Abstract
The pregnane X receptor (PXR) is a key regulator of genes encoding several major types of cytochrome P450 enzymes and transporters (e.g., multidrug resistance-1, MDR1); therefore, PXR contributes significantly to drug-drug interactions. PXR binds to response elements and confers transactivation. Several target genes such as CYP3A4 and 3A7 contain two PXR elements (distant and proximal) that are separated by more than 7000 nucleotides in the genome. Disruption of the distant element causes a 73% decrease of the reporter activity, whereas inactivation of the proximal element decreases by only 53%. This study was undertaken to test the hypothesis that PXR differentially binds to the elements with the distant enhancer being bound to a higher extent. To test this hypothesis, a stable transfected line (hPXR-HRE) was prepared to constitutively express human PXR and harbor a chromatinized CYP3A4-ER6 reporter. This line responded to rifampicin and dexamethasone similarly as hepatocytes based on the relative potency and activation kinetics. Contrary to the hypothesis, chromatin immunoprecipitation experiments showed that the genomic fragment harboring the proximal element was preferably precipitated over the fragment containing the distant element in the CYP3A4 gene, but the opposite was true with the CYP3A7 gene. In addition, the promoters from the MDR1 and CYP2B6 genes were abundantly present in the PXR immunocomplexes from the vehicle-treated cells. However, such abundant interactions were markedly diminished when cells were treated with PXR activator rifampicin. These findings suggest that PXR binding is dependent on the genomic context and PXR activators modulate such bindings.
Drug metabolism and efflux are two major pharmacokinetic determinants that contribute significantly to drug-drug interactions. The cytochrome P450 (P4502) system is involved in the metabolism of more than 90% drugs, whereas transporter proteins such as multidrug resistance-1 (MDR1) are known to mediate xenobiotic efflux and thus dictate absorption, presystemic elimination, and intracellular concentration of many therapeutic agents (Shou et al., 2001; Skatrud, 2002). The pregnane X receptor (PXR) is a key regulator of genes encoding several major types of P450 enzymes and transporter proteins (Goodwin et al., 2002). PXR structurally belongs to a superfamily of the nuclear receptors and regulates target gene transcription in a ligand-dependent manner. Unlike many other nuclear receptors, PXR has a rather large ligand-binding pocket, which is spherical in shape, extremely hydrophobic, and expandable (Watkins et al., 2001). Such structural features allow PXR to interact with a wide range of chemicals with dissimilar structure. These chemicals include prescription drugs (e.g., rifampicin), herbal supplements (e.g., hyperforin), and pesticides (e.g., trans-nonachlor) (Moore et al., 2000; Goodwin et al., 2002). In addition, many endogenous compounds are also shown to activate PXR, notably steroidal compounds (e.g., corticosterone) and bile acids (e.g., lithocholic acid). Interestingly, both glucocorticoids (e.g., dexamethasone) and antiglucocorticoids (e.g., pregnenolone 16α-carbonitrile) are shown to activate PXR (Kliewer et al., 1998; Goodwin et al., 2002).
In addition to the promiscuous nature on ligand interactions, PXR exhibits a broad binding activity toward varying DNA elements (Geick et al., 2002; Goodwin et al., 2002; Kast et al., 2002; Toell et al., 2002). Most elements contain a half-site AG(G/T)TCA; however, several distinct but related half-sites also form PXR response elements (PXREs) that effectively interact with PXR-RXRα heterodimers (Goodwin et al., 2002). These half-sites are arranged into various configurations such as direct, inverted, and everted repeats that are spaced by different numbers of nucleotides. In some PXR target genes, more than one PXRE is located (Goodwin et al., 1999; Pascussi et al., 1999; Bertilsson et al., 2001). For example, the CYP3A4 gene, encoding the most abundant P450 enzyme, has one PXRE located in the proximal promoter region (proximal PXRE) and a second one (distal PXRE) located ∼8 kb upstream of the transcription initiation site (Goodwin et al., 1999). Transient cotransfection experiments with reporter constructs demonstrate that both elements are required for maximum activation. Selective disruption of the distant or proximal element causes decreases in responding to PXR-mediated transactivation by 73 and 53%, respectively (Goodwin et al., 1999). A minimal promoter (containing the proximal element), although supporting the basal transcription, shows little ability to confer PXR-mediated induction (Goodwin et al., 1999). These findings suggest that both elements work in a coordinated manner and the distal but not the proximal element is the primary target for PXR interaction. The distal and proximal elements are also located in the CYP3A7 gene, which encodes a major fetal P450 enzyme (Pascussi et al., 1999; Bertilsson et al., 2001).
This study was undertaken to test the hypothesis that PXR differentially binds to the proximal and distal elements with the distant element being bound to a higher extent. To test this hypothesis, a stable transfected line (hPXR-HRE) was prepared with human PXR; chromatin immunoprecipitation (CHIP) was performed with an antibody specific to the transfected PXR; and the immunoprecipitates were analyzed for the relative abundance of genomic fragments containing the PXR elements among several target genes. In the CYP3A4 gene, the proximal element-containing fragment was precipitated to a markedly higher extent than the fragment containing the distant element, whereas the opposite was true with the CYP3A7 gene. In the absence of rifampicin, the promoter fragments from the MDR1 and CYP2B6 genes were abundantly precipitated. However, such abundant interactions were markedly diminished when cells were treated with rifampicin. These findings suggest that PXR binding is dependent on the genomic context and PXR activators modulate such bindings. These findings also provide a molecular explanation on the widely observed differential inducibility of the target genes by PXR activators.
Materials and Methods
Supplies and Plasmid Constructs. Human liver cDNA library, cDNA positive selection kit, LipofectAMINE, and culture medium were purchased from Invitrogen (Carlsbad, CA); rifampicin, dexamethasone, FLAG-tagged vector, the anti-FLAG monoclonal antibody, and its agarose-bead conjugates were obtained from Sigma-Aldrich (St. Louis, MO); kits for luciferase detection and the pRL-TK plasmid were purchased from Promega (Madison, WI). Delipidated and normal fetal bovine sera were from Hyclone Laboratories (Logan, UT). Unless otherwise indicated, all other reagents were purchased from Fisher Scientific Co. (Pittsburgh, PA). Full-length cDNA encoding human PXR was isolated from a cDNA library by a cDNA-trapping procedure with a biotinylated oligonucleotide ACAGTGGCTGTGAGCTTCCAG as described previously (Hu and Yan, 1999; Zhang et al., 1999). The coding sequence was amplified by PCR and subcloned to the FLAG-cytomegalovirus vector to express the FLAG-hPXR. The CYP3A4-ER6 reporter was constructed by inserting two copies of CYP3A4 response element (5′-GATCAGACAGTTCATGAAGTTCATCTAGATC-3′) into the pGL3 promoter luciferase vector. Two additional reporters were prepared with the pGL3 basic vector to contain the proximal element alone (–362 to + 53, CYP3A4-P) or the proximal fused to the distal element (–362 to + 53 and –7836 to –7208, CYP3A4-DP) (Goodwin et al., 1999). The fragments harboring these elements were amplified by PCR with primers that were extended to include appropriate endonucleases to facilitate subsequent ligation. All constructs were subjected to sequencing analyses.
Transient Cotransfection Experiment. Cells (CV-1 and HEK293T) were plated in 24-well plates in Dulbecco's modified Eagle's medium supplemented with 10% delipidated fetal calf serum at a density of 8 × 104 cells per well. Transfection was conducted by lipofection with LipofectAMINE as described previously (Li et al., 2003). Transfection mixtures contained 200 ng of hPXR plasmid, 100 ng of reporter plasmid, and 10 ng of pRL-TK plasmid. Cells were transfected for 4 h, and the medium was replaced with fresh medium. After 24 h, media were changed again with the same media containing rifampicin or solvent DMSO. The transfected cells were incubated for an additional 24 h and washed once with PBS and collected by scraping. The collected cells were subjected to two cycles of freeze/thaw. The reporter enzyme activities were assayed with a dual-luciferase reporter assay system as described by the manufacturer. This system contained two substrates, which were used to determine the activity of two luciferases sequentially. The firefly luciferase activity, which represented the reporter gene activity, was initiated by mixing an aliquot of lysates (10 μl) with luciferase assay reagent II. Then the firefly luminescence was quenched, and the Renilla luminescence was simultaneously activated by adding Stop & Glo reagent to the sample tube. The firefly luminescence signal was normalized based on the Renilla luminescence signal, and the ratio of rifampicin to DMSO treatment served as fold induction.
Stable Transfection. Stable transfection lines were prepared with 293T cells to constitutively express hPXR and harbor the CYP3A4-ER6 reporter in the genome. To facilitate the selection of genomic integration, the FLAG-hPXR construct was modified by inserting a Zeocin-resistant gene, whereas the CYP3A4-ER6 reporter construct was altered by a blasticidin-resistant gene. The initial transfection was conducted with the FLAG-hPXR expression plasmid, whereas the subsequent transfection was performed with the CYP3A4-ER6 reporter construct. Cells were seeded at 50% confluence and transfected by LipofectAMINE as described previously (Li et al., 2002b). The transfected cells were cultured in full media for 24 h and split into fresh media. The split cells were seeded to 10% confluence and cultured in media containing Zeocin (300 μg/ml). Positive foci resistant to Zeocin were picked up, expanded, and assayed for their ability to respond to rifampicin on activating the CYP3A4-ER6 reporter by cotransfection experiments. One subline exhibiting a high responsiveness was transfected again with the CYP3A4-ER6 reporter construct. The selection was performed against both blasticidin (5 μg/ml) and Zeocin (300 μg/ml). Blasticidin was designed to select stable transfectants of the CYP3A4-ER6 reporter construct, whereas Zeocin was designed to maintain the integration of the previously transfected FLAG-hPXR construct. The resultant sublines containing both transfected constructs were tested for their ability to respond to PXR ligand-mediated activation. In this case, only the firefly luminescence signal was determined and normalized based on the amount of the protein used.
Immunocytochemistry. Stable transfected cells were seeded in chamber slides at ∼40% confluence and cultured overnight. Media were replaced with fresh media containing rifampicin, dexamethasone, or DMSO. After an additional 24-h incubation, cells were fixed in ice-cold 10% acetone for 10 min, then air-dried. The chambers were removed, and the entire slide was immersed in PBS for rehydration. The slides were then incubated in 1% goat serum for 2 h to block nonspecific binding followed by incubation in the anti-FLAG monoclonal antibody (1:1000 dilution with PBS) for 30 min. The slide was washed five times with PBS. Subsequently, the slide was incubated with fluorescent isothiocyanate-conjugated anti-mouse IgG antibody for an additional 30 min. The slide was mounted and visualized by a confocal microscope (Xie et al., 2003).
Formaldehyde Cross-Linked Chromatin Preparation. The dual transfected line (hPXR-HRE), highly responding to rifampicin stimuli, was cultured in a 75-ml flask to an 80% confluence and treated with DMSO or rifampicin (10 μM) for 24 h. The cells were then cross-linked for 10 min at room temperature with formaldehyde solution (11% formaldehyde, 50 mM HEPES, pH 8.0, 100 mM NaCl, 1 mM EDTA, 0.5 mM EGTA). The cross-linking was terminated by adding glycine to a final concentration of 125 mM. The fixed cells were washed three times with ice-cold PBS and collected in lysis buffer (10 mM Tris-HCl, pH 8.0, 0.25% Triton X-100, 1 mM EDTA, 0.5 mM EGTA, 10 mM sodium butyrate, 100 mM NaCl, 20 mM β-glycerophosphate, 100 μM sodium orthovanadate, and protease inhibitors A and B). After a 10-min incubation on ice, the cell suspensions were centrifuged at 3000 rpm for 4 min under refrigeration. The nuclear pellets were resuspended in wash buffer (lysis buffer minus Triton X-100) and centrifuged again. The resultant pellets were resuspended in 1.5 ml of wash buffer and sonicated with a Branson Sonifier 250 (Branson Ultrasonics Corporation, Danbury, CT) with an output of 40 at 20% cycle for 30 s. The sonications were repeated five times. To the sonicated chromatins, SDS was added to a final concentration of 1%, and the chromatin solutions were rotated at room temperature for 1 h. The soluble supernatant chromatins were prepared by centrifuging at room temperature for 30 min (13,000g). After the cross-linking was reversed, the soluble chromatins were analyzed by agarose gel electrophoresis to ensure that the average size of the chromatins was ∼1.5 kb.
Chromatin Immunoprecipitation. The chromatin solutions were diluted 10-fold to reduce the concentration of SDS to 0.1% in RIPA buffer (10 mM Tris-HCl, pH 8.0, 1% Triton X-100, 0.1% sodium deoxycholate, 1 mM EDTA, 0.5 mM EGTA, 10 mM sodium butyrate, 140 mM NaCl, 20 mM β-glycerophosphate, 100 μM sodium orthovanadate, and protease inhibitor cocktails). The diluted chromatin solutions (one fourth) were mixed with 50 μl of anti-FLAG-coated agarose beads preincubated with bovine serum albumin and sonicated lambda phage DNA. As a negative control, the agarose beads were preincubated with lysates from non-cross-linked cells. The chromatin/agarose beads were rotated at room temperature for 4 h and centrifuged at 1000g for 30 min. The beads were washed as follows: 3 × 1 ml of RIPA buffer; 3 × 1 ml of RIPA plus 500 mM NaCl; 3 × 1 ml of RIPA containing 250 mM LiCl, 1% NP40, and 1% sodium deoxycholate; and 3 × 1 ml of Tris-EDTA, pH 7.5, containing 10 mM sodium butyrate, 20 mM β-glycerophosphate, and protease inhibitor cocktails. The antibody bound chromatins were eluted by TE washing buffer containing 1.5% SDS. The input, unbound, and antibody bound chromatins were incubated at 65°C for 7 h to reverse the cross-linking. The DNA was precipitated with ethanol, resuspended in water, and treated with RNase A (50 μg/ml) for 1 h at 37°C followed by proteinase K treatment (100 μg/ml) in the presence of 0.25% SDS at 37°C overnight. The DNAs were subjected to phenol-chloroform extraction once and chloroform extraction once, precipitated with ethanol in the presence of glycogen (10 μg), and redissolved in water. The presence of the genomic fragments containing PXR elements in the CYP2B6, 3A4, 3A7, and MDR1 genes was analyzed by PCR with primers listed in Table 1.
Other Assays. Protein concentration was determined with a BCA protein assay kit (Pierce Chemical, Rockford, IL) as described by the manufacturer. Western analyses were performed as described previously (Xie et al., 2002). Data are presented as mean ± S.D. of at least three separate experiments, except where results of blots are shown in which case a representative experiment is depicted in the figures. Comparisons between two values were made with Student's t test at p < 0.05.
Results
Transactivation of CYP3A4 Enhancer and Promoter Reporters in CV-1 and 293T Cells. To study intracellular interactions between PXR and its response elements as a function of PXR-mediated transactivation, stable transfectants were prepared to constitutively express human PXR and to harbor a CYP3A4 reporter in the genome. The initial effort was made to select the type of reporters (e.g., enhancer and native promoter) for stable transfection. Three reporters were tested, including CYP3A4-ER6, CYP3A4-P, and CYP3A4-DP. The CYP3A4-ER6 reporter contained two copies of the ER6 element in the proximal promoter of the CYP3A4 gene; the CYP3A4-P contained the proximal promoter (–362 to + 53); and the CYP3A4-DP had the DR3-containing fragment (–7836 to –7208) fused to the proximal promoter. Both wild-type and FLAG-tagged PXR constructs were tested for the ability to transactivate these reporters.
The reporter activities in response to rifampicin are summarized in Fig. 1A. In both CV-1 and 293T, the CYP3A4-ER6 and CYP3A4-PD reporters were markedly transactivated, with the CYP3A4-ER6 showing a slightly higher activity (∼4versus 3-fold). In contrast, the CYP3A4-P reporter, although containing the ER6 element, exhibited little activity. The insignificant responsiveness of the CYP3A4-P reporter was not due to an insufficient amount of this construct, since little activation was consistently observed even though two times as much as the other reporters were used (data not shown). These findings are consistent with previous reports that the proximal ER6 element in the CYP3A4 native promoter has no inherent ability to respond to PXR-mediated transactivation but coordinates the distal element-mediated activity (Goodwin et al., 1999, 2002). In addition, the FLAG-tagged PXR was slightly less active than the wild-type PXR toward all reporters. Generally, CV-1 cells supported a somewhat higher transactivation than 293 cells, although both cell lines were derived from the kidney (monkey and human, respectively).
We next examined whether the slightly higher activation in CV-1 cells was due to higher expression of the transfected PXR. Western analyses were performed with both nuclear and cytosolic fractions. In contrary to the prediction, both nuclear and cytosolic PXR were readily detected in 293T but not in CV-1 cells (Fig. 1B). In fact, no cytosolic PXR was detected in CV-1 cells. Even in the nucleus, the level of PXR in 293T cells was more than 20 times of that in CV-1 cells (Fig. 1B). The levels of β-actin, however, were comparable, suggesting that CV-1 cells indeed support lower expression of the transfected PXR or the transfection efficiency is lower in CV-1 cells under the conditions used. In addition, treatment with rifampicin caused no changes on the expression of PXR or the relative abundance between the nuclear and cytoplasmic compartments (data not shown).
Stable Transfectants Exhibited Same Activation Profiles as Primary Cultured Hepatocytes. Transient transfection experiments are widely used to determine reporter activity (Goodwin et al., 2002; Li et al., 2003); however, many factors are shown to affect the overall transactivation activity, such as the copy number of the inserted response element. Recently, transiently introduced and chromatinized reporters are found to differentially respond to ligands of RAR and RXRα, two receptors that are structurally related to PXR (Lefebvre et al., 2002). We next tested chromatinized CYP3A4 reporter in response to PXR activators. Cells (293T) were transfected stably to express hPXR, and clones responding to rifampicin were stably transfected again with the CYP3A4-ER6 reporter. The resultant dually transfected clones, designated hPXR-HRE cells, were tested for the ability to respond to rifampicin and dexamethasone, two prototypical CYP3A inducers that have a different potency toward human PXR (Pascussi et al., 2001). Figure 2A shows the representative results from hPXR-HRE cells. Both rifampicin and dexamethasone caused a concentration-dependent activation of the reporter. At higher concentrations (≥5 μM), rifampicin caused a significantly higher activation than dexamethasone, with the highest activity being ∼7- and 4-fold, respectively. In contrast, at nanomolar concentrations (0.05 and 0.1 μM), dexamethasone caused higher activation than rifampicin. Overall, dexamethasone caused a biphasic activation pattern, whereas rifampicin caused a monophasic activation pattern (Fig. 2A). The differences on the activation profiles and relative potency between dexamethasone and rifampicin were similar to those reported on the induction of CYP3A4 in primary hepatocytes (Pascussi et al., 2001). In addition, the stable transfected cells supported a higher activation than transient cotransfection (3.5versus 5.8-fold at 10 μM of rifampicin) (Figs. 1A and 2A).
Several investigators reported that the biphasic induction of CYP3A4 by dexamethasone in hepatocytes is achieved through two distinct mechanisms: initial induction of PXR (phase I) and subsequent activation of this receptor (phase II) (Huss and Kasper., 2000; Pascussi et al., 2000, 2001). Next we examined whether dexamethasone increases the expression of the FLAG-tagged PXR in the stably transfected cells as seen with the PXR endogenous gene in hepatocytes. The samples used for assaying reporter activities (10 μM shown only) were subjected to Western analyses with an anti-FLAG antibody. As shown in Fig. 2B, cells treated with rifampicin expressed similar levels of PXR as DMSO-treated cells. In contrast, dexamethasone-treated cells expressed two times as much. Such an increased expression was detected in both nuclear and cytosolic fractions. Consistent with the immunoblotting analyses, immunocytochemical studies visualized by confocal microscope revealed that dexamethasone-treated cells exhibited a generally brighter staining than cells treated with rifampicin or DMSO. Although some cells showed higher staining than others, the overall distribution of PXR between the nucleus and cytoplasm was comparable among all treatments (Fig. 2C). The confocal images were obtained with a chamber slide, which was immunochemically stained in a single container to ensure that cells in different chambers were uniformly processed. The experiments were repeated four times and consistent results were obtained. No staining was observed in the vector-transfected cells or when the primary antibody was replaced with control IgG (data not shown).
PXR Bindings Are Dependent on the Genomic Context, and Rifampicin Modulates Such Bindings. The hPXR-HRE cells responded to rifampicin and dexamethasone similarly as cultured hepatocytes in terms of relative potency and activation profiles (Pascussi et al., 2001), suggesting that these cells provide a valid model to study intracellular interactions between PXR and its response elements. We next performed CHIP to determine whether promoters targeted by PXR were selectively precipitated by an anti-FLAG antibody (recognizing the transfected PXR). Several target genes were chosen including CYP2B6, CYP3A4, CYP3A7, and MDR1. The promoters of these genes contain one or more PXREs (e.g., DR3, ER6) and have been shown to respond to PXR-mediated transactivation by cotransfection experiments (Goodwin et al., 2001, 2002; Geick et al., 2002; Kast et al., 2002; Toell et al., 2002). The hPXR-HRE cells were treated with DMSO or rifampicin (10 μM), and a small portion of the cells was collected for assaying reporter activity (data not shown), whereas the remaining cells were cross-linked with formaldehyde. The nuclei were isolated, and the chromatins were sonicated. The sonicated chromatins were subjected to centrifugation to remove large DNA fragments and analyzed by agarose gel electrophoresis to ensure that the sonicated chromatins in the supernatants had an average length of ∼1.5 kb (data not shown). The resultant chromatins were subjected to immunoprecipitation with an anti-FLAG antibody-coated agarose beads and analyzed by PXR for the relative abundance of genomic fragments targeted by PXR.
The results of the CHIP experiments are summarized in Fig. 3A. For all target genes, PCR amplifications with DNA input as the template resulted in a product with a predicted size (Table 1). The relative intensities of the amplified products were comparable among the target sequences, suggesting that the genes analyzed have a similar copy number in the hPXR-HRE cells (Fig. 3A). In contrast, the relative intensities varied markedly among the amplified products when the immunocomplexed DNA was used as the template (Fig. 3A, middle and bottom). The genomic fragments containing the CYP2B6, CYP3A4-P, MDR1, or the integrated CYP-ER6 element were amplified to a markedly higher extent than those containing the CYP3A4-D, CYP3A7-P, or CYP3A7-D element. Interestingly, when the precipitated DNA from the rifampicin-treated cells was used as the template, the abundant amplification of MDR1 and CYP2B6 sequences was markedly diminished (Fig. 3). It should be emphasized that more PCR cycles were performed with precipitated than input DNA (32 versus 40).
Several control experiments were conducted to validate the CHIP experiments. The DEC2 fragment, derived from a gene that is not a PXR target based on transient cotransfection experiments (Li et al., 2003), was detected only when the DNA input was used as the template. The anti-FLAG-coated beads, saturated by FLAG-tagged PXR lysates before being used for CHIP, failed to precipitate chromatins even with excessive cycles of PCR amplification (data not shown). The PCR analyses were performed with more than one pair of primers on each target gene, and consistent results were obtained. Finally, a master tube containing all components except the primers was prepared and equally distributed among all PCR tubes to minimize intertube variations. These controls ensured the specificity of the CHIP experiments.
Discussion
This study reports the intracellular interactions of PXR with its varying response elements in the presence or absence of rifampicin, a potent PXR activator. A stable transfected line hPXR-HRE, constitutively expressing human PXR and harboring the CYP3A4-ER6 reporter in the genome, responds to rifampicin and dexamethasone similarly as cultured hepatocytes in terms of relative potency and activation kinetics (Pascussi et al., 2001). CHIP experiments with input DNA produce comparable PCR amplifications among the target genes analyzed. In contrast, CHIP experiments with PXR-complexed DNA results in markedly differential amplifications, suggesting that this receptor differentially interacts with its response elements intracellularly. The proximal element in the CYP3A4 gene is preferably bound over the distal element, whereas the opposite is true with the CYP3A7 gene.
The binding preference toward the CYP3A4-ER6 (proximal) over the CYP3A4-DR3 (distal) element is likely due to the difference on the genomic context rather than the binding affinity. Electrophoretic mobility shift assays have demonstrated that in vitro translated PXR-RXRα heterodimers bind comparably to both elements, and the binding of radiolabeled ER6 element is equally competed by unlabeled CYP3A4-DR3 or -ER6 element, suggesting that PXR has a similar binding affinity toward both elements (Goodwin et al., 1999). In this study, the CYP3A4-ER6-containing fragment is consistently precipitated preferably over that of the CYP3A4-DR3 element by a ratio of 5:1 (Fig. 3A), and the regions harboring these two elements exhibit little sequence identity, suggesting that the difference on the genomic contexts is responsible for such differentially intracellular interactions. The precise mechanism on the preferable binding of CYP3A4-ER6 remains to be determined. It is likely that the CYP3A4-ER6 element is located in the region where the basal transcriptional apparatus is present (Goodwin et al., 1999), which provides a platform for better interactions with nuclear proteins.
The CYP3A7 elements (distal and proximal), in contrast to their CYP3A4 counterparts, exhibit an opposite binding preference, although the overall interactions are markedly diminished in the CYP3A7 gene (Fig. 3). The CYP3A7-DR3 is identical to the CYP3A4-DR3 element, whereas the CYP3A7-ER6 differs from the CYP3A4-ER6 (Table 2). Two forms (ER6-Itoh and ER6-jmp) of the CYP3A7-ER6 element have been reported with the ER6-Itoh having a single base deletion in the 3′-half-site of the element (Itoh et al., 1992). Even in the case of the CYP3A7-ER6-jmp, a single nucleotide substitution is located compared with the CYP3A4-ER6 element (TTAACT versus TGAACT) (Goodwin et al., 1999; Pascussi et al., 1999). In vitro translated PXR-RXRα heterodimers are shown to effectively bind to the CYP3A7-ER6-jmp but not the CYP3A7-ER6-Itoh. However, such a binding is readily displaced by the CYP3A4-ER6, suggesting that the CYP3A7-ER6, even in the form of ER6-jmp, represents a lower affinity-binding site for PXR (Pascussi et al., 1999). Given the importance of the ER6 element on CYP3A induction (Goodwin et al., 1999), the markedly less intracellular interactions of the CYP3A7-ER6 established by the CHIP experiments provide a molecular explanation that CYP3A4 is a more sensitive target of PXR than CYP3A7 (Rae et al., 2001).
It is interesting to notice that rifampicin modulates the overall PXR binding among its target genes, particularly on the MDR1 gene. In the absence of rifampicin, the MDR1 promoter shows the most PXR binding; however, such an abundant binding is markedly diminished in cells treated with rifampicin (Fig. 3). This drastic reduction is likely due to the clustered elements in the MDR1 gene (Geick et al., 2002). In the distal regulatory region, a PXR-binding cluster (only 48 nucleotides long) is located to include three DR4 elements, one DR3, and one ER6. These elements overlap at least by one half-site. Gel mobility shift assays show that several of the elements bind to PXR-RXRα heterodimers, although the binding affinity varies from element to element (Geick et al., 2002). Nevertheless, the clustered elements likely increase the likelihood to interact with PXR, resulting in an abundant presence of the MDR1 promoter in the PXR-complexed chromatins from the DMSO-treated cells. However, mechanisms for rifampicin-induced reduction on the interactions remain to be determined. It is likely that interactions between PXR and clustered elements are weak and may not tolerate conformation changes as a result of dissociation or recruitment of cofactors initiated by the PXR-ligand binding. In addition to MDR1, CYP2B6 has exhibited the rifampicin-induced reduction on the interactions. Interestingly, among the target genes analyzed, only MDR1 and CYP2B6 contain DR4 elements (Table 2).
The most intracellular interactions of the CYP3A4-ER6 detected by CHIP in the presence of rifampicin are highly consistent with numerous reports that the CYP3A4 gene is the most sensitive target of PXR among the genes analyzed (Greuet et al., 1996; Schuetz et al., 1996; Rae et al., 2001), suggesting that the presence of a liganded PXR in the promoter determines induction specificity and magnitude. However, overexpression of PXR does not necessarily change the equation toward a higher induction. As a matter of fact, in some cases, high levels of PXR expression cause even lower transactivation. In this study, 293T cells support markedly higher expression of transfected PXR than CV-1 cells; however, the overall activation is inversely correlated with the PXR levels (3.6versus 4.3-fold) (Fig. 1A). Even in 293T cells, dexamethasone but not rifampicin markedly increases the expression of the transfected PXR; however, rifampicin causes a higher activation (5.7 versus 3.8 at 10 μM) (Fig. 2A). The precise mechanisms remain to be determined on such a disproportional relationship. It is likely that high levels of the cytoplasmic PXR in 293T cells are readily accessible to PXR ligands and thus decrease the amount of ligands for the nuclear PXR. Alternatively, free PXR molecules (e.g., unbound to a DNA element) retain the ability to interact a ligand, and such PXR-ligand complexes recruit then deplete coactivators (e.g., steroid receptor coactivator-1), which are required for PXR-mediated transcription activation (Kliewer et al., 1998).
Unliganded PXR, if bound to a DNA element, may actually exert repressive activity. Several lines of evidence support this possibility. First, mice selectively lacking functional PXR have higher mRNA levels of CYP3A than the corresponding wild-type mice, providing direct evidence that unliganded PXR represses basal transcription of its target genes (Xie et al., 2000; Staudinger et al., 2001). Second, PXR and its functionally related protein constitutive androstane receptor have been recently shown to interact with Pit-1 in the absence of a ligand, and such an interaction is transcriptionally repressive (Gonzalez and Carlberg, 2002). Finally, it has been well established that many hormone receptors such as thyroid hormone receptor, structurally related to PXR, exert repressive activity in the absence of a ligand (Li et al., 2002a). Therefore, the availability of corepressors, coactivators, and ligands (exogenous and endogenous) collectively determines the overall transactivation activity mediated by PXR.
The biphasic activation on the reporter by dexamethasone but not rifampicin suggests that this steroid acts on more than one pathway to regulate the expression of PXR target genes. In primary hepatocytes and some hepatic cell lines, dexamethasone causes a biphasic pattern of CYP3A4 induction (Huss and Kasper, 2000; Pascussi et al., 2000, 2001). It has been proposed that the initial phase of CYP3A induction is due to induction of PXR, and the induced PXR in turn interacts with unknown ligand and causes induction of CYP3A. In this study, the stable transfected hPXR-HRE cells constitutively express high levels of PXR, and such an abundant expression is sufficient to confer even higher activation. In support of this notion, rifampicin causes higher activation than dexamethasone but does not increase the PXR expression (Fig. 2). Interestingly, we have not observed such a biphasic activation with a transiently transfected reporter (data not shown), suggesting that chromatin environment is responsible for the biphasic activation in the stable transfected cells. Similar observation has been made with RAR and RXRα, two receptors structurally related to PXR. A transiently introduced reporter supports continuous, cooperative activation of transcription by RAR and RXR ligands, whereas a chromatinized reporter exhibits a complex biphasic activation pattern (Lefebvre et al., 2002).
In summary, our work points to several important conclusions. First, PXR elements (e.g., CYP3A4-ER6 and -DR3), even with similar affinity toward this receptor, are differentially bound by PXR intracellularly, suggesting that the genomic context plays an important role in PXR-DNA interactions. Second, PXR-mediated constitutive binding (e.g., MDR1) can be modulated by a PXR ligand, providing an additional mechanism by which PXR regulates its target genes. Third, the relative abundance of an element present in the PXR-precipitated chromatins from rifampicin-treated cells is correlated well with the sensitivity to PXR-mediated induction among the genes analyzed, suggesting that PXR plays an essential role in determining induction specificity and magnitude. Fourth, dexamethasone but not rifampicin causes a biphasic activation in the chromatinized reporter, suggesting that this steroid acts on more than one pathway to regulate the expression of PXR target genes (probably by modulating chromatin environment). Given the fact that PXR is found to regulate an increasing number of genes in assays such as transient cotransfection, the findings presented in this study provide important molecular insight that only the intracellular complexes between a liganded PXR and its element likely confer transcription activation.
Footnotes
-
↵2 Abbreviations used are: P450, cytochrome P450; MDR1, multidrug resistance-1; PXR, pregnane X receptor; PXRE, PXR response element; RXRα, retinoid X receptor-α; kb, kilobase; CHIP, chromatin immunoprecipitation; ER6, everted repeat spaced by six nucleotides; PCR, polymerase chain reaction; DMSO, dimethyl sulfoxide; PBS, phosphate-buffered saline; RIPA, radioimmunoprecipitation assay; DR3, direct repeat spaced by three nucleotides; RAR, retinoic acid receptor.
-
This work was partially supported by Grant R01GM61988 from the National Institute of General Medical Sciences and a grant from the Center of Alternatives to Animal Testing, Johns Hopkins University.
-
↵1 Present address: Yale Medical School, Section of Neuropathology, 310 Cedar Street, New Haven, CT 06510.
- Received June 6, 2003.
- Accepted September 8, 2003.
- The American Society for Pharmacology and Experimental Therapeutics