Tm JOURNAL OF BIOWGICAL C H E M I ~ Y 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 269, No. 29, Issue of July 22, pp. 19028-19033, 1994 Printed in U.S.A.

Agonistic and Antagonistic Effects of a-Naphthoflavone on Dioxin Receptor Function ROLE OF THE BASIC REGION HELIX-LOOP-HELIX DIOXIN RECEPTOR PARTNER FACTOR Amt*

(Received for publication, November 12, 1993, and in revised form, March 31, 1994)

Anna WilhelmssonS, Murray L. Whitelaw, J a n - h e Gustafsson, and Lorenz PoellingerP From the Department of Medical Nutrition, Karolinska Institute, Huddinge University Hospital F-60, Novum, S-141 86 Huddinge, Sweden

The dioxin receptor is a ligand-dependent transcrip- tion factor that binds to target DNA sequences (xenobi- otic responsive elements, XREs) following ligand-de- pendent dimerization with its partner factor, Arnt (aryl hydrocarbon receptor nuclear translocator). Both fac- tors contain an N-terminal basic region helix-loop-helix motif mediating dimerization and subsequent DNA binding. In this study we investigate the possible role of Arnt in agonistic and antagonistic effects of the dioxin receptor ligand a-naphthoflavone (ANI?). Using specific antisera for the ligand binding dioxin receptor and Arnt, respectively, we show that exposure of the dioxin recep- tor to ANJ? in vitro induced recruitment of Arnt, thus stimulating binding of the heteromeric complex to XRE. In transient transfection assays, ANF at high concentra- tions stimulated expression of an XRE-driven reporter gene. This agonistic effect of ANF is, therefore, most likely attributable to ANF stimulation of dioxin recep- tor-Arnt heterodimerization and subsequent binding of the complex to XRE. Using a minimal XRE-driven re- porter gene construct, we could further confirm earlier studies showing that ANF antagonizes the effect of a dioxin receptor agonist, 2,3,7,8-tetrachlorodibenzo-p-di- oxin. Next we employed chimeric receptor constructs containing amino acids 1-500 of the human glucocorti- coid receptor fused to dioxin receptor fragments lacking the very N-terminal basic region helix-loop-helix dimer- ization and DNA binding motif. These chimeric receptor constructs show dioxin responsiveness upon transient transfection into mutant Arnt-deficient hepatoma cells and are, thus, functionally uncoupled from Arnt. Impor- tantly, dioxin-dependent activation of the chimeric re- ceptors was inhibited in the presence of A N F , demon- strating that dimerization of dioxin receptor with Arnt was not necessary for manifestation of the antagonistic effect of ANF. Rather, dioxin receptor sequences, which confer dioxin regulation upon a heterologous DNA bind- ing and transactivating domain, also mediated the an- tagonistic effects of ANF.

2,3,7,8-Tetrachlorodibenzo-p-dioxin (dioxin, TCDD)’ induces

ety. The costs of publication of this article were defrayed in part by the * This work was supported by grants from the Swedish Cancer Soci-

“advertisement” in accordance with 18 U.S.C. Section 1734 solely to payment of page charges. This article must therefore be hereby marked

indicate this fact. i Present address: KaroBio Inc.. BlickagAngen 6. P. 0. Box 4032,

S-141 04 Huddinge, Sweden. ” .

5 To whom correspondence should be addressed. Tel.: 468-779-7124; F&: 468-711-6659.

chlorodibenzo-p-dioxin; bHLH, basic region helix-loop-helix; XRE, xe- The abbreviations and trivial names used are: TCDD, 2,3,7,8-tetra-

nobiotic responsive element; hsp90, 90-kDa heat shock protein; Amt,

transcription of a battery of target genes including the cyto- chrome P450IA1 gene via the intracellular basic helix-loop- helix (bHLH) dioxin receptor, also termed aryl hydrocarbon receptor (Ema et al., 1992; Burbach et al., 1992). Upon exposure to dioxin, the receptor acquires an increased affinity for target regulatory DNA sequences (xenobiotic responsive elements, XREs) that confer dioxin regulation on target promoters. Acti- vation of the receptor to an XRE binding form is accompanied by its apparent translocation from the cytoplasm to the nucleus of target cells (reviewed by Poellinger et al. (1992)). The recep- tor activation pathway is not fully understood but is clearly a multistep process.

In nonstimulated cells the receptor is recovered as a -300- kDa complex containing the ligand binding dioxin receptor and the molecular chaperon hsp90 (90-kDa heat shock protein) (De- nis et al., 1988; Perdew, 1988). There may be other components of this large receptor form (Perdew, 1992), but these have so far not been identified. Binding of ligand induces release of hsp90 from the -100-kDa receptor (Wilhelmsson et al . , 1990) and dimerization (Whitelaw et al., 199313; Matsushita et al., 1993; Probst et al., 1993) with the structurally related bHLH factor Arnt (Hoffman et al . , 1991). The nuclear receptor-Arnt het- erodimer shows strong XRE binding activity (Reyes et al., 1992; Dolwick et al., 1993; Whitelaw et al., 199313) and, through an unclear mechanism, alters the transcription rate of target genes (for a review, see Fujii-Kuriyama et al. (1992)).

Among dioxin receptor ligands there seems to be a good cor- relation between the affinity of a given ligand, its in vitro ac- tivating capacity (measured as induction of XRE binding activ- ity of the receptor), and its potency to induce transcription of the cytochrome P450IA1 gene in vivo (Cuthill et al., 1991). a-Naphthoflavone (ANF) binds to the dioxin receptor but shows no agonistic effect, or it does so only at very high (in the milli- molar range) concentrations (Blank et al., 1987; Santostefano et al., 1993 and references therein). Exposure of target cells to both dioxin and ANF or exposure of the dioxin receptor to both ligands in vitro has indicated that ANF antagonizes the effect of dioxin (Blank et al., 1987; Gasiewicz and Rucci, 1991; Berghard et al., 1992; Merchant et al., 1992; Santostefano et al., 1993). The mechanism of action of ANF remains, however, un- clear. In the present study we have, therefore, investigated regulation of a minimal XRE-driven reporter gene construct by ANF in vivo and examined the effect of ANF on the DNA bind- ing function of the dioxin receptor in an in vitro reconstituted system. Finally, dioxin receptor fragments fused to a heterolo- gous DNA binding motif were used in transfection experiments with wild-type and mutant Arnt-deficient hepatoma cells to

aryl hydrocarbon receptor nuclear translocator; ANF, 7,8-benzoflavone (a-naphthoflavone); BNF, 5,6-benzoflavone (p-naphthoflavone); hGH, human growth hormone; CHO, Chinese hamster ovary; DBD, DNA binding domaids); MMTV, mouse mammary tumor virus.

19028

Regulation of Dioxin Receptor Function by a-Naphthoflavone 19029

investigate whether ANF-mediated responses functionally re- quired the dioxin receptor partner factor Arnt.

MATERIALS AND METHODS Plasmids-The reporter gene constructs pMMTV-AF (generously

provided by Dr. Stefan Nilsson, KaroBio Inc., Stockholm, Sweden) and pXRE-MMTV-hGH have been described previously (Alksnis et al., 1991; Whitelaw et al., 1993b). Construction of pGemArnt (Whitelaw et al., 1993a), pAmtAbHLH (designated pArntAHLH-2 in Mason et al. (1994)), and pDFUATGBS has been described previously. The metallothionein promoter-driven vectors pMT-TDBD, pMT-~DBD/83-805, and pMT- ~DBD/83-593 (Whitelaw et al., 1993a) contain chimeric glucocorticoid- dioxin receptor derivatives spanning amino acids 1-500 (here desig- nated TDBD) of the human glucocorticoid receptor (Hollenberg et al., 1985) fused to amino acids 83-805 and 83-593, respectively, of the murine dioxin receptor (Burbach et al., 1992; Ema et al., 1992).

Cell Culture and fiansfections-Hepa lclc7 cells and a phenotypi- cally dioxin receptor nuclear translocation-deficient variant of it (Hepa c4) were kept in an atmosphere of 6% CO, in minimal essential medium supplemented with 10% fetal calf serum, 2 mM L-glutamine, 100 IU of penicillidml, and 100 mg of streptomycidml. CHO cells were cultured in Ham's F-12 medium with the same supplements as for Hepa lclc7 cells. All media and supplements were from Life Technologies, Inc. For transient transfections, cells were seeded at a density of 105/35 x 10-mm dish and grown to -50% confluence (approximately 20 h). Reporter gene constructs (1 pg) were transfected with 5 pg of DOTAP (Boehringer Mannheim) or 10 pl of lipofectin (Life Technologies, Inc.) in the absence or presence of metallothionein promoter-driven expression vectors (1 pg) in 1 ml of medium as described by the manufacturer. After 12 h of transfection, the medium was replaced with fresh medium containing the indicated concentrations of dioxin receptor ligands dissolved in di- methyl sulfoxide (final concentration of dimethyl sulfoxide, 0.1% (dv)). Ligand treatment was continued for 72 h. Secreted alkaline phospha- tase activity was determined by a colorimetric assay as described pre- viously (Gottlicher et al., 1992). Secreted hGH levels were assayed by a radioimmunoassay (Pharmacia Biotech Inc.).

Electrophoretic Mobility Shift Assay-Gel mobility shift analysis of the DNA binding activity of the Hepa lclc7 dioxin receptor was per- formed as described previously (Hapgood et al., 1989). To this end, a cytosolic extract was prepared from untreated Hepa lclc7 cells in ETG buffer (20 mM Tris, pH 7.4, 10% (w/v) glycerol, 1 mM EDTA, 2 mM 2-mercaptoethanol) as described (Wilhelmsson et al., 1990). Treatment with TCDD (Chemsyn Laboratories, Lenexa, KS) or ANF (Serva, Hei- delberg, Germany) dissolved in dimethyl sulfoxide was for 2.5 h a t 25 "C. In experiments where antiserum was included in the reaction mixture, it was added together with the Hepa lclc7 cytosolic extract.

Dioxin Receptor and Arnt Antisera-Peptides corresponding to amino acids 12-31 (Burbach et al., 1992; Ema et al., 1992) of the murine dioxin receptor or amino acids 39-58 (HoEman et al., 1991) of human Arnt were used to generate specific antisera in rabbits. Briefly, the peptides were coupled to ovalbumin via a cysteine residue added to their C termini using a heterobifunctional cross-linking reagent (SPDP, Phar- macia). Conjugates were injected into rabbits, and serum was collected 10-14 days after dosing. A peptide corresponding to amino acids 762- 777 of the human glucocorticoid receptor (Hollenberg et al., 1985) was used as control in epitope competition experiments.

In Vitro Expression of Dioxin Receptor, Wild-type, or Mutant Arnt Proteins-Dioxin Receptor, wild-type, or mutant Arnt mRNA was syn- thesized in vitro from pDWATGBS, pGemArnt, or pArntAbHLH using T3 or T7 polymerase and subsequently translated in rabbit reticulocyte lysates (Promega, Madison, WI) in the presence of either [35 Slmethi- onine (DuPont NEN) or 20 PM unlabeled methionine under conditions suggested by the manufacturer.

Safety Precautions-Due to the toxicity of dioxin receptor ligands, their use requires special handling as outlined previously (Wilhelmsson et al. 1990 and references therein). Contaminated materials were dis- posed of by high temperature incineration.

RESULTS AND DISCUSSION Agonistic and Antagonistic Effects of A N F on Activity of a

Minimal XRE-Driven Reporter Gene Construct-To investigate effects ofANF on dioxin receptor function we transiently trans- fected wild-type Hepa lclc7 cells with the XRE-containing pXRE-MMTV-hGH reporter gene. As schematically repre- sented in Fig. lA, this reporter gene contains a minimal mouse mammary tumor virus (MMTV) promoter in which endogenous

A MMTV n

-I XRE 1 I hGH

. . g g

FIG. 1. Transient transfection of an XRE-driven reporter gene into wild-type Hepa lclc7 cells. A, schematic representation of re- porter gene pXRE-MMTV-hGH in which endogenous glucocorticoid re-

of the cytochrome P4501A1 promoter. B, reporter gene pXRE-MMTV- sponsive sequences have been substituted with a single XRE 1 element

hGH was transfected into wild-type Hepa lclc7 cells. Cells were treated as indicated for 72 h. Levels of secreted hGH were measured by radio- immunoassay. hGH secretion in vehicle-treated cells was arbitrarily set to 1. Results from a representative experiment are shown.

glucocorticoid response elements have been substituted with a single XRE (XRE1) (Fujisawa-Sehara et al., 1987). In Hepa lclc7 cells this reporter gene shows very low basal activity in the absence of treatment and strong transcriptional activation following the addition of 1-10 n~ dioxin (Whitelaw et al., 199313) (Fig. 1B). Activation of the reporter gene was also ob- served upon the addition ofANF, although only a fraction of the maximal dioxin induction reponse was detected in the presence of 10-50 p~ ANF (Fig. 1B). Co-administration of dioxin and ANF, however, resulted in an ANF dose-dependent inhibition of the dioxin induction response (Fig. 1B). Thus, in agreement with earlier data on cytochrome P450IA1 mRNA and enzyme induction (Merchant et al., 19921, these experiments demon- strate that ANF has both agonistic (at high doses) and antago- nistic effects on dioxin receptor function in Hepa lclc7 cells. ANF appears to compete with dioxin for the same ligand bind- ing site in ligand binding experiments in vitro (Blank et al., 1987; Merchant et al . , 1990). However, its affinity for the re- ceptor is approximately 40 times lower than that of dioxin (Blank et al., 1987). In our experimental system it is clear that even 50,000-fold higher concentrations of ANF relative to TCDD did not induce full activity of the dioxin receptor and that the same concentration ofANF antagonized the induction response produced by dioxin. ANF is, therefore, appropriately defined as a partial dioxin receptor antagonist.

A N F Stimulates XRE Binding Activity of Dioxin Receptor in Vitro by Inducing Recruitment of bHLH Factor Arnt- Following the addition of dioxin (10 nM) or ANF (1 p ~ ) to a cytosolic extract from nontreated Hepa lclc7 cells in vitro, a ligand-induced protein-XRE complex can be detected in a gel mobility shift assay using a labeled XRE fragment as specific

19030 Regulation of Dioxin Receptor Function by a-Naphthoflavone

A

B

Cytosolic Extract:

Ligand:

Serum:

Ligand:

Serum:

Peptide:

+ U u)

A bHLH P I S

Arnt .

I M I r J . : 1

AmtAbHLH IS In

ANF: n-

" " + - + - .

(L C R

1 2 3 4 5 6 7 8 9

z I 0 TCDD - z

I D TCDD -

a-DR a-Amt

" "

10

1 3 5 7 9 1 1 1 3

FIG. 2. Detection of dioxin receptor and Arnt in the XRE-bind- ing dioxin receptor form generated by TCDD or ANF treatment of Hepa lclc7 cytosol. A, Iirpa lclc7 cytosol \vas incuhatcd i n [ i t ro with 10 nlv TCTII) or 1 p~ ASF and annlyzrd in grl mohility shift experiments in the presence or ahsence of antisera against the ligand binding dioxin rrceptor or Arnt. h n r 2 rrprrsrnts vrhicle dimethyl sulfoxide (DMSO )-treated c.ytosol in the ahsrnce of serum. I a n m 3 - 6 , TU)I)-treatrd cytosol: Ionrs 7-10, ANF-trrated cytosol. Antiserum against the ligand binding dioxin receptor (n-DN) is included in lnnes 4 nnd 8: Innrs :i nnd 9 contain Amt antisrrum I w A m t 1. h n r s 6 and 10 are ligand-treated control samplrs annlyzed in the presence of preim- munc srrum Cl?I.S. 1. 13. epitopr competition experiments. A cytosolic extract from nontreated Hrpn lclc7 cells was trratcd with dioxin as described above, and DNA rcwtions were assembled in the presence or ahsence of either dioxin or Arnt antisera, as indicated. Synthetic pep- tides ( 10 ng) spanning regions of the dioxin receptor (Dl?; Innas 5 and I 1 I. glucocorticoid rrceptor (GR; Innrs 6 nnd 12 ). or Arnt ( Ianrs 7 nnd 1 3 ) were added to certain reactions. Lanes 2 and 8 show the DNA hinding activity of the extract treated with vehicle dimethyl sulfoxide alone. In hnth panels, lnnr I shows the mohility of the XRE prohe in the ahsrnce of any added protein. R and F drnotc migration of the ligand- induced receptor complex and free prohe. respectively.

probe (Fig. 2 4 ) . To characterize this complex, we employed monospecific antibodies against the dioxin receptor (Whitelaw et al . , 1993b) and Arnt (Mason et al., 1994), respectively. As expected from earlier observations (Reyes et al., 1992; Matsus- hita et al., 1993; Whitelaw et al., 1993b3, dioxin treatment resulted in induction of a n XRE complex, the formation of which was inhibited in the presence of dioxin receptor antibod-

1 2 3 4 5 6 7 0

FIG. 3. Reconstitution of ANF-induction of the DNA binding activity of the dioxin receptor. role of the bHLH motif of Arnt. Pnnd A, schrmatic rrpresrntation of wild-type Amt and the narrow h1.II.H delrtion mutant. PASS dcwotes the Per-Xrnt-Sim homology region contnining the two P.4S rrpeats A and B I /iIIad haws I . I'nnal R. DSA binding rractions wrre assrmhled with the ligand-free i n vitro trans-

Arnt hHLH deletion mutant AmtlhHLH individually or mixed to- Iatrd dioxin rccrptor (Dl? ), in ritro translatrd wild-t?.pe Amt. or the

gether, as indicated. The material was trrated with 1 p~ AVF, and XRE hinding activity was analyzed as descrihrd in thr legend to Fig. 2.

ies. Addition of antibodies against the bHLH partner factor Arnt resulted in lower mobility of the ligand-induced complex (Fig. 2A ). In contrast, XRE complex formation was not signifi- cantly affected by the addition of preimmune serum. By anal- ogy, formation of the ANF-induced XRE complex was specifi- cally inhibited by the receptor antibodies, and Arnt antibodies retarded the ligand-dependent XRE complex. The specificity of the antibodies was examined in competion experiments em- ploying the synthetic dioxin receptor and Arnt peptides that were used as epitopes for immunization, and an unrelated pep- tide of similar length and charge spanning a fragment of the hormone binding domain of the glucocorticoid receptor. In the presence of the dioxin receptor petide, no inhibition of DNA binding activity by the dioxin receptor antibodies was observed (Fig. 2R, compare lanes 4 6 ), whereas the addition of the Arnt or glucocorticoid receptor peptides did not alter the inhibitory effect of the antibodies (lanes 6 and 7). Conversely, the Arnt peptide inhibited formation of a supershifted protein-DNAcom- plex by the Arnt antibodies. However, the antibodies produced the supershift in the presence of the dioxin and glucocorticoid receptor peptides (Fig. 2R, compare lanes 9-13 ).

We next reconstituted activation of the dioxin receptor by ANF using in vitro expressed dioxin receptor and Arnt proteins. In contrast to a recent report (Dolwick et al., 1993). these pro- teins did not show any XRE binding activity when incubated together in the absence of ligand (Fig. 3R, lane 5). However, ANF treatment induced XRE binding activity of the dioxin receptor-Arnt complex (lane 6) . In control experiments, a nar- row Arnt deletion mutant, ArntlbHLH, which lacks the mini-

Regulation of Dioxin Receptor Function by a-Naphthoflavone 19031

A

Regulatory factors:

rDBD

rDBDI83-805

rDBD/83-593

B

C

hGR 1-500 mDR I I

21 DBD I

PAS Q . . . . . . . . . \\\A a

I 83 805

. . . . . . . . \\\Y I i

83 593

n

> v Factor: - 2 DMT-rDBDI83-805

a l - ?!l

B 0

l-

m

Y Factor: % pMT-tDBDI83-593

l o T 7

constructs by TCDD and ANF'. A, schematic representations of chi- FIG. 4. Regulation of chimeric glucocorticoid-dioxin receptor

meric receptor constructs and the reporter gene construct pMMTV-AF (top). Chimeras contain the N-terminal 500 amino acids of the human glucocorticoid receptor (hGR), including the transcriptional activation domain 71 and the DNA binding domain (DBD), fused to amino acids 83-805 or 83-593 of the murine dioxin receptor (mDR). Dioxin receptor structures lack the N-terminal bHLH motif of the receptor but contain a region with homology to Arnt and the Drosophila proteins Per and

mal bHLH domain (as schematically represented in Fig. 3 A ) and fails to dimerize with the dioxin receptor (Mason et al., 1994), did not enable the dioxin receptor to bind to DNA in the presence ofANF (Fig. 3B, compare lanes 7 and 8). We conclude from these experiments that the ANF-occupied dioxin receptor is competent to recruit the Arnt partner factor, resulting in XRE binding activity, and that both ANF and dioxin treatment of the receptor resulted in stimulation of receptor-Arnt het- erodimerization. In strong support of this model, the Arnt de- letion mutant ArntAbHLH lacking the minimal bHLH dimer- ization motif failed to induce the DNA binding activity of the ANF-stimulated dioxin receptor in a reconstituted cell-free sys- tem (Fig. 3B). Finally, in agreement with the functional data in vivo (Fig. 11, the employed dose (1 p ~ ) of ANF was less potent than 10 nM dioxin in inducing XRE binding activity in vitro by the receptor-Arnt heterodimer (Fig. 2 A ) , possibly due to a lower efficacy in recruitment of the Arnt partner factor.

Interestingly, we failed to detect any differences in the rela- tive mobility of the XRE complex generated by either the ANF- or the dioxin-activated dioxin receptor-Arnt heterodimers, ar- guing that no conformational difference in the activated recep- tor forms was detectable by this method. In contrast, binding of certain antiestrogens and antiprogestins to the estrogen and progesterone receptors, respectively, has been reported to re- sult in receptor-target DNA complexes with different relative mobilities as compared with complexes generated by agonist- activated receptors (Kumar and Chambon, 1988; Brown and Sharp, 1990; Meyer et al., 1990; Klein-Hitpass et al., 1991; Beck et al., 1993).

Antagonism of Dioxin Receptor Function by ANF Does Not Require bHLH Dioxin Receptor Partner Factor Arnt-Our re- sults indicate that the ANF-occupied dioxin receptor is tran- scriptionally active in vivo (Fig. lB) and is able to recruit the Arnt co-factor via the bHLH motif to acquire XRE binding activity in vitro (Figs. 2 and 3). To address the mechanistically important question of whether dioxin receptor-Amt dimeriza- tion is a target for the antagonistic activity of ANF, we used chimeric glucocorticoid-dioxin receptor derivatives containing the first 500 amino acids of the human glucocorticoid receptor fused to various fragments of the murine dioxin receptor. As schematically shown in Fig. 4A, the employed glucocorticoid receptor fragment spans the major transactivation (71) (GiguBre et al., 1986; Hollenberg and Evans, 1988) and DNA binding domains (DBD) but not the C-terminal hormone bind- ing domain of the receptor. This glucocorticoid receptor deriva- tive (here termed TDBD) shows constitutive transcriptional ac- tivity (Godowski et al., 1987). Importantly, the dioxin receptor fragments that were fused to TDBD all lack the very N-terminal bHLH motif of the dioxin receptor.

Upon transient transfection into CHO cells, the chimeric glucocorticoid-dioxin receptor derivatives pMT-~DBD/83-805 and pMT-7DBDI83-593 show repression of the activity of the glucocorticoid response element-driven reporter gene pMMTV- AF. However, this negative regulation is reversed following

Sim (PAS). Q denotes a glutamine-rich region with similarity to tran- scriptional activation domains of other transcription factors. B, Co- transfection of pMT-~DBD/83-805 and pMMTV-AF into CHO cells. Cells were treated with the indicated concentrations of ligand for 72 h prior to analysis. In control experiments, CHO cells were cotransfected with expression vector lacking insert (pCMV) and the reporter gene. Results from a representative experiment are shown. C, Cotransfection of pMT-~DBD/83-593 and pMMTV-AF into CHO cells. Cells were treated as above. All columns represent cotransfection of pMMTV-AF with ~MT-TDBD, pMT-~DBD/83-593, or pMT-~DBD/83-805 as indi- cated. Results from a representative experiment are shown. Inpanels B and C, phosphatase activity in the cell medium of vehicle-treated cells was arbitrarily set to 1.

19032 Regulation of Dioxin Receptor Function by a-Naphthoflavone

8 T . .

7 -- 6

--

C g 5" < 4 "

; 3 " - U

2

" 1

--

0 . I - , UJgzYzYzYzY

E S d S B Treatment: DMSO e 2 2 2 2

- + 7 g g r

FIG. 5. Cotransfection of pMMTV-AF and pMT-~DBDh33-805 into Arnt-deficient Hepa c4 cells. Cells were trasiently transfected and treated as indicated. For details, see legend to Fig. 4. Results of a representative experiment are shown. DMSO, dimethyl sulfoxide.

treatment with dioxin (Whitelaw et al., 1993a) (Fig. 4, B and C). This effect was observed with both chimeric receptor forms. The dioxin receptor residues deleted in the shorter construct (amino acids 593-805 of the murine dioxin receptor) contain a glutamine-rich region bearing similarity to transcriptional ac- tivation domains in other transcription factors (Ema et al., 1992; Burbach et al., 1992; Mitchell and Tjian, 1989). The li- gand binding domain of the dioxin receptor (contained within amino acids 230421) (Dolwick et al., 1993; Whitelaw et al., 1993a) is, however, intact, demonstrating that this region of the dioxin receptor confers dioxin regulation upon a heterologous transcription factor (Whitelaw et al., 1993a). Treatment with 1 w ANF resulted in no or a very minor induction of chimeric receptor activity (Fig. 4, B and C). When both ANF and dioxin were administered together, ANF suppressed dioxin induction in a dose-dependent manner (Fig. 4, B and C). Adioxin receptor agonist, BNF (which is structurally very similar to ANF), in- duced expression of the reporter gene irrespective of the chi- meric receptor construct used. Unlike ANF, BNF did not an- tagonize the effect of TCDD in co-administration experiments (data not shown). In conclusion, the antagonistic effect of ANF is observed in dioxin receptor derivatives lacking the bHLH motif, and the 212 most C-terminal amino acids of the dioxin receptor (possibly harboring a transactivation domain) (Whitelaw et al., 1993a) are not necessary for the antagonistic activity.

Given the background that deletion of the bHLH motif of Arnt abrogates dimerization with the dioxin receptor in co- immunoprecipitation assays (Whitelaw et al., 199313; Mason et al., 1994) and detectable interaction in DNA binding assays (Fig. 3B), there is reason to believe that the bHLH dioxin receptor deletion mutants do not functionally interact with Arnt. However, it has recently been shown that the PAS domain of the Drosophila Per protein functions as a protein dimeriza- tion motif (Huang et al., 1993). A similar PAS domain (sche- matically represented in Fig. 4A) is also present in the dioxin receptor and Arnt, raising the possibility that the bHLH region may not be the only motif involved in dimerization between the two proteins.

To unambiguously test the role of Arnt, if any, in the antago- nistic activity of ANF, we cotransfected pMT-~DBD/83-805 to- gether with the GRE-containing reporter gene into the mutant

hepatoma cells Hepa c4, which are deficient in Arnt function (Hoffman et al., 1991). These cells do not mediate any dioxin induction of XRE-driven reporter genes including pXRE- MMTV-hGH (Whitelaw et al., 1993b and references therein). In contrast, MMTV promoter-driven reporter gene activity was increased by the chimeric glucocorticoid-dioxin receptor ap- proximately 6-fold upon treatment with dioxin (1 nM). Treat- ment with 1 or 10 w ANF did not significantly alter the activ- ity of the reporter gene. When TCDD (1 m), and ANF (1 or 10 p ~ ) were added together, ANF completely inhibited the induc- tion response observed with dioxin treatment alone (Fig. 5) . Since agonistic effects ofANF become apparent when very high doses are used (Santostefano et al., 1993; Berghard et al., 1992), we also tested the effect of 50 p~ ANF on the functional activity of the chimeric receptor construct. In our system, this concen- tration had no effect on expression of the reporter gene. Fur- thermore, as expected, BNF had an agonistic effect when added alone and did not significantly decrease the response to dioxin of the chimeric receptor when the two compounds were added together (data not shown). Taken together, the present data strongly argue that Arnt is not required for the antagonistic activity of ANF and that dioxin receptor-Amt dimerization is not a primary target for the antagonistic action of ANF. Dioxin receptor motifs located between amino acids 83 and 593 and, thus, including the PAS domain, appear rather to contain the structural determinants for this effect to occur. We have ob- served that this region mediates both dioxin binding and stable association with hsp90 (Whitelaw et al., 1993a). I t is, therefore, possible that ANF is less efficient than TCDD at promoting release of hsp90 from the ligand binding dioxin receptor sub- unit. Alternatively, we cannot rule out the possibility that the dioxin receptor contains at least one ligand-dependent tran- scriptional activation domain between amino acids 83 and 805 and that ANF is incapable of inducing its activity. In a similar fashion, it has been shown that the hormone-dependent trans- activation function, TAF-2, of the progesterone receptor is in- active in the presence of the anti-hormone RU486 (Meyer et al., 1990). However, in analogy to our limited understanding of ANF action, the detailed mechanism of glucocorticoid and pro- gesterone receptor antagonism by RU486 is not yet understood, and several mechanisms including effects prior to, during, and subsequent to activation of the DNA binding activity of these receptors have been proposed to mediate antagonism by RU486 (for review, see Baulieu (1991)). Clearly, the present chimeric receptor system provides a tool that will be helpful in the iden- tification and dissection of dioxin receptor structures that me- diate the antagonistic activity of ANF.

Acknowledgments-We thank Kristina Sjoblom for excellent techni- cal assistance, Dr. Oliver Hankinson (UCLA) for originally supplying wild-type and c4 mutant Hepa cells, and Dr Stefan Nilsson (KaroBio Inc., Stockholm, Sweden) for the pMMTV-AF reporter gene construct.

REFERENCES

Alksnis, M., Barkhem, T., Stromstedt, P.-E., Ahola, H., Kutoh, E., Gustafsson, J.-A., Poellinger, L., and Nilsson, S. (1991) J. Biol. Chem. 266, 10078-10085

Baulieu, E.-E. (1991) Dends Endocrinol. Metab. 2, 233-239 Beck, C. A,, Weigel, N. L., Moyer, M. L., Nordeen, S. K., and Edwards, D. P. (1993)

Berghard, A., Gradin, K, and Toftgbrd, R. (1992) Carcinogenesis 13,651455 Blank, J. A,, Tucker,A. N., Sweatlock, J., Gasiewicz, T. A., and Luster, M. I. (1987)

Proc. Natl. Acad. Sei. U. S. A. 90,4441-4445

Brown, M., and Sharp, P. A. (1990) J. Biol. Chem. 266,1123&11243 Burbach. K. M.. Poland. A,. and Bradfield. C. A. (1992) Proc. Natl. Acad. Sci.

Mol. Pharmacol. 32, 168-172

~ ~~ ~

U. S. A:&, 8185-8189 '

Cuthill, S., Wilhelmsson, A., and Poellinger, L. (1991) Mol. Cell. B i d . 11,401-411 Denis, M., Cuthill, S., Wikstrom, A,-C., Poellinger, L., and Gustafsson, J . d . (1988)

Dolwick, K. M., Swanson, H. I., and Bradfield, C. A. (1993) Proc. Natl. Acad. Sci. Biochem. Biophys. Res. Commun. 166,801-807

Ema, M., Sogawa, K., Watanabe, N., Chujoh, Y., Matsushita, N., Gotoh, O., Funae, U. S. A. 90,856643570

Y., and Fujii-Kuriyama, Y. (1992) Biochem. Biophys. Res. Commun. 184, 246- 253

Regulation of Dioxin Receptor Function by a-Naphthoflavone 19033 Fujii-Kuriyama, Y., Imataka, H., Sogawa, K., Yasumoto, K.-I., and Kikuchi, Y.

Fujisawa-Sehara, A,, Sogawa, K., Yamane, M., and Fujii-Kuriyama, Y. (1987)

Gasiewicz, T. A., and Rucci, G. (1991) Mol. Pharrnacol. 40, 607412 Gigube, V., Hollenberg, S. M., Rosenfeld, M. G., and Evans, R. M. (1986) Cell 46,

Godowski, P. J., Rusconi, S., Miesfeld, R., and Yamamoto, K. R. (1987) Nature 326,

Gttlicher, M., Widmark, E., Li, Q., and Gustafsson, J.-A. (1992) Proc. Natl. Acad.

Hapgood, J., Cuthill, S., Denis, M., Poellinger, L., and Gustafsson, J.d. (1989)

Hoffman, E. C., Reyes, H., Chu, E-F., Sander, E , Conley, L. H., Brooks, B. A., and

Hollenberg, S. M., Weinberger, C., Ong, E. S., Cerelli, G., Oro, A,, Lebo, R., Thomp- Hollenberg, S. M., and Evans, R. M. (1988) Cell 66,899-906

son, E. B., Rosenfeld, M. G., and Evans, R. M. (1985) Nature 318, 635441 Huang, Z. J., Edery, I., and Rosbash, M. (1993) Nature 364,259-262 Klein-Hitpass, L., Cato, A. C. B., Henderson, D., and Ryffel, G. U. (1991) Nucleic

Kumar, V., and Chambon, P. (1988) Cell 66, 145-156 Mason, G. G. F., Witte, A,, Whitelaw, M. L., Antonsson, C., McGuire, J., Wilhelms-

4449 son, A,, Poellinger, L., and Gustafsson, J.-A. (1994) J. Biol. Chem. 269.443%

(1992) FMEB J. 6,706710

Nucleic Acids Res. 16, 4179-4191

645-652

365-368

Sci. U. S. A. 89,4653-4657

Proc. Natl. Acad. Sci. U. S. A. 86, 6044

Hankinson, 0. (1991) Science 252,954-958

Acids Res. 19,1227-1234

Matsushita, N., Sogawa, K., Ema, M.,Yoshida,A., and Fujii-Kuriyama, Y. (1993) J.

McGuire, J . , Whitelaw, M. L., Pongratz, I., Gustafsson, J.-A., and Poellinger, L.

Merchant, M.,Arellano, L., and Safe, S. (1990)Arch. Biochem. Biophys. 281,8449 Merchant, M., Morrison, V., Santostefano, M., and Safe, S. (1992)Arch. Biochem.

Meyer, M.-E., Pornon, A,, Ji, J., Bocquel, M.-T., Chambon, P., and Gronemeyer, H.

Mitchell, P. J., and Tjian, R. (1989) Science 246, 371378 Perdew, G. H. (1988) J. Biol. Chem. 263, 13802-13805 Perdew, G. H. (1992) Biochem. Biophys. Res. Commun. 182,55-62 Poellinger, L., Gttlicher, M., and Gustafsson, J . a . (1992) Dends Pharrnacol. Sci.

Probst, M. R., Reisz-Porszasz, S., Agbunag, R. V., Ong M. S., and Hankinson, 0.

Reyes, H., Reiz-Porszasz, S., and Hankinson, 0. (1992) Science 266, 1193-1195 Santostefano, M., Merchant, M., Arellano, L., Morrison, V., Denison, M. S., and

Whitelaw, M., Gttlicher, M., Gustafsson, J.-A., and Poellinger, L. (1993a) EMBO

Whitelaw, M., Pongratz, I., Wilhelmsson, A,, Gustafsson, J.-A., and Poellinger, L.

Wilhelmsson, A., Cuthill, S., Denis, M., Wikstrom, A X . , Gustafsson, J.-A., and

Biol. Chem. 268,21002-21006

(1994) Mol. Cell. Biol. 14, 2438-2446

Biophys. 298, 389394

(1990) EMBO J. 9,3923-3932

13,241-245

(1993) Mol. Pharmacol. 44,511-518

Safe, S. (1993) Mol. Pharrnacol. 43,200-206

J. 12,4169-4179

(1993b) Mol. Cell. Biol. 13, 2504-2514

Poellinger, L. (1990) EMBO J. 9,69-76