Communication Vol. 268, No. 26, Issue of September 15, pp. 19169-19172, 1993 THE JOURNAL OF BIOLIXICAL CHEMISTRY

0 1993 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

Chicken Ovalbumin Upstream Promoter Transcription Factor (COUP-TF) Binds to a Peroxisome Proliferator-responsive Element and Antagonizes Peroxisome Proliferator-mediated Signaling*

(Received for publication, June 10, 1993, and in revised form, June 30, 1993)

Keqji 5. MiyataS, Baowei ZhangS, Sandra L. Marcus$, John P. Caponell, and Richard A. Rachubinskill From the Department of Biochemistry, McMaster University, Hamilton, L8N 325 Ontario, Canada

Peroxisome proliferators form a family of diverse xe- nobiotic compounds that includes hypolipidemic agents, herbicides, and plasticizers. These compounds activate transcription of a subset of nuclear genes in- cluding those encoding peroxisomal fatty acid P-oxida- tion enzymes, whose elevated activities can lead to hepatocarcinogenesis. Induction of the genes encoding fatty acyl-CoA oxidase and hydratase-dehydrogenase, the first and second enzymes of the pathway, is mediated by peroxisome proliferator-activated nuclear receptors (PPARs) that bind to upstream responsive elements (PPREs) through heterodimerization with retinoid X re- ceptors. We demonstrate that the chicken ovalbumin up- stream promoter transcription factor 1 (COUP-TFl), an- other member of the nuclear hormone receptor superfamily, binds to the hydratase-dehydrogenase PPRE in vitro and in vivo and antagonizes PPAR-depen- dent signaling. These data suggest that members of the COUP-TF family play a role in modulating receptor-me- diated activation of peroxisome proliferator-responsive genes.

Peroxisomes are essential for normal lipid homeostasis (1,2). A number of xenobiotics, including amphipathic carboxylates used in the treatment of hyperlipidemia, induce peroxisome proliferation and ultimately hepatocarcinogenesis in rodents (3). These so-called peroxisome proliferators are nongenotoxic carcinogens that promote tumor growth apparently by modu- lating genes involved in cell growth and differentiation (4, 5).

*This work was supported in part by grants from the Heart and Stroke Foundation of Ontario (to J. P. C. and R. A. R.). 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.

$ These authors should be considered equivalent first authors. 5 Recipient of a Medical Research Council of Canada studentship. 11 National Cancer Institute of Canada Senior Scientist. To whom

correspondence should be addressed. Tel.: 416-525-9140 (ext. 2774); Fax: 416-522-9033.

ence should be addressed. Tel.: 416-525-9140 (ext. 2316); Fax: 416-522- I1 Medical Research Council of Canada Scientist. To whom correspond-

9033.

The administration of peroxisome proliferators leads to the coordinated transcriptional induction of the nuclear genes en- coding the enzymes of the peroxisomal P-oxidation pathway: fatty acyl-CoA oxidase, enoyl-CoA hydratasd3-hydroxyacyl- CoA dehydrogenase (HD),’ and 3-ketoacyl-CoA thiolase (6-8). cis-Acting peroxisome proliferator-responsive elements (PPREs) have been identified in the 5”flanking regions of the genes encoding fatty acyl-CoA oxidase and HD (9-12). Both PPREs contain direct repeats of the sequence TGACCT, which is the consensus binding site for several members of the nuclear hormone receptor superfamily (13, 141, including the vitamin DB, thyroid, and retinoic acid receptors. These receptors bind to their cognate sites through cooperativity with the retinoid X receptor, RXR (15-17). Signal transduction by peroxisome pro- liferators is mediated by a similar mechanism involving ligand- activated receptors called peroxisome proliferator-activated re- ceptors (PPARs) that belong to this family of transcription factors (18-21). PPARs bind cooperatively to PPREs through heterodimerization with the 9-cis-retinoic acid receptor, RXRa

The transcriptional activation mediated by steroid receptors is subject to negative regulation. For example, the chicken oval- bumin upstream promoter-transcription factors (COUP-TFs), which are orphan members of the nuclear hormone receptor superfamily, have been shown to repress hormonal induction of vitamin D3, thyroid, and retinoic acid receptor target genes by competing with these receptors for their binding sites (27, 28) and by heteromerization with RXR (28). COUP-TFs bind as homodimers to diverse response elements consisting of TGACCT repeats which exhibit wide variation in spacing and orientation (27); however, they have the greatest affinity for direct repeats separated by 1 base pair (DR1). The HD PPRE consists of three imperfect direct repeats of the TGACCT motif separated by 2 base pairs (DR2) and 1 base pair (DRl), respec- tively (Fig. lA; Refs. 11 and 12). Herein we show that human (h) COUP-TF1 (which is identical to the human orphan recep- tor Ear3; Ref. 29) binds to the HD PPRE in vitro and represses peroxisome proliferator induction mediated by PPARs in vivo.

MATERIALS AND METHODS

(12, 22-26).

Cells-Rat hepatoma H4IIEC3 cells were cultured as monolayers in Dulbecco’s modified Eagle’s medium containing 10% horse serum and 5% fetal bovine serum. HeLa cells were maintained in suspension in Joklik’s modified medium plus 5% fetal bovine serum. BSC4O cells were maintained in Dulbecco’s modified Eagle’s medium plus 10% calf serum.

Plasmids-pCPSZuc is a luciferase expression vector containing the minimal promoter for the gene encoding rat liver carbamoyl phosphate synthetase (11). Plasmids for expression in BSC40 cells were con- structed by cloning an oligonucleotide corresponding to the HD PPRE ( 5 ’ - g a t C C T C T C C T T T G A C C T A ’ I T G A A C T A C A and its complement 5 ‘ - g a t c T C A A A T G T A G C A G T A G G T C A A A G - GAGAG or mutant forms of this PPRE, into the unique BamHI site of pCPSZuc located upstream of the carbamoyl phosphate synthetase pro- moter (25). pHD(X1)luc and pHD(X3)Zuc contain one copy and three direct tandem copies of the HD PPRE oligonucleotide, respectively, in the same 5’ + 3’ orientation found in the natural HD promoter (25). Rat (r) PPAR and RXRa vectors for in vivo and in vitro expression have

hydroxyacyl-CoA dehydrogenase; COUP-TF, chicken ovalbumin up- The abbreviations used are: HD, enoyl-CoA hydratasd3-

stream promoter factor transcription factor; EMSA, electrophoretic mo- bility shift analysis; PPAR, peroxisome proliferator-activated receptor; PPRE, peroxisome proliferator-responsive element; RXR, retinoid X re- ceptor; RXRa, 94s-retinoic acid receptor; r, rat; h, human.

19169

19170 COUP-TFl Antagonizes Signaling by Peroxisome Proliferators been described (26). A plasmid expressing hCOUP-TF1 was constructed by ligating the HindIIYXbaI fragment from pCOUP-TFl (27) into the corresponding sites of pWCMV (Invitrogen). A plasmid encoding a truncated form of hCOUP-TF1 (tCOUP-TF1; Ref. 27) was constructed by cloning the SmYAatII fragment from pCOUP-TFl into the EcoRV and AatII sites of pGEM-5Zff +) (Promega). Transcription and transla- tion of this plasmid in vitro generated a truncated hCOUP-TFl protein lacking 51 amino acids a t i t s amino terminus.

Dannfections and Measurement of Luciferase Activity-Transfections of H4IIEC3 cells were done by the calcium phosphate method followed by a dimethyl sulfoxide shock (11). BSC40 cells were transfected simi- larly except that cells were incubated for 24 h before and during trans- fection in medium without phenol red and containing 5% charcoal- stripped fetal bovine serum (26). Transfections typically contained 5 pg of a reporter gene construct and, where indicated, 2 pg of rPPAR, 2 pg of RXRa, and 0-4 pg of hCOUP-TFl expression plasmids. Effector plas- mid dosage was kept constant by the addition of appropriate amounts of the corresponding empty expression vector (pSG5 or pRC/CMV). The total amount of DNA was kept at 20 pg with sonicated salmon sperm DNA. Ciprofibrate ( x 100 stock in dimethyl sulfoxide) was added to fresh medium to a final concentration of 0.5 mM. Extracts were prepared 48-h post-transfection, and luciferase activity was measured.

ing rPPAFt, RXRa, HNF4, hCOUP-TFl, and tCOUP-TFl and subse- In Vitro Danscription /Dunslation-Transcription of cDNAs encod-

quent translation in rabbit reticulocyte lysate were performed using a commercially available kit (Promega). Translations of proteins for use in electrophoretic mobility shift assays (EMS&) were done with unla- beled methionine.

Electrophoretic Mobility Shift Analysis-Nuclear extracts from HeLa cells were prepared from cell suspensions (30). Nuclear extracts were prepared from monolayer cultures of H4IIEC3, BSC40, and BSC4O cells transfected with various expression plasmids as described (30, 31). EMSA was performed as described (11.26). Double-stranded probes of the wild-type and mutant forms of the HD PPRE were end-labeled with [a-D2PldATP and the Klenow fragment of DNA polymerase I. Binding reactions were analyzed by electrophoresis a t 4 “C on prerun 3.5% poly- acrylamide gels (30:l acrylamiddNJV”methylenebisacrylamide weight ratio) with 22 mM Tris base, 22 mM boric acid, 1 mM EDTA as running buffer. For binding reactions done with in vitro synthesized protein, 1-2 pl of translation mixture was incubated with labeled probe in a final volume of 15 pl. The total amount of reticulocyte lysate was kept con- stant in each reaction by the addition of unprogrammed lysate.

RESULTS AND DISCUSSION The HD PPRE consists of three imperfect direct repeats of

the consensus nuclear hormone binding motif TGACCT sepa- rated by 2 base pairs (DR2) and l base pair (DRl), respectively (Fig. L4). We have demonstrated that an oligonucleotide cor- responding to the HD PPRE interacts with cellular factors pre- sent in H4IIEC3 cells, a rat hepatoma cell line responsive to peroxisome proliferators (11, 25), and that the PPRE-binding proteins include rPPAR and RXRa (26). To explore whether members of the COUP-TF family are present in these com- plexes, EMSA was camed out with radiolabeled HD PPRE probe and nuclear extracts prepared from H4IIEC3 cells in the presence of antibody to hCOUP-TF1 (hereafter called COUP- TF1). This antibody recognizes both COUP-TF1 and a related receptor COUP-TF2 (27), which is also called ARP-1 (32). As shown in Fig. 1B, inclusion of anti-hCOUP-TF1 resulted in the formation of supershifted complexes (lane f ) , demonstrating that the PPRE-binding proteins include COUP-TF-related fac- tors. The major protein-DNA complex formed between HD PPRE and nuclear extracts of HeLa cells was almost quantita- tively supershifted with anti-hCOUP-TF1 (lane c). Therefore, COUP-TF-related factors in both rat hepatoma and HeLa nu- clear extracts bind to the HD PPRE.

To determine whether COUP-TF1 could bind directly to the HD PPRE, EMSA was camed out with in vitro translated re- ceptor. As shown in Fig. 2A, in oitro translated COUP-TF1 bound to the HD PPRE (lane e; the identity of COUP-TF1 in these complexes was established using anti-hCOUP-TF1 se- rum; data not presented). COUP-TF1 also bound to the PPRE of the gene encoding fatty acyl-CoA oxidase (data not shown).

A: om DRl

HD PPRE: CCTCTCCTTTGACC‘TATTGAA~TATTAC~~ACATTTCA

B: HeLn amt nopmtomm

a b c d e f FIG. 1. A COUP-TF-related factor is present in protein-DNA

complexes formed on the HD PPRE. A. sequence of the HD PPRE. The arrows designate the TCACCT direct repeat motifs. E . electro- phoretic mobility shift assays. Nuclear extracts prepared from H e h cells or rat hepatoma H41IEC3 cells were incubated with radiolabeled HD PPRE probe alone (lanes a and d ) or with addition of preimmune serum (lanes b and e ) or of polyclonal anti-hCOUP-TF1 serum (lane8 c and f 1, as indicated. The brackets indicate novel complexes of reduced mobility generated in the presence of anti-hCOUP-TF1 serum. Where indicated, 0.25 pl of preimmune serum ( P I ) or of anti-hCOUP-TF1 serum was added to the binding reactions, which were then preincu- bated for 5 min prior to addition of probe.

Several distinct PPARa bind to the HD PPRE cooperatively through heteromerization with RXRa (26). To determine the DNA sequence requirements for binding COUP-TF1 versus the cooperative binding of PPAR and RXRa on the HD PPRE. we used oligonucleotides in which each of the core TGACCT motifs was individually mutated. Fig. 2B shows that mutation of the first repeat (M4) had no effect on binding of in vitro translated rPPAR/RXRa (lane c) but completely abolished COUP-TF1 binding (lane d ) . A single mutation in the second repeat (M3) reduced significantly binding of both rPPAWRXRa (lane e ) and COUP-TF1 (lanef). Mutation of the third repeat (M5) inhibited rPPAR/RXRa binding (lane g) but did not affect COUP-TF1 binding (lane h) . Therefore, COUP-TF1 and rPPAR/RXRa rec- ognize distinct, yet overlapping, core elements within the HD PPRE, with the first two repeata being the important determi- nant for COUP-TF1 binding and the second and third repeats being necessary for rPPAR/RXRa binding. The DR2 spacing of the first two repeats is not necessary for COUP-TF1 binding, as an oligonucleotide (M6) in which the DR2 was converted to a DR1 bound COUP-TF1 strongly (lane j). Therefore, the pri- mary sequence of the first two repeats, and not their relative spacing, are important for binding of COUP-TF1.

The mutated PPREs were assessed for their ability to confer peroxisome proliferator responsiveness onto a luciferase re- porter gene in transient transfection assays of rat H4IIEC3 cells. As shown in Fig. 3, the activities of the wild-type HD PPRE reporter constructs pHD(X1)luc and pHMX3)luc were induced approximately 3- and l&fold, respectively, in the pres- ence of the peroxisome proliferator ciprofibrate, similar to what has been shown previously (25). Reporter constructs containing either one copy or three tandem copies of the different mutant

COUP-TFl Antagonizes Signaling by Peroxisome Proliferators 19171

a b c d e

M5 Me

5 ~ ~ ~ ~ r ~ ~ l l i ~ ~ l A r ~ M l T ; t r r n r c Z T l c r r r i c r

5 ~ ~ ~ ~ ~ ~ ~ l l ~ ( ~ 1 ~ ~ A l l G U r ~ l A l l A C C l A C A l l l G A 15 5 ~ ~ I c ~ c c I I ~ G ~ C C ~ A I ~ G U G ~ A ~ ~ ~ ~ ~ C A C A ~ I ~ G ~

5 CClc:cc~I1Gmxl llGUCIIllACClACAIl~G*

u3 C C I C T C C I I T G ~ C C ~ A ~ T G U ~ I ~ ~ I A C C I A C A ~ ~ ~ G A

FIG. 2. COUP-TFl translated in a i h binds to the HD PPRE. A, rPPAR, human RXRa, and COUP-TF1 RNAs were translated in uitro, incubated with labeled HD PPRE probe, and analyzed by EMSA. Ad- ditions were as indicated at the top of the figure. Lane d contained a mixture of equal volumes of rPPAR and RXRa translation reactions to show that efficient binding to the HD PPRE requires both receptors. B , EMSAs were carried out with the wild-type HD PPRE or with mutant oligonucleotides in which the first (M4), second (M3), or third (M5) TGACCT repeats were individually mutated or in which the A between the first and second repeats was deleted (M6). The probes were incu- bated with in uitro translated rPPAR and RXRa (lanes a , c, e, g. and i) or COUP-TF1 (lanes b, d, f, h, and j ) , a s indicated.

PPREs were unresponsive to ciprofibrate. These results extend our previous observation (25) that multimerization of the wild- type HD PPRE serves only to amplify the behavior of the re- sponse element to ciprofibrate and has little effect on basal level transcription with either the wild-type or mutant HD PPREs. Importantly, the results of Fig. 3 demonstrate that the sequences required for COUP-TF1 binding in uitm are also necessary for peroxisome proliferator responsiveness in uiuo. Moreover, they confirm our finding that PPAR/RXRa binding is necessary but not sufficient for peroxisome proliferator medi- ated induction via the HD PPRE (26) and demonstrate that all three repeats of the HD PPRE and the particular spacing of the repeats are involved in the responsiveness to proliferators, sug-

pHDMG(X3)Iuc ~ - 1

5 10 15 2 0

Relative Luciferase Activity

FIG. 3. Mutant HD PPREs are unable to confer peroxisome proliferator responsiveness in duo. Luciferase reporter constructs containing the wild-type or mutant HD PPREs M3. M4. MS. and M6 were individually cloned as single copies (X1 ) or as three direct tandem copies (X3) upstream of reporter plasmid pCPSluc and transfected into H4IIEC3 cells, which were subsequently treated with dimethyl sulfox- ide alone or with the peroxisome proliferator ciprofibrate. The values were normalized to the activity of control transfections done with pCPSluc in the absence of ciprofibrate. which was taken as 1. The values are from two independent transfections done in duplicate and did not vary by more than 15%.

gesting the involvement of additional transcription factors in regulating the response to ciprofibrate.

COUP-TFs repress induction by hormone receptors through the formation of inactive DNA-binding heterodimers (28). Since COUP-TF1.DNA complexes had the same electrophoretic mo- bility vis a uis those formed with rPPAR/RXRa (see Fig. 2A,

lanes d and e ) , we constructed a truncated form of COUP-TF1 (tCOUP-TF1) that permitted better resolution of potential het- eromeric complexes. As seen in Fig. 4, tCOUP-TF1 formed a complex with the HD PPRE (lane d ) that migrated faster than the complex formed between full-length COUP-TF1 and HD PPRE (lane c). Amixture of tCOUP-TF1 and full-length COUP- TF1 generated a complex of intermediate size (lane h, a m w - head), showing that COUP-TF1 dimerizes on the HD PPRE. Mixing tCOUP-TF1 with RXRa (fane f), rPPAR (fane i ) . or HNF4 (lane g), another member of the nuclear hormone recep tor family (33). did not generate any novel complexes, indicat- ing that COUP-TF1 does not form PPRE-binding heteromera with these receptors. In addition, an excess of RXRa itself was unable to inhibit COUP-TF1 binding (data not presented), s u g gesting that COUP-TF1 homodimer binding to the PPRE is not appreciably affected by the formation of COUP-TF1.RXRu het- erodimers in solution.

We wished to determine whether COUP-TFl could down- regulate PPAR/RXRa-mediated transactivation in uiuo. The pHD(X3)Zuc reporter plasmid was cotransfected with expres- sion vectors for rPPAR, RXRa, and COUP-TF1 into BSC4O cells. BSC40 cells are unresponsive to peroxisome proliferatora in the absence of co-transfected PPAR and RXRa (Fig. 5 ) and do not contain factors that bind to the HD PPRE (11). A 10-fold induction in luciferase activity was observed in the presence of both rPPAR and RxRa, which was further increased by the addition of the ciprofibrate. Addition of increasing amounta of COUP-TF1 expression vector resulted in the progressive re- pression of rPPAR/RXRa-mediated induction with pHD(X3)luc. Repression of induction was specific, as the transfection of COUP-TF1 expression vector alone had no effect on the basal

19172 COUP-TF1 Antagonizes Signaling by Peroxisome Proliferators RXRa + + + HNF4 + + +

CCUP-TFl + ICCUP-TFl + + + + +

rPPAR + + + + - A-b--&-* - - . *.. .

a b c d e f g h i j k no. 4. COUP-TF1 does not form DNA-binding heteromera In

vitro translated RXRa, rPPAR, HNF4, COUP-TF1, and tCOUP-TF1 were incubated as indicated with the wild-type HD PPRE probe and analyzed by EMSA. The arrow indicates the complex formed by a COUP-TFldCOUP-TF1 heterodimer.

,ppAR - - + + + + RXR" - + + + + +

Clproflbrata - + - +

COUP-TF - - - - 0.6 2 4

Fro. 5. COUP-TF1 antagoniee~~ peroxieome proliferator-medi- ated dgnaling. pHWX3)luc was transfected into BSC4O cells along with effector plasmids expressing rPPAR, RxRa, and COUP-TFl, as indicated. The values shown are relative activities from repeat trans- fections done in duplicate and were normalized to the value obtained for ciprofibrate-treated cells cotransfected with rPPAR and RXRa expres- sion plasmids, which was taken as 100%. The numbers for COUP-TFl refer to the amount (in pg) of cotransfected plasmid expressing COUP- TF1.

activity of pHD(X3)Zuc. Expression of reporter constructs con- taining mutant HD PPREs was not stimulated by cotransfec- tion with plasmids expressing rPPAR and RxRa in either the presence or absence of ciprofibrate (data not shown), in agree- ment with the transfection results presented in Fig. 3. There- fore, COUP-TF1 antagonizes PPAR-mediated peroxisome pro- liferator signaling in vivo. This finding, combined with the results of Fig. 4, suggests that COUP-TF1-mediated antago- nism occurs via competition with PPAFURXRa for occupancy of the DNA target site on the HD PPRE.

The promiscuous binding of members of the COUP-TF recep tor family to diverse response elements composed of TGACCT

repeats, with its consequences on transcriptional induction me- diated by cognate receptors, has positioned COUP-TFs as cen- tral regulators of hormone-responsive networks. COUP-TFs have been shown to play an important role in the modulation of expression of apolipoprotein genes, whose products are in- volved in lipid homeostasis (34). Our demonstration that COUP-TF1 binds to a PPRE and antagonizes PPAR-mediated signaling implicates members of the COUP-TF family an play- ing a role in lipid homeostasis also through the modulation of peroxisome-mediated metabolism of fatty acids and in the cel- lular response to peroxisome proliferators and other xenobiotic compounds.

Acknowledgments-We thank Dr. M.J. %ai for the giRs of $OW- TF1 and anti-hCOUP-TF1 serum, Dr. F. M. Sladek for the gift of HNF4 cDNA, Dr. D. Noonan for the gift of rPPAR cDNA, and Dr. R. Evans for the gift. of RXRa cDNA. Ciprofibrate was a kind gift from Sterling Drug. We greatly appreciate the commenta of Dr. S. Subramani on the manu- script.

REFERENCES 1. Lazarow, P. B., and Fujiki, Y. (1985) Annu. Rev. CcN Bid . 1.489-630 2. Vamecq, J.. and Draye. J. P. (1989) Esaays Biahem. 24. 116225 3. Rao. M. S., and Reddy, J. K (1991) Enuiron. Health Perupect. 93,206209 4. Cattley. R. C., and Popp. J. A. (1989) Cancer Rea. 49.324&3251 5. Marsmann, D. S.. Cattley, R. C.. Conway, J. G.. and Popp. J. A. (1988) Cancer

6. Furuta, S., Miyazawa, S.. and Hashimoto. T. (1982) J. Biochem. (7bkyo) 92, Res. 48.6739-6744

7. €teddy. J. K, Coel, S. K, Nemali. M. R.. Camno. J. J.. Laffler. T. G.. Reddy. M. 319-326

K., Sperbeck. S. J., Osumi. T.. Hashimoto. T., Lalwani. N. D.. and Rao. M.

8. Sharma, R. K.. Lake, B. G.. Makowski. R.. Bradohaw. T.. Earnahan. D.. Dale. S. (1986) P m . Natl. Acari. Sci. U. S. A 83, 1747-1751

J. W.. and Gibson, G. G. (1988) Eur. J. Blochem. 184.69-78 9. Osumi, T., Wen, J.-K, and Hashimoto. T. (1991) Biochern. Biophys. h a . Com-

rnun. 175,866-871 10. lhgwood, J. D., I semam, I., Anderson, R C.. Bundell. K R. McPheaf W. L..

and Green, S. (1992) EMBO J . 11,433-439 11. Zhang, B.. Manus, S. L.. S a a d i . F. C.. Alvares. K. Reddy. J. K. Subramani.

S.. Rachubinski. R. A.. and Capone. J. P. (1992) P m . Notl. Acad. Sci.

12. Bardot. 0.. Aldridge. T. C.. LatrufTe, N.. and Green, S. (1993) Biochem. Bio-

13. Evans. R. (1988) Science 240,889-895 14. Umesono, K., Murakami, K. K. Thompeon. C. C.. and Evans. R. M. (1991)Cell

66, 1255-1266 15. Yu, V. C., Delsert. C.. Andersen. B., Holloray, J. M.. Devary. 0. V.. Nllllr. A M.,

Kim, S. Y.. Boutin, J.-M.. Glass. C. K. and Roemfeld. M. G. (1991) Cell 87.

16. h i d , M., Kastner. P., Lyons, R. Nalrshatri. H., Saundm. M., Zacharapski, T.. 1251-1263

Chen. J.-Y.. Staub, A. Gamier, J.-M.. Mader. S.. and Chambon. P. (1992) Cell as, 377-395

17. Kliewer, S. A. UmeMmo. K, MangeMod, D. J.. and Evans. R M. (1992) Nature 356,446-449

18. Imemann, I., and Green, 5. (1990) Nature 847.646-660 19. Dreyer, C., Krey, G.. Keller, H., Givel. F.. Helftenbein. G.. and Wahli. W. (1992)

20. COttlicher, M.. Widmark, E., Li. Q.. and Gustafeeon. J.-A (1992) P m . Natl. Cell as, 879487

21. Schmidt, A.. Endo. N.. Rutledge. S. J.. Vogel. R. S h i m . D.. and Rodan. G. A Acad. Sei. U. S. A. 88.46534657

22. Kliewer, S. A. Umesono. K.. Noonan. D. J.. Heyman, R. A. and Evans. R M. (1992) Mol. Endocrinol. 6. 1634-1641

23. Gearing, K. L., GOttlicher, M., Teboul, M., Widmark, E., and Custafsson. J.-k (1992) Nature 368, 771-774

24. Keller, H., Dmyer, C., Medin. J.. Mahf0udi.A.. Ozato. K, and Wahli. W. (1993) (1993) P m . Nafl. Acari. Sci. U. S. A. 90, 1440-1444

25. Zhang, B., Marcus. S. L.. Miyata. K S.. Subramani, S., Capone. J. P.. and P m . Natl. Acad. Set. U. S. A. 90,2160-2164

26. Marcus, S. L., Miyata. K S.. Zhang. B.. Subramani. S., Raehubiruki. R.A. and Rachubinski. R. A. (1993) J. BIOI. Chem. 268, 12939-12945

27. Cooney. A. J., n a i . S. Y., O'Malley. B. W.. and T~ai . M . J . (1992) Mol. Cell. Bid . Capone, J. P. (1993) P m . Null. A d . Sci. L! S. A. 90.5723-6727

28. Cwney.AJ.. Leng.X,Tmi. S.Y., O'Malley, B. W., andTbai. M . J . (1993)J. Bid. 12.4153463

29. Miysjima. N.. Kadowaki, Y.. Fulrushige. S.. Shimizu. S., Semb., K. Yma- Chern. 288,41524160

nashi. Y.. Mataubara. K, lbymhima. K. and Yamamato. T. (1988) Nueleic Acids Res. 16. 11057-11074

30. Dingam, J. D., Lebovitz. R. M., and Rocder, R. C. (1983)NucleicAridrh1.11, 1475-1489

31. Andrews, N. C.. and Faller, D. V. (1991) Nucleic Acidr &I. 19, 2499 32. Ladias. J. A. A. and Karathanasis. S. K (1991) Science 261, 661-666 33. Sladek. F. M., Zhong. W.. Lai, E.. and Darnell. J. E., Jr. (1990) Gena & Dnt. 4,

34. Ladias, J. A. A., Hadmpoulou-Cladarar, M., Kardassi. D., Cardof P.. Cheng.

U. S. A. 89.7541-7545

phys. Rea. Cornrnun. 192.3745

2353-2365

J.. Zannis, V., and Cladam. C. (1992) J. Biol. Chem. 267. 1581%1Mlgo