Proc. Natl. Acad. Sci. USA Vol. 92, pp. 4432-4436, May 1995

Biochemistry

Dual transcriptional control by Ear3/COUP: Negative regulation through the DR1 direct repeat and positive regulation through a sequence downstream of the transcriptional start site of the mouse mammary tumor virus promoter
YASUNORI KADOWAKI*, KUMAO TOYOSHIMAtS, AND TADASHI YAMAMOTO* *Department of Oncology, The Institute of Medical Science, University of Tokyo 4-6-1, Shirokanedai, Minato-ku, Tokyo 108, Japan; and tDepartment of Oncogene Research, Research Institute for Microbial Diseases, Osaka University, Suita, Osaka 565, Japan Communicated by Keith R. Yamamoto, University of California, San Francisco, CA, January 3, 1995 (received for review April 15, 1994)
also binds with a similar affinity to the palindromic thyroid hormone response element (TRE). This binding preference of Ear3/COUP is same as that of the retinoid X receptor (RXR), which is another member of the superfamily. In the present study, we identified a sequence responsible for Ear3/COUPmediated transactivation in the region downstream of the transcription start site of the mouse mammary tumor virus promoter. This cis-acting sequence was unresponsive to RXR. When the DR1 or TRE sequence was added upstream of the promoter, transactivation by Ear3/COUP was completely abolished, whereas RXR enhanced transcription from the promoter. The mode of action of Ear3/COUP could be utilized to control complex gene expressions in morphogenesis, homeostasis, and development.
Ear3/COUP is an orphan member of the steroid/thyroid hormone receptor superfamily of transcription factors and binds most tightly to a direct repeat of AGGTCA with 1 nucleotide in between (DR1). Ear3/COUP

ABSTRACT

response element (TRE) (see Fig. la). In contrast, Ear3/COUP bound weakly to an estrogen response element (ERE) and hardly bound to a glucocorticoid response element (GRE). This binding preference of Ear3/COUP was almost the same as that of the retinoid X receptor (RXR) (8), a member of the STR family that responds to retinoic acid (15). Thus, Ear3/COUP and RXR are thought to be involved in regulation of the same target genes. We have examined the transcriptional effects of Ear3/ COUP on these HREs. During the course of the study, we identified a DNA sequence to which Ear3/COUP bound with high affinity, in the region downstream of the transcriptional start site of the mouse mammary tumor virus (MTV) promoter. This binding resulted in transcriptional activation. The modes of action of Ear3/COUP on upstream DR1 and on this MTV sequence were compared.

hormone

MATERIALS AND METHODS
Plasmids. Plasmid pRS-hRXRa (15) and plasmids AMTVCAT and MTV-CAT (16) were gifts from R. M. Evans (Salk Institute, San Diego). AMTV-ERE-CAT, AMTV-TRE-CAT, and AMTV-DR1-CAT were constructed by introducing synthetic oligonucleotides encoding ERE, TRE, and DR1, respectively (9), into the unique HindIIl site of AMTV-CAT. The reporter construct pBLCAT2 (17), in which the herpes simplex viral thymidine kinase (TK) promoter is linked to the chloramphenicol acetyltransferase (CAT) reporter gene, is denoted as TK-CAT in the text. For construction of TK-DR1-CAT and TK-2xDR1-CAT, one or two copies of the DR1 oligonucleotide were cloned into the HindIII site of TK-CAT. For construction of deletion mutants of MTV-CAT, unique restriction sites that flank the major cis-acting elements (Bgl II site at -61; BamHI site at -45) were created by oligonucleotide-directed mutagenesis. Each mutant plasmid was cleaved at both the newly created restriction site and the unique Sal I site of the vector. The resulting plasmids were used to generate the deletion mutants MTV-OCT-CAT and MTV-TATACAT. MTV-NF-OCT-CAT was generated by removing the Sac I-Sal I fragment from MTV-CAT. For construction of M-MCAT, a Pst I site was created at + 10 of MTV-TATA-CAT by in vitro mutagenesis. The Pst I-Xho I fragment of TK-CAT was replaced by the Pst I-Xho I fragment of M-M-CAT to generate T-M-CAT. The Sal I-Pst I fragment of TK-CAT was replaced by the Sal I-Pst I fragment of M-M-CAT to generate M-TAbbreviations: STR,
Steroids, retinoids, and thyroid hormones play crucial roles in vertebrate development and homeostasis. The receptors for these hormones constitute a large family of ligand-dependent transcription factors, the steroid/thyroid hormone receptor superfamily (STR family), which mediate activation of gene transcription by interacting with specific cis-acting elements called hormone response elements (HREs) (1-4). HREs of the STR family are composed of tandem or inverted repeats of a hexameric half-site (AGGTCA or AGAACA), and their consensus sequences are classified into nine types (5-9). Differences in the orientations and spacings of these half-sites within HREs dictate selective transcriptional responses to members of the STR family. Usually, HREs are located upstream of promoters, and the efficiencies of their binding to receptors are parallel to those of their transactivation efficiencies (7). The ear3 gene, obtained by cross-hybridization to the v-erbA sequence, encodes a protein that belongs to the STR family as judged by its structural characteristics (10). The ear3 gene product is identical to COUP (11), a transcription factor originally characterized by its interaction with the so-called COUP sequence which is a part of the chicken ovalbumin gene promoter (12). A remarkable feature of the ear3 gene product (Ear3/COUP) is that its amino acid sequence is highly conserved (>90% identity) in Drosophila (13) and sea urchin (14), suggesting its important role throughout evolution. In previous studies (8, 9), the binding specificities of Ear3/ COUP to nine HREs were assessed by gel mobility-shift experiments. Ear3/COUP exhibited a strong preference for binding to the DR1 direct repeat sequence and to a thyroid

publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. 1734 solely to indicate this fact.
steroid/thyroid hormone receptor; HRE, horresponse element; ERE, estrogen response element; GRE, glucocorticoid response element; TRE, thyroid hormone response element; RXR, retinoid X receptor; MTV, mammary tumor virus; CAT, chloramphenicol acetyltransferase; TK, thymidine kinase. tPresent address: The Center for Adult Diseases, 1-3, Nakamichi, Higashinan-ku Osaka 537, Japan. To whom reprint requests should be addressed.
Biochemistry: Kadowaki et al.
Proc. Natl. Acad. Sci. USA 92 (1995)

AGGTCANAGGTCA

TRE AGGTCATGACCT ERE AGGTCANNNTLJACCT GRE AGAACANNNTGTTCT..

TK-CAT

'NFlsite oOCTsite VTATAbox

MTV-CAT

TK-DR1-CAT i; TK-2xDR1-CATIDR IDRI

AMTV-CAT

.100 b
AMTV-ERE-CAT AMTV-TRE-CAT AMTV-DR -CAT

ERE TRE

Ox- lx- 2xDRI DRI DRI

-xOx- IxDR1 DR

x2xDRI DR

A ERE TRE DR1

A ERETREDRI
FIG. 1. Positive and negative transcriptional regulation of the AMTV promoter by Ear3/COUP. (a) Structures of HREs examined in this study. The hexameric half-sites and their relative orientations are indicated by arrows. The affinities of Ear3/COUP and RXR for each HRE are from previous reports (8, 9, 15) and are indicated as + + + (high), + (low), or - (no affinity). (b) Schematic representation of reporter plasmids. The authentic GREs (open bars) of MTV-CAT were replaced by a HindIII site to generate AMTV-CAT (19). A single copy of ERE, TRE, or DR1 was cloned into the unique HindIII site of AMTV-CAT to generate AMTV-ERE-CAT, AMTV-TRE-CAT, or AMTV-DR1-CAT, respectively. (c) Effects of Ear3/COUP on the AMTV promoter. Each of the reporter constructs described in b was cotransfected into CV-1 cells with the expression plasmid pSV-Ear3/ COUP (open bars) or expression vector pSV2H with no insert (black bars). CAT activities are shown as percentages of the Ear3/COUPinduced activity obtained with AMTV-CAT. (d) Effects of RXR on the AMTV promoter. CV-1 cells were cotransfected with the same set of reporter constructs used in c, together with expression plasmid pRS-hRXRa, in the presence (open bars) or absence (gray bars) of 10 I,M retinoic acid. CAT activity is shown as a percentage of the Ear3/COUP-induced activity obtained with AMTV-CAT in c.
FIG. 2. Effects of Ear3/COUP on the TK promoter. (a) Schematic representation of the reporter plasmid TK-CAT and its derivatives. One or two copies of DR1 were cloned into the unique HindIII site in the multiple cloning region of TK-CAT, to generate TK-DR1-CAT and TK-2xDR1-CAT. (b) Transcriptional effects of Ear3/COUP. CV-1 cells were cotransfected with one of the reporter constructs described above, together with expression plasmid pSV-Ear3/COUP (open bars) or expression vector pSV2H with no insert (black bars). The RXR-induced activity obtained with TK-2xDR1-CAT in c is arbitrarily set at 100%. (c) Transactivation by RXR. CV-1 cells were cotransfected with the same set of reporter constructs used in b, together with expression plasmid pRS-hRXRa in the presence (open bars) or absence (gray bars) of 10 ,uM retinoic acid.

CAT. Mutants of MTV-TATA-CAT carrying a deletion in the 3' sequence were obtained by digestion with exonucleases III and VII from the Pst I site in the 5'-to-3' direction. A mutant of MTV-TATA-CAT lacking the entire 3' sequence was obtained by removing the Pst I-Xho I fragment of M-T-CAT. For construction of pSV-Ear3/COUP, the HindIII fragment of pSRa-Ear3/COUP (9) was excised by digestion with HindIII and ligated into the unique HindIII site of pSV-H, a derivative of pSV2neo (18). CAT Assay and Gel Shift Assay. The CAT assay was performed as described (16). Precipitates used for transfection typically contained 5 ,ug of expression plasmid, 2.5 ,tg of reporter plasmid, 0.3 A,g of 3-galactosidase reporter plasmid driven by the promoter of the /-actin gene (p/3act-3gal), and 7.5 ,tg of pUC119 as a carrier. Lysate samples containing equal amounts of ,3-galactosidase activity were assayed for CAT activity. The gel shift assay was performed exactly as described (9). Complexes were resolved by electrophoresis in a 5%

polyacrylamide gel.

RESULTS
Differential Effects of Ear3/COUP and RXR on a Modified MMTV Promoter. An Ear3/COUP expression plasmid was
cotransfected into CV-1 monkey kidney cells together with CAT reporter plasmids containing various HREs. In the reporter construct, the CAT gene was linked to a derivative of mouse MTV promoter that is devoid of its own enhancer element GRE (AMTV-CAT) (Fig. lb) (19). This plasmid has been used as a basal promoter to test the activities of various HREs (16, 19-22). A single copy of the HRE (DR1, TRE, or ERE) was inserted into the unique HindIII site of the plasmid, generating the reporter plasmids AMTV-DR1-CAT, AMTVTRE-CAT, and AMTV-ERE-CAT. To our surprise, Ear3/COUP stimulated efficient transcription from the "empty" AMTV promoter (Fig. ic). When an ERE to which Ear3/COUP binds with low affinity was present (AMTV-ERE-CAT), stimulation of transcription by Ear3/ COUP was reduced to half the level observed with AMTVCAT (Fig. lc). Moreover, Ear3/COUP did not stimulate transcription of AMTV-DR1-CAT or AMTV-TRE-CAT, both of which contain high-affinity binding sequences for Ear3/ COUP (Fig. lc). Thus, Ear3/COUP is an unusual transcriptional regulator: it transactivates the AMTV promoter by an unknown mechanism and self-suppresses or diminishes the transactivation when the apparent binding element is present in the context of the promoter. Since RXR has the same binding sequence as Ear3/COUP, we examined the mode of action of RXR on the AMTV-based promoters. The above same sets of reporter constructs and an expression plasmid encoding human RXRa (pRS-hRXRa) were cotransfected into CV-1 cells that were grown in the presence or absence of retinoic acid. RXR could not stimulate transcription from AMTV-CAT, but expression from reporter plasmids containing high-affinity binding sites (AMTV-DR1CAT and AMTV-TRE-CAT) was efficiently induced by RXR in the presence of retinoic acid (Fig. ld). A slight increase in CAT activity was seen when AMTV-ERE-CAT was employed

Biochemistry: Kadowaki et al
Proc. Natl. Acad ScL USA 92 (1995)

AMTV CAT MTV CAT

------

i HiinII1

MTV-NF-OC('T-CAT MTV-OCT-CAT MTV-TATA-CAT MTV-NF-CAT
MOCK Ear3/COUP. 1 'rEar3,/COL P

' -ssss------ -.

CCATAAiATAAAA

MTV-TATA-CAT

(M-M-CAT)

T-T-CAT -.- ^.^.^:

P TATAst ,.r". site ^..-\\

\\\\\co

competitor (-) (-)

_ -ATI IM/C_

M-T-CAT

--J-T-M-CtAT - - - tX

T-T-CAT
FIG. 3. Localization of the region responsible for Ear3/COUP-mediated transactivation in the MTV promoter. (a) Transcriptional activation by a series of deletion mutants of the MTV promoter constructed from the parental plasmid MTV-CAT. The deletion constructs were cotransfected with pSV-Ear3/COUP (open bars) or expression vector pSV2H with no insert (black bars) into CV-1 cells. Ear3/COUP-induced activity with MTV-NF-OCT-CAT is arbitrarily set at 100%. (b) Chimeric analysis of MTV-TATA-CAT and TK-CAT. A naturally occurring Pst I site located 13 bp downstream of the transcription start site of the TK promoter and a Pst I site created 10 bp downstream of the transcription start site of MTV-TATA-CAT by site-directed mutagenesis were utilized to divide the promoter into two pieces. Exchange of the corresponding fragment of each plasmid resulted in two hybrid constructs, M-T-CAT and T-M-CAT. These reporter plasmids were cotransfected into CV-1 cells, together with expression plasmid pSV-Ear3/COUP (open bars) or expression vector pSV2H with no insert (black bars). The CAT activities of M-M-CAT and M-T-CAT are expressed as percentages of the Ear3/COUP-induced activity with M-M-CAT. The CAT activities of T-T-CAT and T-M-CAT are expressed as percentages of the Ear3/COUP-induced activity with T-M-CAT. (c) Gel shift analysis. Whole-cell extracts were prepared from COS cells transfected with the Ear3/COUP expression plasmid (Ear3/COUP) or with an expression vector with no insert (MOCK). Then the lysates were incubated with a 32P-labeled DNA probe which corresponded to the entire 125-bp 3' sequence. The reaction mixture was separated by polyacrylamide gel electrophoresis. Unlabeled competitors (DR1, ERE, GRE, or the 3' sequence itself) were added in 10-fold molar excess over the labeled.-probe. as a reporter plasmid (Fig. ld). Thus, the ability of RXR to transcription start site of the MTV promoter (3' sequence) was

cause transactivation through a HRE parallels closely its ability to bind the HRE. This property has also been observed in other members of the STR family (20, 21). Ear3/COUP Is Silent in the Reporter Construct Based on the TK Promoter. To address whether Ear3/COUP acts similarly on promoters other than the AMTV promoter, we chose the TK promoter (17) containing one or two copies of the DR1 element (Fig. 2a). In contrast to the results with the
of the basal TK promoter (Fig. 2b). Furthermore, no activation or repression of CAT activity was observed with reporter constructs containing high-affinity binding sites (TK-DR1CAT and TK-2xDR1-CAT) (Fig. 2b). On the other hand, the transcriptional activities of RXR on these reporters (Fig. 2c) were virtually identical to those with the AMTV promoter. These results indicated that positive and negative regulations were present for Ear3/COUP and that the regulation depended on the type of promoter. Localization of an Ear3/COUP-Responsive Sequence Downstream of the TATA Box. To localize the region in the AMTV promoter responsible for the positive transcriptional effect of Ear3/COUP, a series of deletion mutations were introduced into the parental reporter plasmid MTV-CAT (Fig. 3a). Ear3/COUP-dependent transactivation was maintained even when the promoter sequence was deleted to a site 6 bp upstream of the TATA box (Fig. 3a, MTV-TATA-CAT). Thus, we concluded that the sequence responsible for Ear3/COUPmediated transactivation lay near the TATA box or further downstream. To identify its location more precisely, we constructed chimeras between the MTV-TATA promoter and TK promoter (Fig. 3b). Upon cotransfection of these reporter constructs and the Ear3/COUP expression plasmid, Ear3/ COUP failed to activate transcription from M-T-CAT but efficiently activated transcription from T-M-CAT (Fig. 3b). These findings suggested that the sequence downstream of the

AMTV promoter,

Ear3/COUP did not activate transcription
responsible for transactivation by Ear3/COUP. To evaluate whether the observed functional effects involved binding of Ear3/COUP to the 3' sequence, we performed an electrophoretic mobility-shift assay using extracts prepared from COS cells transiently transfected with an Ear3/COUP expression plasmid (Ear3/COUP extract). A shifted band was observed with the Ear3/COUP extract but not with the extract from mock transfectants, suggesting that the Ear3/COUP protein could form a complex with the 3' sequence (Fig. 3c). Next, we investigated the relative affinity of Ear3/COUP for the 3' sequence and the other HREs. Unlabeled DNA probes corresponding to the 3' sequence and three representative HREs (GRE, ERE, and DR1) were employed as competitors to compete with the radiolabeled DNA probe for binding to Ear3/COUP. Complex formation between the 3' sequence and Ear3/COUP was eliminated in the presence of a 10-fold molar excess of the competitors containing either DR1 sequence or the 3' sequence itself (Fig. 3c). In contrast, the same concentration of ERE or GRE competitor could not quench the complex formation (Fig. 3c). These data suggested that Ear3/COUP bound to the 3' sequence with high affinity, comparable to the affinity of Ear3/COUP for the DR1 sequence. Mapping of the Ear3/COUP Response Sequence in the 3' Region. To confine the region essential for transactivation by Ear3/COUP, deletions were introduced in the 3' sequence of the reporter plasmid MTV-TATA-CAT. The resultant reporter constructs (Fig. 4a) were used for cotransfection assay as above. Deletions of the first 81 bp from the 5' end had no effect on Ear3/COUP-induced transactivation; but further deletion eliminated the transactivating function. These data suggested that the responsible element was localized between +91 and +135. A series of subfragments of the 125-bp 3' sequence (Fig. 4b; fragments A-G) were tested for their ability to bind Ear3/

Biochemistry: Kadowaki et ail
Proc. Natl. Acad. Sci. USA 92

(1995)

.d-*,r.

7='-,---r

iL ii-

Im'' iiin

4~~~~~~~~l1-

lU -iL+

J w-^ 400~~~~~~~~El iil

Ear3/COUP L _*

Ear3/COUP -->-

Ear3/COUP

competitor (-)(-) A

B C D E F G

vector without an insert
FIG. 5. Regulatory effects of RXR (a and c), Ear3/COUP (b and d), and combination of RXR and Ear3/COUP (e) on TK-type (a, b, and e) and MTV-type (c and d) promoters. In a, c, and e, bars indicate activity without (-) and with (+) ligand. In b and d, bars indicate
(-) vs. vector with an insert (+).

i( (.(.(((( 'A(;A''

('('( A

*;( '(

( 1('A diT
FIG. 4. Confinement of the putative sequence responsible for transactivation by Ear3/COUP. (a) Effect of deletion of the 3' sequence. Plasmids containing the different deletions of the 3' sequence of-the reporter MTV-TATA-CAT were cotransfected into CV-1 cells with expression plasmid pSV-Ear3/COUP (open bars) or expression vector pSV2H with no insert (black bars). CAT activities are expressed as percentages of the Ear3/COUP-induced activity with the construct deleted of the region from +10 to +91. (b) Gel shift analysis. Lysates were prepared as in Fig. 3c. The 45-bp fragment C, corresponding to the sequence from +91 to +135 as shown in a, was 32P-labeled and used as a probe. The indicated competitors containing the sequence of each subfragment (A-G) were added in 20-fold molar excess over the labeled probe.
COUP. When the radiolabeled fragment, which corresponded region from +91 to +135 (see Fig. 4a), was used as a probe an intense band was seen with Ear3/COUP-transfected lysate (Fig. 4b). The formation of this binding complex was blocked specifically by the unlabeled oligonucleotides C and G, but not by the other oligonucleotides. These results clearly demonstrated that the sequence between positions +102 and +135 (fragment G) of the 3' region was essential for binding to the Ear3/COUP protein. Inspection of the sequence of fragment G revealed that the sequence from + 104 to + 120, 5'-CGGTCACCCTCAGGTCG-3', closely resembled a typical DR5 direct repeat (Fig. 4b). Previous studies (9, 23) showed that Ear3/COUP was able to bind tightly not only to the DR1 sequence but also to the other direct-repeat sequences, although the affinity was slightly lower. This was consistent with the observation that DR1 was a slightly better competitor than the 3' sequence itself (Fig. 3c). These data strongly suggested that the DR-like sequence was responsible for Ear3/COUP-mediated transactivation.

to the

DISCUSSION
The STR family members generally activate transcription upon binding to their corresponding HREs, which are located upstream of the transcriptional start site (7) (Fig. 5 a ii and c ii). The present study shows that Ear3/COUP has a unique transactivating function: Ear3/COUP stimulated transcription
through binding to a novel sequence and this transactivation was suppressed through its binding to DR1 present upstream of the same promoter sequence. (Fig. lc, Fig. 5d). There are reports of positive effects of Ear3/COUP in an in vitro transcription system (12) or with a chimera of the ligandbinding domain of Ear3/COUP and the DNA-binding domain of the progesterone receptor (24). Our present data clearly show that Ear3/COUP can activate transcription from the MTV promoter in cells. On the other hand, it was reported that Ear3/COUP did not affect transcription of promoters containing its binding element DR1 (23-25). This is consistent with our results with the TK promoter (Fig. 2b, Fig. 5b ii) but not with our results with the AMTV promoter (Fig. lc, Fig. 5d). Therefore, Ear3/COUP is unique among the members of the STR family in that it is bifunctional in regulatory function. Deletion analyses using AMTV promoters revealed that the responsible sequence was downstream of the transcription start site (3' sequence) (Fig. 3, Fig. 5d i). The responsible sequence was further narrowed down to 34 bp, spanning from +102 to +135, which included a DR-like element (Fig. 4). Ear3/COUP bound to this sequence with high affinity (Fig. 3c, Fig. 4b). To our knowledge, this is the first report to show that transactivation by an STR family member is mediated through a sequence downstream of the transcriptional start site. In addition, the 3' sequence of MTV promoter was specific to Ear3/COUP, since other members of the STR family did not affect transcription from the AMTV promoter (Fig. ld, Fig. 5c i) (16, 20-22). A question was whether the 3' sequence itself was a position-independent cis-acting enhancer sequence specific for Ear3/COUP or whether its location was rather important for Ear3/COUP function. Results of a recent experiment favor the latter possibility: Ear3/COUP failed to transactivate a reporter construct in which the entire 3' sequence was placed upstream of the TK promoter in either orientation (data not shown). The precise mechanism(s) by which Ear3/COUP transactivates from the AMTV promoter remains to be investigated. At present, we speculate that the mechanism(s) involves interactions between Ear3/COUP and the general transcription factors assembled in the transcriptional initiation complex. Consistent with this speculation, a report by Ing et at (26) has shown that Ear3/COUP interacts with one such factor, TFIIB. Ear3/COUP might have a unique interaction surface, distinct from those of other members of the STR family, including RXR. When Ear3/COUP is positioned at a site downstream of the transcriptional start, this surface might

Proc. NatL. Acad. Sci. USA 92 (1995)
We thank R. M. Evans for gifts of AMTV-CAT, MTV-CAT, and RS-hRXRa. We thank J. Inoue, S. Kato, and K. Umesono for valuable suggestions. We thank H. Umemori, M. Sudol, and A. Tanaka for critical reading of the manuscript. Y.K. was supported by postdoctoral fellowships in cancer research from the Japan Society for the Promotion of Science.

1. 2. 3. 4. 5. 6.

interact with the factors in a way which favors transcriptional initiation. When Ear3/COUP is positioned upstream of the AMTV promoter, it may interact with the factors in a different manner, resulting in no activation of the promoter. The molecular basis of the Ear3/COUP binding to the 3' sequence is not known. A previous report showed that the RXR-Ear3/ COUP heterodimer could interact with DR1 (8). Similarly, the RXR-Ear3/COUP heterodimer or even the RXR homodimer could bind to the DR-like element in the 3' sequence. Nevertheless, the surface of RXR would be unable to interact with the transcription factors in a way to evoke transcriptional activation. A typical example of negative regulation by Ear3/COUP is suppression of RXR-mediated transactivation of DR1containing promoters (8, 23, 25) (Fig. 5e). This suppression is thought to be due to competitive DNA binding by the Ear3/ COUP and RXR. The negative regulation by Ear3/COUP shown in the present study (Fig. 5d ii) is distinct from the Ear3/COUP-mediated suppression of RXR activity (Fig. 5e), in that the interaction of Ear3/COUP with DR1 suppresses Ear3/COUP-mediated transcriptional activation that is mediated through the sequence downstream of the transcription initiation site (Fig. 5d i). Because Ear3/COUP did not downregulate basal transcription from TK-DR1-CAT (Fig. 2b, Fig. 5b ii) (8, 23, 25), we conclude that Ear3/COUP is not simply a DR1-dependent transcriptional repressor. Thus, our observation implicates a novel mechanism of transcriptional suppression by Ear3/COUP. Binding of Ear3/COUP to the upstream DR1 sequence may deteriorate transcription systems required for Ear3/COUP-dependent activation through the 3' element and RXR-dependent activation through the upstream DR1 sequence. No specific ligand for Ear3/COUP has been identified. The present study demonstrates that Ear3/COUP can function as a transactivator without exogenously added ligand (Fig. lc, Fig. 5d i), suggesting that calf serum contains a sufficient concentration of the ligand. In our previous report (9), we speculated that Ear3/COUP and RXR might have almost the same functions, since they share the same binding elements. However, the mirror image of transcriptional activity presented in Fig. 1 c and d implicates entirely the opposite effect of the two receptors. Interestingly, a report by Cooney et al (23) suggested that Ear3/COUP downregulated the activity of not only RXR but also several other members of STR family, including the vitamin D3 receptor and the thyroid hormone receptor. On the other hand, other reports showed that RXR could upregulate the activity of the same members of the STR family, including the vitamin D3 and thyroid hormone receptors (27). Ear3/COUP and RXR may function as key molecules in the network of gene regulation by the STR family: each function is directed in the opposite direction.

11. 12.

13. 14. 15. 16. 17. 18. 19. 20.

21. 22.

23. 24. 25. 26.
Yamamoto, K. R. (1985) Annu. Rev. Genet. 19, 209-252. Evans, R. M. (1988) Science 240, 879-895. Green, S. & Chambon, P. (1988) Trends Genet. 4, 309-314. Beato, M. (1989) Cell 56, 335-344. Klock, G., Strahle, U. & Schutz, G. (1987) Nature (London) 329, 734-736. Glass, C. K., Holloway, J. M., Devary, O. V. & Rosenfeld, M. G. (1988) Cell 54, 313-323. Umesono, K., Murakami, K. K., Thompson, C. C. & Evans, R. M. (1991) Cell 65, 1255-1266. Kliewer, S. A., Umesono, K., Heyman, R. A., Mangelsdorf, D. J., Dyck, J. A. & Evans, R. M. (1992) Proc. Natl. Acad. Sci. USA 89, 1448-1452. Kadowaki, Y., Toyoshima, K. & Yamamoto, T. (1992) Biochem. Biophys. Res. Commun. 183, 492-498. Miyajima, N., Kadowaki, Y., Fukushige, S., Shimizu, S., Semba, K., Yamanashi, Y., Matsubara, K., Toyoshima, K. & Yamamoto, T. (1988) Nucleic Acids Res. 183, 492-498. Wang, L., Tsai, S. Y., Cook, R. G., Beattie, W. G., Tsai, M. & O'Malley, B. W. (1989) Nature (London) 340, 163-166. Tsai, S. Y., Sagami, I., Wang, H., Tsai, M. & O'Malley, B. W. (1987) Cell 50, 701-709. Mlodzik, M., Hiromi, Y., Weber, U., Goodman, C. S. & Rubin, G. M. (1990) Cell 60, 211-224. Chan, S., Xu, N., Niemeyer, C. C., Bone, J. R. & Flytzanis, C. N. (1992) Proc. Natl. Acad. Sci. USA 89, 10568-10572. Mangelsdorf, D. J., Ong, E. S., Dyck, J. A. & Evans, R. M. (1990) Nature (London) 345, 224-229. Umesono, K., Giguere, V., Glass, C. K., Rosenfeld, M. G. & Evans, R. M. (1988) Nature (London) 336, 262-265. Luckow, B. & Schutz, G. (1987) Nucleic Acids Res. 15, 5490. Akiyama, T., Matsuda, S., Namba, Y., Saito, T., Toyoshima, K. & Yamamoto, T. (1991) Mol. Cell. Biol. 11, 833-842. Hollenberg, S. M. & Evans, R. M. (1988) Cell 55, 899-906. Thompson, C. C. & Evans, R. M. (1989) Proc. Natl. Acad. Sci. USA 86, 3494-3498. Umesono, K. & Evans, R. M. (1989) Cell 57, 1139-1146. Yao, T., Segraves, W. A., Oro, A. E., McKeown, M. & Evans, R. M. (1992) Cell 71, 63-72. Cooney, A. J., Tsai, S. Y., O'Malley, B. W. & Tsai, M.-J. (1992) Mol. Cell. Biol. 12, 4153-4163. Power, R. F., Lydon, J. P., Conneely, O. M. & O'Malley, B. W. (1991) Science 252, 1546-1548. Mietus-Snyder, M., Sladek, F. M., Ginsburg, G. S., Kuo, C. F., Ladias, J. A. A., Darnell, J. E., Jr., & Karathanasis, S. K. (1992) Mol. Cell. Biol. 12, 1708-1718. Ing, N. H., Beekman, J. M., Tsai, S. Y., Tsai, M.-J. & O'Malley, B. (1992) J. Biol. Chem. 267, 17617-17623. Yu, V. C., Delsert, C., Andersen, B., Holloway, J. M., Devary, O. V., Naar, A. M., Kim, S. Y., Boutin, J., Glass, C. K. & Rosenfeld, C. K. (1991) Cell 67, 1251-1266.