Introduction

Cytosolic sulfotransferases (SULTs) play an important role in the phase II metabolism of several xenobiotics. Sulfation can result in either inactivation or the metabolic activation of sulfate acceptor substrates [1, 2]. Functional polymorphisms of the SULT1As may influence individual susceptibility to adverse drug reactions involving drug substrates and to chemical carcinogenesis [3–6]. The human SULT1A subfamily contains several members, which share over 92% identity at the amino acid level [7–9]. SULT1A1 is an important enzyme in xenobiotic metabolism because it has broad substrate specificity with a high affinity for many compounds and it is expressed widely in the body, particularly in the liver, brain, jejunum, and platelets [10–12]. Importantly, SULT1A1 is subject to a common functional polymorphism [13, 14], with the variant allele (SULT1A1*2, allele frequency 32% in Caucasians) encoding an allozyme with a single amino acid change (Arg213His) that is uniformly associated with low SULT activity and low thermal stability [15]. Another single amino acid change, Met223Val, in SULT1A1*3 is also commonly detected in several ethnic groups, especially African-Americans [14, 15]. Like SULT1A1, SULT1A2 is highly polymorphic, and several polymorphisms have been detected in humans [9, 14, 16]. SULT1A2*2 differs from SULT1A2*1 at two amino acid residues (Ile7Thr and Asn235Thr) [15]. The change at codon 235 appears to be particularly important functionally. Threonine in that position drastically reduces the affinity for the substrate 4-nitrophenol [15–17].

It has been reported that the gene SULT1A2 is usually in linkage disequilibrium with SULT1A1 and that the high-activity alloenzymes (SULT1A1*1 and SULT1A2*1) are associated in several ethnic groups [18, 19]. Moreover, the genes for the low-affinity alloenzymes (SULT1A1*2 and SULT1A2*2) are also associated.

In this study, we determined the allele frequencies of SULT1A1 and SULT1A2 using polymerase chain reaction (PCR)-based restriction fragment length polymorphism (RFLP) assays designed to detect common variant alleles of SULT1A1 and SULT1A2. We then assessed the linkage disequilibrium between these two genes in a Korean population. Finally, we compared the results with those for other ethnic populations.

Materials and methods

Study population

After written informed consent was obtained from participants, blood samples were collected from 234 healthy Korean subjects. Genomic DNA was isolated from peripheral blood using the QIAamp Blood Kit (Qiagen, Chatsworth, CA, USA). All the samples were collected under conditions that had been reviewed and approved by the Institutional Review Board of Gil Medical Center, Gachon Medical School, Incheon, Korea.

Identification of SULT1A1 and SULT1A2 polymorphisms

The presence of alleles of both SULT1A1*2 or 1A1*3 and SULT1A2*2 or 1A2*3 was determined by PCR-RFLP analysis as previously described, with a slight modification [20]. HhaI, Hsp92II, BstEII, and BstU1 were used as restriction enzymes for SULT1A1*2 or SULT1A1*3 and SULT1A2*2 or SULT1A2*3, respectively.

Statistical analysis

Statistical analysis was performed with SPSS version 12.0 (SPSS, IL, USA), using a chi-square (χ2) test. Linkage analysis was performed with the Estimating Haplotype Frequency Test using the EH program (http://linkage.rockefeller.edu/soft/) described by Terwilliger and Ott [21] with the population-based sampling method. D′ values, a standard measure of linkage disequilibrium, were calculated as D/D max , where D is the coefficient of linkage disequilibrium, indicating the amount of disequilibrium (−0.25< D<0.25) [22]. Because D depends on allele frequencies, the standardized value D′ is preferred [23]. D′ values can vary from +1 when alleles at the two loci studied are maximally positively associated (that is, always occurring together in the same haplotype) to −1 when the alleles are maximally negatively associated (that is, never occurring together in the same haplotype).

Results

The SULT1A1 and SULT1A2 genotypes were determined in 234 healthy Korean subjects (195 males, 39 females). The observed allelic frequencies of SULT1A1*1 and SULT1A1*2 were 0.876 and 0.124, respectively. However, we found no subject carrying the SULT1A1*3 allele in this population. The allelic frequencies of SULT1A2*1 and SULT1A2*2 were 0.885 and 0.115, respectively, but no subject carried the SULT1A2*3 allele (Table 1). The expected allelic frequencies of SULT1A1 and SULT1A2 estimated at Hardy–Weinberg equilibrium were quite similar to the observed distributions in the population (χ2 =4.475, P=0.107 for SULT1A1, χ2 =3.378, P=0.185 for SULT1A2). There were no significant gender differences in allele frequencies for SULT1A1 or SULT1A2 (χ2 =0.0269, P=0.870 for SULT1A1, χ2 =0.0031, P=0.956 for SULT1A2).

Table 1 Allele frequencies of SULT1A1 and SULT1A2 determined in the present study, compared with those from other studies (data in parentheses represent 95% confidence intervals)
Full size table

We estimated the allele frequencies of SULT1A1 and SULT1A2 to determine whether the two alleles mapped together or independently. The results showed that the observed subject numbers were quite similar to the expected numbers, assuming that the two genes are associated (Table 2). The statistical hypothesis that these alleles are independently distributed was rejected (χ2 =159.2, P<0.001), indicating that the two genes are strongly linked. Furthermore, the calculated value of D′, the coefficient of linkage disequilibrium, showed a strong positive linkage between SULT1A1*2 and SULT1A2*2 (D′=0.8919, χ2 =343.24, P=0.0034) in this population.

Table 2 Observed distributions of the combined SULT1A1 and SULT1A2 genotypes: comparison of observed numbers with the expected numbers, assuming an independent distribution of the alleles or assuming linkage
Full size table

Discussion

In this study we compared the frequencies of all six common SULT1A1 and SULT1A2 alleles that occur in the Korean population with the frequencies of those same alleles in other ethnic groups, and we observed striking ethnic differences compared with African-American and Caucasian populations [19]. However, the allele frequencies for SULT1A1 and SULT1A2 were similar to those of Chinese and Japanese populations [19]. We could not find any allele of SULT1A1*3 or SULT1A2*3 in this population. Similarly, other Asian ethnic groups, including the Chinese and Japanese, had very low frequencies of these alleles, whereas African-Americans had higher frequencies (0.23 for SULT1A1 and 0.11 for SULT1A2) [19]. Even though we found no subject with the allele of SULT1A1*3 or SULT1A2*3 in this population, the small sample size might have limited the detection of these alleles. Therefore, the study should be conducted with larger populations.

In this study, the allele frequencies of SULT1A2*1 and *2 were very similar to those of the SULT1A1*1 and *2 alleles. Because the expected allele frequencies of SULT1A1 and SULT1A2 estimated at Hardy–Weinberg equilibrium was quite similar to the observed distribution, we assessed whether SULT1A1 and SULT1A2 are associated by estimating their haplotype frequencies [24]. The frequencies of the alleles SULT1A1*1 (0.876) and 1A2*1 (0.885) and of the haplotype SULT1A1*1/1A2*1 (0.752) were comparable, suggesting that nearly all SULT1A1*1 alleles (97.1%) are associated with SULT1A2*1 and, conversely, that most SULT1A1*2 alleles (84.5%) are associated with SULT1A2*2. On the other hand, the frequencies of the other two haplotypes, SULT1A1*1/1A2*2 and SULT1A1*2/1A2*1, were very low in the Korean population (3.9 and 2.1%, respectively). The association between SULT1A1 and SULT1A2 is consistent with previous data reported [19]. Furthermore, the calculated D′ value is close to +1.0 (D′=0.8919), indicating that the two genes are strongly and positively linked. This value is similar to that for the Chinese population but higher than those for the Caucasian and African-American populations [19]. The SULT1A genes, including SULT1A1 and SULT1A2, are located in the chromosomal region 16p11.2–12.1 [25]. The encoded proteins have the same number of amino acid residues (295 amino acids) and 93–96% identity in the amino acid sequences, but they differ substantially in their substrate specificity and regulation [26]. In addition, a previous study in which the SULT1A1 gene was cloned suggested that SULT1A1 and SULT1A2 are approximately 45 kb apart and oriented head-to-tail [25]. These findings may explain the strong positive linkage disequilibrium of the two genes. However, until now, there has been no clear demonstration of this association at a mechanistic level.

SULT1A1 and SULT1A2 are the focus of intense research in the fields of cancer, drug metabolism, and pharmacogenetics because genetic variation may predispose carriers to lung cancers [27], protect against colorectal cancers [4], and affect the age of onset of breast cancer [5]. In the case of drug metabolism, SULT1A1 and SULT1A2 are involved in the sulfation of various drugs and chemicals [28, 29]. Rossi et al. [30] have shown that the sulfation rate for 4-nitrophenol, a substrate of SULT1A, by isoenzyme SULT1A1*2/*2 was significant lower than that by that of SULT1A1*1/*1 or *1/*2. Therefore, it would be important to determine the difference in biochemical properties based on the genetic polymorphisms of SULT1A1 and SULT1A2 in humans. As mentioned above, the allele frequencies in this population are very close to those observed in other Asian populations but are fairly different from those observed in Caucasians and African-Americans. Therefore, sulfation of certain substance may be clinically significantly different between these populations and thus affect the occurrence of sulfation-related toxicities and diseases. However, little is known of the roles of the genetic polymorphisms of these two genes, and future research should be directed to that end.

In summary, we have demonstrated that SULT1A1*2 and SULT1A2*2 are major allelic variants but that SULT1A1*3 and SULT1A2*3 are rare in the Korean population. Furthermore, SULT1A1 and SULT1A2 are strongly and positively linked in this population.