Cytochromes P450 (P450s) comprise a superfamily of metabolic enzymes that play important roles in the oxidative metabolism of xenobiotics and endogenous substrates (Gonzalez and Gelboin, 1994). The human CYP2C subfamily is composed of four isoforms (CYP2C8, CYP2C9, CYP2C18, and CYP2C19) that account for about 20% of the total human adult liver P450 contents (Shimada et al., 1994). Among the CYP2C subfamily isoforms, CYP2C9 and CYP2C19 play critical roles in the metabolism of clinically used drugs (Goldstein and de Morais, 1994). It has been reported that the expression level of the CYP2C9 gene in the human liver is about 20 times higher than that of the CYP2C19 gene (Furuya et al., 1991; Romkes et al., 1991; Inoue et al., 1997), indicating that there are some differences between the regulatory mechanisms of CYP2C9 and CYP2C19 gene transcriptions. It has been reported that pregnane X receptor, constitutive androstane receptor, glucocorticoid receptor, and hepatocyte nuclear factor 3γ participate in the basal expression of CYP2C9 and CYP2C19 genes (Ferguson et al., 2002; Gerbal-Chaloin et al., 2002; Chen et al., 2003; Bort et al., 2004). However, the factor(s) responsible for the difference between the expression levels of CYP2C9 and CYP2C19 genes is still unclear.

Hepatocyte nuclear factor 4α (HNF4α) is a member of the nuclear receptor superfamily (Sladek et al., 1990) and is expressed at high levels in the liver, kidney, pancreas, and small intestine (Sladek et al., 1990; Thomas et al., 2001). HNF4α appears to be an important factor for liver differentiation and function because it is involved in regulation of the expression of numerous liver-enriched genes, such as those related to glucose or lipid metabolism (Watt et al., 2003), those related to synthesis of blood coagulation factors (Sladek and Seidel, 2001), and drug-metabolizing enzymes, including CYP3A4, CYP2A6, CYP2C9, and CYP2D6 (Jover et al., 2001). It is thought that HNF4α binds to a specific DNA sequence called a direct repeat 1 (DR1) element as a homodimer to stimulate transcription of these genes (Cairns et al., 1996; Tirona et al., 2003; Pitarque et al., 2005). However, HNF4α does not always transactivate all the genes that have a DR1 element. For example, it has been reported that a DR1 element exists in the CYP2C18 promoter but that HNF4α does not bind to the DR1 element of the CYP2C18 gene and does not transactivate this promoter (Ibeanu and Goldstein, 1995).

It has been reported that there are two DR1 elements in the promoter region of the CYP2C9 gene, and HNF4α can activate the transcription of this gene via the DR1 element (Ibeanu and Goldstein, 1995; Chen et al., 2005). We also identified two DR1 elements in the CYP2C19 promoter, but it is not clear whether these elements are functional. Therefore, to clarify the mechanism determining the difference between the expression levels of CYP2C9 and CYP2C19 genes, we hypothesized that HNF4α contributes to the difference between the expression levels of CYP2C9 and CYP2C19 in the human liver. The 5′-flanking regions from -2 kilobase pairs (kbp) to the translation start site of these genes were analyzed by electrophoretic mobility shift assays (EMSAs), cotransfection assays, mutagenesis, and chromatin immunoprecipitation (ChIP) assays. Our results suggest that HNF4α participates in the regulation of CYP2C9 gene transcription but not in that of the CYP2C19 gene despite the fact that the same DR1 elements exist in both gene promoters.

Materials and Methods

EMSAs and Supershift Assays. EMSAs were performed using double-stranded DNA labeled with [γ-³²P]dATP (GE Healthcare Bio-Sciences, Piscataway, NJ) and 10 μg of the nuclear extracts as described previously (Furihata et al., 2004). The following is the sequence of the oligonucleotides used as probes, wild-type, or mutated specific cold competitors: 5′-ACAAGACCAAAGGACATTT-3′ for the DR1-A WT, 5′-ACACCCCCAAAGGACATTT-3′ for the DR1-A MT, 5′-AGTGGGTCAAAGTCCTTTC-3′ for the DR1-B WT, 5′-AGTCCCTCAAAGTCCTTTC-3′ for the DR1-B MT, 5′-TCGAGCGCTGGGCAAAGGTCACCTGC-3′ for the HNF4 WT, and 5′-TCGAGCGCTAGGCACCGGTCACCTGC-3′ for the HNF4 MT. Only the sequences of the sense strands are displayed above, and mutated nucleotides are underlined. Nuclear extracts were prepared from HepG2 cells by using a CelLytic Nuclear Extraction Kit (Sigma-Aldrich, St. Louis, MO) according to the manufacturer's protocol. After extracting nuclear contents, the protein concentration was determined by using a Bio-Rad Dc Protein Assay Kit (Bio-Rad Laboratories, Hercules, CA). The nuclear extracts were stored at -80°C until used.

For competition experiments, unlabeled competitive double-stranded DNA was added to the binding reaction mixture at a 50-fold excess of the probe amount before addition of the probe. For supershift assays, either 2 μg of IgG against HNF4α (2ZK9218H; Perseus Proteomics, Tokyo, Japan) or control mouse IgG (sc-2025; Santa Cruz Biotech, Santa Cruz, CA) was added to the binding reaction mixture at room temperature for 30 min before addition of the probe.

Plasmids. The 5′-flanking regions of the CYP2C9 and CYP2C19 genes were isolated by polymerase chain reaction (PCR) with the common sense primer 5′-ACCTCTAGATTGCTTTTCTTTGCCCTGTAT-3′ (for CYP2C9 and CYP2C19) and the antisense primer 5′-GAGGACCTGAAGCCTTCTCTTCTTGTTA-3′ (for CYP2C9) or 5′-GGGGACCTGAAGCCTTCTCCTCTTGTTA-3′ (for CYP2C19) using human genomic DNA as a template. The amplicons were subcloned into a pGEM-T-easy vector (Promega, Madison, WI). After XbaI and BamHI digestion, the fragment was ligated into a pGL3-basic vector (Promega). These constructs are hereafter referred to as 2C9 -2k and 2C19 -2k, respectively. The nucleotide sequences were determined using a Dye Terminator Cycle Sequencing-Quick Start Kit (Beckman Coulter, Fullerton, CA) and a CEQ 2000 DNA Analysis System (Beckman Coulter). Fourteen deletion constructs were generated by nested PCR of the primary clone using the following sense primers: 5′-TCTCTAGAGGTTAATCTAAATCTAAGAATTCA-3′ (2C9 -380 and 2C19 -380), 5′-ATTTCTAGAGCATCAGATTATTTACTTCA-3′ (2C9 -340), 5′-ATTACGCGTGCATCAGATTGTTTACTTCA-3′ (2C19 -340), 5′-TCTAGAGTGCTCTCAATTATGATGGTG-3′ (2C9 -320), 5′-TCTAGACAGTGCTCTCAATTATGAC-3′ (2C19 -320), 5′-TTTTCTAGAAATACCTAGGCTCCAACCAAGT-3′ (2C9 -255), 5′-TCTAGAATTACCAATACCTAGGCTTCAA-3′ (2C19 -255), 5′-ATACGCGTAAGGAGAACAAGACCAAAGGAC-3′ (2C9 -195 and 2C19 -195), 5′-TTTCTAGATATCAGTGGGTCAAAGTCCT-3′ (2C9 -160 and 2C19 -160), and 5′-ATCTAGATTTCAGAAGGAGCATATAGT-3′ (2C9 -140 and 2C19 -140). The antisense primer used was the same as that used in genome cloning. The obtained 5′-deletion fragments except for 2C9 -195, 2C19 -340, and 2C19 -195 were transferred into the pGL3-basic vector as described above. 2C9 -195, 2C19 -340, and 2C19 -195 were inserted into the pGL3-basic vector by MluI and BamHI digestion. All the constructs are named as shown in parentheses.

The cDNA clone of mouse HNF4α2 was isolated from mouse liver cDNA by PCR amplification and was subcloned into pTARGET mammalian expression vector (Promega) by EcoRI digestion, resulting in pHNF4α2 as described elsewhere (Furihata et al., 2006).

Site-Directed Mutagenesis. Site-directed mutagenesis was carried out as described elsewhere (Furihata et al., 2004). To introduce mutations into the reporter plasmids, complementary primers harboring a few mutations were designed for each target site as follows: 5′-GGAGAACAAGACCT _ _GGACATTTTATTTTTATCTGTATCAGTGGG-3′ and 5′-CCCACTGATACAGATAAAAATAAAATGTCCA _ _GGTCTTGTTCTCC-3′ for the CYP2C9 DR1-Amt; 5′-CTGTATCAGTGGGTCT _ _GTCCTTTCAGAAGGAGCATATAGTGG-3′ and 5′-CCACTATATGCTCCTTCTGAAAGGACA _ _GACCCACTGATACAG-3′ for the CYP2C9 DR1-Bmt; 5′-CGAAGGAGAACAAGACCT _ _GGACATTTTATTTTTATCTCTATCAGTGG-3′ and 5′-CCACTGATAGAGATAAAAATAAAATGTCCA _ _GGTCTTGTTCTCCTTCG-3′ for the CYP2C19 DR1-Amt; 5′-CTCTATCAGTGGGTCT _ _GTCCTTTCAGAAGGAGCATATAGTGGG-3′ and 5′-CCCACTATATGCTCCTTCTGAAAGGACA_ _GACCCACTGATAGAG-3′ for the CYP2C19 DR1-Bmt. The mutagenic sites are underlined, and spaces indicate deletions of 2-bp nucleotides.

Cell Culture, Transient Transfection, and Dual Luciferase Assay. FLC7 cells (Kawada et al., 1998), a human hepatocellular carcinoma cell line, were provided by Dr. S. Nagamori (Kyorin University, Tokyo, Japan). FLC7 cells were maintained at 37°C with 5% CO₂ in Dulbecco's modified Eagle's medium/F-12 (Invitrogen, Carlsbad, CA) supplemented with 10% heat-inactivated fetal bovine serum and 50 U/ml penicillin and 50 μg/ml streptomycin.

FLC7 cells were plated at a density of 1.8 × 10⁵ cells/well in 24-well plates 1 day before transfection. The reporter plasmids (200 ng/well) were cotransfected with pHNF4α (100 ng/well) and phRL-TK vector (Promega, 4 ng/well) into FLC7 cells by TransIT-LT1 (Mirus, Madison, WI). Twenty-four hours after transfection, luciferase reporter activities were measured as described previously (Kobayashi et al., 2004). The Renilla luciferase activity derived from the control plasmid phRL-TK was used to normalize the results of the firefly luciferase activity of reporter plasmids. Experiments were performed in triplicate, and each value is the mean ± S.D. from three or four separate assays.

ChIP Assays. ChIP assays were performed by using a ChIP-IT kit (Active Motif, Carlsbad, CA) according to the manufacturer's protocol. Human liver (from a 56-year-old Caucasian male) was supplied by the National Disease Research Interchange (Philadelphia, PA) through HAB Research Organization (Tokyo, Japan), and this study was approved by the Ethics Committee of Chiba University (Chiba, Japan). The human liver tissue (2.4 g) was isolated and chopped on ice and then cross-linked by 1% formaldehyde for 12 min. Cross-linking was stopped by the addition of glycine solution. The chromatin was sheared by using an ultrasonic disruptor UD-201 (TOMY SEIKO, Tokyo, Japan) at 25% power with 14 pulses. Nine micrograms of the sheared chromatin was immunoprecipitated with either control mouse IgG or anti-HNF4α IgG (2ZH1415H). After incubating for 4 h at 4°C with gentle rotation, salmon sperm DNA/protein G agarose was added to the mixture, and it was further incubated for 1.5 h under the same conditions. The DNA fragment was purified and used as a template for PCR. The DNA sequences around DR1 elements of the CYP2C9 and CYP2C19 genes were amplified by using the sense primers 5′-CAACCAAGTACAGTGAAACTG-3′ (for CYP2C9) and 5′-CAGAATGTACAGAGTGGGCAC-3′ (for CYP2C19) and the antisense primers 5′-TAACACTCCATGCTAATTCGG-3′ (for CYP2C9) and 5′-AACACTCCATGCTAATTAAGT-3′ (for CYP2C19). The specificity of the CYP2C9 and CYP2C19 primers was verified by the lack of amplification from sheared genomic DNA than the intended target. PCR conditions were as follows: 94°C for 2 min, followed by 94°C for 30 s, 47°C (for CYP2C9) or 50°C (for CYP2C19) for 30 s, and 72°C for 30 s, 40 cycles. The amplicons were visualized by ethidium bromide staining, and the sequence of each amplicon was confirmed by direct DNA sequence.

Determination of mRNA Levels. To measure the CYP2C9 and CYP2C19 mRNA levels, cDNA prepared from total RNA of the same human liver used for ChIP assays was subjected to quantitative real-time PCR with an ABI Prism 7000 Sequence Detection System (Applied Biosystems, Foster City, CA). The mRNA levels of CYP2C9 and CYP2C19 were determined by using Gene Expression Assays (Applied Biosystems) gene expression products for CYP2C9 and CYP2C19, respectively. The mRNA levels were normalized against glyceraldehyde-3-phosphate dehydrogenase mRNA determined by PreDeveloped TaqMan Assay Reagents for glyceraldehyde-3-phosphate dehydrogenase (Applied Biosystems).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Nucleotide sequences of the promoter regions of the CYP2C9 and CYP2C19 genes. Nucleotides are arbitrarily numbered in negative numbers from the ATG coding for the initiation codon (+1). Differences in nucleotide sequence are highlighted in bold letters, and putative HNF4α binding sites (DR1-A and DR1-B) are indicated by arrows.

Statistical Analyses. Data are presented as mean ± S.D. The p values for each experimental comparison were determined using Student's t test.

Results

Identification of Two DR1 Elements in the CYP2C19 Gene. A comparison of the 5′-flanking regions of the CYP2C9 and CYP2C19 genes is shown in Fig. 1. The 5′-flanking region from -2 kbp to the translation start site of the CYP2C9 gene was 88.8% identical to that of the CYP2C19 gene. We searched for the DR1 element in the CYP2C19 promoter by using a searching program for nuclear receptor binding sites (http://www.nubiscan.unibas.ch/; Podvinec et al., 2002) and found two putative DR1 elements (score, ≥0.75). No other DR1 element was identified with this score in this region. The upstream and downstream elements identified in the CYP2C19 promoter had the same sequences as those of two DR1 elements of the CYP2C9 promoter to which it has been reported that HNF4α can bind (Ibeanu and Goldstein, 1995; Chen et al., 2005). The upstream elements and the downstream elements in both genes are hereafter referred to as the DR1-A element and the DR1-B element, respectively.

Binding of HNF4α to the DR1-A and DR1-B Elements of the CYP2C9 and CYP2C19 Promoters in Vitro. EMSAs were performed to examine whether HNF4α could bind to the DR1-A and DR1-B elements in the CYP2C9 and CYP2C19 promoters. We used a nuclear extract prepared from HepG2 cells because it has been reported that this cell line endogenously expressed HNF4α (Ihara et al., 2005; Furihata et al., 2006). As shown in Fig. 2A, specific protein-DNA complexes were formed when the radiolabeled probe containing either the DR1-A element (DR1-A WT) or the DR1-B element (DR1-B WT) was incubated with HepG2 nuclear extracts (lanes 3 and 9, respectively). These complexes migrated at the same position as that of the one formed with the radiolabeled probe of HNF4α consensus (HNF4 WT, lane 1). The formation of the complexes was eliminated by the addition of self-competitors (DR1-A WT, lane 4; DR1-B WT, lane 10) or unlabeled HNF4 WT (lanes 6 and 12). However, complex formation was not inhibited in the presence of mutated competitors (DR1-A MT, lane 5; DR1-B MT, lane 11; HNF4 MT, lanes 7 and 13).

To determine the proteins forming these complexes, supershift assays were performed using IgG against HNF4α. The addition of anti-HNF4α IgG to the mixture resulted in generation of a supershifted band when either DR1-A WT or DR1-B WT was used as a probe (Fig. 2B, lanes 3 and 7, respectively). Control mouse IgG, used as a negative control, did not affect the formation of any complexes (lanes 4 and 8). These results indicate that HNF4α can bind to both the DR1-A and DR1-B elements of the CYP2C9 and CYP2C19 promoters.

Effects of HNF4α on the Transcriptional Activity of the CYP2C9 and CYP2C19 Promoters in FLC7 Cells. Cotransfection analyses were performed by using human hepatocarcinoma FLC7 cells to examine whether HNF4α played different roles in the transactivation of the CYP2C9 and CYP2C19 promoters. We have determined that this cell line does not express endogenous HNF4α (Furihata et al., 2006). Several constructs containing various lengths of the CYP2C9 promoter region and the HNF4α expression vector were cotransfected into FLC7 cells (Fig. 3, left). The levels of the transcriptional activities of the five constructs containing two DR1 elements (2C9 -2k, 2C9 -380, 2C9 -340, 2C9 -320, and 2C9 -255) in the presence of HNF4α were increased to 4.9-, 2.4-, 4.2-, 4.0-, and 3.5-fold, respectively, compared with those in the absence of HNF4α. However, the transcriptional activity of 2C9 -195 was not increased by HNF4α despite the fact that this construct contained two DR1 elements. Deletion of the DR1-A element (2C9 -160) or both elements (2C9 -140) from the promoter region abolished its response for transactivation by HNF4α. The same experiments were also performed using eight different deletion constructs of the CYP2C19 promoter (Fig. 3, right). In contrast to the results obtained from the CYP2C9 constructs, the levels of the transcriptional activities of CYP2C19 constructs were not increased in the presence of HNF4α. Deletion of the DR1-B elements of CYP2C9 (2C9 -140) and CYP2C19 (2C19 -140) from the promoter regions abolished the transcriptional activities in the presence and absence of HNF4α.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Binding of HNF4α to the DR1-A and DR1-B elements of the CYP2C9 and CYP2C19 promoters in vitro. A, EMSAs were performed using nuclear extracts prepared from HepG2 cells with the following probes: HNF4 WT in lane 1, DR1-A WT in lanes 2 through 7, and DR1-B WT in lanes 8 through 13. Oligonucleotide competitors were added with 50-fold excess amounts of the following probes: DR1-A WT in lane 4, DR1-A MT in lane 5, DR1-B WT in lane 10, DR1-B MT in lane 11, HNF4 WT in lanes 6 and 12, and HNF4 MT in lanes 7 and 13. Symbols (+) and (-) indicate the presence and absence of the nuclear extracts or competitors, respectively. B, supershift assays were performed using antibodies specific for HNF4α. Anti-HNF4α IgG (2 μg) was added to the reaction mixtures in lanes 3 and 7. Control IgG (2 μg) was used as a negative control (lanes 4 and 8). Symbols (+) and (-) indicate the presence and absence of the nuclear extracts or IgG, respectively. The arrow indicates supershifted bands.

Mutation analyses were performed to examine whether HNF4α required two DR1 elements for its transactivation ability (Fig. 4). As for the CYP2C9 constructs, HNF4α could stimulate the level of the promoter activity of the wild-type construct (2C9 -2k) to approximately 4-fold, but the introduction of mutation of each DR1 element resulted in complete loss of transactivation of the mutated CYP2C9 promoter (2C9 DR1-Amt, 2C9 DR1-Bmt) by HNF4α (Fig. 4, left). On the other hand, the levels of the transcriptional activities of constructs of the CYP2C19 promoter (2C19 -2k, 2C19 DR1-Amt, and 2C19 DR1-Bmt) were not increased by HNF4α (Fig. 4, right). These results indicate that HNF4α can increase the level of transcriptional activity of the CYP2C9 promoter but not that of the CYP2C19 promoter and that this activation occurred only when two DR1 elements of the CYP2C9 promoter were simultaneously functional. Introduction of mutation of DR1-B elements decreased transcriptional activities of the CYP2C9 and CYP2C19 promoters in the presence and absence of HNF4α.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Effects of HNF4α on transcriptional activity of the CYP2C9 and CYP2C19 promoters in FLC7 cells. Deletion constructs (200 ng) of the CYP2C9 or CYP2C19 promoter were cotransfected with 100 ng of HNF4α expression vector (pHNF4α, open bars) or 100 ng of an empty vector (pT, closed bars). Two HNF4α binding sites are shown in circles. Each value is the mean ± S.D. of relative activity (firefly/Renilla) for four separate experiments, each performed in triplicate. *, p < 0.05 and **, p < 0.01 compared with the empty vector. Luc, luciferase.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Mutation analysis for two HNF4α binding sites of the CYP2C9 and CYP2C19 promoters in FLC7 cells. Reporter constructs (200 ng) of the CYP2C9 or CYP2C19 promoter were cotransfected with 100 ng of HNF4α expression vector (pHNF4α, open bars) or 100 ng of an empty vector (pT, closed bars). Two HNF4α binding sites are shown in circles, and mutations are indicated by crosses. Each value is the mean ± S.D. of relative activity (firefly/Renilla) for three separate experiments, each performed in triplicate. *, p < 0.05 and **, p < 0.01 compared with the empty vector. Luc, luciferase.

Binding of HNF4α to the CYP2C9 Promoter but Not to the CYP2C19 Promoter in Vivo. ChIP assays were performed using human liver to examine whether HNF4α could bind to the CYP2C9 and CYP2C19 gene promoters in vivo (Fig. 5). After DNA extraction of the immunoprecipitated chromatin, PCR was performed to detect the occupancy of DR1 elements of the CYP2C9 and CYP2C19 genes by HNF4α. As for CYP2C9, the DR1 elements were much more abundant in DNA extracted from chromatin immunoprecipitated with anti-HNF4α IgG than in that with control mouse IgG (Fig. 5, top). On the other hand, no DNA fragment around the DR1 elements of the CYP2C19 gene was detected in both extracted DNA samples (Fig. 5, bottom). We also determined the expression levels of CYP2C9 and CYP2C19 mRNA in the same liver used for ChIP assays by using quantitative real-time PCR. The expression level of CYP2C9 mRNA was 82.5 times higher than that of CYP2C19 mRNA.

Discussion

The present study showed that two DR1 elements were located in the CYP2C9 promoter (Fig. 1) and that the transcriptional activities of the CYP2C9 promoter were increased by exogenous HNF4α (Fig. 3). The introduction of mutation to each DR1 element resulted in complete loss of transactivation (Fig. 4). These results are consistent with the results presented in a recent report (Chen et al., 2005). We also performed the same experiments for the CYP2C19 promoter. In contrast to the case of the CYP2C9 promoter, transactivation by HNF4α was not observed in the CYP2C19 promoter despite the existence of two DR1 elements (Fig. 3). In addition, HNF4α could bind to the DR1 elements located in the CYP2C9 promoter but not to those in the CYP2C19 promoter in vivo (Fig. 5). These results suggest that HNF4α participated in the transactivation of at least -2 kbp of the CYP2C9 promoter but not that of the CYP2C19 promoter.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Binding of HNF4α to the DR1-A and DR1-B elements of the CYP2C9 and CYP2C19 promoters in vivo. ChIP assays were performed using the sheared genomic DNA extracted from human liver (9 μg), control mouse IgG (3 μg), and anti-HNF4α IgG (3 μg). M, DNA size marker; input, control sheared genomic DNA; HNF4α, sheared genomic DNA immunoprecipitated with anti-HNF4α IgG; control, sheared genomic DNA immunoprecipitated with control mouse IgG; N.C., nontemplate control.

Significant decreases in the levels of CYP3A4, CYP3A5, CYP2A6, CYP2B6, CYP2C9, and CYP2D6 mRNA have been observed in HNF4α-deficient human hepatocytes (Jover et al., 2001), and several studies have shown that transcription of CYP3A4, CYP2A6, and CYP2D6 genes are regulated by HNF4α via DR1 elements located in their promoters (Cairns et al., 1996; Tirona et al., 2003; Pitarque et al., 2005). Our results and the results of a recent study by Chen et al. (2005) showed that HNF4α was involved in the expression of the CYP2C9 gene. Therefore, these findings suggest that HNF4α plays important roles in regulation of the expression of these P450 genes in the human liver. On the other hand, it has been reported that HNF4α is not involved in transactivation of the CYP2C18 promoter, although a DR1 element is located in this promoter (Ibeanu and Goldstein, 1995). Considering the fact that the expression level of CYP2C18 mRNA in the human liver is very low compared with the expression levels of other genes of the CYP2C subfamily (Goldstein and de Morais, 1994), it is possible that the lack of a functional DR1 element in the CYP2C18 promoter contributed to this low level of expression of CYP2C18 mRNA in the human liver. Accordingly, the same idea would also explain why the expression level of CYP2C19 is lower than that of CYP2C9. That is, the existence of “functional” DR1 elements in the regulatory region of the CYP2C9 gene would be crucial factors for its higher level of expression than that of the CYP2C19 gene in the human liver.

The reason for the different effects of HNF4α on transactivation of the CYP2C9 and CYP2C19 genes is currently unknown. However, the results obtained from our study provided some clues for understanding this difference. HNF4α could not transactivate the CYP2C9 promoter in the absence of the region from -255 to -195 bp (-255/-195 bp), although two DR1 elements were still present in the promoter (Fig. 3), suggesting that the region -255/-195 bp of the CYP2C9 promoter is necessary for HNF4α to up-regulate the transcription of the CYP2C9 gene. One possible explanation for these results is that other HNF4α binding sites exist in the region -255/-195 bp of the CYP2C9 gene, and they can help the action of HNF4α that is recruited to the downstream elements. However, no DR1 elements were found in this region of the CYP2C9 promoter by a searching program for nuclear receptor binding sites (http://www.nubiscan.unibas.ch/; Podvinec et al., 2002), and HNF4α could not bind to this region in EMSA by using an oligonucleotide probe ranging from -255 to -195 bp (data not shown). Therefore, effects of the region -255/-195 bp on transcription of the CYP2C9 gene are unlikely to be mediated by the direct binding of HNF4α to this region. Another possibility is that a certain factor, which assists with HNF4α-mediated transactivation of the CYP2C9 promoter, specifically binds to the region -255/-195 bp of the CYP2C9 gene but not to the CYP2C19 gene. Actually, there are 8-bp differences between the region -255/-195 bp of the CYP2C9 promoter and the region -257/-197 bp of the CYP2C19 promoter. The factor that binds to the region -255/-195 bp of the CYP2C9 promoter may stabilize the binding of HNF4α to the DR1 element of the CYP2C9 promoter, or it may recruit cofactors that are required for function of HNF4α. However, no complexes were formed in EMSAs using HepG2 nuclear extracts and an oligonucleotide probe ranging from -255 to -195 bp (data not shown). A searching program for transcriptional factors could not identify any factors that fulfill these requirements. Thus, further detailed study is needed to elucidate the role of the region -255/-195 bp of the CYP2C9 promoter in HNF4α function.

Deletion and mutation of DR1-B elements decreased transcriptional activities of the CYP2C9 and CYP2C19 promoters in the presence and absence of HNF4α (Figs. 3 and 4). A putative binding site of CCAAT enhancer-binding protein was found in the downstream of DR1-B elements partly overlapped. Therefore, the binding of CCAAT enhancer-binding protein to the CYP2C9 and CYP2C19 promoters may be inhibited by deletion and mutation of DR1-B elements, resulting in the decrease of basal activities of CYP2C9 and CYP2C19 promoters.

In conclusion, we showed that HNF4α is one of the important factors regulating promoter activity of the CYP2C9 gene but not that of the CYP2C19 gene in the human liver. The direct binding of HNF4α to two DR1 elements of the CYP2C9 promoter is essential for HNF4α-mediated transactivation of the CYP2C9 promoter. In addition, this transactivation requires certain factors that facilitate the function of HNF4α via the region from -255/-195 bp of the CYP2C9 promoter. The results of the present study suggest that HNF4α is one of the determinants for the difference between expression levels of CYP2C9 and CYP2C19 in the human liver.