J Biol Chem, Vol. 274, Issue 10, 6043-6046, March 5, 1999

COMMUNICATION
The Repressed Nuclear Receptor CAR Responds to Phenobarbital in Activating the Human CYP2B6 Gene^*

Tatsuya Sueyoshi, Takeshi Kawamoto, Igor Zelko, Paavo Honkakoski, and Masahiko Negishi^§

From the Pharmacogenetics Section, Laboratory of Reproductive and Developmental Toxicology, NIEHS, National Institutes of Health, Research Triangle Park, North Carolina 27709 and the  Department of Pharmaceutics, University of Kuopio, FIN-70211 Kuopio, Finland

	ABSTRACT

Top Abstract Introduction References

The endogenous CYP2B6 gene becomes phenobarbital (PB) inducible in androstenol-treated HepG2 cells either transiently or stably transfected with a nuclear receptor CAR expression vector. The PB induction mediated by CAR is regulated by a conserved 51-base pair element called PB-responsive enhancer module (PBREM) that has now been located between 1733 and 1683 bp in the gene's 5'-flanking region. An in vitro translated CAR acting as a retinoid X receptor  heterodimer binds directly to the two nuclear receptor sites NR1 and NR2 within PBREM. In a stably transfected HepG2 cell line, both PBREM and NR1 are activated by PB and PB-type compounds such as chlorinated pesticides, polychlorinated biphenyls and chlorpromazine. In addition to PBREM, CAR also transactivates the steroid/rifampicin-response element of the human CYP3A4 gene in HepG2 cells. Thus, activation of the repressed nuclear receptor CAR appears to be a versatile mediator that regulates PB induction of the CYP2B and other genes.

	INTRODUCTION

Top Abstract Introduction References

Metabolism by the liver microsomal CYPs¹ plays an important role in the detoxification of xenochemicals such as pharmaceutical drugs and environmental contaminants. Ironically it can often result in activation of xenochemicals to more toxic and/or carcinogenic products (1). Inducible gene transcription by exposure to xenochemicals is characteristic for CYPs, which increases the organism's defense capability against toxicity and carcinogenicity. Since the discovery of PB induction 35 years ago (2), PB has served as a prototype for a large group of structurally and functionally diverse xenochemicals that induce CYP2B genes. A PB-responsive enhancer activity was first associated with a 177-bp DNA sequence (2318/2155 bp) of the rat CYP2B2 gene in primary hepatocytes (3) and was later confirmed using an in situ injection method (4). Furthermore, the corresponding sequence was also found in the mouse Cyp2b10 gene (5). The enhancer activity of the 177-bp sequence has now been delimited to a 51-bp element PBREM in Cyp2b10 (6) and similar DNA element in CYP2B2 (7, 8). The nuclear orphan receptor heterodimer CAR-RXR has been identified as a transactivator of PBREM (9).

The nuclear orphan receptor CAR was originally characterized as a constitutive activator of an empirical set of retinoic acid response elements (RAREs) (10). Consistently, CAR-mediated PBREM activity observed in HepG2 cells was PB-independent (9). Thus, CAR remains to be established as a true signal mediating PB induction. Most recently, the CAR-mediated transactivation of RAREs has been reported to be repressed by 3-androstenol and androstanol (11). Considering the functional properties of CAR, we have examined whether CAR can activate the endogenous CYP2B gene and whether activation is PB-dependent in a stable HepG2 cell line transfected with a CAR expression vector. We have also located the PBREM sequence in the human CYP2B6 gene and demonstrated PB responsiveness of PBREM in HepG2 cells.

	EXPERIMENTAL PROCEDURES

Chemicals-- TCPOBOP was synthesized using the method of Kende et al. (12). 3-Androstenol was obtained from Steraloids (Wilton, NH); o,p'-DDT, methoxychlor, and PCB were from AccuStandard (New Haven, CT). All other chemicals were purchased from Sigma.

Plasmids-- CAR- cDNA was amplified as described in a previous report (9) and cloned into between the BamHI and XhoI site of pCR3 vector. In the case of the luciferase reporter plasmid, the 160-bp thymidine kinase (tk) promoter was amplified from pBLCAT2 plasmid and cloned into the BglII and HindIII sites of pGL3-Basic vector containing the firefly luciferase reporter gene (Promega). The complementary oligonucleotides of human and mouse PBREM and the triplicate human NR1 ((5'-gatcACTGTACTTTCCTGACCTTGgatc-3')₃) sequences were synthesized and were cloned in front of the tk promoter (BglII site), resulting in PBREM-tk-luciferase and (NR1)₃-tk-luciferase reporter gene plasmids. The (ER6)₃-tk-CAT reporter gene plasmid was constructed by insertion of (5'-GATCAATATGAACTCAAAGGAGGTCAGTG-3')₃ into the pBLCAT-2 basic vector (13).

Cells and Transfection Assay-- HepG2 cells were cultured in minimal essential medium supplemented with 10% fetal bovine serum and antibiotics (9). A stable transfected HepG2 cell line (g2car-3) was transfected with a pCR3-mCAR expression vector and selected by neomycin resistance. The enhancer-containing plasmids (0.1 µg) were co-transfected with pRL-SV40 (0.2 µg) into g2car-3 cells (17-mm well) using calcium phosphate co-precipitation and incubated for 24 h. Luciferase activity was measured simultaneously using the Dual-Luciferase Reporter Assay System (Promega). Promoter activities were determined from three independent transfections and calculated from firefly luciferase activities normalized against Renilla luciferase activities of an internal control pRL-SV40 plasmid. For the CAT assays, a CYP3A4 (ER6)₃-tk-CAT reporter gene plasmid (1.0 µg) was co-transfected with pCMVgal (Promega) as an internal control. 3-Androstenol (4 µM) was included in the culture medium, which produced maximal repression of the CYP2B6 gene in the CAR-transfected HepG2 cells (EC₅₀  1 µM).

RT-PCR-- To measure CYP2B6 mRNA, total RNA was subsequently prepared from HepG2 and g2car-3 cells using TRIZOL reagent (Life Technologies, Inc.). One-twentieth of each cDNA synthesized from 2 µg of RNA was subjected to 37- and 29-cycle amplifications of CYP2B6 mRNA from HepG2 and g2car-3 cells, respectively. The primers (5'-CACCCTAACACCCATGACCG-3' and 5'-GATCACACCATATCCCCGGA-3') were designed from the reported sequences (GenBank^TM accession number J02864) across intron 1 of the CYP2B6 gene. PCR reactions were performed using Taq DNA polymerase (Promega) with TaqStart antibody (CLONTECH) at 94 °C for 30 s, 55 °C for 30 s, and 72 °C for 40 s. The amplified DNA (276 bp) was separated on a 1.5% agarose gel and visualized with staining by ethidium bromide. -Actin mRNA was amplified as an internal control.

Gel-shift Assays-- hCAR and RXR were translated using TNT-coupled reticulocyte lysate system (Promega). The double-stranded oligonucleotides of the human NR1, NR2 (5'- gatcCATGGACTTTCCTGAACCAAgatc-3'), or ER6 were labeled using [³²P]dCTP and DNA polymerase Klenow fragment. Gel-shift assays were performed as described previously (9). For competition, 200-fold excess of each cold probe was included in the incubation.

DNA Sequencing-- The cosmid clone F17390 containing the humanCYP2B6 gene was generously provided by Dr. H. W. Mohrenweiser (14). A 5 kbp of the 5'-flanking region was amplified by PCR using primers generated from exon 1 and the cosmid vector, digested to smaller fragments (0.5-1.0 kbp) by sonication, blunt-ended with T4 DNA polymerase, and cloned into the SmaI site the pUC18 vector. Over 100 pUC18 clones were sequenced, and sequence data were assembled by Sequencher to complete the 1.8-kbp 5'-flanking region of CYP2B6 gene. The GenBank^TM accession number of this sequence is AF081569.

	RESULTS AND DISCUSSION

The levels of CYP2B6 mRNA were measured in the transfected HepG2 cells to obtain more direct evidence that CAR can mediate PB induction of the CYP2B gene (Fig. 1A). First, HepG2 cells are transiently transfected with CAR-expressing plasmids and then treated with 3-androstenol alone or both 3-androstenol and TCPOBOP (the most potent PB-type inducer known). RT-PCR assay reveals that endogenous CYP2B6 mRNA is present in CAR-transient transfected HepG2 cells but not in nontransfected HepG2 cells. Androstenol treatment represses CYP2B6 mRNA in CAR-transfected HepG2 cells. Subsequently, TCPOBOP dramatically induces CYP2B6 mRNA in 3-androstenol-treated CAR-transfected HepG2 cells. These transient transfections showed that the endogenous CYP2B6 gene can be regulated by CAR and PB.

View larger version (41K): [in this window] [in a new window]   Fig. 1.   CAR-mediated induction of CYP2B6 mRNA in HepG2 cells. A, HepG2 cells were transfected with pCR3-mCAR (1.0 µg/21-mm well) 24 h prior to treatment with TCPOBOP (250 nM) in the presence or absence of 3-androstenol (4 µM). For all experiments, total RNAs were prepared from duplicated cultures treated separately and were subjected to RT-PCR and agarose gel electrophoresis. DNA bands were visualized by ethidium bromide. B, dose-dependent induction by PB or TCPOBOP. Various concentrations (indicated on the top of each line) of PB or TCPOBOP were added at 70% confluent g2car-3 cells in 12-well dish with 1-ml culture medium containing 4 µM 3-androstenol, followed by further 24-h incubation before performance of RT-PCR.

We then constructed a stable HepG2 cell line (named g2car-3) transfected with a CAR-expressing plasmid in order to measure mRNA more consistently. First, stable expression of CAR resulted in a dramatic increase of CYP2B6 mRNA in the g2car-3 cells as indicated by intense bands after just 29 cycles (Fig. 1B). However, 37 cycles of PCR does not amplify mRNA from HepG2 (left two lanes in Fig. 1A). The expression levels of CYP2B6 mRNA in g2car-3 cells are roughly comparable with that of Cyp2b10 mRNA in PB-treated mouse livers. Using this stable cell line, dose-dependent induction was examined with PB and TCPOBOP. As expected, CYP2B6 mRNA is induced in a dose-dependent manner by the inducers in 3-androstenol-treated g2car-3 cells (Fig. 1B). The results show unequivocally that CAR mediates induction of the endogenous CYP2B6 gene in response to PB and TCPOBOP.

The mouse PBREM is a 51-bp DNA enhancer consisting of two nuclear receptor DR4 motifs (NR1 and NR2) flanking a NF1 binding site (6). The corresponding 51-bp DNA is found in a 1.7-kbp region of the human CYP2B6 gene, while its 5' to 3' orientation is opposite from the mouse and rat PBREM sequences (Fig. 2). The 16-bp NR1 site of human PBREM differs by only 1 base from mouse NR1, making NR1 the most conserved sequence between human, mouse, and rat PBREM elements. The NR1 site, thus, may be the key element conferring enhancer activity to PBREM. Since the human NR2 is 87% sequence identical to NR1 and also binds to in vitro translated CAR-RXR heterodimer (Fig. 3A), the function of NR2 may be redundant. On the other hand, the least conserved NF1 site may not be critical for PB responsiveness in agreement with our previous findings (9). Consistent with our previous findings (9), neither PB or 3-androstenol (10 µM) affected binding of in vitro translated CAR (data not shown). Moore et al. have also shown that androstenol and androstanol do not inhibit binding of CAR to RARE sequences, while they appear to dissociate SRC-1 from CAR (11). Therefore, it remains to be investigated whether PB can recruit transcriptional co-activators to activate PBREM.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Localization and sequence of the human PBREM. The antisense sequence of the human PBREM is shown in comparison with mouse PBREM and the corresponding sequences of PB-inducible rat CYP2B1 and CYP2B2 genes (GenBankTM accession numbers U30327 and S51970, respectively). The 5'-flanking sequence of the human CYP2B6 gene is tentatively numbered using the putative transcription start site of the mouse Cyp2b10 gene. NR1 and NR2 represent two DR4 nuclear receptor motifs, which are boxed, and their direct repeats are indicated by arrows, while NF1 denotes a putative nuclear factor 1 binding site. Indirect repeats (ER6) in CYP3A4 gene are indicated by arrows with solid lines and a possible DR4 motif within ER6 is indicated by arrows with broken lines.

View larger version (27K): [in this window] [in a new window]   Fig. 3.   PB responsiveness of CAR-mediated transactivation in g2car-3 cells. A, binding of in vitro translated CAR and RXR to NR1 and NR2. Gel shift was performed with either 32P-labeled NR1 or NR2 probes. For competition, 200-fold cold probe was included in gel-shift assay. B, the promoter activities were calculated from three independent transfections and are shown by firefly luciferase activities normalized against Renilla luciferase activities for the internal control pRL-SV40 plasmid. The arrows indicate the 5' to 3' orientations of the PBREM sequences. Andro., 3-androstenol. Luci, luciferase.

We then examined whether CAR-mediated transactivation of PBREM is PB-inducible. For this, the PBREM-tk-luciferase reporter gene was transfected into 3-androstenol-treated g2car-3 cells, which were then induced by TCPOBOP. CAR-mediated transactivation was enhanced by 5-10-fold in g2car-3 cells (Fig. 3B). NR1 alone appears to be sufficient to confer PB responsiveness to CAR-mediated transactivation. Moreover, the (NR1)₃-tk-luciferase reporter gene displays a high TCPOBOP responsiveness (Fig. 3B), suggesting that PB and TCPOBOP probably induce CAR-mediated transactivation by displacing 3-androstenol. These transactivations by CAR are totally consistent with the induction of the endogenous CYP2B6 gene (Fig. 1). Thus, the nuclear receptor CAR can mediate PB-inducible transcription of the CYP2B6 gene through its binding to the NR1 site of PBREM. 3-Androstenol is one of the 16-androstenes in human urine and is also present in circulation (15). The reported peripheral plasma concentration (nanomolar) of 3-androstenol in men is substantially lower than the micromolar concentration required to repress CAR-mediated transactivation in HepG2 cells. Although 16-androstenes may accumulate to higher levels and/or steroids may repress CAR at substantially lower concentration in liver, how CAR is actually repressed in liver remains as a question of major interest.

PBREM is a versatile enhancer that responds to a great number of PB-type inducers with very diverse structures (6). Therefore, known PB-type inducers o,p'-DDT, methoxychlor, PCB, and chlorpromazine were selected to examine whether induction by these inducers is also CAR-mediated. All four xenochemicals are found to induce CYP2B6 mRNA in 3-androstenol-treated g2car-3 cells (Fig. 4A). The (NR1)₃-tk-luciferase reporter activity is also increased by these PB-type inducers (Fig. 4B). In contrast, 3-methylcholanthrene (3-MC), dexamethasone (DEX), and clofibrate, which are classical CYP1A, CYP3A, and CYP4A inducers, respectively, neither transactivate PBREM nor induce CYP2B6 mRNA (Fig. 4, A and B). A slight increase by 3-MC of mRNA may not be mediated by CAR, since 3-MC did not activate PBREM. DEX actually represses both mRNA and PBREM activity. It is known that DEX treatment decreases CYP2B mRNA in rat hepatoma-derived Fao cells (16), while DEX induces the Cyp2b gene in mouse primary hepatocytes (17). The mechanism by which DEX represses PBREM remains to be investigated. CAR, nevertheless, displays a versatility to mediate transactivation by a number of known PB-type inducers and a specificity which corresponds only to that of PB-type inducers.

View larger version (26K): [in this window] [in a new window]   Fig. 4.   Versatility of CAR to respond to various PB-type inducers. A, duplicates of g2car-3 cells (21-mm well) were treated by chlorpromazine (CPZ) (25 µM), o,p'-DDT (50 µM), methoxychlor (50 µM), or PCB (50 µM) for 24 h. For controls, cells were also treated with DEX (50 µM), 3-MC (1 µM), or clofibrate (200 µM) for 24 h. Total cellular RNA was subjected to RT-PCR. These concentration of inducers were selected from our previous paper (6). The amplified DNAs were separated on an agarose gel and stained by ethidium bromide. B, for transactivation, the (NR1)3-tk-luciferase gene reporter plasmid (0.1 µg) was co-transfected with pRL-SV40 plasmid into g2car-3 cells. The cells were incubated with various chemicals for 24 h, harvested, and subjected to luciferase assay.

PB is a pleiotropic agent that induces various drug-metabolizing enzymes, including the CYP3A gene (18, 19). As shown in Fig. 5, A and B, in vitro translated CAR-RXR heterodimer binds to the steroid/rifampicin-responsive ER6 element of human CYP3A4 gene and mediates PB-inducible transactivation of the response element in the g2car-3 cells. Therefore, the pleiotropic PB actions can, in part, be explained by the ability of CAR-RXR heterodimer to bind to distinct nuclear receptor motifs. Alternatively, since ER6 is reported to be activated by PXR and PB (13), PB may activate multiple nuclear orphan receptors to induce various CYP genes. These effects of PB are concerted to increase hepatic metabolic capability. For example, proliferation of endoplasmic reticulum increases membranes into which CYP protein is inserted, while heme synthesis and glucose metabolism supply the prosthetic group protoheme and the reducing agent NADPH, respectively. Does CAR play a role in endoplasmic reticulum proliferation and/or regulate glucose metabolism? Whether CAR is responsible for these pleiotropic actions of PB remains a focus of future investigation.

View larger version (32K): [in this window] [in a new window]   Fig. 5.   Induction of the human CYP3A4 gene in g2car-3 cells. A, gel-shift assay of the binding of in vitro translated mCAR and RXR to 32P-labeled ER6 as probe. For competition, 3A4 denotes cold ER6 oligonucleotide. In NR1m (5'-gatcATTATACTTTTCTTGAACTTGgatc-3'), 3 bases (underlined) were mutated from the original NR1 sequence. B, (ER6)3-tk-CAT reporter gene plasmid (1.0 µg) was transfected into g2car-3 cells (60-mm dish), which were treated with TCPOBOP (250 nM) for 24 h in the culture medium. Cells were then harvested for and CAT activity determined. pSVgal (Promega) was co-transfected as an internal control.

Nuclear orphan receptors have now emerged as major mediators of xenochemical-inducible transcription of CYP genes. In addition to CAR, pregnenolone X receptor, and peroxisome proliferator-activated receptor  activate xenochemical-induced transcription of the CYP3A and CYP4A genes, respectively (20-23). As a family of proteins with a low affinity to diverse ligands and with a binding flexibility to various DNA motifs, the nuclear receptors may collectively induce a large set of CYP and conjugation enzymes in response to myriads of xenochemicals. Induction, in turn, could result in increased metabolism of xenochemicals and endogenous substrates (e.g. steroids and lipids) to produce other potential ligands of nuclear receptors. For example, the pesticide methoxychlor is metabolized by CYP2B enzyme to estrogenic products (24). As methoxychlor also activates CAR and CYP2B expression, CAR could be directly involved in increasing availability of ligands for estrogen receptors. A functional cross-talk between CYP and nuclear receptor superfamilies may have important consequences in the physiological and pathological responses to xenochemicals.

	FOOTNOTES

^* The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ To whom all correspondence should be addressed: Pharmacogenetics Section, Laboratory of Reproductive and Developmental Toxicology, NIEHS, National Institutes of Health, Research Triangle Park, NC 27709. Tel.: 919-541-2404; Fax: 919-541-0696; E-mail: negishi{at}niehs.nih.gov.

	ABBREVIATIONS

The abbreviations used are: CYP, cytochrome P450; CAT, chloramphenicol acetyltransferase; PB, phenobarbital; PBREM, PB-responsive enhancer module; tk, thymidine kinase; RARE, retinoic acid response element; RT-PCR, reverse transcriptase-polymerase chain reaction; TCPOBOP, 1,4-bis[2-(3,5-dichloropyridyloxy)]benzene; NR, nuclear receptor; 3-androstenol, 16-(5)-androsten-3-ol; o, p'-DDT, 1,1,1-trichloro-1,2-bis(o,p'-chlorophenyl)ethane; PCB, 2,3,3',4',5',6-hexachlorobiphenyl; bp, base pair(s); kbp, kilobase pair(s); 3-MC, 3-methylcholanthrene; DEX, dexamethasone; RXR, retinoid X receptor; PBREM, phenobarbitalresponsive enhancer module; CAR, constitutively activated receptor.

	REFERENCES

Top Abstract Introduction References

1. Conney, A. H. (1982) Cancer Res. 42, 4875-4917[Free Full Text]
2. Remmer, H., and Merker, H. (1963) Science 142, 1657-1658[Abstract/Free Full Text]
3. Trottier, E., Belzil, A., Stoltz, C., and Anderson, A. (1995) Gene (Amst.) 158, 263-268[CrossRef][Medline] [Order article via Infotrieve]
4. Park, Y., Li, H., and Kemper, B. (1996) J. Biol. Chem. 271, 23725-23728[Abstract/Free Full Text]
5. Honkakoski, P., and Negishi, M. (1997) J. Biol. Chem. 272, 14943-14949[Abstract/Free Full Text]
6. Honkakoski, P., Moore, R., Washburn, K. A., and Negishi, M. (1998) Mol. Pharmacol. 53, 597-601[Abstract/Free Full Text]
7. Stoltz, C., Vachon, M.-H., Trottier, E., Dubois, S., Paquet, Y., and Anderson, A. (1998) J. Biol. Chem. 273, 8528-8535[Abstract/Free Full Text]
8. Kim, J., and Kemper, B. (1997) J. Biol. Chem. 272, 29423-29425[Abstract/Free Full Text]
9. Honkakoski, P., Zelko, I., Sueyoshi, T., and Negishi, M. (1998) Mol. Cell. Biol. 18, 5652-5658[Abstract/Free Full Text]
10. Baes, M., Gulick, T., Choi, H.-S., Martinoli, M., Simha, D., and Moore, D. D. (1994) Mol. Cell. Biol. 14, 1544-1552[Abstract/Free Full Text]
11. Forman, B. M., Tzameli, I., Choi, H.-S., Chen, J., Simha, D., Seol, W., Evans, R. M., and Moore, D. D. (1998) Nature 395, 612-615[CrossRef][Medline] [Order article via Infotrieve]
12. Kende, A. S., Ebetino, F. H., Drendel, W. B., Sundaralingam, M., Glover, E., and Poland, A. (1985) Mol. Pharmacol. 28, 445-453[Abstract]
13. Lehmann, J. M., McKee, D. D., Watson, M. A., Willson, T. M., Moore, J. T., and Kliewer, S. A. (1998) J. Clin. Invest. 102, 1016-1023[Medline] [Order article via Infotrieve]
14. Hoffman, S. M., Fernandez-Salguero, P., Gonzalez, F. J., and Mohrenweiser, H. W. (1995) J. Mol. Evol. 41, 894-900[Medline] [Order article via Infotrieve]
15. Gower, D. B., and Ruparelia, B. A. (1993) J. Endocrinol. 137, 167-187[Medline] [Order article via Infotrieve]
16. de Waziers, I., Bouguet, J., Beaune, P. H., Gonzalez, F. J., Ketterer, B., and Barouki, R. (1992) Pharmacogenetics 2, 12-18[CrossRef][Medline] [Order article via Infotrieve]
17. Honkakoski, P., Moore, R., Gynther, J., and Negishi, M. (1996) J. Biol. Chem. 271, 9746-9753[Abstract/Free Full Text]
18. Waxman, D. J., and Azaroff, L. (1992) Biochem. J. 281, 577-592
19. Honkakoski, P., and Negishi, M. (1998) J. Biochem. Mol. Toxicol. 12, 3-9[CrossRef][Medline] [Order article via Infotrieve]
20. Kliewer, S. A., Moore, J. T., Wade, L., Staudinger, J. L., Watson, M. A., Jones, S. A., McKee, D. D., Oliver, B. B., Willson, T. M., Zetterstrom, R. H., Perlmann, T., and Lehmann, J. M. (1998) Cell 92, 73-82[CrossRef][Medline] [Order article via Infotrieve]
21. Huss, J. M., Wang, S. I., Astrom, A., McQuiddy, P., and Kasper, C. B. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 4666-4670[Abstract/Free Full Text]
22. Lee, S. S.-T., Pineau, T., Drago, J., Lee, E. J., Owens, J. W., Kroetz, D. L., Fernandez-Salguero, P. M., Westphal, H., and Gonzalez, F. J. (1995) Mol. Cell. Biol. 15, 3012-3022[Abstract]
23. Johnson, E. F., Palmer, C. A., Griffin, K. J., and Hsu, M.-H. (1996) FAESB J. 10, 1241-1248
24. Bulger, W. H., Feil, V. J., and Kupfer, D. (1985) Mol. Pharmacol. 27, 115-124[Abstract]