17β-Estradiol Inhibits Phosphorylation of Stromal Interaction Molecule 1 (STIM1) Protein

Background: It is unclear how acute exposure to estrogen inhibits Ca2+ influx and Ca2+-activated Cl− secretion in airway epithelia. Results: Estrogen inhibited oligomerization and basal phosphorylation of STIM1, a key regulator of Ca2+ influx. Conclusion: Acute exposure to estrogen inhibits Ca2+ influx by disrupting STIM1. Significance: Knowing how estrogen inhibits Ca2+ signaling in the airways increases our understanding of sex-specific differences in lung diseases. Sex plays a significant role in the development of lung diseases including asthma, cancer, chronic bronchitis, and cystic fibrosis. In cystic fibrosis, 17β-estradiol (E2) may inhibit store-operated Ca2+ entry (SOCE) to impinge upon airway secretions, leaving females at greater risk of contracting lung infections. Stromal interaction molecule 1 (STIM1)-mediated SOCE is essential for cell homeostasis and regulates numerous processes including cell proliferation, smooth muscle contraction, and secretion. E2 can signal nongenomically to modulate Ca2+ signaling, but little is known of the underlying mechanisms. We found that E2 exposure inhibited STIM1 translocation in airway epithelia, preventing SOCE. This correlated with a decrease in STIM1-STIM1 FRET and STIM1 mobility in E2-exposed HEK293T cells co-expressing estrogen receptor α. We also examined the role of STIM1 phosphorylation in E2-mediated inhibition of STIM1 mobility. STIM1 is basally phosphorylated at serine 575, which is required for SOCE. Exposure to E2 significantly decreased STIM1 serine phosphorylation. Mutating serine 575 to an alanine blocked STIM1 phosphorylation, reduced basal STIM1 mobility, and rendered STIM1 insensitive to E2. These data indicate that E2 can signal nongenomically by inhibiting basal phosphorylation of STIM1, leading to a reduction in SOCE.

Sex plays a significant role in the development of lung diseases including asthma, cancer, chronic bronchitis, and cystic fibrosis. In cystic fibrosis, 17␤-estradiol (E2) may inhibit store-operated Ca 2؉ entry (SOCE) to impinge upon airway secretions, leaving females at greater risk of contracting lung infections. Stromal interaction molecule 1 (STIM1)-mediated SOCE is essential for cell homeostasis and regulates numerous processes including cell proliferation, smooth muscle contraction, and secretion. E2 can signal nongenomically to modulate Ca 2؉ signaling, but little is known of the underlying mechanisms. We found that E2 exposure inhibited STIM1 translocation in airway epithelia, preventing SOCE. This correlated with a decrease in STIM1-STIM1 FRET and STIM1 mobility in E2-exposed HEK293T cells co-expressing estrogen receptor ␣. We also examined the role of STIM1 phosphorylation in E2-mediated inhibition of STIM1 mobility. STIM1 is basally phosphorylated at serine 575, which is required for SOCE. Exposure to E2 significantly decreased STIM1 serine phosphorylation. Mutating serine 575 to an alanine blocked STIM1 phosphorylation, reduced basal STIM1 mobility, and rendered STIM1 insensitive to E2. These data indicate that E2 can signal nongenomically by inhibiting basal phosphorylation of STIM1, leading to a reduction in SOCE.
Calcium is an important second messenger that controls cellular processes as varied as contraction, ion transport, regulation of transcription, growth, and cell division (1). Thus, precise regulation of cytoplasmic Ca 2ϩ is required to maintain normal cellular homeostasis. This is achieved mainly by sequestering Ca 2ϩ to internal stores, such as the endoplasmic reticulum (ER) 3 and mitochondria (2). Following receptor-mediated ER Ca 2ϩ depletion, store-operated Ca 2ϩ entry (SOCE) is used to further modulate cytoplasmic Ca 2ϩ levels and to replenish depleted stores (3). SOCE requires the ER Ca 2ϩ -sensing protein stromal interaction molecule 1 (STIM1) (4,5) and the Ca 2ϩ release-activated Ca 2ϩ channel subunit Orai1 (6,7). The N-terminal domain of STIM1 contains an EF-hand that protrudes into the lumen of the ER and senses changes in ER Ca 2ϩ levels (3,8). Upon depletion of ER Ca 2ϩ , STIM1 undergoes a conformational change and oligomerizes into puncta at the ER-plasma membrane junction, where it activates Orai1 to induce Ca 2ϩ influx (3,8). In airway epithelia, one important consequence of agonist-dependent increases in [Ca 2ϩ ] i is the subsequent activation of transepithelial Cl Ϫ secretion through the Ca 2ϩ -activated Cl Ϫ channel (CaCC). In patients with cystic fibrosis (CF), who the lack the CF transmembrane conductance regulator (CFTR) Cl Ϫ channel, CaCC is still present and represents a "rescue" channel that may help maintain airway hydration in the absence of the CFTR (9).
Gender can greatly influence lung health (10). For example, CF women have a poorer prognosis and are more likely to experience an acute exacerbation than CF men (11,12). Additionally, adult women are more likely to be diagnosed with early onset chronic obstructive pulmonary disease (COPD) (13), asthma (14), and adenocarcinoma of the lung (15). We have previously shown that 17␤-estradiol (E2) inhibits agonist-mediated Ca 2ϩ signaling through estrogen receptor ␣ (ESR1) in a nongenomic fashion in airway epithelia, which ultimately prevents CaCC activation and leads to airway dehydration (16). Here, we tested the hypothesis that E2 inhibits key components of SOCE. We found that STIM1 was specifically inhibited by E2/ESR1, leading to an inhibition of SOCE but not ER Ca 2ϩ release.
Cell Culture and Transfections-Human excess donor lungs and excised recipient lungs were obtained at the time of lung transplantation from portions of main stem or lumbar bronchi, and cells were harvested by enzymatic digestion as described previously under a protocol approved by the University of North Carolina Institutional Review Board (17). Human bronchial epithelial cells (HBECs) were plated on either glass coverslips to perform siRNA knockdown experiments or on polyester membrane Transwells (Corning) to induce polarization. HBECs that were plated on Transwells were grown in an airliquid interface for 3 weeks prior to experiments. HEK293T cells were maintained in minimum Eagle's medium ␣ supplemented with 10% fetal bovine serum and 1ϫ penicillin/streptomycin solution. HEK293T cells were typically used 2-3 days after seeding. Cultures were transfected for 4 -6 h using Lipofectamine 2000 (Life Technologies) according to the manufacturer's instructions. After transfecting, cultures were washed and placed in phenol-free medium and allowed to incubate in 5% CO 2 at 37°C overnight.
siRNA Knockdown-STIM1 and Orai1 were transiently knocked down in HBECs using the Amaxa Nucleofector system according to the manufacturer's instructions with at least two different siRNA sequences obtained from Dharmacon. STIM1 and Orai1 knockdown were verified by quantitative PCR and at the functional level by measuring changes in intracellular Ca 2ϩ with Fura-2.
Measurements of Intracellular Ca 2ϩ -Intracellular Ca 2ϩ imaging experiments were performed as described previously (16). Briefly, HEK293T and nonpolarized HBEC cultures were loaded with 2 M Fura-2 AM at 37°C for 15 min. Polarized HBECs were loaded with 5 M Fura-2 AM while in the presence of 1 mM probenecid at 37°C for 1 h. Cultures were washed with a standard Ringer's solution and then with Ca 2ϩ -free Ringer's solution. Cultures were then placed in Ca 2ϩ -free Ringer's solution, and images were collected with a 60 ϫ 1.2 NA water objective on a Nikon Ti-S inverted microscope. Fura-2 fluorescence was acquired alternately at 340 and 380 nm (emission Ͼ450 nm) using LUDL filter wheels, obtained with an Orca CCD camera (Hammamatsu), and controlled with SimplePCI software. Background light levels were measured and subtracted from the corresponding signal measured in Fura-2-loaded cells before calculating the ratio (340/380). The background light levels in polarized HBECs were obtained by measuring the signals in non-Fura-2-loaded HBECs.
Immunofluorescence-Polarized HBECs were fixed bilaterally with 4% paraformaldehyde for 30 min at room temperature. After washing three times with PBS, HBECs were permeabilized with 0.1% Triton X-100 for 10 min and incubated in blocking buffer (1% IgG-free BSA, 1% normal goat serum) at room temperature for 1 h. HBECs were then incubated with an antibody against STIM1 in blocking buffer overnight at 4°C. After washing with PBS, HBECs were incubated with Alexa Fluor 568 goat anti-rabbit secondary antibody (Life Technologies) for 1 h at room temperature. Actin was stained with phalloidin-488 (Invitrogen) according to the manufacturer's instructions. All antibodies and stains were applied bilaterally. Images were analyzed using ImageJ (National Institutes of Health).
STIM1 Puncta Quantification-HEK293T cells expressing STIM1-mCherry and ESR1-CFP were treated with vehicle, thapsigargin (TG), E2, or E2 ϩ TG and then fixed with 4% paraformaldehyde for 10 min at room temperature. Each culture was assigned a random number that was unknown to the researcher to reduce bias. Multiple images were collected, and each cell was visually analyzed. A cell was deemed positive for puncta if any puncta were visible.
Confocal Microscopy Assays-A Leica SP5 confocal microscope with a 63ϫ glycerol immersion objective was used for all confocal-based assays. Acceptor-photobleaching fluorescence resonance energy transfer (FRET) studies were performed on HEK293T cells expressing ESR1-CFP, STIM1-YFP, and STIM1-mCherry as described previously (18). Basal FRET was measured in cells that were treated with dimethyl sulfoxide vehicle or 10 nM E2 for 15 min. 2 M thapsigargin was used to deplete ER-Ca 2ϩ stores, and FRET measurements were collected. FRET images were analyzed using ImageJ. The FRET efficiency (%E) was calculated as: ((dono postbleach Ϫ donor prebleach )/donor postbleach )*100. Fluorescent recovery after photobleaching (FRAP) was measured in HEK293T cells expressing ESR1-mOr and YFP-STIM1, YFP-STIM1 570STOP , or YFP-STIM1 S575A , unless stated otherwise. Prebleach images were obtained before a circular region of interest was bleached. Subsequent images were collected every 30 s until maximum recovery had occurred. STIM1 FRAP curves were calculated by measuring the average fluorescence intensity in the bleached region of interest per time point and dividing it by the average fluorescence intensity of a nonbleached region of interest within the same cell per time point. To account for variations in protein expression and photobleaching, the average intensity ratio was normalized to the initial prebleach ratio.
Immunoprecipitation and Western Blotting-Immunoprecipitation was performed in HEK293T cells transfected with ESR1-mOrange and YFP-STIM1 or YFP-STIM1-570STOP truncation mutant (STIM1 570STOP ). Cells were washed in 500 l of Nonidet P-40 lysis buffer (0.1% Nonidet P-40, 50 mM Tris-HCl, pH 7.4, 10 mM NaMoO 4 , 150 mM NaCl) supplemented with 1ϫ complete EDTA-free protease inhibitors (Roche Applied Science) and 2ϫ Halt phosphatase inhibitor mixture (Thermo Scientific). Lysates were then centrifuged at 16,000 ϫ g for 10 min at 4°C, and the supernatants were collected. Pro-tein concentrations were determined by the BCA protein assay kit (Pierce), and equal amounts of protein (600 g) were diluted to 1.2 g l Ϫ1 with Nonidet P-40 lysis buffer in spin columns (Pierce) and mixed with anti-GFP polyclonal antibody (Abcam; 1:500) and rotated at 4°C overnight. Protein A/G-agarose beads (30 l; Pierce) were added to lysate and rotated at 4°C for 4 h. Beads were washed three times with Nonidet P-40 lysis buffer and centrifuged at 1000 ϫ g for 2 min at 4°C, and protein was eluted by boiling in sample buffer (2ϫ NuPAGE LDS sample buffer and 1ϫ reducing agent (Life Technologies)). Protein was resolved on a 4 -12% gradient NuPAGE Bis-Tris gels (Life Technologies) and transferred to nitrocellulose using iBlot (Life Technologies). For Western blotting, immunoblots were blocked in 3% BSA in PBS-T for 1 h at room temperature, incubated overnight at 4°C with primary antibodies (anti-mpm-2, 1:500 (Millipore); anti-GFP 1:2000 (Abcam), and incubated with secondary antibodies (horseradish peroxidase-linked anti-mouse or ant-rabbit 1:10,000 (GE Healthcare)) for 1 h at room temperature. Proteins were detected using SuperSignal West Pico or Femto Chemiluminescent Substrate kit (Pierce).
Statistical Analysis-All experiments were conducted on at least three separate occasions. Data were analyzed using the method stated in the figure legend. All statistical analyses were calculated using GraphPad Prism software.

E2 Inhibits Apical Ca 2ϩ Influx in Human Bronchial
Epithelial Cells-We have shown previously that E2 indirectly inhibits ligand-evoked Ca 2ϩ -activated Cl Ϫ secretion due to a reduction in Ca 2ϩ signaling in human airway epithelia (16). To test whether this depression of cytoplasmic Ca 2ϩ was due to an E2-mediated reduction in SOCE, we performed Ca 2ϩ imaging experiments in polarized, Fura-2-loaded HBECs. HBECs were placed in Ca 2ϩ -free Ringer's bilaterally and treated apically with thapsigargin (2 M) to deplete ER Ca 2ϩ (Fig. 1A). Because E2 would physiologically reach the epithelium through the blood, we applied it basolaterally. A 15-min pretreatment with E2 (10 nM) had no effect on ER Ca 2ϩ release (Fig. 1A). We then added 2 mM Ca 2ϩ to the apical surface to induce SOCE. Cells that were pretreated with E2 displayed a significantly reduced change in the Fura-2 ratio emission upon Ca 2ϩ addition relative to untreated cells, suggesting that basolateral E2 specifically inhibits apical Ca 2ϩ entry in polarized HBECs without affecting ER Ca 2ϩ release (Fig. 1A).
Little is known regarding the molecular identity of the proteins involved in SOCE in the airway. Therefore, we next determined whether STIM1 and Orai1 were essential for SOCE in HBECs. Nonpolarized HBECs were transfected with STIM1, Orai1, or scrambled siRNA constructs or treated with vehicle (control). Quantitative PCR measurements indicated that both genes were successfully knocked down (Fig. 1B). We then performed Ca 2ϩ imaging in these cells. Nonpolarized HBECs were placed in a Ca 2ϩ -free Ringer's solution and stimulated with thapsigargin to specifically analyze ER Ca 2ϩ dynamics. Knockdown of STIM1 or Orai1 had no effect on ER Ca 2ϩ release compared with scrambled siRNA (Fig. 1C). However, knockdown of either STIM1 or Orai1 completely abolished SOCE, whereas scrambled siRNA was without effect (Fig. 1C). This indicated that STIM1 and Orai1 were essential for SOCE in HBECs. Although these observations were made in nonpolarized HBECs, it is reasonable to assume that STIM1 and Orai1 are necessary for SOCE in polarized HBECs.

STIM1-dependent Ca 2؉ influx in HBECs is inhibited by E2.
A, mean changes in Fura-2 emission ratio in polarized HBECs under control conditions or exposed to 10 nM E2 for 15 min. HBECs were placed in bilateral Ca 2ϩ -free Ringer's and stimulated with 2 M thapsigargin to monitor ER Ca 2ϩ depletion. 2 mM Ca 2ϩ was added back to the Ringer's solution on the apical surface only to induce SOCE. B, relative expression of STIM1 or Orai1 after siRNA knockdown, as indicated by quantitative PCR. Nonpolarized HBECs were treated with transfection reagent alone (Ctrl) or transfected with scrambled (Scr) or targeted knockdown (KD) siRNA constructs. C, Fura-2 emission ratio in nonpolarized HBECs that were transfected with scrambled, STIM1, or Orai1 siRNA constructs. SOCE was measured by adding extracellular Ca 2ϩ back to the Ringer's solution after ER Ca 2ϩ depletion with 2 M thapsigargin. Data are shown as mean Ϯ S.E. NOVEMBER 22, 2013 • VOLUME 288 • NUMBER 47 E2 Inhibits Translocation of STIM1 to the Apical Membrane in HBECs-We have shown previously that GPCR-mediated IP 3 formation and ER Ca 2ϩ release are unaffected by E2 (16). Because STIM1 is next in the SOCE pathway, we tested whether this protein was sensitive to E2. Under basal conditions, STIM1 was found to be diffusely located in polarized HBECs ( Fig. 2A). However, after thapsigargin-induced ER Ca 2ϩ depletion, STIM1 translocated toward the apical membrane. HBECs that were basolaterally pretreated with E2 for 15 min displayed no difference in STIM1 localization compared with untreated cells. However, when ER Ca 2ϩ was depleted in cultures pretreated with E2, STIM1 failed to translocate apically ( Fig. 2A). Data from several experiments were quantified by measuring STIM1 fluorescence intensity using an actin counterstain as a marker of polarity (Fig. 2B). Analysis of these data revealed that the subapical concentration of STIM1 was signif-icantly reduced after E2 exposure and suggested that E2 interrupted normal STIM1 function.

E2 Inhibits STIM1
E2 Inhibits SOCE and STIM1 Oligomerization in HEK293T Cells-To further study the effects of E2/ESR1 on STIM1 function, we performed additional experiments in HEK293T cells. HEK293T cells were transfected with ESR1-mOrange because they do not endogenously express estrogen receptors (19). As seen in HBECs, E2 did not affect ER Ca 2ϩ release, but significantly reduced Ca 2ϩ influx (Fig. 3, A and B), suggesting that ESR1-transfected HEK293T cells were a valid model for studying the effects of E2/ESR1 on STIM1 function.
Oligomerization of STIM1 is an essential step in the initiation of SOCE after ER Ca 2ϩ depletion (20,21). Therefore, we sought to determine whether STIM1 oligomerization after ER Ca 2ϩ depletion was inhibited by E2. To characterize the effect of E2 on STIM1 oligomerization, we performed acceptor photobleaching FRET. As a positive control, we linked YFP to mCherry with 5 glycines, which returned a FRET efficiency of ϳ42% using acceptor photobleaching, which is close to the theoretical maximum of 50% that can be measured with GFP-type proteins (n ϭ 21). Conversely, unlinked YFP and mCherry did not undergo FRET (n ϭ 10). We then performed FRET between STIM1-YFP and STIM1-mCherry in HEK293T cells. Note: these cells were also expressing ESR1-CFP. However, this construct was only used to verify ESR1 localization and was not excited during acceptor photobleaching. In agreement with previous studies (20,(22)(23)(24), we found that STIM1-STIM1 FRET significantly increased after thapsigargin-induced ER Ca 2ϩ depletion, suggesting that STIM1 had oligomerized (Fig. 3, C and D). A 15-min pretreatment with E2 alone did not affect STIM1-STIM1 FRET. However, when cells were pretreated with E2 and then stimulated with thapsigargin, no significant increase in FRET was observed (Fig. 3, C and D). These data suggest that acute exposure to E2 inhibits STIM1 oligomerization.
E2 Disrupts STIM1 Aggregation and Redistribution-Following oligomerization, STIM1 redistributes to the ER-plasma membrane junction where it interacts with Orai1 and can been seen as puncta (20,25). To determine whether E2 affects STIM1 puncta formation, we transfected HEK293T cells with STIM1-mCherry and ESR1-CFP and collected images before and after ER Ca 2ϩ depletion in the presence/absence of E2 (Fig. 4A). We did not observe STIM1 puncta under basal conditions, whereas after ER Ca 2ϩ depletion, STIM1 formed puncta in ϳ90% of the cells (Fig. 4, A and B). E2 alone did not induce STIM1 puncta formation. However, in the presence of E2, STIM1 puncta formation was significantly decreased after ER Ca 2ϩ depletion, suggesting that its ability to oligomerize had been impaired (Fig. 4, A and B).
We next performed FRET studies between STIM1 and Orai1. Under control conditions, STIM1-Orai1 FRET significantly increased after ER Ca 2ϩ depletion as described previously (23). This indicated that STIM1 and Orai1 were in close proximity. However, this increase in FRET was abolished by a 15-min E2 pretreatment, suggesting that E2 prevented STIM1 from interacting with Orai1 (Fig. 4, C and D). Because STIM1 oligomerization precedes the STIM1-Orai1 interaction, it is likely that loss of STIM1-Orai1 FRET in cells pretreated with E2 is a result of inhibiting STIM1 oligomerization. However, we cannot exclude the possibility that E2/ESR1 also independently affect Orai1. STIM1 Mobility within the ER Is Inhibited by E2-Because E2 impaired the ability of STIM1 to oligomerize and interact with Orai1 (Figs. 3 and 4), we characterized the effects of E2 on STIM1 mobility. Previous studies have utilized FRAP to demonstrate that STIM1 is highly mobile under basal conditions (20). However, upon ER Ca 2ϩ depletion, STIM1 mobility significantly decreased (20). FRAP experiments were performed on HEK293T cells co-expressing YFP-STIM1 and ESR1-CFP (Fig. 5A). Under basal conditions, we observed an ϳ80% recovery of STIM1 into the bleached area (Fig. 5, B and C). However, when ER Ca 2ϩ was depleted with thapsigargin, the recovery of STIM1 decreased significantly by ϳ50% postbleaching, indicating that STIM1 became less mobile (Fig. 5, B and C), in agreement with previous reports (20,26). We then repeated this protocol on cells pretreated with E2. Under basal conditions (i.e. with full ER stores), E2 significantly reduced STIM1 mobility (Fig. 5, B and C), suggesting that E2 exposure disrupts STIM1 mobility before STIM1 can sense changes in ER Ca 2ϩ levels. Following E2 exposure, thapsigargin only exerted a small additional effect on STIM1 mobility (Fig. 5, B and C). To ensure that E2 inhibited STIM1 mobility through ESR1, we performed the same experiment on cells only expressing YFP-STIM1 but not ESR1 (Fig. 5, D and E). Under these conditions, we found that E2 did not affect STIM1 mobility; however, the thapsigargin response was normal. This coincides with our previous observation that E2 decreases cytoplasmic Ca 2ϩ specifically through ESR1 in baby hamster kidney cells (16).
EB1 Mobility Is Not Affected by E2-STIM1 translocates along microtubules by associating with the microtubule plusend tracking protein EB1 (27). Therefore, it is possible that the decrease in STIM1 mobility after E2 exposure was due to a decrease in EB1 mobility. To test this, we performed FRAP studies on HEK293T cells co-expressing EB1-GFP and ESR1-CFP. We found no change in EB1 mobility after ER Ca 2ϩ depletion or exposure to E2 (Fig. 6). These data suggest that E2/ESR1 is directly impacting STIM1 without affecting microtubule turnover.
E2 Inhibits Basal Phosphorylation of STIM1-STIM1 contains several serine phosphorylation sites within its C-terminal domain that affect STIM1 function (28,29). To determine whether E2/ESR1 exerted their influence on STIM1 by modulating the phosphorylation status of the latter, we used an anti-GFP antibody that recognizes YFP to immunoprecipitate YFP- STIM1 from HEK293T cells that co-expressed ESR1. The serine phosphorylation status of STIM1 was then probed using the mpm-2 antibody that specifically recognizes phosphorylated serines that are followed by a proline as described previously (29). All known serine phosphorylation sites in STIM1 are followed by a proline; therefore, the mpm-2 antibody should theoretically detect phosphorylation on every known serine phosphorylation site. In the absence of E2, we observed a distinct band at the predicted mass for YFP-STIM1 (97 kDa, Fig.  7A), indicating that STIM1 was phosphorylated. Cells were then exposed to E2 for 5, 10, and 15 min to determine whether the serine phosphorylation status was altered. After a 10-min exposure to E2, we observed a significant decrease in STIM1 phosphorylation, and by 15 min, phosphorylation was almost completely abolished (Fig. 7, A and B). Although we observed a rapid inhibition of STIM1 phosphorylation, it was also important to determine whether the inhibitory action of E2 was persistent. Therefore, we performed an additional experiment where cells were exposed to E2 for up to 2 h (Fig. 7, C and D). STIM1 phosphorylation was fully inhibited by 15 min and remained inhibited during the entire time course, suggesting that E2-mediated inhibition of STIM1 phosphorylation is not transient.
To determine where the phosphorylation site(s) were located, we measured phosphorylation in a STIM1 truncation mutant that terminated at amino acid 570 (STIM1 570STOP ) (29). We observed a significant reduction in phosphorylation both basally and following E2 exposure with this mutant, indicating that the serine phosphorylation site was located toward the C terminus of STIM1 (Fig. 7, E and F). The lack of phosphorylation with STIM1 570STOP also acted as a control to confirm that the YFP tag was not being phosphorylated. Because we observed a decrease in phosphorylation after E2 pretreatment, we hypothesized that basal phosphorylation was being inhibited. Serines 575, 608, and 621 are phosphorylated and are required to allow activation of SOCE (28). As a first step toward testing whether these serines were affected by E2/ESR1, we constructed a YFP-STIM1 point mutant where serine 575 was  Dashed circles indicate the area that was bleached. B, normalized recovery of YFP-STIM1 after photobleaching in HEK293T cells co-expressing ESR1-CFP under basal conditions (n ϭ 15) and after thapsigargin stimulation (n ϭ 13) or when pretreated with E2 (n ϭ 17) and after E2 with thapsigargin (n ϭ 16). C, mobile fraction of YFP-STIM1 in HEK293T cells co-expressing ESR1 under each condition. D, normalized recovery of YFP-STIM1 after photobleaching in HEK293T under basal conditions (n ϭ 13) and after thapsigargin stimulation (n ϭ 10) or when pretreated with E2 (n ϭ 9) and after thapsigargin stimulation (n ϭ 10). E, mobile fraction of YFP-STIM1 in HEK293T cells under each condition. *, significant difference of p Ͻ 0.05; **, significant difference of p Ͻ 0.001 (one-way ANOVA followed by Tukey's multiple comparison test). Data are shown as mean Ϯ S.E. Scale bar represents 10 m.
We next performed FRAP studies on STIM1 S575A to determine whether removal of this phosphorylation site affected STIM1 mobility (Fig. 7, G and H). STIM1 S575A was significantly less mobile than WT STIM1 under basal conditions (Fig. 7H). Furthermore, treatment with thapsigargin or E2 had no additional effect on STIM1 S575A mobility. These data suggest that loss of STIM1 phosphorylation at Ser-575 by E2/ESR1 induces impaired STIM1 mobility, leading to reduced SOCE. The mobility of the triple point mutant, STIM1 S575A/S608A/S621A , was not different from STIM1 S575A (data not shown). Overall, these data indicate that phosphorylation of Ser-575 is crucial for normal STIM1 function and that a change in the STIM1 phosphorylation status by E2/ESR1 is a mode of regulation for this protein.

DISCUSSION
Many biological processes are initiated by rises in cytoplasmic Ca 2ϩ . Thus, regulation of intracellular Ca 2ϩ levels is extremely important for normal cellular function (1). We have shown previously that E2 inhibits Ca 2ϩ activated Cl Ϫ secretion in airway epithelia by obstructing Ca 2ϩ influx (16); however, the pathway utilized by E2 to disrupt this process was not known. Here, we show that acute exposure to E2 inhibits translocation of STIM1 to the apical membrane region in polarized HBECs after ER Ca 2ϩ depletion (Fig. 2). Together with calcium imaging experiments in polarized HBECs (Fig. 1), our findings suggest that SOCE occurs at the apical membrane in polarized HBECs and is specifically inhibited by E2. These findings complement previously published data that indicated that SOCE can occur at the apical membrane in nasal epithelial cells (30). Paradiso et al. demonstrated that Ca 2ϩ influx was itself polarized and was ipsilateral to the stimulus (30).
Previous studies have shown that under resting conditions, STIM1 exists as a dimer that oligomerizes into a higher order complex immediately after ER Ca 2ϩ depletion (20,31). Several studies have utilized FRET to measure oligomerization (20,25). An increase in FRET efficiency after ER Ca 2ϩ depletion is suggestive of STIM1 oligomerization. However, FRET only shows proximity and does not confirm an interaction. In our studies, we found that acute exposure to E2 inhibited the thapsigarginmediated FRET increase, suggesting that oligomerization was hindered (Fig. 3). It is important to note that E2 alone had no effect on STIM1-STIM1 FRET under resting conditions; however, STIM1 mobility was significantly decreased (Fig. 5), suggesting that the resting oligomeric structure of STIM1 was not affected by E2. Therefore, it is possible that E2 only affects STIM1 aggregation after ER Ca 2ϩ depletion and not the resting oligomeric structure. Following oligomerization, STIM1 translocates to the ER-plasma membrane junction where it aggregates into distinct puncta and interacts with Orai1. Our results indicated that E2 inhibited the formation of puncta after ER Ca 2ϩ depletion, which coincided with a decrease in STIM1-Orai1 FRET (Fig. 4). Because STIM1 oligomerization precedes puncta formation and the STIM1-Orai1 interaction, it is likely that the observations in Fig. 4 are a result of inhibiting STIM1 oligomerization. STIM1 is also known to interact with members of the transient receptor potential (TRP) superfamily, which can form calcium-permeable channels (32). The exact role of TRP channels in SOCE is not fully understood. However, TRPC1, 3, and 4 have been shown to enter into a complex with Orai1 and may facilitate Ca 2ϩ entry (33,34). Therefore, we speculate that E2-mediated inhibition of STIM1 oligomerization may also inhibit STIM1-TRP channel interactions.
In our FRAP studies, we observed a significant decrease in the mobile fraction of STIM1 after thapsigargin-induced ER Ca 2ϩ depletion, which is consistent with previously published data (20). This finding may be due to STIM1 aggregates interacting with the plasma membrane and/or Orai1. To our surprise, E2 alone decreased the mobile fraction of STIM1 without inducing puncta formation. Because STIM1 aggregation precedes translocation to the ER-plasma membrane junction, it is unlikely that E2 decreased STIM1 mobility by inducing a plasma membrane interaction. However, it is possible that E2 altered a STIM1-protein interaction that decreased mobility under resting conditions. Several studies have shown that cytoskeletal proteins, such as tubulin and actin, can interact with or impact STIM1 function (27,(35)(36)(37). Previous studies have shown that the microtubule plus-end tracking protein EB1 links STIM1 to the microtubule network and is likely involved in reorganizing the ER (27,36). Therefore, it is possible that E2/ESR1 was impacting EB1 function and not STIM1. However, our FRAP studies indicated that E2/ESR1 had no effect on EB1 mobility, suggesting a more direct impact on STIM1 (Fig.  6). Nevertheless, it is possible that E2 could affect interactions between STIM1 and the cytoskeleton to inhibit mobility under resting conditions.

JOURNAL OF BIOLOGICAL CHEMISTRY 33515
E2/ESR1 inhibited basal phosphorylation of STIM1 on Ser-575, which occurred rapidly and was long lasting (Fig. 7, A-F). Furthermore, mutation of this residue to an alanine significantly decreased STIM1 mobility (Fig. 7, G and H), replicating the effects of E2/ESR1. However, we cannot rule out the possibility that other serine phosphorylation sites are being affected by E2/ESR1, but may not be detectable with the mpm-2 antibody. Furthermore, the exact mechanism utilized by E2/ESR1 to FIGURE 7. E2 inhibits phosphorylation of STIM1 serine residues. A, E2 rapidly inhibits STIM1 serine phosphorylation. YFP-STIM1 was immunoprecipitated (IP) from HEK293T cells co-expressing ESR1 before and after short term E2 exposure. Membranes were probed with the mpm-2 antibody that specifically recognizes phosphorylated serines. The blot was stripped and re-probed with a GFP antibody to reveal total immunoprecipitated STIM1. IB, immunoblotting. B, relative density of STIM1 phosphorylation for each time point, all n ϭ 3. Relative density for phosphorylated STIM1 was calculated by normalizing to total immunoprecipitated STIM1 and comparing it with WT STIM1 under basal conditions. C, long lasting inhibition of STIM1 serine phosphorylation. YFP-STIM1 was immunoprecipitated from HEK293T cells co-expressing ESR1 before and after long term E2 exposure. D, relative density of STIM1 phosphorylation for each time point, all n ϭ 4. E, WT YFP-STIM1, STIM1 570STOP , and STIM1 S575A immunoprecipitated from HEK293T cells co-expressing ESR1. F, relative density of STIM1 phosphorylation for each STIM1 construct under basal conditions or after E2 treatment. YFP-STIM1, n ϭ 11; STIM1 570STOP , n ϭ 4; STIM1 S575A , n ϭ 6. G, mean recovery of STIM1 S575A into the bleached area following FRAP. Basal, n ϭ 19; thapsigargin, n ϭ 12; E2, n ϭ 11; E2 ϩ thapsigargin, n ϭ 10. H, mobile fraction of STIM1 S575A after photobleaching (black bars) compared with WT STIM1 (white bars). All FRAP data were collected concurrently. *, significant difference of p Ͻ 0.05; **, significant difference of p Ͻ 0.001(one-way ANOVA followed by Tukey's multiple comparison test). Data are shown as mean Ϯ S.E. inhibit serine phosphorylation remains unknown. Estrogen receptors have typically been described as regulators of gene transcription (38). However, it is well established that estrogen receptors can also signal nongenomically (38,39). For example, signaling through ESR1 has been shown to impact the PI3K/ Akt and ERK1/2 cascades (39,40). Conversely, E2/ESR1 may activate a phosphatase to dephosphorylate STIM1.
Gender is a risk factor for most lung diseases, including asthma, CF, and COPD. Bronchoconstriction (41), mast cell degranulation (42), and mucus hypersecretion (43) are characteristics of asthma that are Ca 2ϩ -regulated, and E2/ESR1 has been shown to act as a bronchodilator in human airway smooth muscle cells by decreasing intracellular Ca 2ϩ levels (44). This suggests that E2 is airway-protective in asthma. However, there are many other characteristics of asthma that may be affected by E2/ESR1, including airway smooth muscle hyperplasia and hypertrophy, goblet cell hyperplasia, and inflammation. COPD also disproportionately affects women, and 80% of early onset COPD patients are female (13). Characteristics of COPD include abnormal ion transport, goblet cell hyperplasia, and mucus dehydration (45). Women with CF are also known to suffer more severely than their male counterparts (11,12,46,47), have a higher mortality (11,47), and are more likely to experience acute exacerbations when E2 is highest during the periovulatory period (12,46). Sex differences such as those seen in asthma, COPD, and CF are likely to be very complex. However, E2/ESR1-mediated inhibition of STIM1 may play a role in affecting the severity of all of these diseases. In the case of bronchoconstriction, reduced Ca 2ϩ levels may be protective. Conversely, in the case of altered ion transport, limiting CaCC activity by decreasing Ca 2ϩ may further impair the ability to hydrate the airways and so may exacerbate the disease in females (9).
In conclusion, we have demonstrated that E2/ESR1 can inhibit SOCE by regulating STIM1 function. Under normal conditions, phosphorylated STIM1 can sense ER Ca 2ϩ depletion, oligomerize, and activate Orai1 (Fig. 8A). However, we propose that acute E2 exposure activates ESR1, which signals nongenomically, to inhibit basal phosphorylation of STIM1 (Fig. 8B). Consequently, STIM1 is unable to oligomerize after ER Ca 2ϩ depletion and fails to activate Orai1, leading to a decrease in SOCE. A better understanding of this process may lead to the development of novel therapeutic strategies to reverse the sex bias in chronic lung disease.