E6AP and Calmodulin Reciprocally Regulate Estrogen Receptor Stability*

Estrogen promotes the proliferation of human breast epithelial cells by interacting with the estrogen receptor (ER). Physiological responses of cells to estrogen are regulated in part by degradation of the ER. Previous studies revealed that calmodulin binds directly to the ER, thereby enhancing its stability. Consistent with these findings, cell-permeable calmodulin antagonists dramatically reduced the number of ER in MCF-7 human breast epithelial cells. Here we investigated the molecular mechanism by which calmodulin attenuates ER degradation. MG132 and lactacystin, inhibitors of the ubiquitin-proteasome pathway, prevented the calmodulin antagonist CGS9343B from reducing the amount of ER in MCF-7 cells. In contrast, protease inhibitors afforded no protection. Moreover, CGS9343B enhanced ER ubiquitination. A point mutant ER construct that is unable to bind calmodulin, termed ERΔCaM, is ubiquitinated to a greater extent than wild type ER. The ubiquitin-protein isopeptide ligase E6-associated protein (E6AP) associated with and promoted the degradation of ER. The possible convergence of calmodulin and E6AP on ER degradation was examined. ERΔCaM bound E6AP with higher affinity than that of wild type ER. Moreover, calmodulin attenuated the in vitro interaction between ER and E6AP in a Ca2+-dependent manner. Collectively, our data reveal that E6AP is a component of ER degradation via the ubiquitin-proteasome pathway and that Ca2+/calmodulin modulates this degradation mechanism. These results have potential implications for the development of selectively targeted therapeutic agents for breast cancer.

Estrogen receptors (ERs) 2 are ligand-activated transcription factors that mediate the actions of estrogen, thereby participating in metabolism, cell growth and proliferation, reproduction, and development (1)(2)(3). Like the other members of the steroid hormone receptor superfamily, the classical mode of action of ER involves binding of the ligandreceptor complex to specific DNA response elements within the promoters of target genes. Transcriptional activation is subsequently activated or repressed as a result of DNA-bound receptors recruiting chromatin remodeling complexes, coactivator and corepressor pro-teins, and the transcriptional initiation machinery (4). It has been known for some time that estradiol (E 2 ) significantly decreases the concentration of ER in cultured breast epithelial cells (5). More recent evidence from several investigators revealed that E 2 -induced degradation of ER is mediated by the ubiquitin-proteasome pathway (6 -9). Most interestingly, the mechanism underlying ER destruction produced by E 2 is distinct from that produced by the selective ER modulators fulvestrant (ICI 182,780) or GW5638 (10).
A growing body of evidence strongly supports a role for Ca 2ϩ and calmodulin in several aspects of ER function (11). Initial findings indicated that ER isolated from rat uterus cystol bound calmodulin-Sepharose in vitro in the presence of Ca 2ϩ (12,13). Subsequently, we documented by coimmunoprecipitation that endogenous ER binds endogenous calmodulin in human breast epithelial cells (11). In addition, calmodulin was shown to increase the affinity of ER for the estrogen-response element (ERE) (14,15). Previous work from our laboratory documented that nuclear calmodulin is required for E 2 to stimulate transcription by ER (16). In addition, we demonstrated that calmodulin binds directly to ER, thereby stabilizing ER (11). In this study, we set out to identify the molecular mechanism by which calmodulin promotes ER stability. We observed that inhibitors of the proteasome pathway, but not protease inhibitors, prevented the cell-permeable calmodulin antagonist CGS9343B from reducing the amount of ER in cell lysates. Consistent with these findings, ubiquitination of ER was enhanced when the interaction between ER and calmodulin was attenuated either with CGS9343B or with a point mutant ER construct that is unable to bind calmodulin. Moreover, reduction of binding to calmodulin enhanced the binding of ER to the E3 ubiquitin ligase E6-associated protein (E6AP). Thus, calmodulin modulates ER degradation via the ubiquitin-proteasome pathway, at least in part by regulating the interaction between ER and E6AP.

EXPERIMENTAL PROCEDURES
Materials-Tissue culture reagents were purchased from Invitrogen, and fetal bovine serum (FBS) was obtained from BioWhittaker. Charcoal-treated FBS was from Cocalico Biologicals Inc. MCF-7 and T47D breast epithelial cells and COS-7 green monkey kidney cells were obtained from American Type Culture Collection. CGS9343B was generously donated by Drs. E. Moret and B. Schmid (Novartis, Switzerland). Lactacystin, MG132, calpeptin, and the calpain inhibitor II N-acetylleucine-leucine-methioninal (ALLM) were purchased from Calbiochem. Protein A-and protein G-Sepharose were from Amersham Biosciences. FuGENE 6 was obtained from Roche Applied Science. Polyvinylidene difluoride (PVDF) membrane was purchased from Millipore Corp. Nonimmune rabbit serum (NIRS), 17␤-estradiol (E 2 ), and trifluoperazine (TFP) were from Sigma. Purified human ER and pig brain calmodulin were purchased from Panvera and Ocean Biologics, respectively.
Antibodies-Anti-ER monoclonal (Ab-15) and polyclonal antibodies were from Neomarkers and Santa Cruz Biochemistry, respectively. Anti-His and anti-HA (12CA5) antibodies were from Roche Applied Science. Anti-Myc antibody was from Maine Biotechnology Inc. The anti-E6AP antibody has been described previously (17).
A deletion mutant and a point mutant ER unable to bind calmodulin were developed. These constructs are described in detail elsewhere (22). Briefly, ER⌬298 -317 was generated by deleting amino acid residues 298 -317 from pcDNA3-myc-ER by using PCR. Site-directed mutagenesis of pcDNA3-myc-ER was performed with the QuikChange site-directed mutagenesis kit (Stratagene). The mutant cDNA was amplified by PCR with Pfu turbo DNA polymerase using the oligonucleotides 5Ј-CCAAGCCCGCTCATGGACGAACGCTCTAAGAAGA-3Ј (mutated residues are underlined). These changes result in replacement of Ile-298 with Glu and Lys-299 with Asp. The construct is termed ER⌬CaM. The sequence of both constructs was confirmed by DNA sequencing.
Cell Culture and Transfection-MCF-7 and COS-7 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% (v/v) FBS. T47D cells were grown in RPMI 1640 medium supplemented with 10% (v/v) FBS. Cells were plated in 100-mm dishes at 37°C in 5% CO 2 , and 24 h later, 4 -8 g of DNA was transiently introduced into cells using FuGENE 6 according to the manufacturer's instructions. When cells were incubated with E 2 , the culture medium was replaced 24 h prior to the treatment with phenol red-free Dulbecco's modified Eagle's medium containing 10% charcoal-treated FBS (11). Cells were lysed between 24 and 48 h post-transfection in buffer A (50 mM Tris-HCl, pH 7. 5, 150 mM NaCl, 0.1% Triton X-100, 0.1% protease inhibitor mixture (Sigma), and 1 mM phenylmethylsulfonyl fluoride (PMSF)) containing 1 mM CaCl 2 . Lysates were incubated on ice for 5 min and sonicated. Debris was cleared by centrifugation for 5 min at 4°C.
Treatment with Cell-permeable Agents-Where indicated, cells were incubated with 40 M CGS9343B, 10 M TFP, 20 M ALLM, or 10 M calpeptin for 16 h. When proteasome inhibitors were used, cells were preincubated for 2 h with 10 M lactacystin or 10 M MG132 prior to other treatments.
Immunoprecipitation-For immunoprecipitation, cells were lysed in buffer B (50 mM Tris-HCl, pH 8, 0.5 mM NaCl, 1% Nonidet P-40, 1 mM dithiothreitol, 0.1% protease inhibitor mixture, and 1 mM PMSF), and immunoprecipitation was performed essentially as described previously (23). Briefly, after pre-clearing with protein A-Sepharose beads, equal amounts of protein lysate were immunoprecipitated with anti-ER or anti-E6AP antibodies for 2 h at 4°C. NIRS was used as a control. Immune complexes were collected for 2 h with 40 l of protein A-Sepharose and washed five times with buffer B, and samples were processed by Western blotting as described below.
Western Blotting-Equal amounts of protein lysate and immune complexes were resolved by SDS-PAGE and transferred to PVDF membrane. Membranes were blocked with 5% nonfat powdered milk in TBS-T buffer (25 mM Tris, pH 8. 0, 140 mM NaCl, 2.5 mM KCl, and 0.05% Tween 20) and probed with the antibodies as indicated in the figure legends. Complexes were visualized with the appropriate horse-radish peroxidase-conjugated secondary antibody and developed by enhanced chemiluminescence.
In Vivo Ubiquitination-COS-7 cells were cotransfected with pMPT107 (His 6 -tagged ubiquitin) and pcDNA3-ER or pcDNA3-ER⌬CaM. Forty eight hours post-transfection, cells were treated for 2 h with 10 M MG132 or vehicle, followed by further incubation with 40 M CGS9343B or an equal volume of vehicle (ethanol) for 16 h. Lysates were harvested in 1 ml of buffer C (50 mM Tris, pH 7. 5, 150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, 0.1% protease inhibitor mixture, and 1 mM PMSF). After preclearing with 20 l each of protein G-and protein A-Sepharose beads, equal amounts of protein lysate were immunoprecipitated with anti-His or anti-ER antibodies for 4 h at 4°C. Immune complexes were collected and washed five times with buffer C, and the samples were processed by Western blotting.
Reaction mixtures contained 10 mM HEPES, pH 7. 9, 0. 5 mM EDTA, 5 mM MgCl 2 , 2 mM NaF, 2 mM ATP, 60 mM KCl, 10 M biotin-tagged ubiquitin (BostonBiochem), 200 nM E1-His, 5 M UbcH5c-His, 1 g of GST-E6AP, and 1 g of ER. Where indicated, ER was preincubated with 100 ng of calmodulin in the presence of 1 mM CaCl 2 at 4°C for 10 min. The reaction mixture was incubated at 37°C for 30 min, and the assays were stopped by adding SDS-PAGE solubilization buffer. Samples were processed by Western blotting, and ubiquitinated proteins were visualized with streptavidin fused to horseradish peroxidase, followed by enhanced chemiluminescence.
GST Pull-down Assays-GST fusion constructs of wild type E6AP (20) were expressed in E. coli and isolated with glutathione-Sepharose. For pull-down from cell lysates, COS-7 cells were transiently transfected with pcDNA3-ER or pcDNA3-ER⌬CaM and 48 h later were lysed in buffer B. Equal amounts of lysate were incubated for 2 h at 4°C with 500 ng of GST-E6AP on glutathione-Sepharose beads. GST alone (10 g) was used as control. After sedimentation by centrifugation, samples were washed with buffer B and resuspended in SDS-PAGE sample buffer (20 mM Tris-HCl, pH 7, 5.2% (w/v) SDS, 2% (v/v) ␤-mercaptoethanol, 0.01% (w/v) bromphenol blue, 0. 25 M sucrose, and 2 mM EDTA). Samples were heated at 100°C for 5 min and processed by immunoblotting. For competition analysis with pure proteins, 1. 5 g of ER was preincubated for 1 h at 4°C with 0, 0.25, or 1.25 g of calmodulin or 5 g of bovine serum albumin in buffer A in the presence of 1 mM CaCl 2 . GST-E6AP (5 g) on glutathione-Sepharose beads was added, and samples were incubated for an additional 1 h (10 g of GST alone was used as control). Samples were processed by Western blotting as described above.
Miscellaneous-Densitometry of enhanced chemiluminescence signals was analyzed with UN-SCAN-IT software (Silk Scientific Corp.). Statistical analysis was performed by Students's t test, using InStat software (GraphPad Software, Inc.). Protein concentrations were determined with the DC Protein Assay (Bio-Rad).

Calmodulin Antagonists Reduced the Amount of ER Protein in Breast
Epithelial Cells-We have documented previously (11) that incubation of MCF-7 cells with calmodulin antagonists significantly reduced the amount of ER␣ protein. This reduction was due to enhanced degrada-tion of ER␣; gene expression was unaltered (11). The effect is specific for ER␣. The stability of ER␤, which does not bind calmodulin (24), is unaffected by calmodulin antagonists (16). Therefore, all further analyses were confined to ER␣. Here we examined the effect of two structurally distinct, cell-permeable calmodulin antagonists, CGS9343B and TFP, in another ER-positive breast epithelial cell line. Exposure to 40 M CGS9343B reduced the total amount of ER in T47D cells (Fig. 1A). The reduction was not altered by E 2 . Note that E 2 alone decreased ER concentration ( Fig. 1A), consistent with prior observations (11). Incubation with TFP similarly reduced the amount of ER (Fig. 1), indicating that calmodulin stabilizes ER in diverse breast epithelial cell lines.
Inhibition of the interaction between ER and calmodulin by calmodulin antagonists results in degradation of ER (11). Therefore, we hypothesized that mutant ER constructs that cannot bind to calmodulin would be unaffected by CGS9343B. To test this hypothesis, we examined degradation of ER⌬298 -317 and ER⌬CaM. Neither ER⌬298 -317, which lacks the calmodulin-binding region of ER, nor the point mutant construct ER⌬CaM binds calmodulin (22). COS-7 cells, which lack endogenous ER, were transfected with wild type ER, ER⌬298 -317, or ER⌬CaM and incubated with CGS9343B. Analogous to the observation with endogenous ER, CGS9343B substantially reduced the amount of transfected full-length ER (Fig. 1B). In contrast, the levels of ER⌬298 -317 and ER⌬CaM were not significantly altered by antagonism of calmodulin with CGS9343B. These data confirm that the effect of CGS9343B on ER stability is specific and is mediated by disrupting the interaction of calmodulin with ER.
Effect of Proteasome and Protease Inhibitors on ER Degradation Induced by Antagonism of Calmodulin-In order to elucidate the molecular mechanism by which calmodulin stabilizes ER, we used a panel of inhibitors. The ubiquitin-proteasome degradation pathway is the major system for degradation of short lived proteins (25,26). Therefore, the possible involvement of this pathway was evaluated with the proteasome inhibitors MG132 and lactacystin. MCF-7 cells were preincubated with lactacystin or MG132 for 2 h prior to treatment with CGS9343B. Analysis by Western blotting revealed that lactacystin prevented CGS9342B from reducing the amount of ER ( Fig. 2A). Lactacystin alone, without CGS9343B, had no significant effect on the ER content of the cells. Essentially identical results were obtained with MG132 ( Fig. 2A).
Similar analysis was conducted with protease inhibitors. In contrast to the protective effect of the proteasome inhibitors, neither the calpain inhibitor II ALLM nor calpeptin had any effect on the reduction of ER produced by incubating cells with CGS9343B (Fig. 2B). Collectively, these data imply that calmodulin protects the ER from degradation in the ubiquitin-proteasome pathway.
The Calmodulin Antagonist CGS9343B Enhanced Ubiquitination of ER-The data in Fig. 2 suggested that calmodulin may inhibit ER ubiquitination. To test this hypothesis, COS-7 cells were cotransfected with ER and His-tagged ubiquitin, and the extent of ER ubiquitination was examined by probing anti-His immunoprecipitates for ER. Ubiquitin is conjugated to multiple lysine residues of target proteins and forms polyubiquitin chains (27). Therefore, ubiquitin-tagged proteins can sometimes be seen as a ladder of higher molecular weight species on SDS-PAGE (6). Incubation of cells with MG132 enhanced ER ubiquitination ( Fig. 3; ubiquitinated forms of ER are visible as a smear of higher molecular weight bands). Similarly, incubation with CGS9343B alone slightly increased ER ubiquitination but to a lesser extent than MG132 (Fig. 3A). When cells were treated with both MG132 and CGS9343B, ER ubiquitination was further enhanced. Essentially identical results were obtained when ubiquitinated proteins were isolated by immunoprecipitating with anti-ER antibodies and identified by probing blots for Hisubiquitin (data not shown). These observations suggest that CGS9343B promotes ER degradation by enhancing ER ubiquitination.
Ubiquitination of Wild Type ER and ER⌬CaM-The role of calmodulin in ER ubiquitination was explored further with ER⌬CaM, a point mutant ER construct that cannot bind calmodulin (22). COS-7 cells were transiently transfected with wild type ER or ER⌬CaM, and the extent of ubiquitination of the two ER proteins was compared. When cells were incubated with MG132, ubiquitination of ER⌬CaM was substantially greater than that of wild type ER (Fig. 3B). The expression levels of wild type ER and ER⌬CaM were comparable as shown in the Western blot of cellular lysates (Fig. 3B, lower panel).
ER Is a Substrate of E6AP-Ubiquitination is a three-step process in which ubiquitin is activated by a ubiquitin-activating enzyme (E1), transferred to a ubiquitin-conjugating enzyme (E2) and then transferred to the target by a ubiquitin-protein ligase (E3) (26,28). E6AP was initially identified as a ubiquitin ligase that cooperates with the human papillomavirus (HPV) E6 oncoprotein to degrade the tumor suppressor p53 (18). A few years ago, the human progesterone receptor was shown to directly interact with E6AP (29). Therefore, we wondered whether E6AP participates in regulation of ER. Initial analysis was to determine whether the E3 ligase activity of E6AP is capable of degrading endogenous ER. Transient transfection of MCF-7 cells with wild type E6AP decreased endogenous ER levels by 21 Ϯ 3.3% (mean Ϯ S.E., n ϭ 3; p Ͻ 0.05) (Fig. 4). Further analysis was performed with a catalytically inactive mutant form of E6AP (termed E6AP C833A), which has a cysteine  to alanine substitution at amino acid 833 in the active site, rendering it incapable of forming a thiol ester with ubiquitin (17). In contrast to the wild type enzyme, E6AP C833A did not significantly reduce the amount of ER in the cells (Fig. 4). Note that the expression levels of E6AP and C833A were comparable. These data reveal that the E3 ligase activity of E6AP contributes to degradation of ER.
E6AP Associates Specifically with ER-The possible interaction of ER with E6AP was examined by two approaches, namely immunoprecipitation and pull-down with GST fusion proteins. In the first approach, endogenous E6AP was immunoprecipitated from COS-7 cells expressing ER. ER coimmunoprecipitated with E6AP (Fig. 5A). Specificity was confirmed by the absence of ER from samples precipitated in parallel with NIRS. Analogous experiments were conducted with MCF-7 cells from which endogenous ER was immunoprecipitated. This analysis revealed that both wild type and mutant E6AP coimmunoprecipitated with ER (Fig. 5B). Although the expression levels of wild type and mutant E6AP were equivalent (see lysate in Fig. 5B), the amount of E6AP C833A that coimmunoprecipitated with ER was substantially more than the amount of wild type E6AP. Similar observations were made in COS-7 cells, which were cotransfected with ER and HA-tagged E6AP (Fig. 5C). Both wild type E6AP and E6AP C833A coimmunoprecipitated with transfected ER. Analogous to the findings with MCF-7 cells, the amount of E6AP C833A that coimmunoprecipitated with ER from COS-7 cells was greater than that of wild type E6AP (Fig. 5C).
The specificity of the interaction between ER and E6AP was evaluated by examining the possible association of ER with Nedd4, another HECT domain E3 ligase (30). ER was immunoprecipitated from COS-7 cells, which had been cotransfected with ER and Nedd4. In contrast to the findings with E6AP, no Nedd4 was detected in immunoprecipitates, despite the presence of abundant ER (Fig. 5D). Note that Nedd4 expressed well in the cells but did not induce ER degradation. A catalytically inactive Nedd4 construct also failed to bind ER (Fig. 5D). When Nedd4 was immunoprecipitated from cells under similar conditions, no ER was detected in the precipitates (data not shown). Collectively, these findings reveal that ER binding to E6AP is specific because ER did not bind another related HECT domain E3 ubiquitin ligase.
Effect of Calmodulin on the Binding of ER to E6AP-In the second approach, GST-E6AP was incubated with lysates derived from COS cells transfected with ER. GST-E6AP bound ER in COS cell lysates (Fig.  6A). Binding was specific as no ER was detected in lysates incubated  with GST alone. The enhanced ubiquitination of ER in the absence of calmodulin binding, coupled with the association between E6AP and ER, led to the hypothesis that calmodulin may modulate the ER-E6AP interaction. Three interrelated experimental methods were used to test this hypothesis. In the first strategy, the binding of ER⌬CaM to GST-E6AP was compared with that of wild type ER. Binding of ER⌬CaM to E6AP was substantially greater than that of wild type ER (Fig. 6A).
The second strategy was to perform direct in vitro competition assays to assess ER binding to E6AP in the presence of calmodulin. Pure ER bound to GST-E6AP (Fig. 6B). When ER was preincubated with calmodulin in the presence of Ca 2ϩ , the binding of ER to GST-E6AP was substantially reduced (Fig. 6B). In contrast, when Ca 2ϩ was chelated with EGTA, calmodulin did not attenuate ER binding to GST-E6AP (data not shown). These findings are consistent with our prior observations that calmodulin binds ER in Ca 2ϩ -dependent manner (11).
The effect of the cell-permeable calmodulin antagonist CGS9343B was the third strategy. Incubating MCF-7 cells transfected with E6AP with CGS9343B increased by 2.5-fold the amount of wild type E6AP that coimmunoprecipitated with endogenous ER (Fig. 7). CGS9343B also increased the interaction of E6AP C833A with ER (Fig. 7), although the magnitude of enhancement was much less than that seen with wild type E6AP. Because CGS9343B decreases the binding of calmodulin to ER (11), these results suggest a model in which calmodulin and E6AP compete for ER binding. Disrupting the association of calmodulin with ER facilitates the ER-E6AP interaction.
Calmodulin Reduced ER Ubiquitination in Vitro-An in vitro ubiquitination assay was used to determine whether E6AP directly ubiquitinates ER. We were able to reconstitute ER ubiquitination using pure proteins. In the presence of E1, E2 UbcH5c, and ubiquitin, E6AP catalyzed the ubiquitination of ER (Fig. 8). No ubiquitinated ER was detected when E6AP or ubiquitin was omitted from the assay. Preincubation of ER with Ca 2ϩ /calmodulin markedly reduced ubiquitination of ER by E6AP (Fig. 8). Calmodulin had no effect on E6AP-mediated ubiquitination of ER when Ca 2ϩ was chelated with EGTA (data not shown). These data reveal that E6AP is directly responsible for ER ubiquitination and that calmodulin attenuates the reaction.

DISCUSSION
Several lines of evidence couple Ca 2ϩ and calmodulin to breast carcinoma and ER function. These include the following. (i) Calmodulin concentrations are increased in malignant human mammary tissue (31). . Cells were lysed 24 h later, and equal amounts of protein lysate were immunoprecipitated (IP) with anti-ER antibodies. Western blots were probed with anti-HA (to detect the HA-E6AP) and anti-ER antibodies. C, COS-7 cells were cotransfected with pcDNA3-ER and either pCMV4 vector, E6AP, or E6AP C833A. After 48 h, cells were lysed and processed as described for B. D, COS-7 cells were cotransfected with pcDNA3-ER and catalytically inactive (C744A) or wild type (WT) pcDNA-Myc-Nedd4. Equal amounts of protein lysate were resolved by Western blotting (lysate). In addition, equal amounts of protein lysate were immunoprecipitated (IP) with anti-ER antibody or NIRS. Western blots were probed with anti-Myc (to detect the Myc-Nedd4) and anti-ER antibodies. All data are representative of at least three independent experimental determinations.

FIGURE 6. Effect of calmodulin on the interaction between ER and E6AP.
A, COS-7 cells were transiently transfected with pcDNA3-ER (WT) or pcDNA3-ER⌬CaM (⌬CaM). Forty eight hours later, cells were lysed, and an equal amount of protein lysate was incubated with 500 ng of GST-E6AP or 10 g of GST alone on glutathione-Sepharose beads for 2 h at 4°C. After sedimentation by centrifugation, samples were washed and resolved by SDS-PAGE. The gel was cut in half; the top portion (containing GST-E6AP) was stained with Coomassie Blue, and the bottom half was processed by blotting for ER. Cell lysates were also applied directly to the gel and processed by Western blotting as lysate input control (Lysate). Data are representative of three independent experiments. B, purified ER was preincubated with 0, 0.25, or 1.25 g of calmodulin or 5 g of bovine serum albumin (BSA) in the presence of Ca 2ϩ for 1 h at 4°C. GST-E6AP or GST on glutathione-Sepharose beads was added for additional 1 h at 4°C, and samples were processed by Western blotting. A representative blot is shown in the upper panel. The amount of ER bound to GST-E6AP was quantified by densitometry. Data, expressed relative to the amount of ER bound to GST-E6AP without calmodulin, represent the means Ϯ S.D. (n ϭ 2) (lower panel).
(ii) Ca 2ϩ /calmodulin stimulates E 2 binding to the ER (32). (iii) Calmodulin binds to the ER in a Ca 2ϩ -dependent manner (11,15,33), and this increases the K d value of estradiol binding (14). (iv) Ca 2ϩ /calmodulin stimulates tyrosine phosphorylation and activation of the ER (32). (v) Association of calmodulin with the activated estrogen-ER complex increases the interaction of the complex with ERE (14). (vi) Calmodulin is required for the formation of the ER⅐ERE complex and for activation of an estrogen-responsive promoter (15). (vii) Calmodulin antagonists prevent E 2 from stimulating ER transcription (16,24). (viii) Binding to calmodulin is required for the normal transcriptional function of the ER (22). Together these findings suggest that Ca 2ϩ /calmodulin may participate in ER-induced transcriptional activation and the mitogenic effects of estrogen. An intriguing publication revealed that E 2 administration to ovariectomized mice up-regulated, independently of ER, the expression of calmodulin mRNA in the uterus by 4.5-fold (34). These data reveal a reciprocal regulation between calmodulin and estrogen.
An additional biological role of calmodulin in estrogen function is to modulate ER degradation (11). The binding of calmodulin to ER enhanced the stability of ER both in vitro and in intact cells (11). The present study was undertaken to elucidate the molecular mechanism by which calmodulin mediates this effect. Two complementary strategies were adopted. These were the use of a selective, cell-permeable calmodulin antagonist CGS9343B (35) and a point mutant ER (termed ER⌬CaM) that is unable to bind calmodulin (22). We developed ER⌬CaM by selectively mutating two amino acids (Ile-298 and Lys-299) in the calmodulin-binding domain of ER, thereby abrogating calmodulin binding. CGS9343B enhanced proteolysis of wild type ER but had no effect on the level of ER⌬CaM in cells. These data verify that the calmodulin antagonist CGS9343B promotes ER degradation by disrupting the interaction between calmodulin and ER.
The ubiquitin-proteasome pathway is responsible for E 2 -dependent down-regulation of ER levels (6 -9). We observed that the proteasome inhibitors lactacystin and MG132 blocked the degradation of ER produced by CGS9343B. Despite a previous report that ER is a substrate of calpain (36), inhibition of calpain activity did not attenuate the enhanced ER proteolysis that occurred in cells incubated with CGS9343B. These results imply that calmodulin stabilizes ER by decreasing its degradation in the ubiquitin-proteasome pathway. This hypothesis is supported by the finding that disruption of the interaction between ER and calmodulin augments ER ubiquitination. Two distinct, complementary strategies yielded essentially identical results. Incubation of MCF-7 cells with CGS9343B, which blocks the association of calmodulin with ER, enhanced ubiquitination of ER. Similarly, ER⌬CaM was ubiquitinated to a greater extent than wild type ER.
Ubiquitination of protein substrates is a multistep process that involves the concerted action of at least three classes of enzymes as follows: E1, E2, and E3 (26,37). E1 activates ubiquitin, a highly conserved 76-residue polypeptide, in an ATP-dependent reaction to generate a high energy thiol ester intermediate. One of at least 25 E2 enzymes transfers the activated ubiquitin moiety from E1 to the substrate that is specifically bound to one of a group of E3 ligases. E3s catalyze the attachment of ubiquitin to the substrate, the final step in the pathway. The sequential addition of activated ubiquitin moieties onto a lysine residue of the previously conjugated ubiquitin molecule produces a polyubiquitin chain (26). This chain is recognized by the 26 S proteasome, resulting in degradation of the ubiquitinated protein. The E3 ubiquitin ligases are key components that provide selectivity to the pathway by interacting directly with different substrates. E6AP is the prototype of a family of E3 ligases called HECT proteins, all of which contain a domain homologous to the E6AP C terminus (38). This domain catalyzes ubiquitin transfer, although substrate specificity is provided by other domains. The first documentation of a role for E6AP in nuclear hormone function was by O'Malley and co-workers (29) who described a hormone-dependent regulation of transcriptional activity of the progesterone receptor by E6AP. This effect, which was independent of the ligase function of E6AP, was also noted for glucocorticoid receptor and ER (29). However, neither the interaction of E6AP with ER nor its role in ER ubiquitination was examined in that study.
Published work from several laboratories over the last 5 years has

FIGURE 8. Calmodulin inhibits E6AP-mediated ubiquitination of ER in vitro.
Purified ER was preincubated in vitro without (Ϫ) or with (ϩ) calmodulin (CaM). After 10 min, E1 and E2 were added in the absence or presence of 10 M biotin-tagged ubiquitin and 1 g of E6AP, and the reaction was allowed to proceed for 30 min. The assay was stopped by adding SDS-PAGE solubilization buffer, and samples were processed by Western blotting. Ubiquitinated protein products were visualized with horseradish peroxidase-conjugated streptavidin. Ubiquitinated ER (ER-ub) is labeled. The identity of the ubiquitinated protein ϳ50 kDa is unknown. Molecular weight markers are depicted on the left. Data are representative of two independent experiments.
shown that the proteasome is responsible for E 2 -mediated ER degradation (6 -9). However, mammalian cells contain hundreds of E3 ubiquitin ligases, each of which binds to a specific protein substrate that has been targeted for degradation (39), and the E3 ligase(s) responsible for destruction of ER had not been identified previously. In this study we document that E6AP significantly reduced ER levels in MCF-7 cells. Enzymatic activity is necessary for ER degradation because a catalytically inactive mutant form of E6AP (E6AP C883A) did not alter the amount of ER, despite a documented ability to bind ER. Endogenous E6AP specifically coimmunoprecipitated with ER from cell lysates and ER bound to GST-E6AP. The binding appears to be specific for E6AP as Nedd4, another HECT domain E3 ligase that is homologous to E6AP in its C terminus (30), neither coimmunoprecipitated with ER nor induced ER degradation. Collectively, these data reveal that E6AP is a component of a degradation pathway responsible for ER proteolysis. E 2 -mediated down-regulation of ER has been linked to its transcriptional activity (6,10). For example, the p160 coactivator AIB1 both recruits transcription factors involved in ER gene activation and mediates ER degradation by E 2 (10). Calmodulin also has a dual role in regulating ER by affecting both transcription and degradation. However, the effects of calmodulin are different from those of AIB1. Calmodulin binding is necessary for E 2 -stimulated transcription (16,22) and blocks ER degradation. Our findings provide a molecular mechanism to explain the prior observations that calmodulin is overexpressed in breast cancer (31) and that calmodulin antagonists reduce the growth of breast cancer cells (41,42) and synergistically augment anti-estrogen therapy (43). Estrogen has been linked to the progression of the majority of human breast cancers. An increase in cellular calmodulin would increase both the amount of ER and its E 2 -stimulated transcriptional activity, thereby enhancing the growth-promoting effects of estrogen. Disrupting the interaction between ER and calmodulin will attenuate the effects of estrogen at two points in its signaling pathway, by decreasing ER concentrations and preventing E 2 from stimulating ER transcriptional activation.
E6AP appears to participate in the mechanism by which calmodulin regulates the fate of ER. One model that could explain our data is that calmodulin competes with E6AP for binding to ER. Evidence consistent with this scheme includes the observation that disrupting the interaction between calmodulin and ER either by CGS9343B or by mutating the calmodulin-binding domain of ER produced the same outcome, namely an enhanced association of ER with E6AP. The inhibition of in vitro ER binding to E6AP mediated by Ca 2ϩ /calmodulin further supports this model. Finally, calmodulin attenuated E6AP-mediated ubiquitination of ER in vitro. Collectively, our data reveal that calmodulin binding reduces the interaction of ER with E6AP, thereby protecting ER from degradation in the proteasome pathway. A common theme underlying calmodulin function is that binding of calmodulin alters the conformation of its target proteins (44). It is tempting to speculate that when calmodulin associates with ER, the tertiary conformation of ER is modified. This would change the interaction of ER with other targets, such as E6AP, thereby altering ER degradation. The ability of calmodulin to induce a conformational change in ER can be inferred from the protection afforded by calmodulin to in vitro proteolysis of ER (11). It is important to define exactly where on ER the E6AP-binding domain is located to directly test the competition model in vivo. Detailed insight into the reciprocal regulation of ER stability by calmodulin and E6AP might prove useful in the development of selectively targeted strategies to block the proliferation of breast carcinoma cells.
A direct connection has been established between HPV and cervical neoplasia, and HPV DNA has been detected in nearly all cervical cancer (45). One of the mechanisms underlying the malignant transformation is the inactivation of the p53 tumor suppressor gene by the HPV early gene E6 (17). E6AP has a prominent role in this process by cooperating with the E6 oncoprotein (17). Additional insight into the molecular mechanisms of HPV-induced cervical carcinogenesis has been obtained using HPV16 transgenic mice. Transgenic mice expressing the oncogene of HPV16 under the control of the human keratin-14 promoter developed cervical carcinoma only when treated with E 2 (46). More recent analysis revealed that E 2 contributes to the onset, persistence, and progression of cervical cancer in this HPV transgenic mouse model (47). Thus, E 2 is an essential cofactor for the induction of cervical carcinoma by HPV16 oncogenes. These findings raise the question as to how E 2 and HPV16 interact to induce carcinogenesis. It is tempting to speculate that the interaction of E6AP with ER identified in this study may contribute to this process. Further work is necessary to test this provocative hypothesis.