Phenobarbital induction mediated by a distal CYP2B2 sequence in rat liver transiently transfected in situ.

The promoter activities of the genes for cytochrome P450 2B1 (CYP2B1) and cytochrome P450 2C1 (CYP2C1) have been assayed by direct injection of promoter-luciferase chimeric genes into rat liver. Activities of minimal promoters for CYP2C1 and CYP2B1 were detectable in untreated animals but were not increased by treatment of the animals with phenobarbital. After insertion to the 5' side of the minimal promoters of one to three copies of the CYP2B2 sequence from -2318 to -2155, a phenobarbital-responsive element in primary hepatocyte cultures (Trottier, E., Belzil, A., Stoltz, C., and Anderson, A. (1995) Gene (Amst.) 158, 263-268), phenobarbital treatment induced the activity of the CYP2C1 promoter by 5-15-fold and the CYP2B1 promoter by 2.5-5-fold. Mutation of a basal transcription element-like motif and a CCAAT/enhancer binding protein element in the CYP2B1 proximal promoter region reduced expression, but 3-4-fold induction by phenobarbital was retained. Mutation of the "Barbie box," a putative phenobarbital-responsive element (He, J.-S., and Fulco, A. J. (1991) J. Biol. Chem. 266, 7864-7869) in the CYP2B1 proximal promoter did not reduce the relative response to phenobarbital. These results demonstrate that direct injection of DNA into rat liver may be used to assay phenobarbital responsiveness of cytochrome P450 genes. In this system, a distal CYP2B2 element mediates a response to phenobarbital, and proximal elements, including the Barbie box, are not required for the induction.

The induction of drug-metabolizing enzymes by phenobarbital (PB) 1 was reported almost 40 years ago (1). Cytochromes P450 are the primary enzymes responsible for phase I metabolism of drugs, and PB has been shown to increase the amounts of P450 enzyme and mRNA by stimulation of gene expression (2). PB-responsive P450 genes have been isolated and characterized, but progress in determining the molecular mechanism by which this drug activates gene expression has been impeded by the lack of a continuous cell line that is competent for PB induction. Transcription in cell-free systems of a CYP2B1/2 mini-gene containing 179 bp of the 5Ј-flanking region was induced in liver extracts from PB-treated rats (3,4), and this same sequence responded to PB when introduced into rat liver cells in vivo as a complex with asialoglycoproteins (5). PB-dependent changes in binding of nuclear proteins to proximal elements have been detected by gel shift and DNase I footprinting assays and have led to a model in which positive and negative regulatory elements are involved in the PB response (5). In other studies, a sequence motif was identified as a PB-response element ("Barbie box") in CYP genes in Bacillus megaterium, and similar motifs were found in other PB-responsive genes in several species including the proximal promoter of mammalian CYP2B1/2 and CYP2C1 (6 -8). Proteins in extracts from B. megaterium and rat liver nuclei bound to these sequencesinthebacterialandmammaliangenesinabarbituratedependent manner (6). Mutation of a Barbie box in the rat ␣ 1 -acid glycoprotein promoter eliminated a 1.6-fold induction by PB in transfected hepatocytes in primary culture (9).
In contrast to these studies implicating a proximal promoter sequence as a PB-responsive element, other studies have suggested that elements 5Ј distal to the promoter mediate PB induction. A distal enhancer-like element in the chick CYP2H1 gene was reported to mediate PB induction of a heterologous promoter in primary cultures of chicken hepatocytes (10). Transgenes in mice, which contained the CYP2B2 gene with 800 bp of 5Ј-flanking region including the proximal promoter elements discussed above, were expressed constitutively at high levels and were not induced by PB, while expression of transgenes containing 19 kb of 5Ј-flanking region was dependent on PB treatment (11). Consistent with this result a CYP2B2 sequence from Ϫ2318 to Ϫ2155 was shown to mediate PB induction in primary rat hepatocyte cultures with its homologous promoter and a heterologous promoter (12). In studies on the mouse Cyp2b10 gene, deletion of sequences from Ϫ1404 to Ϫ971 resulted in elimination of a 2-3-fold response to PB in transfected primary hepatocytes (13). Interestingly, this region in Cyp2b10 is highly similar to the corresponding region in CYP2B2, but the constructions of the CYP2B2 promoter in which the Ϫ2318 to Ϫ2155 region were deleted, but the sequences similar to the Ϫ1404 to Ϫ971 region in the mouse gene were retained, were not induced by PB in transfected primary hepatocytes (12). These results indicate that the proximal promoter alone in CYP2B genes is not sufficient for PB induction, but that distal elements are also required.
We now report that direct injection of DNA into rat liver is an effective in situ transient transfection method to assay CYP promoter activity. The distal PBRE reported by Trottier et al. (12) mediates PB responsiveness in this system and mutation of proximal promoter elements, including the Barbie box, does not eliminate the response.

EXPERIMENTAL PROCEDURES
Oligonucleotides-The following oligonucleotides were synthesized on an Applied Biosystems model 380A DNA synthesizer at the Biotech-* This study was supported by National Institutes of Health Grant GM39360. 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.
Plasmid Constructions-Construction of pC1-272LUC (14) and pB1-110LUC (15) have been described. A fragment of CYP2B2 from Ϫ2343 to Ϫ2095 was isolated by PCR using rat genomic DNA with 2B2PBRE-A as a 5Ј-primer and 2B2PBRE-B as a 3Ј-primer and cloned into EcoRI/BamHI-digested pTZ18R (16) to produce pTZ-PBRE. The sequence of the Sau3AI fragment from Ϫ2318 to Ϫ2155 of CYP2B2 (designated PBRE) differed from the published sequence only by having 3 rather than 4 Ts at Ϫ2226, but the published numbering is used in this report (17). To construct hybrid plasmids of PBRE and CYP2C1 or CYP2B1 5Ј-flanking regions, the Ϫ272 to 1 CYP2C1 and the Ϫ110 to 1 CYP2B1 fragments were first excised from pC1-2000LUC (18) and pB1-110LUC by BamHI and HindIII digestion and cloned into pTZ18R to produce pTZC1-271 and pTZB1-110. The EcoRI-BamHI CYP2B2 fragment from pTZ-PBRE was isolated and digested with Sau3AI. The resulting Sau3AI fragment was ligated to itself for 1 h, then to BamHIdigested pTZC1-272. Orientation of single, double, and triple copies of the PBRE in pTZC1-272 was identified by sequencing. The hybrid PBRE-C1-272 fragments were excised by digestion with SmaI and HindIII and inserted into the promoterless luciferase vector pA 3 LUC (19) digested with the same enzymes. Single and triple copies of the PBRE in the natural orientation in these CYP2C1 constructions were amplified by PCR using a T7 promoter primer and primer 2C1-257. The PCR products were digested with EcoRI and BamHI and inserted into EcoRI/BamHI-digested wild type pTZB1-110 or its BTE and C/EBP mutations (15). For insertion of a double copy of the PBRE into pTZB1-110, the double copy was amplified as above from a CYP2C1 construction; however, the orientation of the copies in the CYP2C1 construction was reversed. To obtain the proper orientation, the EcoRI/BamHIdigested PCR product was inserted into EcoRI/BamHI-digested pTZ18R, reexcised by digestion with SmaI and HincII, and cloned into the SmaI site of pTZB1-110. The correct orientation was confirmed by sequencing. The Barbie box was mutated by PCR using BBM1 or BBM2 as a 5Ј-primer, 2BA as a 3Ј-primer, and pB1-110LUC as a template. Mutated PCR products were digested with BamHI and HindIII, inserted into pTZ18R to form pTZBBM1 amd pTZBBM2, and three copies of the PBRE were inserted into these vectors as described above for pTZB1-110. The PBRE and CYP2B1 hybrid sequences were excised from the pTZ18R vector with SmaI and HindIII and inserted into pA 3 LUC digested with the same enzymes.
Direct DNA Hepatic Injections-Direct hepatic injections of DNA were performed as described by Malone et al. (20) with minor modifications. DNA of plasmids containing the CYP promoter-luciferase genes was purified by centrifugation twice in CsCl gradients, followed by dialysis against 1 mM Tris-HCl, pH 8.0, 0.1 mM EDTA. Male Sprague-Dawley rats weighing 250 -350 g were anesthetized with ether by inhalation, and a single liver lobe was pulled out through a ventral midline incision. DNA from CYP2B1 (350 g) or CYP2C1 (400 g) promoter constructions in 1.5 ml of Dulbecco's modified Eagle's medium were injected into three different spots of a single liver lobe with a 25-gauge 5 ⁄8-inch needle. One mg of dexamethasone/kg was injected subcutaneously 24 h before surgery as an anti-inflammatory agent that increases expression of the injected gene (20). Shortly after rats recovered from surgery, 1 mg of dexamethasone/kg, subcutaneously, and either saline or 100 mg of PB/kg, intraperitoneally, were injected. Rats were sacrificed 24 h later to obtain liver tissues for luciferase assay.
Luciferase Assays-Luciferase assays were performed as described with minor modifications (21). Rats were killed by ether or CO 2 overexposure, and liver tissue containing the sites of injection was excised, minced, and homogenized in 1.5 ml of 1% Triton X-100, 25 mM glycylglycine, pH 7.8, 15 mM MgSO 4 , 4 mM EGTA, 1 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl sulfate, and 0.5 g/ml leupeptin in a Dounce tissue grinder with 20 strokes with a tight pestle. The homogenates were centrifuged for 20 min at 4°C in a microcentrifuge (Beckman Instruments). Supernatants were then transferred to fresh 1.5-ml Eppendorf tubes, and 0.1 ml was used for each luciferase assay. A background of 300 arbitrary light units, observed in liver extracts after pA 3 LUC injection, was subtracted, and total activity was determined based on the total volume of the supernatant. Statistical significance between the untreated and PB-treated samples was tested by one-tailed Student's t test.

Mediation of PB Induction by a Distal Sequence in Rat Liver
Transfected in Situ-Injection of DNA of chimeric genes containing Ϫ3500 (data not shown) or Ϫ272 bp of CYP2C1 5Јflanking region fused to a luciferase reporter gene into rat liver resulted in a low level of expression of luciferase about 2-fold greater than that in liver injected with pA 3 LUC, a promoterless vector (Fig. 1A). Treatment of rats with 100 mg of PB/kg for 24 h before sacrifice did not increase activity for the CYP2C1 construction. To determine if the CYP2B2 sequence from Ϫ2318 to Ϫ2155 (PBRE) (12) could confer PB responsiveness to the CYP2C1 promoter, a single copy or three copies of this sequence were inserted 5Ј of the CYP2C1 promoter. Insertion of the PBRE had little effect on CYP2C1 promoter activity in untreated rats, but activity was increased about 5-and 15-fold over controls by PB treatment for the constructions containing single and triple copies of the sequence, respectively.
To confirm the role of the PBRE in PB induction, single, The open boxes within the proximal promoter regions represent the Barbie box (BB), BTE, and C/EBP (C/E) motifs and the ϫ's in these boxes indicate that the motif was mutated. DNA was injected into the liver, luciferase activity was assayed, and total activity in the liver was calculated as described under "Experimental Procedures." The background in the luciferase assay measurements was about 300 arbitrary light units, and for reference, the means of measured activities for the control and PB-treated samples were 1762 and 1328, respectively, for the CYP2B1 construction without a PBRE. Animals were treated with saline (solid filled bars) or with 100 mg PB/kg (stippled bars). The numbers of rats injected in each treatment group are shown at the tops of the bars in parentheses, the fold induction by PB is indicated, and the standard error of the mean is shown. The assays for CYP2B1 promoter constructions with the single and double copies of the PBRE were done several months after the rest of the assays and resulted in higher activities for the untreated animals than in the earlier studies for unknown reasons. Activities of PB-treated groups were significantly different from the corresponding untreated groups (p Ͻ 0.05, Student's t test) for each construction containing the PBRE except for BBM2. double, or triple copies of the sequence were also inserted before the Ϫ110 to 1 CYP2B1 minimal promoter. The Ϫ110 promoter sequence, without the trimer inserted, was about 5-fold more active than the Ϫ272 CYP2C1 promoter. Insertion of the PBRE sequences had little effect on promoter activity in untreated animals, but activity was increased by treatment with PB about 2.5-fold for the single copy and 4 -5-fold for the double and triple copies (Fig. 1B). These results indicate that the PBRE sequence is either a PB-responsive enhancer or that it is a general enhancer which increases basal and PB-dependent transcription mediated by elements in the proximal promoter.
Role of the CYP2B1 Proximal Promoter in PB Induction-To examine the role of the proximal promoter elements in the responsiveness to PB, mutations in the BTE-like motif, the C/EBP element, and the Barbie box sequence were introduced in the context of the Ϫ110 to 1 CYP2B1 promoter with the PBRE trimer inserted to the 5Ј side (Fig. 2). These mutations in the BTE and C/EBP elements each inhibit CYP2B1 promoter activity by about 70% in transfected HepG2 cells (15). The mutations in the Barbie box include the core sequence, AAAG, which is highly conserved in this proposed PB-responsive element (8). Mutation of the BTE-like element resulted in about a 65% reduction in activity in untreated animals (Fig. 1B). Activity was increased about 3-fold by treatment with PB, which is similar to the increase observed with the wild-type Ϫ110 to ϩ1 promoter. Mutation of the C/EBP element resulted in a decrease of about 80% in untreated animals, and treatment with PB increased activity about 4-fold, again similar to induction of the wild-type promoter. The BTE-like element and the C/EBP element, therefore, function as PB-independent positive regulators of CYP2B1 promoter activity.
Similar PB responsiveness was observed when the Barbie box sequence was mutated. The activity of the BBM1 mutated promoter in the untreated animals was increased about 2-fold (Fig. 1B). However, treatment with PB again increased activity about 5-fold, indicating that the Barbie box sequence was not required for PB responsiveness. The increased activity in the untreated animal suggested that this region might have a modest negative regulatory activity; however, it is also possible that the increase was due to the unintentional creation of a positive regulatory element by the mutation. A second mutation was constructed in this region (BBM2) which gave similar results, although the variability in the experiment was larger than normally observed, and the difference between the PB-treated and untreated groups was not statistically significant in this one case (Fig. 1B). The similar results with two mutations indicate that the small increase in activity is due to the mutated Barbie box and not to inadvertent creation of a new positive regulatory site. The continued PB response with each of the BTE, C/EBP, and Barbie box mutations suggests that the PBRE is a PB-dependent enhancer and not a general enhancer dependent on proximal promoter elements for PB responsiveness. DISCUSSION A major impediment to analysis of the genetic mechanisms involved in the PB induction of cytochrome P450 and other drug-metabolizing enzymes has been the lack of a continuous cell culture model system in which promoter activity could be assayed. Primary cultures of hepatocytes (12,13), cell-free transcription (3,4,17,22,23), and targeting to the liver of DNA complexed to asialoglycoproteins (5) have been reported as methods in which PB responsiveness is retained. These methods are relatively complex, and the responsiveness in cell-free systems has been very modest compared to the induction observed in vivo. DNA injected directly into rat liver has been shown to be taken up by hepatocytes and expressed (20), and as reported here, is a relatively simple, reproducible method for analysis of PB responsiveness of CYP genes. With the luciferase reporter, results can be obtained less than 2 days after the injection of the DNA, and the range of variability is usually about 2-fold, so that changes of 4 -15-fold as observed here are easily detected with small numbers of animals in each treatment group. There are two shortcomings of this system at present. The uptake and expression of DNA is inefficient so that 300 -500 g must be injected in each liver. The low efficiency has precluded the used of an internal standard to allow normalization of the efficiency of uptake and expression of the injected DNA, since only the luciferase assay has sufficient sensitivity for detection. Sensitive chemiluminescent assays of ␤-galactosidase (24) may provide a second reporter that could be used for normalization, and this possibility is being examined. The second disadvantage is that dexamethasone treatment is required to increase the efficiency of expression of the injected genes by the liver (20), which might alter the response to other ligands, such as PB. Even with these disadvantages, the simplicity and speed of the assay should permit rapid identification of the elements required for PB induction in CYP and other drug metabolizing enzyme genes.
These in situ transfection experiments provide strong support for the role in PB-dependent gene expression of the distal PBRE identified by transfections of primary hepatocyte cultures (12) and are consistent with CYP2B2 transgenic studies (11) which indicated that sequences distal to Ϫ800 were required for PB induction. Induction of up to 15-fold with triple copies of the PBRE and 2.5-fold with single copies of the PBRE were observed which are less than the 20 -100-fold increases of mRNA levels observed for CYP2B2 and CYP2B1 in vivo (25)(26)(27). The low induction may be due to the relatively high expression observed in the injected liver in untreated animals, particularly for CYP2B1, since expression of this gene in vivo is not detectable in untreated animals (28). Sequences which mediate the inhibition in vivo might be missing from the constructions or normal chromatin structure may be required for the suppression which is lacking in genes transiently transfected. The PBRE was able to confer PB induction on minimal promoters of its homologous gene as well as a heterologous gene, CYP2C1, as assayed by injection in situ. In primary hepatocyte cultures, this sequence conferred PB responsiveness to the heterologous herpes simplex thymidine kinase gene (12). In the hepatic direct injection assay, like the primary  (12). At the bottom the sequence of the region from Ϫ89 to Ϫ40 of the CYP2B1 proximal promoter is shown. The overlines indicate the positions of the proposed Barbie box sequence and the BTE-like sequence. A consensus C/EBP sequence is shown above the functional C/EBP site in the CYP2B1 sequence. The mutations introduced into the Barbie box, the BTE-like motif, and the C/EBP element are shown below the sequence. hepatocyte culture (12), the induction was orientation-independent (data not shown), and induction was proportional to increased copies of the PBRE, characteristics common to enhancer sequences. The ability of this element to confer PB responsiveness to heterologous promoters and the continued responsiveness when proximal promoter sequences are mutated establish that this upstream enhancer-like element is a PB-responsive element. The demonstration that this distal region is hypersensitive to DNase I treatment of intact chromatin and that the relative sensitivity is altered by PB treatment provide further support for the regulatory role of this region (29).
In contrast to the distal elements, the in situ transfection experiments do not provide support for a role of the proximal promoter region of CYP2B1 and CYP2C1 in PB induction. It has been suggested that upstream elements might be nonspecific enhancers, with PB regulation exerted at the proximal promoter region (5). However, in the absence of the PBRE, promoter fragments from Ϫ3500 to ϩ1 (data not shown), Ϫ272 to ϩ1 for CYP2C1, Ϫ1400 to ϩ1 (data not shown), and Ϫ110 to ϩ1 for CYP2B1 did not exhibit any positive response to PB treatment. Likewise, PB induction was not observed in transfected primary hepatocytes which contained CYP2B2 promoter fragments from Ϫ2015 or Ϫ1680 to 1 (12). Further, mutation of the known positive regulatory elements in the CYP2B1 promoter (15) and the proposed PB-responsive element, Barbie box (8), did not substantially alter the relative induction by PB. Mutation of the BTE and the C/EBP reduced activity 65 and 80%, respectively, in untreated animals, which is similar to the reduction in activity of 70 -75% observed for these mutations in HepG2 cells (15). Interestingly, the Barbie box mutations increased activity 2-3-fold, suggesting that this element might have modest negative regulatory activity. Studies with the mouse Cyp2b10 gene, which is 83% similar in the 5Ј-flanking region to CYP2B1, transfected in primary hepatocytes, also indicated that distal elements mediate PB induction (13). In this mouse gene the Barbie box is split by a 42-bp insertion, and this lack of motif conservation suggests that the Barbie box sequence is not a common PB element in P450 genes.
It is possible that PB responsiveness of proximal elements is not detected in the in situ assays either because the assay is not sensitive enough or because of the conditions of the assay compared to other assays. In some cell-free studies, PB responsiveness was modest (17), less than 2-fold, and this level of induction might be difficult to detect in the in situ injection assay in which induction mediated by the PBRE was considerably less than induction of CYP2B1 in vivo. However, 8-fold and higher induction with only 179 nucleotides of CYP2B2 5Ј-flanking region was observed in other cell-free studies and when DNA was targeted to the liver as a complex with asialoglycoproteins (3,5). The reason for the different results observed with the in situ injection assay is not immediately obvious. A second caveat is that dexamethasone treatment is required for successful in situ transfection, and dexamethasone has been reported to both increase or decrease PB induction of P450 genes (2). Dexamethasone has been reported to reduce PB induction mediated by the 179-bp CYP2B1 promoter fragment (4,30). Earlier studies (30) used doses of dexamethasone 10 -50-fold higher than the 1 mg/kg used in these in situ injection experiments, but the more recent studies (4) reported partial inhibition of PB induction by 0.1 mg of dexamethasone/kg. However, in transgenic mice, dexamethasone is not required, and therefore the lack of PB responsiveness of CYP2B2 transgenes containing 800 bp of 5Ј-flanking sequence (11) argues against dexamethasone masking of PB-proximal elements as an explanation for their lack of PB responsiveness in the in situ injection assay. The data with the in situ injection assay show that the proximal elements are not required for the PB induction mediated by the distal elements and are most consistent with a major role for the 5Ј distal PBRE as a PB-response element and a modest role, if any, for proximal promoter sequences.