The active metabolite of tamoxifen, 4-hydroxytamoxifen (4-OHT), is used in the laboratory for mechanistic studies of antiestrogen action. This compound binds to the estrogen receptor α (ER) and silences activating function 2 (AF2) in the ligand binding domain, but activating function 1 (AF1) at the other end of the ER remains constitutive and is considered to be ligand independent. Amino acid D351 in the ligand binding domain appears to be critical for interactions with the antiestrogenic side chain of antiestrogens. We have devised an assay to evaluate the biological activity of 351 mutant ERs and antiestrogens at the transforming growth factor α(TGFα) gene in situ (J. I. MacGregor Schafer et al., Cancer Res., 59:4308–4313, 1999). The substitution of glycine for aspartate at position 351 results in the conversion of the 4-OHT:ER complex from estrogen-like to completely antiestrogenic. In cells stably expressing D351G ER, the ER retains responsiveness to estradiol (E2)and also retains antiestrogenic responsiveness to both raloxifene and ICI 182,780. The relative binding affinity of E2 for D351G ER (0.77 ± 0.17 × 10−9m) is comparable with wild-type ER (0.42 ± 0.08 × 10−9m). In addition,the D351G ER retains the ability to bind SRC-1 in the presence of E2, thus D351G ER AF2 activity has not been compromised. We also used a cell line stably expressing an ER with a triple mutation in helix 12 (D538A, E542A, and D545A) that ablated AF2 activity, which resulted in decreased effects of E2, suggesting that both AF1 and AF2 activity are required for maximal estrogen activity in MDA-MB-231 cells. Interestingly, the triple mutation also completely reduced the estrogen-like actions of 4-OHT. We propose that a specific mutation at amino acid 351 can allosterically silence AF1 in the 4-OHT:ER complex by either preventing the binding of coactivators or encouraging the binding of a corepressor molecule. We suggest that the 4-OHT-specific site responsible for estrogen-like actions can be referred to as AF2b. This binding site would consist of at least four carboxylic acids at amino acids 351 and 538, 542 and 545 in helix 12 to permit coactivator docking for gene activation. The AF2b site is distinct from AF2 for E2 action. Further studies will provide insight into the estrogen-like actions of tamoxifen in select tissues and breast tumors and identify a significant mechanism of drug resistance to tamoxifen.

Tamoxifen, a nonsteroidal antiestrogen, is the endocrine treatment of choice for patients with ER3-positive breast cancer(1) and the first drug found to reduce the incidence of breast cancer in high-risk women (2). Tamoxifen is not a pure antiestrogen and has significant estrogen-like properties that are exhibited in some estrogen target tissues, such as bone, uterus, and liver (3), and can stimulate tumor growth in laboratory models (4, 5). The precise molecular mechanism of action of tamoxifen is unknown (3, 6), but the elucidation of the structure-function relationships of the tamoxifen:ER complex would aid in the development of new SERMs and potentially provide a target to subvert the development of drug resistance.

The active metabolite of tamoxifen, 4-OHT (7, 8, 9), is used in the laboratory as an antiestrogen to study the molecular actions of tamoxifen. The metabolite has a high binding affinity for the ER and competes with E2 for the LBD (7, 9, 10). Two transcriptional domains, AF1 and AF2 (11),have been identified in the ER to bind coactivators and initiate gene transcription. The binding of E2 to the ER causes activation of AF1 and AF2, but 4-OHT silences AF2 activity in the LBD(12, 13, 14). It is hypothesized that AF1 activity is constitutive and ligand independent.

The recent resolution of the crystal structure of the ER LBD with Ral(15) and 4-OHT (16) has provided an invaluable insight into the activation of AF2 in ER by E2 and silencing by antiestrogens. An estrogen binds within the hydrophobic pocket of the LBD and is sealed inside by helix 12. This conformational change in the LBD activates AF2. In contrast, 4-OHT and Ral prevent helix 12 from sealing the binding pocket. The inappropriate repositioning of helix 12 prevents the binding of the coactivator GRIP1 and silences AF2 (16).

Antiestrogens require an alkylaminoethoxy side chain to block estrogen action (17). Early hypotheses proposed that the antiestrogenic side chain must specifically interact with the LBD to prevent sealing of the hydrophobic pocket (18, 19). The crystal structure of the Ral:ER complex demonstrates that the antiestrogenic side chain interacts with aspartate at position 351 in the ER (15), thereby suggesting the importance of this site for antiestrogen action. The crystal structure of the 4-OHT:ER complex shows the side chain having a loose interaction with amino acid 351 (Ref. 16; Fig. 1). By coincidence, a D351Y ER, isolated from a tamoxifen-stimulated tumor line, had already been identified (20) and found to change the pharmacology of Ral from a complete antiestrogen to a partial estrogen (21, 22).

The 4-OHT:ER complex is more estrogen-like in target tissues than the Ral:ER complex. We have developed an assay of receptor function by stably transfecting cDNAs for wild-type (23) and D351Y(24) ER into MDA-MB-231 cells to examine the molecular pharmacology of antiestrogens under controlled conditions. The biological activity of a drug:ER complex can be assessed using the induction of mRNA for the TGFα gene in situ(25). Surprisingly, the 4-OHT:ER complex is as effective as the E2:ER complex at inducing mRNAs for TGFα(26). In contrast, Ral is a complete antiestrogen with the wild-type receptor (22). We propose that the difference in actions of 4-OHT and Ral results from distinct differences in the crystallographic structure of the 4-OHT:ER complex and the Ral:ER complex. The side chain of Ral interacts with D351, but the side chain of 4-OHT is in a different position so that the interaction with D351 is tenuous (Fig. 1). Because we have demonstrated previously that a natural substitution at D351 can enhance estrogenicity of an antiestrogen receptor complex (21, 22), we believed that this amino acid position in the ER was significant for further study.

We have substituted aspartate with glycine at amino acid 351 in the ER to determine the impact on the agonist activity of partial antiestrogens by removing the charge and the side chain at position 351. Our hypothesis was that the receptor complex could lose the estrogenic property that partial antiestrogens contain. We were surprised to discover that a glycine substitution silences both AF1 and AF2 when the efficacy of the 4-OHT:ER D351G is evaluated at the TGFα gene in situ. Antiestrogenic activity of D351G is retained.

Cell Lines and Tissue Culture.

ER-negative MDA-MB-231 cells were obtained originally from American Type Culture Collection (Rockville, MD). S30 (MDA-MB-231 cells stably transfected with wild-type ER; Ref. 27) and JM-6(MDA-MB-231 cells stably transfected with D351G ER) were grown in phenol red-free minimal essential medium supplemented with 5% 3×dextran-coated, charcoal-treated calf serum, 2 mmglutamine, 6 ng/ml bovine insulin, 100 units/ml penicillin/100 μg/ml streptomycin, nonessential amino acids, and 500 μg/ml G418. Cells were passaged twice per week at 1:10 with 0.5% trypsin. Cells were grown at 37°C in a 5% CO2 incubator.

Site-directed Mutagenesis.

The glycine mutation (pSG5D351GER) at amino acid 351 was introduced using QuickChange Site-Directed Mutagenesis Kit (Stratagene, La Jolla,CA) using pSG5HEGO (a wild-type ER expression vector generously provided by Professor Pierre Chambon, University of Strasbourg,France) as the template. The primers constructed are as follows: 5′ primer (5′-GGGCTTACTGACCAACCTGGC AGGCAGGGAGCTGGTTCAC-3′; the underlined nucleotide was changed to make the D351G mutation) and a 3′ primer(5′-CCCGAATGACTGGTTG GACCGTCCGTCCCTCGACCAAGTG-3′). pSG5ERα/G3m (D538A/E542A/D545A) expression vector was generated as described above (28). The pair of primers used was:5′-GCCCCTCTATGCCCTGCTGCTGGCGATGCTGGCCGCCCACCGCC-3′/5′-GGCGGTGGGCGGCCAGCATCGCCAGCAGCAGGGCATAGAGG-GGC-3′. The point mutations are underlined. pGEX-ΔSRC-1 is an expression vector for a GST fusion protein containing amino acids 759-1061 of SRC-1 (29). The SRC-1 fragment with BamHI and EcoRI sites was generated by PCR and inserted into pGEX-2TK(Pharmacia). All sequences were confirmed by sequencing analysis (ABI automated sequence and USB sequencing kit).

Stable Transfections.

MDA-MB-231 cells were stably transfected by electroporation. Cells(5 × 106) were mixed with 10 μg of pSG5D351GER or pSG5ERα/G3m (an ER with an inactivating triple mutation in AF2) mutant ER expression vector and 1 μg of constitutive neomycin resistance expression vector (pBK-CMV) and were electroporated (950 μF; 330V) using a Bio-Rad Gene Pulser II. Briefly, after electroporation, cells were cultured in estrogen-free media for 48 h to recover. Then, neomycin-resistant cells were selected by using media containing 500 μg/ml G418. Colonies were isolated and propagated in G418-containing media. Neomycin-resistant clones were screened, identified, and characterized for ER expression levels using both Northern and Western blot analysis.

Western Blot Analysis.

Cells were seeded into T-75 cm2 tissue culture flasks and treated for 24 h with hormone or drug. At the conclusion of the experiment, cells were trypsinized and pelleted. The pellet was resuspended in protein extraction buffer [0.5% NP40, 2%glycerol, 1 mm DTT, 1 mm EDTA, 150 mm NaCl, 50 mm Tris (pH 7.4), 1 mmEGTA, 3 mm phenylmethylsulfonyl fluoride, 25 μg/ml leupeptin, 9 μg/ml aprotinin, 25 μg/ml trypsin inhibitor, and 25μg/ml t-chymotrypsin]. Samples were incubated on ice with intermittent vortexing for 30 min and then pelleted. Supernatant was then collected and stored at −80°C. Protein concentration was measured using the Bio-Rad Protein Assay kit, and equal amounts of protein were run in a standard Western blot protocol. The ER primary antibody used was AER311 (Neomarkers, Fremont, CA), and β-actin antibody AC-15 (Sigma Chemical Co., St. Louis, MO) was used to standardize loading. The appropriate secondary antibody conjugated with horseradish peroxidase kit was used to visualize bands using an ECL visualization kit (Amersham Corp., Arlington Heights, IL). The membrane was wrapped in plastic wrap and exposed to Kodak X-OMAT film for 10 s to 1 h.

Northern Blot Analysis.

Analysis of TGFα mRNA expression was assessed by Northern blots. Briefly, total RNA was isolated from S30, JM-6, and G3m cells using the TRIZOL reagent (Life Technologies, Inc.) after treatment with a range of concentrations of E2, 4-OHT, Ral, and ICI 182,780 for 24 h. A human TGFα cDNA probe derived by EcoRI digestion of a TGFα-containing plasmid was a generous gift from Dr. R. Derynck, (Genentech, San Francisco, CA). Bands were quantitated densitometrically using ImageQuaNT analysis(Sunnyvale, CA). Comparisons in TGFα induction relative to β-actin induction were analyzed by ANOVA. The two tailed t test was used to analyze differences between treatments using StatMost 2.5(Datamost Corp., Salt Lake, UT).

In Vitro Protein-Protein Interaction.

A GST-pull down assay was performed as described previously (30, 31) using 35S-labeled wild-type ER and 35S-labeled D351G ER, which were made from pSG5HEGO and pSG5D351GER, respectively, using an in vitrotranscription-coupled translation system (Promega Corp., Madison, WI).

Ligand Binding Assay.

Ligand binding assays in stable transfectants were performed as described previously (32). For saturation binding assays,the stable transfectants were incubated with increasing concentrations of [3H]E2 (Amersham Corp., Arlington Heights, IL) for 2 h at room temperature. To determine nonspecific binding, each concentration of[3H]E2 was competed with 400-fold excess of unlabeled DES. The specific binding was obtained by subtracting the nonspecific binding from the total binding. For competition binding assays, the stable transfectants were incubated with 10−9m[3H]E2 with increasing concentrations of different ligands including 4-OHT, Ral, or ICI 182,780 for 2 h at room temperature. The data were analyzed by GraphPad Prism (GraphPad Prism Software, Inc., San Diego, CA).

We stably transfected D351G ER into MDA-MB-231 cells. The initial screen of 10 prospective clones for ER by Northern blot yielded three positive clones (data not shown). When protein expression was determined using Western blot analysis, one clone (JM-6) was isolated,which expressed acceptable levels of ER when compared with the S30 cells (Fig. 2). In addition, to determine the effects of E2 and antiestrogens on D351G ER expression, we treated S30 (Fig. 3,A) and JM-6 cells (Fig. 3,B) for 24 h and found that 4-OHT treatment has no effect on D351G ER expression. Both E2 and ICI 182,780 reduced ER expression(Fig. 3).

Although S30 and JM-6 cells express comparable levels of ER, the single mutation in D351G ER could affect ligand binding for various compounds and affect the interpretation of results. We used a hormone binding assay to determine the relative binding affinity of wild-type and D351G ER for E2 and the EC50 for 4-OHT, Ral, and ICI 182,780 (Table 1). We found that the ligand affinity of wild-type ER compared with D351G ER was not statistically different for E2(Kd 0.42 ± 0.08 versus0.77 ± 0.17 nm). Ral was the only antiestrogen that had a significantly lower binding affinity for D351G ER compared with the wild type (EC501.04 ± 0.22 versus 13.24 ± 1.45 nm; P < 0.05). From these studies, it is also important to point out that the concentrations of 4-OHT used for the pharmacology studies saturated the binding sites(Fig. 4).

To study the transcriptional activity of D351G ER at a target gene in situ, we performed concentration response experiments using E2. These results show that E2 induces TGFα mRNA expression beginning at 10−10m in JM-6 cells(Fig. 5). These data are comparable with results published previously(21) with D351 ER in S30 cells.

To investigate the agonist action of antiestrogens on TGFα mRNA induction, we first compared single concentrations(10−7m) of 4-OHT, Ral, and ICI 182,780 with E2 (10−9m). In S30 cells, 4-OHT induced TGFα transcription to the same extent as E2 (Fig. 6,A, B). In contrast, although E2 was capable of inducing TGFα in JM-6 cells, 4-OHT was not estrogenic, as evidenced by a diminished ability to induce TGFα gene transcription(Fig. 6,C, D). In both S30 and JM-6 cells, neither Ral nor ICI 182,780 induced TGFα mRNA transcription. These data alone,however, do not prove that 4-OHT, Ral, and ICI 182,780 are, in fact,binding to the ER. This could be an explanation for the lack of agonist activity. To address this question, we used combinations of E2 (10−9m) and antiestrogens (10−7m) and determined whether the antiestrogen blocked TGFα mRNA induction (Fig. 7). We found that 4-OHT and ICI 182,780 continued to be antiestrogens because they inhibit E2 induction of TGFα. The sum of the data from Figs. 3 and 4 and Table 1 demonstrate that differences in ligand binding affinity or ER protein stability cannot explain the difference in the intrinsic activity of 4-OHT between wild-type and D351G ER. The substitution of glycine at position 351 silences transcriptional activity of the 4-OHT:ER complex.

To further characterize the actions of 4-OHT, Ral, and ICI 182,780 to inhibit E2 induction completely in JM-6 cells, we performed concentration-response experiments. Higher concentrations of 4-OHT, Ral, and ICI 182,780 were able to inhibit E2-induced transcription to control levels in JM-6 cells (Fig. 8, A–C). Both 4-OHT and ICI 182,780 were effective antiestrogens at 10−8m;however, a concentration of 10−7m Ral was required to block E2 action (Fig. 8,D). This was consistent with a reduced EC50 for Ral (Table 1).

To establish that the E2 LBD of the D351G was still capable of AF2 activation, we used GST-ΔSRC-1 pull-down experiments using 35S-labeled D351 ER and D351G ER after treatment with vehicle, E2, 4-OHT, or ICI 182,780 (Fig. 9 ). After vehicle and E2 treatments, an increase in binding to SRC-1 occurred compared with antiestrogen treatments. Similar binding patterns were observed for D351 and D351G ERs. A possible explanation for the SRC-1 binding in the vehicle lane is that rabbit reticulocyte lysate was used that may contain E2. This is supported by the fact that all antiestrogens used prevented binding of SRC-1. These data support the view that the AF2 activity is still functioning in the D351G ER, and loss of 4-OHT agonist activity is not a result of the loss of AF2 function in the D351G ER as a whole.

Overall, these data suggest that D351G ER in JM-6 cells cannot activate AF1 with 4-OHT, whereas wild-type ER in S30 cells maintains AF1 activity with 4-OHT. To examine this observation further, we constructed a mutant ER with an inactivating triple mutation in AF2 and generated stable transfectants (G3m). We used Northern blot analysis of TGFα mRNA to compare S30 and JM-6 cells with G3m cells (Fig. 10). We treated these cells with vehicle, E2, or 4-OHT and measured TGFα transcription and found that 4-OHT was unable to activate TGFα transcription through an ER with an inactivated AF2,suggesting that an intact helix 12 is also necessary for the agonist activity of 4-OHT. In addition, when treated with E2, the TGFα induction in G3m cells was reduced, demonstrating that E2 requires both AF1 and AF2 for full activity; thus, the induction by E2 was through the AF1 alone.

We have shown that the simple substitution of glycine for aspartate at position 351 in the human ER retains robust estrogen action with E2 but silences the well-documented,estrogen-like action of 4-OHT (13, 26) that apparently resides in constitutive AF1 activity (12). Simply stated,a glycine substitution in the LBD silences AF1 activity in the distant A/B domain of ER and converts the estrogen-like 4-OHT:ER complex into an antiestrogenic or inhibitory complex at a TGFα gene target in situ.

We have focused our attention on amino acid 351 because of the finding that a natural mutation D351Y, isolated from a tamoxifen-stimulated breast tumor (20), could convert Ral from an antiestrogenic complex to an estrogenic complex at a TGFαgene target in the context of the MDA-MB-231 cells (21, 22). Naturally, we wondered whether the reverse would be true, i.e., by removing the charge from amino acid 351, could the 4-OHT:ER complex convert from being estrogen-like to antiestrogenic? If so, we could dissect out differences between the promiscuous 4-OHT:ER complex and the E2:ER complex.

In an early publication (25), we proposed an assay that could distinguish between tamoxifen-like, Ral-like, and ICI 182,780-like compounds. The assay methodology compared and contrasted the activation of the TGFα gene in MDA-MB-231 cells stably transfected with cDNAs from the wild-type and D351Y ER. Unfortunately, it was not possible, at that time, to distinguish the tamoxifen-like compounds from estrogens because both groups of compounds fully activated the TGFα gene target(26) with either wild-type or D351Y ER. With the validation of the D351G assay system, we can now select estrogens,tamoxifen-like, Ral-like, and ICI 182,780-like compounds by comparing the wild type transfectant with the D351Y and D351G transfectants. Although we believe that the application of our observation will help to identify new SERMs for clinical applications, we suggest that our finding provides an important new tool to identify novel coactivators and corepressors of ER function. Identification of a molecular mechanism to explain the promiscuous estrogenic actions of the 4-OHT:ER complex could potentially provide one of perhaps many reasons for the target site-specific actions of SERMs. Additionally, if a molecular mechanism for the estrogen-like action of tamoxifen at the ER could be established, then perhaps this knowledge could be used to counter the development of tamoxifen-stimulated breast cancer growth as a form of drug resistance.

Previous studies (12, 13) suggest that the agonist actions of 4-OHT may be a function of the target tissue as well as the promoter structure of the target gene. For convenience, several groups have used yeast cells to conduct transient transfection experiments(10, 33), but sometimes variable and inconsistent results are reported. An expression system that uses a yeast copper metallothionine promoter and ubiquitin fusion technology to express ER in yeast was unable to show that antiestrogens could block E2 action at cotransfected β-galactosidase reporter (34). However, the antiestrogen action alone did not produce a full estrogen-like response; therefore, the explanation for the negative findings is unclear. In contrast, Wrenn and Katzenellenbogen (35) noted that 4-OHT was a full agonist using a β-galactosidase reporter gene; however, the ER used for these pharmacology studies (34, 35) was the G400V ER, which is known to be unstable (33). G400V ER has a reduced affinity for E2 compared with wild-type ER and an enhanced estrogenicity with 4-OHT in stable transfectants of MDA-MB-231 cells(36). Paradoxically, 4-OHT is an antiestrogen with no estrogen-like properties when G400V is transiently transfected into Chinese hamster ovary cells using a catechol acetyl transferase reporter gene (35).

To establish assays of potential physiological relevance, the C3 promoter system has been ligated to a luciferase reporter gene. Importantly, the complex promoter has enhanced estrogen-like properties in the context of human hepatocellular carcinoma (HepG2) cells cotransfected with wild-type ER. The promoter system is physiologically relevant because E2 stimulated expression of the complement component in luminal epithelial cells of the rat uterus, but expression is not prevented by the coadministration of a number of antiestrogens including tamoxifen (37). In fact, the triphenylethylene type antiestrogens stimulates C3 mRNA(37), although the pure antiestrogen ICI 164,384 could block both E2 and 4-OHT-stimulated C3(38).

We initially used transient transfection of ER cDNAs into ER negative breast cancer cell lines and used a simple ERE luciferase reporter system to study ligand activity. Results were variable, and 4-OHT was not a full agonist, even when MDA-MB-231 cells were used (data not shown). It is clear that the agonist actions of a SERM:ER complex can only become apparent in the correct cellular context if a complex gene promoter such as C3 is used (13). We have taken another approach by studying ER actions at a TGFα gene target in vitro in the context of an MDA-MB-231 breast cancer cell(23). We reasoned that these cells would be replete with relevant transcription factors to regulate the TGFα gene,a known growth factor involved in cell proliferation. As a result, we stably transfected MDA-MB-231 breast cancer cells with cDNAs from D351G to compare and contrast with our established cell lines(23).

The wild-type and D351G ER have similar affinities for E2, and there is a concentration-related inhibition of [3H]E2 with antiestrogens. However, the fact that the stable transfectant, JM-6,with D351G ER had no agonist actions with 4-OHT compared with wild-type ER mandated that the stability of the ER was determined. Earlier studies with G400V ER demonstrated that the mutant receptor was turned over rapidly compared with the wild-type ER (33);therefore, it was possible that rapid receptor destruction was responsible for the loss of agonist activity for 4-OHT. This approach was particularly important because the pure antiestrogen ICI 182,780 is known to facilitate the rapid destruction of ER(39). As a result, ICI 182,780 probably exerts its pure antiestrogenic activity by producing both an inappropriate shape for the receptor complex and rapid destruction of complexes. However, we show the D351G is not destroyed (Fig. 3) by 4-OHT; therefore, this is unlikely to be the explanation for a decrease in the estrogen-like properties of the complex. A D351V mutation has been analyzed previously in the context of yeast cells (35). The D351V mutant was isolated as one of a series of mutations generated randomly by formic acid treatment (35). Interestingly, whereas 4-OHT was a full agonist with wild-type ER, the D351V ER did not possess estrogen-like actions. Although the ligand binding affinity or the antiestrogenic activity was not determined in yeast, D351V ER was antiestrogenic with 4-OHT in Chinese hamster ovary cells.

We have noted previously that the antiestrogens EM-652(25) and Ral (22) silence both AF1 and AF2 in the wild-type ER in the context of the MDA-MB-231 cells. It is,therefore, not inconsistent that we now find that the mutation D351G in the LBD can silence AF1 and AF2 activity with 4-OHT. The allosteric interaction of AF2 with AF1 was further investigated using an ER with an inactivating triple mutation in AF2 (G3m) stably transfected into MDA-MB-231 cells. We show that this mutation, which changes three amino acids that are carboxylic acids to uncharged alanines, also resulted in the ablation of 4-OHT agonist activity. This led us to deduce that both aspartate at 351 and an intact helix 12 are required for the partial agonist activity of 4-OHT (Fig. 10). This result with helix 12 mutations is consistent with previous results using HepG2 cells and a C3 promoter (13). In addition, the agonist activity of E2 in the G3m cells is reduced compared with D351G, suggesting that classical AF2 activity maybe affected in JM-6 cells when E2 is a binding ligand (Fig. 10).

We thought that it is important to compare and contrast the D351 and D351G ERs to be confident that the silencing of AF1 with 4-OHT in D351G is really the result of the single amino acid change in the LBD. We have shown that the binding characteristics for E2, 4-OHT, and ICI 182,780 are similar, but we took the comparison one step further. We show, using GST-pull down experiments, that D351G ER is still capable of binding the coactivator SRC-1 when bound to E2 but not when ligated with 4-OHT (Fig. 9). This provides evidence that AF2 is still functional in JM-6 cells for E2 action but is silenced for 4-OHT. For the future, the important goal is to discover the precise relationship of the AF1 and AF2 sites in the ER to identify site interactions and coactivator clusters. Unfortunately, the crystallographic structure is unhelpful, because only the LBD minus the F region has been crystallized (15, 16). At present, it is also not possible to perform pull-down assays with the whole ER. A method is required to investigate the interaction of coactivators and corepressors with the whole receptor to isolate novel proteins. Nevertheless, on the basis of our experimental evidence, we believe it is appropriate to propose a model of ER pharmacology as a framework for further study.

Recently, Norris et al.(27) have identified peptides that can distinguish between E2:ER complexes and 4-OHT:ER complexes. Different peptides can block E2 and 4-OHT agonist activities, thereby,demonstrating that agonism occurs by different mechanisms at different surfaces. In other words, coactivator binding could be ER complex specific. The repositioning of helix 12 to seal E2 into the LBD enables coactivators such as GRIP1 to bind in the AF2 domain (16). The blocking of this site by helix 12 with either 4-OHT (16) or Ral(15) as the binding ligand raises the issue of a second site that could bind a coactivator molecule. It is interesting to speculate that coactivator binding for 4-OHT:ER complexes could involve a point attachment in the region 300–351 with AF1. The region 282–351 has been noted previously by Norris et al.(40)to be a potential transactivating site on ER that they have termed AF2a.

The antiestrogens 4-OHT and Ral have a side chain that interacts with aspartate 351 (Fig. 1). However, the interaction of the tertiary nitrogen of 4-OHT with the carboxylic acid is tenuous at best; thus, we propose that unlike Ral, the charge on the aspartate 351 is not neutralized. Recently, we have resolved the space-filling structure of the ER dimer with either DES or 4-OHT as the binding ligand (Fig. 11). We have identified aspartate 351 as a surface amino acid that can potentially interact with other coactivator molecules. Fig. 11illustrates the profound difference on the external surface of the DES:ER or 4-OHT:ER complex. We suggest that the difference in the positions of the side chain of 4-OHT and Ral results in the creation of a potential docking site for coactivators for the 4-OHT:ER complex. Although it is unlikely that this is the only docking site,which may also include the AF1 site and helix 12, we suggest that the loss of charge at aspartate 351 by the substitution of glycine is critical for the loss of estrogenicity in the 4-OHT complex. We are currently addressing the structure-function relationships of this novel target on the surface of the ER.

Fig. 1.

Diagram based on the crystal structure of the ER LBD complexed with either E2, 4-OHT, or Ral (15, 16). The amino acids that interact with the ligand have been identified, and bond lengths, when important, have been noted. The significant difference between E2 and the antiestrogens is the interaction with aspartate 351. However, the antiestrogenic side chain of Ral appears to be closer than is possible for the side chain of 4-OHT. We propose that this is an important difference between 4-OHT and Ral complexes that could result in an increased E2-like action of the 4-OHT:ER complex.

Fig. 1.

Diagram based on the crystal structure of the ER LBD complexed with either E2, 4-OHT, or Ral (15, 16). The amino acids that interact with the ligand have been identified, and bond lengths, when important, have been noted. The significant difference between E2 and the antiestrogens is the interaction with aspartate 351. However, the antiestrogenic side chain of Ral appears to be closer than is possible for the side chain of 4-OHT. We propose that this is an important difference between 4-OHT and Ral complexes that could result in an increased E2-like action of the 4-OHT:ER complex.

Close modal
Fig. 2.

Western blot screen of ER protein expression in MDA-MB-231 cells stably expressing D351 (S30) or D351G(JM-6). Western blot analysis was performed as described in “Materials and Methods.” β-actin protein was measured to ensure even loading. Western blot is representative of three independent experiments.

Fig. 2.

Western blot screen of ER protein expression in MDA-MB-231 cells stably expressing D351 (S30) or D351G(JM-6). Western blot analysis was performed as described in “Materials and Methods.” β-actin protein was measured to ensure even loading. Western blot is representative of three independent experiments.

Close modal
Fig. 3.

Western blot screen of ER protein expression in MDA-MB-231 cells stably expressing D351 (S30; A) or D351G (JM-6; B). Cells were treated for 24 h with vehicle (ethanol), 10−9mE2, 10−7m 4-OHT, or 10−7m ICI 182,780. Western blot analysis was performed as described in “Materials and Methods.” β-actin protein was measured to ensure even loading. Western blots are representative of four independent experiments.

Fig. 3.

Western blot screen of ER protein expression in MDA-MB-231 cells stably expressing D351 (S30; A) or D351G (JM-6; B). Cells were treated for 24 h with vehicle (ethanol), 10−9mE2, 10−7m 4-OHT, or 10−7m ICI 182,780. Western blot analysis was performed as described in “Materials and Methods.” β-actin protein was measured to ensure even loading. Western blots are representative of four independent experiments.

Close modal
Fig. 4.

Ligand binding assay was performed as described in“Materials and Methods.” Competition assays were performed in S30 and JM-6 cells. The data shown are the combination of three independent experiments, each done in duplicate; bars, SD.

Fig. 4.

Ligand binding assay was performed as described in“Materials and Methods.” Competition assays were performed in S30 and JM-6 cells. The data shown are the combination of three independent experiments, each done in duplicate; bars, SD.

Close modal
Fig. 5.

Northern blot analysis of TGFα mRNA in JM-6 cells treated with E2. A, JM-6 cells grown in E2-free stripped media were treated for 24 h with increasing doses of E2. Northern blot analysis of total RNA was performed using TGFα cDNA probe as described in “Materials and Methods.” β-actin mRNA was measured to ensure even loading. Northern blot is representative of three independent experiments. B, densitometric analysis of Northern blots was analyzed by ANOVA, and E2 significantly induced TGFα in a concentration-dependent manner (P < 0.05). Bars, SD.

Fig. 5.

Northern blot analysis of TGFα mRNA in JM-6 cells treated with E2. A, JM-6 cells grown in E2-free stripped media were treated for 24 h with increasing doses of E2. Northern blot analysis of total RNA was performed using TGFα cDNA probe as described in “Materials and Methods.” β-actin mRNA was measured to ensure even loading. Northern blot is representative of three independent experiments. B, densitometric analysis of Northern blots was analyzed by ANOVA, and E2 significantly induced TGFα in a concentration-dependent manner (P < 0.05). Bars, SD.

Close modal
Fig. 6.

Northern blot analysis of TGFα mRNA in S30(A, B) and JM-6 (C, D) cells treated with E2 or various antiestrogens. S30 and JM-6 cells grown in E2-free stripped media were treated for 24 h with vehicle (ethanol), 10−9mE2, 10−7m 4-OHT,10−7m Ral, or 10−7mICI 182,780. Northern blot analysis of total RNA was performed using TGFα cDNA probe as described in “Materials and Methods.”β-actin mRNA was measured to ensure even loading. Northern blot is representative of three independent experiments. Densitometric analysis of Northern blots was analyzed by Student’s t test. The differences between control and E2 and control and 4-OHT were significant (P = 0.02), but there was no significant difference between E2 and 4-OHT in S30 cells (P > 0.05; B). The difference between control and E2 was significant(P < 0.05), but control and 4-OHT was not significantly different in JM-6 cells (P = 0.4; D). Bars, SD.

Fig. 6.

Northern blot analysis of TGFα mRNA in S30(A, B) and JM-6 (C, D) cells treated with E2 or various antiestrogens. S30 and JM-6 cells grown in E2-free stripped media were treated for 24 h with vehicle (ethanol), 10−9mE2, 10−7m 4-OHT,10−7m Ral, or 10−7mICI 182,780. Northern blot analysis of total RNA was performed using TGFα cDNA probe as described in “Materials and Methods.”β-actin mRNA was measured to ensure even loading. Northern blot is representative of three independent experiments. Densitometric analysis of Northern blots was analyzed by Student’s t test. The differences between control and E2 and control and 4-OHT were significant (P = 0.02), but there was no significant difference between E2 and 4-OHT in S30 cells (P > 0.05; B). The difference between control and E2 was significant(P < 0.05), but control and 4-OHT was not significantly different in JM-6 cells (P = 0.4; D). Bars, SD.

Close modal
Fig. 7.

Effects of the pure antiestrogen ICI 182,780 TGFα mRNA expression in combination with either E2 or antiestrogen treatment. JM-6 cells grown in E2-free stripped media were treated for 24 h with vehicle (ethanol), 10−9m E2, 10−7m 4-OHT,10−7m Ral, or 10−7mICI 182,780 or combinations. β-actin mRNA was measured to ensure even loading. Northern blot (A) is representative of three independent experiments. Densitometric analysis of Northern blots(B) was analyzed by Student’s t test. The difference between control and E2 was significant(P < 0.05), and control and 4-OHT was not significant (P > 0.05). Bars, SD.

Fig. 7.

Effects of the pure antiestrogen ICI 182,780 TGFα mRNA expression in combination with either E2 or antiestrogen treatment. JM-6 cells grown in E2-free stripped media were treated for 24 h with vehicle (ethanol), 10−9m E2, 10−7m 4-OHT,10−7m Ral, or 10−7mICI 182,780 or combinations. β-actin mRNA was measured to ensure even loading. Northern blot (A) is representative of three independent experiments. Densitometric analysis of Northern blots(B) was analyzed by Student’s t test. The difference between control and E2 was significant(P < 0.05), and control and 4-OHT was not significant (P > 0.05). Bars, SD.

Close modal
Fig. 8.

Dose-responsive increase in TGFα transcript expression in JM-6 cells after treatment with 4-OHT (A), Ral(C), or ICI 182,780 (E) measured by Northern blot analysis. JM-6 cells grown in E2-free stripped media were treated for 24 h with 10−9m E2 and increasing doses of 4-OHT, Ral, or ICI 182,780. β-actin mRNA was measured to ensure even loading. Northern blot is representative of three independent experiments. Densitometric analysis of Northern blots was analyzed by ANOVA, and each antiestrogen significantly reduced TGFα in a concentration-dependent manner. B, 4-OHT; P = 0.002; bars, SD. D, Ral; P = 0.003; bars, SD. F, ICI 182,780; P < 0.05; bars, SD.

Fig. 8.

Dose-responsive increase in TGFα transcript expression in JM-6 cells after treatment with 4-OHT (A), Ral(C), or ICI 182,780 (E) measured by Northern blot analysis. JM-6 cells grown in E2-free stripped media were treated for 24 h with 10−9m E2 and increasing doses of 4-OHT, Ral, or ICI 182,780. β-actin mRNA was measured to ensure even loading. Northern blot is representative of three independent experiments. Densitometric analysis of Northern blots was analyzed by ANOVA, and each antiestrogen significantly reduced TGFα in a concentration-dependent manner. B, 4-OHT; P = 0.002; bars, SD. D, Ral; P = 0.003; bars, SD. F, ICI 182,780; P < 0.05; bars, SD.

Close modal
Fig. 9.

The recruitment of SRC-1 to the LBD of D351 and D351G. Control and E2, but not antiestrogens, aid in SRC-1 binding. The pull-down assay was conducted as described in “Materials and Methods.” Five μl of 35S-labeled wild-type ER or 35S-labeled D351G ER were incubated with GST or GST-ΔSRC-1 with vehicle or different ligands (10−8m E2, 10−6m 4-OHT, or 10−6m Ral). The bound 35S-labeled wild-type ER or 35S-labeled D351G ER was resolved by 7.5%SDS-PAGE. The pull-down assay is representative of two independent experiments.

Fig. 9.

The recruitment of SRC-1 to the LBD of D351 and D351G. Control and E2, but not antiestrogens, aid in SRC-1 binding. The pull-down assay was conducted as described in “Materials and Methods.” Five μl of 35S-labeled wild-type ER or 35S-labeled D351G ER were incubated with GST or GST-ΔSRC-1 with vehicle or different ligands (10−8m E2, 10−6m 4-OHT, or 10−6m Ral). The bound 35S-labeled wild-type ER or 35S-labeled D351G ER was resolved by 7.5%SDS-PAGE. The pull-down assay is representative of two independent experiments.

Close modal
Fig. 10.

A, Northern blot analysis of TGFαmRNA in S30, JM-6 cells, and G3m cells treated with E2 or 4-OHT. S30, JM-6 cells, and G3m grown in E2-free stripped media were treated for 24 h with vehicle (ethanol),10−9m E2, 10−7m 4-OHT. Northern blot analysis of total RNA was performed using TGFα cDNA probe as described in “Materials and Methods.”β-actin mRNA was measured to ensure even loading. Northern blot is representative of three independent experiments. The combined data set is illustrated in the histogram, and the response of each clone is to E2 and 4-OHT, analyzed by Student’s t test. The difference between control and E2 was significant in S30 (P < 0.05) and JM-6(P < 0.05); the difference between control and E2 in G3m was not significant(P = 0.1). The difference between control and 4-OHT was significant S30 (P < 0.05); the difference between control and E2 in was not significant in JM-6 (P = 0.67) and G3m(P = 0.29). Bars, SD.

Fig. 10.

A, Northern blot analysis of TGFαmRNA in S30, JM-6 cells, and G3m cells treated with E2 or 4-OHT. S30, JM-6 cells, and G3m grown in E2-free stripped media were treated for 24 h with vehicle (ethanol),10−9m E2, 10−7m 4-OHT. Northern blot analysis of total RNA was performed using TGFα cDNA probe as described in “Materials and Methods.”β-actin mRNA was measured to ensure even loading. Northern blot is representative of three independent experiments. The combined data set is illustrated in the histogram, and the response of each clone is to E2 and 4-OHT, analyzed by Student’s t test. The difference between control and E2 was significant in S30 (P < 0.05) and JM-6(P < 0.05); the difference between control and E2 in G3m was not significant(P = 0.1). The difference between control and 4-OHT was significant S30 (P < 0.05); the difference between control and E2 in was not significant in JM-6 (P = 0.67) and G3m(P = 0.29). Bars, SD.

Close modal
Fig. 11.

A, the water-accessible surface view of the ER dimer liganded with the synthetic estrogen DES. The ligand cannot be seen because it is sealed within the hydrophobic LBD by helix 12, indicated in yellow. The position of helix 12 allows GRIP, a coactivator, to bind at part of the AF2 site. The three amino acids that are mutated to neutralize carboxylic acids(D538A/E542A/D545A; G3m) are indicated in orange on helix 12. Aspartate 351 lies on the surface next to helix 12. When the ER is liganded with 4-OHT (B), only the side chain is visible protruding from the complex near to aspartate 351 on the surface. Helix 12 is repositioned to block AF2 binding at the GRIP site. Because G3m blocks the agonist action of 4-OHT, then helix 12 may be involved, with its free carboxylic acids, in coactivator binding in AF2b. The representations in C and D are the water-accessible surface views in the presumed D351 and D351G proteins with 4-OHT as the binding ligand. No charge is available to bind in the AF2b binding site. The structural model of dimeric human ER bound to DES or 4-OHT was constructed from 3ERTpdb (16)using crystallographic symmetry operations. After removing all of the water molecules except the ordered water-forming hydrogen bonds with DES or 4-OHT, the model was minimized into the consistent valence force field using Discover (Molecular Simulations, Inc., San Diego, CA).

Fig. 11.

A, the water-accessible surface view of the ER dimer liganded with the synthetic estrogen DES. The ligand cannot be seen because it is sealed within the hydrophobic LBD by helix 12, indicated in yellow. The position of helix 12 allows GRIP, a coactivator, to bind at part of the AF2 site. The three amino acids that are mutated to neutralize carboxylic acids(D538A/E542A/D545A; G3m) are indicated in orange on helix 12. Aspartate 351 lies on the surface next to helix 12. When the ER is liganded with 4-OHT (B), only the side chain is visible protruding from the complex near to aspartate 351 on the surface. Helix 12 is repositioned to block AF2 binding at the GRIP site. Because G3m blocks the agonist action of 4-OHT, then helix 12 may be involved, with its free carboxylic acids, in coactivator binding in AF2b. The representations in C and D are the water-accessible surface views in the presumed D351 and D351G proteins with 4-OHT as the binding ligand. No charge is available to bind in the AF2b binding site. The structural model of dimeric human ER bound to DES or 4-OHT was constructed from 3ERTpdb (16)using crystallographic symmetry operations. After removing all of the water molecules except the ordered water-forming hydrogen bonds with DES or 4-OHT, the model was minimized into the consistent valence force field using Discover (Molecular Simulations, Inc., San Diego, CA).

Close modal

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.

1

Supported in part by Department of Defense Breast Cancer Training Grants DAMD 17-94-J-4466, DAMD 17-96-1-6169, and Breast Cancer Program Development Grants P20 CA65764 and RO1-CA56143. We are deeply grateful for support from the Lynn Sage Breast Cancer Research Foundation of Northwestern Memorial Hospital and the Avon Breast Cancer Research Fund.

3

The abbreviations used are: ER, estrogen receptor α; 4-OHT, 4-hydroxytamoxifen; Ral, raloxifene; TGF,transforming growth factor; SERM, selective estrogen receptor modulator; AF1, activating function 1; AF2, activating function 2; LBD,ligand binding domain; EC50, effective 50% inhibitory concentration; GST, glutathione S-transferase; DES,diethylstilbestrol.

Table 1

Binding affinity of E2 and EC50 for antiestrogens in S30 and JM-6 cellsa

Kd (nm) E2EC50 (nm)
4-OHTRalICI 182,780
S30 0.42 ± 0.08 2.14 ± 1.52 1.04 ± 0.22 4.38 ± 1.49 
JM-6 0.77 ± 0.17 5.87 ± 0.37 13.24 ± 1.45 3.16 ± 0.96 
Kd (nm) E2EC50 (nm)
4-OHTRalICI 182,780
S30 0.42 ± 0.08 2.14 ± 1.52 1.04 ± 0.22 4.38 ± 1.49 
JM-6 0.77 ± 0.17 5.87 ± 0.37 13.24 ± 1.45 3.16 ± 0.96 
a

The difference in binding affinity between S30 and JM-6 was not significant. The EC50s of compounds were similar except for Ral(P < 0.05) by Student’s ttest.

1
Early Breast Cancer Trialists’ Collaborative Group. Tamoxifen for early breast cancer: an overview of the randomised trials.
Lancet
,
351
:
1451
-1467,  
1998
.
2
Fisher B., Costantino J. P., Wickerham D. L., Redmond C. K., Kavanah M., Cronin W. M., Vogel V., Robidoux A., Dimitrov N., Atkins J., Daly M., Wieand S., Tan-Chiu E., Ford L., Wolmark N. Tamoxifen for prevention of breast cancer: report of the National Surgical Adjuvant Breast and Bowel Project P-1 Study.
J. Natl. Cancer Inst.
,
90
:
1371
-1388,  
1998
.
3
MacGregor J. I., Jordan V. C. Basic guide to the mechanisms of antiestrogen action.
Pharmacol. Rev.
,
50
:
151
-196,  
1998
.
4
Gottardis M. M., Jordan V. C. Development of tamoxifen-stimulated growth of MCF-7 tumors in athymic mice after long-term antiestrogen administration.
Cancer Res.
,
48
:
5183
-5187,  
1988
.
5
Gottardis M. M., Jiang S-Y., Jeng M-H., Jordan V. C. Inhibition of tamoxifen-stimulated growth of an MCF-7 tumor variant in athymic mice by novel steroidal antiestrogens.
Cancer Res.
,
49
:
4090
-4093,  
1989
.
6
Levenson A. S., Jordan V. C. Selective oestrogen receptor modulation: molecular pharmacology for the millennium.
Eur. J. Cancer
,
35
:
1628
-1639,  
1999
.
7
Jordan V. C., Collins M. M., Rowsby L., Prestwich G. A monohydroxylated metabolite of tamoxifen with potent antioestrogenic activity.
J. Endocrinol.
,
75
:
305
-316,  
1977
.
8
Borgna J. L., Rochefort H. Hydroxylated metabolites of tamoxifen are formed in vivo and bound to estrogen receptor in target tissues.
J. Biol. Chem.
,
256
:
859
-868,  
1981
.
9
Coezy E., Borgna J. L., Rochefort H. Tamoxifen and metabolites in MCF-7 cells: correlation between binding to estrogen receptor and cell growth inhibition.
Cancer Res.
,
42
:
317
-323,  
1982
.
10
Metzger D., White J. H., Chambon P. The human oestrogen receptor functions in yeast.
Nature (Lond.)
,
334
:
31
-36,  
1988
.
11
Kumar V., Green S., Stack G., Berry M., Jin J. R., Chambon P. Functional domains of the human estrogen receptor.
Cell
,
51
:
941
-951,  
1987
.
12
Berry M., Metzger D., Chambon P. Role of the two activating domains of the oestrogen receptor in the cell-type and promoter-context dependent agonistic activity of the anti-oestrogen 4-hydroxytamoxifen.
EMBO J.
,
9
:
2811
-2818,  
1990
.
13
Tzukerman M. T., Esty A., Santiso-Mere D., Danielian P., Parker M. G., Stein R. B., Pike J., McDonnell D. P. Human estrogen receptor transactivational capacity is determined by both cellular and promoter context and mediated by two functionally distinct intramolecular regions.
Mol. Endocrinol.
,
94
:
21
-30,  
1994
.
14
Kraus W. L., McInerney E. M., Katzenellenbogen B. S. Ligand-dependent, transcriptionally productive association of the amino- and carboxyl-terminal regions of a steroid hormone nuclear receptor.
Proc. Natl. Acad. Sci. USA
,
92
:
12314
-12318,  
1995
.
15
Brzozowski, A. M., Pike, A. C. W., Dauter, Z., Hubbard, R. E., Bonn, T., Engström, O., Öhman, L., Greene, G. L., Gustafsson, J. Å., and Carlquist, M. Molecular basis of agonism and antagonism in the oestrogen receptor. Nature (Lond.), 389: 753–758, 1997.
16
Shiau A. K., Barstad D., Loria P. M., Cheng L., Kushner P. J., Agard D. A., Greene G. L. The structural basis of estrogen receptor/co-activator recognition and the antagonism of this interaction by tamoxifen.
Cell
,
95
:
927
-937,  
1998
.
17
Jordan V. C. Biochemical pharmacology of antiestrogen action.
Pharmacol. Rev.
,
36
:
337
-350,  
1984
.
18
Lieberman M. E., Gorski J., Jordan V. C. An estrogen receptor model to describe the regulation of prolactin synthesis by antiestrogens in vitro.
J. Biol. Chem.
,
258
:
4741
-4745,  
1983
.
19
Tate A. C., Greene G. L., DeSombre E. R., Jensen E. V., Jordan V. C. Differences between estrogen- and antiestrogen-estrogen receptor complexes from human breast tumors identified with an antibody raised against the estrogen receptor.
Cancer Res.
,
44
:
1012
-1018,  
1984
.
20
Wolf D. M., Jordan V. C. The estrogen receptor from a tamoxifen stimulated MCF-7 tumor variant contains a point mutation in the ligand binding domain.
Breast Cancer Res. Treat.
,
31
:
129
-138,  
1994
.
21
Levenson A. S., Catherino W. H., Jordan V. C. Estrogenic activity is increased for an antiestrogen by a natural mutation of the estrogen receptor.
J. Steroid Biochem. Mol. Biol.
,
60
:
261
-268,  
1997
.
22
Levenson A. S., Jordan V. C. The key to the antiestrogenic mechanism of raloxifene is amino acid 351 (aspartate) in the estrogen receptor.
Cancer Res.
,
58
:
1872
-1875,  
1998
.
23
Jiang S-Y., Jordan V. C. Growth regulation of estrogen receptor negative breast cancer cells transfected with cDNAs for estrogen receptor.
J. Natl. Cancer Inst.
,
84
:
580
-591,  
1992
.
24
Catherino W. H., Wolf D. M., Jordan V. C. A naturally occurring estrogen receptor mutation results in increased estrogenicity of a tamoxifen analog.
Mol. Endocrinol.
,
9
:
1053
-1063,  
1995
.
25
MacGregor Schafer J. I., Liu H., Tonetti D. A., Jordan V. C. The interaction of raloxifene and the active metabolite of the antiestrogen EM-800 (SC 5705) with the human estrogen receptor.
Cancer Res.
,
59
:
4308
-4313,  
1999
.
26
Levenson A. S., Tonetti D. A., Jordan V. C. The oestrogen-like effect of 4-hydroxytamoxifen on induction of transforming growth factor α mRNA in MDA-MB-231 breast cancer cells stably expressing the oestrogen receptor.
Br. J. Cancer
,
77
:
1812
-1819,  
1998
.
27
Norris J. D., Paige L. A., Christensen D. J., Chang C. Y., Huacani M. R., Fan D., Hamilton P. T., Fowlkes D. M., McDonnell D. P. Peptide antagonists of the human estrogen receptor.
Science (Washington DC)
,
285
:
744
-746,  
1999
.
28
Smith C. L., Nawaz Z., O’Malley B. W. Coactivator and corepressor regulation of the agonist/antagonist activity of the mixed antiestrogen, 4-hydroxytamoxifen.
Mol. Endocrinol.
,
11
:
657
-666,  
1997
.
29
Onate S. A., Tsai S. Y., Tsai M. J., O’Malley B. W. Sequence and characterization of a coactivator for the steroid hormone receptor superfamily.
Science (Washington DC)
,
270
:
1354
-1357,  
1995
.
30
Halachmi S., Marden E., Martin G., MacKay H., Abbondanza C., Brown M. Estrogen receptor-associated proteins: possible mediators of hormone-induced transcription.
Science (Washington DC)
,
264
:
1455
-1458,  
1994
.
31
Hanstein B., Eckner R., DiRenzo J., Halachmi S., Liu H., Searcy B., Kurokawa R., Brown M. p300 is a component of an estrogen receptor coactivator complex.
Proc. Natl. Acad. Sci. USA
,
93
:
11540
-11545,  
1996
.
32
Liu H., Wormke M., Safe S. H., Bjeldanes L. F. Indolo[3,2-b]carbazole: a dietary-derived factor that exhibits both antiestrogenic and estrogenic activity.
J. Natl. Cancer Inst.
,
86
:
1758
-1765,  
1994
.
33
Tora L., White J., Brou C. The human estrogen receptor has two independent nonacidic transcriptional activation functions.
Cell
,
59
:
477
-487,  
1989
.
34
Lyttle C. R., Damian-Matsumura P., Juul H., Butt T. R. Human estrogen receptor regulation in a yeast model system and studies on receptor agonists and antagonists.
J. Steroid. Biochem. Mol. Biol.
,
42
:
677
-685,  
1992
.
35
Wrenn C. K., Katzenellenbogen B. S. Structure-function analysis of the hormone binding domain of the human estrogen receptor by region-specific mutagenesis and phenotypic screening in yeast.
J. Biol. Chem.
,
268
:
24089
-24098,  
1993
.
36
Jiang S. Y., Parker C. J., Jordan V. C. A model to describe how a point mutation of the estrogen receptor alters the structure-function relationship of antiestrogens.
Breast Cancer Res. Treat.
,
26
:
139
-147,  
1993
.
37
Sundstrom S. A., Komm B. S., Xu Q., Boundy V., Lyttle C. R. The stimulation of uterine complement component C3 gene expression by antiestrogens.
Endocrinology
,
126
:
1449
-1456,  
1990
.
38
Galman M. S., Sundstrom S. A., Lyttle C. R. Antagonism of estrogen- and antiestrogen-induced uterine complement component C3 expression by ICI 164,384.
J. Steroid. Biochem.
,
36
:
281
-286,  
1990
.
39
Dauvois S., White R., Parker M. G. The antiestrogen ICI 182780 disrupts estrogen receptor nucleocytoplasmic shuttling.
J. Cell Sci.
,
106
:
1377
-1388,  
1993
.
40
Norris J. D., Fan D., Kerner S. A., McDonnell D. P. Identification of a third autonomous activation domain within the human estrogen receptor.
Mol. Endocrinol.
,
11
:
747
-754,  
1997
.