Regulation of the rat UGT1A6 by glucocorticoids involves a cryptic glucocorticoid response element.

Glucocorticoids precociously induce fetal rat UGT1A6 and potentiate polycyclic aromatic hydrocarbon (PAH)-dependent induction of this enzyme in vivo and in isolated rat hepatocytes. To establish whether induction was due to glucocorticoid receptor (GR), luciferase reporter vectors were tested in transfection assays with HepG2 cells. Using a reporter construct containing approximately 2.26 kilobases of the 5'-flanking region of the UGT1A6-noncoding leader exon (A1*), dexamethasone increased basal activity 3- to 7-fold in cells cotransfected with an expression plasmid for GR. PAH increased gene expression 23-fold, but the presence of dexamethasone only induced PAH-dependent expression by 1.5-fold, suggesting interaction between GR and the aryl hydrocarbon (Ah) receptor. Furthermore, the GR antagonist RU 38486 [17beta-hydroxy-11beta-(4-dimethylamino-phenyl)-17alpha-(prop-1-ynyl)-estra-4,9-dien-3-one] was a partial agonist that increased, rather than inhibited, basal activity 3-fold. 5'-deletion analysis defined the 5'-boundary for a functional glucocorticoid-responsive unit between base pairs -141 and -118 relative to the transcription start site. This region contains the Ah receptor response element (AhRE), and both PAH and glucocorticoid-dependent gene activation were lost when this area was deleted. Mutation of a single base pair located in the AhRE region simultaneously reduced induction by PAH and increased glucocorticoid induction. Thus, the sequences of both the AhRE and glucocorticoid response elements seem to overlap, suggesting that Ah receptor binding may decrease glucocorticoid-dependent induction due to interactions of these two cis-acting elements. Mutation of a putative GRE located between base pair -81 and -95 reduced, but did not completely eliminate, glucocorticoid-dependent induction of the reporter, suggesting that a nonclassic mechanism of induction is involved in this response.

Structural heterogeneity at the UDP-glucuronosyltransferase 1 locus: functional consequences of three novel missense mutations in the human UGT1A7 gene.

One of the most important mechanisms involved in host defense against xenobiotic chemicals and endogenous toxins is the glucuronidation catalysed by UDP-glucuronosyltransferase enzymes (UGT). The role of genetic factors in determining variable rates of glucuronidation is not well understood, but phenotypic evidence in support of such variation has been reported. In the present study, six single nucleotide polymorphisms were discovered in the first exon of the UGT1A7 gene, which codes for the putative substrate-binding domain, revealing a high structural heterogeneity at the UGT1 gene locus. The new UGT1A7 proteins differ in their primary structure at amino acid positions 129, 131 and 208, creating four distinct UGT1A7 allelic variants in the human population: UGT1A7*1 (N129 R131 W208), *2 (K129 K131 W208), *3 (K129 K131 R208), and *4 (N129 R131 R208). In functional studies, HEK cells stably transfected to express the four allelic UGT1A7 variants exhibited significant differences in catalytic activity towards 3-, 7-, and 9-hydroxy-benzo(a)pyrene. UGT1A7*3 exhibited a 5.8-fold lower relative Vmax compared to wild-type *1, whereas *2 and *4 had a 2.6- and 2.8-fold lower relative Vmax than *1, respectively, suggesting that these mutations confer slow glucuronidation phenotype. Kinetic characterization suggested that these differences were primarily attributable to altered Vmax. Additionally, it suggested that each amino acid substitutions can independently affect the UGT1A7 catalytic activity, and that their effects are additive. The expression pattern of UGT1A7 studied herein and its catalytic activity profile suggest a possible role of UGT1A7 in the detoxification and elimination of carcinogenic products in lung. A population study demonstrated that a considerable proportion of the population (15.3%) was found homozygous for the low activity allele containing all three missense mutations, UGT1A7*3. These findings suggest that further studies are needed to investigate the impact of the low UGT1A7 conjugator genotype on individual susceptibility to chemical-induced diseases and responses to therapeutic drugs.

Induction of UDP-glucuronosyltransferase UGT1A1 by the flavonoid chrysin in the human hepatoma cell line hep G2.

The UDP-glucuronosyltransferases (UGTs) have long been known to be inducible by various chemicals, including drugs, although the extent of induction in general has been modest. In the present study, we determined the ability of the dietary flavonoid chrysin to induce UGT activity, protein and mRNA. When pretreating human hepatoma Hep G2 cells with 25 microM chrysin, the glucuronidation of chrysin itself increased 4.2-fold when measured in the intact cell and 14-fold in the cell homogenate, i.e., autoinduction. Microsomes from chrysin-treated cells probed with specific antibodies in Western analyses showed marked induction of the UGT1A family of proteins. Isoform-specific induction of the important hepatic UGT1A1 protein was observed but not of UGT1A6 or UGT2B7. The strong induction of UGT1A1 was confirmed by Northern analyses of total RNA as well as mRNA, using a specific probe. UGT1A1 message as well as protein was detectable also in untreated Hep G2 cells. In catalytic activity assays with recombinant UGT1A1, 1A4, 1A6 and 1A9, chrysin was found to be a high affinity substrate for UGT1A1 (K(m) 0.35 microM). Catalytic activity was also found for UGT1A9 and 1A6 but not for 1A4. Further studies demonstrated a 20-fold induction of the glucuronidation of bilirubin by the chrysin-treated cells and a 7. 9-fold induction of the glucuronidation of the oral contraceptive drug ethinylestradiol, two of the best known and specific UGT1A1 substrates, demonstrating the potential importance of this induction. In view of these findings, it will be important to extend these studies to other dietary flavonoids.

Involvement of hepatocyte nuclear factor 1 in the regulation of the UDP-glucuronosyltransferase 1A7 (UGT1A7) gene in rat hepatocytes.

UDP-glucuronosyltransferase 1A7 (UGT1A7) is a major UGT contributing to the glucuronidation of xenobiotic phenols in rats. Its expression in rat liver is tightly regulated, with low constitutive and high inducible expression in response to aryl hydrocarbon receptor ligands and oltipraz. Previously, we reported the absence of 3-methylcholanthrene- or oltipraz-responsive elements in the 1.6-kbp region flanking the UGT1A7 promoter. However, potential binding sites were noted for several liver-enriched transcription factors. Here we show that deletion of the hepatic nuclear factor (HNF)3, HNF4, and CCAAT-enhancer binding protein-like binding sites had no effect on the expression of a UGT1A7 reporter plasmid, p(-965/+56)1A7-Luc, in primary rat hepatocytes. The full activity of the promoter was contained in the region between bases -157 and +76. Two sites of binding by rat liver nuclear proteins were detected in this region by DNase footprinting. PR-1 corresponded to the HNF1-like binding site between bases -52 and -38, whereas PR-2 was located between -30 to -6. Gel retardation studies supported the presence of HNF1alpha in the PR-1 DNA-liver nuclear protein complex. Mutation of PR-1 inhibited binding in the gel shift assay, prevented activation by overexpressed HNF1 in human embryonic kidney cells, and reduced by >80% the maximal luciferase activities expressed from basal and 3-methylcholanthrene-responsive UGT1A7 gene reporter constructs in primary rat hepatocytes. These data provide evidence for an important stimulatory role of HNF1 in promoting UGT1A7 gene expression in rat liver.

Differential protection by rat UDP-glucuronosyltransferase 1A7 against Benzo[a]pyrene-3,6-quinone- versus Benzo[a]pyrene-induced cytotoxic effects in human lymphoblastoid cells.

UDP-glucuronosyltransferase 1A7 (UGT1A7) is a polyaromatic hydrocarbon (PAH)-inducible UGT with activity toward various benzo[a]pyrene (B[a]P) metabolites. To investigate the influence of rat UGT1A7 on B[a]P-induced cytotoxicity, human lymphoblastoid L3 cells were transfected with pMF6 (control expression vector), p167Dtk2 (microsomal epoxide hydrolase expression vector), or p167Dtk2-1A7 (epoxide hydrolase/UGT1A7 coexpression vector), and the cell populations were compared for sensitivity to B[a]P-induced effects. B[a]P inhibited cell proliferation and decreased relative cell survival of p167Dtk2 and p167Dtk2-1A7 cells to a similar extent. Metabolism studies using [(3)H]B[a]P revealed increased formation of glucuronide conjugates of B[a]P-4,5-diol, 3-OH-, or 9-OH-B[a]P and an unidentified metabolite by p167Dtk2-1A7 cells, but the presence of unconjugated metabolites suggested that glucuronidation capacity may be limited. No differences between p167Dtk2 and p167Dtk2-1A7 L3 cells were observed in the growth inhibitory effects of 3-OH-B[a]P or B[a]P-7,8-diol, but p167Dtk2-1A7-expressing cells were found to be less sensitive to B[a]P-3,6-quinone-induced effects on cell proliferation and relative cell survival. The effect was also observed in AHH-1 lymphoblastoid cells expressing UGT1A7 without epoxide hydrolase. The UGT1A7-expressing AHH-1 cells were also less sensitive to growth inhibition by B[a]P-1,6-quinone and B[a]P-6,12-quinone. Flow cytometric analysis of vehicle and B[a]P-3, 6-quinone-exposed cell populations showed an association between UGT1A7 expression and resistance to B[a]P-3,6-quinone-induced apoptosis and loss of cell viability. These data suggest that UGT1A7 may be preferentially active toward B[a]P-quinones and that UGT1A7 may represent the PAH-inducible UGT activity previously implicated in protection against toxic redox cycling by B[a]P-3,6-quinone.

Roles of glucuronidation and UDP-glucuronosyltransferases in xenobiotic bioactivation reactions.

Glucuronide conjugates represent one of the major types of naturally occurring phase 2 metabolites of xenobiotics and endobiotics. The process underlying their formation, glucuronidation, is normally considered detoxifying, because glucuronides usually possess less intrinsic biological or chemical activity than their parent aglycones and they are rapid excreted. However, a number of glucuronide conjugates are known that are active and may contribute to pharmacological activities or toxicities associated with their parent compounds. These include two classes of glucuronides with electrophilic chemical reactivity (N-O-glucuronides of hydroxamic acids and acyl glucuronides of carboxylic acids) and several types of glucuronides that impart biological effects through non-covalent interactions (morphine 6-O-glucuronide, retinoid glucuronides, and D-ring glucuronides of estrogens). Glucuronides may thus contribute to clinically significant effects, including environmental arylamine-induced carcinogenesis, drug hypersensitivity and other toxicities associated with carboxylic acid drugs, morphine analgesia, and cholestasis from estrogens. This review summarizes the rat and human UDP-glucuronosyltransferases that may be involved in the formation of bioactive glucuronides, including their substrate- and tissue-specificity and genetic and environmental influences on their activity. This knowledge may be useful for enhancing the therapeutic efficacy and minimizing the risk of adverse effects associated with xenobiotics that undergo bioactivating glucuronidation reactions.

Liposome-mediated gene transfer in rat lung transplantation: A comparison between the in vivo and ex vivo approaches.

OBJECTIVE: We compared the efficacy of in vivo and ex vivo liposome transfection in rat lung transplantation. METHODS: (1) Chloramphenicol acetyltransferase group: Fischer rats underwent isogeneic transplantation (n = 4 per group). Recipients were put to death on postoperative day 2 for chloramphenicol acetyltransferase activity. Ex vivo setting: Grafts received cDNA complexed or not with liposomes and were transplanted after 1.5 or 10 hours at 10 degreesC. In vivo setting: Donors were intravenously injected with cDNA complexed or not with liposomes. Lungs were harvested after 1.5 or 10 hours, preserved at 10 degreesC, and transplanted. (2) Transforming growth factor-beta1 group: Brown-Norway rats served as donors and Fischer rats as recipients. All grafts were preserved for 3 hours at 10 degreesC. On postoperative day 5, arterial oxygenation and histologic rejection scores were assessed. Ex vivo setting: Grafts received transforming growth factor-beta1 sense (n = 8) or antisense (n = 7) complexed with liposomes or cDNA alone (n = 5). In vivo setting: Donors were intravenously injected with liposome:transforming growth factor-beta1 sense cDNA (n = 7). Exposure time was 3 hours. RESULTS: (1) Chloramphenicol acetyltransferase-transfection was superior in the ex vivo group but was not statistically different for longer exposure times. (2) Transforming growth factor-beta1-arterial oxygenation was superior in the ex vivo liposome:sense group. cDNA alone was inefficient. Rejection scores were not statistically different between ex vivo and in vivo liposome:sense groups but were better when the ex vivo liposome:sense group was compared with the cDNA alone or the antisense groups. CONCLUSIONS: (1) With current liposome technology, the ex vivo route is superior to the in vivo approach; (2) cDNA alone does not provide transgene expression at levels to produce a functional effect.

Transcriptional activation of the UDP-glucuronosyltransferase 1A7 gene in rat liver by aryl hydrocarbon receptor ligands and oltipraz.

UDP-glucuronosyltransferase UGT1A7 catalyzes the glucuronidation of benzo(a)pyrene metabolites and other bulky aromatic compounds. Both UGT1A7 mRNA and an associated enzyme activity (benzo(a)pyrene7, 8-dihydrodioltransferase activity) are markedly increased in livers of rats treated with beta-naphthoflavone or 4-methyl-5-pyrazinyl-3H-1,2-dithiole-3-thione (oltipraz). Nuclear runoff assays show that the effects of both inducers are primarily due to transcriptional activation. A 27-kilobase region that included the UGT1A7/UGT1A6 promoter regions was cloned. Primer extension and RNase protection studies indicated >/=30 transcription start sites in five clusters between bases -85 and -40 respective to the translation start codon. There was no recognizable TATA box, but the promoter region is TA-rich. Sequence analysis revealed potential binding sites for CCAAT enhancer-binding protein, activator protein 1, and hepatic nuclear factors 1, 3, and 4, but no xenobiotic response elements or antioxidant response elements, implicated in the regulation of other genes by beta-naphthoflavone or oltipraz, were found. A UGT1A7 gene reporter plasmid directed strong constitutive expression in transient transfection assays using primary rat hepatocytes. Treatment with 3-methylcholanthrene or oltipraz had no effect compared with similarly treated pGL3-Basic-transfected cells. These results suggest that the regulatory elements controlling xenobiotic inducibility of UGT1A7 transcription are located either 5' or 3' of bases -1600 to +54. One possibility is that the polycyclic aromatic-mediated regulation of UGT1A7 occurs via the xenobiotic response element flanking the UGT1A6 locus 7 kilobase pairs downstream.

Aryl hydrocarbon receptor-inducible or constitutive expression of human UDP glucuronosyltransferase UGT1A6.

Transcriptional regulation of human UGT1A6, a UDP glucuronosyltransferase isoform conjugating a wide variety of planar phenols, has been studied using transfection experiments with plasmids containing its 3-kb 5' upstream region and chloramphenicol acetyltransferase as reporter gene. Previously, two modes of expression of the isoform have been described: in colon carcinoma Caco-2 cells UGT1A6 was found to be TCDD-inducible, whereas in lung carcinoma A549 cells it was constitutively expressed. Therefore functional analysis of UGT1A6 regulation was carried out using these two cell lines. In the upstream region of human UGT1A6 one xenobiotic-responsive element (XRE) was found between-1498 and -1502 bp. In Caco-2 cells the reporter gene activity of the entire plasmid and of deletion mutants containing the XRE were TCDD-inducible, in contrast to experiments with a deletion mutant which did not contain the XRE. TCDD induction was marginal in transfection studies with A549 cells. Gel mobility shift analysis indicated that the aryl hydrocarbon receptor and its partner Arnt bind to the XRE. Furthermore, primer extension studies suggest cell-specific use of multiple TATA boxes. Hence, regulation of human UGT1A6 appears to be cell-specific including both constitutive and aryl hydrocarbon receptor-controlled expression.

Identification of a rat oltipraz-inducible UDP-glucuronosyltransferase (UGT1A7) with activity towards benzo(a)pyrene-7,8-dihydrodiol.

Previous work has shown that polycyclic aromatic hydrocarbons and oltipraz both induce an unidentified rat liver UDP-glucuronosyltransferase with activity toward benzo(a)pyrene-7, 8-diol, the proximate carcinogenic form of benzo(a)pyrene. Here we report the isolation of a benzo(a)pyrene-7,8-diol transferase-encoding cDNA, LC14, from an adult rat hepatocyte-derived cell line (RALA255-10G LCS-3). The predicted amino acid sequence of LC14 is nearly identical (5 differences out of 531 residues) to that deduced from UGT1A7, recently cloned at the genomic DNA level (Emi, Y., Ikushiro, S., and Kyanagi, T. (1995) J. Biochem. (Tokyo) 117, 392-399). Northern analysis of RNA from female F344 rat liver and LCS-3 cells revealed over a 40-fold and 4.4-fold enhancement by oltipraz treatment, respectively. Benzo(a)pyrene-7, 8-diol glucuronidating activity was detected (0.4 nmol/10(6) cells/16 h) in AHH-1 cells transfected with the LC14 expression vector, pMF6-LC14-3. The LC14-encoded transferase exhibited even higher activity toward certain benzo(a)pyrene phenols, including the major 3- and 9-phenol metabolites (4.1 and 2.8 nmol/10(6) cells/16 h, respectively). The Km of the enzyme for (-)-trans benzo(a)pyrene-7, 8-diol and 3-OH-BP was 15.5 and 12.3 microM, respectively. Northern analyses of total RNA revealed expression of LC14 or LC14-like RNA in all extrahepatic tissues tested. Marked inducibility by oltipraz was observed only in liver and (to a lesser extent) intestine. The results suggest that induction of UGT1A7 may explain the increased glucuronidating activities toward benzo(a)pyrene-7,8-diol and other metabolites that occur following treatment with polycyclic aromatic hydrocarbon-type inducing agents and oltipraz. UGT1A7 appears to represent an important cellular chemoprotective enzyme which mediates conjugation and elimination of toxic benzo(a)pyrene metabolites.

Induction of a rat liver benzo[a]pyrene-trans-7,8-dihydrodiol glucuronidating activity by oltipraz and beta-naphthoflavone.

We report here that rats possess a hitherto unrecognized xenobiotic-inducible hepatic 7,8-dihydro-7,8-diol-benzo[a]pyrene (BPD) UDP-glucuronosyltransferase (UGT) activity. BPD UGT activity is induced in female F344 rat liver by treatment with the selective Phase 2 conjugation enzyme inducer oltipraz [4-methyl-5-(2-pyrazinyl)-1,2-dithiole-3-thione at 75-450 mg/kg per day for 3 days] and also by a polycyclic aromatic hydrocarbon-type inducer, beta-naphthoflavone (80 mg/kg per day for 3 days). Incubations of oltipraz-treated rat liver microsomes with racemic trans BPD (100 microM) resulted in formation of two fluorescent glucuronides that were resolved by silica thin layer chromatography (Rf 0.5 and 0.6). Incubations with either the (-) or (+) trans BPD isomers resulted in selective formation of the Rf 0.5 [designated -DS, for (-) diol specific] or Rf 0.6 [designated +DS, for (+) diol specific] glucuronide, respectively. The -DS and +DS BPD glucuronides were fluorescent under long wave ultraviolet irradiation, dependent on the presence of UDP-glucuronic acid in the incubation, and were beta-glucuronidase-sensitive. The inducing effect of oltipraz on BPD UGT activity was dose-dependent. The mean BPD UGT activity of the vehicle-treated control group was 0.05 +/- 0.02 nmol/mg per min compared with 0.53 +/- 0.07 nmol/mg per min in the group treated with oltipraz (450 mg/kg per day for 3 days) (P < 0.001). The apparent Km of the induced BPD UGT for BPD was 20 microM, suggesting that the enzyme has the capacity to bind and turnover BPD under physiological conditions. Pretreatment with beta-naphthoflavone, but not phenobarbital, induced BPD UGT activity to approximately the same extent as oltipraz. Neither oltipraz nor beta-naphthoflavone exhibited induction of BPD UGT in livers of homozygous Gunn rats, which lack functional UGT1-encoded isozymes. We conclude that the oltipraz- and polycyclic hydrocarbonresponsive BPD UGT is a member of the UGT1 family. The role of this isoform as a modifier of susceptibility to carcinogenesis elicited by B[a]P remains to be determined.

Two human liver cDNAs encode UDP-glucuronosyltransferases with 2 log differences in activity toward parallel substrates including hyodeoxycholic acid and certain estrogen derivatives.

Two human liver UDP-glucuronosyltransferase cDNA clones, HLUG25 [Jackson, M. R., et al. (1987) Biochem. J. 242, 581-588] and UDPGTh-2 [Ritter, J. K., et al. (1990) J. Biol. Chem. 266, 7900-7906] have previously been shown to encode isozymes active in the glucuronidation of hyodeoxycholic acid (HDCA) and certain estrogen derivatives (estriols and 3,4-catechol estrogens), respectively. Here we report that the UDPGTh-2-encoded isoform (udpgth-2) and the HLUG25-encoded isoform (udpgth-1) have parallel aglycon specificities. Following expression in COS-1 cells, each isoform metabolized three types of dihydroxy- or trihydroxy-substituted ring structures, including the 3,4-catechol estrogen (4-hydroxyestrone), estriol and 17-epiestriol, and HDCA, but the udpgth-2 isozyme is 100-fold more efficient than udpgth-1. udpgth-1 and udpgth-2 are 86% identical overall (76 differences out of 528 amino acids), including 55 differences in the first 300 amino acids of the amino terminus, a domain which confers isoform substrate specificity. The data indicate that a high level of conservation in the amino terminus is not required for the preservation of substrate selectivity. Analysis of glucuronidation activity encoded by UDPGTh-1/UDPGTh-2 chimeric cDNAs constructed at their common restriction sites, SacI (codon 297), NcoI (codon 385), and HhaI (codon 469), showed that nine amino acids between residues 385 and 469 are important for catalytic efficiency, suggesting that this region represents a domain which is critical for catalysis but distinct from that responsible for aglycon selection.(ABSTRACT TRUNCATED AT 250 WORDS)

Mouse pulmonary cytochrome P-450 naphthalene hydroxylase: cDNA cloning, sequence, and expression in Saccharomyces cerevisiae.

We have isolated a cDNA clone, Nah-2, encoding the cytochrome P-450Nah (naphthalene hydroxylase) from a mouse lung lambda ZAP cDNA library using anti-cytochrome P-450Nah IgG as a probe. This same antibody selectively blocked [Nagata, K., Martin, B.M., Gillette, J.R., & Sasame, H.A. (1990) Drug Metab. Dispos. 18, 557-564] the cytochrome P-450 in mouse lung microsomes that catalyzed the conversion of naphthalene to (1R,2S)-naphthalene 1,2-oxide, which has been postulated as a causative agent in the naphthalene-induced tissue-specific necrosis of Clara cells in mouse lung. The toxic effect is seen in mouse and not in rat. The cDNA encodes a polypeptide of 491 amino acids with a molecular mass of 50 kDa. Northern blot analysis with an Nah-2-specific probe revealed that the mRNA is expressed in a species- and tissue-specific manner, present only in mouse lung and liver and not in that of rat. The mRNA encoding Nah-2 is constitutively expressed and is not induced by either phenobarbital, pyrazole, pregnenolone 16 alpha-carbonitrile, or 3-methylcholanthrene. Comparative amino acid sequence analyses with other documented members of the P-450 gene superfamily revealed that this encoded protein is in the IIF subfamily. To analyze its substrate specificity, the cDNA was inserted into the vector, pAAH5, and expressed in the Saccharomyces cerevisiae strain, AH22. The presence of cytochrome P-450Nah in the microsomes isolated from transformed cells and analyzed by Western blot was confirmed by immunocomplexing product with anti-cytochrome P450Nah IgG. Furthermore, activity toward naphthalene in the microsomes from the transformed cells established that this clone encodes a naphthalene hydroxylase.(ABSTRACT TRUNCATED AT 250 WORDS)

A putative AP-2 binding site in the 5' flanking region of the mouse POMC gene.

Using extracts of AtT-20 cell nuclei, protein binding sites on the POMC gene 5'-flanking region were examined with an exonuclease protection approach. One such binding site, located from -119 to -106 bp upstream from the mouse POMC gene transcription initiation site, which exhibited a close homology to the activator protein-2 (AP-2) site [1]. A double-stranded oligonucleotide containing this site was subsequently used in gel shift assays to demonstrate AP-2 consensus sequence binding activity in extracts of AtT-20 cell nuclei. Gel shift competition experiments using both homologous and heterologous competitor DNA sequences revealed that the AP-2 like factor(s) exhibited specific binding to the mouse AP-2 consensus sequence. Furthermore, AP-2 factor binding was also modulated by a CTF/NF1-like factor. Pretreatment of AtT-20 cell nuclear extracts with alkaline phosphatase prior to inclusion in gel shift assays led to a reduction in the intensities of AP-2 factor-specific bands, indicating a potential involvement of protein phosphorylation in AP-2 factor binding in AtT-20 cells.