| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Departments of Chemistry (V.S.L., J.A.K.) and Molecular and Integrative Physiology (F.S., K.K., B.S.K.), University of Illinois at Urbana-Champaign, Urbana, Illinois 61801
Address all correspondence and requests for reprints to: John A. Katzenellenbogen, Department of Chemistry, University of Illinois, 600 South Mathews Avenue, Urbana, Illinois 61801. E-mail: jkatzene{at}uiuc.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
and ß (ER and ERß) through genomic (transcription regulation) and nongenomic pathways (activation of signaling cascades, endothelial nitric oxide synthase, and ion channels) (1). ER is thus an important target in human medicine for treatment of breast cancer. Although hormone therapy with antiestrogens or aromatase inhibitors is initially effective in most ER-positive tumors, treatment typically fails with time as tumors become hormone resistant (2). Although many mechanisms for hormone resistance have been proposed, recent evidence suggests that cellular adaptations resulting in altered posttranslational modifications (PTMs) of ER and of its interacting coregulatory partners play major roles in recurrence and aggressive tumor growth. As a result of these PTM changes, tamoxifen no longer inhibits ER action and can even act as an agonist, encouraging tamoxifen-dependent growth of tumors (3, 4). Although resistance to aromatase inhibitors typically develops more slowly, related PTM changes that enable ER to act in a ligand-independent fashion likely underlie this form of resistance as well (2). Thus, understanding the effects of these PTMs on ER function is an essential step in understanding drug resistance and in formulating combinatorial drug treatment for better management of breast cancer. ER is known to be the target of several PTMs: phosphorylation, glycosylation, and acetylation (5). Of these, ER phosphorylation is considered most important; it results from the activation of various cellular kinases and alters ER action. A number of different factors have been demonstrated to activate kinases that phosphorylate ER (2, 6). Notably, the nongenomic actions of estrogen are associated with the activation of a number of kinases, including Src, MAPK, and AKT, all of which phosphorylate ER (5). In addition, peptide growth factor receptors, including epidermal growth factor and HER2 that are typically overexpressed in about 30% of invasive cancers, activate Src, MAPK, and phosphatidylinositol 3-kinase/AKT pathways that can modify ER, even in the absence of estrogen (7).
A number of phosphorylation sites on ER have been characterized by biochemical methods (Fig. 1A
) (5). Phosphorylations at serine 102, 104 [cdk (8, 9)] and 118 [MAPK (9, 10)], 167 [RSK (11), AKT (12)] are located in the N-terminal A/B domain. Interestingly, this region is responsible for the ligand-independent transactivation functions of ER (13). Also, tamoxifen-resistant breast tumors exhibit higher phosphorylation at both serine 118 and 167 (14, 15, 16). The phosphorylation at serine 236 in the DNA binding domain has been shown to alter DNA binding as studied by mutagenesis and gel shift assays using cell extracts (9, 17), and in vitro phosphorylation by PKA at serine 305 at the junction of the hinge region (D domain) and the ligand binding domain (LBD) has been demonstrated and might be involved in tamoxifen resistance (18). In addition, the LBD of ER can be phosphorylated at tyrosine 537 by Src kinase (19, 20, 21), and mutagenesis studies indicate a role of this phosphorylation in modulation of ligand and DNA binding, resulting in increased transcription. These prior studies used cell extracts, intact cells, or mutagenesis methods to determine the effect of phosphorylation on ER function and estrogen action. However, because they failed to delineate the precise effect of these individual phosphorylation events on important receptor functions, such as ligand binding, DNA binding, and coactivator binding, we undertook the current study to examine how receptor phosphorylation alters these important determinants of ER action.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
, left). In the parental MCF-7 cells, phosphorylation at both S118 and at S167 was ligand dependent. Very low levels of phosphorylation at either of these positions were observed in vehicle control cells, whereas in E2-treated cells, a high level of phosphorylation at S118 and a low level of phosphorylation at S167 were detected. In the presence of TOT, however, the pattern of phosphorylation was different: a low level of phosphorylation at S118 and a stronger phosphorylation signal at S167 were observed, indicating a ligand-dependent and ligand-specific preference for phosphorylation of ER in parental MCF-7 cells.
In contrast, in the tamoxifen-resistant MCF-7 Her2/neu cells, both S118 and S167 exhibited similar levels of phosphorylation in the presence of vehicle or E2 or TOT (Fig. 1B, right). Thus, a notable difference between the two cell lines is the very low level of endogenous ER phosphorylation in vehicle-treated parental MCF-7 cells compared with high receptor phosphorylation levels in comparably vehicle-treated MCF-7 Her2/neu cells, indicating a loss of ligand-dependent and differential phosphorylation of S118 and S167 in the tamoxifen-resistant cells. An increase in S118 phosphorylation of ER in response to both E2 and TOT has been reported by Shou et al. (22) in MCF-7 and MCF-7 Her2/neu; however, their cell treatment included serum starvation, which may have decreased Her2 signaling due to lack of heregulin. The changes in phosphorylation levels we observe can be attributed to the increased activation of kinases by the overexpressed Her2 receptors in these cells. (The level of endogenous ER detected by the D12 antibody is very comparable in different treatments.) Thus, overexpression of Her2 receptor leads to a loss of both ligand-dependent and ligand-specific phosphorylation of ER in MCF-7Her2/neu cells. Because posttranslational modifications have been indicated to modify receptor function, this change in phosphorylation status could cause altered ligand response and cell physiology in the MCF-7 Her2/neu cells (22).
In Vitro Phosphorylation of ER: Establishing Specificity and Stoichiometry of ER Modification
To determine whether receptor phosphorylation modulates receptor function, we first characterized the phosphorylation of ER using purified full-length receptor phosphorylated in vitro by either MAPK, AKT, Src, PKA, or a no enzyme (control), as described in Materials and Methods. After phosphorylation of the receptor by the kinases in vitro, the control and the phosphorylated proteins were fractionated on SDS-PAGE, transferred to a nylon membrane, and immunoblotted using phospho-specific or anti-FLAG antibodies (Fig. 2A). A phosphorylation-specific signal was observed only when MAPK, AKT, Src, or PKA was included in the assay. Immunoblotting by anti-FLAG antibody represents total ER in both unphosphorylated control and in kinase-phosphorylated ER lanes. Importantly, the control ER in the absence of kinase did not yield any signal for phosphorylation by Western immunoblotting, indicating absence of phosphorylation of control ER at the sites under study. Whereas others have reported that ER expressed in baculovirus is phosphorylated (23), under our expression conditions, conducted in serum-free medium, we find no evidence of phosphorylation (also see comment below on mass spectrometric analysis of our expressed ER).
|
We verified the specificity of our antibodies further by confirming loss of immunodetection of mutant ERs in which specific kinase phosphorylation sites were removed by mutation (23). Thus, we generated mutant FLAG-tagged ERs that had serine to alanine (S118A, S167A, S236A, S305A, and double mutant S236/305A) or tyrosine to alanine (Y537A) substitutions at the phosphorylation sites of interest. Wild-type ER and these mutants were then expressed in human embryonic kidney (HEK)293T cells and purified to homogeneity (Fig. 2B, bottom). In this expression system, wild-type ER is hyperphosphorylated, as confirmed by the absence of further phosphorylation of this receptor in vitro upon treatment with certain of the kinases and 
-[32P]ATP. All of the kinases were shown to be active under our treatment conditions because inclusion of purified insect cell-expressed ER resulted in active phosphate incorporation, confirmed by immunoblot analysis (data not shown).
The purified, hyperphosphorylated wild-type and mutant ERs expressed in HEK293T cells were then analyzed for phosphorylation events by Western immunoblotting using phospho-specific antibodies (Fig. 2B). For each of the antibodies used, a strong band was visualized for wild-type ER, whereas no signal was observed for the S118A, S167A, S236A, and S236/305A mutant receptors, confirming the specificity of the antibodies for detection of these sites. A very low-intensity band was visualized for S305A. Although of lower intensity than wild-type ER, a strong band was apparent with Y537A ER, indicating that Src phosphorylates at sites in ER other than Y537 (see below).
To quantify the stoichiometry of kinase-specific phosphorylation of ER, incorporation of [32P]phosphate from -[32P]ATP into baculovirus-expressed wild-type ER was measured in three independent experiments (as described in Materials and Methods); the levels of phosphorylation achieved are expressed as (picomoles phosphate/picomoles ER) x 100 in Fig. 2C. All of the four kinase-specific phosphorylations were robust, with 80, 93, 192, and 85% of the ER being phosphorylated by MAPK, AKT, Src, and PKA, respectively. Interestingly, the high level of ER phosphorylation by Src (192%) indicates more than one phosphorylation event per ER molecule (see above). In addition, in preliminary mass spectrometric analysis of ER phosphorylation, we found no phosphorylation of baculovirus-expressed wild-type ER in the absence of kinase treatment; single phosphorylation events on ER when receptor was treated with MAPK or AKT, and two phosphorylation events upon treatment with Src (Likhite, V. S., M. T. Boyne, N. L. Kelleher, and J. A. Katzenellenbogen, unpublished data).
Taken together, our results suggest that under our experimental conditions baculovirus-expressed wild-type ER is unphosphorylated. MAPK and AKT phosphorylate ER specifically and stoichiometrically at the previously known sites. ER is phosphorylated by Src at at least two sites, including tyrosine 537. Although Src has been previously reported to phosphorylate ER at tyrosine 537 (24), our findings appear to support the recent work by Kushner and co-workers (25), which indicates the Src-mediated action of ER involves the AF1 domain of the receptor, a region that is rich in tyrosines. PKA phosphorylates ER at serine 236 and 305 sites, as mutations at these sites abolish phosphorylation. The specificity of ER phosphorylation by PKA is similar to that previously observed by Michalides et al. (18). Our use of rapid and sensitive methodologies may have enabled detection and identification of phosphorylation events not previously detected.
ER Phosphorylation Alters Ligand Binding
Ligand binding is an important function that alters receptor action and cellular responses. Modulation of this function by PTM might allow for integration of cellular signals with the estrogenic responses. To determine the effect of phosphorylation on ligand binding, equilibrium-based fluorescence polarization assays were carried out.
We first established that ER phosphorylation did not affect the binding of the tracer fluorescent ligand Fluormone, on which this fluorescence polarization assay is based. Increasing concentrations of control or phosphorylated receptor were incubated with Fluormone, and equivalent fluorescence polarization curves were obtained (data not shown). Because phosphorylation did not alter the binding of Fluormone to the receptor under the experimental conditions, we could use it as a tracer in competition binding assays to examine whether ER phosphorylation altered the binding of the physiologically important natural ligand, E2, and TOT, an antiestrogen used for breast cancer treatment.
Ligand displacement assays were carried out by incubating 1 nM Fluormone with 15 nM purified full-length ER (control and phosphorylated, Fig. 3A) and titrating in increasing concentrations of E2. Polarization of the fluorescein emission was measured at 525 nm and was plotted against log of E2 concentrations. Interestingly, phosphorylation status of the ER influences binding curves and IC50 values for E2. The E2 binding curve for Src (IC50, 2.5 nM) and PKA-phosphorylated receptor (IC50 2.5 nM) were shifted to the left, with a decrease in EC50 value for E2 binding as compared with the control, unphosphorylated ER (IC50, 18 nM), indicating that these kinases increase the binding of E2. MAPK- and AKT-phosphorylated receptor did not show any shift in binding curve as compared with the control, consistent with no change in E2 binding.
|
), a consistent right shift in binding curves was noted, with an increase in EC50 for MAPK (IC50, 150 nM) and Src (IC50, 76 nM) phosphorylated ER as compared with the control, unphosphorylated ER (IC50, 37 nM). Thus, the action of these kinases reduces TOT binding. Receptor phosphorylation by AKT and PKA did not alter the TOT binding profile.
Figure 3C summarizes the ligand binding data obtained from these experiments as IC50 values of the control and the phosphorylated receptor for E2 and TOT, and error bars are shown. Where kinase treatment changes ligand binding affinity, E2 binding is found to increase and TOT binding to decrease. Specifically, E2 binding is increased as a result of Src and PKA phosphorylation, whereas TOT binding is decreased by MAPK and Src treatment. The other ligand-kinase combinations showed no effect of phosphorylation on ligand binding affinity. Thus, phosphorylation events on the receptor at the A/B, C, and the E domains can differentially modulate affinity of E2 and TOT, important regulators of ER function in breast cancer progression and treatment.
DNA Binding Is Modulated by Receptor Phosphorylation
Although ER has a modular structure, signals can affect ER globally by alterations in protein conformation. As we have shown that phosphorylation of kinase sites in the A/B, C, and the E domains can alter ligand binding in the E domain, it is possible that these phosphorylation events might also modulate DNA binding, a function critical for the genomic regulatory activity of the receptor. Therefore, we examined the effect of phosphorylation on interaction of ER with a consensus estrogen response element (ERE) using fluorescence polarization assays.
Fluorescein-labeled oligonucleotides containing the vitellogenin A2 ERE (1 nM) were combined with increasing concentrations of control ER or phosphorylated ER, in the absence of ligand or in the presence of E2 or TOT, and fluorescence polarization was measured to calculate anisotropy. In the absence of ligand (apo-receptor), only the MAPK-phosphorylated ER resulted in a right-shifted curve (EC50, 140 nM) as compared with the control (EC50, 34 nM; Fig. 4A). ER phosphorylation by AKT, Src, and PKA did not result in a significant change in DNA binding for the apo-receptor. In the presence of E2 (Fig. 4B), Src phosphorylation of ER resulted in right-shifted curve, increasing the EC50 (97 nM) as compared with control (EC50, 42 nM). Interestingly, AKT phosphorylation of ER decreased the EC50 (14 nM) for binding to the ERE. The EC50 values for receptor phosphorylated by MAPK and PKA, however, did not vary from the control, although the binding curves showed lower maximum values.
|
). In the presence of TOT, MAPK- and Src-phosphorylated receptors demonstrated right-shifted curves; the EC50 for Src was 82 nM, but the curve for MAPK was so right shifted that the EC50 could only be estimated (
490 nM). Further, AKT phosphorylation of ER resulted in a small but significant decrease in EC50 (12 nM), but no change in DNA binding for the PKA-phosphorylated receptor was observed.
The effects of receptor phosphorylation of E2 or TOT-occupied ER on binding to the ERE are summarized as EC50 values in Fig. 4D, where error bounds are also shown. The control receptor had very similar affinities to the ERE in the absence of ligand or in the presence of E2 and TOT. The most noticeable changes are the modulation of ligand effects by MAPK phosphorylation. The MAPK-phosphorylated unliganded and TOT-bound receptors both demonstrated a dramatic decrease in affinity for the ERE, whereas the receptor affinity for the ER, when bound to E2, as shown in Fig. 4B, remains unaltered. Furthermore, when bound to either E2 or TOT, phosphorylation of the receptor by AKT at S167 showed a small but statistically significant increase in ERE binding affinity, whereas Src-phosphorylated ER exhibits a small but a significant decrease in affinity. Surprisingly, ER phosphorylation by PKA in the DNA binding domain did not alter affinity of the receptor for A2 ERE in the absence of ligand or in the presence of either E2 and TOT.
Interestingly, in some cases the different receptor-phosphorylated states exhibited distinct maximum anisotropy, with the highest values for control, MAPK-, and Src-phosphorylated receptor, followed by AKT-phosphorylated receptor, with the lowest anisotropy observed with PKA-phosphorylated receptor (Fig. 4, A–C). These differences are present with apo-receptor (Fig. 4A) but are more pronounced in the presence of E2 (Fig. 4B) and TOT (Fig. 4C). These changes suggest that phosphorylation alters not only the affinity of receptor-DNA interaction, but alters the conformation or apparent size of the ER-DNA complex.
ER Phosphorylation Modulates Coactivator Binding
In addition to ligand and DNA binding, coactivator recruitment by the ER is an important step in formation of a preinitiation complex and gene-regulatory effects in response to E2. To examine the effect of receptor phosphorylation on interaction of coactivator, we used a sensitive, homogeneous time-resolved fluorescence resonance energy transfer assay (TR-FRET), shown schematically in Fig. 5.
|
). With this assay, we were able to demonstrate coactivator recruitment only with E2-liganded ER; regardless of phosphorylation state, apo- and TOT-liganded ER did not recruit the SRC3 RID.
|
) , suggesting probable differences in the stoichiometry or conformation of complexes formed upon receptor phosphorylation, as we proposed to explain Bmax differences in some of the DNA binding experiments (see above and Discussion). Importantly, the coactivator binding curves for MAPK-, AKT-, and PKA-phosphorylated receptor were left shifted, with a corresponding decrease in EC50 values as compared with the unphosphorylated ER control.
Selective modulation of the coactivator binding function of ER by MAPK, AKT, and Src phosphorylation is evident from the EC50 data plotted in Fig. 6B. The control, unphosphorylated receptor bound to SRC3 RID with an EC50 of 15 nM; the MAPK-, AKT-, and PKA-phosphorylated receptors bound to the SRC3 RID with a statistically significant increase in affinity, reflected in EC50 values of 3 nM, 6 nM, and 1 nM, respectively. There was no difference in the binding of the SRC3 RID to the control receptor and the Src-phosphorylated receptor. As noted above, the maximum FRET signal observed with the Src-phosphorylated receptor, however, was much higher than the control, indicating differences in complex formation despite similar affinities. Thus, ER phosphorylation affects coactivator recruitment, as well as ligand and DNA binding.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
In this study, we have examined the state of ER phosphorylation in parental, tamoxifen-sensitive, and in tamoxifen-resistant MCF-7 human breast cancer cells, and we developed convenient fluorescence-based assay systems through which we could quantify how site-specific phosphorylation of ER affects ligand binding, DNA binding, and coactivator recruitment. We observed differences in endogenous ER phosphorylation in the tamoxifen-sensitive and -resistant cells, consonant with prior work (22); we also found that specific kinase-mediated phosphorylations of ER result in distinctive changes in ER ligand, DNA, and coactivator interactions. Globally, these changes would be expected to result in enhanced activity of E2 and decreased effectiveness of TOT as an antagonist.
Phosphorylation Status of Endogenous ER from MCF-7 Cells
Our analysis of the phosphorylation status of endogenous ER reveals ligand-dependent and ligand-specific phosphorylation at both S118 and S167 in parental MCF-7 cells. Although E2- and TOT-induced phosphorylation at both S118 and S167 have been observed individually (9, 27, 28), differences in phosphorylation patterns have not been compared. The observed differences in phosphorylation levels with E2 and TOT treatment suggest that unique signaling pathways activated by E2 and TOT could underlie ligand-dependent differences in cross talk between these pathways and ligand-induced changes in ER conformation that affect ER activity as a function of its distinctive protein kinase-altered phosphorylation state.
A number of investigators have noted ligand-independent phosphorylation of ER at both S118 and S167 in response to peptide growth factors, including epidermal growth factor, IGF, and heregulin (14, 22, 29, 30), and we show that phosphorylation at S118 and S167 in the tamoxifen-resistant MCF-7 Her2 cells was similar in the absence of ligand, or in the presence of E2 or TOT. This loss of ligand-dependent phosphorylation of ER in these cells might be due to up-regulated MAPK and PI3/AKT activity associated with the overexpressed HER2 signaling in these cells (7, 22). These HER2 overexpressing cells have previously been demonstrated to proliferate in response to TOT, and in these cells ER-TOT recruits SRC3 to the pS2 promoter, as documented by chromatin immunoprecipitation assays (22). Alternately, the differences in ER phosphorylation patterns in these cell lines might be attributed to the activity of PP5, a serine/threonine protein phosphatase that has been demonstrated to regulate serine 118 ER phosphorylation (31).
The phosphorylation status of ER in clinical breast tumors and its association with up-regulation of growth factor receptors have been under intense investigation to determine whether ER phosphorylation correlates with drug therapy resistance and might suggest approaches to more effective therapy management (14, 16, 22, 32, 33, 34). To investigate further the relationship between phosphorylation status and receptor function—and ultimately to better understand drug resistance—we examined how these individual phosphorylation events alter basic ER functions, including ligand, DNA, and coactivator binding. The results of our study are summarized in Fig. 7.
|
). Notably, in this regard, activation of MAPK has been proposed to be involved in hypersensitization of long-term estrogen-deprived MCF-7 cells to E2 (36). In addition to increased coactivator binding affinity, we observe a higher Bmax with the MAPK-phosphorylated ER-E2 complex compared with the control complex in this FRET-based assay (Fig. 6A). This could result from an increase in the number of SRC3 RID complexes recruited to the DNA-bound phosphorylated ER, or to a change in complex conformation that reduces the distance between the acceptor and the donor. Importantly, in the context of TOT action, we found that phosphorylation of ER at S118 results in decreased ligand binding and reduced binding of the ER-TOT complex to DNA. This decrease in affinity for TOT might be playing a role in tamoxifen resistance. Although, in our system, which uses coactivator RID, we were unable to detect coactivator binding to the phosphorylated receptor in the presence of TOT, phosphorylation of S118 might allow interaction with the endogenous full-length coactivator even in the presence of TOT, through the generation of an additional coactivator interaction site at the phosphorylated S118 in the N-terminal region of ER (37). It is difficult, however, to obtain this large coactivator protein in quantities sufficient for these experiments. The phosphorylation status of the coactivator is also known to modulate receptor-coactivator interactions (4), leading to coactivator recruitment by ER in the presence of TOT, as Schiff and co-workers (22) observed by chromatin immunoprecipitation assays in MCF-7 HER2/neu cells.
Phosphorylation of ER by AKT
Phosphorylation at S167 has also been reported to be E2 dependent or ligand independent (5, 11, 12) and to play a role in E2-dependent transcription, apoptosis, and proliferation (5, 38, 39). Through our in vitro studies, we show that ER phosphorylation at S167 does not alter the affinity of the receptor for the ligands E2 or TOT, but results in an increase in affinity of the phosphorylated ER for the DNA with either ligand (Fig. 7). Furthermore, this phosphorylation increased affinity of ER for SRC3 RID, but only in the presence of E2 (Fig. 7). Thus, ER phosphorylation at S167 might increase E2 agonism by enhancing binding of the E2-bound ER to the DNA and increasing coactivator recruitment. Although AKT phosphorylation might also increase interaction of ER-TOT with DNA, whether this would lead to increased TOT antagonism or agonism would depend on the coactivator status of the cell.
ER Phosphorylation by Src Kinase
Some groups have reported ER phosphorylation on tyrosine residues (40, 41, 42, 43), but other groups have been unable to observe this (5, 9). These discrepancies might be due to the transient nature of tyrosine phosphorylation under the various conditions studied.
We observed specific phosphorylation of ER LBD at Y537 by Src kinase in vitro that did not occur with the mutant Y537S ER LBD (our unpublished data). Functional analysis of Src-phosphorylated receptor demonstrates an increase in affinity of the receptor for E2 but a decrease in affinity for TOT (Fig. 7). A similar increase in affinity for E2 was also observed by other investigators upon receptor phosphorylation by Src (40, 44). Furthermore, we observed that Src-phosphorylated receptor exhibited a decrease in affinity for DNA in the presence of E2 and TOT (Fig. 7), although others, using gel shift assays, observed an increase in DNA binding (40). Although Src has not been demonstrated to phosphorylate ER in the DNA-binding domain, functional modulation of distal receptor regions by ligand and DNA binding have been documented, accenting the fluidity of the receptor in transmitting signal across the receptor domains (45, 46). In coactivator recruitment assays, Src phosphorylation of the receptor did not alter interaction of ER-E2 with SRC3 RID. It is possible that the transient Src phosphorylation of the receptor plays a role in sensitizing ER to low concentrations of E2 and might be involved in priming ER for nongenomic actions, as well as in hypersensitization of MCF-7 cells to estrogen depletion during treatment with tamoxifen or aromatase inhibitors.
Phosphorylation by PKA
Prior studies have demonstrated that PKA phosphorylates ER, with S236 and S305 implicated as phosphorylation sites (9); phosphorylation at S236 is reported to alter receptor dimerization and DNA binding (47), whereas phosphorylation at S305 by PKA has been implicated in tamoxifen resistance (18). We find that ER phosphorylation by PKA increases affinity of the receptor for E2 without altering affinity for TOT, again suggesting cross talk between receptor domains. However, we did not observe any alterations in DNA binding with either ligand when ER was phosphorylated by PKA (Fig. 7). S236 is in the DNA-binding domain, and phosphorylation at this site has been previously shown to regulate ER dimerization and DNA binding (47). These observations, however, were made with a very different assay system that used ER mutants and cell extracts. It is highly likely that phosphorylations at these sites enhanced interactions with protein partners that modulate ER dimerization or DNA interactions.
We also find that ER phosphorylated by PKA has an increase in affinity for SRC3 RID in the presence of E2; however, we could observe no interaction of the receptor with the SRC3 RID in the presence of TOT (Fig. 7). Michalides et al. had previously shown, in U2OS cells, that overexpressed SRC1 interacted with ER phosphorylated by PKA in the presence of TOT (18). Again, these differences might be due to differences in the assay systems: our coactivator interaction assay involved use of in vitro phosphorylated, full-length ER, DNA (A2 ERE) and SRC3 RID, as opposed to the cell-based assays of Michalides, which detected interaction of full-length coactivator with ER, but did not necessarily involve DNA-bound receptor.
Implications of Phosphorylation Status to ER Signaling, Cross Talk with Growth Factors, and Endocrine Drug Resistance
In this study, we have examined the results of ER modification by individual kinases in isolation, whereas under physiological conditions, it is very likely that the receptor is phosphorylated by multiple kinases, depending on the cell stimuli. Endogenous ER is thus likely to have multiple modifications that tune its actions in response to cell stimuli. Currently, a number of studies are directed at the analysis of breast cancer biopsies for S118 and S167 phosphorylation, but the lack of specific antibodies for S236, S305 and Y537 make histochemical analysis of modification at these sites difficult (14, 15, 16). A proteomic approach using mass spectroscopy might enable the determination of global status of ER phosphorylation and might serve as a useful analytical tool to eventually help in making the medical decisions needed to select the most effective drug therapy for an individual breast cancer patient.
In spite of drug resistance to tamoxifen and aromatase inhibitors, ER still plays a crucial role in the growth of the majority of the breast tumors. Recent research indicates adaptations by breast cancer cells that involve the hijacking of signaling pathways eventually succeed in overcoming tamoxifen antagonism or estrogen deprivation (26, 48, 49). These signaling pathways are thought to alter ER function through receptor phosphorylation at unique sites, as well as PTM of ER coregulatory proteins.
In this study, we have observed a shift from ligand-dependent ER phosphorylation in parental MCF-7 cells to a ligand-independent receptor phosphorylation in the HER2 over expressing, tamoxifen-resistant, MCF-7 HER2/neu cells. Additionally, we find that receptor function is modulated by phosphorylation in ways that could increase E2 agonism and decrease TOT action. Thus, receptor phosphorylation could be overcoming drug resistance through two different mechanisms: 1) increased affinity of phosphorylated ER for E2 and coactivators that would sensitize tumor cells to low E2 concentrations, or 2) decreased affinity of the phosphorylated ER for TOT and of the TOT-ER complex for the DNA. Our study thus indicates that each receptor phosphorylation can alter receptor function in a distinctive manner, either enhancing E2 agonism or decreasing TOT action, and that these alterations might be used by breast cancer cells individually or in combination to overcome drug resistance.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
were prepared. Site-directed mutagenesis was performed on pCMV5-Flag-ER using the QuikChange kit (Stratagene) according to the manufacturer’s instructions. The oligonucleotides used (mutational site underlined) were: S118A: 5'-CCG CAG CTG GCG CCT TTC CTG C; S167A: 5'-C AGA GAA AGA TTG GCC GCT ACC AAT GAC AAG GG; S236A: 5'-GAT AAA AAC AGG AGG AAG GCC TGC CAG GCC TGC; S305A: 5'-C ATG ATC AAA CGC TCT AAG AAG AAC GCC CTG GCC TTG TCC CTG AC and Y537A: 5'-G AACGTG GTG CCC CTC GCT GAC CTG CTG CTG GA. All plasmids were sequenced to confirm introduction of the desired mutation.
Cell Culture and Preparation of Nuclear Proteins
MCF-7 and MCF-7 Her2/neu human breast cancer cells were cultured as previously described (3, 50). Four days before ligand treatment, cells were switched to treatment media, phenol red-free MEM (Sigma, St. Louis, MO) containing 5% charcoal dextran-treated calf serum for MCF-7 cells and improved MEM containing 10% charcoal dextran-treated fetal bovine serum for MCF-7 Her2/neu cells. Media were changed on d 2 and d 4 of culture. Cells were harvested, resuspended in the respective treatment media, and treated with vehicle (N,N-dimethylformamide), 1 µM E2, or TOT for 15 min at 37 C, and then cells were collected by centrifugation at 1000 rpm for 5 min. Nuclear extracts were prepared as mentioned in Ref. 46 , and protein concentration was determined using Bio-Rad’s protein assay reagents (Bio-Rad Laboratories, Hercules, CA).
Protein Expression
Full-length Flag-tagged ER was expressed in SF9 cells and purified as described by Kraus and Kadonaga (51). For TR-FRET, SRC3 RID (amino acid residues 627–829) was expressed, purified, and labeled with Cy5 as described (52). For analysis of phosphorylation specificity, the FLAG-tagged and mutant ERs were expressed in HEK293T cells that were maintained in DMEM (Life Technologies, Inc., Gaithersburg, MD) supplemented with 10% heat-inactivated calf serum (Hyclone Laboratories, Logan, UT) and 1% antibiotics. Before transient transfections, the cells were maintained in phenol red-free DMEM containing 5% charcoal-stripped fetal bovine serum for a minimum of 4 d with the medium being changed every other day. The cells were transfected in six-well plates using Lipofectamine 2000 reagent (Invitrogen, San Diego, CA) and 1.7 µg/well of pCMV5-Flag-ER of wild-type ER or the various ER mutants according to the manufacturer’s instructions. Cells were lysed in lysis buffer containing, 50 mM Tris (pH 7.5), 150 mM NaCl, 0.05% Nonidet P-40, 2 mM dithiothreitol, and protease inhibitors (Upstate Cell Signaling Solutions, Charlottesville, VA) on ice and the cell lysate was collected. The FLAG-tagged receptor was then purified as described by Kraus and Kadonaga (51).
In Vitro Kinase Assays
Activated kinases including MAPK, AKT, Src, and PKA (Upstate Cell Signaling Solutions) were used to phosphorylate ER. In vitro kinase assays were carried out in kinase buffer (50 mM Tris, pH 7.5; 50 mM MgCl2; 100 µM ATP; 1 mM dithiothreitol) and ER in the presence of either kinase or equivalent amounts of 10% glycerol as control. After incubation at 30 C for 30 min, ER was either mixed with sample buffer for Western immunoblot analysis, or diluted immediately in appropriate assay buffers (pH 8) and used for functional assays. Quantification of kinase-specific phosphorylation was carried out using P81 paper as mentioned in Upstate quantification protocol and then counted using a counter. The disintegration per min counts were then used to calculate picomoles of [32P]phosphate, and the results were then expressed as (picomoles phosphate/picomoles ER) x 100.
Western Immunoblotting
To determine phosphorylation status of endogenous ER under hormonal stimulation, 100 µg of nuclear protein from parental MCF-7 and tamoxifen-resistant MCF-7 Ner2/neu cells treated with vehicle, E2, or TOT were mixed in 1x sodium dodecyl sulfate sample buffer, fractionated on SDS-PAGE, and transferred to nylon membrane. Western immunoblot analysis was then performed using antibodies against phosphoserine 118 ER or phosphoserine 167 ER (Cell Signaling Technology, Inc., Beverly, MA) to determine phosphorylation at these two sites. To determine the total amount of ER in the nuclear proteins loaded on the gel, an ER-specific antibody D12 (Sc8005, Santa Cruz Biotechnology, Inc., Santa Cruz, CA) was used. Secondary antibodies (antimouse and antirabbit) used for immunobloting were purchased from Cell Signaling Technology, and West Femto horseradish peroxidase substrate was from Pierce Chemical Co. (Rockford, IL).
To determine phosphorylation of ER in the in vitro kinase assays, control unphosphorylated receptor and ER phosphorylated with MAPK, AKT, Src, or PKA were run on SDS-PAGE, transferred to nylon membranes, and probed with antibodies against phosphoserine 118 ER, phosphoserine 167 ER, phosphoserine for PKA substrate (Cell Signaling Technology, Inc.), or phosphotyrosine Src substrate (Upstate Cell Signaling Solutions). The secondary antibodies used were antirabbit (antiphosphotyrosine Src-substrate) and antimouse (antiphosphoserine Src substrate) horseradish peroxidase conjugates (Cell Signaling Technology, Inc.).
Phosphorylation of wild-type and mutant ER expressed in HEK293T cells was monitored after fractionating the purified receptor and analyzing by Western immunoblot as described above.
Fluorescence Polarization Assays
Ligand displacement assays were carried out using Fluormone (PanVera, Invitrogen, Carlsbad, CA), as mentioned in the assay directions. Fluormone is a conjugate of E2 and fluorescein (undisclosed structure) that retains good binding affinity for ER and ERß. Briefly, 1 nM Fluormone was combined with 15 nM purified full-length baculovirus-expressed Flag-tagged ER, and increasing concentrations of E2 or TOT were titrated and incubated for 1 h at 27 C in 96-well high-efficiency plates (Molecular Devices, Sunnyvale, CA). Fluorescence polarization for fluorescein was measured at 545 nm using Wallac Victor II plate reading fluorometer (PerkinElmer, Wellesley, MA), and data were plotted as log of E2 concentration on the x-axis and polarization values on the y-axis using PRISM 4.0 (Graph Pad Software, San Diego, CA). The IC50 values were calculated and statistically analyzed using Student’s t test.
DNA polarization assays were carried out using 1 nM fluorescein-labeled A2 ERE (Integrated DNA Technologies, Coralville, IA) and increasing concentrations of control or phosphorylated ER in DNA binding buffer (20 mM HEPES, 100 mM K glutamate, 0.3 mg/ml of ovalbumin, and 2 mM ß-mercaptoethanol). DNA polarization assays were carried out in the absence of ligand or in the presence of E2 or TOT, and polarization was measured as described. Anisotropy values were calculated (53), and the data were plotted as log ER concentration to anisotropy values obtained. The EC50 values were determined and statistically analyzed by Student’s t test to determine phosphorylation-induced differences.
TR-FRET Assay
TR-FRET assays were carried out using 10 nM biotinylated A2 ERE (Integrated DNA Technologies), 2.5 nM europium-labeled streptavidin (PerkinElmer), 20 nM full-length control or phosphorylated ER, and increasing concentrations of Cy5-labeled SRC3 RID in the presence of 100 nM E2 in TGNK buffer (50 mM Tris, pH 8; 10% glycerol; 0.05% Nonidet P-40; and 50 mM KCl) containing 0.3 mg/ml of ovalbumin and 2 mM ß-mercaptoethanol. The components were combined and incubated at 27 C for 1 h, and TR-FRET was measured after 50 µsec delay for donor (Europium emission, 615 nm) and acceptor (Cy5 emission, 670 nm). The TR-FRET was expressed as ratio of acceptor emission to donor emission, and data were plotted with (acceptor/donor) x 1000 on the y-axis against log of ER concentration on the x-axis. The EC50 values were determined and then statistically analyzed.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
First Published Online August 31, 2006
Abbreviations: E2, Estradiol; ER, estrogen receptor; ERE, estrogen response element; FRET, fluorescence resonance energy transfer; HEK, human embryonic kidney; LBD, ligand-binding domain; PKA, protein kinase A; PTM, posttranslational modification; RID, receptor interacting domain; SRC3, steroid receptor coactivator 3; TOT, trans-hydroxytamoxifen; TR-FRET, time-resolved FRET.
Received for publication February 9, 2006. Accepted for publication August 22, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
and estrogen receptor ß: regulation by selective estrogen receptor modulators and importance in breast cancer. Breast Cancer Res 2:335–344[CrossRef][Medline] transcriptional activation through phosphorylation of serines 104 and 106 by the cyclin A-CDK2 complex. J Biol Chem 274:22296–22302 after PKA activation in breast cancer. Cancer Cell 5:597–605[CrossRef][Medline]. Nucl Recept Signal 4:e007 by S118 phosphorylation involves a ligand-dependent interaction with TFIIH and participation of CDK7. Mol Cell 6:127–137[CrossRef][Medline] at serine 118 by two distinct signal transduction pathways revealed by phosphorylation-specific antisera. Oncogene 21:4921–4931[CrossRef][Medline] expression is associated with better disease outcome in women treated with tamoxifen. Clin Cancer Res 10:5902–5906 serine 167 is predictive of response to endocrine therapy and increases postrelapse survival in metastatic breast cancer. Breast Cancer Res 7:R753–R764 and is involved in the regulation of receptor activity. J Biol Chem 280:33006–33014 transactivation function 1 (AF-1) and AF-2 mediated by steroid receptor coactivator protein-1: requirement for the AF-1 -helical core and for a direct interaction between the N- and C-terminal domains. Mol Endocrinol 15:1953–1970: a new model for anti-estrogen resistance. J Biol Chem 276:9817–9824 by protein kinase A regulates dimerization. Mol Cell Biol 19:1002–1015 and EGFR signalling pathways regulates tamoxifen-resistant growth. Breast Cancer Res Treat 96:131–146[CrossRef][Medline]NURSA Molecule Pages Link:
This article has been cited by other articles:
 |
|
 |
|
  H. Pick, S. Etter, O. Baud, R. Schmauder, L. Bordoli, T. Schwede, and H. Vogel Dual Activities of Odorants on Olfactory and Nuclear Hormone Receptors J. Biol. Chem., October 30, 2009; 284(44): 30547 - 30555. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Bhat-Nakshatri, G. Wang, H. Appaiah, N. Luktuke, J. S. Carroll, T. R. Geistlinger, M. Brown, S. Badve, Y. Liu, and H. Nakshatri AKT Alters Genome-Wide Estrogen Receptor {alpha} Binding and Impacts Estrogen Signaling in Breast Cancer Mol. Cell. Biol., December 15, 2008; 28(24): 7487 - 7503. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Kharode, P. V. N. Bodine, C. P. Miller, C. R. Lyttle, and B. S. Komm The Pairing of a Selective Estrogen Receptor Modulator, Bazedoxifene, with Conjugated Estrogens as a New Paradigm for the Treatment of Menopausal Symptoms and Osteoporosis Prevention Endocrinology, December 1, 2008; 149(12): 6084 - 6091. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Narayanan, C. C. Coss, M. Yepuru, J. D. Kearbey, D. D. Miller, and J. T. Dalton Steroidal Androgens and Nonsteroidal, Tissue-Selective Androgen Receptor Modulator, S-22, Regulate Androgen Receptor Function through Distinct Genomic and Nongenomic Signaling Pathways Mol. Endocrinol., November 1, 2008; 22(11): 2448 - 2465. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Yamashita, M. Nishio, T. Toyama, H. Sugiura, N. Kondo, S. Kobayashi, Y. Fujii, and H. Iwase Low phosphorylation of estrogen receptor {alpha} (ER{alpha}) serine 118 and high phosphorylation of ER{alpha} serine 167 improve survival in ER-positive breast cancer Endocr. Relat. Cancer, September 1, 2008; 15(3): 755 - 763. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. S. Fenne, T. Hoang, M. Hauglid, J. V. Sagen, E. A. Lien, and G. Mellgren Recruitment of Coactivator Glucocorticoid Receptor Interacting Protein 1 to an Estrogen Receptor Transcription Complex Is Regulated by the 3',5'-Cyclic Adenosine 5'-Monophosphate-Dependent Protein Kinase Endocrinology, September 1, 2008; 149(9): 4336 - 4345. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Chen, T. Dang, R. D. Blind, Z. Wang, C. N. Cavasotto, A. B. Hittelman, I. Rogatsky, S. K. Logan, and M. J. Garabedian Glucocorticoid Receptor Phosphorylation Differentially Affects Target Gene Expression Mol. Endocrinol., August 1, 2008; 22(8): 1754 - 1766. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Paulmurugan, A. Tamrazi, J. A. Katzenellenbogen, B. S. Katzenellenbogen, and S. S. Gambhir A Human Estrogen Receptor (ER){alpha} Mutation with Differential Responsiveness to Nonsteroidal Ligands: Novel Approaches for Studying Mechanism of ER Action Mol. Endocrinol., July 1, 2008; 22(7): 1552 - 1564. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X. Long, M. Fan, R. M. Bigsby, and K. P. Nephew Apigenin inhibits antiestrogen-resistant breast cancer cell growth through estrogen receptor-{alpha}-dependent and estrogen receptor-{alpha}-independent mechanisms Mol. Cancer Ther., July 1, 2008; 7(7): 2096 - 2108. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. Hammoud, B. Tan, S. Badve, and R. M Bigsby Estrogen promotes tumor progression in a genetically defined mouse model of lung adenocarcinoma Endocr. Relat. Cancer, June 1, 2008; 15(2): 475 - 483. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Tan, X. Long, H. Nakshatri, K. P Nephew, and R. M Bigsby Striatin-3{gamma} inhibits estrogen receptor activity by recruiting a protein phosphatase J. Mol. Endocrinol., May 1, 2008; 40(5): 199 - 210. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Andre, R. Nahta, R. Conforti, T. Boulet, M. Aziz, L. X. H. Yuan, F. Meslin, M. Spielmann, G. Tomasic, L. Pusztai, et al. Expression patterns and predictive value of phosphorylated AKT in early-stage breast cancer Ann. Onc., February 1, 2008; 19(2): 315 - 320. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. H. Ing, D. A. Massuto, and L. A. Jaeger Estradiol Up-regulates AUF1p45 Binding to Stabilizing Regions within the 3'-Untranslated Region of Estrogen Receptor {alpha} mRNA J. Biol. Chem., January 18, 2008; 283(3): 1764 - 1772. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. De los Santos, O. Martinez-Iglesias, and A. Aranda Anti-estrogenic actions of histone deacetylase inhibitors in MCF-7 breast cancer cells Endocr. Relat. Cancer, December 1, 2007; 14(4): 1021 - 1028. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. L. Weigel and N. L. Moore Steroid Receptor Phosphorylation: A Key Modulator of Multiple Receptor Functions Mol. Endocrinol., October 1, 2007; 21(10): 2311 - 2319. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Grisouard, S. Medunjanin, A. Hermani, A. Shukla, and D. Mayer Glycogen Synthase Kinase-3 Protects Estrogen Receptor {alpha} from Proteasomal Degradation and Is Required for Full Transcriptional Activity of the Receptor Mol. Endocrinol., October 1, 2007; 21(10): 2427 - 2439. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. R. Schultz-Norton, K. A. Walt, Y. S. Ziegler, I. X. McLeod, J. R. Yates, L. T. Raetzman, and A. M. Nardulli The Deoxyribonucleic Acid Repair Protein Flap Endonuclease-1 Modulates Estrogen-Responsive Gene Expression Mol. Endocrinol., July 1, 2007; 21(7): 1569 - 1580. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
indicates increase in affinity; 
a decrease in affinity; 