Function of Nuclear Co-repressor Protein on Thyroid Hormone Response Elements Is Regulated by the Receptor A/B Domain

Recently, a family of nuclear co-repressor proteins (TRACs) have been identified that interact with thyroid hormone (TR) and retinoic acid receptors to mediate ligand-independent repression of gene transcription. In this report, we have cloned and characterized a human TRAC, which when expressed as a truncated protein lacking its repressing domains, can abolish endogenous cellular TRAC activity. Use of this inhibitor has uncovered a differential function of TRACs on negative versus positive thyroid hormone response elements and has demonstrated the importance of the TR A/B domain in modulating TRAC function. Thus, isoform-specific functions of the TR may be mediated by their functional interaction with co-repressor proteins. Recently, a family of nuclear co-repressor proteins (TRACs) have been identified that interact with thyroid hormone (TR) and retinoic acid receptors to mediate ligand-independent repression of gene transcription. In this report, we have cloned and characterized a human TRAC, which when expressed as a truncated protein lacking its repressing domains, can abolish endogenous cellular TRAC activity. Use of this inhibitor has uncovered a differential function of TRACs on negative versus positive thyroid hormone response elements and has demonstrated the importance of the TR A/B domain in modulating TRAC function. Thus, isoform-specific functions of the TR may be mediated by their functional interaction with co-repressor proteins. INTRODUCTIONThe thyroid hormone receptor (TR), 1The abbreviations used are: TRthyroid hormone receptorTREthyroid response elementRXRretinoid X receptorT3triiodothyronineAAamino acid(s)Tracthyroid hormone and retinoic acid-receptor-associated co-repressors. a member of the steroid/thyroid hormone receptor superfamily, mediates gene regulation through its ability to bind to specific thyroid hormone response elements (TREs) (1Desvergne B. Petty K.J. Nikodem V.M. J. Biol. Chem. 1991; 266: 1008-1013Abstract Full Text PDF PubMed Google Scholar, 2Brent G.A. Harney J.W. Chen Y. Warne R.L. Moore D.D. Larsen P.R. Mol. Endocrinol. 1989; 3: 1996-2004Crossref PubMed Scopus (182) Google Scholar, 3Norman M.F. Lavin T.N. Baxter J.D. West B.L. J. Biol. Chem. 1989; 264: 12063-12073Abstract Full Text PDF PubMed Google Scholar, 4Umesono K. Murakami K.K. Thompson C.C. Evans R.M. Cell. 1991; 65: 1255-1266Abstract Full Text PDF PubMed Scopus (1488) Google Scholar) either as a monomer, homodimer, or as a heterodimer with RXR isoforms (5Yu V.C. Delsert C. Andersen B. Holloway J.M. Devary O.V. Naar A.M. Kim S.Y. Boutin J.M. Glass C.K. Rosenfeld M.G. Cell. 1991; 67: 1251-1266Abstract Full Text PDF PubMed Scopus (1051) Google Scholar, 6Heyman R.A. Mangelsdorf D.J. Dyck J.A. Stein R.B. Eichele G. Evans R.M. Thaller C. Cell. 1992; 68: 397-406Abstract Full Text PDF PubMed Scopus (1558) Google Scholar, 7Mangelsdorf D.J. Ong E.S. Dyck J.A. Evans R.M. Nature. 1990; 345: 224-229Crossref PubMed Scopus (1252) Google Scholar). Through its interactions with specific TREs and other protein co-activators, the TR can either positively or negatively regulate gene expression in the presence of its ligand, triiodothyronine (T3) (8Hollenberg A.N. Monden T. Flynn T.R. Boers M.-E. Cohen O. Wondisford F.E. Mol. Endocrinol. 1995; 9: 540-550Crossref PubMed Google Scholar, 9Saatcioglu F. Deng T. Karin M. Cell. 1993; 75: 1095-1105Abstract Full Text PDF PubMed Scopus (104) Google Scholar, 10Wondisford F.E. Farr E.A. Radovick S. Steinfelder H.J. Moates J.M. McClaskey J.H. Weintraub B.D. J. Biol. Chem. 1989; 264: 14601-14604Abstract Full Text PDF PubMed Google Scholar). The TR also has significant ligand-independent activity in both cell culture and in vitro transcription systems (11Casanova J. Helmer E. Selmi-Ruby S. Qi J.-S. Au-Fliegner M. Desai-Yajnik V. Koudinova N. Yarm F. Raaka B.M. Samuels H.H. Mol. Cell. Biol. 1994; 14: 5756-5765Crossref PubMed Scopus (122) Google Scholar, 12Tong G.-X. Jeyakumar M. Tanen M. Bagchi M.K. Mol. Cell. Biol. 1996; 16: 1909-1920Crossref PubMed Google Scholar). On positive TREs, the TR is a potent repressor of transcription in the absence of ligand (13Baniahmad A. Kohne A.C. Renkawitz R. EMBO J. 1992; 11: 1015-1023Crossref PubMed Scopus (238) Google Scholar, 14Sap J. Munoz A. Schmitt J. Stunnenberg H. Vennstrom B. Nature. 1989; 340: 242-244Crossref PubMed Scopus (295) Google Scholar, 15Damm K. Evans R.M. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 10668-10672Crossref PubMed Scopus (55) Google Scholar), while on negative TREs the TR activates transcription (9Saatcioglu F. Deng T. Karin M. Cell. 1993; 75: 1095-1105Abstract Full Text PDF PubMed Scopus (104) Google Scholar, 16Wondisford F.E. Steinfelder H.J. Nations M. Radovick S. J. Biol. Chem. 1993; 268: 2749-2754Abstract Full Text PDF PubMed Google Scholar, 17Tomie-Canie M. Day D. Samuels H.H. Freedberg I.M. Blumenberg M. J. Biol. Chem. 1996; 271: 1416-1423Abstract Full Text Full Text PDF PubMed Scopus (65) Google Scholar, 18Bigler J. Eisenman R.N. EMBO J. 1995; 14: 5710-5723Crossref PubMed Scopus (43) Google Scholar). T3 reverses these ligand-independent activities and causes up-regulation on positive TREs and down-regulation on negative TREs. Recently, two separate proteins, termed nuclear co-repressors (TRAC), have been identified which interact with the TR in the absence of ligand to mediate ligand-independent repression on positive TREs (19Seol W. Choi H.S. Moore D.D. Mol. Endocrinol. 1995; 9: 72-85Crossref PubMed Google Scholar, 20Horlein A.J. Naar A.M. Heinzel T. Torchia J. Gloss B. Kurokawa R. Ryan A. Kamei Y. Soderstrom M. Glass C.K. Rosenfeld M.G. Nature. 1995; 377: 397-404Crossref PubMed Scopus (1695) Google Scholar, 21Chen J.D. Evans R.M. Nature. 1995; 377: 454-457Crossref PubMed Scopus (1699) Google Scholar). These proteins possess a carboxyl-terminal domain which mediates their interaction with the TR and amino-terminal domains, which presumably interact with the initiation complex to mediate ligand-independent repression. Their role in ligand-independent activation on negative TREs is unknown.Further complexity in thyroid hormone signaling is mediated by the presence of 3 separate TR isoforms (α1, β1, and β2) which activate or repress transcription in the presence of ligand (22Lazar M.A. Endocr. Rev. 1993; 14: 184-193Crossref PubMed Scopus (811) Google Scholar). The TRβ isoforms are alternatively spliced variants which differ only in their amino terminus, while TRα1 is the product of a separate gene but shares strong homology in its DNA and ligand binding domains but differs also in its amino terminus. Recent data from a number of different groups suggest that isoform-specific differences in both ligand-independent and -dependent activities exist on both positive and negative TREs and that these differences involve the separate amino termini which allow the TRs to either bind TREs differently, interact with members of the initiation complex with different affinities, or possess separate activation domains as is the case with the TRβ2 amino terminus (23Hadzic E. Desai-Yajnik V. Helmer E. Guo S. Wu S. Koudinova N. Casanova J. Raaka B.M. Samuels H.H. Mol. Cell. Biol. 1995; 15: 4507-4517Crossref PubMed Google Scholar, 24Hollenberg A.N. Monden T. Wondisford F.E. J. Biol. Chem. 1995; 270: 14274-14280Abstract Full Text Full Text PDF PubMed Scopus (72) Google Scholar, 25Sjoberg M. Vennstrom B. Mol. Cell. Biol. 1995; 15: 4718-4726Crossref PubMed Google Scholar). These isoforms are also differentially expressed such that TRβ2 is primarily expressed in the hypothalamus and pituitary, suggesting that its primary role may be in mediating negative regulation by T3.To explore the role of nuclear co-repressors in both negative regulation and isoform specificity, we isolated, using the yeast two-hybrid system (26Fields S. Song O. Nature. 1989; 340: 245-247Crossref PubMed Scopus (4822) Google Scholar, 27Chien C.T. Bartel P.L. Sternglanz R. Fields S. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 9578-9582Crossref PubMed Scopus (1222) Google Scholar), 3.1 kilobases of the carboxyl-terminal portion of the human analog of NCoR originally identified by Seol et al. (19Seol W. Choi H.S. Moore D.D. Mol. Endocrinol. 1995; 9: 72-85Crossref PubMed Google Scholar) and Horlein et al. (20Horlein A.J. Naar A.M. Heinzel T. Torchia J. Gloss B. Kurokawa R. Ryan A. Kamei Y. Soderstrom M. Glass C.K. Rosenfeld M.G. Nature. 1995; 377: 397-404Crossref PubMed Scopus (1695) Google Scholar). In this study, we report that this portion of NCoR functions as a powerful dominant inhibitor of endogenous cellular co-repressor activity on TREs and substantially increases ligand-dependent negative regulation by the TRβ1 and α1 isoforms. However, this function was isoform-specific in that the ligand-independent and -dependent activity of the β2 isoform was unaffected by the presence of this inhibitor. Gel-mobility shift assays demonstrate that the β2 amino terminus does not impair the ability of TRβ2 to interact physically with NCoR, but a functional interaction is prevented, perhaps by coactivators which interact with the β2 amino terminus (25Sjoberg M. Vennstrom B. Mol. Cell. Biol. 1995; 15: 4718-4726Crossref PubMed Google Scholar).RESULTS AND DISCUSSIONUsing the yeast two-hybrid system, as described above, we isolated a 3.1-kilobase cDNA clone from a human placental library which showed strong nucleic acid and amino acid homology to the carboxyl-terminal region of murine NCoR (20Horlein A.J. Naar A.M. Heinzel T. Torchia J. Gloss B. Kurokawa R. Ryan A. Kamei Y. Soderstrom M. Glass C.K. Rosenfeld M.G. Nature. 1995; 377: 397-404Crossref PubMed Scopus (1695) Google Scholar) (Fig. 1). Functional analysis of this clone in yeast demonstrated that 1μM T3 could prevent it from interacting with the carboxyl terminus of TRβ. As this original clone contained the interacting domain of NCoR and surrounding amino acids (AA 1539-2453), but lacked the two repressor domains present between AA 1-1120 of murine NCoR, we hypothesized that it would function as a dominant inhibitor of endogenous TRAC activity and thus termed it, NCoRI. The function of this inhibitor in eukaryotic cell lines was then tested on both positive and negative TREs.As demonstrated in Fig. 2A, TRβ1 represses basal transcription on a positive TRE reporter containing two copies of a DR+4 upstream of TK109 in the absence of ligand. Co-transfection of NCoRI, as expected, competes for endogenous co-repressor activity and causes ligand-independent repression to be lost and in the case of the DR+4 reporter leads to partial activation. Ligand-dependent activation was also blocked on the DR+4 reporter (expressed as fold T3 activation) in the presence of NCoRI, primarily because of the loss of ligand-independent repression, suggesting that relief of repression is an important mechanism of thyroid hormone action (data not shown). To rule out an independent activating function of NCoRI, we cloned AA 1539-2453 in-frame with the Gal4 DNA-binding domain and showed that it not alter expression of a UAS-TK Luc reporter construct (data not shown). NCoRI itself also had little effect on basal activity in CV-1 cells or on a cytomegalovirus promoter when co-transfected in isolation.Fig. 2A dominant inhibitor of the nuclear co-repressor family effects ligand-independent activity of the TR in an isoform-specific fashion on positive and negative thyroid hormone response elements. CV-1 cells were co-transfected with 1.67 μg of reporter and 83 ng of pKCR2-TR and/or 83 ng of pKCR2 and/or 83 ng of pKCR2-NCoRI alone to determine ligand-independent activity and -dependent activity. TRΔββ lacks the A/B domain (24Hollenberg A.N. Monden T. Wondisford F.E. J. Biol. Chem. 1995; 270: 14274-14280Abstract Full Text Full Text PDF PubMed Scopus (72) Google Scholar) while TRE220R and TRAHT contain amino acid changes in the hinge region (20Horlein A.J. Naar A.M. Heinzel T. Torchia J. Gloss B. Kurokawa R. Ryan A. Kamei Y. Soderstrom M. Glass C.K. Rosenfeld M.G. Nature. 1995; 377: 397-404Crossref PubMed Scopus (1695) Google Scholar) of TRβ1. A, ligand-independent activity was determined on the DR+4-luciferase reporter using a number of TR isoforms and mutants. The data are quantified as fold repression on the DR+4 reporter (mean ± S.E.) In this case, 1 represents basal activity either in the presence of pKCR2 alone (stippled bars) or pKCR2-NCoRI (black bars). B, ligand-independent activity was determined, as above, on the TRH luciferase promoter. The data are quantified as fold activation on the TRH reporter (mean ± S.E.). In this case, 1 represents basal activity either in the presence of pKCR2 alone (stippled bars) or pKCR2-NCoRI (black bars). C, the TRH-luciferase reporter was transfected in a similar paradigm as above with the three TR isoforms. 24 h after transfection, 10 nM T3 was added. The data are presented as fold basal activity (mean ± S.E.) where 1 (indicated by the arrow) is the basal reporter activity either in the presence of pKCR2 alone (stippled bars) or pKCR2-NCoRI (black bars). The numbering within the graph represents fold repression in the presence of ligand.View Large Image Figure ViewerDownload Hi-res image we have demonstrated TRα1 is a potent ligand-independent repressor TRβ1 on positive TREs. as in Fig. 2A, of NCoRI competes for endogenous co-repressor activity on the DR+4 in the presence of TRβ2 is the other two isoforms in ligand-independent activity on positive TREs, its activity is not by NCoRI. In which lacks the amino terminus of the TRβ isoforms and is to TRβ2 in ligand-independent a significant in the presence of co-transfected NCoRI These data that TRΔββ with NCoRI and that the of the TRβ2 amino terminus reverses this functional In the TRβ1 and TRα1 amino termini a functional interaction with NCoRI. of the hinge region also functional interactions with NCoRI its to bind NCoR, while with NCoRI and is to bind the NCoR (20Horlein A.J. Naar A.M. Heinzel T. Torchia J. Gloss B. Kurokawa R. Ryan A. Kamei Y. Soderstrom M. Glass C.K. Rosenfeld M.G. Nature. 1995; 377: 397-404Crossref PubMed Scopus (1695) Google Scholar). with and TR isoform were on either an J.H. A.M. Harney J.W. G.A. 1995; PubMed Scopus Google Scholar) or TRE (data not that the TRβ2 amino terminus blocked functional interactions with the functional role of the co-repressor family on thyroid hormone we the effect of co-transfected NCoRI on the negative TREs present in the human TRH promoter. In to its ligand-independent activity on positive TREs, the TR causes ligand-independent activation on negative TREs. As in Fig. basal ligand-independent activation by both TRβ1 and TRα1 is by NCoRI and to In the ligand-independent activation by TRβ2 is not by co-transfected NCoRI, with its isoform-specific effect on the positive TREs. which is a ligand-independent a potent to in the presence of NCoRI, that the TRβ2 amino terminus either interactions or the function of the repressing domains of the co-repressor these data it also that ligand-independent activation is important for negative regulation as ligand-independent effects to co-transfected NCoRI to a of fold repression by TRs which interact with NCoRI and Fig. fold repression is only by NCoRI in of as NCoRI can both ligand-independent activation and ligand-dependent repression of the TRH reporter construct by the TRα1 and TRβ1 the that of endogenous nuclear co-repressor activity could ligand-independent and -dependent regulation of negative TREs by these isoforms TRβ2 as the only TR isoform of mediating negative would be with TRβ2 expressed primarily in the and where negative regulation by T3 is R.A. M.A. Larsen P.R. Moore D.D. 1989; PubMed Scopus Google functional data suggest that isoform-specific interactions with the nuclear co-repressor we the different TR isoforms could interact with NCoRI when to In vitro of the NCoRI cDNA a primary of with its (data not shown). of in vitro NCoRI with in vitro TRβ1 in the presence of a DR+4 a to TR homodimer, Fig. and a in the TR homodimer, with a interaction can be in the presence of nM T3 Fig. demonstrate that NCoRI does not bind the in the absence of TR shown). TRβ2 (Fig. as as TRα1 also interact with NCoRI as does (Fig. As expected, does not bind NCoRI (Fig. These data that functional interactions between TRβ2 and NCoRI not to the two proteins can physically interact on the response with the domain of the TR the primary region for interaction (20Horlein A.J. Naar A.M. Heinzel T. Torchia J. Gloss B. Kurokawa R. Ryan A. Kamei Y. Soderstrom M. Glass C.K. Rosenfeld M.G. Nature. 1995; 377: 397-404Crossref PubMed Scopus (1695) Google Scholar, 21Chen J.D. Evans R.M. Nature. 1995; 377: 454-457Crossref PubMed Scopus (1699) Google Scholar). the TRβ2 amino terminus either through the of other proteins or through changes in the of the TR the repressing function of the nuclear co-repressor from nuclear co-repressor inhibitor to TRβ1 in a ligand-independent fashion in the presence of a Gel-mobility shift assays were out using of in vitro proteins from and a DR+4 The of T3 used was The DR+4 was with in vitro and NCoRI or Large Image Figure ViewerDownload Hi-res image nuclear co-repressor to TR isoforms. The DR+4 was with in vitro and TR and NCoRI as A, B, C, Large Image Figure ViewerDownload Hi-res image of the interactions between NCoRI and the TR in the shift demonstrate that NCoRI to interact with the TR or with TR as the complex Fig. is not in either or by NCoRI. the of the is with RXR with NCoRI for binding to the To this we a between and As Fig. the of RXR competes for the TR binding to NCoRI and the complex These data that the heterodimer does not interact with NCoRI in the presence of DNA in this and that the interaction between TR or and NCoR may be important for co-repressor nuclear co-repressor inhibitor the TR The DR+4 was with TRβ1 and NCoRI. the the an of was added. The were on a by Large Image Figure ViewerDownload Hi-res image we have demonstrated that endogenous nuclear co-repressor activity can be for with a dominant negative of which lacks the two repressing domains, blocked ligand-independent repression by the TR on positive TREs and ligand-independent activity on negative TREs. However, co-repressor activity to be isoform-specific as the ligand-independent function of TRβ2 is not by NCoRI. is demonstrated on negative TREs where ligand-independent activation and repression by TRβ2 is not by NCoRI its interaction with NCoRI is Thus, the A/B domain of the TRβ2 can nuclear co-repressor function through either or the of which the function of the the that endogenous may exist which could thyroid hormone and retinoic acid action in a Recently, an endogenous truncated isoform of the co-repressor was identified S. Mol. Endocrinol. 1996; PubMed Scopus Google Scholar). protein to function as an to be other isoforms of the human NCoR INTRODUCTIONThe thyroid hormone receptor (TR), 1The abbreviations used are: TRthyroid hormone receptorTREthyroid response elementRXRretinoid X receptorT3triiodothyronineAAamino acid(s)Tracthyroid hormone and retinoic acid-receptor-associated co-repressors. a member of the steroid/thyroid hormone receptor superfamily, mediates gene regulation through its ability to bind to specific thyroid hormone response elements (TREs) (1Desvergne B. Petty K.J. Nikodem V.M. J. Biol. Chem. 1991; 266: 1008-1013Abstract Full Text PDF PubMed Google Scholar, 2Brent G.A. Harney J.W. Chen Y. Warne R.L. Moore D.D. Larsen P.R. Mol. Endocrinol. 1989; 3: 1996-2004Crossref PubMed Scopus (182) Google Scholar, 3Norman M.F. Lavin T.N. Baxter J.D. West B.L. J. Biol. Chem. 1989; 264: 12063-12073Abstract Full Text PDF PubMed Google Scholar, 4Umesono K. Murakami K.K. Thompson C.C. Evans R.M. Cell. 1991; 65: 1255-1266Abstract Full Text PDF PubMed Scopus (1488) Google Scholar) either as a monomer, homodimer, or as a heterodimer with RXR isoforms (5Yu V.C. Delsert C. Andersen B. Holloway J.M. Devary O.V. Naar A.M. Kim S.Y. Boutin J.M. Glass C.K. Rosenfeld M.G. Cell. 1991; 67: 1251-1266Abstract Full Text PDF PubMed Scopus (1051) Google Scholar, 6Heyman R.A. Mangelsdorf D.J. Dyck J.A. Stein R.B. Eichele G. Evans R.M. Thaller C. Cell. 1992; 68: 397-406Abstract Full Text PDF PubMed Scopus (1558) Google Scholar, 7Mangelsdorf D.J. Ong E.S. Dyck J.A. Evans R.M. Nature. 1990; 345: 224-229Crossref PubMed Scopus (1252) Google Scholar). Through its interactions with specific TREs and other protein co-activators, the TR can either positively or negatively regulate gene expression in the presence of its ligand, triiodothyronine (T3) (8Hollenberg A.N. Monden T. Flynn T.R. Boers M.-E. Cohen O. Wondisford F.E. Mol. Endocrinol. 1995; 9: 540-550Crossref PubMed Google Scholar, 9Saatcioglu F. Deng T. Karin M. Cell. 1993; 75: 1095-1105Abstract Full Text PDF PubMed Scopus (104) Google Scholar, 10Wondisford F.E. Farr E.A. Radovick S. Steinfelder H.J. Moates J.M. McClaskey J.H. Weintraub B.D. J. Biol. Chem. 1989; 264: 14601-14604Abstract Full Text PDF PubMed Google Scholar). The TR also has significant ligand-independent activity in both cell culture and in vitro transcription systems (11Casanova J. Helmer E. Selmi-Ruby S. Qi J.-S. Au-Fliegner M. Desai-Yajnik V. Koudinova N. Yarm F. Raaka B.M. Samuels H.H. Mol. Cell. Biol. 1994; 14: 5756-5765Crossref PubMed Scopus (122) Google Scholar, 12Tong G.-X. Jeyakumar M. Tanen M. Bagchi M.K. Mol. Cell. Biol. 1996; 16: 1909-1920Crossref PubMed Google Scholar). On positive TREs, the TR is a potent repressor of transcription in the absence of ligand (13Baniahmad A. Kohne A.C. Renkawitz R. EMBO J. 1992; 11: 1015-1023Crossref PubMed Scopus (238) Google Scholar, 14Sap J. Munoz A. Schmitt J. Stunnenberg H. Vennstrom B. Nature. 1989; 340: 242-244Crossref PubMed Scopus (295) Google Scholar, 15Damm K. Evans R.M. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 10668-10672Crossref PubMed Scopus (55) Google Scholar), while on negative TREs the TR activates transcription (9Saatcioglu F. Deng T. Karin M. Cell. 1993; 75: 1095-1105Abstract Full Text PDF PubMed Scopus (104) Google Scholar, 16Wondisford F.E. Steinfelder H.J. Nations M. Radovick S. J. Biol. Chem. 1993; 268: 2749-2754Abstract Full Text PDF PubMed Google Scholar, 17Tomie-Canie M. Day D. Samuels H.H. Freedberg I.M. Blumenberg M. J. Biol. Chem. 1996; 271: 1416-1423Abstract Full Text Full Text PDF PubMed Scopus (65) Google Scholar, 18Bigler J. Eisenman R.N. EMBO J. 1995; 14: 5710-5723Crossref PubMed Scopus (43) Google Scholar). T3 reverses these ligand-independent activities and causes up-regulation on positive TREs and down-regulation on negative TREs. Recently, two separate proteins, termed nuclear co-repressors (TRAC), have been identified which interact with the TR in the absence of ligand to mediate ligand-independent repression on positive TREs (19Seol W. Choi H.S. Moore D.D. Mol. Endocrinol. 1995; 9: 72-85Crossref PubMed Google Scholar, 20Horlein A.J. Naar A.M. Heinzel T. Torchia J. Gloss B. Kurokawa R. Ryan A. Kamei Y. Soderstrom M. Glass C.K. Rosenfeld M.G. Nature. 1995; 377: 397-404Crossref PubMed Scopus (1695) Google Scholar, 21Chen J.D. Evans R.M. Nature. 1995; 377: 454-457Crossref PubMed Scopus (1699) Google Scholar). These proteins possess a carboxyl-terminal domain which mediates their interaction with the TR and amino-terminal domains, which presumably interact with the initiation complex to mediate ligand-independent repression. Their role in ligand-independent activation on negative TREs is unknown.Further complexity in thyroid hormone signaling is mediated by the presence of 3 separate TR isoforms (α1, β1, and β2) which activate or repress transcription in the presence of ligand (22Lazar M.A. Endocr. Rev. 1993; 14: 184-193Crossref PubMed Scopus (811) Google Scholar). The TRβ isoforms are alternatively spliced variants which differ only in their amino terminus, while TRα1 is the product of a separate gene but shares strong homology in its DNA and ligand binding domains but differs also in its amino terminus. Recent data from a number of different groups suggest that isoform-specific differences in both ligand-independent and -dependent activities exist on both positive and negative TREs and that these differences involve the separate amino termini which allow the TRs to either bind TREs differently, interact with members of the initiation complex with different affinities, or possess separate activation domains as is the case with the TRβ2 amino terminus (23Hadzic E. Desai-Yajnik V. Helmer E. Guo S. Wu S. Koudinova N. Casanova J. Raaka B.M. Samuels H.H. Mol. Cell. Biol. 1995; 15: 4507-4517Crossref PubMed Google Scholar, 24Hollenberg A.N. Monden T. Wondisford F.E. J. Biol. Chem. 1995; 270: 14274-14280Abstract Full Text Full Text PDF PubMed Scopus (72) Google Scholar, 25Sjoberg M. Vennstrom B. Mol. Cell. Biol. 1995; 15: 4718-4726Crossref PubMed Google Scholar). These isoforms are also differentially expressed such that TRβ2 is primarily expressed in the hypothalamus and pituitary, suggesting that its primary role may be in mediating negative regulation by T3.To explore the role of nuclear co-repressors in both negative regulation and isoform specificity, we isolated, using the yeast two-hybrid system (26Fields S. Song O. Nature. 1989; 340: 245-247Crossref PubMed Scopus (4822) Google Scholar, 27Chien C.T. Bartel P.L. Sternglanz R. Fields S. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 9578-9582Crossref PubMed Scopus (1222) Google Scholar), 3.1 kilobases of the carboxyl-terminal portion of the human analog of NCoR originally identified by Seol et al. (19Seol W. Choi H.S. Moore D.D. Mol. Endocrinol. 1995; 9: 72-85Crossref PubMed Google Scholar) and Horlein et al. (20Horlein A.J. Naar A.M. Heinzel T. Torchia J. Gloss B. Kurokawa R. Ryan A. Kamei Y. Soderstrom M. Glass C.K. Rosenfeld M.G. Nature. 1995; 377: 397-404Crossref PubMed Scopus (1695) Google Scholar). In this study, we report that this portion of NCoR functions as a powerful dominant inhibitor of endogenous cellular co-repressor activity on TREs and substantially increases ligand-dependent negative regulation by the TRβ1 and α1 isoforms. However, this function was isoform-specific in that the ligand-independent and -dependent activity of the β2 isoform was unaffected by the presence of this inhibitor. Gel-mobility shift assays demonstrate that the β2 amino terminus does not impair the ability of TRβ2 to interact physically with NCoR, but a functional interaction is prevented, perhaps by coactivators which interact with the β2 amino terminus (25Sjoberg M. Vennstrom B. Mol. Cell. Biol. 1995; 15: 4718-4726Crossref PubMed Google Scholar).