- Email: [email protected]
Transcriptional Regulation by Competition between ELP Isoforms and Nuclear Receptors
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS ARTICLE NO.
230, 407–412 (1997)
RC965972
Transcriptional Regulation by Competition between ELP Isoforms and Nuclear Receptors Naoe Kotomura,1 Yasuharu Ninomiya, Kazuhiko Umesono,* and Ohtsura Niwa Department of Molecular Pathology, Research Institute for Radiation Biology and Medicine, Hiroshima University, 1-2-3 Kasumi, Mimami-ku, Hiroshima 734, Japan; and *Molecular and Developmental Biology, Graduate School of Biological Sciences, Nara Institute of Science and Technology, 8916-5 Takayama, Ikoma, Nara 630-01, Japan
Received December 3, 1996
ELP is a transcription factor belonging to the nuclear receptor superfamily. The consensus binding sequence for ELP contains a half site of the nuclear receptor recognition element. We demonstrated previously that ELP1, the repressor type isoform of ELP, competes for binding with the retinoic acid receptor and represses retinoic acid-induced transactivation. In this study, competitive repression by ELP1 was investigated for several other nuclear receptors. As in the case of the retinoic acid receptor, binding of vitamin D receptor, thyroid hormone receptor, and estrogen receptor could be competed by ELP1, resulting in repression of their ligand-dependent transactivation. Interestingly, the activator-type ELP isoforms were capable of repressing retinoic acid-induced transactivation through binding to the retinoic acid receptor binding element. These data suggest that competition for target DNA binding is a general mechanism of transcriptional repression by ELP isoforms. q 1997 Academic Press
In undifferentiated embryonal carcinoma (EC) cells, replication of a variety of viruses is repressed. Moloney murine leukemia virus has been used as a model to analyze this repression (1-9). ELP, the embryonal long terminal repeat (LTR) binding protein, was identified as a repressor which binds to the negative regulatory region of the Moloney murine leukemia virus LTR in EC cells (10). cDNA cloning and structure analysis revealed that ELP is a mouse homologue of Drosophila FTZ-F1 and a member of monomeric orphan receptors of the nuclear receptor superfamily (11, 12). The consensus sequence for ELP binding, YCAAGGYCR, was identified through analysis of cellular targets of ELP (13). This consensus sequence contains a half site of the nuclear receptor binding element. ELP represses 1 To whom correspondence should be addressed. Fax: 81-82-2567102.
the transcription of reporter plasmids driven by the Moloney murine leukemia virus LTR as well as those containing chromosomal binding sites for ELP (13). As for the mechanism of repression by ELP, competitive binding between ELP and transactivators was suggested, because ELP repressed retinoic acid (RA) induced transactivation by competing for binding to retinoic acid responsive elements (14). The ELP gene has been mapped on mouse chromosome 2 (15). Extensive analysis has revealed four isoforms of ELP which included ELP1, ELP2, ELP3 and Ad4BP/SF-1 (16). These isoforms are produced by alternative promoter usage and differential splicing of the ELP gene. ELP1 is the original ELP isolate and functions as a repressor. On the other hand, ELP2, ELP3 and Ad4BP/SF-1 function as activators (16). In this study, we have demonstrated that ELP1 could repress transcriptional activities of a variety of nuclear receptors by competitive binding to the target elements, and activator type ELP isoforms also repress RA-induced transactivation in the same manner as ELP1 does. MATERIALS AND METHODS Cell lines. NIH3T3 and COS7 cells were described previously (14). NIH3T3 cells were used for CAT assay, and COS7 cells were used for preparation of nuclear extracts after transfection with plasmids carrying nuclear receptor genes. Oligonucleotide elements. Four oligonucleotide elements designated as DR3/ELP site, DR4/ELP site, DR5/ELP site and ERE/ELP site were synthesized. The sequences of these oligonucleotides are shown in Fig. 1. Underline indicates the binding sequence for ELP (TCAAGGTCA), and arrow indicates the half site for nuclear receptor binding (AGGTCA). In this study, four nuclear receptors, vitamin D receptor (VDR), thyroid hormone receptor (TR), retinoic acid receptor (RAR) and estrogen receptor (ER), were tested. They bind to the target sequences as hetero- or homo-dimers. The binding specificity of the sequence elements was determined by the orientation and the spacing between two half sites; VDR binds to direct repeats separated by 3 nucleotides (DR3), TR binds to DR4, RAR binds to DR5, ER binds to an inverted repeat separated by 3 nucleotides (17, 18). Therefore, it 0006-291X/97 $25.00
407
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
BBRC 5972
/
6916$$$761
12-24-96 01:49:42
bbrcg
AP: BBRC
Vol. 230, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
is predicted that DR3/ELP site is the target for ELP and VDR binding, DR4/ELP site for ELP and TR, DR5/ELP site for ELP and RAR, and ERE/ELP site for ELP and ER. Reporter plasmids. pDSVCAT carrying the CAT reporter gene downstream of the SV40 early promoter was used to construct various reporter plasmids of the present experiment (19). One copy of each oligonucleotide element was inserted at the Hind III site 5* to the SV40 early promoter, generating pDR3/ELP-DSVCAT, pDR4/ ELP-DSVCAT, pDR5/ELP-DSVCAT and pERE/ELP-DSVCAT. Effector plasmids. pCMX carrying the CMV promoter and the poly(A) signal of the SV40 T antigen was used to construct effector plasmids (19). The coding sequences for firefly luciferase (LUC), human VDR, human TRb, human RARa, human ER and ELP isoforms, ELP1, ELP2 and ELP3, and the truncated ELP mutant, E-ZF, were placed at the polylinker site between the promoter and the poly(A) signal, generating pCMX-LUC, pCMX-hVDR, pCMX-hTRb, pCMXhRARa, pCMX-hER, pCMX-ELP1, pCMX-ELP2, pCMX-ELP3 and pCMX-E-ZF, respectively (16, 19, 20). pCMX-LUC was used as the negative control. pTB-1 which carries the bacterial b-galactosidase gene downstream of the RSV promoter, was used as an internal control to monitor the efficiency of transfection. CAT assay. NIH3T3 cells were seeded at a density of 21105 cells per 6-cm plate 24 h prior to transfection. Ten micrograms of the plasmid mixture was transfected into the cells. The mixture contained 2 mg of a reporter plasmid, 5, 15 or 100 ng of an effector plasmid, and 2 mg of pTB-1 as an internal control. pUC119 was used as a carrier to adjust the total amount of DNA to 10 mg. Transfection was performed by the calcium-phosphate precipitation method (10). Cells were incubated for 12 h with the precipitate, the medium was then replaced and cultures were further incubated for 24 h with or without the ligand. Final concentrations of each ligand were 100 nM for VD3 , T3 and estradiol, and 1 mM for RA. Thereafter, the cells were harvested and enzymatic activity in cell extracts was measured by using the standard procedure. Radioactivity of reacted products was quantified by use of an imaging analyzer (BAS2000, Fuji). CAT assay was repeated at least twice for each experiment.
FIG. 2. Binding specificity of ELP1 and nuclear receptors. [a-32P]dCTP-labeled DR3/ELP site, DR4/ELP site, DR5/ELP site, and ERE/ ELP site were used as probes. In each reaction, nuclear extract containing 1.5 mg protein was used. C, E, V, T, R, and ER denote nuclear extracts of COS7 cells transfected with none, pCMX-ELP1, pCMXhVDR, pCMX-hTRb, pCMX-hRARa, and pCMX-hER, respectively.
ligand were 100 nM for VD3 , T3 and estradiol. Thereafter, the cells were harvested and nuclear extracts were prepared. Procedures for the preparation of nuclear extracts and the gel shift assay were described previously (6, 19).
RESULTS
Gel shift assay. COS7 cells were transfected with effector plasmids, and those nuclear extracts were used in the gel shift assay. Briefly, COS7 cells were seeded at a density of 7.51105 cells per 10-cm plate 24 h prior to transfection. Four or 5 plates were prepared for each sample. Twenty micrograms of the effector plasmid was transfected into the cells by the calcium-phosphate precipitation method. Cells were incubated for 24 h with the precipitate, the medium was then replaced and cultures were further incubated for 24 h. Three hours before the end of incubation, appropriate ligands were added to the culture transfected with pCMXhVDR, pCMX-hTRb and pCMX-hER. Final concentrations of each
Binding of ELP1 and nuclear receptors to the oligonucleotide elements. Oligonucleotide elements shown in Fig. 1 were used as probes, and the binding of ELP1 and other nuclear receptors to these elements was tested by gel shift assay using nuclear extracts of COS7 cells transfected with effector plasmids. As predicted from the sequences, these oligonucleotide elements served as efficient targets for binding of nuclear receptors; DR3/ELP site was bound by ELP1 and VDR (Fig. 2, lane 2 and 3), DR4/ELP site by ELP1 and TRb (Fig. 2, lane 5 and 6), DR5/ELP site by ELP1 and RARa (Fig. 2, lane 8 and 9), and ERE/ELP site by ELP1 and ER (Fig. 2, lane 11 and 12).
FIG. 1. Sequences of oligonucleotides. Each oligonucleotide has the Hind III linker on both sides. An arrow indicates the half site of direct or inverted repeats, and an underline indicates the binding sequence for ELP.
Ligand induced transactivation and its repression by ELP1. To test whether each oligonucleotide elements function as ligand response elements, reporter plasmids carrying one copy of DR3/ELP site, DR4/ELP site, DR5/ELP site or ERE/ELP site were transfected into NIH3T3 cells together with effector plasmids coding for LUC or target nuclear receptors. Cultures were treated with corresponding ligands and the CAT activities were analyzed. Data are shown in Fig. 3A. DR3/ELP site and DR5/ELP site responded to their ligands efficiently without co-transfection of VDR or RARa. We assumed that this was due to the presence of endogenous VDR and RAR. In the case of DR4/ELP site and ERE/ELP site, corresponding ligands activated transcription when co-transfected with pCMX-hTRb and pCMX-
408
AID
BBRC 5972
/
6916$$$762
12-24-96 01:49:42
bbrcg
AP: BBRC
Vol. 230, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 3. Ligand-induced transactivation and its repression by ELP1. (A) Two micrograms of reporter plasmids (pDR3/ELP-DSVCAT, pDR4/ELP-DSVCAT, pDR5/ELP-DSVCAT, and pERE/ELP-DSVCAT) was transfected into NIH3T3 cells together with effector plasmids. Nuclear receptor minus cultures were those transfected with pCMX-LUC. Amounts of each effector plasmid were 100 ng of pCMX-LUC, pCMX-hVDR, and pCMX-hRARa, 15 ng of pCMX-hTRb, and 5 ng of pCMX-hER. After incubation for 24 h with or without addition of ligands (VD3 , T3 , and estradiol, 100 nM; RA, 1 mM), CAT activities were measured. Each CAT activity was normalized to that of ligand and nuclear receptor minus culture. (B) Two micrograms of reporter plasmids was transfected into NIH3T3 cells with or without 100 ng of pCMX-ELP1. pDR4/ELP-DSVCAT and pERE/ELP-DSVCAT were analyzed under co-transfection of 15 ng of pCMX-hTRb and 5 ng of pCMX-hER, respectively. In ELP1 minus cultures of pDR3/ELP-DSVCAT and pDR5/ELP-DSVCAT, 100 ng of pCMX-LUC was co-transfected. Culture conditions and measurement of CAT activities are the same as those of (A). Each CAT activity was normalized to that of ligand and ELP1 minus culture.
hER, respectively. These data indicate that DR3/ELP site functions as the VD3 response element, DR4/ELP site as the T3 response element, DR5/ELP site as the RA response element, and ERE/ELP site as the estradiol response element. We have reported previously that ELP1 represses RA-induced transactivation by competition with RARa (14). Similarly, ELP1 may function as a repressor to other nuclear receptors and represses ligand-induced transactivation. A series of reporter plasmids used in Fig. 3A were transfected into NIH3T3 cells together with or without pCMX-ELP1, and cultures were treated with corresponding ligands. Since the ligand response of DR4/ELP site and ERE/ELP site required the presence of pCMX-hTRb and pCMX-hER, respectively (Fig. 3A), those effector plasmids were co-transfected when testing the effect of ELP1. Data are shown
in Fig. 3B. In DR3/ELP site, ELP1 repressed VD3-induced transactivation significantly (2.1 fold repression), and repression by ELP1 was little when VD3 was absent (1.4 fold repression). Similarly ELP1 specifically repressed T3-induced transactivation in DR4/ELP site (2.3 and 1.3 fold repression with or without T3). In DR5/ ELP site and ERE/ELP site, ELP1 repressed transcription under ligand free condition (1.7 and 2.0 fold repression, respectively), and the repression was more significantly when ligand were present (2.7 and 4.0 fold repression, respectively). Repression of RA-induced transactivation by ELP1 of the present analysis was consistent with the previous report (14). In addition, ELP1 was found to repress transcriptional activation by VDR, TR and ER. ELP isoforms repress RA-induced transactivation by competitive binding. Isoforms of ELP, ELP1, ELP2,
409
AID
BBRC 5972
/
6916$$$762
12-24-96 01:49:42
bbrcg
AP: BBRC
Vol. 230, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 4. Competitive binding and repression of DR5/ELP site by ELP isoforms and a truncated mutant. (A) Structures of ELP isoforms and the ELP truncated mutant. ZF and F1 indicate the zinc finger region and the FTZ-F1 box, respectively (26). (B) Binding specificity of ELP isoforms and a truncated mutant to DR5/ELP site. [a-32P]dCTP-labeled DR5/ELP site was used as a probe. In each reaction, nuclear extract containing 1.5 mg protein was included. C, E, 2, 3, Z, and R denote nuclear extracts of COS7 cells transfected with none, pCMXELP1, pCMX-ELP2, pCMX-ELP3, pCMX-E-ZF, and pCMX-hRARa, respectively. (C) Effector plasmids (100 ng each) were transfected into NIH3T3 cells with 2 mg of pDR5/ELP-DSVCAT. After incubation for 24 h with or without addition of 1 mM RA, CAT activities were measured. Each CAT activity was normalized to that of pCMX-LUC with addition of RA. (D) Dose-dependent competition of E-ZF to RARaDNA complex. [a-32P]dCTP-labeled DR5/ELP site was used as probe. In each reaction, 2 mg of nuclear extracts of COS7 cells transfected with none (C) or pCMX-hRARa (R) was included. Nuclear extract of COS7 cells transfected with pCMX-E-ZF (Z) was added into the above reactions (from 0.8 ng to 2 mg).
ELP3 and Ad4BP/SF-1, all bound to the same sequence element, yet ELP1 functioned as a repressor and other three as activator. The activator type ELP isoforms are also expected to compete for binding to the sequence elements of nuclear receptors when the elements contain the consensus sequence of ELP binding. Therefore, we have studied the effect of ELP isoforms on RAR mediated transactivation by using DR5/ELP site as the RA response element. Fig. 4A shows the structures of three ELP isoforms, ELP1, ELP2 and ELP3, and a truncated mutant of ELP, E-ZF. E-ZF contains only the DNA binding region of the gene. Binding activity of those ELP isoforms and E-ZF to DR5/ELP site was analyzed. As is clear from Fig. 4B, ELP2, ELP3 and E-ZF formed complexes with DR5/ELP site as did ELP1 and RARa. To test the effect of ELP isoforms and E-ZF on RAR mediated transactivation, effector plasmids carrying ELP isoforms or E-
ZF were transfected into NIH3T3 cells together with pDR5/ELP-DSVCAT. Data are shown in Fig. 4C. ELP2 and ELP3 did not activate the transcription under RA free condition. Furthermore, they repressed RA-induced transactivation of DR5/ELP site. There was no significant difference in the magnitude of repression among ELP isoforms (2.6, 2.5 and 2.3 fold repression by ELP1, ELP2 and ELP3, respectively). In addition, the truncated mutant of ELP, E-ZF, also repressed RAinduced transactivation. Above data suggest that ELP isoforms compete with RARa and repressed RA-induced transactivation, and DNA binding region of ELP is sufficient for the competitive repression. To confirm this suggestion, it was analyzed whether the complex formation between DR5/ELP site and RARa was abrogated by E-ZF. The binding of RARa to DR5/ELP site was indeed competed in proportion to the dose of addition of E-ZF (Fig. 4D).
410
AID
BBRC 5972
/
6916$$$762
12-24-96 01:49:42
bbrcg
AP: BBRC
Vol. 230, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
DISCUSSION In this study, we showed that ELP1 could compete with nuclear receptors, VDR, TR, RAR and ER, and repressed their ligand-dependent transactivation (Fig. 3B). ELP1 lacked the region III, the putative activator domain conserved among nuclear receptors, and functioned as a repressor (11). On the other hand, ELP3 possessed that region, and functioned as an activator (16). Truncated mutant of ELP3 which lacks the region III could not activate transcription (data not shown). These findings suggest that ELP1 is an inactive form, and competitive binding is a general mechanism of the ELP1 mediated repression. ELP3-induced transactivation was repressed by ELP1 efficiently (data not shown), and ELP1 was expressed with other ELP isoforms in cell lines and tissues (16). Therefore, it is possible that the major targets of ELP1 are other ELP isoforms, and ELP1 modulates the transactivation by activator type ELP isoforms. Such competitive binding was reported for Drosophila FTZ-F1 in which FTZ-F1a and FTZ-F1b compete for binding to a site in the fushi tarazu gene (21). ELP3 is identical in amino acid sequence with mouse SF-1 and bovine Ad4BP (16). SF-1 and Ad4BP are known as key regulators of steroidgenic P-450 genes, and activate the transcription of those genes in adrenal cells (22, 23). ELP2 shares the same N-terminus as ELP1, but carries the C-terminal transactivation domain as in the case of ELP3 (16). ELP2 and ELP3 activated transcription of a reporter plasmid carrying 8 copies of the ELP binding element of the Moloney leukemia virus LTR (16), but a single copy of that element was not sufficient for transcriptional activation (data not shown). In addition, ELP2 and ELP3 did not activate the transcription of reporter plasmid carrying a single copy of DR5/ELP site (Fig. 4C). A similar phenomenon was reported on the Xenopus homologue of FTZ-F1 (24). For transcriptional activation of steroidgenic P-450 genes, several elements are need in addition to Ad4BP binding element (25). It is likely that ELP isoforms bind to the target elements as a monomer, because FTZ-F1 binds to the target sequence as a monomer (26). These findings raise an interesting possibility that transcriptional activation mediated by the ELP isoforms is strictly dependent upon the context of the target promoter and other transcription factors available in a given cell. In contrast to the mechanism of transactivation, repression by ELP is mediated simply by competition. ELP2 and ELP3 repressed RA-induced transactivation on DR5/ELP site by competing for the binding site as in the case of ELP1 and RARa (Fig. 4C). This suggests that even the activator type of ELP isoforms repress transactivation by other nuclear receptors when they share the binding sequence motif. Indeed, there are some data suggesting the negative function of ELP iso-
forms. Although ELP3 was the major form of the isoforms expressed in undifferentiated EC cells, the ELP binding site of the Moloney murine leukemia virus genome nevertheless functioned as a negative element (6, 16). ELP isoforms functioned negatively to ELP binding site of the ELP gene, and this ELP binding site overlapped with the consensus sequences for other transcription factors binding (27). Consensus sequence for ELP binding have been found in several genes, and some of these genes are proposed as the target of ELP isoforms (28, 29). Further analysis of the function of ELP on these genes may verify the roles of ELP isoforms in transactivation and repression. ACKNOWLEDGMENTS We thank Dr. S. Ishii for the pTB-1 plasmid. We also thank Mr. T. Nishioka for photographic work and Ms. T. Matsuura for secretarial work. This work was supported by a Grant-in-Aid from the Ministry of Education, Science, and Culture, Japan.
REFERENCES 1. Niwa, O., Yokota, Y., Ishida, H., and Sugahara, T. (1983) Cell 32, 1105–1113. 2. Linney, E., Davis, E., Overhauser, D. J., Chao, E., and Fan, H. (1984) Nature 308, 470–472. 3. Niwa, O. (1985) Mol. Cell. Biol. 5, 2325–2331. 4. Loh, T. P., Sievert, L. L., and Sctt, R. W. (1990) Mol. Cell. Biol. 10, 4045–4057. 5. Tsukiyama, T., Niwa, O., and Yokoro, K. (1990) Virology 177, 772–776. 6. Tsukiyama, T., Niwa, O., and Yokoro, K. (1991) J. Virol. 65, 2979–2986. 7. Petersen, R., Kempler, G., and Barklis, E. (1991) Mol. Cell. Biol. 11, 1214–1221. 8. Akgun, E., Ziegler, M., and Grez, M. (1991) J. Virol. 65, 382– 388. 9. Yamauchi, M., Freitag, B., Khan, C., Berwin, B., and Barklis, E. (1995) J. Virol. 69, 1142–1149. 10. Tsukiyama, T., Niwa, O., and Yokoro, K. (1989) Mol. Cell. Biol. 9, 4670–4676. 11. Tsukiyama, T., Ueda, H., Hrise, S., and Niwa, O. (1992) Mol. Cell. Biol. 12, 1286–1291. 12. Mangelsdorf, D. J., Thummel, C., Beato, M., Herrlich, P., Shutz, G., Umesono, K., Blumberg, B., Kastner, P., Mark, M., Chambon, P., and Evans, R. M. (1995) Cell 83, 835–839. 13. Tsukiyama, T., and Niwa, O. (1992) Nucleic Acids Res. 20, 1477– 1482. 14. Kotomura, N., Okada, M., Ninomiya, Y., Tsukiyama, T., Umesono, K., Evans, R. M., and Niwa, O. (1994) J. Biochem. 116, 1309–1316. 15. Taketo, M., Parker, K. L., Howard, T. A., Tsukiyama, T., Wong, M., Niwa, O., Morton, C. C., Miron, P. M., and Seldin, M. F. (1995) Genomics 25, 565–567. 16. Ninomiya, Y., Okada, M., Kotomura, N., Suzuki, K., Tsukiyama, T., and Niwa, O. (1995) J. Biochem. 270, 7453–7461. 17. Mangelsdorf, D. J., and Evans, R. M. (1995) Cell 83, 841–850. 18. Beato, M., Herrlich, P., and Schutz, G. (1995) Cell 83, 851–857. 19. Umesono, K., Murakami, K. K., Thompson, C. C., and Evans, R. M. (1991) Cell 65, 1255–1266.
411
AID
BBRC 5972
/
6916$$$762
12-24-96 01:49:42
bbrcg
AP: BBRC
Vol. 230, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
20. Waterman, M. L., Adler, S., Nelson, C., Greene, G. L., Evans, R. M., and Rosenfeld, M. G. (1988) Mol. Endocrinol. 2, 14–21. 21. Ohno, C. K., Ueda, H., and Petkovich, M. (1994) Mol. Cell. Biol. 14, 3166–3175. 22. Ikeda, Y., Lala, D. S., Luo, X., Kim, E., Moisan, M-P., and Parker, K. L. (1993) Mol. Endocrinol. 7, 852–860. 23. Morohashi, K., Zanger, U. M., Honda, S., Hara, M., Waterman, M. R., and Omura, T. (1993) Mol. Endocrinol. 7, 1196–1204. 24. Ellinger-Ziegelbauer, H., Hihi, A. K., Laudet, V., Keller, H., Wahli, W., and Dreyer, C. (1994) Mol. Cell. Biol. 14, 2786–2797.
25. Takayama, K., Morohashi, K., Honda, A., Hara, N., and Omura, T. (1994) J. Biochem. 116, 193–203. 26. Ueda, H., Sun, G.-C., Murata, T., and Hirose, S. (1992) Mol. Cell. Biol. 12, 5667–5672. 27. Ninomiya, Y., Kotomura, N., and Niwa, O. (1996) Biochem. Biophys. Res. Commun. 222, 632–638. 28. Shen, W.-H., Moor, C. C. D., Ikeda, Y., Parker, K., and Igraham, H. A. (1994) Cell 77, 651–661. 29. Burris, T., Guo, W., Le, T., and McCabe, R. B. (1995) Biochem. Biophys. Res. Commun. 214, 576–581.
412
AID
BBRC 5972
/
6916$$$762
12-24-96 01:49:42
bbrcg
AP: BBRC
230, 407–412 (1997)
RC965972
Transcriptional Regulation by Competition between ELP Isoforms and Nuclear Receptors Naoe Kotomura,1 Yasuharu Ninomiya, Kazuhiko Umesono,* and Ohtsura Niwa Department of Molecular Pathology, Research Institute for Radiation Biology and Medicine, Hiroshima University, 1-2-3 Kasumi, Mimami-ku, Hiroshima 734, Japan; and *Molecular and Developmental Biology, Graduate School of Biological Sciences, Nara Institute of Science and Technology, 8916-5 Takayama, Ikoma, Nara 630-01, Japan
Received December 3, 1996
ELP is a transcription factor belonging to the nuclear receptor superfamily. The consensus binding sequence for ELP contains a half site of the nuclear receptor recognition element. We demonstrated previously that ELP1, the repressor type isoform of ELP, competes for binding with the retinoic acid receptor and represses retinoic acid-induced transactivation. In this study, competitive repression by ELP1 was investigated for several other nuclear receptors. As in the case of the retinoic acid receptor, binding of vitamin D receptor, thyroid hormone receptor, and estrogen receptor could be competed by ELP1, resulting in repression of their ligand-dependent transactivation. Interestingly, the activator-type ELP isoforms were capable of repressing retinoic acid-induced transactivation through binding to the retinoic acid receptor binding element. These data suggest that competition for target DNA binding is a general mechanism of transcriptional repression by ELP isoforms. q 1997 Academic Press
In undifferentiated embryonal carcinoma (EC) cells, replication of a variety of viruses is repressed. Moloney murine leukemia virus has been used as a model to analyze this repression (1-9). ELP, the embryonal long terminal repeat (LTR) binding protein, was identified as a repressor which binds to the negative regulatory region of the Moloney murine leukemia virus LTR in EC cells (10). cDNA cloning and structure analysis revealed that ELP is a mouse homologue of Drosophila FTZ-F1 and a member of monomeric orphan receptors of the nuclear receptor superfamily (11, 12). The consensus sequence for ELP binding, YCAAGGYCR, was identified through analysis of cellular targets of ELP (13). This consensus sequence contains a half site of the nuclear receptor binding element. ELP represses 1 To whom correspondence should be addressed. Fax: 81-82-2567102.
the transcription of reporter plasmids driven by the Moloney murine leukemia virus LTR as well as those containing chromosomal binding sites for ELP (13). As for the mechanism of repression by ELP, competitive binding between ELP and transactivators was suggested, because ELP repressed retinoic acid (RA) induced transactivation by competing for binding to retinoic acid responsive elements (14). The ELP gene has been mapped on mouse chromosome 2 (15). Extensive analysis has revealed four isoforms of ELP which included ELP1, ELP2, ELP3 and Ad4BP/SF-1 (16). These isoforms are produced by alternative promoter usage and differential splicing of the ELP gene. ELP1 is the original ELP isolate and functions as a repressor. On the other hand, ELP2, ELP3 and Ad4BP/SF-1 function as activators (16). In this study, we have demonstrated that ELP1 could repress transcriptional activities of a variety of nuclear receptors by competitive binding to the target elements, and activator type ELP isoforms also repress RA-induced transactivation in the same manner as ELP1 does. MATERIALS AND METHODS Cell lines. NIH3T3 and COS7 cells were described previously (14). NIH3T3 cells were used for CAT assay, and COS7 cells were used for preparation of nuclear extracts after transfection with plasmids carrying nuclear receptor genes. Oligonucleotide elements. Four oligonucleotide elements designated as DR3/ELP site, DR4/ELP site, DR5/ELP site and ERE/ELP site were synthesized. The sequences of these oligonucleotides are shown in Fig. 1. Underline indicates the binding sequence for ELP (TCAAGGTCA), and arrow indicates the half site for nuclear receptor binding (AGGTCA). In this study, four nuclear receptors, vitamin D receptor (VDR), thyroid hormone receptor (TR), retinoic acid receptor (RAR) and estrogen receptor (ER), were tested. They bind to the target sequences as hetero- or homo-dimers. The binding specificity of the sequence elements was determined by the orientation and the spacing between two half sites; VDR binds to direct repeats separated by 3 nucleotides (DR3), TR binds to DR4, RAR binds to DR5, ER binds to an inverted repeat separated by 3 nucleotides (17, 18). Therefore, it 0006-291X/97 $25.00
407
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
BBRC 5972
/
6916$$$761
12-24-96 01:49:42
bbrcg
AP: BBRC
Vol. 230, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
is predicted that DR3/ELP site is the target for ELP and VDR binding, DR4/ELP site for ELP and TR, DR5/ELP site for ELP and RAR, and ERE/ELP site for ELP and ER. Reporter plasmids. pDSVCAT carrying the CAT reporter gene downstream of the SV40 early promoter was used to construct various reporter plasmids of the present experiment (19). One copy of each oligonucleotide element was inserted at the Hind III site 5* to the SV40 early promoter, generating pDR3/ELP-DSVCAT, pDR4/ ELP-DSVCAT, pDR5/ELP-DSVCAT and pERE/ELP-DSVCAT. Effector plasmids. pCMX carrying the CMV promoter and the poly(A) signal of the SV40 T antigen was used to construct effector plasmids (19). The coding sequences for firefly luciferase (LUC), human VDR, human TRb, human RARa, human ER and ELP isoforms, ELP1, ELP2 and ELP3, and the truncated ELP mutant, E-ZF, were placed at the polylinker site between the promoter and the poly(A) signal, generating pCMX-LUC, pCMX-hVDR, pCMX-hTRb, pCMXhRARa, pCMX-hER, pCMX-ELP1, pCMX-ELP2, pCMX-ELP3 and pCMX-E-ZF, respectively (16, 19, 20). pCMX-LUC was used as the negative control. pTB-1 which carries the bacterial b-galactosidase gene downstream of the RSV promoter, was used as an internal control to monitor the efficiency of transfection. CAT assay. NIH3T3 cells were seeded at a density of 21105 cells per 6-cm plate 24 h prior to transfection. Ten micrograms of the plasmid mixture was transfected into the cells. The mixture contained 2 mg of a reporter plasmid, 5, 15 or 100 ng of an effector plasmid, and 2 mg of pTB-1 as an internal control. pUC119 was used as a carrier to adjust the total amount of DNA to 10 mg. Transfection was performed by the calcium-phosphate precipitation method (10). Cells were incubated for 12 h with the precipitate, the medium was then replaced and cultures were further incubated for 24 h with or without the ligand. Final concentrations of each ligand were 100 nM for VD3 , T3 and estradiol, and 1 mM for RA. Thereafter, the cells were harvested and enzymatic activity in cell extracts was measured by using the standard procedure. Radioactivity of reacted products was quantified by use of an imaging analyzer (BAS2000, Fuji). CAT assay was repeated at least twice for each experiment.
FIG. 2. Binding specificity of ELP1 and nuclear receptors. [a-32P]dCTP-labeled DR3/ELP site, DR4/ELP site, DR5/ELP site, and ERE/ ELP site were used as probes. In each reaction, nuclear extract containing 1.5 mg protein was used. C, E, V, T, R, and ER denote nuclear extracts of COS7 cells transfected with none, pCMX-ELP1, pCMXhVDR, pCMX-hTRb, pCMX-hRARa, and pCMX-hER, respectively.
ligand were 100 nM for VD3 , T3 and estradiol. Thereafter, the cells were harvested and nuclear extracts were prepared. Procedures for the preparation of nuclear extracts and the gel shift assay were described previously (6, 19).
RESULTS
Gel shift assay. COS7 cells were transfected with effector plasmids, and those nuclear extracts were used in the gel shift assay. Briefly, COS7 cells were seeded at a density of 7.51105 cells per 10-cm plate 24 h prior to transfection. Four or 5 plates were prepared for each sample. Twenty micrograms of the effector plasmid was transfected into the cells by the calcium-phosphate precipitation method. Cells were incubated for 24 h with the precipitate, the medium was then replaced and cultures were further incubated for 24 h. Three hours before the end of incubation, appropriate ligands were added to the culture transfected with pCMXhVDR, pCMX-hTRb and pCMX-hER. Final concentrations of each
Binding of ELP1 and nuclear receptors to the oligonucleotide elements. Oligonucleotide elements shown in Fig. 1 were used as probes, and the binding of ELP1 and other nuclear receptors to these elements was tested by gel shift assay using nuclear extracts of COS7 cells transfected with effector plasmids. As predicted from the sequences, these oligonucleotide elements served as efficient targets for binding of nuclear receptors; DR3/ELP site was bound by ELP1 and VDR (Fig. 2, lane 2 and 3), DR4/ELP site by ELP1 and TRb (Fig. 2, lane 5 and 6), DR5/ELP site by ELP1 and RARa (Fig. 2, lane 8 and 9), and ERE/ELP site by ELP1 and ER (Fig. 2, lane 11 and 12).
FIG. 1. Sequences of oligonucleotides. Each oligonucleotide has the Hind III linker on both sides. An arrow indicates the half site of direct or inverted repeats, and an underline indicates the binding sequence for ELP.
Ligand induced transactivation and its repression by ELP1. To test whether each oligonucleotide elements function as ligand response elements, reporter plasmids carrying one copy of DR3/ELP site, DR4/ELP site, DR5/ELP site or ERE/ELP site were transfected into NIH3T3 cells together with effector plasmids coding for LUC or target nuclear receptors. Cultures were treated with corresponding ligands and the CAT activities were analyzed. Data are shown in Fig. 3A. DR3/ELP site and DR5/ELP site responded to their ligands efficiently without co-transfection of VDR or RARa. We assumed that this was due to the presence of endogenous VDR and RAR. In the case of DR4/ELP site and ERE/ELP site, corresponding ligands activated transcription when co-transfected with pCMX-hTRb and pCMX-
408
AID
BBRC 5972
/
6916$$$762
12-24-96 01:49:42
bbrcg
AP: BBRC
Vol. 230, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 3. Ligand-induced transactivation and its repression by ELP1. (A) Two micrograms of reporter plasmids (pDR3/ELP-DSVCAT, pDR4/ELP-DSVCAT, pDR5/ELP-DSVCAT, and pERE/ELP-DSVCAT) was transfected into NIH3T3 cells together with effector plasmids. Nuclear receptor minus cultures were those transfected with pCMX-LUC. Amounts of each effector plasmid were 100 ng of pCMX-LUC, pCMX-hVDR, and pCMX-hRARa, 15 ng of pCMX-hTRb, and 5 ng of pCMX-hER. After incubation for 24 h with or without addition of ligands (VD3 , T3 , and estradiol, 100 nM; RA, 1 mM), CAT activities were measured. Each CAT activity was normalized to that of ligand and nuclear receptor minus culture. (B) Two micrograms of reporter plasmids was transfected into NIH3T3 cells with or without 100 ng of pCMX-ELP1. pDR4/ELP-DSVCAT and pERE/ELP-DSVCAT were analyzed under co-transfection of 15 ng of pCMX-hTRb and 5 ng of pCMX-hER, respectively. In ELP1 minus cultures of pDR3/ELP-DSVCAT and pDR5/ELP-DSVCAT, 100 ng of pCMX-LUC was co-transfected. Culture conditions and measurement of CAT activities are the same as those of (A). Each CAT activity was normalized to that of ligand and ELP1 minus culture.
hER, respectively. These data indicate that DR3/ELP site functions as the VD3 response element, DR4/ELP site as the T3 response element, DR5/ELP site as the RA response element, and ERE/ELP site as the estradiol response element. We have reported previously that ELP1 represses RA-induced transactivation by competition with RARa (14). Similarly, ELP1 may function as a repressor to other nuclear receptors and represses ligand-induced transactivation. A series of reporter plasmids used in Fig. 3A were transfected into NIH3T3 cells together with or without pCMX-ELP1, and cultures were treated with corresponding ligands. Since the ligand response of DR4/ELP site and ERE/ELP site required the presence of pCMX-hTRb and pCMX-hER, respectively (Fig. 3A), those effector plasmids were co-transfected when testing the effect of ELP1. Data are shown
in Fig. 3B. In DR3/ELP site, ELP1 repressed VD3-induced transactivation significantly (2.1 fold repression), and repression by ELP1 was little when VD3 was absent (1.4 fold repression). Similarly ELP1 specifically repressed T3-induced transactivation in DR4/ELP site (2.3 and 1.3 fold repression with or without T3). In DR5/ ELP site and ERE/ELP site, ELP1 repressed transcription under ligand free condition (1.7 and 2.0 fold repression, respectively), and the repression was more significantly when ligand were present (2.7 and 4.0 fold repression, respectively). Repression of RA-induced transactivation by ELP1 of the present analysis was consistent with the previous report (14). In addition, ELP1 was found to repress transcriptional activation by VDR, TR and ER. ELP isoforms repress RA-induced transactivation by competitive binding. Isoforms of ELP, ELP1, ELP2,
409
AID
BBRC 5972
/
6916$$$762
12-24-96 01:49:42
bbrcg
AP: BBRC
Vol. 230, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 4. Competitive binding and repression of DR5/ELP site by ELP isoforms and a truncated mutant. (A) Structures of ELP isoforms and the ELP truncated mutant. ZF and F1 indicate the zinc finger region and the FTZ-F1 box, respectively (26). (B) Binding specificity of ELP isoforms and a truncated mutant to DR5/ELP site. [a-32P]dCTP-labeled DR5/ELP site was used as a probe. In each reaction, nuclear extract containing 1.5 mg protein was included. C, E, 2, 3, Z, and R denote nuclear extracts of COS7 cells transfected with none, pCMXELP1, pCMX-ELP2, pCMX-ELP3, pCMX-E-ZF, and pCMX-hRARa, respectively. (C) Effector plasmids (100 ng each) were transfected into NIH3T3 cells with 2 mg of pDR5/ELP-DSVCAT. After incubation for 24 h with or without addition of 1 mM RA, CAT activities were measured. Each CAT activity was normalized to that of pCMX-LUC with addition of RA. (D) Dose-dependent competition of E-ZF to RARaDNA complex. [a-32P]dCTP-labeled DR5/ELP site was used as probe. In each reaction, 2 mg of nuclear extracts of COS7 cells transfected with none (C) or pCMX-hRARa (R) was included. Nuclear extract of COS7 cells transfected with pCMX-E-ZF (Z) was added into the above reactions (from 0.8 ng to 2 mg).
ELP3 and Ad4BP/SF-1, all bound to the same sequence element, yet ELP1 functioned as a repressor and other three as activator. The activator type ELP isoforms are also expected to compete for binding to the sequence elements of nuclear receptors when the elements contain the consensus sequence of ELP binding. Therefore, we have studied the effect of ELP isoforms on RAR mediated transactivation by using DR5/ELP site as the RA response element. Fig. 4A shows the structures of three ELP isoforms, ELP1, ELP2 and ELP3, and a truncated mutant of ELP, E-ZF. E-ZF contains only the DNA binding region of the gene. Binding activity of those ELP isoforms and E-ZF to DR5/ELP site was analyzed. As is clear from Fig. 4B, ELP2, ELP3 and E-ZF formed complexes with DR5/ELP site as did ELP1 and RARa. To test the effect of ELP isoforms and E-ZF on RAR mediated transactivation, effector plasmids carrying ELP isoforms or E-
ZF were transfected into NIH3T3 cells together with pDR5/ELP-DSVCAT. Data are shown in Fig. 4C. ELP2 and ELP3 did not activate the transcription under RA free condition. Furthermore, they repressed RA-induced transactivation of DR5/ELP site. There was no significant difference in the magnitude of repression among ELP isoforms (2.6, 2.5 and 2.3 fold repression by ELP1, ELP2 and ELP3, respectively). In addition, the truncated mutant of ELP, E-ZF, also repressed RAinduced transactivation. Above data suggest that ELP isoforms compete with RARa and repressed RA-induced transactivation, and DNA binding region of ELP is sufficient for the competitive repression. To confirm this suggestion, it was analyzed whether the complex formation between DR5/ELP site and RARa was abrogated by E-ZF. The binding of RARa to DR5/ELP site was indeed competed in proportion to the dose of addition of E-ZF (Fig. 4D).
410
AID
BBRC 5972
/
6916$$$762
12-24-96 01:49:42
bbrcg
AP: BBRC
Vol. 230, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
DISCUSSION In this study, we showed that ELP1 could compete with nuclear receptors, VDR, TR, RAR and ER, and repressed their ligand-dependent transactivation (Fig. 3B). ELP1 lacked the region III, the putative activator domain conserved among nuclear receptors, and functioned as a repressor (11). On the other hand, ELP3 possessed that region, and functioned as an activator (16). Truncated mutant of ELP3 which lacks the region III could not activate transcription (data not shown). These findings suggest that ELP1 is an inactive form, and competitive binding is a general mechanism of the ELP1 mediated repression. ELP3-induced transactivation was repressed by ELP1 efficiently (data not shown), and ELP1 was expressed with other ELP isoforms in cell lines and tissues (16). Therefore, it is possible that the major targets of ELP1 are other ELP isoforms, and ELP1 modulates the transactivation by activator type ELP isoforms. Such competitive binding was reported for Drosophila FTZ-F1 in which FTZ-F1a and FTZ-F1b compete for binding to a site in the fushi tarazu gene (21). ELP3 is identical in amino acid sequence with mouse SF-1 and bovine Ad4BP (16). SF-1 and Ad4BP are known as key regulators of steroidgenic P-450 genes, and activate the transcription of those genes in adrenal cells (22, 23). ELP2 shares the same N-terminus as ELP1, but carries the C-terminal transactivation domain as in the case of ELP3 (16). ELP2 and ELP3 activated transcription of a reporter plasmid carrying 8 copies of the ELP binding element of the Moloney leukemia virus LTR (16), but a single copy of that element was not sufficient for transcriptional activation (data not shown). In addition, ELP2 and ELP3 did not activate the transcription of reporter plasmid carrying a single copy of DR5/ELP site (Fig. 4C). A similar phenomenon was reported on the Xenopus homologue of FTZ-F1 (24). For transcriptional activation of steroidgenic P-450 genes, several elements are need in addition to Ad4BP binding element (25). It is likely that ELP isoforms bind to the target elements as a monomer, because FTZ-F1 binds to the target sequence as a monomer (26). These findings raise an interesting possibility that transcriptional activation mediated by the ELP isoforms is strictly dependent upon the context of the target promoter and other transcription factors available in a given cell. In contrast to the mechanism of transactivation, repression by ELP is mediated simply by competition. ELP2 and ELP3 repressed RA-induced transactivation on DR5/ELP site by competing for the binding site as in the case of ELP1 and RARa (Fig. 4C). This suggests that even the activator type of ELP isoforms repress transactivation by other nuclear receptors when they share the binding sequence motif. Indeed, there are some data suggesting the negative function of ELP iso-
forms. Although ELP3 was the major form of the isoforms expressed in undifferentiated EC cells, the ELP binding site of the Moloney murine leukemia virus genome nevertheless functioned as a negative element (6, 16). ELP isoforms functioned negatively to ELP binding site of the ELP gene, and this ELP binding site overlapped with the consensus sequences for other transcription factors binding (27). Consensus sequence for ELP binding have been found in several genes, and some of these genes are proposed as the target of ELP isoforms (28, 29). Further analysis of the function of ELP on these genes may verify the roles of ELP isoforms in transactivation and repression. ACKNOWLEDGMENTS We thank Dr. S. Ishii for the pTB-1 plasmid. We also thank Mr. T. Nishioka for photographic work and Ms. T. Matsuura for secretarial work. This work was supported by a Grant-in-Aid from the Ministry of Education, Science, and Culture, Japan.
REFERENCES 1. Niwa, O., Yokota, Y., Ishida, H., and Sugahara, T. (1983) Cell 32, 1105–1113. 2. Linney, E., Davis, E., Overhauser, D. J., Chao, E., and Fan, H. (1984) Nature 308, 470–472. 3. Niwa, O. (1985) Mol. Cell. Biol. 5, 2325–2331. 4. Loh, T. P., Sievert, L. L., and Sctt, R. W. (1990) Mol. Cell. Biol. 10, 4045–4057. 5. Tsukiyama, T., Niwa, O., and Yokoro, K. (1990) Virology 177, 772–776. 6. Tsukiyama, T., Niwa, O., and Yokoro, K. (1991) J. Virol. 65, 2979–2986. 7. Petersen, R., Kempler, G., and Barklis, E. (1991) Mol. Cell. Biol. 11, 1214–1221. 8. Akgun, E., Ziegler, M., and Grez, M. (1991) J. Virol. 65, 382– 388. 9. Yamauchi, M., Freitag, B., Khan, C., Berwin, B., and Barklis, E. (1995) J. Virol. 69, 1142–1149. 10. Tsukiyama, T., Niwa, O., and Yokoro, K. (1989) Mol. Cell. Biol. 9, 4670–4676. 11. Tsukiyama, T., Ueda, H., Hrise, S., and Niwa, O. (1992) Mol. Cell. Biol. 12, 1286–1291. 12. Mangelsdorf, D. J., Thummel, C., Beato, M., Herrlich, P., Shutz, G., Umesono, K., Blumberg, B., Kastner, P., Mark, M., Chambon, P., and Evans, R. M. (1995) Cell 83, 835–839. 13. Tsukiyama, T., and Niwa, O. (1992) Nucleic Acids Res. 20, 1477– 1482. 14. Kotomura, N., Okada, M., Ninomiya, Y., Tsukiyama, T., Umesono, K., Evans, R. M., and Niwa, O. (1994) J. Biochem. 116, 1309–1316. 15. Taketo, M., Parker, K. L., Howard, T. A., Tsukiyama, T., Wong, M., Niwa, O., Morton, C. C., Miron, P. M., and Seldin, M. F. (1995) Genomics 25, 565–567. 16. Ninomiya, Y., Okada, M., Kotomura, N., Suzuki, K., Tsukiyama, T., and Niwa, O. (1995) J. Biochem. 270, 7453–7461. 17. Mangelsdorf, D. J., and Evans, R. M. (1995) Cell 83, 841–850. 18. Beato, M., Herrlich, P., and Schutz, G. (1995) Cell 83, 851–857. 19. Umesono, K., Murakami, K. K., Thompson, C. C., and Evans, R. M. (1991) Cell 65, 1255–1266.
411
AID
BBRC 5972
/
6916$$$762
12-24-96 01:49:42
bbrcg
AP: BBRC
Vol. 230, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
20. Waterman, M. L., Adler, S., Nelson, C., Greene, G. L., Evans, R. M., and Rosenfeld, M. G. (1988) Mol. Endocrinol. 2, 14–21. 21. Ohno, C. K., Ueda, H., and Petkovich, M. (1994) Mol. Cell. Biol. 14, 3166–3175. 22. Ikeda, Y., Lala, D. S., Luo, X., Kim, E., Moisan, M-P., and Parker, K. L. (1993) Mol. Endocrinol. 7, 852–860. 23. Morohashi, K., Zanger, U. M., Honda, S., Hara, M., Waterman, M. R., and Omura, T. (1993) Mol. Endocrinol. 7, 1196–1204. 24. Ellinger-Ziegelbauer, H., Hihi, A. K., Laudet, V., Keller, H., Wahli, W., and Dreyer, C. (1994) Mol. Cell. Biol. 14, 2786–2797.
25. Takayama, K., Morohashi, K., Honda, A., Hara, N., and Omura, T. (1994) J. Biochem. 116, 193–203. 26. Ueda, H., Sun, G.-C., Murata, T., and Hirose, S. (1992) Mol. Cell. Biol. 12, 5667–5672. 27. Ninomiya, Y., Kotomura, N., and Niwa, O. (1996) Biochem. Biophys. Res. Commun. 222, 632–638. 28. Shen, W.-H., Moor, C. C. D., Ikeda, Y., Parker, K., and Igraham, H. A. (1994) Cell 77, 651–661. 29. Burris, T., Guo, W., Le, T., and McCabe, R. B. (1995) Biochem. Biophys. Res. Commun. 214, 576–581.
412
AID
BBRC 5972
/
6916$$$762
12-24-96 01:49:42
bbrcg
AP: BBRC