Orphan nuclear receptors constitutive androstane receptor and pregnane X receptor share xenobiotic and steroid ligands.

Xenobiotics induce the transcription of cytochromes P450 (CYPs) 2B and 3A through the constitutive androstane receptor (CAR; NR1I3) and pregnane X receptor (PXR; NR1I2), respectively. In this report, we have systematically compared a series of xenobiotics and natural steroids for their effects on mouse and human CAR and PXR. Our results demonstrate dual regulation of PXR and CAR by a subset of compounds that affect CYP expression. Moreover, there are marked pharmacological differences between the mouse (m) and human (h) orthologs of both CAR and PXR. For example, the planar hydrocarbon 1, 4-bis[2-(3,5-dichloropyridyl-oxy)]benzene activates mCAR and hPXR but has little or no activity on hCAR and mPXR. In contrast, the CAR deactivator androstanol activates both mouse and human PXR. Similarly, the PXR activator clotrimazole is a potent deactivator of hCAR. Using radioligand binding and fluorescence resonance energy transfer assays, we demonstrate that several of the compounds that regulate mouse and human CAR, including natural steroids, bind directly to the receptors. Our results suggest that CAR, like PXR, is a steroid receptor that is capable of recognizing structurally diverse compounds. Moreover, our findings underscore the complexity in the physiologic response to xenobiotics.

The cytochromes P450 (CYPs) 1 comprise a superfamily of heme-containing monooxygenases that play a crucial role in the metabolism of natural compounds and xenobiotics, including environmental pollutants and drugs. Transcription of several of the CYP gene families is induced by xenobiotics. For example, the barbiturate phenobarbital (PB) and the antibiotic rifampicin are classic inducers of CYP2B and CYP3A gene expression, respectively (1,2).
Two orphan members of the steroid/retinoid/thyroid hormone receptor superfamily of ligand-activated transcription factors, CAR and PXR, have recently been implicated in mediating the effects of xenobiotics on CYP2B and CYP3A gene expression (1)(2)(3). CAR and PXR are members of the same nuclear receptor subfamily (NR1), sharing ϳ40% amino acid identity in their ligand binding domains (LBDs). CAR is most abundantly expressed in liver and has strong constitutive activity in cell-based reporter assays in the absence of any added ligand (4,5). In HepG2 cells or other cell lines, exogenously expressed CAR can enter the nucleus and regulate the expression of target genes. This constitutive activity can be inhibited by superphysiological concentrations of the testosterone metabolites androstanol and androstenol (6). These androstanes inhibit the interaction of CAR with the steroid receptor coactivator 1 (SRC-1), suggesting that "deactivation" is mediated by direct binding to the orphan receptor. In contrast to transfected cell lines, CAR is not present in the nucleus of primary hepatocytes but is instead sequestered in the cytoplasm. Treatment of primary hepatocytes with either PB or the planar hydrocarbon TCPOBOP results in the translocation of CAR into the nucleus, where it binds to its cognate DNA response elements as a heterodimer with the 9-cis-retinoic acid receptor (RXR) and activates the transcription of target genes, including CYP2B (7)(8)(9). CAR/RXR binding sites have been identified in the PB-responsive regions of the mouse, rat, and human CYP2B genes. The effects of PB on CYP2B expression are blocked by the phosphatase inhibitor okadaic acid (8), suggesting that dephosphorylation of CAR, rather than direct ligand binding, is involved in its translocation into the nucleus.
Transcription of CYP3A is induced by a remarkable diversity of xenobiotics, including many widely used drugs. Several lines of evidence indicate that PXR mediates the induction of CYP3A expression by these compounds. First, PXR is abundantly and selectively expressed in the liver and intestine, the same tissues in which induction of CYP3A occurs (10 -13). Moreover, PXR binds as a heterodimer with RXR to xenobiotic response elements in the CYP3A gene promoters and is activated in cell-based reporter assays by the range of xenobiotics that are known to induce CYP3A expression (10 -13). Finally, PXR orthologs from different species are differentially activated by xenobiotics with a profile that correlates with the speciesspecific induction of CYP3A gene expression (11,14).
Although CAR and PXR have been implicated as primary regulators of CYP2B and CYP3A expression, respectively, there is evidence for cross-talk between these two nuclear receptor signaling pathways. First, the CYP2B and CYP3A genes are regulated in rodent and/or human hepatocytes by * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ several of the same xenobiotics, including PB, clotrimazole, and rifampicin (15,16). Moreover, CAR was recently shown to transactivate through the CYP3A4 xenobiotic response element that serves as a PXR/RXR binding site (9). In order to clarify the relative functions of CAR and PXR, we have systematically compared the activities of various xenobiotics on the human (h) and mouse (m) orthologs of CAR and PXR using cell-based transactivation and in vitro radioligand binding assays. Our results demonstrate an unexpected overlap in the compounds that regulate these two orphan receptors.

EXPERIMENTAL PROCEDURES
Reagents-RU486 was purchased from Biomol (Plymouth Meeting, PA). SR12813 was synthesized in house. All other xenobiotics and steroids used in the transfection and binding assays were purchased from either Sigma or Steraloids, Inc. (Wilton, NH). [ 3 H]Clotrimazole (specific activity 78 Ci/mmol) was prepared by Amersham Pharmacia Biotech (Cardiff, United Kingdom).
Cotransfection Assays-CV-1 cells were plated in 96-well plates at a density of 20,000 cells/well in Dulbecco's modified Eagle's medium high glucose medium supplemented with 10% charcoal/dextran treated fetal bovine serum (HyClone, Logan, UT). Transfection mixes contained 5 ng of receptor expression vector, 20 ng of reporter plasmid, 12 ng of ␤-actin SPAP as internal control, and 43 ng of carrier plasmid. Human and mouse PXR expression plasmids and the XREM-CYP3A4-LUC reporter, containing the enhancer and promoter of CYP3A4 driving luciferase expression, were as described previously (10,11,17). Transfections were performed with LipofectAMINE (Life Technologies, Inc.) essentially according to the manufacturer's instructions. Drug dilutions were prepared in phenol red-free Dulbecco's modified Eagle's medium/ F-12 medium with 15 mM HEPES supplemented with 10% charcoalstripped, delipidated calf serum (Sigma). Cells were incubated for 24 h in the presence of drugs, after which the medium was sampled and assayed for alkaline phosphatase activity. Luciferase reporter activity was measured using the LucLite assay system (Packard Instrument Co., Meriden, CT).
Overexpression and Purification of CAR and PXR LBDs-The LBDs of mouse and human CAR were expressed in Escherichia coli strain BL21(DE3) as polyhistidine-tagged fusion proteins. Sequence encoding a modified polyhistidine tag (MKKGHHHHHHG) was fused in frame to sequence encoding residues 113-358 of mCAR (5) or 103-348 of hCAR (4) and subcloned into the pRSETa expression vector (Invitrogen). Cells were grown at 25°C in LB medium with 0.1 mg/ml carbenicillin for approximately 20 h before harvesting. Routinely, 60 -70 g of cell paste were resuspended in 400 -450 ml of TBS (25 mM Tris, 150 mM NaCl, pH 8.0). Cells were lysed by passing three times through an APV Rannie MINI-lab homogenizer, and cell debris was removed by centrifugation (30 min, 20,000 ϫ g, 4°C). After adjusting the imidazole concentration to 50 mM, the lysate was loaded onto a column packed with Sepharose (Ni 2ϩ -charged) chelation resin (Amersham Pharmacia Biotech) and preequilibrated with TBS plus 50 mM imidazole. The column was washed with approximately one column volume of TBS containing 72.5 mM imidazole. The recombinant protein was eluted with a gradient from 50 to 500 mM imidazole. Column peak fractions were pooled immediately and diluted 5-7-fold with 25 mM Tris, pH 8.0, containing 5% 1,2propanediol, 0.5 mM EDTA, and 5 mM DTT. The diluted protein sample was then loaded onto a column packed with Poros HQ resin (PerSeptive Biosystems) and eluted with a gradient from 0 to 500 mM NaCl. Peak fractions were pooled and concentrated using Centriprep YM-10 filter devices (Amicon) and subjected to size exclusion, using a column packed with Superdex-75 resin (Amersham Pharmacia Biotech) pre-equilibrated with TBS containing 5% 1,2-propanediol, 0.5 mM EDTA, and 5 mM DTT. The purified CAR LBDs were biotinylated for use in scintillation proximity assays. hPXR was coexpressed with an 88-amino acid fragment of SRC-1 and purifed to homogeneity for use in radioligand binding assays as described previously (14).
Fluorescence Resonance Energy Transfer (FRET) Ligand Sensing Assay-The FRET ligand sensing assay was performed by modification of a previously published procedure (18). Biotinylated mouse and human CAR LBDs were labeled with the steptavidin-conjugated fluorophore allophycocyanin. Labeled CAR LBDs were incubated with a peptide that included the second LXXLL motif of SRC-1 (amino acids 676 -700) labeled with europium chelate. Data were collected with a Wallac Victor fluorescence reader in a time-resolved mode as described previously (18) and the fluorescence ratio calculated.
Scintillation Proximity Binding Assay (SPA)-Streptavidin polyvi-nyltoluene SPA beads (Amersham Pharmacia Biotech) were resuspended in assay buffer (50 mM Tris, 50 mM KCl, 1 mM CHAPS, 0.1 mg/ml bovine serum albumin, 10 mM DTT, pH 8.0). Biotinylated hCAR was added to a final concentration of 50 nM. The receptor/bead mixture was allowed to incubate at room temperature for 15 min with gentle agitation. Uncoupled receptor was removed by centrifuging the receptor/bead mixture at 2500 rpm for 10 min and pouring off the supernatant. Receptor-coupled beads were resuspended in assay buffer and a second wash performed. All experiments were run in Costar 96-well SPA plates using an assay volume of 100 l. Typically, reactions included receptor-coated beads, 10 nM [ 3 H]clotrimazole, and the indicated concentrations of competitor compounds. SPA plates were incubated for 1 h at room temperature and counted in a Wallac 1450 Microbeta counter. PXR scintillation proximity binding assays were performed as described previously (14).

Activity of Xenobiotics and Steroids on CAR and PXR-
Previous studies have revealed an overlap in the xenobiotics that regulate the expression of CYP2B and CYP3A (15,16). We sought to determine whether there was a corresponding overlap in the activities of xenobiotics on CAR and PXR. Accordingly, the mouse and human CAR and PXR were tested in transfection assays in CV-1 cells with xenobiotics that are well established inducers of CYP2B and/or CYP3A gene expression. The xenobiotics evaluated include PB, TCPOBOP, the antibiotic rifampicin, the synthetic pregnane pregnenolone 16␣-carbonitrile (PCN), the synthetic glucocorticoid dexamethasone, the antiprogestin RU486, the antimycotic clotrimazole, and the cholesterol-lowering drug SR12813. We also included the sex steroid metabolites 5␤-pregnane-3,20-dione and 5␣,17␣-androstanol, which activate and deactivate PXR and CAR, respectively. As a reporter, we used the recently described XREM-CYP3A4-LUC construct containing the enhancer (nucleotides Ϫ7836 to Ϫ7208) and promoter (nucleotides Ϫ362 to ϩ53) of CYP3A4 driving luciferase gene expression (17). This reporter construct is responsive to both CAR and PXR (see below). 2 All compounds were initially screened at a concentration of 10 M, except for PB, which was tested at 0.5 mM. As described previously (14), rifampicin, RU486, clotrimazole, SR12813, PB, and 5␤-pregnane-3,20-dione were efficacious activators of hPXR (Fig. 1A). PCN and dexamethasone were relatively weak activators of hPXR as compared with the other compounds. PCN, dexamethasone, RU486, and clotrimazole all activated mPXR on the XREM-CYP3A4-luciferase reporter with comparable efficacy. Interestingly, SR12813 also induced expression from the XREM-CYP3A4-luciferase reporter. We previously showed that this compound had little activity on mPXR when the (DR3) 2 -tk-CAT reporter was used (14). Thus, the effects of SR12813 are promoter-specific.
PXR assays performed with xenobiotics known to regulate CAR revealed several interesting results. First, the CAR deactivator androstanol activated both the hPXR and mPXR ( Fig.  1A and Table I). Moreover, TCPOBOP was an efficacious activator of hPXR but had only weak effects on mPXR. Thus, two compounds with established CAR activity also had marked effects on PXR.
The same panel of compounds was next tested on CAR. As expected, androstanol was an efficacious deactivator of mCAR (Fig. 1B). Androstanol also deactivated hCAR, although the effect was not as pronounced as on mCAR. By contrast, the antimycotic clotrimazole was an efficacious deactivator of hCAR but not mCAR. Full dose-response analysis revealed that clotrimazole deactivated hCAR with an EC 50 value of ϳ700 nM ( Fig. 1C and Table I). The PXR activators rifampicin, PCN, dexamethasone, RU486, and SR12813 had little or no effect on CAR activity (Fig. 1B). PB caused a weak but reproducible 2 B. Goodwin and C. Liddle, unpublished data. deactivation of both hCAR and mCAR at high concentrations.
TCPOBOP was a potent (EC 50 ϳ100 nM) and efficacious activator of mCAR but had virtually no activity on hCAR (Fig.  1, B and D, and Table I). TCPOBOP was previously shown to activate mCAR, albeit weakly, on a reporter construct driven by a PB-responsive enhancer module from the CYP2B gene (8). Interestingly, 5␤-pregnane-3,20-dione activated hCAR ϳ2-fold (Fig. 1, B and D) but had only weak effects on mCAR. These data suggest that natural ligands may exist that increase CAR activity above its high constitutive activity.
Compounds Induce Conformational Changes in the CAR LBD-Androstanol has been shown previously to inhibit the interaction between mCAR and a polypeptide derived from SRC-1 (6). We next tested whether the xenobiotics and steroids that modulated CAR activity in the transfection assay also affected CAR/SRC-1 interactions. A FRET ligand-sensing assay (18) was developed for hCAR. The LBD of hCAR was expressed in E. coli, purified to homogeneity, biotinylated, and labeled with the steptavidin-conjugated fluorophore allophycocyanin. Labeled hCAR LBD was incubated with a peptide that included the second LXXLL motif of SRC-1 (amino acids 676 -700) labeled with europium chelate. As expected, hCAR LBD interacted efficiently with SRC-1 in the absence of compound, and androstanol disrupted this interaction ( Fig. 2A). Clotrimazole was a potent inhibitor of the hCAR/SRC-1 interaction (IC 50 ϳ 100 nM) ( Fig. 2A). We note that clotrimazole was ϳ10-fold more potent in the FRET assay ( Fig. 2A) than in transfection assays (Fig. 1C). This discrepancy may be due to cell permeability or the metabolism of clotrimazole in cells. Consistent with its ability to activate hCAR in transfection assays, 5␤pregnane-3,20-dione caused a marked increase in the binding of SRC-1 to hCAR ( Fig. 2A). Thus, compounds can be identified that induce conformational changes in the hCAR LBD and either enhance or disrupt its basal interaction with SRC-1.
A similar FRET assay was developed for mCAR. In contrast to the hCAR assay, no basal interaction was observed between mCAR and the SRC-1 peptide. We also failed to detect any basal interaction between mCAR and three other peptides con-PXR (A) or mouse or human CAR (B) and the XREM-CYP3A4-luciferase reporter. Cells were treated with 10 M amounts of each compound, except for phenobarbital, which was tested at 0.5 mM. Cell extracts were subsequently assayed for luciferase activity. Data represent the mean of assays performed in triplicate Ϯ S.E. and are plotted as fold activation relative to transfected cells treated with vehicle alone. Cotransfection of hCAR or mCAR expression plasmids with the CYP3A4-XREMluciferase plasmid increased reporter levels 3.3-and 5.2-fold, respectively, relative to transfection with reporter plasmid alone. Full dose-response curves are shown for deactivators (C) or activators (D) of human or mouse CAR.  taining LXXLL motifs derived from either SRC-1 or the coactivator CBP (data not shown). The reason for this difference between the mouse and human CAR is unclear, but it may be due in part to an amino acid change (Gln 3 Gly) in the AF-2 helix. The lack of a basal interaction between these coactivator peptides and mCAR precluded us from testing whether androstanol disrupted coactivator-mCAR interactions. However, TCPOBOP resulted in a marked increase in the interaction of SRC-1 with mCAR (Fig. 2B). These data provide strong evidence that TCPOBOP activates mCAR by binding directly to the receptor's LBD. We note that PB had no effect on SRC-1 interactions with either mouse or human CAR at concentrations up to 1 mM (data not shown), suggesting that PB and TCPOBOP may activate CAR through distinct mechanisms. Xenobiotics and Steroids Bind hCAR-The potency of clotrimazole in hCAR cell-based reporter and in vitro FRET assays suggested that [ 3 H]clotrimazole might be used to develop a radioligand binding assay for this orphan receptor. To test this idea, an SPA (19) was developed using the purified, biotinylated hCAR LBD and streptavidin-coated polyvinyltoluene SPA beads. [ 3 H]Clotrimazole bound specifically to hCAR with a K i of ϳ100 nM (Fig. 3A). Dexamethasone, which has no effect on hCAR in transfection assays, failed to compete with [ 3 H]clotrimazole for binding to hCAR (Fig. 3A). No specific binding of [ 3 H]clotrimazole to the SPA beads was detected in the absence of hCAR (data not shown). These data suggest that clotrimazole mediates its effects on hCAR activity through direct interactions with the receptor LBD.
The availability of radioligand binding assays for both hCAR and hPXR allowed us to directly compare the binding of compounds to these receptors. In agreement with the transfection data, the sex steroid metabolites androstanol and 5␤-pregnane-3,20-dione competed efficiently with [ 3 H]clotrimazole for binding to hCAR. TCPOBOP, dexamethasone RU486, rifampicin, PCN, and SR12813, which had little or no activity on hCAR in the transfection assay, did not compete in the binding assay (Fig. 3B). PB, which deactivated hCAR in the transfection studies, did not compete at a 1 mM concentration with [ 3 H]clotrimazole for binding to hCAR, although it was a weak competitor of [ 3 H]SR12813 for binding to hPXR. These data suggest that PB does not mediate its effects on hCAR through direct interactions with the receptor, but rather through cellular metabolism of PB to a compound that can bind CAR or via an indirect mechanism. The latter hypothesis is supported by the finding that PB effects on CYP2B gene expression are blocked by the phosphatase inhibitor okadaic acid (8). As expected, RU486, clotrimazole, SR12813, and 5␤-pregnane-3,20dione competed efficiently with [ 3 H]SR12813 for binding to hPXR (Fig. 3B). The hCAR ligands androstanol and TCPOBOP also bound to hPXR (Fig. 3B). We conclude that many of the structurally diverse compounds that modulate CAR and PXR activity do so through direct interactions with the LBDs of these orphan receptors.
PXR Is the Dominant Regulator of CYP3A4 in Vitro-Since both clotrimazole and androstanol are ligands that activate hPXR and deactivate hCAR, we used these compounds to test which receptor would be dominant on the XREM-CYP3A4luciferase reporter. CV-1 cells were transfected with a fixed amount of hCAR expression plasmid (5 ng) and increasing amounts of the hPXR expression plasmid (0.1-5 ng), and the cells were treated with 10 M clotrimazole or androstanol. Even at a 50-fold excess of hCAR to hPXR expression plasmid, the negative effects of clotrimazole and androstanol on hCAR were neutralized by their positive effects on hPXR (Fig. 4). At a 20-fold excess of hCAR to hPXR, there was a net activation of reporter activity in response to clotrimazole (Fig. 4). Thus, the negative effects of these compounds on hCAR activity are not sufficient to overcome their positive effects on hPXR. We conclude that, in a cell-based assay in which both hCAR and hPXR are present in the nucleus, hPXR is the dominant regulator of CYP3A4 expression. These data agree with previous results showing that clotrimazole is a net activator of CYP3A4 gene expression in primary human hepatocytes (20). We note that the regulation of CYP3A4 expression is likely to be much more complex in the intact liver, where CAR has been reported to be sequestered in the cytoplasm in the absence of activators (8). DISCUSSION The orphan nuclear receptors CAR and PXR have recently been implicated in the induction of CYP expression (1-3). However, the relative contributions of these orphans in the regulation of particular CYPs have remained unclear. In this report, we have systematically compared the effects of a panel of xenobiotics with established activities on CYP expression on the the mouse and human orthologs of CAR and PXR. Our results demonstrate a surprising degree of overlap in the compounds that are capable of binding and regulating the activities of these orphan receptors. In some cases, xenobiotics have opposing effects on these orphans; for example, clotrimazole activates hPXR but deactivates hCAR. Since CAR and PXR can bind to some of the same response elements and are likely to share many target genes, the net effect of a xenobiotic on gene transcription will often be complex and depend on its effects on both CAR and PXR.
Nuclear receptor orthologs generally share Ͼ90% amino acid identity in their LBDs. However, comparison of PXR from four different species shows that this receptor has diverged considerably during the course of evolution (14). The human, rabbit, and rodent PXR are all roughly equally divergent and share only ϳ70% amino acid identity. This divergence in PXR is an important component of cross-species differences in the regu-lation of CYP3A expression by xenobiotics (14). Similarly, the human and rodent CAR share only ϳ70% amino acid identity in their LBDs (4,5) and demonstrate marked differences in their responses to xenobiotics. For example, clotrimazole is an efficacious deactivator of hCAR but has little or no activity on mCAR. Conversely, TCPOBOP is a potent activator of mCAR but lacks any activity on hCAR. The divergence in amino acid sequences across CAR orthologs undoubtedly contributes to cross-species differences in the physiologic effects of xenobiotics.
We have developed a high throughput scintillation proximity binding assay for hCAR and used it to demonstrate that clotrimazole and the sex steroid metabolites androstanol and 5␤pregnane-3,20-dione bind directly to this orphan receptor. These data establish CAR as a bona fide receptor. Interestingly, PB, which induces CYP2B gene expression in human hepatocytes (7,9), does not bind to hCAR at concentrations up to 1 mM. These data suggest that PB may activate CAR through an indirect mechanism such as phosphorylation. The recent report that the phosphodiesterase inhibitor okadaic acid blocks the CAR-mediated effects of PB on CYP2B expression supports this idea (8). Thus, there appear to be multiple mechanisms for the regulation of CAR activity.
Does CAR have a natural ligand, or has it evolved as part of a defense mechanism against xenobiotics? The androstanes were originally shown to deactivate mCAR in cell-based assays, lowering its transcriptional activity below the basal level (6). We now show that the progesterone metabolite 5␤-pregnane-3,20-dione binds to hCAR and activates the receptor above its high basal level of activity. Although superphysiological concentrations of 5␤-pregnane-3,20-dione and androstanol are required for these effects, the data raise the intriguing possibility that physiological steroidal ligands exist that can either activate or deactivate CAR. From an evolutionary perspective, it is interesting that both CAR and PXR regulate the expression of CYPs, which are involved in the oxidative metabolism of natural steroid hormones as well as xenobiotics. It may be that CAR and PXR are components of a regulatory network that governs endocrine hormone homeostasis. The availability of high throughput binding assays and robust cell-based assays for CAR and PXR will aid in the search for the natural ligands for these orphan receptors.