Smad3 Is Involved in the Intracellular Signaling Pathways That Mediate the Inhibitory Effects of Transforming Growth Factor-ß on StAR Expression
C. Brand, S. Souchelnytskiy,* E. M. Chambaz, J-J. Feige, and S. Bailly1
INSERM Unité 244, Commissariat à l’Energie Atomique, Département de Biologie Moléculaire et Structurale, Biochimie des Régulations Cellulaires Endocrines, 17 rue des martyrs, F-38054 Grenoble, France; and * Ludwig Institute for Cancer Research, Box 595, S-75 124 Uppsala, Sweden
Received October 8, 1998
Transforming growth factor ßs (TGFBs) constitute a family of dimeric proteins that regulate growth and differentiation of many cell types. TGFß1 is also a potent autocrine regulator of adrenocortical steroido- genesis. We have recently shown that in primary cul- tures of bovine fasciculo-reticularis cells, the main tar- get of TGFß is the steroidogenic acute relay protein (StAR), a key protein necessary for intramitochon- drial cholesterol transport. Here, we show that StAR expression is also inhibited by TGFß1 in the human adrenocortical carcinoma cell line NCI-H295R. This inhibitory effect is mediated by Smad proteins. In- deed, we found that overexpression of wild-type Smad3 inhibited endogenous StAR mRNA expression while overexpression of a dominant negative Smad3 protein reversed the inhibitory effect of TGFß1 on StAR mRNA expression. Taken together, these results demonstrate that the Smad3 protein is involved in TGFß-dependent regulation of steroidogenesis. @ 1998 Academic Press
Transforming growth factor-ß1 belongs to a family of peptides regulating cell growth, differentiation and ex- tracellular matrix production (1). Its roles in growth inhibition of many epithelial cells, potentiation of wound repair, angiogenesis and immunomodulation are among its many documented actions. TGFß1 has also been identified as a physiological autocrine/ paracrine regulator of adrenocortical steroidogenic functions (2, 3). TGFß1 has been found to modify sev-
eral steps in the steroidogenic pathway (see 2 for re- view). Recently, we have reported that TGFß inhibits the expression of the steroidogenic acute regulatory protein (StAR) in bovine adrenocortical cells (4). This effect of TGFß is of high importance, as StAR is a key regulatory protein implicated in cholesterol transport from the cytoplasm to the inner mitochondrial mem- brane where it is metabolized into pregnenolone by P450scc (see 5 for review). The intramitochondrial transport of cholesterol is the rate-limiting step in ste- roidogenesis and the main site for regulation by phys- iological stimuli during acute stimulation of steroid production. The evidence that StAR is critical for ste- roid hormone production has been derived, in part, from the demonstration that mutations in the StAR gene cause congenital lipoid adrenal hyperplasia, a disease in which adrenal and gonadal steroid synthesis are severely impaired at the P450scc step (6). Targeted disruption of the murine StAR gene results in a phe- notype in homozygous null mutants similar to congen- ital lipoid adrenal hyperplasia in humans (7).
TGF& family members exert their cellular effects by binding transmembrane receptors that possess serine/ threonine kinase activity (8, 9, 10). Upon ligand bind- ing, a heteromeric receptor complex consisting of two type II and two type I receptors is formed. Within the complex, the type I receptor is phosphorylated and activated by the type II receptor. Recently, our under- standing of TGFß intracellular signaling has improved dramatically through studies of TGFß-like signaling in genetically accessible species (11). Genetic studies in Drosophila melanogaster and Caenorhabditis elegans have led to the identification of a conserved family of proteins termed Smads, which perform essential roles in intracellular signaling downstream of serine/ threonine kinase receptors. In vertebrates, at least nine homologues have been identified (8, 9, 10). Smad proteins identified thus far can be divided into three
1 Corresponding author. INSERM U244, DBMS/BRCE, CEA-G, 17 rue des martyrs, 38054 Grenoble Cedex 9, France. Fax: 33 476 88 50 58. E-mail: Sbailly@geant.ceng.cea.fr.
This work was supported in part by INSERM, the Commissariat à l’Energie Atomique (CEA/DSV/DBMS), the Ligue Nationale Contre le Cancer, and the Association pour la Recherche contre le Cancer. Céline Brand is a recipient of a doctoral grant from the CEA.
groups: pathway-restricted Smads which each interact with only one TGFß subtype (Smad2 and Smad3 are specific mediators of TGFß and activin pathways, whereas Smad1, Smad5, Smad8 and MADH6/Smad9 are involved in BMP signaling); mediator Smads (Smad4), and inhibitory Smads (Smad6 and Smad7). Upon heteromeric complex formation, the pathway- restricted Smads interact transiently with specific ac- tivated type I receptors and thus become phosphory- lated at their C-terminus. Following phosphorylation these pathway-restricted Smads interact with the com- mon mediator Smad4 and translocate to the nucleus to regulate gene transcription. Smad2/Smad4, together with Fast-1, a winged-helix DNA binding protein, as- sociate into activin response factor (ARF) that binds to the Xenopus laevis Mix.2 promoter (12). Recently a Smad3/Smad4 binding sequence termed the ‘CAGA’ box was identified in the PAI-1 (13) and JunB promot- ers (14).
In this report, we have investigated the regulation of StAR expression in the human adrenocortical tumor cell line NCI-H295R. We show that TGF61 inhibits StAR expression in this cell line at the level of both mRNA and protein. Furthermore, we show that tran- sient overexpression of a dominant-negative Smad3 partially blocked the action of TGF1. Overexpression of the wild type Smad3 was sufficient to mimic the inhibitory effect of TGFß1 on StAR expression. This report demonstrates the involvement of Smad proteins in the TGFß1-dependent regulation of steroidogenesis.
EXPERIMENTAL PROCEDURES
Materials. Recombinant TGFß1 was purchased from R&D Sys- tems (Abingdon, UK). Ham’s F12, DMEM, Trypsin/EDTA 0.25% were purchased from Life Technologies, Inc. (Cergy-pontoise, France), ITS+ from Beckton Dickinson Labware (Bedford, USA) and Ultroser SF from Biosepra (Villeneuve la Garenne, France). Forsko- lin was purchased from Sigma (Saint Guentin Fallavier, France).
p3TPLux was a generous gift from Dr. Joan Massagué (New York, NY) and pcDNA3mycSmad3 from Dr. Masahiro Kawabata (Tokyo, Japan). pcDNA3flagSmad2, and pcDNA3flagSmad4 are N-terminally tagged Smads and were a generous gift from Dr. Serhiy Souchelnytskyi (Uppsala, Sweden). pcDNA3Smad3AC is a carboxy-terminally trun- cated Smad3 which acts as a dominant-negative (15) and was kindly provided by Dr. R. Derynck (San Francisco, CA).
Cell culture. NCI-H295R cells (16) were kindly supplied to us by Dr. W. Rainey (Dallas, TX). They were maintained in Dulbecco’s Modified Eagle’s-Ham’s F-12 medium (DMEM/F-12) supplemented with 1 % ITS+ (insulin 1 µg/ml, transferrin 1 µg/ml, selenium 1 ng/ml, linoleic acid 1 µg/ml final concentrations), 2 % Ultroser SF and antibiotics.
RNA preparation and Northern blot analysis. Total RNA was isolated from cells using the RNAgents kit (Promega, Charbonnières, France). 20 µg of total RNA were separated by electrophoresis through a 1% agarose gel containing 1.9 % formaldehyde and trans- ferred to a Hybond-N membrane (Amersham, Les Ulis, France). Blots were hybridized sequentially with a partial bovine StAR cDNA, a full-length human PAI-1 cDNA (a generous gift from Denis Vivien (Caen, France)) and an 18S rRNA probe that were labeled by random priming with [a-32P]dCTP (111 Tbq/mmol, ICN Pharmaceuticals,
Orsay, France) using the Radprime DNA labeling kit (Life Technol- ogies, Inc., Cergy-pontoise France). Prehybridization and hybridiza- tion at 65°℃ were performed in Rapid-Hyb buffer (Amersham, Les Ulis, France). Hybridizing bands were visualized on a ß-imager (Phosphorimager, Molecular Dynamics, Sunnyvale, CA) and quanti- fied using the ImageQuantTM program (Molecular Dynamics, Sunnyvale, CA). Values for StAR and PAI-1 mRNAs were normal- ized to values for the 18S rRNA.
Western blotting. NCI-H295R cells were scraped and homoge- nized with a Potter-Kontes homogenizer (1200 rpm, 35 strokes) in 5 mM Tris-HCl, pH 7.4, 275 mM sucrose. The homogenate was centri- fuged at 500 x g for 15 min to remove large debris and nuclei. Mitochondria were collected by centrifugation at 10,000 x g for 10 min and washed once with the same buffer. Mitochondrial proteins (20 µg/lane) were resolved on 12% SDS-PAGE denaturing gels and transferred to PVDF (Polyvinylidene difluoride) polyscreen mem- brane (NEN Life Science Products, Boston, USA) by electroblotting. Following transfer, the membrane was incubated with an antiserum against a peptide fragment (amino acids 88-98) of the murine 30 kDa StAR (1/1000, 17). StAR protein was detected with a secondary antibody coupled to peroxidase and visualized with chemilumines- cence. Quantitations were done by densitometry.
Transfections. NCI-H295R cells were transfected by electropora- tion (250 mV, 450 µF) in a buffer containing 250 mM K2HPO4, 100 mM CH3COOK, KOH 100 mM, 26 mM MgSO4, pH 7.4 using a gene pulser apparatus (Eurogentec, Angers, France).
For reporter gene activity determination, transfections were per- formed with 2 µg of the p3TPLux plasmid and 1.2 µg of pCMV5BGal or the different Smad expressing plasmids and pSG5 for a total of 6 µg DNA. Twenty-four hours after transfections, cells were incubated with or without TGFB1 (2 ng/ml) for 24 hours. The luciferase activ- ities using luciferin and renilla as substrate were measured on a LUMAT LB 9507 luminometer (EGG-Berthold, Bad Wildbad, Ger- many). The luciferase activity from luciferin as substrate was nor- malized to that from renilla as substrate to compensate for varia- tions in transfection efficiency, and the results are shown as the mean ± SEM of three samples.
For cell-sorting experiments, transfections were performed with 2 µg of the pEGFP vector (Clontech Laboratories, Inc, Palo Alto, USA) and 4 µg of the Smad cDNA constructs (pCDNA3mycSmad3 or pCDNA3Smad34C) or pCMV5ßGal plasmid. Twenty four hours af- ter transfection, EGFP-positive cells were sorted using a FACSstar cell sorter (Beckton Dickinson, Pont de Claix, France) with a stan- dard excitation wavelength of 488 nm. The EGFP-positive cells were put back into culture overnight and then treated with forskolin (25 p.M) in the presence or the absence of TGFß1 (2 ng/ml) for 12 hours.
RESULTS
TGFß1 inhibits StAR mRNA expression. We re- cently observed that TGF31 inhibited the expression of StAR in primary cultures of bovine adrenocortical (fasciculata-reticularis) cells (4). In order to generalize this observation, we studied the effect of TGFß1 on NCI-H295R cells, a human adrenocortical carcinoma cell line (16). Northern blot analysis of StAR mRNA expression in NCI-H295R cells revealed the presence of a single mRNA transcript of 1.6 kb. NCI-H295R cells were then treated for 12 or 24 h with TGFB1 (2 ng/ml) in the presence or the absence of forskolin (25 M). As shown in Figure 1, TGFß1 inhibited StAR mRNA ex- pression in both unstimulated (43% and 41% inhibition at 12 and 24 h, respectively) and forskolin-stimulated cells (20% and 32% inhibition at 12 and 24 h, respec-
A
Time (h)
12
24
12
24
TGFß1
-
+
+
- + - +
Forskolin
-
-
-
+ + + +
1.6 kb
18S
B
100
% StAR mRNA level
300
80
60
200
40
100
20
0
CTL
TGFß
CTL
TGFß
0
Forskolin
Forskolin + TGFß
Forskolin
Forskolin + TGFß
12 h
24 h
12 h
24 h
tively). Additional experiments indicated that the in- hibition was observed as of 6 h of TGFB1 treatment (data not shown).
TGFß1 inhibits StAR protein expression. To deter- mine whether TGFß1 affects StAR protein expression, we performed immunoblot analysis on mitochondrial proteins from NCI-H295R cells that had been incu- bated for 12 or 24 h with or without forskolin in the presence or the absence of TGFß1. As shown in Figure 2, TGFß1 inhibited StAR protein expression in both unstimulated (37% and 37% inhibition at 12 and 24 h, respectively) and forskolin-stimulated cells (23% and 25% inhibition at 12 and 24 h, respectively). Altogether these results show that TGF31 inhibits StAR mRNA and protein expression in the human NCI-H295R cell line.
Smad3 is involved in the TGFß-induced decrease of StAR expression. Smad proteins have been found to be major components of intracellular TGFß signaling. Thus,
we studied involvement of Smads in TGFß signaling in NCI-H295R. We first tested whether transfection of TGFß-specific Smad expression vectors (Smad2, Smad3 or Smad4) into NCI-H295R cells would modify the ex- pression of p3TPLux, a TGFß-responsive luciferase re- porter gene which bears a part of the plasminogen- activator inhibitor-1 (PAI-1) promoter. Smad3 was the most potent among them (Table 1). However, cotransfec- tion of Smad expression constructs and a luciferase re- porter gene driven by the human StAR promoter led to non-specific “squelching” in NCI-H295R. Therefore, we chose to study the effects of Smad3 on endogenous StAR mRNA levels. As a control for Smad3 activity in our experimental conditions, we looked at its effect on endog- enous PAI-1 mRNA levels. To do this, NCI-H295R cells were transiently co-transfected with either pcDNA3mycSmad3 or pCMV5ßgalactosidase (as a con- trol for expression of an unrelated protein) expression plasmids and an EGFP expression plasmid. After 24 h of
A
Time (h)
12
24
12
24
TGFß1
-
+
-
+
-
+
-
+
Forskolin
-
-
-
-
+
+
+
+
28.5 kDa
B
100
400
IOD (% control)
80
300
60
200
40
20
100
0
CTL
TGFB
CTL
TGFB
0
Forskolin
Forskolin + TGFß
Forskolin
Forskolin + TGFß
12 h
24 h
12 h
24 h
transfection, the EGFP-positive cells were sorted using a FACS sorter and cultured overnight. The cells were then incubated for 12 h with forskolin (25 (M) in the presence or the absence of TGFB1 (2 ng/ml). Total RNA was then extracted and Northern Blot analysis was performed us- ing StAR and PAI-1 cDNA probes. Results are shown as mRNA/18S ratios normalized to the same ratio from un- treated cells at 12 h. Table 2 shows that overexpression of
the wild type mycSmad3 mimics the effect of TGFß on endogenous PAI-1 expression, although the level of PAI-1 mRNA induction was slightly lower than that observed with 2 ng/ml TGFß1 (6.4 versus 10.8, respectively). More interestingly, we found that wild type Smad3 also mim- icked the inhibitory effect of TGFß1 on endogenous StAR mRNA expression. As was found for PAI-1, the effect of the wild type Smad3 on StAR mRNA expression was
| CTL | TGFß1 | |||
|---|---|---|---|---|
| RLUª (mean ± sd) | fold induction by Smadsb | RLUª (mean ± sd) | fold induction by Smadsb | |
| none | 850 ± 294 | 1 | 11612 ± 670 | 13.7 |
| flagSmad2 | 1277 ± 182 | 1.5 | 19854 ± 629 | 23.4 |
| mycSmad3 | 22375 ± 3457 | 26.3 | 25465 ± 1688 | 30.0 |
| flagSmad4 | 9114 ± 503 | 10.7 | 18076 ± 913 | 21.3 |
| Smad3AC | 1135 ± 134 | 1.3 | 4365 ± 254 | 5.1 |
Note. p3TPLUX was transfected into NCI-H295R cells together with the indicated Smad constructs. Twenty-four hours after transfection, the cells were incubated in the absence or the presence of TGF31 (2 ng/ml) for another 24 h.
a Results are expressed as relative light units of luciferase activity as the mean ± sd of one experiment carried out with triplicate cultures. b Fold induction with the different Smad expression plasmids compared to CTL without Smad constructs.
| Plasmid co-transfected with EGFP | StAR expression (StAR mRNA/18S) | PAI-1 expression (PAI-1 mRNA/18S) | ||
|---|---|---|---|---|
| pCMV5ßgal | pcDNA3 mycSmad3 | pCMV5ßgal | pcDNA3 mycSmad3 | |
| TGFß1 | - + | - | - + | - |
| Experiment 1 | 1 0.50 | 0.71 | 1 11.5 | 5.4 |
| Experiment 2 | 1 0.66 | 0.75 | 1 10 | 7.3 |
Note. NCI-H295R cells were co-transfected with pEGFP and pCMV5ßgal or pcDNA3mycSmad3. After 24 h, EGFP-positive cells were sorted using a FACS scan and put back in culture overnight. These cells were then incubated with or without TGFß1 (2 ng/ml) for 12h in the presence of forskolin (25 uM). Total RNA was hybridized sequentially with 32P-labeled partial bovine StAR cDNA, full-length human PAI-1 cDNA and 18S rRNA probes. Hybridization signals were quantified using a ß-imager. Results are expressed as ratios of StAR/18S or PAI-1/18S hybridization signals normalized to those obtained for the control condition (i.e. transfection of pCMV5ßgal plasmid in absence of TGFß1).
slightly lower than that observed with TGFß1 (0.73 ver- sus 0.58, respectively). We then tested the effect of a carboxy-terminally truncated TGFß-dominant negative Smad3 (15), pcDNASmad3AC, in the presence or absence of TGFB1. Table 3 shows that overexpression of pcDNA3smad3AC partially blocked the inhibitory effect of TGFß1 on endogenous StAR expression (29% inhibi- tion with TGFß1 vs 13% inhibition in the presence of TGFß1 and Smad3AC).
DISCUSSION
This work demonstrates that the inhibitory effect of TGFß1 on StAR expression previously observed in bo- vine fasciculo-reticularis (4) and glomerulosa (C. Brand, unpublished data) cells can be extended to hu- man adrenocortical cells.
Smad proteins are a new family of transcription fac- tors that have been specifically implicated in the
| Plasmid co-transfected with EGFP | StAR expression (StARmRNA/18S) | ||
|---|---|---|---|
| pCMV5ßgal | pcDNA3Smad34C | ||
| TGFß1 | - + | - | + |
| Experiment 1 | 1 0.66 | 1 | 0.90 |
| Experiment 2 | 1 0.76 | 1 | 0.83 |
Note. NCI-H295R cells were co-transfected with pEGFP and pCMV5ßgal or pcDNA3Smad3AC. After 24 h, EGFP positive cells were sorted using a FACS scan and put back in culture overnight. These cells were then incubated with or without TGFB1 (2 ng/ml) for 24 h in the presence of forskolin (25 µM). Total RNA was hybridized sequentially with 32P-labeled partial bovine StAR cDNA and 18S rRNA probes. Hybridization signals were quantified using a B-imager. Results are expressed as ratios of StAR/18S hybridization signals normalized to those obtained for the control condition (i.e. absence of TGF(1).
TGFß-family transduction pathway (8, 9, 10). How- ever, despite intensive study, there is no evidence so far of an implication of Smad2/3 as intermediates in TGFß signaling pathways leading to inhibitory effects of TGFß on gene expression. Most studies have concen- trated on the induction of genes encoding either extra- cellular matrix proteins such as PAI-1, or cell cycle regulatory proteins (18). Further, most of these obser- vations were made in a very limited number of cell types: Mv1Lu, HepG2, NIH-3T3 and MDA-MB468, a breast carcinoma cell line lacking Smad4. We investi- gated the involvement of Smad proteins in an inhibi- tory effect of TGF, namely StAR inhibition, in the human adrenocortical cell line, NCI-H295R.
We found that overexpression of a dominant- negative Smad3 partially blocked the action of TGFß1. Overexpression of the wild type Smad3 was sufficient to mimic the inhibitory effect of TGFß1 on StAR ex- pression. Our data demonstrate that Smad3 is a signal transduction intermediate in TGFß’s inhibition of StAR expression in adrenocortical cells.
How does Smad3 regulates StAR expression? Sev- eral hypotheses can be made. The first possibility could be that Smad3 interacts directly with the human StAR promoter. A Smad3/Smad4 DNA binding sequence was recently reported as a “CAGA” box within the human PAI-1 (13) and the mouse JunB (14) promoters. By sequence analogy we have found a “CAGA” box in the human promoter of StAR. Further experiments will be needed to see if Smad proteins can bind to this se- quence. An interaction between Smad3 and the CREB binding protein (CBP) or its analog p300 (19) has re- cently been described. Utilization of CBP/p300 as a coactivator places the Smad proteins in a network of transcription factors that all compete for limited amounts of CBP/p300 and may thus influence each other’s activity by squelching. Among the many inter- actants that have been described for CBP/p300 are members of the nuclear receptors of the steroid hor- mone superfamily, and in particular SF-1 (20), which
plays a fundamental role in StAR expression (21). In our case, we could imagine that Smad proteins inhibit SF-1 activation and therefore StAR expression by de- pletion of the available CBP/p300.
However, both models suggest a direct effect of Smads on the promoter of StAR while we have previously shown that TGFß1 inhibition of StAR mRNA expression re- quires de novo protein synthesis. Since Smad proteins described so far seem to be regulated through phosphor- ylation, it is more likely that TGFß activates the Smad pathway that in turn induces the synthesis of a protein. One potential candidate for this relay protein could be the immediate early gene JunB. Indeed, JunB has been shown to be specifically induced by TGFß-family mem- bers in several cell types (22) and in particular in bovine adrenocortical cells (23) and this was later reinforced by the presence of a “CAGA” box within its promoter (14). Furthermore, overexpression of JunB has been shown to mimic BMP2-induction of myogenic differentiation (22) and TGFß-induction of collagenase (24).
Altogether, this work demonstrates the involvement of Smad proteins as the first step in the pathway of TGFB inhibition on StAR mRNA expression. Further work will be needed to continue down the signaling cascade. This work demonstrates, for the first time, the implication of Smad in inhibitory effect of TGFß, and in a different cell type than usually studied, adrenocortical cells.
ACKNOWLEDGMENTS
We are grateful to Dr. N. Cherradi (Faculty of Medicine, Geneva) for the generous gift of anti-StAR antibody. The expert technical assistance of P. Claustre for FACS cell sorting is gratefully aknowl- edged. We thanks Dr. Anna Chinn for careful review of this manu- script.
REFERENCES
1. Roberts, A., and Sporn, M. (1993) Growth Factors 8, 1-9.
2. Feige, J-J., Vilgrain, I., Brand, C., Bailly, S., and Souchelnytskiy, S. (1998) J. Endocrinol. 158, 7-19.
3. Le Roy, C., Leduque, P., Dubois, P. M., Saez, M., and Langlois, D. (1996) J. Biol. Chem. 271, 11027-11033.
4. Brand, C., Cherradi, N., Defaye, G. Chinn, A., Chambaz, E. M., Feige, J-J., and Bailly, S. (1998) J. Biol. Chem. 273, 6410-6416.
5. Stocco, D. M., and Clark. B. J. (1997) Steroids 62, 29-36.
6. Lin, D., Sugawara, T., Strauss III, J. F., Clark, B. J., Stocco, D. M., Saenger, P., Rogol, A., and Miller, W. L. (1995) Science 267, 1828-1831.
7. Caron, K. M., Soo, S-C., Wetsel, W. C., Stocco, D. M., Clark, B. J., and Parker, K. (1997) Proc. Natl. Acad. Sci USA 94, 11540- 11545.
8. Kretzschmar, M., and Massagué, J. (1998) Current Opinion in Genetics & Development 8, 103-111.
9. Attisano, L., and Wrana, J. L. (1998) Current Opinion in Cell Biology 10, 188-194.
10. Derynck, R., Feng X. H. (1997) Biochim. Biophys. Acta 1333, F105-F150.
11. Padgett, R. W., Das, P., and Krishna, S. (1998) BioEssays 20, 382-391.
12. Chen, X., Weisberg, E., Fridmacher, V., Watanabe, M., and Whitman, M. (1997) Nature 389, 85-89.
13. Dennler, S., Itoh, S., Vivien, D., ten Dijke, P., Huet, S., and Gauthier, J-M. (1998) EMBO J. 17, 3091-3100.
14. Jonk, L. J. C., Itoh, S., Heldin, C-H., ten Dijke, P., and Kruijer, W. (1998) J. Biol. Chem. 273, 21145-21152.
15. Zhang, Y., Feng, X-H., Wu, R-Y., and Derynck, R. (1996) Nature 383, 168-172.
16. Rainey, W. E., Bird, I. M., and Mason, J. I. (1994) Mol. Cell. Endocrinol. 100, 45-50.
17. Cherradi, N., Rossier, M. F., Valloton, M. B., Timberg, R., Fried- berg, I., Orly, J., Wang, X. J., Stocco, D. M., and Capponi, A. M. (1997) J. Biol. Chem. 272, 7899-7907.
18. Liu, X., Sun, Y., Constantinescu, S. N., Karam, E., Weinberg, R. A., and Lodish, H. F. (1997) Proc. Natl. Acad. Sci. USA 94, 10669-10674.
19. Janknecht, R., Wells, N. J., and Hunter, T. (1998) Genes & Dev. 12, 2114-2119.
20. Carlone, D. L., and Richards, J. S. (1997) Mol. Endocrinol. 11, 292-304.
21. Sugawara, T., Kiriakidou, M., McAllister, J. M., Holt, J. A., Arakane, F., and Strauss III, J. F. (1997) Steroids 62, 5-9.
22. Chalaux, E., Lopez-Rovira, T., Rosa, J. L., Bartrons, R., and Ventura, F. (1998) J. Biol. Chem. 273, 537-543.
23. Viard, I., Jaillard, C., and Saez, J. M. (1993) FEBS Lett. 328, 94-98.
24. Mauviel, A., Chung, K-Y., Agarwal, A., Tamai, K., and Uitto, J. (1996) J. Biol. Chem. 271, 10917-10923.