Estradiol-induced Phosphorylation of Serine 118 in the Estrogen Receptor Is Independent of p42/p44 Mitogen-activated Protein Kinase*

Phosphorylation of Ser118 of human estrogen receptor α (ER) enhances ER-mediated transcription and is induced by hormone binding and by activation of the mitogen-activated protein kinase (MAPK) pathway. We discovered that phosphorylation of Ser118 reduces the electrophoretic mobility of the ER. Using this mobility shift as an assay, we determined the in vivo stoichiometry and kinetics of Ser118 phosphorylation in response to estradiol, ICI 182,780, epidermal growth factor (EGF), and phorbol 12-myristate 13-acetate (PMA). In human breast cancer MCF-7 cells, estradiol induced a steady state phosphorylation of Ser118 within 20 min with a stoichiometry of 0.67 mol of phosphate/mol of ER. Estradiol did not activate p42/p44 MAPK, and basal p42/p44 MAPK activity was not sufficient to account for phosphorylation of Ser118 in response to estradiol. In contrast, both EGF and PMA induced a rapid, transient phosphorylation of Ser118 with a stoichiometry of ∼0.25, and the onset of Ser118 phosphorylation correlated with the onset of p42/p44 MAPK activation by these agents. Either the EGF- or PMA-induced Ser118 phosphorylation could be inhibited without influencing estradiol-induced Ser118phosphorylation. The data suggest that a kinase other than p42/p44 MAPK is involved in the estradiol-induced Ser118phosphorylation. We propose that the hormone-induced change in ER conformation exposes Ser118 for phosphorylation by a constitutively active kinase.

Estrogen receptor ␣ (ER) 1 is a member of a superfamily of nuclear receptors that act as transcription factors when bound to specific lipophilic hormones (1). In common with other steroid hormone receptors, the ER has an N-terminal domain with a hormone-independent transcriptional activation function (AF-1), a central DNA-binding domain, and a C-terminal ligand-binding domain with a hormone-dependent transcriptional activation function (AF-2) (1,2). In addition to being activated by binding to its ligand, estradiol, the ER can also be activated as a transcription factor by several nonsteroidal agents (2).
Phosphorylation of the ER increases in response to estradiol (2). A major site of estradiol-induced phosphorylation is serine 118 (Ser 118 ), which is located in AF-1 (3)(4)(5). In a number of cell types, mutation of Ser 118 to Ala causes a reduction in estradioldependent transcriptional activation (4, 6 -8). Ser 118 is located within the sequence PPQLSPFLQ which has a high degree of homology with the optimum peptide substrate identified for p44 MAPK (9). Ser 118 is also phosphorylated in response to EGF and PMA, possibly via p42/p44 MAPK (6,7). Transcriptional activation of the ER by EGF is diminished when Ser 118 is mutated to Ala (6 -8). Thus both estradiol binding and activation of the MAPK pathway lead to phosphorylation of Ser 118 , and this phosphorylation is involved in regulating ER-mediated transcription.
The evidence that p42/p44 MAPK directly phosphorylates Ser 118 is based on in vitro observations that activated p42/p44 MAPK phosphorylated the N-terminal domain of the ER but not a mutant of this construct in which Ser 118 was replaced with Ala (6). Additional supporting evidence is provided by the observation that in vivo a constitutively activated MAPK kinase (MAPKK) stimulated phosphorylation of the N-terminal domain of the ER but did not stimulate phosphorylation of a mutant of this construct in which Ser 118 was replaced with Ala (6). However, it has still not been established whether p42/p44 MAPK directly phosphorylates Ser 118 in the context of the full-length ER in vivo. We demonstrate that the onset of Ser 118 phosphorylation correlates with the onset of p42/p44 MAPK activation in response to EGF and PMA. These data are consistent with the hypothesis that p42/p44 MAPK phosphorylates Ser 118 . However, our data demonstrate that p42/p44 MAPK is not responsible for the observed Ser 118 phosphorylation in response to estradiol binding by the ER. Our results raise the possibility that Ser 118 may be the site of phosphorylation by either two or more kinases in response to various stimuli.

EXPERIMENTAL PROCEDURES
Chemicals-ICI 182,780 was a gift from Zeneca Pharmaceuticals, Cheshire, United Kingdom.
Expression Vectors and Receptor Mutants-Pierre Chambon (Unité 184 de Biologie Moléculaire et de Génétique de I'INSERM, Strasbourg Cédex, France) kindly provided pSG5-ER (the human estrogen receptor cDNA (10)). The mutants S118A-ER, S104/106/118A-ER, S154A-ER, and S167A-ER were created by site-directed mutagenesis of pSG5-ER using the megaprimer method (11). The luciferase reporter plasmid, 2ERE-MpGL2, has been described (8). pSI-␤-galactosidase contains the ␤-galactosidase gene in the pSI vector from Promega. * This work was supported by National Institutes of Health Grant R29CA-55887 and National Science Foundation Grant OSR-935-0540. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Cell Culture, Transfection, Cell Harvest, and Immunoblotting-COS-1, MCF-7, T-47D, and NMuMG cell lines were obtained from American Type Culture Collection, Manassas, VA. All media used in this study were phenol-red free, and all fetal bovine serum used was charcoal-dextran-treated (CFBS). COS-1, T-47D, and NMuMG cells were maintained in DMEM/Ham's Nutrient Mix F-12 supplemented with 10% CFBS. COS-1 media contained 50 units/ml penicillin and 50 g/ml streptomycin. T-47D and NMuMG cell media contained antibiotic-antimycotic (Life Technologies, Inc.). MCF-7 cells were maintained in DMEM supplemented with 5% CFBS, 50 units/ml penicillin, and 50 g/ml streptomycin. Cells were incubated 24 h in media without serum prior to treatment with steroids or other agents unless indicated otherwise.
COS-1 cells were transfected by the diethylaminoethyl-dextran method as modified by Reese and Katzenellenbogen (12). Each 6-cm plate was transfected with 0.5 g of pSG5-ER or pSG5-ER mutant plus 2 g of pGL2-Basic Vector (Promega) as carrier DNA. Following dimethyl sulfoxide treatment, the cells were incubated for 24 h in serumfree medium before treatment with estradiol or PMA and harvesting.
Following treatment with steroids or agents as indicated in the figure legends, the cells (on 6-cm plates) were washed with warm PBS and 0.5 ml of 2ϫ SDS-PAGE sample buffer (0.12 M Tris-Cl, pH 6.8, 4% SDS, 20% glycerol, 0.2 M DTT, 0.008% bromphenol blue) near 100°C was then added to the plate. The cells were scraped into a 1.5-ml centrifuge tube in a heat block at 100°C. The tubes were heated for 10 min, cooled on ice, and stored at Ϫ20°C until samples were analyzed on SDS-PAGE.
At the time of electrophoresis, samples were reheated for 5 min at 100°C, and 30-l aliquots were layered on 7% SDS-PAGE and electrophoresed until the 43-kDa marker reached the bottom of the gel. Proteins on the gel were transferred to nitrocellulose membranes in 25 mM Tris-HCl, pH 8.3, 192 mM glycine, and 20% methanol. For detection of the ER, immunoblotting was performed with monoclonal antibody ␣78 -1-C6 (mAb 78) (13) for COS-1 extracts or monoclonal antibody EVG F9 (mAb F9) (14) for MCF-7, T-47D, or NMuMG extracts. For detection of activated p42/p44 MAPK, immunoblotting was performed with anti-ACTIVE MAPK (Promega). For detection of total p42/p44 MAPK, immunoblotting was performed with ANTI-pan ERK (Transduction Laboratories). All immunoblots used either anti-mouse or anti-rabbit IgG conjugated to horseradish peroxidase as a secondary antibody. Antigenantibody interactions were detected with NEN Life Science Products Western chemiluminescence reagent. Varying exposures of the film were taken. A Molecular Dynamics Personal Densitometer with Image-Quant software was used to determine the relative amount of ER in the immunoblot bands. To ensure that a linear response over the range of ER concentrations obtained from the experiments was achieved, a standard curve relating the amount of ER to the densitometric reading was performed. The relative intensities of the ER bands were quantitated from various exposures of the film.
Immunoprecipitation and Protein Phosphatase 1 Treatment-MCF-7 cells were treated for 38 min with 1 nM estradiol and then rinsed with cold PBS. The cells were lysed in buffer A (50 mM Tris-Cl, pH 8, 0.15 M NaCl, 1% Nonidet P-40 (Nonidet P-40), 20 mM Na 2 MoO 4 , 20 mM NaF, 2 mM Na 3 VO 4 , 20 mM ␤-glycerophosphate, 1 mM phenylmethylsulfonyl fluoride, 2 g/ml each of pepstatin, leupeptin, and aprotinin, 200 nM okadaic acid, 10 mM MgCl 2 , and 100 units/ml DNase 1 (Life Technologies, Inc.)). The ER was immunoprecipitated with mAb ␣78 as described (5). The immunoprecipitates were washed with protein phosphatase 1 (PP1) buffer (New England Biolabs), and two aliquots were suspended in a final volume of 30 l of PP1 buffer with 1 mM MnCl 2 . One tube received 1 unit of PP1, and both tubes were incubated for 5 min at 30°C. Reactions were stopped by addition of SDS-PAGE sample buffer and heating for 5 min at 100°C. Sample proteins were analyzed on SDS-PAGE, electrotransferred, and immunoblotted as described above.
Transcriptional Analysis-For transcriptional analysis, MCF-7 cells were seeded at a density of 5 ϫ 10 5 cells per 6-cm dish in DMEM containing 5% CFBS and antibiotic-antimycotic the day before transfection. Cells were transfected by 15 l of lipofectAMINE in 2 ml of DMEM according to the instructions provided by Life Technologies, Inc. with 0.5 g of 2ERE-MpGL2 (luciferase reporter plasmid), 1.0 g of pSI-␤-galactosidase (for normalizing luciferase data), and 1.5 g of pGL2-Basic Vector (carrier DNA). The cells were left overnight in the incubator and washed once with PBS before the addition of DMEM with the appropriate concentration of estradiol, modulators, or vehicle. After treatment for 5.5 h, the cells were lysed with Reporter Lysis Buffer (Promega). Luciferase and ␤-galactosidase activity were measured as described previously (5).

Development of an Assay for the Detection of Ser 118
Phosphorylation-Many proteins migrate more slowly on SDS-PAGE due to phosphorylation of particular sites. Estradiol addition to cells causes the ER to migrate as a doublet on SDS-PAGE (5,8,16). We showed that in human ER expressed in COS-1 cells the major difference in the phosphorylation of the upper and lower bands was phosphorylation of Ser 118 in the upper band. Furthermore, a mutant ER in which Ser 118 was replaced with Ala was phosphorylated at a level only 20% that of the wild type (5) and did not upshift when cells were treated with estradiol (5). Thus, phosphorylation of Ser 118 appeared to be required for the electrophoretic upshift. This electrophoretic upshift could be used to monitor the phosphorylation of Ser 118 , greatly facilitating the characterization of kinases that phosphorylate Ser 118 . Therefore, we wanted to confirm that only the phosphorylation of Ser 118 was required for the electrophoretic upshift.
Phosphorylation of the ER occurs only on Ser residues (3,17,18) with the exception of a low level of phosphorylation on Tyr 537 in the human ER (19). Le Goff et al. (3) reported that a mutant of the ER lacking the A/B region (amino acids 1-175) showed no estradiol-induced phosphorylation, indicating that the hormone-induced phosphorylation sites are predominantly in the A/B region. To determine whether phosphorylation of another Ser in the A/B region in addition to Ser 118 is required for the electrophoretic upshift, we prepared mutants in which other Ser residues in the A/B region known or suspected to be phosphorylated were replaced by Ala residues (2). Ser-104 and Ser-106 were replaced with Ala residues to create the mutant S104/106A-ER. Ser-118, Ser-154, and Ser-167 were individually mutated to Ala to create the mutants S118A-ER, S154A-ER, and S167A-ER. The wild type and mutant ERs were transfected into COS-1 cells. Cell lysates were immunoblotted with an anti-ER antibody. Data displayed in Fig. 1, lanes 1-4 show that after subtraction of the control, estradiol converted 30% of the wild type ER to the upshifted form, PMA converted about 65% of the receptor, and estradiol plus PMA converted 83% of the receptor. In contrast, Fig. 1, lanes 5-8 show that neither estradiol nor PMA caused a change in mobility with the mutant S118A-ER. All of the other mutants, S167A-ER (lanes 9 -12), S154A-ER (lanes [13][14][15][16], and S104/106A-ER (lanes [17][18][19][20] responded to estradiol and PMA similarly to the wild type. Thus, the change in mobility of the ER did not require phosphorylation of Ser residues 104, 106, 154, or 167. Ser 118 Phosphorylation of the ER in Mammary Cell Lines-To determine whether estradiol causes phosphorylation of Ser 118 in endogenous ER in mammary cell lines, MCF-7 (human breast carcinoma), NMuMG (mouse mammary gland normal epithelium), and T-47D (human breast duct carcinoma, pleural effusion), cells were incubated for 30 min with estradiol or vehicle. Estradiol induced an electrophoretic upshift of the ER in all three mammary cell lines ( Fig. 2A, immunoblot).
To confirm that the upshift was due to phosphorylation, MCF-7 cells were treated with estradiol, and the ER was immunoprecipitated. Aliquots of the immunoprecipitate were incubated with and without PP1. ER incubated without PP1 (Fig.  2B, lane 1) had 35% of the ER upshifted, whereas only 3% of the ER was upshifted (lane 2) in ER incubated with PP1. The amount of ER in the lower band in lane 2 is equivalent to the amount of ER in the two bands in lane 1 combined, indicating that dephosphorylation of the ER in the upper band converted it into the lower band. The fraction of ER recovered in the upper band following immunoprecipitation was less than when the cells were treated directly with hot SDS sample buffer ( Fig.  2A), presumably due to the action of phosphatases during immunoprecipitation. These results together with the data with mutants confirm that Ser 118 was phosphorylated in endogenous ER in human breast cells and most likely on the analogous Ser 122 in mouse mammary cells.
Phosphorylation of Ser 118 in Response to Estradiol and ICI 182,780 of the ER-To determine the dose response of Ser 118 phosphorylation to estradiol in vivo, MCF-7 cells were treated with increasing concentrations of estradiol. Estradiol-induced phosphorylation of Ser 118 was maximal at approximately 1 nM estradiol, with an average of 67% of the receptor phosphorylated on Ser 118 (Fig. 3). Thus the stoichiometry of Ser 118 phosphorylation is higher in cells that normally express ER as compared with heterologously expressed ER in COS-1 cells.
To determine whether ICI 182,780 induces Ser 118 phosphorylation, MCF-7 cells were treated with various concentrations of ICI 182,780. Interestingly, although ICI 182,780 is a pure ER antagonist (20), 30 nM of this compound induced phosphorylation of Ser 118 to the same extent as estradiol (Fig. 3).
We found that ICI 182,780 caused a rapid decrease in total ER protein, an observation in agreement with reports for ICI 164,384 (20,21). Treatment of MCF-7 cells with 10 nM estradiol for 1 h did not significantly alter the total amount of ER protein, whereas treatment with 10 nM ICI 182,780 for 1 h reduced the total amount of ER protein by 70% compared with the control level (data not shown).
Induction of Ser 118 Phosphorylation in Response to Estradiol, EGF, and PMA-Ser 118 is also phosphorylated in response to EGF and PMA, possibly via p42/p44 MAPK (6 -8). Therefore, we wanted to compare the time course and stoichiometry of Ser 118 phosphorylation in response to both EGF and PMA as compared with estradiol. Within 5 min of estradiol treatment, phosphorylation of Ser 118 was already noticeable (Fig. 4A) and was maximal within 20 min, remaining at that maximal level at 60 min. In response to EGF, approximately 24% of the receptor was phosphorylated on Ser 118 , and this maximal level was reached within 5 min after addition of ligand (Fig. 4B). In response to PMA, approximately 26% of the receptor was phosphorylated on Ser 118 , and phosphorylation reached a maximum with either wild type ER or with one of four mutants: S118A-ER, S154A-ER, S167A-ER, or S104/106A-ER. Twenty-four h after transfection, cells were treated for 30 min with 50 nM estradiol (E), 100 nM PMA (P), 50 nM estradiol plus 100 nM PMA (EϩP), or vehicle (C). Cells were harvested by addition of 2ϫ SDS-loading buffer at 100°C. Soluble cell lysate was resolved on SDS-PAGE, transferred to nitrocellulose, and the ER was detected by immunoblotting.

FIG. 2. Ser 118 phosphorylation of the ER in mammary cell lines.
A, serum-starved MCF-7, NMuMG, and T-47D cells were treated for 30 min with 100 nM estradiol (E) or vehicle (C). Lysates were analyzed as described in Fig. 1. B, serum-starved MCF-7 cells were treated for 38 min with 1 nM estradiol. ER was immunoprecipitated and divided into two aliquots that were incubated for 5 min at 30°C with or without PP1. The reaction was stopped, the immunoprecipitates were resolved on SDS-PAGE and transferred to nitrocellulose, and the ER was detected by immunoblotting. level within 10 min (Fig. 4C). Thus, the extent and time course of induction of Ser 118 phosphorylation differed considerably between estradiol, EGF, and PMA. Estradiol causes a rapid but prolonged phosphorylation, whereas EGF and PMA cause rapid and transient phosphorylation of Ser 118 .
Next, to ensure that we were studying Ser 118 phosphorylation under conditions in which EGF and PMA influence ERmediated transcription, we determined the effect of both EGF and PMA on ER-dependent transcription in MCF-7 cells. These cells were cotransfected with a reporter construct that contained a luciferase gene regulated by two palindromic estrogen response elements (ERE) and a plasmid that constitutively expresses ␤-galactosidase. This latter construct allows the luciferase data to be normalized for transfection efficiency. In parallel experiments, the cells were cotransfected with a reporter construct that did not contain the EREs. These experiments were necessary to eliminate transcriptional responses that were independent of the ERE. Following transfection, the cells were treated for 5.5 h with PMA, EGF, or vehicle with or without estradiol. PMA and EGF caused a small steroid-independent activation of transcription from the ERE-containing reporter plasmid of 1.9-and 1.8 times, respectively (Fig. 5). Estradiol alone stimulated transcription by 5.5 times. Adding PMA and estradiol together resulted in 11.7 times activation, indicating that their effects are synergistic. A combination of EGF and estradiol modestly increased transcription by 1.2 times compared with estradiol alone. Thus ligand-dependent and -independent activation of ER-dependent transcription occurs in MCF-7 cells.
Estradiol Does Not Activate p42/p44 MAPK, whereas EGF and PMA Both Activate p42/p44 MAPK-It has been reported that estradiol can stimulate signal transduction pathways that lead to p42/p44 MAPK activation in MCF-7 cells (22). To investigate the possibility that p42/p44 MAPK was responsible for Ser 118 phosphorylation in response to estradiol, serumstarved MCF-7 cells were treated with either estradiol, EGF, or PMA, and Ser 118 phosphorylation and p42/p44 MAPK activation were determined. As Fig. 6A shows, the onset of Ser 118 phosphorylation paralleled the onset of activation of p42/p44 MAPK by EGF. The activated form of p42/p44 MAPK was detected by using an antibody that specifically recognizes the dually phosphorylated form of p42/p44 MAPK (23). The rapid onset of phosphorylation of Ser 118 also paralleled the onset of p42/p44 MAPK activation by PMA. However, activation of p42/ p44 MAPK was not observed in response to estradiol. The immunoblot with an antibody that recognizes both activated and unactivated forms of p42/p44 MAPK demonstrated that approximately equal levels of p42/p44 MAPK were present in each lane. In other experiments, extended exposure of the anti-active MAPK immunoblot demonstrated the presence of basal p42/p44 MAPK activity. However, this low level of activity was not enhanced by estradiol (Fig. 7A).
As an alternative approach to determine whether estradiol activates p42/p44 MAPK, an in-gel kinase assay was performed using extracts from MCF-7 cells that had been treated with estradiol, PMA, EGF, or vehicle. The substrate embedded in the gel was myelin basic protein. Both PMA and EGF stimulated a protein kinase of ϳ42 kDa (Fig. 6B, lanes 1, 3, and 5  versus 4 and 6), which indicates that p42/p44 MAPK was active in the in-gel kinase assay. However, estradiol did not stimulate p42/p44 MAPK or any other protein kinases to phosphorylate To correct for differences in transfection efficiency, the luciferase activities were normalized to the ␤-galactosidase activities. To correct for transcriptional responses not due to the ER, the normalized luciferase activity of cells transfected with MpGL2 was subtracted from the normalized luciferase activity of cells transfected with 2ERE-MpGL2. The data are reported as the % of response obtained compared with the response obtained with estradiol. The means Ϯ S.E. are shown. ** indicates a p value of Ͻ0.005, and * indicates a p value of Ͻ 0.05. For p values, PMA and EGF were compared with control, and estradiol with either EGF or PMA were compared with estradiol. myelin basic protein (Fig. 7B, lanes 1, 3, and 5 versus 2). Thus two different approaches both indicated that estradiol did not activate p42/p44 MAPK.

EGF-and PMA-induced Ser 118 Phosphorylation Can Be Distinguished from Estradiol-induced Ser 118 Phosphorylation-
The above results suggest that activation of p42/p44 MAPK may be involved in the phosphorylation of Ser 118 in response to EGF and PMA but not in response to estradiol. To further test this hypothesis we compared the effects of various inhibitors on either the EGF-or PMA-induced Ser 118 phosphorylation versus the estradiol-induced Ser 118 phosphorylation. Fig. 7A shows that PD98059, an inhibitor of MAPK-activating enzyme (MEK1/2) (24), diminished the EGF-induced Ser 118 phosphorylation but did not affect the estradiol-induced Ser 118 phosphorylation. Extended exposure of the anti-active MAPK immunoblot showed faint bands corresponding to active p42 MAPK and p44 MAPK. This basal activity of p42/p44 MAPK diminishes in extracts taken from cells pretreated with PD98059 before addition of estradiol or EGF. Estradiol did not increase p42/p44 MAPK activity above the basal level. Importantly, even though PD98059 was able to diminish basal p42/ p44 MAPK activity, PD98059 did not effect estradiol-induced Ser 118 phosphorylation.
It has been suggested that there are two MEK1/2-independent pathways for activating p42/p44 MAPK; one pathway involves PKC, and the other pathway involves phosphatidylinositol 3Ј-kinase (25). Both staurosporine and GF109203X, inhibitors of PKC (26,27), completely prevented phosphorylation of Ser 118 in response to PMA but had no effect on Ser 118 phosphorylation induced by estradiol (Fig. 7, B and C). Pretreatment of MCF-7 cells for 30 min with 100 nM wortmannin, conditions shown to inhibit phosphatidylinositol 3Ј-kinase (28), did not reduce estradiol-induced Ser 118 phosphorylation (data not shown). An inhibitor combination of GF109203X, PD98059, and wortmannin also did not affect estradiol-induced Ser 118 phosphorylation (data not shown). Pretreatment of MCF-7 cells for 30 min with 20 M SB203580, conditions shown to inhibit the activation of p38 MAPK (29), had no effect on the estradiolinduced Ser 118 phosphorylation (data not shown). These results provide additional evidence that estradiol does not activate a signal transduction pathway that leads to p42/p44 MAPK or p38 MAPK activation.
Of interest to note is that Trowbridge et al. (30) have suggested that the cyclin A⅐cyclin-dependent kinase 2 (cdk2) complex phosphorylates the ER. However, staurosporine, which inhibits a number of kinases including PKC and cdk2 at similar concentrations (26), had no affect on estradiol-induced phosphorylation (Fig. 7C). Moreover, pretreatment of MCF-7 cells for 30 min with 100 M olomoucine, a condition that inhibits cdk1 and cdk2 (31), also did not affect estradiol-induced Ser 118 phosphorylation (data not shown). These data do not support a model in which cdk2 is the kinase responsible for Ser 118 phosphorylation in response to estradiol binding by the ER.
Basal Activity of p42/p44 MAPK Is Not Responsible for Estradiol-induced Ser 118 Phosphorylation-Our results demonstrate that p42/p44 MAPK is not activated in response to estradiol. It could be argued, however, that the estradiol-bound receptor is an extremely good substrate for p42/p44 MAPK and that even the activity of p42/p44 MAPK in the presence of PD98059 is sufficient for 67% of the ER molecules to be phosphorylated in response to estradiol. To test this hypothesis, we determined whether increasing p42/p44 MAPK activity would enhance estradiol-induced Ser 118 phosphorylation. Increasing amounts of EGF were added to MCF-7 cells that had been pretreated with a high concentration of estradiol for an extended period of time. This pretreatment was performed to ensure that ER binding to estradiol was not rate-limiting for Ser 118 phosphorylation. The lowest dose of EGF did not effect FIG. 6. p42/p44 MAPK is not activated in MCF-7 cells in response to estradiol. A, serum-starved MCF-7 cells were treated for the min indicated with 100 nM estradiol, 100 nM PMA, 100 ng/ml EGF, or vehicle (for 0 min value). Lysates were resolved on either 7% SDS-PAGE (for ER) or 10% SDS-PAGE (for p42/p44 MAPK). After transfer to nitrocellulose, the ER was detected with an anti-ER antibody, and active p42/p44 MAPK was detected with an antibody that recognizes only the dually phosphorylated form of p42/p44 MAPK. This later blot was stripped and reprobed with anti-p42/p44 MAPK antibody that detects both activated and nonactivated p42/p44 MAPK. B, extracts were made from MCF-7 cells that were serum-starved for 24 h and then treated for 5 min with 10 nM estradiol (E), 100 ng/ml EGF, 100 nM PMA, or vehicle (C). These extracts were used in an in-gel kinase assay containing myelin basic protein in the gel. The relative positions of the molecular mass standards are shown. The arrow indicates the position of p42/p44 MAPK activated by PMA and EGF. estradiol-induced Ser 118 phosphorylation even though p42/p44 MAPK activity was significantly increased over the basal level (Fig. 8). Higher levels of EGF which dramatically enhanced p42/p44 MAPK activity only increased Ser 118 phosphorylation to the same extent as EGF alone and did not induce 100% of the estradiol-bound ER molecules to be phosphorylated (Fig. 8). DISCUSSION Phosphorylation of Ser 118 is involved in regulating the transcriptional activity of AF-1 in response to either estradiol or EGF (6 -8). Although Ser 118 had been identified as a putative substrate for p42/p44 MAPK, no information concerning the stoichiometry and kinetics of Ser 118 phosphorylation in response to estrogen agonists or antagonists, EGF, or PMA was known. We developed a convenient and quantitative electrophoretic mobility assay to monitor specifically the in vivo phosphorylation of Ser 118 . Using this assay, we provide the first evidence that phosphorylation of Ser 118 occurs in human breast cancer cells, a physiologically relevant target tissue. The rapidity of Ser 118 phosphorylation in response to estradiol suggests that the phosphorylation is a direct response to hormone binding. Le Goff et al. (3) have shown that ICI 164,384 increases phosphorylation of the ER but they did not determine the sites of phosphorylation. Interestingly, we have found that both estradiol and ICI 182,780 induce Ser 118 phosphorylation to a similar extent. The receptor is known to have a different conformation after it binds agonist as opposed to antagonist (32). However, the conformational difference induced by estradiol and ICI 182,780 does not seem to affect the N-terminal region of the receptor.
Although EGF and PMA only transiently induce Ser 118 phosphorylation, these agents are able to stimulate ER-mediated transcription. In agreement with published reports, we observed that EGF and PMA with or without estradiol enhanced ER-mediated transcription (7,(33)(34)(35). Phosphorylation of Ser 118 appears to be important for ER-mediated transcription since a mutant ER in which Ser 118 is replaced with Ala has a diminished transcriptional ability compared with the wild type ER (4, 6 -8). One possible model to explain the mechanism by which EGF and PMA can enhance ER-mediated transcription is that transient Ser 118 phosphorylation is sufficient to induce a cascade of events that eventually lead to transcriptional activation. Of interest to note is that a combination of EGF and estradiol additively enhance ER-mediated transcription, whereas a combination of PMA and estradiol synergistically FIG. 7. The inhibitors PD98059, staurosporine, and GF109203X prevent Ser 118 phosphorylation induced by either PMA or EGF but not that induced by estradiol. Serum-starved MCF-7 cells were used in all experiments. In panel A, cells were pretreated with 100 M PD98059 or vehicle for 30 min. Then either 1 nM estradiol or vehicle was added for 20 min, or 100 ng/ml EGF or vehicle was added for 5 min. In panel B, cells were pretreated 30 min with 5 M staurosporine or vehicle. Then 1 nM estradiol, 100 nM PMA, or vehicle was added for 15 min. In panel C, cells were pretreated for 30 min with 10 M GF109203X or vehicle. Then 1 nM estradiol, 100 nM PMA, or vehicle was added for 15 min. Lysates were analyzed as described in Fig. 6. The percent upshift of the ER was determined from immunoblots by densitometry. The means Ϯ S.D. are shown.
FIG. 8. Basal activity of p42/p44 MAPK is not responsible for estradiol-induced Ser 118 phosphorylation. Serum-starved MCF-7 cells were pretreated for 30 min with 10 nM estradiol. Varying amounts of EGF were then added for 5 min. Lysates were analyzed as described in Fig. 6. The percent upshift of the ER was determined from immunoblots by densitometry. The means Ϯ S.D. are shown. effect transcription. It is possible that EGF only enhances Ser 118 phosphorylation, whereas PMA can also influence another step in ER-mediated transcription. Thus the effects of EGF and estradiol are additive because both agents are influencing the same event, whereas the effects of PMA and estradiol are synergistic because these agents are influencing multiple steps that are involved in ER-mediated transcription.
Our results with EGF and PMA are consistent with Ser 118 being a substrate for activated p42/p44 MAPK (6, 7). In contrast to our results with EGF and PMA, we obtained three different lines of evidence that p42/p44 MAPK is not the kinase responsible for the estradiol-induced Ser 118 phosphorylation. First, activation of p42/p44 MAPK activity was not detectable in serum-starved MCF-7 cells treated with estradiol. These results are in disagreement with the work of Miagliaccio et al. (22). Second, basal level of p42/p44 MAPK is not sufficient to account for the estradiol-induced Ser 118 phosphorylation. This reasoning is based on the observation that decreasing the basal p42/p44 MAPK activity with PD98059 had no effect on estradiol-induced Ser 118 phosphorylation. Furthermore, if basal p42/ p44 MAPK activity was responsible for the estradiol-induced Ser 118 phosphorylation, then elevation of p42/p44 MAPK activity should result in 100% of the estradiol-bound ER molecules being phosphorylated. This result was not observed. Third, inhibitors that diminished either the EGF or PMA enhancement of p42/p44 MAPK activity and induction of Ser 118 phosphorylation did not affect the estradiol-induced phosphorylation. Taken together, our results strongly suggest that in vivo Ser 118 of the human ER is a target for two or more kinases that are regulated differentially by second messengers.
Trowbridge et al. (30) reported that cdk2 enhances both ligand-dependent and -independent transcription by ER. An in vitro kinase assay demonstrated that Ser 104 and Ser 106 of the human ER were the likely targets of cdk2 activity. However, we did not observe any effect of the inhibitors staurosporine and olomoucine which inhibit cdk2 on estradiol-induced Ser 118 phosphorylation. Like Ser 118 , Ser 104 and Ser 106 are adjacent to proline residues. In our 32 P-labeling experiments, Ser 104 and Ser 106 were minor sites of phosphorylation compared with Ser 118 (5,8).
This paper provides the first evidence that more than one kinase regulates Ser 118 phosphorylation of the human ER. One simple model to explain the estradiol-induced phosphorylation of Ser 118 is that, upon hormone binding, the ER undergoes a conformational change that makes the ER a better substrate for a constitutively active kinase. Phosphorylation of Ser 118 may allow the ER to integrate signals from hormone binding and activation of the MAPK pathway.