A p56(lck) ligand serves as a coactivator of an orphan nuclear hormone receptor.

Chicken ovalbumin upstream promoter transcription factor II (COUP-TFII), an orphan member of the nuclear hormone receptor superfamily, acts as a transcriptional repressor by antagonizing the functions of other nuclear hormone receptors and by actively silencing transcription. However, in certain contexts, COUP-TFII stimulates transcription directly. A cellular factor, isolated by interaction cloning, bound COUP-TFII in vitro and allowed COUP-TFII to function as a transcriptional activator in mammalian cells. This factor is identical to a recently described ligand of the tyrosine kinase signaling molecule p56lck, suggesting that it mediates cross-talk between mitogenic and nuclear hormone receptor signal transduction pathways.

The nuclear hormone receptor superfamily comprises a large group of ligand-activated transcription factors important for the normal development and functioning of an organism. These receptors mediate transcriptional responses to steroids, retinoids, vitamin D, thyroid hormone, and fatty acids/peroxisome proliferators by binding as homodimers or heterodimers to characteristic DNA hormone response elements in target genes (1,2). The largest subgroup within this superfamily is the orphan receptors, so called because their cognate ligands have not been identified or may not exist (3). Accordingly, the mechanisms of action and physiological roles of orphan receptors remain poorly characterized.
Among the most studied of the orphan receptors is the chicken ovalbumin upstream promoter transcription factor (COUP-TF) 1 subfamily, which includes COUP-TFI (Ear3) and COUP-TFII (Arp1) (4). COUP-TFs are spatially and temporally regulated transcription factors implicated in many fundamental biological processes, including myogenesis, neurogenesis, organogenesis, determination of cell fate, and metabolic homeostasis. The importance of the COUP-TFs is underscored by their evolutionary conservation and by the observation that ablation of either COUP-TFI or COUP-TFII in the mouse is lethal (5).
COUP-TFs act principally as repressors of ligand-mediated hormone receptor signaling pathways via both protein-protein and protein-DNA interactions (5)(6)(7)(8). For example, COUP-TFs bind promiscuously to hormone response elements recognized by other nuclear receptors, thereby competing with them for their target sites. COUP-TFs can also titrate the common heterodimerization partner, retinoid X receptor (RXR), that is required for high affinity DNA binding of most members of the thyroid hormone/retinoic acid receptor subfamily (9 -11). In addition to these indirect mechanisms of repression, COUP-TFs can also actively silence basal and activated transcription (5), likely through direct interactions with TFIIB or other general transcription factors (12). COUP-TFs therefore antagonize cellular responses to multiple hormone signaling pathways and can have profound effects on numerous biological processes.
Paradoxically, both COUP-TFI and COUP-TFII can activate transcription in certain cell types and promoter contexts (13)(14)(15)(16)(17)(18). Moreover, COUP-TFII has been shown to function as a constitutive transactivator in vitro (19) and in yeast (20), suggesting that it possesses intrinsic activation potential. The mechanisms underlying this functional duality are unknown. Evidence suggests that the function of many nuclear hormone receptors is dependent upon, or modulated by, the actions of an increasing number of both common and distinct receptor binding cofactors that differentially recognize liganded and unliganded receptors (21)(22)(23)(24)(25)(26)(27)(28). Most of the auxiliary factors so far identified act as corepressors or negative modulators of receptor function (e.g. NCoR, SMRT, TRUP, and TRIP1). However, in a few cases, receptor-selective, positively acting coactivators (e.g. RIP140, SRC-1, and CBP/p300) have been identified. To determine if differential COUP-TF activity is mediated through the actions of auxiliary proteins, we used the yeast two-hybrid interaction cloning system to identify novel COUP-TFII interacting proteins. We identified a factor that bound COUP-TFII in vitro and allowed COUP-TFII to act as a transcriptional activator in mammalian cells. This factor is a recently reported ligand of the tyrosine kinase signaling molecule p56 lck . Our results suggest that this factor mediates cross-talk between mitogenic and nuclear hormone receptor signal transduction pathways.

MATERIALS AND METHODS
Two-hybrid Library Screening-Two-hybrid analysis was carried out using the Matchmaker System (Clontech) as described (29). pGBD-COUP-TFII contains the full-length cDNA for rat COUP-TFII (20) in the Gal4 DNA binding domain expression plasmid pGBT9 (Clontech). This bait plasmid was transformed into yeast HF7c cells along with a human liver cDNA library constructed in the Gal4 activation domain vector pGAD10 (Clontech). Transformants (1 ϫ 10 6 ) were plated onto synthetic complete plates lacking histidine, leucine, and tryptophan, and His ϩ /Leu ϩ /Trp ϩ colonies were recovered and assayed for ␤-galactosidase activity using filter assays. Library plasmids were rescued by * This work was supported by grants from the Heart and Stroke Foundation of Canada and the Medical Research Council of Canada (MRC). 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 1 The abbreviations used are: COUP-TF, chicken ovalbumin upstream promoter transcription factor; DBD, DNA binding domain; MBP, maltose binding protein; ORCA, orphan receptor coactivator; PPAR, peroxisome proliferator-activated receptor; PPRE, peroxisome proliferator-response element; RXR, retinoid X receptor.
electroporation into Escherichia coli, retransformed into yeast strain SFY526, and tested for specificity against pGBD-COUP-TFII, pGBT9, and several irrelevant Gal4 DNA binding domain fusion expression plasmids. Of the positive clones recovered, six independent isolates contained the same 2.1-kilobase pair insert based on restriction enzyme analysis. Two of these clones were sequenced and shown to encode a 440-amino acid long protein, which we refer to as ORCA (orphan receptor coactivator). The predicted amino acid sequence of ORCA is identical to the published sequence of the p56 lck -interacting protein p62 (30) (GenBank accession no. U46751).
Protein Binding Assays-The COUP-TFII cDNA was cloned as an EcoRI fragment into the EcoRI site of pMal-c2 (New England Biolabs), and the maltose binding protein (MBP) chimera was purified from induced cultures of E. coli according to the manufacturer's instructions. Control MBP was purified under identical conditions. Protein binding assays were carried out as described previously (29) using proteins synthesized in vitro with a coupled transcription/translation system (InVitrogen). Full-length ORCA cDNA was cloned into the EcoRI site of pSG5 (Stratagene) to generate pORCA/SG5, which is suitable for in vitro and in vivo expression. pORCA ⌬258 -440 (numbers refer to amino acid residues) was constructed by inserting a double-stranded oligonucleotide (5Ј-GCGTAATTAATTAATTACGC) containing termination codons in all three reading frames into the blunt-ended ClaI site of pORCA/SG5. pORCA ⌬128 -163 was constructed by site-directed deletion mutagenesis using 5Ј-GTGCACCCCAATGTGATCACCAAGCTCG-CATTCCCC and single-stranded DNA prepared from pORCA/SG5. Mutagenesis was carried out following standard procedures (31), and accuracy was confirmed by DNA sequencing in each case.
Transient Transfections and Measurement of Luciferase Activity-The luciferase reporter plasmid pHD(X3)luc, containing the rat hydratase-dehydrogenase PPRE, and effector plasmids expressing fulllength cDNAs for rat PPAR␣, human RXR␣, human COUP-TFI, and rat COUP-TFII nuclear receptors have been described (20,(32)(33)(34). pORCA/ SG5 is described above. BSC40 cells (10-cm subconfluent dishes) were transfected, and luciferase activity was measured as reported previously (32) using conditions described in the figure legends. Plasmid and promoter dosage was kept constant by addition of the appropriate amount of corresponding empty vector. A ␤-galactosidase reporter plasmid (pCH110, Pharmacia Biotech Inc.) was included as an internal transfection standard. Where indicated, the peroxisome proliferator Wy-14,643 was added to a final concentration of 0.1 mM from a 100 ϫ concentrated stock solution in dimethyl sulfoxide.

ORCA Interacts with COUP-TFII in Yeast and in Vitro-
Using rat COUP-TFII fused to the Gal4 DNA binding domain (Gal4 DBD) as bait to screen a human liver cDNA library fused to the Gal4 acidic activation domain, we isolated several clones that specifically interacted with COUP-TFII but not with the Gal4 DBD itself or with various control Gal4 DBD fusion proteins. Sequence analysis of the 2.1-kilobase pair insert of two of these clones showed that they encoded a predicted 440-amino acid protein unrelated to other known nuclear receptor-interacting factors. However, the deduced protein was identical to a recently described human phosphoprotein (p62) originally identified by its ability to interact with the SH2 domain of the tyrosine kinase signaling protooncogene molecule, p56 lck (30). We refer to this protein as ORCA.
To determine if ORCA bound directly to COUP-TFII, we made use of pull-down assays with in vitro synthesized ORCA (Fig. 1B, top panel) and a MBP-COUP-TFII fusion protein.
ORCA showed binding to MBP-COUP-TFII (Fig. 1B, middle  panel), with little or no binding to MBP itself (Fig. 1B, bottom   FIG. 1. ORCA binds to COUP-TFII in vitro and functions as a coactivator in vitro. Panel A, the full-length cDNA encoding ORCA was obtained from the yeast two-hybrid screen and encodes a 440-amino acid long protein identical to the p56 lck interacting protein, p62 (30). The positions of a putative Cys-finger that shares homology with the coactivators CBP and p300 and a Ser-rich domain are indicated. Derivatives lacking amino acid residues 128 -163 or lacking amino acid residues carboxyl to position 258 were constructed as described under "Materials and panel). A luciferase control did not bind to MBP-COUP-TFII. ORCA contains a cysteine-rich, zing finger-like motif (residues 128 -163), which could serve as a protein interaction motif, and a Ser-rich domain (downstream of residue 266) that may be a target for protein kinases (Fig. 1A) (30). To determine the importance of these regions in mediating interaction with COUP-TFII, the domains were altered by site-directed mutagenesis and tested for activity in protein interaction experiments. Derivatives truncated at amino acid 258 or missing residues 128 -163 interacted with COUP-TFII, indicating that these regions are not necessary for binding. Interestingly, the protein truncated at amino acid 258 bound to COUP-TFII much more avidly than wild-type ORCA. Therefore, interaction determinants are contained within the amino-terminal 258 amino acids of ORCA, but downstream elements appear to modulate its binding activity.
ORCA Is a Selective Coactivator of COUP-TFII in Mammalian Cells-To determine what effects ORCA might have on COUP-TFII activity in vivo, we carried out cotransfection experiments in mammalian cells with a luciferase reporter gene that contained the peroxisome proliferator-response element (PPRE) from the rat hydratase-dehydrogenase gene (33). COUP-TFII has been shown to bind avidly to this PPRE as a homodimer, but it has little effect on basal transcriptional activity of a linked reporter gene (Fig. 1C) (20). Cotransfection of ORCA on its own had no effect on basal level expression. However, coexpression of ORCA and COUP-TFII resulted in a 30 -40-fold induction in luciferase expression. The 128 -163 deletion and the carboxyl-terminal truncation derivatives also stimulated transactivation by COUP-TFII. Stimulation of COUP-TFII-mediated transactivation by ORCA was also observed with a reporter construct that contained a COUP-TFII binding response element from the rat ornithine transcarbamylase gene (35) (data not shown).
To examine the specificity of ORCA, we examined its effects on COUP-TFI, a highly related receptor that also binds strongly to the hydratase-dehydrogenase PPRE (32). COUP-TFI and COUP-TFII are nearly identical in their DNA binding and putative ligand binding domains but diverge in their respective amino termini. Transfections were carried out in parallel with COUP-TFI and COUP-TFII in the presence of various amounts of ORCA expression plasmid ( Fig. 2A). ORCA had a stimulatory effect on COUP-TFI activity, but the effect was much less pronounced than that observed with COUP-TFII and was seen only with relatively high concentrations of ORCA ( Fig. 2A).
To further investigate the in vivo selectivity of ORCA, we examined its effect on transactivation by the peroxisome proliferator-activated receptor (PPAR)/RXR heterodimer. ORCA had no effect on transcriptional activation mediated by PPAR/ RXR heterodimers, either in the absence or presence of the PPAR activator, Wy-14,643 (Fig. 2B). As we have previously shown (20), COUP-TFII antagonized transactivation mediated by PPAR/RXR and decreased Wy-14,643-dependent, PPAR/ RXR-mediated activation by 50 -60% (Fig. 2B). However, in the presence of coexpressed ORCA, repression by COUP-TFII was completely relieved. Indeed, transactivation by PPAR/RXR in the presence of both COUP-TFII and ORCA was approximately twice that observed with PPAR/RXR alone. This additive effect is expected if the transcriptional response is the combination of PPAR/RXR-and COUP-TFII/ORCA-mediated positive effects. Our findings indicate that ORCA is a selective coactivator of COUP-TFII and allows COUP-TFII to function as a positive transcriptional activator in mammalian cells.
The central role of nuclear hormone receptors in cell proliferation, differentiation, and development implies intuitively that their function must somehow be integrated with other signal transduction pathways important in the multifactorial regulation of these processes. Indeed, several lines of evidence indicate that COUP-TF activity is regulated by, or linked with, cell surface signaling pathways and second messenger activation. For example, the Drosophila COUP-TFII homologue seven-up, which is necessary for photoreceptor cell determination, requires an active ras signal transduction pathway for its activity (36,37). Moreover, a fusion between the COUP-TFI ligand binding domain and the progesterone receptor DNA binding domain was activated by the catecholamine neurotransmitter dopamine, suggesting that COUP-TFI can be regulated by cAMP-dependent protein kinase A phosphorylation cascades (38). Our finding that a p56 lck -interacting partner also serves as a COUP-TFII transcriptional coactivator suggests that this factor may link COUP-TFII and cell surface signal transduction pathways. This integrating role may be similar to what is observed with the cAMP response element binding protein CBP and the related protein p300, which functions as a coactivator of AP-1 and cAMP response element binding transcription factor families (39). CBP has recently been shown to be a constituent of a multicomponent coactivator complex that is necessary for activation of several ligand-dependent nuclear hormone receptors, including the retinoic acid and thyroid hormone receptors (40). Therefore, CBP family proteins play a role in integrating cAMP second messenger and nuclear hormone receptor signal transduction pathways. Interestingly, ORCA shares a small region of homology with CBP (residues 136 -154 of ORCA and residues 1715-1722 of CBP/ p300), suggesting a potential similarity in their mechanism of FIG. 2. ORCA is a selective coactivator. Panel A, transfections were carried out as in Fig. 1C with either COUP-TFI or COUP-TFII expression plasmids (0.5 g each) and increasing amounts of ORCA expression plasmid, as indicated. Luciferase activity was normalized to the value obtained with the reporter gene alone, which was taken as 1. The values shown are averages from three transfections carried out in duplicate. Values from individual transfections did not vary by more than 15%. Panel B, ORCA relieves COUP-TFII-mediated repression of transactivation by PPAR/RXR. BSC40 cells were transfected with 5 g of pHD(X3)luc reporter plasmid, along with expression plasmids for rat PPAR␣ and human RXR␣ (2 g each), COUP-TFII (0.5 g), and ORCA (4 g), as indicated. Wy-14,643 was added to a final concentration of 0.1 mM. Luciferase activity was measured as above and normalized to the value obtained with PPAR/RXR in the presence of Wy-14,643, which was taken as 100%. The values shown are averages from three transfections carried out in duplicate. Values from individual transfections did not vary by more than 15%.

action.
A putative role for ORCA in linking distinct signaling pathways remains to be established, since its function in cell surface signal transduction is not yet known. ORCA/p62 was originally isolated based on its interaction with the p56 lck , a T-cell-specific src family tyrosine kinase required for T-cell signal transduction. However, it is unlikely that ORCA/p62 function is restricted to p56 lck -mediated events, since ORCA/p62 is ubiquitously expressed with at least two known isoforms in humans (30). Moreover, a homolog of unknown function has been identified in mouse (GenBank accession no. U40930), and related proteins exist in Drosophila (41). This suggests that ORCA/p62 may be part of a larger family of factors that play a more general role in signal transduction in the cell. Indeed, p62 has been reported to also bind the Ras-GTPase activating protein (42) and a novel cytokine receptor induced in Epstein-Barr virus-infected B lymphocytes (43).
The mechanism by which ORCA enhances COUP-TFII activity is unknown at present, but several possible scenarios, which are not necessarily mutually exclusive, can be proposed. ORCA may function by binding directly to COUP-TFII to generate a DNA-bound multicomponent activating complex, similar to CBP and retinoic acid receptor (40). However, ORCA does not bind directly to COUP-TFII binding sites, and we have been unable to detect a COUP-TFII⅐ORCA supercomplex in gel retardation experiments, suggesting that if such a ternary complex forms, the COUP-TFII/ORCA/DNA interaction is weak or transient. Alternatively, ORCA may function directly or indirectly by phosphorylating COUP-TFII. This would be consistent with evidence implicating phosphorylation in activation of COUP-TF (38). It is interesting to note in this regard that ORCA/p62 is a phosphoprotein that possesses a tightly associated or intrinsic Ser/Thr protein kinase activity (42) and that both COUP-TFI and COUP-TFII contain a conserved consensus mitogen-activated protein kinase site (PX(S/T)P) in their amino-terminal domains. Finally, ORCA may override the function of a specific COUP-TFII-associated corepressor. This would be consistent with the observation that COUP-TFII constitutively activates transcription in yeast and in vitro, where presumably such a corepressor is not present or is limiting, respectively. In agreement with this possibility, expression of ORCA in yeast did not further potentiate transactivation mediated by COUP-TFII (data not shown).
In summary, we have identified a novel cellular factor that is known to interact with components of cell surface signal transduction pathways and that converts COUP-TFII from a transcriptional repressor into a transcriptional activator in mammalian cells. Our findings illustrate a novel mechanism by which an orphan nuclear hormone receptor can differentially regulate gene expression in an apparently ligand-independent manner. Moreover, our findings point to a role for ORCA and related factors in mediating cross-talk among distinct signal transduction pathways important for cellular growth and differentiation.