- Email: [email protected]
The Interaction of the Vitamin D Receptor with Nuclear Receptor Corepressors and Coactivators
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS ARTICLE NO.
253, 358 –363 (1998)
RC989799
The Interaction of the Vitamin D Receptor with Nuclear Receptor Corepressors and Coactivators Tetsuya Tagami,* Ward H. Lutz,† Rajiv Kumar,† and J. Larry Jameson* *Division of Endocrinology, Metabolism, and Molecular Medicine, Northwestern University Medical School, Chicago, Illinois 60611; and †Departments of Medicine, Biochemistry and Molecular Biology, Mayo Clinic and Foundation, Rochester, Minnesota 55905
Received October 26, 1998
The vitamin D receptor (VDR), thyroid hormone receptor (TR), and retinoic acid receptor (RAR) are ligand-dependent transcription factors that function via the formation of heterodimeric complexes with retinoid X receptor (RXR). Although TR and RAR are known to act as transcriptional repressors in the absence of cognate ligands, it is not clear whether VDR exhibits this property. Recently, transcriptional repression (basal silencing) by TR and RAR was shown to be mediated by nuclear receptor corepressors (CoRs), such as NCoR and SMRT. In this report, we examined the silencing ability of VDR and its interaction with NCoR and SMRT using mammalian twohybrid assays. The Gal4-VDR fusion protein silenced the basal expression of a reporter that contains Gal4 binding sites, but the degree of silencing activity was weaker than that of Gal4-TR. In mammalian twohybrid assays, the interaction of VP16-SMRT or VP16NCoR was also stronger with Gal4-TR than with Gal4VDR. Similar results were obtained when the assay was performed using the opposite configuration. Gal4SMRT or Gal4-NCoR interacted better with VP16-TR than with VP16-VDR. These interactions were disrupted by the addition of cognate ligands. In contrast, VP16-VDR interacted better than VP16-TR when studied with a coactivator, Gal4-SRC1, or with the heterodimeric partner, Gal4-RXR. Consistent with these findings, relatively weak transcriptional silencing by the native VDR was observed using the osteopontin VDRE. Thus, in comparison to TR, VDR exhibits relatively weak ligand-independent transcriptional silencing, but it possesses strong dimerization with RXR and ligand-induced binding to transcriptional coactivators. © 1998 Academic Press
Nuclear receptors act primarily through direct association with specific DNA sequences known as hormone response elements (HREs) (1). The HREs for the nonsteroid members of the receptor superfamily consist 0006-291X/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
of a minimal core heptad consensus sequence, AGGTCA (2). The HREs for the vitamin D receptor (VDR), thyroid hormone receptor (TR), and retinoic acid receptor (RAR) are composed of direct repeats (DRs) of this core half-site spaced by 3, 4, or 5 nucleotides (i.e., DR3, DR4, and DR5, respectively) (3). These receptors usually bind to DNA as heterodimers with a common partner, the retinoid X receptor (RXR). In the absence of ligands, TR and RAR have been shown to act as transcriptional repressors of target genes (4 – 6). Recently, two classes of corepressors (CoRs), termed nuclear receptor corepressor (NCoR) (7,8) or RXR-interacting protein 13 (RIP13) (9) and silencing mediator for retinoid and thyroid hormone receptors (SMRT) (10) or T3 receptor-associating cofactor (TRAC) (11) were shown to mediate ligand-independent repression by these receptors. Although VDR is structurally related to TR and RAR, its silencing ability and potential interactions with CoRs remain controversial (7,12–15). In this study, we examined the silencing ability of VDR and its interaction with CoRs using a sensitive mammalian two-hybrid assay. We also examined the interaction of the VDR with a nuclear coactivator, SRC1, using similar methods. MATERIALS AND METHODS Plasmid constructions. The pCMX-VDR and pCMX-SMRT expression vectors were provided by K. Umesono and R. M. Evans (Salk Institute, San Diego, CA) (3,12). The VP16 constructs for VDR (92427), TRb1 (174-461) and RXRa (199-462) contain the ligand binding domain (LBD) of the receptors downstream of the VP16 activation domain in-frame in pCMX. The pCMX-NCoR expression vector was provided by M. G. Rosenfeld (University of California, San Diego, CA) (7). The SRC-1 cDNA was provided by B. W. O’Malley (Baylor College of Medicine, Houston, TX) (16). The Gal4-NCoR (residues 1552-2453) and Gal4-SRC1 (residues 213-1061) constructs contain the indicated TR interaction domains of these proteins fused downstream of the Gal4 DNA binding domain (DBD) in pSG424 (17). The Gal4 reporter plasmid, UAS-E1BTATA-Luc (18), contains five copies of the UAS element upstream of E1BTATA in pA3-Luc. The osteopontin VDRE-tk-Luc contains three copies of the mouse osteopontin VDRE (19) upstream of a 151 base pair truncated thymidine
358
Vol. 253, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. Silencing ability of nuclear receptors and their interactions with CoRs. The format of the mammalian two-hybrid experiment is shown at the top of the figure. The indicated Gal4 constructs (100 ng) were cotransfected into TSA-201 cells with 300 ng of VP16 constructs together with 100 ng of the reporter gene, UAS-E1BTATA-Luc, in the absence of ligand. (A) Basal expression without VP16 constructs. (B) Interaction with SMRT. (C) Interaction with NCoR. The numbers at the top of each bar indicate fold stimulation mediated by the interaction with VP16-CoR relative to VP16 alone. kinase promoter, inserted between the SacI and HindIII sites of the pGL3 plasmid (Promega, Madison, WI). Transient expression assays. TSA-201 cells, a clone of human embryonic kidney 293 cells (20), and CV1 cells, were maintained in Dulbecco’s modified essential medium (DMEM) containing 10% fetal bovine serum (FBS), penicillin (100 U/mL) and streptomycin (100 mg/mL). Cells were grown in serum-free DMEM supplemented with Nutridoma (Boehringer Mannheim, Indianapolis, IN) and were transfected by the calcium phosphate method as described (21). The total amount of expression plasmid DNA was kept constant in the different experimental groups by adding corresponding amounts of the same plasmids without cDNA inserts. After exposure to the calcium phosphate-DNA precipitate for 8h, DMEM with 1% Nutridoma was added, in the absence or presence of 1, 25-(OH)2 D3 (Calbiochem, San Diego, CA) (100 nM), T3 (1 mM) and/or 9cis-retinoic acid (100 nM) (ICN, Costa Mesa, CA). Cells were harvested after 40h for measurements of luciferase activity (22). Results are expressed as the mean 6 SEM from at least three transfections performed in triplicate.
RESULTS Gal4-VDR can silence transcription and interact with CoRs. The LBD of VDR, TRß or RXRa was fused to the DBD of the yeast transcription factor, Gal4. The reporter gene, UAS-E1BTATA-Luc, contains five Gal4 binding sites and was used to assess the silencing ability of the receptors in the absence of cognate ligands (Fig. 1A). Relative to the Gal4 DBD alone, Gal4-TR showed 75% inhibition in TSA-201 cells. Gal4RXR resulted in 50% inhibition under these conditions. The silencing ability of Gal4-VDR was intermediate between Gal4-TR and Gal4-RXR.
Interactions between nuclear receptors and CoRs were examined using a mammalian two-hybrid assay. The LBD of VDR, TRß or RXRa was fused to the Gal4-DBD to target these proteins to the reporter gene. The carboxyl terminal half of the SMRT or NCoR proteins, which contain the two TR interaction domains, was fused to the transcriptional activation domain of VP16. UAS-E1BTATA-Luc was used to assess in vivo interactions between the Gal4-receptors and VP16CoR. In this manner, interactions between TR and the CoRs result in the transcriptional activation of the reporter by the VP16 domain (Fig. 1A). Gal4-TR was used initially to test the interaction assay. Relative to the Gal4 DBD alone, Gal4-TR was stimulated 149-fold by the addition of VP16-SMRT (Fig. 1B). Gal4-RXR was stimulated 15-fold by VP16-SMRT. The interaction of Gal4-VDR with VP16-SMRT (91-fold) was stronger than that of Gal4-RXR, but weaker than that of Gal4-TR. NCoR interactions were examined with receptors in a similar manner (Fig. 1C). Strong NCoR interactions were detected for TR (1437-fold), whereas the interaction with RXR was less pronounced (10fold). In the case of VDR, Gal4-VDR was stimulated 68-fold by VP16-NCoR. These results indicate that the silencing ability of nuclear receptors correlates with CoR interactions with SMRT and NCoR. The interactions between nuclear receptors with CoRs were also examined using the opposite configuration of
359
Vol. 253, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. Interactions of receptors with CoRs or a CoA, SRC1. The format of the mammalian two-hybrid experiment is shown at the top of the figure. The indicated Gal4 constructs (100 ng) were cotransfected into TSA-201 cells with 300 ng of VP16 constructs together with 100 ng of the reporter gene, UAS-E1BTATA-Luc, in the absence or presence of ligand. (A) Interaction with SMRT. (B) Interaction with NCoR. (C) Interaction with SRC1. The numbers on the top of each bar indicate fold stimulation mediated by the interaction with VP16-cofactor relative to VP16 alone.
fusion proteins (Fig. 2). The C-terminal half of SMRT, NCoR, or the SRC1 coactivator (CoA) protein was fused to the Gal4-DBD to target these proteins to the reporter gene. The LBD of VDR, TRb or RXRa was fused to VP16 to induce transcription when these proteins are recruited to the Gal4 proteins. In this configuration, the effect of the ligand on protein interactions can be assessed, independent of any effects on the Gal4 proteins. Relative to VP16 alone, Gal4-SMRT was stimulated 414-fold by the addition of VP16-TR and this interaction was markedly decreased by the addition of T3, consistent with the known effect of ligand to disrupt TR-CoR interactions (Fig. 2A). In contrast, VP16-VDR stimulated the activity of Gal4-SMRT only 16-fold and the interaction was almost eliminated by the addition of 1, 25-dihydroxyvitamin D3 (VD3). Gal4-NCoR was stimulated 2008-fold by VP16-TR, but only 6-fold by VP16-VDR (Fig. 2B). Again, cognate ligands decreased the interactions between NCoR and the receptors. The coactivator interaction was also examined in this assay (Fig. 2C). In the presence of ligand, Gal4SRC1 was stimulated 100-fold by the addition of VP16-TR and 259-fold by VP16-VDR, indicating that CoA interaction with VDR is stronger than with TR. These findings also confirm that the VP16-VDR construct is fully functional. VDR can heterodimerize with RXR, but not with TR. The receptor-receptor interactions were also examined using a mammalian two-hybrid assay (Fig. 3). Relative
to VP16 alone, Gal4-RXR was stimulated 820-fold by the addition of VDR and 503-fold by VP16-TR, but not with VP16-RXR. Using the opposite combinations, Gal4-VDR was stimulated 456-fold by the addition of VP16-RXR, and Gal4-TR was stimulated 1228-fold by VP16-RXR. No interaction was observed between TR and VDR, consistent with the recent findings of RavalPanda et al. (23). The silencing ability of VDR in the context of a native VDRE. The silencing ability of native VDR was also examined using a reporter gene that contains a vitamin D response element (VDRE). Since the reporter gene, VDRE-tk-Luc, was stimulated by VD3 without cotransfection of VDR expression plasmids in TSA201 cells (data not shown), VDR-negative CV1 cells were used to test the silencing activity of native VDR (14). When VDR was cotransfected with VDRE-tk-Luc, basal transcription was repressed (2-fold) in the absence of VD3 (Fig. 4A). Interestingly, unliganded RXR or TR similarly repressed VDRE mediated transcription, even though this response element is not optimized for binding by these factors. In the presence of VD3, transcription was stimulated only in the presence of cotransfected VDR (Fig. 4B). The RXR ligand, 9 cis-retinoic acid (9cRA), stimulated RXR-mediated activity by 2-fold, whereas no synergistic effect was seen in the presence of liganded VDR and liganded RXR. In contrast, T3 had no effect on the TR-mediated silencing of this reporter gene.
360
Vol. 253, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 3. Mutual interactions among receptors. The format of the mammalian two-hybrid experiment is shown at the top of the figure. The indicated Gal4 constructs (100 ng) were cotransfected into TSA-201 cells with 300 ng of VP16 constructs together with 100 ng of the reporter gene, UAS-E1BTATA-Luc, in the absence of ligand. The numbers at the top of each bar indicate fold-stimulation mediated by the receptor interaction relative to VP16 alone.
DISCUSSION Basal repression by TR and RAR has been demonstrated in the absence of cognate ligand (4,5,24). VDR belongs to the same subfamily of the nuclear hormone receptors that heterodimerize with RXR and mediate ligand-dependent gene transcription. Although VDR also binds to similar response elements (DR3) as those used by TR (DR4) and RAR (DR5), the silencing ability of VDR is still controversial. Yen et al. (14), but not Cheskis and Freedman (25), observed repressed transcription by unliganded VDR using the osteopontin VDRE in CV1 cells. Masuyama et al. (15) also did not observe the silencing effect using the osteocalcin VDRE in either COS7 cells or CV1 cells. It has been shown that CoR proteins, such as NCoR and SMRT, mediate the transcriptional silencing by TR and RAR (7,10). The putative CoR binding domain, in the hinge region of TR and RAR (termed CoR box) (7), is well-conserved in the VDR. Critical residues for CoR interaction have been identified in another subdomain in v-erb A and TRb (26), and these amino acids are also conserved in the VDR. We have found that leucine 454 of the AF2 domain of TRb exerts inhibitory effects on CoR interactions (27), and this residue is also present in VDR.
Given these conserved structural features in the VDR, one might anticipate interactions with CoRs. Although several studies have examined the interactions TR and RAR with CoRs, there is little information about whether VDR binds to these proteins. In the case of NCoR, Horlein et al. (7) showed, using a yeast two hybrid system, that NCoR interacted with TR and RAR, but not with RXRg, VDR, estrogen receptor (ER), or glucocorticoid receptor (GR). In the case of SMRT, it has been reported that it interacts with RAR, TR and, to lesser degree, with RXR, but not with VDR, using mammalian two-hybrid assays (12). However, Li et al. (13) demonstrated later that SMRT interacts weakly with VDR using yeast two-hybrid assays. We showed here that VDR does interact with NCoR and SMRT, but these interactions are weaker than that of TR. Moreover, there is a potential selectivity in the interaction of receptors with the two different CoRs. TR interacts preferentially with NCoR, whereas VDR interacts better with SMRT in our mammalian two-hybrid assays. These features were seen independent of the configuration of the two-hybrid assay. Differential expression of CoRs in various tissues may therefore influence the silencing activity of coexisting receptors (7,12).
361
Vol. 253, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 4. Transcriptional properties of receptors using a reporter that contains the osteopontin VDRE. The reporter gene is shown at the top of the figure. The indicated expression plasmids (10 ng) were transfected into CV1 cells together with 100 ng of the reporter gene, VDRE-tk-Luc (A and B) in the absence or presence of ligand.
Some antagonists have been shown to recruit CoRs to ER (28,29), or to the progesterone receptor (PR) (30), which otherwise do not bind to CoRs in the absence of ligands. Therefore, it is possible that the identification of antagonists for the VDR may provide ligands that could enhance its binding to CoRs and increase silencing. It is of interest that mutations in the human VDR are inherited as autosomal recessive traits (31,32), whereas those in the TR transmitted primarily in an autosomal dominant manner (33). These features probably reflect the fact that the mutant TRs can inhibit the function of the wild type allele in a dominant negative manner, whereas the mutant VDRs are functionally inactive. We (34) and others (35) have previously shown that CoR binding to mutant TRs is necessary for their dominant negative activities. The present study, demonstrating relatively weak interactions between VDR and CoRs, suggests that dominant negative forms of VDR might not exist, or that they might only weakly inhibit VDR-responsive genes. This idea could be tested by generating artificial mutations in the VDR that selectively lack ligand binding or transcriptional activation. Cross-talk between VDR and TR has been suggested because they coexist in some tissues such as pituitary, brain, intestine, skin, thyroid, kidney and bone (36). Several mechanisms have been suggested, including heterodimerization between VDR and TR (37), sequestration of RXR (23), and competition for binding to DNA (14). We observed that TR did not heterodimerize
with VDR in a mammalian two-hybrid assay, but TR was able to weakly silence the basal expression of a VDRE reporter gene. Moreover, T3 did not relieve this silencing, suggesting that TR may act as a constitutive repressor of a VD3 responsive gene. Such differential effects by RAR have been observed on different RAR response elements (8). Further investigation is necessary for understanding the potential cross-talk among these nuclear receptors. In summary, we find that VDR exhibits relatively weak transcriptional silencing, consistent with reduced interactions with NCoR and SMRT. However, it heterodimerizes well with RXR and it binds a coactivator, SRC-1, very well in the presence of ligand. These findings suggest that, in comparison to related family members, like TR and RAR, the function of the VDR is biased towards transcriptional activation as opposed to transcriptional repression. ACKNOWLEDGMENTS We are grateful to M. G. Rosenfeld, R. M. Evans, and B. W. O’Malley, A. for providing plasmids. This work was supported by National Institute of Health (NIH) Grants DK42144 (to J.L.J) and DK25409 (to R. K.).
REFERENCES
362
1. Evans, R. M. (1988) Science 240(4854), 889 – 895. 2. Glass, C. (1994) Endocr. Rev. 15, 391– 407.
Vol. 253, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
3. Umesono, K., Murakami, K. K., Thompson, C. C., and Evans, R. M. (1991) Cell 65, 1255–1266. 4. Brent, G. A., Dunn, M. K., Harney, J. W., Gulick, T., Larsen, P. R., and Moore, D. D. (1989) New Biol. 1(3), 329 –336. 5. Damm, K., Thompson, C. C., and Evans, R. M. (1989) Nature 339(6226), 593–597. 6. Baniahmad, A., Kohne, A. C., and Renkawitz, R. (1992) EMBO J. 11(3), 1015–1023. 7. Horlein, A. J., Naar, A. M., Heinzel, T., Torchia, J., Gloss, B., Kurokawa, R., Ryan, A., Kamei, Y., Soderstrom, M., Glass, C. K., and Rosenfeld, M. G. (1995) Nature 377(6548), 397– 404. 8. Kurokawa, R., Soderstrom, M., Horlein, A., Halachmi, S., Brown, M., Rosenfeld, M. G., and Glass, C. K. (1995) Nature 377(6548), 451– 454. 9. Lee, J. W., Ryan, F., Swaffield, J. C., Johnston, S. A., and Moore, D. D. (1995) Nature 374(6517), 91–94. 10. Chen, J. D., and Evans, R. M. (1995) Nature 377(6548), 454 – 457. 11. Sande, S., and Privalsky, M. L. (1996) Mol. Endocrinol. 10, 813– 825. 12. Chen, J. D., Umesono, K., and Evans, R. M. (1996) Proc. Natl. Acad. Sci. U.S.A. 93, 7567–7571. 13. Li, H., Leo, C., Schroen, D. J., and Chen, J. D. (1997) Mol. Endocrinol. 11, 2025–2037. 14. Yen, P. M., Liu, Y., Sugawara, A., and Chin, W. W. (1996) J. Biol. Chem. 271, 10,910 –10,916. 15. Masuyama, H., Jefcoat, S. C., and MacDoonald, P. N. (1997) Mol. Endocrinol. 11, 218 –228. 16. Onate, S. A., Tsai, S. Y., Tsai, M. J., and O’Malley, B. W. (1995) Science 270(5240), 1354 –1357. 17. Sadowski, I., Ma, J., Tiezenberg, S., and Ptashne, M. (1988) Nature 335, 563–564. 18. Pestell, R. G., Albanese, C., Watanabe, G., Lee, R. J., Lastowiecki, P., Zon, L., Ostrowski, M., and Jameson, J. L. (1996) Mol. Endocrinol. 10, 1084 –1094. 19. Veenstra, T. D., Benson, L. M., A., C. T., Tomlinson, A. J., Kumar, R., and Naylor, S. (1998) Nature Biotechnol. 16, 262– 266.
20. Margolskee, R. F., McHendry-Rinde, B., and Horn, R. (1993) Biotechniques 15(5), 906 –911. 21. Tagami, T., Madison, L. D., Nagaya, T., and Jameson, J. L. (1997) Mol. Cell Biol. 17(5), 2642–2648. 22. deWet, J. R., Wood, K. V., DeLuca, M., and Helinski, D. R. (1987) Mol. Cell Biol. 7, 725–737. 23. Raval-Panda, M., Freedman, L. P., Li, H., and Christakos, S. (1998) Mol. Endocrinol. 12, 1367–1379. 24. Baniahmad, A., Steiner, C., Kohne, A. C., and Renkawitz, R. (1990) Cell 61(3), 505–514. 25. Cheskis, B., and Freedman, L. P. (1994) Mol. Cell Biol. 14, 3329 –3338. 26. Busch, K., Martin, B., Renkawitz, R., and Muller, M. (1997) Mol. Endocrinol. 11, 379 –389. 27. Tagami, T., Gu, W., Peairs, P. T., West, B. L., and Jameson, J. L. Mol. Endocrinol. (in press). 28. Smith, C. L., Nawaz, Z., and O’Malley, B. W. (1997) Mol. Endocrinol. 11, 657– 666. 29. Jackson, T. A., Richer, J. K., Bain, D. L., Takimoto, G. S., Tung, L., and Horowitz, K. B. (1997) Mol. Endocrinol. 11, 693–705. 30. Zhang, X., Jeyakumar, M., Petukhov, S., and Bagchi, M. (1998) Mol. Endocrinol. 12, 513–524. 31. Hughes, M. R., Malloy, P. J., Kieback, D. G., Kesterson, R. A., Pike, J. W., Feldman, D., and O’Malley, B. W. (1988) Science 242, 1702–1705. 32. Whitefield, G. K., Selznick, S. H., Haussler, C. A., Hsieh, J.-C., Galligan, M. A., Jurtka, P. W., Thompson, P. D., Lee, S. M., Zerwekh, J. E., and Haussler, M. R. (1996) Mol. Endocrinol. 10, 1617–1631. 33. Refetoff, S., Weiss, R. E., and Usala, S. J. (1993) Endocr. Rev. 14(3), 348 –399. 34. Tagami, T., and Jameson, J. L. (1998) Endocrinology 139, 640–650. 35. Yoh, S. M., Chatterjee, V. K. K., and Privalsky, M. L. (1997) Mol. Endocrinol. 11, 470 – 480. 36. Walters, M. R. (1992) Endocr. Rev. 13, 1– 46. 37. Schrader, M., Muller, K. M., Nayeri, S., Kahlen, J. P., and Carlberg, C. (1994) Nature 370(6488), 382–386.
363
253, 358 –363 (1998)
RC989799
The Interaction of the Vitamin D Receptor with Nuclear Receptor Corepressors and Coactivators Tetsuya Tagami,* Ward H. Lutz,† Rajiv Kumar,† and J. Larry Jameson* *Division of Endocrinology, Metabolism, and Molecular Medicine, Northwestern University Medical School, Chicago, Illinois 60611; and †Departments of Medicine, Biochemistry and Molecular Biology, Mayo Clinic and Foundation, Rochester, Minnesota 55905
Received October 26, 1998
The vitamin D receptor (VDR), thyroid hormone receptor (TR), and retinoic acid receptor (RAR) are ligand-dependent transcription factors that function via the formation of heterodimeric complexes with retinoid X receptor (RXR). Although TR and RAR are known to act as transcriptional repressors in the absence of cognate ligands, it is not clear whether VDR exhibits this property. Recently, transcriptional repression (basal silencing) by TR and RAR was shown to be mediated by nuclear receptor corepressors (CoRs), such as NCoR and SMRT. In this report, we examined the silencing ability of VDR and its interaction with NCoR and SMRT using mammalian twohybrid assays. The Gal4-VDR fusion protein silenced the basal expression of a reporter that contains Gal4 binding sites, but the degree of silencing activity was weaker than that of Gal4-TR. In mammalian twohybrid assays, the interaction of VP16-SMRT or VP16NCoR was also stronger with Gal4-TR than with Gal4VDR. Similar results were obtained when the assay was performed using the opposite configuration. Gal4SMRT or Gal4-NCoR interacted better with VP16-TR than with VP16-VDR. These interactions were disrupted by the addition of cognate ligands. In contrast, VP16-VDR interacted better than VP16-TR when studied with a coactivator, Gal4-SRC1, or with the heterodimeric partner, Gal4-RXR. Consistent with these findings, relatively weak transcriptional silencing by the native VDR was observed using the osteopontin VDRE. Thus, in comparison to TR, VDR exhibits relatively weak ligand-independent transcriptional silencing, but it possesses strong dimerization with RXR and ligand-induced binding to transcriptional coactivators. © 1998 Academic Press
Nuclear receptors act primarily through direct association with specific DNA sequences known as hormone response elements (HREs) (1). The HREs for the nonsteroid members of the receptor superfamily consist 0006-291X/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
of a minimal core heptad consensus sequence, AGGTCA (2). The HREs for the vitamin D receptor (VDR), thyroid hormone receptor (TR), and retinoic acid receptor (RAR) are composed of direct repeats (DRs) of this core half-site spaced by 3, 4, or 5 nucleotides (i.e., DR3, DR4, and DR5, respectively) (3). These receptors usually bind to DNA as heterodimers with a common partner, the retinoid X receptor (RXR). In the absence of ligands, TR and RAR have been shown to act as transcriptional repressors of target genes (4 – 6). Recently, two classes of corepressors (CoRs), termed nuclear receptor corepressor (NCoR) (7,8) or RXR-interacting protein 13 (RIP13) (9) and silencing mediator for retinoid and thyroid hormone receptors (SMRT) (10) or T3 receptor-associating cofactor (TRAC) (11) were shown to mediate ligand-independent repression by these receptors. Although VDR is structurally related to TR and RAR, its silencing ability and potential interactions with CoRs remain controversial (7,12–15). In this study, we examined the silencing ability of VDR and its interaction with CoRs using a sensitive mammalian two-hybrid assay. We also examined the interaction of the VDR with a nuclear coactivator, SRC1, using similar methods. MATERIALS AND METHODS Plasmid constructions. The pCMX-VDR and pCMX-SMRT expression vectors were provided by K. Umesono and R. M. Evans (Salk Institute, San Diego, CA) (3,12). The VP16 constructs for VDR (92427), TRb1 (174-461) and RXRa (199-462) contain the ligand binding domain (LBD) of the receptors downstream of the VP16 activation domain in-frame in pCMX. The pCMX-NCoR expression vector was provided by M. G. Rosenfeld (University of California, San Diego, CA) (7). The SRC-1 cDNA was provided by B. W. O’Malley (Baylor College of Medicine, Houston, TX) (16). The Gal4-NCoR (residues 1552-2453) and Gal4-SRC1 (residues 213-1061) constructs contain the indicated TR interaction domains of these proteins fused downstream of the Gal4 DNA binding domain (DBD) in pSG424 (17). The Gal4 reporter plasmid, UAS-E1BTATA-Luc (18), contains five copies of the UAS element upstream of E1BTATA in pA3-Luc. The osteopontin VDRE-tk-Luc contains three copies of the mouse osteopontin VDRE (19) upstream of a 151 base pair truncated thymidine
358
Vol. 253, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. Silencing ability of nuclear receptors and their interactions with CoRs. The format of the mammalian two-hybrid experiment is shown at the top of the figure. The indicated Gal4 constructs (100 ng) were cotransfected into TSA-201 cells with 300 ng of VP16 constructs together with 100 ng of the reporter gene, UAS-E1BTATA-Luc, in the absence of ligand. (A) Basal expression without VP16 constructs. (B) Interaction with SMRT. (C) Interaction with NCoR. The numbers at the top of each bar indicate fold stimulation mediated by the interaction with VP16-CoR relative to VP16 alone. kinase promoter, inserted between the SacI and HindIII sites of the pGL3 plasmid (Promega, Madison, WI). Transient expression assays. TSA-201 cells, a clone of human embryonic kidney 293 cells (20), and CV1 cells, were maintained in Dulbecco’s modified essential medium (DMEM) containing 10% fetal bovine serum (FBS), penicillin (100 U/mL) and streptomycin (100 mg/mL). Cells were grown in serum-free DMEM supplemented with Nutridoma (Boehringer Mannheim, Indianapolis, IN) and were transfected by the calcium phosphate method as described (21). The total amount of expression plasmid DNA was kept constant in the different experimental groups by adding corresponding amounts of the same plasmids without cDNA inserts. After exposure to the calcium phosphate-DNA precipitate for 8h, DMEM with 1% Nutridoma was added, in the absence or presence of 1, 25-(OH)2 D3 (Calbiochem, San Diego, CA) (100 nM), T3 (1 mM) and/or 9cis-retinoic acid (100 nM) (ICN, Costa Mesa, CA). Cells were harvested after 40h for measurements of luciferase activity (22). Results are expressed as the mean 6 SEM from at least three transfections performed in triplicate.
RESULTS Gal4-VDR can silence transcription and interact with CoRs. The LBD of VDR, TRß or RXRa was fused to the DBD of the yeast transcription factor, Gal4. The reporter gene, UAS-E1BTATA-Luc, contains five Gal4 binding sites and was used to assess the silencing ability of the receptors in the absence of cognate ligands (Fig. 1A). Relative to the Gal4 DBD alone, Gal4-TR showed 75% inhibition in TSA-201 cells. Gal4RXR resulted in 50% inhibition under these conditions. The silencing ability of Gal4-VDR was intermediate between Gal4-TR and Gal4-RXR.
Interactions between nuclear receptors and CoRs were examined using a mammalian two-hybrid assay. The LBD of VDR, TRß or RXRa was fused to the Gal4-DBD to target these proteins to the reporter gene. The carboxyl terminal half of the SMRT or NCoR proteins, which contain the two TR interaction domains, was fused to the transcriptional activation domain of VP16. UAS-E1BTATA-Luc was used to assess in vivo interactions between the Gal4-receptors and VP16CoR. In this manner, interactions between TR and the CoRs result in the transcriptional activation of the reporter by the VP16 domain (Fig. 1A). Gal4-TR was used initially to test the interaction assay. Relative to the Gal4 DBD alone, Gal4-TR was stimulated 149-fold by the addition of VP16-SMRT (Fig. 1B). Gal4-RXR was stimulated 15-fold by VP16-SMRT. The interaction of Gal4-VDR with VP16-SMRT (91-fold) was stronger than that of Gal4-RXR, but weaker than that of Gal4-TR. NCoR interactions were examined with receptors in a similar manner (Fig. 1C). Strong NCoR interactions were detected for TR (1437-fold), whereas the interaction with RXR was less pronounced (10fold). In the case of VDR, Gal4-VDR was stimulated 68-fold by VP16-NCoR. These results indicate that the silencing ability of nuclear receptors correlates with CoR interactions with SMRT and NCoR. The interactions between nuclear receptors with CoRs were also examined using the opposite configuration of
359
Vol. 253, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. Interactions of receptors with CoRs or a CoA, SRC1. The format of the mammalian two-hybrid experiment is shown at the top of the figure. The indicated Gal4 constructs (100 ng) were cotransfected into TSA-201 cells with 300 ng of VP16 constructs together with 100 ng of the reporter gene, UAS-E1BTATA-Luc, in the absence or presence of ligand. (A) Interaction with SMRT. (B) Interaction with NCoR. (C) Interaction with SRC1. The numbers on the top of each bar indicate fold stimulation mediated by the interaction with VP16-cofactor relative to VP16 alone.
fusion proteins (Fig. 2). The C-terminal half of SMRT, NCoR, or the SRC1 coactivator (CoA) protein was fused to the Gal4-DBD to target these proteins to the reporter gene. The LBD of VDR, TRb or RXRa was fused to VP16 to induce transcription when these proteins are recruited to the Gal4 proteins. In this configuration, the effect of the ligand on protein interactions can be assessed, independent of any effects on the Gal4 proteins. Relative to VP16 alone, Gal4-SMRT was stimulated 414-fold by the addition of VP16-TR and this interaction was markedly decreased by the addition of T3, consistent with the known effect of ligand to disrupt TR-CoR interactions (Fig. 2A). In contrast, VP16-VDR stimulated the activity of Gal4-SMRT only 16-fold and the interaction was almost eliminated by the addition of 1, 25-dihydroxyvitamin D3 (VD3). Gal4-NCoR was stimulated 2008-fold by VP16-TR, but only 6-fold by VP16-VDR (Fig. 2B). Again, cognate ligands decreased the interactions between NCoR and the receptors. The coactivator interaction was also examined in this assay (Fig. 2C). In the presence of ligand, Gal4SRC1 was stimulated 100-fold by the addition of VP16-TR and 259-fold by VP16-VDR, indicating that CoA interaction with VDR is stronger than with TR. These findings also confirm that the VP16-VDR construct is fully functional. VDR can heterodimerize with RXR, but not with TR. The receptor-receptor interactions were also examined using a mammalian two-hybrid assay (Fig. 3). Relative
to VP16 alone, Gal4-RXR was stimulated 820-fold by the addition of VDR and 503-fold by VP16-TR, but not with VP16-RXR. Using the opposite combinations, Gal4-VDR was stimulated 456-fold by the addition of VP16-RXR, and Gal4-TR was stimulated 1228-fold by VP16-RXR. No interaction was observed between TR and VDR, consistent with the recent findings of RavalPanda et al. (23). The silencing ability of VDR in the context of a native VDRE. The silencing ability of native VDR was also examined using a reporter gene that contains a vitamin D response element (VDRE). Since the reporter gene, VDRE-tk-Luc, was stimulated by VD3 without cotransfection of VDR expression plasmids in TSA201 cells (data not shown), VDR-negative CV1 cells were used to test the silencing activity of native VDR (14). When VDR was cotransfected with VDRE-tk-Luc, basal transcription was repressed (2-fold) in the absence of VD3 (Fig. 4A). Interestingly, unliganded RXR or TR similarly repressed VDRE mediated transcription, even though this response element is not optimized for binding by these factors. In the presence of VD3, transcription was stimulated only in the presence of cotransfected VDR (Fig. 4B). The RXR ligand, 9 cis-retinoic acid (9cRA), stimulated RXR-mediated activity by 2-fold, whereas no synergistic effect was seen in the presence of liganded VDR and liganded RXR. In contrast, T3 had no effect on the TR-mediated silencing of this reporter gene.
360
Vol. 253, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 3. Mutual interactions among receptors. The format of the mammalian two-hybrid experiment is shown at the top of the figure. The indicated Gal4 constructs (100 ng) were cotransfected into TSA-201 cells with 300 ng of VP16 constructs together with 100 ng of the reporter gene, UAS-E1BTATA-Luc, in the absence of ligand. The numbers at the top of each bar indicate fold-stimulation mediated by the receptor interaction relative to VP16 alone.
DISCUSSION Basal repression by TR and RAR has been demonstrated in the absence of cognate ligand (4,5,24). VDR belongs to the same subfamily of the nuclear hormone receptors that heterodimerize with RXR and mediate ligand-dependent gene transcription. Although VDR also binds to similar response elements (DR3) as those used by TR (DR4) and RAR (DR5), the silencing ability of VDR is still controversial. Yen et al. (14), but not Cheskis and Freedman (25), observed repressed transcription by unliganded VDR using the osteopontin VDRE in CV1 cells. Masuyama et al. (15) also did not observe the silencing effect using the osteocalcin VDRE in either COS7 cells or CV1 cells. It has been shown that CoR proteins, such as NCoR and SMRT, mediate the transcriptional silencing by TR and RAR (7,10). The putative CoR binding domain, in the hinge region of TR and RAR (termed CoR box) (7), is well-conserved in the VDR. Critical residues for CoR interaction have been identified in another subdomain in v-erb A and TRb (26), and these amino acids are also conserved in the VDR. We have found that leucine 454 of the AF2 domain of TRb exerts inhibitory effects on CoR interactions (27), and this residue is also present in VDR.
Given these conserved structural features in the VDR, one might anticipate interactions with CoRs. Although several studies have examined the interactions TR and RAR with CoRs, there is little information about whether VDR binds to these proteins. In the case of NCoR, Horlein et al. (7) showed, using a yeast two hybrid system, that NCoR interacted with TR and RAR, but not with RXRg, VDR, estrogen receptor (ER), or glucocorticoid receptor (GR). In the case of SMRT, it has been reported that it interacts with RAR, TR and, to lesser degree, with RXR, but not with VDR, using mammalian two-hybrid assays (12). However, Li et al. (13) demonstrated later that SMRT interacts weakly with VDR using yeast two-hybrid assays. We showed here that VDR does interact with NCoR and SMRT, but these interactions are weaker than that of TR. Moreover, there is a potential selectivity in the interaction of receptors with the two different CoRs. TR interacts preferentially with NCoR, whereas VDR interacts better with SMRT in our mammalian two-hybrid assays. These features were seen independent of the configuration of the two-hybrid assay. Differential expression of CoRs in various tissues may therefore influence the silencing activity of coexisting receptors (7,12).
361
Vol. 253, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 4. Transcriptional properties of receptors using a reporter that contains the osteopontin VDRE. The reporter gene is shown at the top of the figure. The indicated expression plasmids (10 ng) were transfected into CV1 cells together with 100 ng of the reporter gene, VDRE-tk-Luc (A and B) in the absence or presence of ligand.
Some antagonists have been shown to recruit CoRs to ER (28,29), or to the progesterone receptor (PR) (30), which otherwise do not bind to CoRs in the absence of ligands. Therefore, it is possible that the identification of antagonists for the VDR may provide ligands that could enhance its binding to CoRs and increase silencing. It is of interest that mutations in the human VDR are inherited as autosomal recessive traits (31,32), whereas those in the TR transmitted primarily in an autosomal dominant manner (33). These features probably reflect the fact that the mutant TRs can inhibit the function of the wild type allele in a dominant negative manner, whereas the mutant VDRs are functionally inactive. We (34) and others (35) have previously shown that CoR binding to mutant TRs is necessary for their dominant negative activities. The present study, demonstrating relatively weak interactions between VDR and CoRs, suggests that dominant negative forms of VDR might not exist, or that they might only weakly inhibit VDR-responsive genes. This idea could be tested by generating artificial mutations in the VDR that selectively lack ligand binding or transcriptional activation. Cross-talk between VDR and TR has been suggested because they coexist in some tissues such as pituitary, brain, intestine, skin, thyroid, kidney and bone (36). Several mechanisms have been suggested, including heterodimerization between VDR and TR (37), sequestration of RXR (23), and competition for binding to DNA (14). We observed that TR did not heterodimerize
with VDR in a mammalian two-hybrid assay, but TR was able to weakly silence the basal expression of a VDRE reporter gene. Moreover, T3 did not relieve this silencing, suggesting that TR may act as a constitutive repressor of a VD3 responsive gene. Such differential effects by RAR have been observed on different RAR response elements (8). Further investigation is necessary for understanding the potential cross-talk among these nuclear receptors. In summary, we find that VDR exhibits relatively weak transcriptional silencing, consistent with reduced interactions with NCoR and SMRT. However, it heterodimerizes well with RXR and it binds a coactivator, SRC-1, very well in the presence of ligand. These findings suggest that, in comparison to related family members, like TR and RAR, the function of the VDR is biased towards transcriptional activation as opposed to transcriptional repression. ACKNOWLEDGMENTS We are grateful to M. G. Rosenfeld, R. M. Evans, and B. W. O’Malley, A. for providing plasmids. This work was supported by National Institute of Health (NIH) Grants DK42144 (to J.L.J) and DK25409 (to R. K.).
REFERENCES
362
1. Evans, R. M. (1988) Science 240(4854), 889 – 895. 2. Glass, C. (1994) Endocr. Rev. 15, 391– 407.
Vol. 253, No. 2, 1998
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
3. Umesono, K., Murakami, K. K., Thompson, C. C., and Evans, R. M. (1991) Cell 65, 1255–1266. 4. Brent, G. A., Dunn, M. K., Harney, J. W., Gulick, T., Larsen, P. R., and Moore, D. D. (1989) New Biol. 1(3), 329 –336. 5. Damm, K., Thompson, C. C., and Evans, R. M. (1989) Nature 339(6226), 593–597. 6. Baniahmad, A., Kohne, A. C., and Renkawitz, R. (1992) EMBO J. 11(3), 1015–1023. 7. Horlein, A. J., Naar, A. M., Heinzel, T., Torchia, J., Gloss, B., Kurokawa, R., Ryan, A., Kamei, Y., Soderstrom, M., Glass, C. K., and Rosenfeld, M. G. (1995) Nature 377(6548), 397– 404. 8. Kurokawa, R., Soderstrom, M., Horlein, A., Halachmi, S., Brown, M., Rosenfeld, M. G., and Glass, C. K. (1995) Nature 377(6548), 451– 454. 9. Lee, J. W., Ryan, F., Swaffield, J. C., Johnston, S. A., and Moore, D. D. (1995) Nature 374(6517), 91–94. 10. Chen, J. D., and Evans, R. M. (1995) Nature 377(6548), 454 – 457. 11. Sande, S., and Privalsky, M. L. (1996) Mol. Endocrinol. 10, 813– 825. 12. Chen, J. D., Umesono, K., and Evans, R. M. (1996) Proc. Natl. Acad. Sci. U.S.A. 93, 7567–7571. 13. Li, H., Leo, C., Schroen, D. J., and Chen, J. D. (1997) Mol. Endocrinol. 11, 2025–2037. 14. Yen, P. M., Liu, Y., Sugawara, A., and Chin, W. W. (1996) J. Biol. Chem. 271, 10,910 –10,916. 15. Masuyama, H., Jefcoat, S. C., and MacDoonald, P. N. (1997) Mol. Endocrinol. 11, 218 –228. 16. Onate, S. A., Tsai, S. Y., Tsai, M. J., and O’Malley, B. W. (1995) Science 270(5240), 1354 –1357. 17. Sadowski, I., Ma, J., Tiezenberg, S., and Ptashne, M. (1988) Nature 335, 563–564. 18. Pestell, R. G., Albanese, C., Watanabe, G., Lee, R. J., Lastowiecki, P., Zon, L., Ostrowski, M., and Jameson, J. L. (1996) Mol. Endocrinol. 10, 1084 –1094. 19. Veenstra, T. D., Benson, L. M., A., C. T., Tomlinson, A. J., Kumar, R., and Naylor, S. (1998) Nature Biotechnol. 16, 262– 266.
20. Margolskee, R. F., McHendry-Rinde, B., and Horn, R. (1993) Biotechniques 15(5), 906 –911. 21. Tagami, T., Madison, L. D., Nagaya, T., and Jameson, J. L. (1997) Mol. Cell Biol. 17(5), 2642–2648. 22. deWet, J. R., Wood, K. V., DeLuca, M., and Helinski, D. R. (1987) Mol. Cell Biol. 7, 725–737. 23. Raval-Panda, M., Freedman, L. P., Li, H., and Christakos, S. (1998) Mol. Endocrinol. 12, 1367–1379. 24. Baniahmad, A., Steiner, C., Kohne, A. C., and Renkawitz, R. (1990) Cell 61(3), 505–514. 25. Cheskis, B., and Freedman, L. P. (1994) Mol. Cell Biol. 14, 3329 –3338. 26. Busch, K., Martin, B., Renkawitz, R., and Muller, M. (1997) Mol. Endocrinol. 11, 379 –389. 27. Tagami, T., Gu, W., Peairs, P. T., West, B. L., and Jameson, J. L. Mol. Endocrinol. (in press). 28. Smith, C. L., Nawaz, Z., and O’Malley, B. W. (1997) Mol. Endocrinol. 11, 657– 666. 29. Jackson, T. A., Richer, J. K., Bain, D. L., Takimoto, G. S., Tung, L., and Horowitz, K. B. (1997) Mol. Endocrinol. 11, 693–705. 30. Zhang, X., Jeyakumar, M., Petukhov, S., and Bagchi, M. (1998) Mol. Endocrinol. 12, 513–524. 31. Hughes, M. R., Malloy, P. J., Kieback, D. G., Kesterson, R. A., Pike, J. W., Feldman, D., and O’Malley, B. W. (1988) Science 242, 1702–1705. 32. Whitefield, G. K., Selznick, S. H., Haussler, C. A., Hsieh, J.-C., Galligan, M. A., Jurtka, P. W., Thompson, P. D., Lee, S. M., Zerwekh, J. E., and Haussler, M. R. (1996) Mol. Endocrinol. 10, 1617–1631. 33. Refetoff, S., Weiss, R. E., and Usala, S. J. (1993) Endocr. Rev. 14(3), 348 –399. 34. Tagami, T., and Jameson, J. L. (1998) Endocrinology 139, 640–650. 35. Yoh, S. M., Chatterjee, V. K. K., and Privalsky, M. L. (1997) Mol. Endocrinol. 11, 470 – 480. 36. Walters, M. R. (1992) Endocr. Rev. 13, 1– 46. 37. Schrader, M., Muller, K. M., Nayeri, S., Kahlen, J. P., and Carlberg, C. (1994) Nature 370(6488), 382–386.
363