- Email: [email protected]
Identification of a Novel Isoform of Estrogen Receptor, a Potential Inhibitor of Estrogen Action, in Vascular Smooth Muscle Cells
JOBNAME: BBRC 219#3 PAGE: 1 SESS: 5 OUTPUT: Mon Apr 1 16:38:35 1996 /xypage/worksmart/tsp000/68953f/19
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS ARTICLE NO.
219, 766–772 (1996)
0308
Identification of a Novel Isoform of Estrogen Receptor, a Potential Inhibitor of Estrogen Action, in Vascular Smooth Muscle Cells Satoshi Inoue,1 Shin-jiro Hoshino, Hideyuki Miyoshi, Masahiro Akishita, Takayuki Hosoi, Hajime Orimo, and Yasuyoshi Ouchi Department of Geriatrics, Faculty of Medicine, The University of Tokyo, Hongo, Bunkyo-ku, Tokyo 113, Japan Received January 16, 1996 Clinical and experimental studies showed that estrogen has antiatherogenic effects. We previously demonstrated that the estrogen receptor (ER) mRNA and protein are expressed in vascular smooth muscle cells (VSMC) derived from rat aorta. Here, the expression of isoforms of the ER was examined in VSMC. Reverse transcriptase-polymerase chain reaction using specific primers for rat ER cDNA was performed from RNA of rat VSMC. This revealed the existence of ER cDNA that is shorter than the wild-type ER cDNA. Sequencing of the amplified products identified three isoforms of the ER and the wild-type ER. These ER mRNA isoforms lacked the region corresponding to exon 4, exon 4 and 5, and exon 3 and 4. Therefore, they were designated as ERD4 isoform, ERD4/5 isoform and ERD3/4 isoform, respectively. Chloramphenicol acetyltransferase assay was performed with these ER isoforms constructed into the expression vector and the reporter plasmid containing the estrogen responsive element. The assay showed that these ER isoforms lost estrogen-dependent transactivation activities and that ERD4/5 isoform has a inhibitory effect on normal estrogen action when it was cotransfected with the wild-type ER. These ER isoforms might be involved in the regulation of VSMC by estrogen. © 1996 Academic Press, Inc.
The protective effects of estrogen from atherosclerosis has been noticed by clinical (1, 2) and experimental (3–5) studies. Premenopausal women have low rate of coronary artery disease compared with men of similar age (6). After natural or surgical menopause, the risk of atherosclerosis increased markedly (6, 7). Estrogen replacement therapy is effective for protection of postmenopausal women from coronary heart disease (1, 2). However, the mechanism of estrogen action on atherosclerosis is poorly understood. Estrogen-related effects on serum lipoprotein levels may explain a part of these mechanisms (3, 4, 8). On the other hand, some experimental studies revealed that the protective effect of estrogen on atherosclerosis could be observed when serum lipoprotein level was unchanged (5). These findings suggest that estrogen may act on the cardiovascular tissues directly. To support this idea, specific binding of estrogen was shown in the heart (9), vascular tissues (10) and vascular cells (11, 12) suggesting the presence of estrogen receptor (ER). In the pathogenesis of atherosclerosis, smooth muscle cell proliferation in vivo plays a central role (13). Atherogenic stimulations induce abnormal proliferation of the vascular smooth muscle cells (VSMC), which lead to the formation of atherosclerotic lesion. Several groups have shown that the proliferation of VSMC is inhibited by estrogen treatment (14, 15). This inhibitory effect on the proliferation of VSMC may contributes to the anti-atherogenic effects of estrogen. We (16) and Karas et al. (17) previously reported that the functional ER was present in VSMC at both the protein and the mRNA levels. These ER receptors may mediate the estrogen action in VSMC. Estrogen has diverse effects on various organs, tissues and cells. These effects were mainly mediated by the ER that exists as a single gene in the mammalian genome (18). To date, several estrogen receptor variants have been isolated in cancerous cells (19–21). These isoforms are mostly splicing variants of the ER that lacks some exons. An isoform of the ER was reported in the normal 1
To whom correspondence should be addressed. Fax: 011-81-3-5689-2483. Abbreviations: FBS, fetal bovine serum; VSMC, vascular smooth muscle cells; ER, estrogen receptor; RT-PCR; Reverse transcriptase-polymerase chain reaction; CAT, Chloramphenicol acetyltransferase. 766 0006-291X/96 $18.00 Copyright © 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
JOBNAME: BBRC 219#3 PAGE: 2 SESS: 4 OUTPUT: Mon Apr 1 16:38:35 1996 /xypage/worksmart/tsp000/68953f/19
Vol. 219, No. 3, 1996
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
tissue, in the brain, for the first time (22). This isoform of the ER lacks exon 4 that include both DNA-binding and estrogen-binding domains. Some of the isoforms of the ER have variant functions compared with the wild-type ER receptor (20, 21). It is possible that isoforms of the ER may modulate the estrogen action in the cardiovascular system. These circumstances prompted us to study the existence of isoforms of the ER in VSMC and possible involvement of the isoforms in the mechanism of estrogen action in VSMC. We performed reverse transcriptase-polymerase chain reaction (RT-PCR) using ER specific primers. Analysis of the PCR products has indicated the presence of three ER mRNA isoforms other than the wild-type ER in VSMC. The function of these ER isoforms was studied by chloramphenicol acetyltransferase (CAT) assay and one of them was shown to have an inhibitory effect on the wild-type ER in VSMC. MATERIALS AND METHODS Cell culture. Eight-week-old female rats (Nippon Bio-Supply Center) were sacrificed under ether anesthesia. VSMC were cultured from medial layer of the thoracic aorta by the method of Chamley et al. (23) Subcultured VSMC (4–6th passage) were used in the experiments. VSMC and A10 cells (a rat aortic smooth muscle cell-line) (24) were maintained in Dulbecco’s modified Eagle medium supplemented with 10% fetal bovine serum (FBS) (Cell Culture Laboratories) at 37°C in a humidified atmosphere of 5% CO2. A10 cells were suitable for transfection experiments using calcium-phosphate precipitation method (25). Reverse transcriptase-polymerase chain reaction (RT-PCR). PCR primers for the rat ER cDNA (18, 22) were synthesized as E-1 (nt 611-630) 59-CTACTACCTGGAGAACGAGC-39; E-2 (nt 762-779) 59-AAGGAGACTCGCTACTGT-39; E-6 (nt 1482-1501) 59-TCAAAGATCTCCACCATGCC-39; E-7 (nt 1650-1669) 59-ATCTTGTCCAGGACTCGGTG-39. RT-PCR was performed as described previously (16). cDNA was synthesized from 10 mg of total RNA of VSMC essentially according to Gubler and Hoffman (26) using the E-7 primer. Samples with or without treatment of avian myeloblastosis virus reverse transcriptase (Seikagaku Kougyo, Tokyo, Japan) were prepared as controls. One-tenth of the resulting cDNA was used as template DNA for the first PCR with E-1 and E-7 primers. One-tenth of the first PCR products was used as template DNA for the second PCR with E-2 and E-6 primers. The reaction was carried out in a final volume 20 mL containing template DNA, 10 pmol of each primer, 200 mmol/L dGTP, dATP, dTTP and dCTP, 10 mmol/L Tris-HCl (pH 8.3), 50 mmol/L KCl, 1.5 mmol/L MgCl2, 0.02% gelatin and 1U Taq DNA polymerase. The temperature program for the amplification was 30 cycles of 1 min at 94°C, 1 min at 57°C, and 2 min at 72°C. DNA sequencing and plasmid construction. PCR products were cloned into PCRII plasmid (Invitrogen) according to manufacturer’s instructions. The resulting recombinant DNA was purified and sequenced by dideoxy method (27) using sequenase (US Biochemical). Utilizing PCR fragments derived from these DNA and full length cDNA of the rat ER (18), the wild-type ER and three ER isoforms were constructed in an expression vector PSSRa (28) as PSSRaER, PSSRaERd4, PSSRaERD4/5, and PSSRaERD3/4. The resulting plasmids were confirmed by sequencing. Chloramphenicol acetyltransferase (CAT) assay. CAT assay was performed as described (29). Briefly, 1 × 106 of A10 cells were plated one day prior to transfection. One hour prior to transfection, the medium was replaced with phenol red free medium containing 10% dextran charcoal-treated FBS. By calcium-phosphate precipitation method (25), cells were transfected with 0.2 mg of expression plasmids (PSSRaER, PSSRaERD4, PSSRaERD4/5 or PSSRaERD3/4), 2 mg of vitERE-CAT reporter plasmid (30) containing the estrogen responsive element and 2 mg of pCH110 b-galactosidase expression plasmid (Pharmacia) used as an internal control to normalize for variations in transfection efficiency. In several experiments, indicated amount of expression plasmids for ER isoforms was added to wild-type PSSRaER expression plasmid (0.2 mg). The total amount of DNA transfected was made up to 20 mg with carrier DNA pGEM3Zf(-) (Promega). After 12 hour incubation the cells were cultured further with medium change in the absence or presence of 1 × 10−8 mol/L 17b-estradiol for 24 hours. The cell extracts were assayed for CAT activities as described (31). The experiment was carried out three times and a representative pattern is shown. To determine relative intensity of the signals, a macintosh computer with a scanner was utilized.
RESULTS Identification of the ER Isoforms in VSMC To investigate the expression of ER isoforms, we used two set of primers to increase the specificity of the amplified products. E-1, E-2, E-6 and E-7 primers are located in exon 1, exon 2, exon 6 and exon 7 of the ER gene, respectively. Total RNA (10 mg) from rat VSMC was converted into corresponding cDNA using reverse transcriptase (RT) and an ER gene-specific primer (E-7). The first PCR was performed utilizing the RT products as template DNA with E-1 and E-7 primers. 767
JOBNAME: BBRC 219#3 PAGE: 3 SESS: 4 OUTPUT: Mon Apr 1 16:38:35 1996 /xypage/worksmart/tsp000/68953f/19
Vol. 219, No. 3, 1996
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Then, the second PCR was carried out using the first PCR products as template DNA with E-2 and E-6 primers. The second PCR products were analyzed by 1% agarose gel electrophoresis (Fig. 1). As a result, other than a band corresponding to the PCR products derived from the wild-type rat cER (Fig. 1; Lane 2, 5: shown by arrowhead), shorter bands were detected in the RT-PCR products derived from VSMC RNA (Fig. 1; lane 2: shown by arrow). There was no band detected in samples without RT (Fig. 1; lane 3) and without template (Fig. 1 lane 4). Amplified PCR products were isolated from agarose gel and cloned into plasmid vector. Then, their sequences were determined by dideoxy method. The sequence of the long band corresponded to the wild-type ER cDNA contained all exon 3, exon 4 and exon 5. The exon-exon junctional sequence were shown in Fig. 2a. One of the sequence of the shorter bands indicated that the 59-termini of exon 5 was connected to the 39-termini of exon 3 (Fig. 2b), another of the sequence of the shorter bands indicated that the 59-termini of exon 6 was connected to the 39-termini of exon 4 (Fig. 2c), and the other of the sequence of the short bands showed that the 59-termini of exon 5 was connected to the 39-termini of exon 2 (Fig. 2d). Thus, they were designated as ERD4 isoform, ERD4/5 isoform and ERD3/4 isoform, respectively. A schema of the relationship of ER isoforms and functional domains of the ER is shown in Fig. 3. The exon-exon junctional sequences and corresponding amino acids are shown for ERD4, ERD4/5 and ERD3/4 isoforms. ERD4 and ERD3/4 isoforms encode the shorter protein that lack amino acids corresponding to those exons. In case of ERD4/5 isoform, a frame-shift occurred at the exon-exon junction and a premature stop codon appears. Thus, ERD4/5 isoform encodes a truncated protein. Inhibitory Effects of the ERD4/5 Isoform on the Wild-Type ER Action To examine the possible function of these ER isoforms, we performed CAT assay using ER isoform expression plasmids. It was shown that the wild-type ER expression plasmid have estrogen dependent transactivating activity and that ERD4, ERD4/5 and ERD3/4 isoforms have lost such activities (Fig. 4). To examine the effect of these isoforms on function of the wild-type ER, each ER isoform expression plasmid was cotransfected with the wild-type ER expression plasmid (Fig. 5A). In case of ERD4/5 the activation of ERE-tk-CAT by wild-type ER was significantly decreased when 50 times amount of isoform ER was added. To confirm this observation, the wild-type ER expression plasmid was cotransfected with increasing amount of ERD4/5 isoform expression plasmid (Fig. 5B). The activation of ERE-tk-CAT by the wild-type ER was decreased dose dependently with the ERD4/5 isoform. Relative values of acetylated chloramphenicol signals were shown in Fig 5B.
FIG. 1. Estrogen receptor (ER) isoforms detected in VSMC. Reverse transcriptase-polymerase chain reaction (RT-PCR) products were analyzed by 1% agarose gel electrophoresis. RT-PCR samples of the VSMC (VSMC) with (lane 2; RT+) and without (lane 3; RT-) reverse transcriptase, and without template (lane 4; NT) and PCR products from the rat ER cDNA (lane 5; rat cER) are shown. Lane 1 (M) shows the l/Hind III size marker. 768
JOBNAME: BBRC 219#3 PAGE: 4 SESS: 4 OUTPUT: Mon Apr 1 16:38:35 1996 /xypage/worksmart/tsp000/68953f/19
Vol. 219, No. 3, 1996
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. The sequence analysis of the reverse transcriptase-polymerase chain reaction products. Autoradiogram of exon–exon junctional sequence determined by dideoxy method for the wild-type estrogen receptor (ER) (a), ERD4 (b), ERD4/5 (c) and ERD3/4 (d) are shown.
DISCUSSION In the previous study (16), we demonstrated the presence of an estrogen receptor in vascular smooth muscle cells at the mRNA level by Northern blot analysis and RT-PCR. The presence of the ER protein was confirmed by immunocytochemistry utilizing anti-ER antibody. We examined
FIG. 3. A schematic representation of the estrogen receptor (ER) isoforms and functional domains of the ER. The position of the primers used in this experiment are shown by arrowheads. The exon–exon junctional sequences and corresponding amino acids of the ER isoforms are shown. Note that a frame-shift is occurred and premature stop codon appears encoding the truncated protein only in the ERD4/5 isoform. 769
JOBNAME: BBRC 219#3 PAGE: 5 SESS: 4 OUTPUT: Mon Apr 1 16:38:35 1996 /xypage/worksmart/tsp000/68953f/19
Vol. 219, No. 3, 1996
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 4. Chloramphenicol acetyltransferase (CAT) assay analysis of three isoforms of the estrogen receptor (ER) mRNA. The wild-type and isoform ER expression plasmids, PSSRaER (ER), PSSRaERD4 (D3/4), PSSRaERD4/5 (D4/5), PSSRaERD3/4 (D3/4) or PSSRa (vector) were transfected into A10 vascular smooth muscle cells with the vitERE-tk-CAT reporter plasmid. After culturing in the absence (−) or presence (+) of 10 nmol/L 17b estradiol (E2), CAT assay was performed.
the regulation of several estrogen responsive genes and found that the transcripts of c-fos protooncogene was regulated by estrogen in these cells. Karas et al. (17) also described that human vascular smooth muscle cells contain the ER both at the mRNA and the protein levels. They have shown by luciferase assay that the ER has estrogen dependent transactivation activity in those cells. These reports suggested that estrogen can act on the vascular smooth muscle cells directly and that
FIG. 5. Inhibitory effect of the ERD4/5 isoform on the wild-type estrogen receptor (ER) action. (A) The wild-type ER expression plasmid PSSRaR was cotransfected with indicated amount (the same or 50 times) of PSSRaERD4 (D4), PSSRaERD4/5 (D4/5) or PSSRaERD3/4 (D3/4). After culturing in the absence (−) or presence (+) of 10 nmol/L 17bestradiol (E2), chloramphenicol acetyltransferase (CAT) assay was performed. (B) The wild-type ER expression plasmid PSSRaER was cotransfected with indicated amount (1×, 3× 9× 27× and 81×) of PSSRaERD4/5 (D4/5). After culturing in the absence (−) or presence (+) of 10 nmol/liter 17b-estradiol (E2), CAT assay was performed. Relative values (transfection with the wild-type ER under the presence of E2 as 100) of acetylated cloramphenicol signals were shown at the top. 770
JOBNAME: BBRC 219#3 PAGE: 6 SESS: 4 OUTPUT: Mon Apr 1 16:38:35 1996 /xypage/worksmart/tsp000/68953f/19
Vol. 219, No. 3, 1996
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
the ER demonstrated in vascular smooth muscle cells can mediate estrogen action in the cardiovascular system. In the present study, we detected three isoforms of ER in the rat VSMC by RT-PCR. One is the isoform that lacks exon 4. Another isoform lacks both exon 4 and exon 5, and the other lacks exon 3 and exon 4. Although it is difficult to compare accurately the amount of mRNA by RT-PCR, the signal of shorter bands are comparable with that of the long band corresponding to the wild-type ER cDNA. From six clones derived from shorter bands, we obtained 3 clones of ERD4, 2 clones of ERD4/5 and 1 clone of ERD3/4. Thus, we estimate that the existence of the ERD4/5 isoform, that was identified for the first time, would not be negligible. In cancerous cells, several isoforms have been demonstrated, such as, ERD3, ERD4, ERD5, ERD7 isoforms. On the other hand, only few reports showed estrogen receptor isoforms in normal organs and tissues. Skipper et al. (22) has been reported the existence of ERD4 in the rat brain. We recently found ERD4 and ERD3/4 isoforms in the rat bone tissues (32). However, the function of these isoforms were still unknown. It has been reported that ERD3 isoform in the human breast cancer cells inhibits estrogen-dependent transcription activation when it is cotransfected with the wild-type ER and reporter plasmid (20). Both exon 3 and exon 4 contain a part of the DNA binding domain, and both exon 4 and exon 5 contain a part of the estrogen-binding domain (22, 33). Therefore, these three isoforms may lose binding activities to estrogen responsive elements and estrogen. CAT assay clearly demonstrated that these isoforms lost estrogen dependent transactivation activity via the estrogen responsive element. Furthermore, CAT assay cotransfected with the wild-type ER revealed one of these isoform, ERD4/5, has an inhibitory effect on normal estrogen action with a dominant negative fashion. It is notable that ERD4/5 encodes truncated protein because of a frame-shift occurred at the exon-exon junction. These results suggest that the truncated protein compete with the wild-type ER having the normal function. The data presented in this report are interesting because estrogen has bidirectional effects in various organs, tissues and cells. For example, estrogen promote the growth of breast cancer MCF-7 cells, whereas estrogen inhibit the growth of VSMC (Akishita et al., submitted). It is possible that the differential effects of estrogen are regulated by the receptor isoforms, such as ERD4/5 isoform that is a potential inhibitor of the wild-type ER function. Alternatively, the inhibitory effect of estrogen action by the ER isoform might be involved in pathophysiological conditions. Further studies should be required to clarify the roles of these ER isoforms in normal and atherosclerotic arteries. ACKNOWLEDGMENTS We thank Ms. M. Watanabe, Ms. H. Yamaguchi and Ms. M. Goto for technical assistance. This work was supported by grants from Foundation and the Ministry of Education, Science and Culture, Japan.
REFERENCES 1. Bush, T. L., Barrett-Conner, E., Cowan, L. D., Criqui, M. H., Wallace, R. B., Suchindran, C. M., Tyroler, H. A., and Rifkind, B. M. (1987) Circulation 75, 1102–1109. 2. Stampfer, M. J., Colditz, G. A., Willett, W. C., Manson, J. E., Rosner, B., Speizer, F. E., and Hennekens, C. H. (1991) N. Engl. J. Med. 325, 756–762. 3. Kushwaha, R. S., and Hazzard, W. R. (1981) Metabolism 30, 359–366. 4. Adams, M. R., Clarkson, T. B., Kaplan, J. R., and Koritnik, D. R. (1989) Transplant. Proc. 21, 3662–3664. 5. Cheng, L. P., Kuwahara, M., Jacobsson, J., and Foegh, M. L. (1991) Transplantation 52, 967–972. 6. Kannel, W. B., Hjortland, M. C., McNamara, P. M., and Gordon, T. (1976) Ann. Intern. Med. 85, 447–452. 7. Wittman, J. C. M., Grobbee, D. E., Kok, F. J., Hoffman, A., and Valkenburg, H. A. (1989) Br. Med. J. 198, 642–644. 8. Walsh, B. W., Schiff, I., Rosner, B., Greenberg, L., Ravnikar, V., and Sacks, F. M. (1991) N. Engl. J. Med. 325, 1196–1204. 9. Stumph, W. E., Sar, M., and Aumüller, G. (1977) Science 196, 319–321. 10. Horwitz, K. B., and Horwitz, L. D. (1982) J. Clin. Invest. 69, 750–758. 11. Colburn, P., and Buonassisi, V. (1978) Science 201, 817–819. 771
JOBNAME: BBRC 219#3 PAGE: 7 SESS: 4 OUTPUT: Mon Apr 1 16:38:35 1996 /xypage/worksmart/tsp000/68953f/19
Vol. 219, No. 3, 1996 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33.
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Nakao, J., Chang, W.-C., Murota, S.-I., and Orimo, H. (1981) Atherosclerosis 38, 75–80. Ross, R. (1993) Nature 362, 801–809. Fischer-Dzoga, K., Wissler, R. W., and Vesselinovitch, D. (1983) Exp. Mol. Pathol. 39, 355–363. Vargas, R., Wroblewska, B., Rego, A., Hatch, J., and Ramwell, P. W. (1993) Br. J. Pharmacol. 109, 612–617. Orimo, A., Inoue, S., Ikegami, A., Hosoi, T., Akishita, M., Ouchi, Y., Muramatsu, M., and Orimo, H. (1993) Biochem. Biophys. Res. Commun. 195, 730–736. Karas, R. H., Patterson, B. L., and Mendelsohn, M. E. (1994) Circulation 89, 1943–1950. Koike, S., Sakai, M., and Muramatsu, M. (1987) Nucleic Acids Res. 15, 2499–2513. Dotzlaw, H., Alkhalaf, M., and Murphy, L. C. (1991) Mol. Endcrinol. 6, 773–785. Wang, Y., and Miksicek, R. J. (1991) Mol. Endcrinol. 5, 1707–1715. Fuqua, S. A. W., Chamnes, G. C., and McGuire, W. L. (1993) J. Cell. Biochem. 51, 135–139. Skipper, J. J., Young, L. J., Bergeron, J. M., and Tetzlaff, M. T. (1992) Proc. Natl. Acad. Sci. USA. 90, 7172–7175. Chamley, J. H., Campbell, G. R., McConnell, J. D., and Gröschel-Stewart, U. (1977) Cell Tiss. Res. 177, 503–522. Kimes, B. W., and Brandt, B. L. (1976) Exp. Cell Res. 98, 349–366. Graham, F. L., and van der Eb, A. J. (1973) Virology 52, 456–467. Gubler, V., and Hoffman, B. J. (1983) Gene 25, 456–467. Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. Acad. Sci. USA. 74, 5463–5467. Takebe, Y., Seiki, M., Fujisawa, J.-I., Hoy, P., Yokote, K., Arai, K.-I., Yoshida, M., and Arai, N. (1988) Mol. Cell. Biol. 8, 466–472. Inoue, S., Orimo, A., Hosoi, T., Kondo, S., Toyoshima, H., Kondo, T., Ikegami, A., Ouchi, Y., Orimo, H., and Muramatsu, M. (1993) Proc. Natl. Acad. Sci. USA. 90, 11117–11121. Inoue, S., Kondo, S., Hashimoto, M., Kondo, T., and Muramatsu, M. (1991) Nucleic Acids Res. 19, 4091–4096. Gorman, C. M., Moffat, L. F., and Howard, B. H. (1982) Mol. Cell. Biol. 2, 1044–1051. Hoshino, S., Inoue, S., Hosoi, T., Saito, T., Ikegami, A., Kaneki, M., Ouchi, Y., and Orimo, H. (1995) Calcif. Tissue Int. 57, 466–468. Ponglikitmongkol, M., Green, S., and Chambon, P. (1988) EMBO J. 7, 3385–3388.
772
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS ARTICLE NO.
219, 766–772 (1996)
0308
Identification of a Novel Isoform of Estrogen Receptor, a Potential Inhibitor of Estrogen Action, in Vascular Smooth Muscle Cells Satoshi Inoue,1 Shin-jiro Hoshino, Hideyuki Miyoshi, Masahiro Akishita, Takayuki Hosoi, Hajime Orimo, and Yasuyoshi Ouchi Department of Geriatrics, Faculty of Medicine, The University of Tokyo, Hongo, Bunkyo-ku, Tokyo 113, Japan Received January 16, 1996 Clinical and experimental studies showed that estrogen has antiatherogenic effects. We previously demonstrated that the estrogen receptor (ER) mRNA and protein are expressed in vascular smooth muscle cells (VSMC) derived from rat aorta. Here, the expression of isoforms of the ER was examined in VSMC. Reverse transcriptase-polymerase chain reaction using specific primers for rat ER cDNA was performed from RNA of rat VSMC. This revealed the existence of ER cDNA that is shorter than the wild-type ER cDNA. Sequencing of the amplified products identified three isoforms of the ER and the wild-type ER. These ER mRNA isoforms lacked the region corresponding to exon 4, exon 4 and 5, and exon 3 and 4. Therefore, they were designated as ERD4 isoform, ERD4/5 isoform and ERD3/4 isoform, respectively. Chloramphenicol acetyltransferase assay was performed with these ER isoforms constructed into the expression vector and the reporter plasmid containing the estrogen responsive element. The assay showed that these ER isoforms lost estrogen-dependent transactivation activities and that ERD4/5 isoform has a inhibitory effect on normal estrogen action when it was cotransfected with the wild-type ER. These ER isoforms might be involved in the regulation of VSMC by estrogen. © 1996 Academic Press, Inc.
The protective effects of estrogen from atherosclerosis has been noticed by clinical (1, 2) and experimental (3–5) studies. Premenopausal women have low rate of coronary artery disease compared with men of similar age (6). After natural or surgical menopause, the risk of atherosclerosis increased markedly (6, 7). Estrogen replacement therapy is effective for protection of postmenopausal women from coronary heart disease (1, 2). However, the mechanism of estrogen action on atherosclerosis is poorly understood. Estrogen-related effects on serum lipoprotein levels may explain a part of these mechanisms (3, 4, 8). On the other hand, some experimental studies revealed that the protective effect of estrogen on atherosclerosis could be observed when serum lipoprotein level was unchanged (5). These findings suggest that estrogen may act on the cardiovascular tissues directly. To support this idea, specific binding of estrogen was shown in the heart (9), vascular tissues (10) and vascular cells (11, 12) suggesting the presence of estrogen receptor (ER). In the pathogenesis of atherosclerosis, smooth muscle cell proliferation in vivo plays a central role (13). Atherogenic stimulations induce abnormal proliferation of the vascular smooth muscle cells (VSMC), which lead to the formation of atherosclerotic lesion. Several groups have shown that the proliferation of VSMC is inhibited by estrogen treatment (14, 15). This inhibitory effect on the proliferation of VSMC may contributes to the anti-atherogenic effects of estrogen. We (16) and Karas et al. (17) previously reported that the functional ER was present in VSMC at both the protein and the mRNA levels. These ER receptors may mediate the estrogen action in VSMC. Estrogen has diverse effects on various organs, tissues and cells. These effects were mainly mediated by the ER that exists as a single gene in the mammalian genome (18). To date, several estrogen receptor variants have been isolated in cancerous cells (19–21). These isoforms are mostly splicing variants of the ER that lacks some exons. An isoform of the ER was reported in the normal 1
To whom correspondence should be addressed. Fax: 011-81-3-5689-2483. Abbreviations: FBS, fetal bovine serum; VSMC, vascular smooth muscle cells; ER, estrogen receptor; RT-PCR; Reverse transcriptase-polymerase chain reaction; CAT, Chloramphenicol acetyltransferase. 766 0006-291X/96 $18.00 Copyright © 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
JOBNAME: BBRC 219#3 PAGE: 2 SESS: 4 OUTPUT: Mon Apr 1 16:38:35 1996 /xypage/worksmart/tsp000/68953f/19
Vol. 219, No. 3, 1996
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
tissue, in the brain, for the first time (22). This isoform of the ER lacks exon 4 that include both DNA-binding and estrogen-binding domains. Some of the isoforms of the ER have variant functions compared with the wild-type ER receptor (20, 21). It is possible that isoforms of the ER may modulate the estrogen action in the cardiovascular system. These circumstances prompted us to study the existence of isoforms of the ER in VSMC and possible involvement of the isoforms in the mechanism of estrogen action in VSMC. We performed reverse transcriptase-polymerase chain reaction (RT-PCR) using ER specific primers. Analysis of the PCR products has indicated the presence of three ER mRNA isoforms other than the wild-type ER in VSMC. The function of these ER isoforms was studied by chloramphenicol acetyltransferase (CAT) assay and one of them was shown to have an inhibitory effect on the wild-type ER in VSMC. MATERIALS AND METHODS Cell culture. Eight-week-old female rats (Nippon Bio-Supply Center) were sacrificed under ether anesthesia. VSMC were cultured from medial layer of the thoracic aorta by the method of Chamley et al. (23) Subcultured VSMC (4–6th passage) were used in the experiments. VSMC and A10 cells (a rat aortic smooth muscle cell-line) (24) were maintained in Dulbecco’s modified Eagle medium supplemented with 10% fetal bovine serum (FBS) (Cell Culture Laboratories) at 37°C in a humidified atmosphere of 5% CO2. A10 cells were suitable for transfection experiments using calcium-phosphate precipitation method (25). Reverse transcriptase-polymerase chain reaction (RT-PCR). PCR primers for the rat ER cDNA (18, 22) were synthesized as E-1 (nt 611-630) 59-CTACTACCTGGAGAACGAGC-39; E-2 (nt 762-779) 59-AAGGAGACTCGCTACTGT-39; E-6 (nt 1482-1501) 59-TCAAAGATCTCCACCATGCC-39; E-7 (nt 1650-1669) 59-ATCTTGTCCAGGACTCGGTG-39. RT-PCR was performed as described previously (16). cDNA was synthesized from 10 mg of total RNA of VSMC essentially according to Gubler and Hoffman (26) using the E-7 primer. Samples with or without treatment of avian myeloblastosis virus reverse transcriptase (Seikagaku Kougyo, Tokyo, Japan) were prepared as controls. One-tenth of the resulting cDNA was used as template DNA for the first PCR with E-1 and E-7 primers. One-tenth of the first PCR products was used as template DNA for the second PCR with E-2 and E-6 primers. The reaction was carried out in a final volume 20 mL containing template DNA, 10 pmol of each primer, 200 mmol/L dGTP, dATP, dTTP and dCTP, 10 mmol/L Tris-HCl (pH 8.3), 50 mmol/L KCl, 1.5 mmol/L MgCl2, 0.02% gelatin and 1U Taq DNA polymerase. The temperature program for the amplification was 30 cycles of 1 min at 94°C, 1 min at 57°C, and 2 min at 72°C. DNA sequencing and plasmid construction. PCR products were cloned into PCRII plasmid (Invitrogen) according to manufacturer’s instructions. The resulting recombinant DNA was purified and sequenced by dideoxy method (27) using sequenase (US Biochemical). Utilizing PCR fragments derived from these DNA and full length cDNA of the rat ER (18), the wild-type ER and three ER isoforms were constructed in an expression vector PSSRa (28) as PSSRaER, PSSRaERd4, PSSRaERD4/5, and PSSRaERD3/4. The resulting plasmids were confirmed by sequencing. Chloramphenicol acetyltransferase (CAT) assay. CAT assay was performed as described (29). Briefly, 1 × 106 of A10 cells were plated one day prior to transfection. One hour prior to transfection, the medium was replaced with phenol red free medium containing 10% dextran charcoal-treated FBS. By calcium-phosphate precipitation method (25), cells were transfected with 0.2 mg of expression plasmids (PSSRaER, PSSRaERD4, PSSRaERD4/5 or PSSRaERD3/4), 2 mg of vitERE-CAT reporter plasmid (30) containing the estrogen responsive element and 2 mg of pCH110 b-galactosidase expression plasmid (Pharmacia) used as an internal control to normalize for variations in transfection efficiency. In several experiments, indicated amount of expression plasmids for ER isoforms was added to wild-type PSSRaER expression plasmid (0.2 mg). The total amount of DNA transfected was made up to 20 mg with carrier DNA pGEM3Zf(-) (Promega). After 12 hour incubation the cells were cultured further with medium change in the absence or presence of 1 × 10−8 mol/L 17b-estradiol for 24 hours. The cell extracts were assayed for CAT activities as described (31). The experiment was carried out three times and a representative pattern is shown. To determine relative intensity of the signals, a macintosh computer with a scanner was utilized.
RESULTS Identification of the ER Isoforms in VSMC To investigate the expression of ER isoforms, we used two set of primers to increase the specificity of the amplified products. E-1, E-2, E-6 and E-7 primers are located in exon 1, exon 2, exon 6 and exon 7 of the ER gene, respectively. Total RNA (10 mg) from rat VSMC was converted into corresponding cDNA using reverse transcriptase (RT) and an ER gene-specific primer (E-7). The first PCR was performed utilizing the RT products as template DNA with E-1 and E-7 primers. 767
JOBNAME: BBRC 219#3 PAGE: 3 SESS: 4 OUTPUT: Mon Apr 1 16:38:35 1996 /xypage/worksmart/tsp000/68953f/19
Vol. 219, No. 3, 1996
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Then, the second PCR was carried out using the first PCR products as template DNA with E-2 and E-6 primers. The second PCR products were analyzed by 1% agarose gel electrophoresis (Fig. 1). As a result, other than a band corresponding to the PCR products derived from the wild-type rat cER (Fig. 1; Lane 2, 5: shown by arrowhead), shorter bands were detected in the RT-PCR products derived from VSMC RNA (Fig. 1; lane 2: shown by arrow). There was no band detected in samples without RT (Fig. 1; lane 3) and without template (Fig. 1 lane 4). Amplified PCR products were isolated from agarose gel and cloned into plasmid vector. Then, their sequences were determined by dideoxy method. The sequence of the long band corresponded to the wild-type ER cDNA contained all exon 3, exon 4 and exon 5. The exon-exon junctional sequence were shown in Fig. 2a. One of the sequence of the shorter bands indicated that the 59-termini of exon 5 was connected to the 39-termini of exon 3 (Fig. 2b), another of the sequence of the shorter bands indicated that the 59-termini of exon 6 was connected to the 39-termini of exon 4 (Fig. 2c), and the other of the sequence of the short bands showed that the 59-termini of exon 5 was connected to the 39-termini of exon 2 (Fig. 2d). Thus, they were designated as ERD4 isoform, ERD4/5 isoform and ERD3/4 isoform, respectively. A schema of the relationship of ER isoforms and functional domains of the ER is shown in Fig. 3. The exon-exon junctional sequences and corresponding amino acids are shown for ERD4, ERD4/5 and ERD3/4 isoforms. ERD4 and ERD3/4 isoforms encode the shorter protein that lack amino acids corresponding to those exons. In case of ERD4/5 isoform, a frame-shift occurred at the exon-exon junction and a premature stop codon appears. Thus, ERD4/5 isoform encodes a truncated protein. Inhibitory Effects of the ERD4/5 Isoform on the Wild-Type ER Action To examine the possible function of these ER isoforms, we performed CAT assay using ER isoform expression plasmids. It was shown that the wild-type ER expression plasmid have estrogen dependent transactivating activity and that ERD4, ERD4/5 and ERD3/4 isoforms have lost such activities (Fig. 4). To examine the effect of these isoforms on function of the wild-type ER, each ER isoform expression plasmid was cotransfected with the wild-type ER expression plasmid (Fig. 5A). In case of ERD4/5 the activation of ERE-tk-CAT by wild-type ER was significantly decreased when 50 times amount of isoform ER was added. To confirm this observation, the wild-type ER expression plasmid was cotransfected with increasing amount of ERD4/5 isoform expression plasmid (Fig. 5B). The activation of ERE-tk-CAT by the wild-type ER was decreased dose dependently with the ERD4/5 isoform. Relative values of acetylated chloramphenicol signals were shown in Fig 5B.
FIG. 1. Estrogen receptor (ER) isoforms detected in VSMC. Reverse transcriptase-polymerase chain reaction (RT-PCR) products were analyzed by 1% agarose gel electrophoresis. RT-PCR samples of the VSMC (VSMC) with (lane 2; RT+) and without (lane 3; RT-) reverse transcriptase, and without template (lane 4; NT) and PCR products from the rat ER cDNA (lane 5; rat cER) are shown. Lane 1 (M) shows the l/Hind III size marker. 768
JOBNAME: BBRC 219#3 PAGE: 4 SESS: 4 OUTPUT: Mon Apr 1 16:38:35 1996 /xypage/worksmart/tsp000/68953f/19
Vol. 219, No. 3, 1996
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. The sequence analysis of the reverse transcriptase-polymerase chain reaction products. Autoradiogram of exon–exon junctional sequence determined by dideoxy method for the wild-type estrogen receptor (ER) (a), ERD4 (b), ERD4/5 (c) and ERD3/4 (d) are shown.
DISCUSSION In the previous study (16), we demonstrated the presence of an estrogen receptor in vascular smooth muscle cells at the mRNA level by Northern blot analysis and RT-PCR. The presence of the ER protein was confirmed by immunocytochemistry utilizing anti-ER antibody. We examined
FIG. 3. A schematic representation of the estrogen receptor (ER) isoforms and functional domains of the ER. The position of the primers used in this experiment are shown by arrowheads. The exon–exon junctional sequences and corresponding amino acids of the ER isoforms are shown. Note that a frame-shift is occurred and premature stop codon appears encoding the truncated protein only in the ERD4/5 isoform. 769
JOBNAME: BBRC 219#3 PAGE: 5 SESS: 4 OUTPUT: Mon Apr 1 16:38:35 1996 /xypage/worksmart/tsp000/68953f/19
Vol. 219, No. 3, 1996
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 4. Chloramphenicol acetyltransferase (CAT) assay analysis of three isoforms of the estrogen receptor (ER) mRNA. The wild-type and isoform ER expression plasmids, PSSRaER (ER), PSSRaERD4 (D3/4), PSSRaERD4/5 (D4/5), PSSRaERD3/4 (D3/4) or PSSRa (vector) were transfected into A10 vascular smooth muscle cells with the vitERE-tk-CAT reporter plasmid. After culturing in the absence (−) or presence (+) of 10 nmol/L 17b estradiol (E2), CAT assay was performed.
the regulation of several estrogen responsive genes and found that the transcripts of c-fos protooncogene was regulated by estrogen in these cells. Karas et al. (17) also described that human vascular smooth muscle cells contain the ER both at the mRNA and the protein levels. They have shown by luciferase assay that the ER has estrogen dependent transactivation activity in those cells. These reports suggested that estrogen can act on the vascular smooth muscle cells directly and that
FIG. 5. Inhibitory effect of the ERD4/5 isoform on the wild-type estrogen receptor (ER) action. (A) The wild-type ER expression plasmid PSSRaR was cotransfected with indicated amount (the same or 50 times) of PSSRaERD4 (D4), PSSRaERD4/5 (D4/5) or PSSRaERD3/4 (D3/4). After culturing in the absence (−) or presence (+) of 10 nmol/L 17bestradiol (E2), chloramphenicol acetyltransferase (CAT) assay was performed. (B) The wild-type ER expression plasmid PSSRaER was cotransfected with indicated amount (1×, 3× 9× 27× and 81×) of PSSRaERD4/5 (D4/5). After culturing in the absence (−) or presence (+) of 10 nmol/liter 17b-estradiol (E2), CAT assay was performed. Relative values (transfection with the wild-type ER under the presence of E2 as 100) of acetylated cloramphenicol signals were shown at the top. 770
JOBNAME: BBRC 219#3 PAGE: 6 SESS: 4 OUTPUT: Mon Apr 1 16:38:35 1996 /xypage/worksmart/tsp000/68953f/19
Vol. 219, No. 3, 1996
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
the ER demonstrated in vascular smooth muscle cells can mediate estrogen action in the cardiovascular system. In the present study, we detected three isoforms of ER in the rat VSMC by RT-PCR. One is the isoform that lacks exon 4. Another isoform lacks both exon 4 and exon 5, and the other lacks exon 3 and exon 4. Although it is difficult to compare accurately the amount of mRNA by RT-PCR, the signal of shorter bands are comparable with that of the long band corresponding to the wild-type ER cDNA. From six clones derived from shorter bands, we obtained 3 clones of ERD4, 2 clones of ERD4/5 and 1 clone of ERD3/4. Thus, we estimate that the existence of the ERD4/5 isoform, that was identified for the first time, would not be negligible. In cancerous cells, several isoforms have been demonstrated, such as, ERD3, ERD4, ERD5, ERD7 isoforms. On the other hand, only few reports showed estrogen receptor isoforms in normal organs and tissues. Skipper et al. (22) has been reported the existence of ERD4 in the rat brain. We recently found ERD4 and ERD3/4 isoforms in the rat bone tissues (32). However, the function of these isoforms were still unknown. It has been reported that ERD3 isoform in the human breast cancer cells inhibits estrogen-dependent transcription activation when it is cotransfected with the wild-type ER and reporter plasmid (20). Both exon 3 and exon 4 contain a part of the DNA binding domain, and both exon 4 and exon 5 contain a part of the estrogen-binding domain (22, 33). Therefore, these three isoforms may lose binding activities to estrogen responsive elements and estrogen. CAT assay clearly demonstrated that these isoforms lost estrogen dependent transactivation activity via the estrogen responsive element. Furthermore, CAT assay cotransfected with the wild-type ER revealed one of these isoform, ERD4/5, has an inhibitory effect on normal estrogen action with a dominant negative fashion. It is notable that ERD4/5 encodes truncated protein because of a frame-shift occurred at the exon-exon junction. These results suggest that the truncated protein compete with the wild-type ER having the normal function. The data presented in this report are interesting because estrogen has bidirectional effects in various organs, tissues and cells. For example, estrogen promote the growth of breast cancer MCF-7 cells, whereas estrogen inhibit the growth of VSMC (Akishita et al., submitted). It is possible that the differential effects of estrogen are regulated by the receptor isoforms, such as ERD4/5 isoform that is a potential inhibitor of the wild-type ER function. Alternatively, the inhibitory effect of estrogen action by the ER isoform might be involved in pathophysiological conditions. Further studies should be required to clarify the roles of these ER isoforms in normal and atherosclerotic arteries. ACKNOWLEDGMENTS We thank Ms. M. Watanabe, Ms. H. Yamaguchi and Ms. M. Goto for technical assistance. This work was supported by grants from Foundation and the Ministry of Education, Science and Culture, Japan.
REFERENCES 1. Bush, T. L., Barrett-Conner, E., Cowan, L. D., Criqui, M. H., Wallace, R. B., Suchindran, C. M., Tyroler, H. A., and Rifkind, B. M. (1987) Circulation 75, 1102–1109. 2. Stampfer, M. J., Colditz, G. A., Willett, W. C., Manson, J. E., Rosner, B., Speizer, F. E., and Hennekens, C. H. (1991) N. Engl. J. Med. 325, 756–762. 3. Kushwaha, R. S., and Hazzard, W. R. (1981) Metabolism 30, 359–366. 4. Adams, M. R., Clarkson, T. B., Kaplan, J. R., and Koritnik, D. R. (1989) Transplant. Proc. 21, 3662–3664. 5. Cheng, L. P., Kuwahara, M., Jacobsson, J., and Foegh, M. L. (1991) Transplantation 52, 967–972. 6. Kannel, W. B., Hjortland, M. C., McNamara, P. M., and Gordon, T. (1976) Ann. Intern. Med. 85, 447–452. 7. Wittman, J. C. M., Grobbee, D. E., Kok, F. J., Hoffman, A., and Valkenburg, H. A. (1989) Br. Med. J. 198, 642–644. 8. Walsh, B. W., Schiff, I., Rosner, B., Greenberg, L., Ravnikar, V., and Sacks, F. M. (1991) N. Engl. J. Med. 325, 1196–1204. 9. Stumph, W. E., Sar, M., and Aumüller, G. (1977) Science 196, 319–321. 10. Horwitz, K. B., and Horwitz, L. D. (1982) J. Clin. Invest. 69, 750–758. 11. Colburn, P., and Buonassisi, V. (1978) Science 201, 817–819. 771
JOBNAME: BBRC 219#3 PAGE: 7 SESS: 4 OUTPUT: Mon Apr 1 16:38:35 1996 /xypage/worksmart/tsp000/68953f/19
Vol. 219, No. 3, 1996 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33.
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Nakao, J., Chang, W.-C., Murota, S.-I., and Orimo, H. (1981) Atherosclerosis 38, 75–80. Ross, R. (1993) Nature 362, 801–809. Fischer-Dzoga, K., Wissler, R. W., and Vesselinovitch, D. (1983) Exp. Mol. Pathol. 39, 355–363. Vargas, R., Wroblewska, B., Rego, A., Hatch, J., and Ramwell, P. W. (1993) Br. J. Pharmacol. 109, 612–617. Orimo, A., Inoue, S., Ikegami, A., Hosoi, T., Akishita, M., Ouchi, Y., Muramatsu, M., and Orimo, H. (1993) Biochem. Biophys. Res. Commun. 195, 730–736. Karas, R. H., Patterson, B. L., and Mendelsohn, M. E. (1994) Circulation 89, 1943–1950. Koike, S., Sakai, M., and Muramatsu, M. (1987) Nucleic Acids Res. 15, 2499–2513. Dotzlaw, H., Alkhalaf, M., and Murphy, L. C. (1991) Mol. Endcrinol. 6, 773–785. Wang, Y., and Miksicek, R. J. (1991) Mol. Endcrinol. 5, 1707–1715. Fuqua, S. A. W., Chamnes, G. C., and McGuire, W. L. (1993) J. Cell. Biochem. 51, 135–139. Skipper, J. J., Young, L. J., Bergeron, J. M., and Tetzlaff, M. T. (1992) Proc. Natl. Acad. Sci. USA. 90, 7172–7175. Chamley, J. H., Campbell, G. R., McConnell, J. D., and Gröschel-Stewart, U. (1977) Cell Tiss. Res. 177, 503–522. Kimes, B. W., and Brandt, B. L. (1976) Exp. Cell Res. 98, 349–366. Graham, F. L., and van der Eb, A. J. (1973) Virology 52, 456–467. Gubler, V., and Hoffman, B. J. (1983) Gene 25, 456–467. Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. Acad. Sci. USA. 74, 5463–5467. Takebe, Y., Seiki, M., Fujisawa, J.-I., Hoy, P., Yokote, K., Arai, K.-I., Yoshida, M., and Arai, N. (1988) Mol. Cell. Biol. 8, 466–472. Inoue, S., Orimo, A., Hosoi, T., Kondo, S., Toyoshima, H., Kondo, T., Ikegami, A., Ouchi, Y., Orimo, H., and Muramatsu, M. (1993) Proc. Natl. Acad. Sci. USA. 90, 11117–11121. Inoue, S., Kondo, S., Hashimoto, M., Kondo, T., and Muramatsu, M. (1991) Nucleic Acids Res. 19, 4091–4096. Gorman, C. M., Moffat, L. F., and Howard, B. H. (1982) Mol. Cell. Biol. 2, 1044–1051. Hoshino, S., Inoue, S., Hosoi, T., Saito, T., Ikegami, A., Kaneki, M., Ouchi, Y., and Orimo, H. (1995) Calcif. Tissue Int. 57, 466–468. Ponglikitmongkol, M., Green, S., and Chambon, P. (1988) EMBO J. 7, 3385–3388.
772