| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Interacts with G13 to Drive Actin Remodeling and Endothelial Cell Migration via the RhoA/Rho Kinase/Moesin Pathway
Molecular and Cellular Gynecological Endocrinology Laboratory (T.S., C.S., P.M., M.S.G., X.-D.F., C.B., S.G., A.C., L.F., A.R.G.), Department of Reproductive Medicine and Child Development, University of Pisa, 56100 Pisa, Italy; Department of Obstetrics and Gynecology (F.N.), New York University, New York, New York 10010; and Department of Obstetrics, Gynecology and Reproductive Sciences (A.F.), Yale University, New Haven, Connecticut 06520
Address all correspondence and requests for reprints to: Tommaso Simoncini, M.D., Ph.D., Molecular and Cellular Gynecological Endocrinology Laboratory, Department of Reproductive Medicine and Child Development, Division of Obstetrics and Gynecology, University of Pisa, Via Roma, 57, 56100 Pisa, Italy. E-mail: t.simoncini{at}obgyn.med.unipi.it.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
with the G protein G13, leading to activation of the small GTPase RhoA and of the downstream effector Rho-associated kinase. The resulting phosphorylation of moesin on Thr558 is the means of moesin’s binding to actin and the remodeling of the actin cytoskeleton. This cascade of events ensues within minutes of estradiol administration and results in changes in cell morphology and to the development of specialized cell membrane structures such as ruffles and pseudopodia that are necessary for cell movement. These findings expand our knowledge of the basis of estrogen’s effects on human cells, including the regulation of actin assembly, cell movement and migration. They highlight novel pathways of signal transduction of estrogen receptor through nontranscriptional mechanisms. Furthermore, exposure of this estrogen receptor-dependent, nongenomic action of estrogen on human vascular endothelial cells is especially relevant to the present interest in the role of estrogen in cardiovascular protection. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
The cardiovascular actions of estrogens are prominent, but incompletely understood (3, 4). The presence of estrogen receptors (ERs) and of enzymes involved in estrogen synthesis in human blood vessels and endothelium has been known for some time (5). A number of studies show that estrogens enhance endothelial proliferation at sites of vascular injury, promoting endothelial recovery after structural damage (6). This has been related to the activation of both long-term, genomic actions as well rapid, presumably nongenomic signaling by estrogens (3, 4, 7), but no mechanistic explanation has been presented. Estradiol has been shown to rapidly affect membrane traffic (8) and observations on breast cancer cells suggest that estrogen exposure leads to ER membrane translocation and to the formation of membrane ruffles and pseudopodia (9), suggesting that exposure to estrogen may result in rapid modifications of the cytoskeleton’s organization and to modified interaction with the surrounding environment. This was further indicated by our findings of the induction of expression of ezrin, an actin-binding protein, in ovarian cancer cells within 3 h of estradiol administration (2).
The actin cytoskeleton forms the backbone of the cell, and its spatial organization is crucial for cell movement and migration. Modification of the form and positioning of actin fibers and their relationship with membrane anchoring structures such as integrins and focal adhesion complexes (10) allows cell movement in the extracellular environment (11). The ezrin/radixin/moesin (ERM) family of regulatory proteins is one of the better characterized families of actin-binding proteins. By interacting with actin, activated ERMs induce actin depolymerization and reassembly toward the cell membrane edge, being responsible for the formation of cortical actin complexes. These complexes sustain the formation of molecular bridges between actin, integrins, and focal adhesion complexes at specialized cell membrane sites such as ruffles and pseudopodia (12, 13). Within this family of regulatory proteins, moesin is prominently expressed in endothelial cells (14). Moesin activation is achieved through phosphorylation on Thr558 by the Rho-associated kinase (ROCK), which sequentially results in a conformational change and in the association with the scaffold protein, ERM-binding protein 50 (EBP50) on moesin’s NH2-terminal end and with F-actin on moesin’s COOH-terminal end to mediate the linkage of microfilaments to membranes in cell surface microvilli.
We have studied nongenomic effects of estrogens on endothelial cell actin rearrangement and endothelial cell movement. This required full characterization of the signaling events from the recruitment of ERs to the activation of moesin. Our studies revealed the activation by ER of G proteins at the cell membrane and the activation of both the small GTPase RhoA and the ROCK, ROCK-2.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
). Vinculin is diffusely distributed throughout the cytoplasm (Fig. 1A). In the presence of a single pulse of unconjugated estradiol, actin and vinculin fibers rapidly change their spatial organization. There is loss of stress fibers and progressive localization of actin and vinculin toward the edge of the cell membrane, where cortical actin complexes and vinculin aggregates are formed (Fig. 1A). This behavior is rapid, time dependent, and transient. It is maximal between 5 and 15 min and after 30–60 min shows a progressive reversal to the pretreatment arrangement (Fig. 1A). There is a parallel time course for the formation of specialized cell membrane structures, such as focal adhesion complexes, pseudopodia, and membrane ruffles (Fig. 1A).
|
). The total cell content of immunoreactive (ir)-moesin does not change during this period, but an increase in ir-moesin protein expression is noticed later on, after 10 h of estradiol treatment, suggesting that moesin may also be regulated at the transcriptional or posttranscriptional level by estradiol, as is ezrin (2) (Fig. 1B). Moesin activation by estradiol is linear through a wide range of doses, including physiological concentrations (Fig. 1C).
To assess whether moesin is required for estrogen-dependent actin remodeling, we silenced moesin signaling by transfection of specific antisense phosphorothioate oligonucleotides (PONs), resulting in near disappearance of ir-moesin in endothelial cells (see Fig. A, published as supplemental data on The Endocrine Society’s Journals Online web site at http://mend. endojournals.org; and Fig. 1D, central lower boxes). Moesin silencing abrogated estradiol-induced remodeling of actin fibers, cell membrane ruffling, and the formation of pseudopodia (Fig. 1D). As control, the transfection of sense (inactive) moesin PONs does not alter estrogen signaling to actin (Fig. 1D).
Estradiol Activates Moesin through a Cell-Membrane ER- and G Protein-Dependent Signaling Pathway
When moesin was assayed in endothelial cells alternatively exposed to estradiol or to preferential agonists to ER [4,4',4''-(4-propyl-[1H]-pyrazole-1,3,5-triyl)trisphenol (PPT); 1 nM] (17) or ERß [2,3-bis(4-hydroxyphenyl)-propionitrile (DPN); 1 nM] (18), moesin phosphorylation was found only in the presence of E2 or of the ER ligand, PPT (Fig. 2A). In contrast, DPN did not activate moesin even at higher concentrations, implying that under physiological conditions ERß is not involved in the activation of this signaling pathway (Fig. 1A). Because the ER selectivity of PPT and DPN is not absolute (at high concentrations they cross-bind to the other receptor), we confirmed this result abrogating ER in endothelial cells. Selective knockout of ER by transfection of small interfering RNAs (siRNAs) results in dramatically reduced phosphorylation of moesin upon estrogen administration (Fig. 2B), in the absence of modifications of wild-type moesin or ERß expression (Fig. 2B).
|
|
). On the other hand, blockade of ERs with the ER antagonist ICI 182,780 or of G proteins with pertussis toxin (PTX) completely abolishes estrogen-dependent moesin phosphorylation (Fig. 2C), implying that this phenomenon is induced through an ER/G protein pathway. As a collateral observation, ICI 182,780 and wortmannin were associated with minor but nonetheless apparent increases of moesin phosphorylation that may have different explanations, including possible weak estrogenic effects of ICI 182,780 and a possible basal inhibitory action of PI3K on moesin.
Moesin was similarly activated upon the exposure of endothelial cells to unconjugated estradiol or to estradiol conjugated to BSA (E2-BSA), a membrane-impermeable form of E2 (Fig. 2D). Indeed, in our cell culture conditions, E2-BSA is unable to activate a heterologous thymidine kinase promoter construct containing a tandem palindromic estrogen-response element (5'-GGTCANNNTGACC-3'), whereas E2 is fully effective, confirming the inability of E2-BSA to enter the cell membrane (see Fig. B, published as supplemental data). Moesin activation by E2-BSA is also prevented by the ER antagonist ICI 182,780 as well as by the G protein inhibitor, PTX (Fig. 2D), suggesting that endothelial moesin activation is achieved through the recruitment of a cell membrane ER/G protein mechanism.
By double immunostaining for wild-type moesin (green) and Thr558-P-moesin (red), we also showed that, as is the case for ezrin (2, 13) during exposure to estradiol, phosphorylated moesin increases in endothelial cells and clusters in cell membrane pseudopodia and ruffles (Fig. 2E). The activation and membrane localization of moesin are largely prevented by ICI 182,780 or PTX (Fig. 2E).
In support of all the above, the dynamic rearrangement of actin and vinculin fibers induced by estradiol is prevented by ICI 182,780 and by PTX, and are entirely reproduced by E2-BSA as well as by the ER agonist, PPT (Fig. 2F). The ERß-selective agonist, DPN, does not alter the arrangement of actin or vinculin fibers (Fig. 2F). Likewise, E2 does not modify the spatial localization of actin or vinculin in endothelial cells deprived of ER by transfection of specific siRNAs (Fig. 2F).
Intracellular Events Linking Activation of ER to Moesin Phosphorylation: No Role of Direct Interaction of ER with Moesin or EBP50
One of the better characterized means of regulation of signaling intermediates by ER is via direct protein-protein interaction. Through this mechanism, ER activates intracellular kinases such as phosphatidylinositol 3-OH kinase (19, 20). We therefore investigated whether there could be a role for direct interaction of ER with moesin or with the moesin-binding protein, EBP50. With coimmunoprecipitation experiments, we were not able to find protein-protein interactions of ER and either moesin, the active form of moesin, or EBP50 (Fig. 3A), not even at different time points (Fig. 3B). In parallel, in the presence of estradiol, EBP50 displays an enhanced interaction with the phosphorylated form of moesin, whereas only a negligible cointeraction with wild-type moesin is seen (Fig. 3C). This shows that after ER recruitment activated moesin and EBP50 come in contact, making actin rearrangement possible (16, 21).
|
), consistent with the concept that in human endothelial cells the EBP50 promoter is controlled by estrogen.
Intracellular Events Linking Activation of ER to Moesin Phosphorylation: Role of ROCK-2
Because ROCK-2 is responsible for Thr558 phosporylation of moesin and for the subsequent cytoskeletal rearrangements (16), we studied whether ROCK-2 is involved in moesin activation by estrogen. Indeed, rapid estradiol-dependent phosphorylation of moesin is prevented by the ROCK-2 inhibitor Y-27632 (Fig. 4A).
|
with ROCK-2 (Fig. 4B). The E2-induced interaction of ER with ROCK-2 is prevented in the presence of the pure ER antagonist, ICI 182,780 (Fig. 4B), indicating that this dynamic interaction is triggered by the occupancy by estradiol of the ER ligand binding pocket. On the contrary, non-cointeraction can be observed between ERß and ROCK-2 (Fig. 4C).
To understand the functional relevance of the ER/ROCK-2 interaction, we studied whether ER-associated ROCK-2 is functionally activated. To do so, we immunoprecipitated ER and used the immunoprecipitates (IPs) to perform kinase assays using dephosphorylated myelin basic protein (MBP) as the target for Thr phosphorylation. MBP phosphorylation was studied with an antibody specifically recognizing Thr-phosphorylated MBP. We found that ER-associated IPs obtained from cells exposed to estradiol actively sustain the Thr phosphorylation of MBP (Fig. 4D). No active phosphorylation of MBP is revealed by ER-associated IPs obtained from samples treated with estradiol in the presence of ICI 182,780 or Y-27632, indicating that ROCK-2 that coimmunoprecipitates with ER is responsible for the enhanced phosphorylation of MBP (Fig. 4D). Moreover, treatment of endothelial cells with PTX does not inhibit MBP activation by ER-associated IPs (Fig. 4D).
To study the activation of non-ER-bound ROCK-2 by estrogen, we immunoprecipitated total ROCK-2 from endothelial cells treated with E2 in the presence or absence of ER, ROCK-2, or G protein inhibitors, and performed kinase assays with the IPs. ROCK-2 is activated upon exposure of endothelial cells to estradiol, and this activation is prevented by treatment with ICI 182,780, Y-27632, or PTX (Fig. 4E). Taken together, the experiments depicted in Fig. 4, D and E, indicate that activation of ROCK-2 by estradiol is achieved through two alternative pathways. First, liganded ER can directly bind and transactivate ROCK-2. This process does not require activation of G proteins (or may alternatively depend on activation of some PTX-insensitive G protein). Second, in parallel, the recruitment of G proteins at the cell membrane by ER leads to the activation of ROCK-2 independent of direct interactions with ER.
Intracellular Events Linking Activation of ER at the Cell Membrane to Moesin Phosphorylation: Role of G13 and RhoA
ROCK-2 activation is usually recruited through a cascade involving the G protein G13 and the subsequent signal transduction to the small GTPase RhoA. This cascade has previously been associated with the activation of other ERM proteins such as ezrin and radixin (23, 24). We therefore studied RhoA activation by assaying RhoA GTP-binding activity. Because active RhoA localizes at the cell membrane, we assayed RhoA GTP-binding activity in purified cell membrane extracts and cytoplasmic fractions (as controls). RhoA is activated by rapid exposure of endothelial cells to estradiol (Fig. 5A). RhoA activation is prevented by the ER antagonist ICI 182,780 and is seen after addition of the ER-selective agonist PPT but not by the ERß-specific ligand, DPN. Perhaps most importantly, RhoA activation is recruited by administration of the membrane-impermeable estradiol-BSA (Fig. 5A). This indicates that a ligand-dependent activation of a cell membrane-associated ER results in RhoA activation in human endothelial cells.
|
|
40%) (see Fig. C, published as supplemental data). Parallel transfections with a constitutively active G13 construct (G13 Q226L) show a significant (and estradiol-independent) activation of RhoA in endothelial cells (Fig. 5B), whereas transfection of a dominant-negative G13 construct (G13 Q226L/D294N) results in a significant reduction of RhoA activation by E2 (Fig. 5B), confirming the link between ER and G13 for the activation of RhoA.
Several signal transducers, including phospholipase C (PLC), protein kinase C (PKC), the p38 MAPK pathway, nitric oxide, and Ras have been implicated in the signaling to RhoA (25). Blockade of PLC with U-73122 results in a complete inhibition of RhoA activation by estradiol (Fig. 5C). Similarly, the transfection of a dominant-negative Ras construct (Ras G12V) results in complete prevention of the effects of estradiol (Fig. 5C), therefore suggesting an involvement of both PLC and Ras in the estradiol-induced signaling of G13 to RhoA. In agreement, a constitutively active Ras construct (Ras S17N) produces a strong (and estrogen-independent) activation of RhoA (Fig. 5C). In contrast, marginal or null effects on estradiol-induced RhoA activity were found in the presence of calphostin C (PKC inhibitor), SB 203580 (p38 inhibitor), or L-NAME (nitric oxide synthase inhibitor).
Because the previous studies coherently indicated that G13 is recruited by estrogens, we studied whether administration of estradiol to endothelial cells results in activation of G13. Indeed, estradiol administration results in enhanced G13 GTP-binding activity, which is prevented by ICI 182,780 (Fig. 5D). A similar effect is obtained by treating endothelial cells with the membrane-impermeable estradiol, E2-BSA, as well as by the ER-selective agonist, PPT (but not by the ERß-specific ligand, DPN) (Fig. 5D).
Transient transfection with the constitutively active G13 construct results in highly increased GTP-binding activity in endothelial cells that is not further augmented by estrogen administration (Fig. 5E). Instead, transfection of the dominant-negative G13 construct results in a significant inhibition of estrogen-dependent G13 activity (Fig. 5E). Control transfection with mutated RhoA constructs does not alter the basal or estrogen-induced G13 GTP-binding activities (see Fig. D, published as supplemental data).
G13 and RhoA Are Necessary for Estrogen-Dependent Moesin Activation
Moesin phosphorylation is significantly induced in cells transfected with G13 or RhoA constitutively active constructs (Fig. 5F) and no additional effect of estrogen are observed (Fig. 5F). However, the transfection of the dominant-negative constructs for G13 or RhoA results in a significant inhibition of estrogen-induced moesin phosphorylation (Fig. 5F), confirming that both G13 as well as RhoA activation are required to achieve moesin recruitment by estrogen.
G13 and RhoA Are Necessary for the Estrogen-Dependent Actin Cytoskeletal Rearrangement
The cytoskeletal modifications induced by rapid estradiol treatment are mimicked by the transfection of the constitutively active G13 or RhoA, which induce obvious accentuations of cortical actin structures, as well as the formation of prominent cell membrane protrusions, with adjacent cells developing intercellular bridges at sites where focal adhesion complexes are formed (Fig. 5G). In these cells, treatment with estradiol does not further alter the cytoskeletal structure (Fig. 5G). In contrast, transfection with the G13 dominant-negative construct completely prevents the cytoskeletal modifications induced by estradiol (Fig. 5G). Dominant-negative RhoA is associated with even more profound effects, resulting in the complete lack of a recognizable peripheral actin cytoskeleton. In addition, the transfection of this construct prevents any estrogen-associated actin modifications (Fig. 5G).
ER Activates G13 via an Indirect Protein-Protein Interaction
To understand the mechanistic basis of the estrogen-dependent activation of G13, we tested whether estradiol triggers G13 activation through the induction of ER/G13 protein-protein interaction. Indeed, with coimmunoprecipitation analysis on purified cell membrane extracts, we show that ER and G13 interact with each other and that the presence of estradiol enhances this interaction (Fig. 6A) in the absence of modifications of the overall amount of G13 in endothelial cells (see Fig. E, published as supplemental data). In addition, ICI 182,780 blocks the estrogen-induced ER/G13 interaction, which is instead not affected by PTX (Fig. 6A). PTX catalyzes the ADP-ribosylation of the subunits of heterotrimeric guanine nucleotide regulatory proteins, therefore restraining the G subunits in their GDP-bound, inactive state. The lack of effect of PTX on ER/G13 interaction therefore suggests that G13 can bind ER independent of its activation status. Parallel co-IP studies show no interaction of G13 with ERß (Fig. 6A), confirming that the ability to recruit this pathway is limited to ER.
|
with G13, we expressed recombinant glutathione S-transferase (GST)-fused ER and His-tagged G13 proteins and performed in vitro GST pull-down assays. Parallel GST pull-down studies using BSA or whole endothelial cell extracts as buffers for protein cointeraction revealed that G13 and ER interact only in the presence of cell extracts, suggesting that the interaction is indirect, possibly requiring one or more scaffold proteins (Fig. 6B).
Moreover, GST pull-downs performed coincubating full-length G13 with GST-bound full-length-ER or with GST-bound ER mutants lacking alternatively the NH2-terminus (
AF-1 ER) or the COOH terminus (534 ER) indicate that the area of ER that interacts with G13 is located in the COOH terminus (Fig. 6C).
Finally, double immunostaining of endothelial cells shows that ER and G13 colocalize in the presence of estradiol or of membrane-impermeable estradiol and that this behavior is blocked in the presence of the ER-antagonist, ICI 182,780 (Fig. 6D). However, the subcellular localization of the cointeraction site cannot be definitely established with direct immunofluorescence.
Estrogen-Dependent Endothelial Migration Depends on an ER/G Protein/ROCK/Moesin Pathway
To understand the relevance for endothelial cell migration of estrogen-dependent moesin activation through the G13/RhoA/ROCK pathway, we performed endothelial migration assays. Endothelial cells grown in the presence of estradiol or E2-BSA show a quicker migration toward the de-endothelialized area of the dish than vehicle-treated cells (Fig. 7). This advantage is, however, obviated by antagonists of the ER, of G proteins, and of ROCK (Fig. 7), suggesting that these mediators are implicated in the migration of endothelial cells induced by estradiol.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
We previously showed that estrogen regulates ezrin-induced cancer cell motility (2, 13). Our present findings extend the understanding of estrogen-regulation of ERM proteins and motility to human vascular integrity. In this report, we show that estrogen’s regulation of moesin occurs by way of a novel cell-signaling pathway that furnishes rapid, nongenomic actions via the ER and G13/RhoA/ROCK. Further studies will be required to determine whether this newly discovered estrogen-signaling pathway is common for all G protein-regulated proteins.
The actions described in this report occur in minutes and are akin to other examples of nongenomic estrogen actions (4). These rapid actions of sex steroids have been found to be relevant in different cell types, including cancer, neurons, and vascular cells (4). Moesin activation after endothelial cell exposure to estradiol is a new example of such nongenomic actions, because it is triggered rapidly (in a matter of minutes) and is thereafter quickly reversed. This plasticity of rapid estrogen action is especially interesting, because it implies adaptive changes of the cell morphology and function such as are necessary for cell motility, proliferation, and invasive behavior (11), as well as the reparative and developmental actions of endothelial cells. In this manner, the kinetics of moesin activation necessary for the rearrangement of the actin cytoskeleton, membrane ruffling, and formation of pseudopodia and lamellipodia, is obtained that supports periodic waves of actin remodeling and develops dynamic anchoring to the extracellular matrix to move in the surrounding environment (11, 26). Therefore, we suggest that, within the broader range of actions of ERs, nongenomic signaling to ERM proteins, particularly moesin, and actin may have special relevance for the determination of estrogen-regulated cell movement in blood vessels. In addition, loss of stress fibers has been associated with cancer cell transformation and metastasis (27). Thus, the cytoskeletal rearrangements induced by estradiol may be related to the carcinogenic actions of this steroid in estrogen-sensitive cancers and eventually with estrogen-dependent cancer cell migration (28). In support of this, we have seen similar actions of estrogen linked to ezrin-driven changes in estrogen-sensitive cancers (Ref. 2 ; and Fadiel, A., and F. Naftolin, unpublished observations).
Our data indicate that ER localized at the cell membrane acts as a G protein-coupled receptor, dynamically interacting with G13 in a ligand-dependent fashion and triggering its activation; this interaction of ER with G13 has not previously been reported. Based on our results, it requires the 62 carboxy-terminal amino acids of ER. This area of the receptor contains a large portion of the hormone-dependent activation function-2 (AF-2) domain, located in the ligand binding domain (29). The AF-2 functional domain includes a highly conserved amphipathic -helix (helix 12) that is essential for ligand-dependent transcriptional activity and interaction with members of the SRC family of coactivators (29). Given the central role played by this domain in the interaction of ER with coactivators and chaperones, the localization of the ER/G13 interaction site to this area couples well with the parallel observation that yet-unidentified protein intermediates are required to bridge the two proteins. This observation is also consistent with recent work indicating that ER is anchored at the cell membrane by scaffold proteins that bridge the receptor to other membrane proteins, including Gi (30) and with other reports indicating that ER can activate G protein-dependent signaling in endothelial cells (31) as well as in other cells such as neurons and breast cancer cells (32, 33).
Our results also indicate that the interaction with G13 does not extend to ERß. In support of this observation, although ERß also contains an AF-2 domain within the C terminus (29), growing evidence indicates that this region may have different roles in the two ER subtypes (29). Furthermore, the present finding that ERß is not involved in the signaling cascade leading to actin rearrangement and endothelial migration is in agreement with the previous observation that ER is necessary to mediate estrogens’ protecting effects during experimental vascular injury in mice whereas ERß is not required (34).
Through the activation of G13 at the cell membrane, ER has a privileged access to the pathways that signal to the actin cytoskeleton. In fact, G13 is responsible for the activation of the RhoA/ROCK-2/moesin cascade pathway (35), and we here show that this connection is also functional in human endothelial cells, providing the first evidence of a regulatory effect of estrogens on this important signaling pathway and providing the indication that the activation of this pathway is implicated in estrogen-induced endothelial migration. These findings shed light on the mechanisms of the protective effects of estrogens through the promotion of endothelial cell migration during vascular injury (6, 36). Through the loss of stress fibers, the induction of waves of actin assembly near the cell membrane and the formation of lamellipodia, pseudopodia, and ruffles, estrogens may facilitate endothelial cell cross talk and migration, contributing to the maintenance of functional vessels. In this regard, previous observations suggest that estrogens may render endothelial cells resistant to metabolic stress due to modulation of the actin cytoskeleton (37), indicating a role for additional pathways, such as the p38ß MAPK cascade, for the regulation of actin arrangement by estrogen during cellular stress (37). Furthermore, the regulation of cellular actin and of endothelial cell movement is crucial during angiogenesis, which is also regulated by estrogens (38, 39).
In conclusion, we have identified a novel rapid, nongenomic signal transduction pathway that is ligand- activated by ER. Through activation at the cell membrane of G13 via an indirect interaction involving the AF-2 domain, ER recruits the RhoA/ROCK-2/moesin cascade, leading to dynamic rearrangements of the actin cytoskeleton that allow the formation of specialized cell membrane structures. This results in cell attachment to the extracellular matrix and cell movement in the surrounding environment. Beyond its relevance in endothelial cells, the characterization of this novel mechanism of action may offer insight into a number of processes regulated by estrogens that involve cell movement, such as wound healing, angiogenesis, atherogenesis, neuronal remodeling, or cancer metastasis, and identifies a previously unknown target for estrogens in human cells, which may in the future represent the target of novel therapeutic applications.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Immunoblotting
Cell lysates were separated by SDS-PAGE. Antibodies against the following proteins were used: moesin (clone 38; Transduction Laboratories, Lexington, KY), Thr558-P-moesin (sc-12895; Santa Cruz Biotechnology, Santa Cruz, CA), ER (clone TE111; NeoMarkers, Union City, CA), ERß (clone 9.88; Sigma-Aldrich), EBP50 (clone 6; Transduction Laboratories), ROCK-2 (C-20; Santa Cruz Biotechnology), RhoA (clone 26C4; Santa Cruz Biotechnology), and G13 (clone A-20; Santa Cruz Biotechnology). Primary and secondary Abs were incubated with the membranes by the standard technique. Immunodetection was accomplished using enhanced chemiluminescence.
Cell Immunofluorescence
HUVEC were grown on coverslips and exposed to treatments. Cells were fixed and permeabilized with methanol at –20 C for 10 min. Blocking was performed with 3% normal serum for 20 min. Cells were incubated with antibodies against G13 (clone A-20; Santa Cruz Biotechnology), human ER (TE111; NeoMarkers), Texas Red-phalloidin (Sigma- Aldrich), or vinculin (clone V284; Upstate, Lake Placid, NY). The nuclei were counterstained with or 4'-6-diamidino-2-phenylindole (DAPI) (Sigma-Aldrich) and mounted with Vectashield mounting medium (Vector Laboratories, Burlingame, CA). Immunofluorescence was visualized using an Olympus BX41 microscope and recorded with a high-resolution DP70 Olympus digital camera.
Coimmunoprecipitation Assays
HUVEC were harvested in 100 mM Tris-HCl (pH 6.8), 4% sodium dodecyl sulfate (SDS), 20% glycerol, 1 mM Na3VO4, 1 mM NaF, and 1 mM phenylmethylsulfonylfluoride (PMSF). Equal amounts of cell lysates were incubated with 1 µg precipitating Ab for 1 h at 4 C under gentle agitation. 25 µl of a 1:1 protein A-agarose slurry were added, and the samples were rolled at 4 C for another hour. The samples were then pelleted, washed, and resuspended in 50 µl 2x Laemmli buffer for immunoblotting.
Kinase Assays
HUVEC were harvested in 20 mM Tris-HCl, 10 mM EDTA, 100 mM NaCl, 0.5% IGEPAL, and 0.1 mg/ml PMSF. Equal amounts of cell lysates were immunoprecipitated with an Ab vs. ER (TE111) or ROCK-2 (C-20; Santa Cruz Biotechnology). The IPs were washed three times with 20 mM Tris-HCl, 10 mM EDTA, 150 mM NaCl, 0.1% IGEPAL, and 0.1 mg/ml PMSF. Two additional washes were performed in kinase assay buffer (20 mM 3[N-morholino]propanesulfonic acid, 25 mM ß-glycerophosphate, 5 mM EGTA, and 1 mM dithiothreitol), and the samples were resuspended in this buffer. Five micrograms of dephosphorylated MBP (Upstate) together with 500 µM ATP and 75 mM MgCl2 were added to each sample and the reaction was started putting the samples at 30 C for 20 min. The reaction was stopped on ice and by resuspending the samples in Laemmli buffer. The samples were separated with SDS-PAGE, and Western analysis was performed using an Ab recognizing Thr98-P-MBP (Upstate).
Rho Activity Assay
HUVEC were exposed to E2 (1 nM), U-73122 (5 µM; PLC inhibitor), calphostin C (1 µM; PKC inhibitor), SB 203580 (1 µM; p38 inhibitor), and L-NAME (1 mM; nitric oxide synthase inhibitor). Ras constitutively active and negative dominant cDNA plasmid (G12V mutant and S17N mutant, respectively; University of Missouri, Rolla, cDNA Resource Center, Rolla, MO) were transfected using Lipofectamine 2000 and OptiMEM (Invitrogen, Carlsbad, CA). Cells were treated 60 h after transfection. RhoA activity was assayed with the Rho Activation Assay kit (Upstate).
GTP-Binding Assays
HUVEC were washed with PBS and suspended in ice-cold lysis buffer (5 mM EDTA, 5 mM EGTA, 100 mM NaCl, 5 mM Tris-HCl (pH 7.4), protease inhibitor, phosphatase inhibitor), disrupted with a Dounce homogenizer (10 strokes twice), and centrifuged at 40,000 x g for 30 min at 4 C. The pellet was washed twice with ice-cold suspension buffer [100 mM Tris-HCl, 5 mM MgCl2, 0.6 mM EDTA (pH 7.4)] and then centrifuged at 2700 x g for 10 min. The pellet was discarded, and the supernatant was centrifuged again at 40,000 x g for 45 min at 4 C. The resulting pellet containing the enriched membrane preparation was then resuspended at a protein concentration of 2 mg/ml as determined by the BCA method. Membrane (30 µg) from HUVEC were incubated for 30 min at 22 C in a buffer containing GTP
S (10 nM), GTP (2 µM), MgCl2 (5 mM), EGTA (0.1 mM), NaCl (50 mM), creatine phosphate (4 mM), phosphocreatine kinase (5 units), ATP (0.1 mM), dithiothreitol (1 mM), leupeptin (100 µg/ml), aprotinin (50 µg/ml), BSA (0.2%), and triethanolamine-HCl (50 mM; pH 7.4). Nonspecific activity was determined in the presence of excess unlabeled GTPS (100 µM). The assay was terminated after 15 min with excess unlabeled GTPS (10 µM). The G13 and RhoA GTP-binding activities were determined by immunoprecipitation of [35S]GTPS-labeled G protein with the G13 Ab (clone A-20; Santa Cruz Biotechnology) or with an anti-RhoA antibody (clone 26C4; Santa Cruz Biotechnology). Samples were resuspended in 200 µl of immunoprecipitation buffer containing Triton X-100 (1%), SDS (0.1%), NaCl (150 mM), EDTA (5 mM), Tris-HCl (25 mM; pH 7.4), protease inhibitor cocktail, and phosphatase inhibitor cocktail II (Sigma-Aldrich). The samples were allowed to incubate for 2 h at 4 C with gentle mixing. The complexes were then incubated with 40 µl protein A-agarose for 1 h at 4 C, and the IP was collected by centrifugation at 12,000 x g for 1 min. The pellets were washed three times in 300 µl buffer containing HEPES (50 mM; pH 7.4), NaF (100 µM), sodium phosphate (50 mM), NaCl (100 mM), Triton X-100 (4%), and SDS (0.1%). The final pellet containing the immunoprecipitated [35S]GTPS-labeled G13 or RhoA was resuspended in 300 µl water and was counted in a liquid scintillation counter (LS 1800; Beckman Instruments, Fullerton, CA).
Transient Transfections
HUVEC were used for all transfection experiments. HUVEC were transfected with each plasmid (15 µg) using the Lipofectin reagent (Invitrogen) according to the manufacturer’s instructions. The transfected plasmids were as follows: G13 Q226L, G13 Q226L/D294N, RhoA T19 and RhoA G14V, Ras G12V and Ras S17N. All the inserts were cloned in pcDNA3.1+. The constructs were obtained from the Guthrie cDNA Resource Center (www.cdna.org). As control, parallel cells were transfected with empty pcDNA3.1+ plasmid. Cells (60–70% confluent) were treated 24 h after transfection, and cellular extracts were prepared according to the experiments to be performed.
Gene Silencing with RNA Interference
Two synthetic siRNA targeting ER (siRNA SMARTpool ESR1; Dharmacon, Lafayette, CO) were used at the final concentration of 100 nM to silence ER according to the manufacturer’s instructions. Endothelial cells were treated 60 h after siRNA transfection. The efficacy of gene silencing was checked with Western analysis and found to be optimal at 60 h.
Endothelial Cell Migration Assays
Endothelial cell migration was assayed with scrape assays. Briefly, confluent endothelial cells were synchronized by replacing the medium with human endothelial serum-free medium (Invitrogen) devoid of growth factors for 24 h. A razor blade was pressed through the endothelial cell monolayer into the plastic plate to mark the starting line. Endothelial cells were swept away on one side of that line. Endothelial cells were washed, and 1.0 ml human endothelial serum-free medium containing gelatin (1 mg/ml) and the test substance was added. Migration was monitored for 24 h. Every 12 h, fresh medium and treatment were replaced. Cells were digitally imaged during phase-contrast microscopy.
Moesin Silencing with Antisense Oligonucleotides
Validated antisense (S-modified) PONs complementary to the 1–15 position of the human moesin gene coding region (41) were obtained. The sequence was 5'-TACGGGTTTTGCTAG-3' for moesin antisense PON. The complementary sense PON was used as control (5'-CTAGCAAAACCCGTA-3'). PON transfections were performed on subconfluent HUVEC in 30-mm petri dishes. PONs were resuspended in serum-free medium with 2% Lipofectin (Invitrogen) and added to the culture medium every 12 h at the final concentration of 4 µM. Every 24 h, HUVEC were washed and fresh medium supplemented with 4 µM PONs was added. Moesin silencing was assessed through protein analysis up to 72 h after transfection.
In Vitro GST Pull-Down Assays
The full-length G13 cDNA was subcloned in the pDNR-1 expression vector with the Pro-Tet 6xHN Bacterial Expression System (BD Creator DNA Cloning; BD Biosciences, Mountain View, CA). N-term-poly-histidine-tagged G13 was expressed in BL21 Star (DE3) Escherichia coli chemically competent cells (Invitrogen) and purified with the Talon purification kit (BD Bioscience). Full-length human ER (WT-ER), AF-1-ER (truncated ER missing amino acids 1–180, corresponding to the NH2-terminal AF-1 domain) and 534-ER (truncated ER missing amino acids 534–596, corresponding to the COOH-terminal region of the receptor and to part of the hormone binding domain/AF-2 domain) cloned in a modified pGEX vector were expressed in BL21 E. coli. The recombinant GST-linked proteins were purified with the Bulk GST kit (Amersham Biosciences, Piscataway, NJ). For the GST pull-down assays, 5 µg purified G13 were mixed with 5 µg of either WT-ER, AF-1-ER, or 534-ER, and incubated for 1 h at 4 C in Bead Binding buffer (1 mM KCl, 50 mM potassium phosphate, 1 mM MgCl2, 10% glycerol, 1% Triton X-100, 50 mM NAF, 0.1 mg/ml PMSF, and 0.3 mg/ml aprotinin). As protein buffer, 100 mg/ml BSA or the same amount of whole endothelial cell protein extracts were added. Glutathione/Sepharose 4B beads (Amersham Biosciences) equilibrated with Bead Binding buffer were added and then incubated for 1 h at 4 C. The beads were washed three times with Bead Binding buffer, and proteins were eluted with SDS-PAGE sample buffer. Coprecipitated G13 was detected using a specific antibody.
Statistical Analysis
All values are expressed as mean ± SD. Statistical differences between mean values were determined by ANOVA, followed by the Fisher’s protected least significance difference.
|
ACKNOWLEDGMENTS |
|---|
protein expression constructs. |
FOOTNOTES |
|---|
First Published Online April 6, 2006
Abbreviations: AF-2, Activation function-2; DAPI, 4'-6-diamidino-2-phenylindole; DPN, 2,3-bis(4-hydroxyphenyl)-propionitrile; EBP50, ERM-binding protein 50; ER, estrogen receptor; ERM, ezrin/radixin/moesin; GST, glutathione S-transferase; HUVEC, human umbilical vein endothelial cell; IP, immunoprecipitate; ir, immunoreactive; MBP, myelin basic protein; PKC, protein kinase C; PLC, phospholipase C; PMSF, phenylmethylsulfonylfluoride; PON, phosphorothioate oligonucleotide; PPT, 4,4',4''-(4-propyl-[1H]-pyrazole-1,3,5-triyl)trisphenol; PTX, pertussis toxin; ROCK, Rho-associated kinase; SDS, sodium dodecyl sulfate; siRNA, small interfering RNA.
Received for publication June 30, 2005. Accepted for publication March 31, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
but not estrogen receptor-ß. Circulation 103:423–428-Shc association and Shc pathway activation. Mol Endocrinol 16:116–127 and epidermal growth factor. Cancer 92:3068–3075[CrossRef][Medline] and estrogen receptor-ß: correlations with biological character and distinct differences among SRC coactivator family members. Endocrinology 141:3534–3545
