Novel CAR-mediated mechanism for synergistic activation of two distinct elements within the human cytochrome P450 2B6 gene in HepG2 cells.

The constitutive active receptor (CAR) regulates the induction of the cytochrome P450 2B6 (CYP2B6) gene by phenobarbital-type inducers, such as 1,4 bis[2-(3,5-dichloropyridyloxy)]benzene (TCPOBOP) via the distal phenobarbital-responsive enhancer module (PBREM, at -1732/-1685 bp). Activation of the PBREM by TCPOBOP generated a 10-fold induction of CYP2B6 mRNA in HepG2 cells stably expressing mouse CAR (Ym17). Co-treatment with the protein phosphatase inhibitor okadaic acid (OA) synergistically increased this induction over 100-fold without directly activating CAR or the PBREM. Although OA synergy required the presence of PBREM, deletion assays delineated the OA-responsive activity to a proximal 24-bp (-256/-233) sequence (OARE) in the CYP2B6 promoter. CAR did not directly bind to the OARE in electrophoretic mobility shift assays. However, both DNA affinity and chromatin immunoprecipitation assays showed a significant increase in CAR association with the OARE after co-treatment with TCPOBOP and OA, indicating the indirect binding of CAR to the OARE. The two cis-acting elements, the distal PBREM and the proximal OARE, within the chromatin structure are both regulated by CAR in response to TCPOBOP and OA, respectively, to maximally induce the CYP2B6 promoter. This functional interaction between the two sites expands the current understanding of the mechanism of CAR-mediated inducible transcription.

The constitutive active receptor (CAR) regulates the induction of the cytochrome P450 2B6 (CYP2B6) gene by phenobarbital-type inducers, such as 1,4 bis[2-(3,5-dichloropyridyloxy)]benzene (TCPOBOP) via the distal phenobarbital-responsive enhancer module (PBREM, at ؊1732/؊1685 bp). Activation of the PBREM by TCPOBOP generated a 10-fold induction of CYP2B6 mRNA in HepG2 cells stably expressing mouse CAR (Ym17). Co-treatment with the protein phosphatase inhibitor okadaic acid (OA) synergistically increased this induction over 100-fold without directly activating CAR or the PBREM. Although OA synergy required the presence of PBREM, deletion assays delineated the OA-responsive activity to a proximal 24-bp (؊256/؊233) sequence (OARE) in the CYP2B6 promoter. CAR did not directly bind to the OARE in electrophoretic mobility shift assays. However, both DNA affinity and chromatin immunoprecipitation assays showed a significant increase in CAR association with the OARE after co-treatment with TCPOBOP and OA, indicating the indirect binding of CAR to the OARE. The two cis-acting elements, the distal PBREM and the proximal OARE, within the chromatin structure are both regulated by CAR in response to TCPOBOP and OA, respectively, to maximally induce the CYP2B6 promoter. This functional interaction between the two sites expands the current understanding of the mechanism of CAR-mediated inducible transcription.
The nuclear receptor CAR 1 transcriptionally modifies the expression of genes involved in the metabolism and elimination of xenobiotics and endogenous compounds, such as cytochrome P450 (CYP), in response to xenochemical exposure. The recep-tor is essential for regulating the induction of a set of CYP genes, particularly the CYP2B subfamily, by phenobarbital (PB) and PB-like inducers (e.g. chlorpromazine (CPZ) and 1,4bis [2-(3,5-dichloropyridyloxy)]benzene (TCPOBOP)) (1)(2)(3)(4)(5). As a heterodimer with the retinoid X receptor (RXR), CAR binds to direct repeats of nuclear receptor half-sites, such as the NR1 within the distal 51-bp PB-responsive enhancer module (PBREM) of Cyp2b10, and activates transcription (1,6,7). Experiments with transgenic mice bearing a mutated PBREM have confirmed that it is an essential enhancer element (8). In addition, several PB-responsive DNA elements have been reported in the proximal promoters of CYP2B genes, but their role as enhancer elements in PB induction is controversial (reviewed in Ref. 9). Thus, the exact mechanism how CAR regulates inducible transcription still remains elusive.
Signal transduction has been repeatedly suggested to regulate PB induction. Various inhibitors of protein kinases or phosphatases often repress PB induction in hepatocytes (10). Translocation of CAR from the cytoplasm to the nucleus is an essential step in PB induction (6) and has recently been shown to involve a phosphorylation/dephosphorylation cascade (11). The protein phosphatase inhibitor okadaic acid (OA) effectively repressed the induction by PB of CYP2B genes in rodent primary hepatocytes (12,13) by preventing CAR nuclear translocation (6,11). However, distinct activation processes occur in the nucleus, for example the Ca 2ϩ /calmodulin-dependent kinase inhibitor KN-62 repressed the induction of CYP2B10 by PB in mouse primary hepatocytes despite CAR accumulation in the nucleus (14,15). It is possible that CAR interacts with other factors in the nucleus that are the target of phosphorylation because direct phosphorylation of CAR has not, as yet, been reported. Indirect mechanisms of gene induction by nuclear receptors, including the retinoic acid receptor (RAR), RXR, and the glucocorticoid receptor, independent of nuclear receptor half-site activation, have been reported previously (16,17) but no such mechanism has been reported for CAR at present.
Dissecting the CAR-mediated response has been challenging because in transformed cell lines CAR spontaneously accumulates in the nucleus and is constitutively active (6). In the present investigation, we used a stable line (Ym17) of HepG2 cells that express mouse CAR-V5-His primarily in the nucleus that can be effectively activated by CAR ligands. These cells provided an opportunity to examine directly the function of the receptor in the nucleus, independent of translocation, and are an excellent system for investigating the nuclear effects of inhibitors (e.g. OA) on CAR-mediated transcription. Here we present experimental considerations to reveal the role of a proximal region of the CYP2B6 gene in synergistically up-regulating CAR-mediated transcription in response to TCPOBOP and OA. These investigations shed light on a unique role for a proximal sequence in maximizing PB-type induction and identify a novel role for CAR in the formation of a proximal protein-DNA complex.
Cloning and Plasmid-The 1.8-kbp 5Ј-flanking DNA of the CYP2B6 gene was cloned between the KpnI and XhoI sites of basic firefly luciferase reporter plasmid pGL3-Basic (Ϫ1.8k-pGL3). To construct deletions, either proper fragments were generated from the 1.8-kbp DNA using convenient restriction sites or from the Ϫ1.8k-pGL3 plasmid using the Stratagene QuikChange® site-directed mutagenesis kit (Cedar Creek, TX) and specific primers spanning the excised bases (and containing the insert sequence for UGT256233). Full-length mouse CAR cDNA was cloned into pBIND vector (pBIND-mCAR). All constructs were verified by their sequences.
Cell Culture-Cells were cultured in minimal essential medium supplemented with 10% fetal bovine serum and antibiotics (100 unit/ml of penicillin and 100 g/ml of streptomycin). Ym17 cells, a stable cell line that expressed mCAR, and Yh18 cells, a stable cell line that expressed hCAR, were established from HepG2 cells that were transfected with a mouse or human pcDNA3.1-CAR-V5-His expression vector, respectively, and were selected for neomycin resistance. During cell treatment, OA was added to the culture medium 30 min before treatment with a given chemical inducer.
Transfection Assays-For transient transfection assays, Ym17 or Yh18 cells in 24-well plates were co-transfected with a given pGL3 luciferase reporter plasmid and pRL-SV40 plasmid (0.1 g/well each) by the calcium phosphate co-precipitation method using CellPhect Transfection Kit (Amersham Biosciences). Sixteen hours after transfection, the cells were treated with chemical inducers. Luciferase activity was measured in cell lysates using the dual-luciferase reporter assay system (Promega, Madison, WI). The receptor-dependent enhancer activity was determined based on firefly luciferase activity normalized against Renilla luciferase activity, except for the Ym17S15 cells when firefly luciferase activity was normalized against protein concentration (g/ ml). Protein concentration was measured at 595 nm using Bio-Rad protein assay reagent (Bio-Rad).
Nuclear Extraction-Ym17 cells were harvested from four confluent 75-cm 2 tissue culture flasks for each treatment group, and nuclear extracts were prepared using the procedure described previously (19). The extracts were dialyzed as described previously (21).
DNase I Protection Assays and Electrophoretic Mobility Shift Assays-DNase I protection assays were carried out as described previously (22). The DNA fragments were separated on standard sequencing gels, which were dried and autoradiographed. Electrophoretic mobility shift assay binding was performed as described previously (1) for 25 min with 5 g of protein to 32 P-labeled double-stranded oligonucleotides corresponding to Ϫ252/Ϫ237 bp of the CYP2B6 5Ј-flanking region (5Ј-catgACCCACACATTCACTT-3Ј) or to the NR1 motif from the Cyp2b10 PBREM (5Ј-gatcTCTGTACTTTCCTGACCTTG-3Ј).
DNA Affinity Chromatography-For affinity purification of CAR, Dynabeads® M-280 streptavidin (Dynal ASA, Oslo, Norway) conjugated with multiple copies of the wild-type or mutant (5Ј-catgAC-CCAAAAAAACACTT-3Ј) Ϫ252/Ϫ237-bp oligonucleotide were prepared as described previously (1). The bound proteins were eluted with 50 l of 20 mM Hepes, pH 7.6, containing 0.5 M NaCl, 20% glycerol, 0.2 mM EDTA, and 1 mM Na 2 MO 4 . Chromatin Immunoprecipitation-6 ϫ 10 6 Ym17 cells per treatment group were incubated with TCPOBOP (250 nM) and/or OA (10 nM) for 24 h. Chromatin immunoprecipitation was performed by using the chromatin immunoprecipitation (ChIP) assay kit (Upstate Biotechnologies, Charlottesville, VA), according to the manufacturer's instructions, and adapted as follows. Proteins and DNA were cross-linked with 1% formaldehyde in the media for 20 min at 37°C. The cells were collected in 450 l of SDS lysis buffer containing protease inhibitors aprotinin (1 g/ml), phenylmethanesulfonyl fluoride (1 M), and pepstatin A (1 g/ml) and were sonicated at power 10 for five 10-s pulses using the Misonix Microson TM XL 2000 ultrasonic cell disrupter. The sonication was optimized as described in the ChIP assay kit instructions and examined by agarose gel electrophoresis to determine generation of DNA fragments between 200 and 1000 bp in length, which avoided nonspecific immunoprecipitation of the OARE via cross-linking of CAR to the PBREM. The diluted lysates were precleared by incubation with 160 l of protein A-agarose salmon sperm slurry, 60 l of normal mouse IgG, and 0.05% bovine serum albumin for 2 h at 4°C. The agarose was collected by centrifugation at 100 ϫ g for 2 min at 4°C. 1 ml of the cleared fraction was subjected to overnight immunoprecipitation with either 5 g of V5 antibody, 5 g of normal mouse IgG, or no antibody at 4°C. 1 ml was retained as an input sample for each treatment group. The immunoprecipitated complexes were recovered, washed, and eluted; the cross-links were reversed and proteins digested as described in the manufacturer's protocol. A phenol chloroform/ethanol precipitation method purified the DNA with a final resuspension in 45 l of water. The DNA was amplified by real time PCR as described under "RT-PCR" using a specific primer and probe set designed to span the OARE between Ϫ287 and Ϫ193-bp upstream of the CYP2B6 transcription start FIG. 2. a, no synergistic induction of the UGT1A1, CYP3A7, and SOD3 genes. Ym17 cells were treated with TCPOBOP (250 nM) and/or OA (10 nM) for 24 h. b, synergistic induction by co-treatment with OA and CPZ or E2. Ym17 cells were treated with CPZ (10 M) or E2 (10 M) with or without OA (10 nM) for 24 h. Total cellular RNAs were prepared from the treated cells and subjected to quantitative real time RT-PCR of UGT1A1, CYP3A7, SOD3, or CYP2B6 mRNA. The levels of target gene mRNAs were normalized by the ␤-actin mRNA levels, and fold induction was calculated relative to the levels in Me 2 SO (DMSO)treated cells.
FIG. 3. a, no synergistic CAR activation in a Gal4-based assay. HepG2 cells were co-transfected with a luciferase reporter gene containing the Gal4 DNA-binding site and mCAR fused with Gal4 (DNAbinding domain) expression plasmid (0.2 g/well each). Cells were treated with TCPOBOP (250 nM) with or without OA (10 nM) for 24 h. Activation activity was expressed as fold induction, calculated relative to the control (Me 2 SO (DMSO)) value of 1. b, OA does not activate CYP2B6 transcription via the PBREM. The luciferase reporter containing five copies of the Cyp2b10 NR1 motif (NR1 5 -tk-pGL3) was cotransfected with pRL-SV40 into Ym17 cells. c, OA synergy is dependent on the 1.8-kbp DNA upstream of the CYP2B6 transcriptional start site and OA-mediated phosphatase inhibition. Ϫ1.8k-pGL3 represents the luciferase reporter containing 1.8-kbp upstream of the CYP2B6 transcription start site. The Ϫ1.8k-pGL3 plasmid was co-transfected with pRL-SV40 into Ym17 cells. The transfected cells were incubated with TCPOBOP (250 nM) and/or OA (10 nM), okadaol (10 nM), or nor-okadaone (NO) (10 nM) for 24 and 48 h (for b and c, respectively), harvested, and assayed for luciferase activity. Activity levels were expressed as fold induction relative to the control (Me 2 SO) value of 1.
site, with the probe located within the OARE sequence. This amplicon was specific for the OARE and did not contain other CAR-binding sites within the CYP2B6 gene, such as the PBREM. The OARE 5Ј-and 3Ј-primers were 5Ј-CAAAAATAGACATACATATACCCACAAACC-3Ј and 5Ј-ATTGGCAGGGATACAGTGGTAGAG-3Ј, respectively, and the probe was 6FAM-CACTTGCTCACCTGGACT-MGB. The precipitated DNA levels were normalized to the input levels using the following equation (2[caret]Ct input Ϫ Ct precipitate ).

RESULTS
Synergistic Induction of the CYP2B6 Gene-In Ym17 cells, the endogenous CYP2B6 gene was induced by TCPOBOP 8 -10fold (Fig. 1a). Unexpectedly, 24 h of co-treatment with OA and TCPOBOP synergistically up-regulated the induction more than 100-fold. OA alone induced CYP2B6 mRNA only 3-fold. This synergistic up-regulation became greater in a time-dependent fashion as the induction by TCPOBOP increased (Fig.  1b) but was not observed in normal HepG2 cells, indicating mCAR was essential for synergistic induction (Fig. 1a). The mRNAs of CAR-regulated UGT1A1, CYP3A7, and SOD3 genes (4,20) were increased by TCPOBOP in Ym17 cells, yet no synergistic up-regulation was observed following co-treatment with OA (Fig. 2a). Thus, OA-dependent synergistic induction occurred specifically in the CYP2B6 gene. OA-dependent synergistic up-regulation, however, was not TCPOBOP-specific and may be a general phenomenon occurring in the CYP2B6 gene in response to PB-type inducers because induction by CPZ and E2 was also synergistically up-regulated from 2-and 4-fold to 27-and 57-fold, respectively, in the presence of OA (Fig. 2b).
Role of OA in Synergistic Induction-To determine whether OA directly activated mCAR, a Gal4-based mammalian onehybrid assay was performed. TCPOBOP enhanced Gal4-reporter activity in HepG2 cells co-transfected with an mCAR-GAL4 DNA binding domain fusion plasmid, whereas additional co-treatment with OA caused no synergistic up-regulation of the reporter gene (Fig. 3a). Thus, OA does not directly regulate mCAR to synergistically induce the CYP2B6 gene. To determine whether OA activated CYP2B6 transcription via the PBREM, a luciferase reporter gene containing five copies of the Cyp2b10 NR1 motif upstream of a thymidine kinase promoter was transfected into Ym17 cells. TCPOBOP enhanced luciferase reporter activity 9-fold, whereas additional co-treatment with OA caused no synergistic up-regulation of the reporter gene (Fig. 3b). However, when the Ϫ1.8-kbp 5Ј-flanking sequence of the CYP2B6 gene that included the PBREM was cloned into a luciferase reporter gene (Ϫ1.8k-pGL3) and transfected into Ym17 cells, TCPOBOP activated the reporter gene 3-fold (Fig. 3c). Moreover, the activity was synergistically upregulated 12-fold after co-treatment with OA (Fig. 3c). The Ϫ1.8-kbp DNA fragment retained its ability to be synergistically activated by OA in the presence of TCPOBOP. No synergistic up-regulation of the reporter gene was observed after co-treatment with TCPOBOP and the inactive analogues of OA, nor-okadaone, or okadaol (Fig. 3c).
Delineation of OA-responsive Activity-To delineate the cisacting region of the CYP2B6 gene directly responsible for the synergistic induction, a series of deletion assays were performed. Initially, a large internal deletion from Ϫ1685 to Ϫ307-bp within the Ϫ1.8k-pGL3 reporter gene was made to remove various elements, such as the glucocorticoid receptor element and activator protein-1 conserved in a region of the CYP2B6 gene corresponding to a Ϫ1.4/Ϫ1.2-kbp region of the rat CYP2B1/2 and mouse Cyp2b10 genes (reviewed in Ref. 9). This joining of the PBREM to the proximal 307-bp (PBREM-307) did not significantly affect the synergy (Fig. 4). However, as expected, the proximal Ϫ307-bp sequence alone did not display either activation by TCPOBOP or a synergistic response to OA, indicating that synergy was PBREM-as well as CAR-dependent.
Subsequently, deletions were constructed within the context For the other nomenclature, PBREM-307, for example, indicates deletion of the sequence between the PBREM and the proximal Ϫ307 bp, and Ϫ307 represents a deletion of the Ϫ1.8-kbp DNA except for the proximal Ϫ307 bp. UGT256233 represents replacement of the OARE with Ϫ256/Ϫ233 bp of the UGT1A1 promoter within the Ϫ1.8k-pGL3 reporter. The Ϫ1.8k-pGL3 and other -pGL3 plasmids were co-transfected with pRL-SV40 into Ym17 cells. The transfected cells were incubated with TCPOBOP (250 nM) and/or OA (10 nM) for 48 h, harvested, and assayed for luciferase activity. Activity levels were expressed as fold induction relative to the control (Me 2 SO (DMSO)) value of 1. Significant synergy was determined by a two-tailed paired Student's t test comparing the sum of induction by TCPOBOP and OA alone to induction by co-treatment; *, p Ͻ 0.05, and ***, p Ͻ 0.005. of the Ϫ1.8-kbp DNA sequence to investigate the activity in the Ϫ306/Ϫ35-bp region. Whereas deletion of Ϫ125/Ϫ35-bp retained the synergistic response, the Ϫ306/Ϫ35-bp deletion lost significant synergistic activity. The proximal Ϫ306/Ϫ125-bp region was further dissected by constructing smaller deletions within the context of the Ϫ1.8-kbp DNA as follows: Ϫ306/Ϫ166, Ϫ256/Ϫ213, and Ϫ256/Ϫ233-bp (Fig. 4), all of which lost significant OA synergistic activity. Deletions within the Ϫ256/ Ϫ233-bp region, such as the Ϫ252/Ϫ237-bp deletion in Fig. 4, all lost significant OA synergistic activity. Thus the OA-response element (OARE) was identified as a proximal 24-bp sequence between Ϫ256 and Ϫ233-bp of the Ϫ1.8-kbp DNA. The 51-bp PBREM was placed in front of the Ϫ256-bp proximal sequence of Ϫ1.8-kbp DNA (PBREM-256), by deleting the Ϫ1685/Ϫ257-bp fragment (Fig. 4), and was found to retain a significant level of synergistic response to OA, whereas deletion of a further proximal 9 bp (Ϫ1685/Ϫ247-bp, PBREM-247) lost synergistic activity, further confirming the role of the proximal Ϫ256-bp sequence in OA-responsive up-regulation of the Ϫ1.8kbp DNA. Replacement of the OARE within the Ϫ1.8k-pGL3 construct with the Ϫ256/Ϫ233-bp sequence from the nonsynergistic gene UGT1A1 to generate a negative control showed a loss of synergy. This supported the Ϫ256/Ϫ233-bp region (OARE) as the OA-responsive site.
Nuclear Protein Binding to the OARE-To investigate protein interactions with this region, DNase I protection of the Ϫ308/Ϫ121-bp fragment incubated with TCPOBOP-and OAtreated Ym17 nuclear extracts was performed. A large protected region was observed at Ϫ264/Ϫ195 bp that included the sequence identified as responsible for OA synergy (Fig. 5, left  panel). However no differences in protein binding were evident either with nuclear extracts from treated Ym17 cells or from HepG2 cells (Fig. 5, right panel). To search for a specific binding site, electrophoretic mobility shift assays were performed using probes spanning the OARE (Ϫ256/Ϫ233-bp) and smaller overlapping regions Ϫ256/Ϫ247, Ϫ252/Ϫ237, and Ϫ242/Ϫ233 bp to dissect any treatment effects on this complex. One protein band showed similar electrophoretic mobility shift profiles for both Ϫ256 probes and corresponded to a single signal observed with the Ϫ252/Ϫ237-bp probe (Fig. 6a, lanes 1 and 3). The protein binding to the Ϫ252/237-bp probe increased in the presence of OA (Fig. 6a, lane 5) and was further enhanced in the presence of both TCPOBOP and OA (lane 7). The binding was specific because it was competed by an excess of unlabeled Ϫ252/Ϫ237-bp probe (Fig. 6a, lanes 2, 4, 6, and 8). Competition increased concentration dependently between 25-and 100-fold excess. For practical reasons, hereafter we will call this putative binding complex OARE-binding protein or OABP.
Involvement of CAR in OABP Binding-As the OA synergistic up-regulation of CYP2B6 appeared to be CAR-dependent, experiments were designed to investigate the role of CAR in the binding of the OABP protein to Ϫ252/Ϫ237 bp. First, electrophoretic mobility shift assays were performed by using nuclear extract from HepG2 cells. No specific protein binding was observed for the Ϫ252/Ϫ237-bp probe, indicating that CAR was necessary for OABP binding (Fig. 6b, lanes 1-3). However, no binding of recombinant mCAR-V5-His or hRXR␣ to the Ϫ252/ Ϫ237-bp probe was observed, either alone or in combination, despite strong binding of the two receptors in combination with a probe consisting of the Cyp2b10 NR1 motif (Fig. 6c, lane 6), suggesting that CAR did not bind directly to the OARE. To investigate the possibility that CAR interacted indirectly with the OARE and whether the interaction was treatment-dependent, DNA affinity chromatography was performed using magnetic beads with multiple copies of the Ϫ252/237-bp wild-type oligonucleotides or mutant oligonucleotides with the central 6 bp mutated, attached by biotinylation, and incubated with Ym17 nuclear extract. 4% of the input nuclear extracts were electrophoresed alongside the eluted fractions and Westernblotted with the V5 antibody to detect the V5-tagged mCAR from Ym17 cells. The eluted fractions from the Ϫ252/Ϫ237-bp wild-type beads indicated a background CAR association with the Ϫ252/Ϫ237-bp region in the presence of Me 2 SO, TCPOBOP, or OA alone (Fig. 7a). However, after co-treatment with TCPOBOP and OA, CAR interaction with the Ϫ252/ Ϫ237-bp region was greatly enhanced. An extremely weak association after a much longer exposure was observed with the mutant beads, which was not changed by treatment. Thus, co-treatment of TCPOBOP and OA significantly increased CAR association with the OARE.
To test whether CAR interacted with the endogenous CYP2B6 OARE in Ym17 cells, ChIP assays were performed with V5 antibody. DNA from Ϫ287 to Ϫ193 bp, encompassing the OARE but not the PBREM, could be amplified by real time PCR from V5 antibody precipitates and at a lower level from control normal mouse IgG or no antibody precipitates. CAR interaction with the OARE was enhanced 27-fold after a 24 h co-treatment with TCPOBOP and OA (Fig. 7b). The data presented in Fig. 7b are representative of three individual experiments, all of which showed greater than 25-fold enhanced interaction of CAR with the OARE after 24 h of co-treatment, whereas treatment effects observed in the IgG and no antibody controls (such as enhancement by TCPOBOP for IgG in Fig. 7b) were not reproducible and thus were considered background variation.
ChIP DNA was also amplified by using specific primers for the unrelated gene ␤-actin, but only input signals were obtained. Moreover, real time PCR using a specific primer and probe set designed to span the PBREM showed no additional binding of CAR with OA treatment, confirming the specificity of the amplification of OARE (not shown). The results, obtained FIG. 6. a, treatment and CAR-dependent binding of a nuclear protein to the central region of the OARE. Electrophoretic mobility shift assay using double-stranded 32 P-end-labeled probe Ϫ252/Ϫ237 with 5 g of nuclear extracts from Ym17 cells incubated with TCPOBOP (250 nM) with or without OA (10 nM) for 48 h. A specific protein denoted OABP and a nonspecific protein bound to the probe. Specificity of binding was confirmed by competition with 100-fold excess of unlabeled Ϫ252/Ϫ237 probe. b, no OABP binding from HepG2 nuclear extracts. Electrophoretic mobility shift assay was as described for a with 5 g of nuclear extracts from HepG2 cells incubated with TCPOBOP (250 nM) with or without OA (10 nM) for 48 h. c, no direct CAR binding to the OARE. Recombinant mouse CAR-V5-His and human RXR␣, generated by TNT® T7 Quick-coupled transcription/translation system (Promega), were used for electrophoretic mobility shift assay using double-stranded 32 P-end-labeled probes Ϫ252/Ϫ237 or NR1 incubated with 4 l of the receptors alone or in combination.

CAR-mediated Synergistic Induction of CYP2B6
from both DNA affinity and ChIP assays, clearly demonstrate that CAR becomes a constituent of the OABP complex following co-treatment with TCPOBOP and OA, although the receptor did not directly bind to the OARE.
OA-responsive Activity Enhanced by Stable Expression of Ϫ1.8k-pGL3-There was a large discrepancy between the level of OA synergistic induction observed with CYP2B6 mRNA in Ym17 cells (Ͼ100-fold) (Fig. 1a) and the Ϫ1.8k-pGL3 construct transiently transfected into Ym17 cells (ϳ20-fold) (Fig. 3). This suggested either the existence of a further enhancer region upstream of 1.8-kbp or that chromatin structure may be involved in the synergistic response. Since the deletion work performed did not support the existence of a further enhancer upstream, the latter theory was tested by stably introducing the Ϫ1.8k-pGL3 construct into Ym17 cells. A colony named Ym17S15 was selected and tested for OA synergistic activity. A 97-fold induction of luciferase activity by TCPOBOP and OA co-treatment was observed, suggesting that chromatin struc-ture is required for a full OA synergistic response (Fig. 8a).
Synergistic Induction of CYP2B6 by OA and a Human CAR Activator-To investigate whether human CAR could mediate the synergistic up-regulation of the CYP2B6 gene, the Ϫ1.8k-pGL3 reporter and the Ϫ256/Ϫ233 OARE deletion construct were transfected into Yh18 cells, a HepG2 cell line stably expressing pcDNA3.1-hCAR-V5-His. Human CAR activator CITCO weakly activated Ϫ1.8k-pGL3 alone, but the activity was synergistically up-regulated to 3-fold when co-treated with OA (Fig. 8b). No enhancement of reporter activity was observed FIG. 7. Co-treatment with TCPOBOP and OA significantly increased CAR association with the OARE. a, DNA affinity purification. 120 g of nuclear extracts from Ym17 cells treated with TCPOBOP (250 nM) and/or OA (10 nM) for 48 h were incubated with magnetic beads conjugated with multiple copies of either wild-type or mutant Ϫ252/ Ϫ237-bp oligonucleotide. 4% of the input nuclear extracts and 25 l of the eluted proteins were resolved on a 4 -12% BisTris-polyacrylamide gel and immunoblotted with V5 antibody (1:5000 dilution), as described previously (19), to detect the association of V5-His-tagged mouse CAR with the central region of the OARE. b, chromatin immunoprecipitation was performed on 6 ϫ 10 6 Ym17 cells treated with TCPOBOP (250 nM) and/or OA (10 nM) for 24 h using either V5-antibody, normal mouse IgG, or no antibody. The purified DNA was amplified by real time PCR using probe and primers specific to the OARE (amplicon Ϫ287 to Ϫ193-bp) and normalized to input samples from each treatment group. DMSO, Me 2 SO.
FIG. 8. a, synergistic activation of Ϫ1.8 kbp-luciferase reporter gene stably incorporated into Ym17 cells. Ϫ1.8k-pGL3 plasmid of the CYP2B6 promoter was co-transfected with pCDNA3.1/Hygro© plasmid into Ym17 cells. Ym17S15 cells were selected using media containing 200 g/ml hygromycin B. The cells were incubated with TCPOBOP (250 nM) and/or OA (10 nM) for 48 h, harvested, and assayed for luciferase activity and total protein concentration. The luciferase activity was normalized per g/ml of protein and is expressed as fold induction relative to the control (Me 2 SO (DMSO)) value of 1. b, synergistic activation of Ϫ1.8 kbp luciferase reporter gene by human CAR activator CITCO and OA. Luciferase reporters containing either the entire 1.8kbp upstream of the CYP2B6 transcription start site (Ϫ1.8k-pGL3) or the 1.8-kbp DNA with the OARE sequence (Ϫ256/Ϫ233-bp) deleted were transiently co-transfected with pRL-SV40 into Yh18 cells. The transfected cells were incubated with CITCO (250 nM) and/or OA (10 nM) for 48 h, harvested, and assayed for luciferase activity. Activity levels were expressed as fold induction relative to the control (Me 2 SO) value of 1.
by OA in the absence of the OARE. Thus the Ϫ1.8-kbp DNA fragment can be synergistically regulated by hCAR. DISCUSSION After the discovery of the distal PBREM and the orphan nuclear receptor CAR, the mechanism of transcriptional activation of the CYP2B genes by PB and PB-type inducers became simplified when both enhancer and receptor were shown to be essential for the induction (4,5,8). Yet several important questions remain unanswered, including the following. Why is the magnitude of CYP2B induction greater than for the majority of other CAR target genes? Why is the induction in vivo in liver greater than in cultured cells? Is there a role for the proximal promoter in regulating induction? The present observations of CYP2B6-specific OA synergistic induction to a level similar to that observed by PB in vivo provide potential answers to these questions, in addition to identifying a novel role for CAR in the induction process. The OARE, a 24-bp proximal sequence, regulated the OA synergistic activation of mouse CAR-mediated transcription of CYP2B6 and was responsible for approximately 90% of the overall induction of the CYP2B6 gene by TCPOBOP and OA in Ym17 cells. The OARE may be activated in vivo as part of the overall or maximal CYP2B6 induction response to PB and PB-type inducers, but inYm17 cells this part of the signal pathway requires augmenting by OA. The fact that human CAR could also mediate the synergistic response through the OARE indicates that it may be a general mechanism for maximal induction.
The OARE is located between Ϫ256 and Ϫ233-bp of the CYP2B6 gene. A previous report (23) showed a nuclear protein binding to the activator protein 1 site at Ϫ1441 bp of the rat CYP2B2 gene was enhanced by PB treatment. In addition, we have published that the Ϫ1404/Ϫ971-bp region of the mouse Cyp2b10 gene was critical for PB inducibility (21). Our present work revealed that the corresponding CYP2B6 region does not play a role in the OA synergistic response and did not significantly affect induction by TCPOBOP alone. In addition, deletion of the region corresponding to the CYP2B1/2-positive element identified by Padmanaban and co-workers, to which protein binding was enhanced by OA or PB in rat (24 -26), did not alter either the TCPOBOP-induced or the OA synergistic activity of the CYP2B6 gene. Moreover, the sequence of the OARE did not resemble those of the previously identified enhancer elements. In light of the necessity for the distal PBREM in PB induction, a large contribution from a proximal PBresponsive enhancer was unlikely. However, our present work shows that the OARE of the CYP2B6 gene plays an essential role in OA synergistic activation and is as critical as the PBREM in determining the overall induction of CYP2B6. Moreover, the chromatin organization of both sites was shown to enhance induction of CYP2B6. It is becoming evident that DNA architecture may be as important as DNA sequence in determining gene activity (27). Chromatin structure and local DNA bending are often involved in stabilizing the interactions within multiple protein complexes or enhanceosomes across varying distances to support maximal gene transactivation (27,28). For example, binding of a fellow NR1I member, the vitamin D receptor, to the osteocalcin distal promoter directly modifies chromatin structure at both the distal and proximal promoter to facilitate functional interactions between the two (29).
CAR was not activated by OA either with or without TCPOBOP, and no synergy was observed with its inactive analogues, nor-okadaone or okadaol, indicating that synergy was not the consequence of nonspecific chemical perturbation by OA but required inhibition of protein phosphatase. Regulation of transcription by phosphorylation of DNA binding factors is common; however, in transformed cell lines, the phosphorylation states of transcription factors can differ compared with in vivo (30). Hyperphosphorylation by OA may allow accumulation of a phosphorylated factor in Ym17 cells that is either usually present or is activated by PB-type inducers in conjunction with CAR in vivo, allowing similar induction to occur. For instance, the ability of OABP to bind to the central region of the OARE may be enhanced because of OA hyperphosphorylation. There is also the potential that activated CAR may enhance transcription of a protein within OABP, thus enhancing binding. As such, identifying the transcription factor or factors that directly bind to the OARE is key to understanding the molecular mechanism of the synergistic transcription of the CYP2B6 gene. The OARE consisted of several putative response elements including an Sp1 site and one nonconsensus (mutated in DNA affinity experiment) and two consensus E box motifs (31) (Fig. 9). E boxes consist of the core hexanucleotide sequence CANNTG and can bind proteins from the basic helix-loop-helix (bHLH) family that includes the aryl hydrocarbon receptor, upstream stimulatory factor, and sterol regulatory elementbinding protein (32). Protein phosphorylation has been shown to either enhance (30,33) or disrupt the DNA binding (34) of some bHLH factors and/or to modify their transcriptional activity (35,36). In fact the bHLH factors heart and neural crest derivatives expressed 1 and 2 interact with the protein phosphatase 2A targeting subunit B56␦ (37). Several groups (38 -40) have reported cooperation between factors binding to multiple E boxes in promoters. They are also able to regulate synergistic gene activation with additional transcription factors, such as Sp1 (41,42), and are known to recruit chromatin remodeling cofactors such as cAMP-response element-binding protein (CBP) and p300/CBP associated factor to E box motifs (43)(44)(45). As such, bHLH factors appear as ideal candidates to play a role in OA synergy. Identification of the exact bHLH and/or other factors involved in the OA synergistic response is the focus of our current research. Eleven candidate bHLH proteins, including those mentioned above, were examined by overexpression and electrophoretic mobility shift assays, but OABP has not been identified at present.
We observed the interaction of CAR with the central region of the OARE by DNA affinity purification and chromatin immunoprecipitation, but we were unable to demonstrate direct FIG. 9. OA synergy involves CAR interaction with two sites, the PBREM and the OARE. Maximal CYP2B6 induction requires two distinct cis-acting elements, the distal PBREM and the proximal OARE. TCPOBOP activated CAR binds to the distal PBREM and induces transcription of the CYP2B6 gene. In the additional presence of OA, TCPOBOP-activated CAR interacts indirectly with the OARE through OABP to form a proximal DNA-protein complex and enhances CYP2B6 transcription a further 90%.
binding of the receptor to this site, suggesting that CAR may form a complex with OABP and interact with the DNA through that protein. In fact, OABP binding appeared CAR-dependent because no binding was observed from HepG2 nuclear extracts in which the levels of endogenous CAR are low. The interaction of CAR with the OARE was significantly increased after cotreatment with TCPOBOP and OA. This indicates that OA synergy requires CAR to be in an active conformation to maximally interact with the OARE and that the OA signal transduction pathway is also required. If CAR interacts with the DNA indirectly via OABP, then enhanced OABP binding to the OARE in the presence of OA may explain the increase in interaction of activated CAR with the site. This is the first evidence that CAR can activate transcription via a nuclear receptor half-site-independent mechanism. Such mechanisms have been reported previously for receptors such as the glucocorticoid receptor, which interacts with signal transducer and activator of transcription 5B to activate genes independent of glucocorticoid response-element binding, in a ligand-dependent co-activator role (17). Nuclear receptor related 1 activates the human dopamine transporter gene by a novel mechanism independent of its nerve growth factor-induced protein I-B response element-like sites (46). Even more relevant to CAR and the OA synergistic response is the ability of its heterodimerization partner RXR and the nuclear receptor RAR to interact with the bHLH protein myogenic factor 3 and to reciprocally activate nuclear receptor half-sites and E box motifs (CANNTG), by co-activating each others' DNA binding-dependent gene activation (16). Hepatocyte nuclear factor 4␣ has been reported to be critical for maximal activation of CYP3A4 and CYP3A23 by CAR or pregnane X receptor. However, unlike the proximal CAR-dependent binding of OABP, hepatocyte nuclear factor 4␣ binds to a DR1 element upstream of the CYP3A4 xenobiotic responsive enhancer module or to site A 8-bp downstream of the CYP3A23 DexRE-2, independent of the presence of CAR and can even increase the basal expression of the CYP3A genes in HepG2 and HeLa cells, respectively (47,48). Most other reports of CAR involvement with additional transcription factors or pathways has involved the competition for or sequestering of co-activators, RXR-and DNA-binding sites, ultimately leading to repression of their target genes. Here we present CAR interaction with a transcription factor to synergistically activate the transcription of its own target gene, which expands the possibilities for gene regulation by CAR. Over 70 potential CAR target genes have been detected (4), so this synergistic activation may be important not only for maximal CYP2B6 induction but as a general mechanism of CARmediated gene activation in vivo, providing an insight into how signals elicited at the PBREM may be transmitted to the proximal promoter.
In conclusion, the maximal CYP2B6 induction requires two distinct cis-acting elements, the distal PBREM and the proximal OARE, and is dependent on their chromatin structure. CAR interacts with OABP to form a proximal DNA-protein complex, thus indirectly interacting with the OARE sequence. Identification of the proteins in the complex and the mechanism of interaction provide exciting targets for our future investigation. Once these are elucidated, it is anticipated that the whole signal pathway and how it relates to induction in vivo will be unveiled.