic polymorphism in humans. Clin References Pharmacol Ther 2001; 68: 72–78. 20 Kwon JT, Nakajima M, Chai S, Yom YK, Kim HK, Yamazaki H, Sohn DR 1 Nakajima M, Yamamoto T, Nunoya K-I, Yokoi T, Nagashima K, Inoue K et al. Nicotine metabolism and CYP2A6 allele frequencies in Koreans. et al. Role of human cytochrome P4502A6 in C-oxidation of nicotine. Pharmacogenetics 2001; 11: 317–323. Drug Metab Dispos 1996; 24: 1212–1217. 21 Miyamoto M, Umetsu Y, Dosaka-Akita H, Sawamura Y, Yokota J, 2 Yamano S, Tatsuno J, Gonzalez FJ. The CYP2A3 gene product catalyzes Kunitoh H et al. CYP2A6 gene deletion reduces susceptibility to lung coumarin 7-hydroxylation in human liver microsomes. Biochemistry cancer. Biochem Biophys Res Commun 1999; 261: 658–660. 1990; 29: 1322–1329. 22 Ariyoshi N, Miyamoto M, Umetsu Y, Kunitoh H, Dosaka-Akita H, 3 Komatsu T, Yamazaki H, Shimada N, Nakajima M, Yokoi T. Roles of Sawamura Y et al. Genetic polymorphism of CYP2A6 gene and tobacco- cytochrome P450 1A2, 2A6 and 2C8 in 5-fluorouracil formation from induced lung cancer risk in male smokers. Cancer Epidemiol Biomarkers tegafur, an anticancer prodrug, in human liver microsomes. Drug Metab Prev 2002; 11: 890–894. Dispos 2000; 28: 1457–1463. 23 Fujieda M, Yamazaki H, Saito T, Kiyotani K, Gyamfi MA, Sakurai M et al. 4 Torchin CD, McNeilly PJ, Kapetanovic IM, Strong JM, Kupferberg HJ. Evaluation of CYP2A6 genetic polymorphisms as determinants of Stereoselective metabolism of a new anticonvulsant drug candidate, smoking behavior and tobacco-related lung cancer risk in male losigamone, by human liver microsomes. Drug Metab Dispos 1996; 24: Japanese smokers. Carcinogenesis 2004; 25: 2451–2458. 1002–1008. 24 Tan W, Chen GF, Ding DY, Song CY, Kadlubar FF, Lin DX. Frequency of 5 Yamazaki H, Murai K, Kawai Y, Kamataki T. Effect of the CYP2A6 CYP2A6 gene deletion and its relation to risk of lung and esophageal polymorphisms on letrozole oxidation in human liver microsomes. cancer in the Chinese population. Int J Cancer 2001; 95: 96–101. Abstracts of the 125th Meeting of the Pharmaceutical Society of Japan, 25 Wang H, Tan W, Hao B, Miao X, Zhou G, He F et al. Substantial Pharmaceutical Society of Japan, Vol. 3. Pharmaceutical Society of Japan, reduction in risk of lung adenocarcinoma associated with genetic Tokyo, 2005 p 117. polymorphism in CYP2A13, the most active cytochrome P450 for the 6 Tiano HF, Hosokawa M, Chulada PC, Smith PB, Wang RL, Gonzalez FJ metabolic activation of tobacco-specific carcinogen NNK. Cancer Res et al. Retroviral mediated expression of human cytochrome P450 2A6 2003; 63: 8057–8061. in C3H/10T1/2 cells confers transformability by 4-(methylnitrosamino)- 26 Yoshida R, Nakajima M, Watanabe Y, Kwon J-T, Yokoi T. Genetic 1-(3-pyridyl)-1-butanone (NNK). Carcinogenesis 1993; 14: 1421–1427. polymorphisms in human CYP2A6 gene causing impaired nicotine 7 Ding S, Lake BG, Friedberg T, Wolf CR. Expression and alternative metabolism. Br J Clin Pharmacol 2002; 54: 511–517. splicing of the cytochrome P-450 CYP2A7. Biochem J 1995; 306: 161– 27 Nakajima M, Yamagishi S, Yamamoto H, Yamamoto T, Kuroiwa Y, Yokoi 166. T. Deficient cotinine formation from nicotine is attributed to the whole 8 Oscarson M, McLellan RA, Gullste´n H, Yue QY, Lang MA, Bernal ML deletion of the CYP2A6 gene in humans. Clin Pharmacol Ther 2000; 67: et al. Characterization and PCR-based detection of a CYP2A6 gene 57–69. deletion found at a high frequency in a Chinese population. FEBS Lett 28 Yoshida R, Nakajima M, Nishimura K, Tokudome S, Kwon J-T, Yokoi T. 1999; 448: 105–110. Effects of polymorphism in promoter region of human CYP2A6 9 Nunoya K, Yokoi T, Takahashi Y, Kimura K, Kinoshita M, Kamataki gene (CYP2A6*9) on expression level of messenger ribonucleic acid T. Homologous unequal cross-over within the human CYP2A gene and enzymatic activity in vivo and in vitro. Clin Pharmacol Ther 2003; 74: cluster as a mechanism for the deletion of the entire CYP2A6 gene 69–76. associated with the poor metabolizer phenotype. J Biochem 1999; 126: 29 Fukami T, Nakajima M, Higashi E, Yamanaka H, Sakai H, McLeod HL 402–407. et al. Characterization of novel CYP2A6 polymorphic alleles (CYP2A6*18 10 Ariyoshi N, Sekine H, Nakayama K, Saito K, Miyamoto A, Kamataki T. and CYP2A6*19) that affect enzymatic activity. Drug Metab Dispos 2005; Identification of deletion-junction site of CYP2A6*4B allele lacking 33: 1202–1210.

The Pharmacogenomics Journal Identification of novel CYP2A7 polymorphic alleles T Fukami et al 412

30 Fukami T, Nakajima M, Yoshida R, Tsuchiya Y, Fujiki Y, Katoh M et al. A a truncated protein lacking enzymatic activity. Biochem Pharmacol novel polymorphism of human CYP2A6 gene CYP2A6*17 has an amino 2005; 70: 801–808. acid substitution (V365M) that decreases enzymatic activity in vitro and 32 Nakajima M, Yamamoto T, Kuroiwa Y, Yokoi T. Improved highly in vivo. Clin Pharmacol Ther 2004; 76: 519–527. sensitive method for determination of nicotine and cotinine in human 31 Fukami T, Nakajima M, Higashi E, Yamanaka H, McLeod HL, Yokoi T. A plasma by high-performance liquid chromatography. J Chromatogr B novel CYP2A6*20 allele found in African-American population produces 2000; 742: 211–215.

The Pharmacogenomics Journal