Overexpression of Orai1 and STIM1 Proteins Alters Regulation of Store-operated Ca2+ Entry by Endogenous Mediators*

Background: Store-operated Ca2+ entry (SOCE) is essential for cell function. Results: We discovered cross-talk between expression of molecules that determine SOCE and demonstrated that the role of endogenous mediators may be altered in overexpressed system. Conclusion: iPLA2β is an important regulator of endogenous SOCE, but its role can be obscured by Orai1 and STIM1 overexpression. Significance: Mediators of endogenous SOCE are important for Ca2+ homeostasis in health and disease. Orai1 and STIM1 have been identified as the main determinants of the store-operated Ca2+ entry (SOCE). Their specific roles in SOCE and their molecular interactions have been studied extensively following heterologous overexpression or molecular knockdown and extrapolated to the endogenous processes in naïve cells. Using molecular and imaging techniques, we found that variation of expression levels of Orai1 or STIM1 can significantly alter expression and role of some endogenous regulators of SOCE. Although functional inhibition of Ca2+-independent phospholipase A2 β (iPLA2β or PLA2g6A), or depletion of plasma membrane cholesterol caused a dramatic loss of endogenous SOCE in HEK293 cells, these effects were attenuated significantly when either Orai1 or STIM1 were overexpressed. Molecular knockdown of iPLA2β impaired SOCE in both control cells and cells overexpressing STIM1. We also discovered important cross-talk between expression of Orai1 and a specific plasma membrane variant of iPLA2β but not STIM1. These data confirm the role of iPLA2β as an essential mediator of endogenous SOCE and demonstrate that its physiological role can be obscured by Orai1 and STIM1 overexpression.


Orai1 and STIM1 have been identified as the main determinants of the store-operated Ca 2؉ entry (SOCE). Their specific roles in SOCE and their molecular interactions have been studied extensively following heterologous overexpression or molecular knockdown and extrapolated to the endogenous processes in naïve cells. Using molecular and imaging techniques, we found that variation of expression levels of Orai1 or STIM1
can significantly alter expression and role of some endogenous regulators of SOCE. Although functional inhibition of Ca 2؉independent phospholipase A 2 ␤ (iPLA 2 ␤ or PLA2g6A), or depletion of plasma membrane cholesterol caused a dramatic loss of endogenous SOCE in HEK293 cells, these effects were attenuated significantly when either Orai1 or STIM1 were overexpressed. Molecular knockdown of iPLA 2 ␤ impaired SOCE in both control cells and cells overexpressing STIM1. We also discovered important cross-talk between expression of Orai1 and a specific plasma membrane variant of iPLA 2 ␤ but not STIM1. These data confirm the role of iPLA 2 ␤ as an essential mediator of endogenous SOCE and demonstrate that its physiological role can be obscured by Orai1 and STIM1 overexpression.
Endogenous store-operated Ca 2ϩ entry (SOCE) 3 has fundamental importance for the vast majority of eukaryotic cells: it ensures timely refilling of ER stores, which is essential for functional integrity and ultimately cell survival (for review, see Refs. 1 and 2). A major breakthrough in understanding SOCE mechanism was triggered by identification of STIM1 (3,4) as a Ca 2ϩ sensor in the ER and Orai1 (5,6) as the plasma membrane channel (historically called CRAC, or CA 2ϩ release-activated CRAC CA 2ϩ channel (1)) that responds to depletion of Ca 2ϩ stores and mediates endogenous SOCE. Molecular down-regulation of either component leads to effective inhibition of SOCE in a wide variety of cell types, and genetic deficiency in either Orai1 or STIM1 was shown to result in major pathologies in mice and humans (for review, see Ref. 7). However, it is important to mention that the cells lacking either Orai1 or STIM1 still remain viable and do not lose the ability to proliferate and sustain most of their functions.
The mechanism of Orai1 and STIM1 involvement in SOCE has been studied extensively in heterologous overexpression systems (for recent reviews, see Refs. 2 and 8 -12). Rigorous studies of overexpressed Orai1 and STIM1 (11)(12)(13)(14)(15)(16)(17)(18)(19), together with in vitro experiments with purified proteins, resulted in the conclusion that direct conformational coupling of STIM1 to Orai1 can be required and is sufficient for SOCE activation in heterologous systems. However, these experiments could not tell much about what else is required for signal transduction from STIM1 to Orai1 in naïve cells when both molecules are expressed at the normal endogenous levels, which are significantly lower than that in heterologous systems. Thus, the question remains open on how closely endogenous SOCE mechanism may be represented by SOCE created in heterologous systems, and if there are any features of endogenous SOCE pathway that may be lost when Orai1 and/or STIM1 are overexpressed. This question is particularly important as additional molecules and mechanisms have been shown to participate intimately in endogenous SOCE (20), but their role in SOCE was questioned in view of a direct conformational coupling model established for overexpressed Orai1 and STIM1.
One such essential component is Ca 2ϩ -independent phospholipase A2 (iPLA 2 ␤ or PLA2G6A). Its important role in endogenous SOCE was demonstrated in vascular smooth muscle cells (24 -28), platelets (21), Jurkat T lymphocytes (21), RBL-2H3 cells (29,30), neuroblastoma/glioma cells (22), as well as in astrocytes (23), keratinocytes (24), skeletal muscle cells (25), fibroblasts (26), prostate cancer cells (27), and endothelial cells (28). In all of these studies, inhibition of the catalytic activity or molecular knockdown of iPLA 2 ␤ caused dramatic impairment of endogenous SOCE, leaving little doubt about its important role in the SOCE process (for review, see Ref. 20). Importantly, results of a genetic screen of Drosophila melanogaster (5), which identified STIM1 and Orai1 as essential components of SOCE, also picked up an ortholog of iPLA 2 ␤ encoded by the CG6718 gene, which is highly homologous to the human PLA2G6 (up to 85% in the main structural domains). Knockdown of the iPLA 2 ␤ ortholog (CG6718) showed a significant impact on SOCE activation, which was identical to that of STIM1 (see supplemental data in Ref. 5). This result clearly confirmed the role of iPLA 2 ␤ as one of the major molecular determinants of endogenous SOCE. However, the importance of iPLA 2 ␤ for SOCE was questioned by the finding that functional inhibition of iPLA 2 ␤ with bromoenole lactone (BEL, a suicidal inhibitor of catalytic activity) failed to inhibit CRAC current in HEK293 cells overexpressing Orai1 and STIM1 (29). This result pointed to the possibility that overexpressed Orai1 and STIM1 create SOCE that may not fully reflect the properties of the endogenous process.
The current study presents new evidence that some properties of endogenous SOCE indeed change in response to overexpression of STIM1 and Orai1. We also tested a novel idea that there may be a cross-talk between expression levels of the major molecular determinants of endogenous SOCE and found that expression of endogenous Orai1 determines and depends upon expression levels of a specific plasma membrane variant of iPLA 2 ␤ but not STIM1. These findings highlight the complexity of the SOCE mechanism and the role of its endogenous mediators, which may be lost in overexpression systems.

EXPERIMENTAL PROCEDURES
Cells and Transfections-Human embryonic kidney 293 (HEK293) cells were obtained from ATCC and grown in Dulbecco's modified Eagle's medium containing 4.5 mg/ml glucose, supplemented with 10% heat-inactivated fetal bovine serum (Invitrogen), 10 units/ml penicillin and 10 mg/ml streptomycin at 37°C, in an atmosphere of 5% CO 2 .
HEK293 cells were transiently transfected with plasmid DNA using jetPEI (PolyPlus) or with siRNA using Lipofectamine 2000 (Invitrogen). In experiments that required down-regulation of one protein along with overexpression of another protein, Lipofectamine 2000 was used for co-transfection of siRNA and plasmid DNA. Transfection factors were used according to the protocols suggested by the manufacturers. The standard transfection rate was ϳ70 -90%, 24 h after transfection.
DNA Constructs-Constructs encoding human Orai1 (Gen-Bank TM accession no. NM032790) and STIM1 (GenBank TM accession no. NM003156) tagged with green fluorescent protein (EGFP) were described previously (30). Briefly, cDNAs of Orai1 and STIM1 were amplified by PCR and cloned in-frame into the pEGFP-N1 and/or pmCherry-N1 vectors (Clontech) using BglII and EcoRI restriction sites.
Total RNA was purified from HEK293 cells with a High Pure RNA isolation kit (Roche Applied Science), and reverse transcription to cDNA was performed using a high capacity RNA-to-cDNA kit (Applied Biosystems). The concentration of RNA during isolation was determined by a NanoDrop ND-1000 spectrophotometer (Thermo Scientific).
Expression of iPLA 2 ␤-L, iPLA 2 ␤-S, STIM1, and Orai1 was determined by quantitative real-time PCR with a StepOnePlus real-time PCR system (Applied Biosystems) with the following TaqMan gene expression assays (Applied Biosystems): Hs00895670_m1 (iPLA2␤-S), Hs00899715_m1 (iPLA2␤-L), Hs00963373_m1 (STIM1), Hs00385627_m1 (Orai1), and 4333767F (human glucouronidase ␤, GUSB). Samples for the analysis were collected before (0 h), 24, 48, and 72 h after transfection. mRNA levels of housekeeping gene (GUSB) were always determined for each target in the same experiment. Data were analyzed with StepOne Software (version 2.0) using the relative standard curve method and normalized to the GUSB and control samples (cells transfected with scrambled siRNA). Data are presented as an average (S.E.) of three to four independent experiments.
Ca 2ϩ Influx Studies-HEK293 cells were grown on glass coverslips and loaded with fura2/AM (Invitrogen), and changes in intracellular Ca 2ϩ (measured as F 340 /F 380 ratio) were monitored as described previously (30). Briefly, a dual-excitation fluorescence imaging system (Intracellular Imaging) was used for studies of individual cells. The changes in intracellular Ca 2ϩ were expressed as ⌬ratio, which was calculated as the difference between the peak values of the ratio after extracellular Ca 2ϩ was added and ratio right before Ca 2ϩ addition. After transfection, the cells were plated on the coverslips and kept in culture. 72 h after transfection, Ca 2ϩ was recorded simultaneously from 20 -40 cells from each coverslip, which after transfection represented a mixed population of GFP-positive (transfected) and GFP-negative (non-transfected) cells. To determine the effect of (S)-BEL and MCD on Ca 2ϩ influx, the cells on each coverslip were pretreated with 50 M (S)-BEL or 10 mM MCD for 30 min at 37°C, and the drugs were washed away prior to the Ca 2ϩ experiment. Data were summarized from the large number of individual cells (as indicated on the bar graphs): 10 -40 cells were analyzed in each single run (representative traces are shown in Figs. 1 and 3), repeated in three to five independent experiments for each condition using at least three different cell transfections. In each experiment, GFP-positive (transfected) and GFP-negative (non-transfected) cells were analyzed and presented separately.

RESULTS
Functional Inhibition of iPLA 2 ␤ Blocks SOCE in Naïve Cells, but Not in Cells Overexpressing Orai1 or STIM1-To compare the role of iPLA 2 ␤ in endogenous SOCE in naïve HEK293 cells with its role in SOCE mediated by overexpressed Orai1 or STIM1, we started with testing the effects of inhibition of catalytic activity of iPLA 2 ␤ on SOCE evoked by either buffering free Ca 2ϩ in ER with TPEN (see Fig. 1) or by depleting ER stores as a result of SERCA inhibition with thapsigargin (TG, Fig. 2). As in our earlier studies (24,25), inhibition of iPLA 2 ␤ activity was achieved by pretreatment of the cells with (S)-BEL (31).
Similarly to a wide variety of cells (for review, see Ref. 20), pretreatment of HEK293 with (S)-BEL (50 M for 30 min) resulted in 86.6 Ϯ 1.9% and 83.8 Ϯ 8.2% inhibition of endogenous SOCE induced by TPEN (1 mM for 5 min) and TG (5 M for 5 min), respectively. Surprisingly, we noticed that when either Orai1 GFP or STIM1 GFP was overexpressed in HEK293 cells, treatment with (S)-BEL produced virtually no effect. To assess this phenomenon in detail and to exclude the possibility that mere transfection of the cells may change their sensitivity to (S)-BEL, we systematically investigated effects of (S)-BEL on SOCE in HEK293 cells transfected with either GFP or Orai1 GFP or STIM1 GFP . In each cell preparation, we simultaneously recorded Ca 2ϩ responses in transfected (GFP-positive) and non-transfected (GFP-negative) cells. We found that transfection of the cells with plain GFP did not change SOCE sensitivity to (S)-BEL, and Fig. 1A shows that TPEN-induced SOCE was inhibited by ϳ80% in both GFP-positive and GFP-negative cells. In contrast, in cells overexpressing either Orai1 GFP (Fig.  1B) or STIM1 GFP (Fig. 1C), (S)-BEL failed to inhibit SOCE in GFP-positive (transfected) cells. Importantly SOCE in adjacent GFP-negative cells (that did not express Orai1 GFP or STIM GFP ) was inhibited, respectively, by 85.3 Ϯ 3.3% and 82.5 Ϯ 1.9%, which was similar to what was observed in naïve cells (Fig. 1A). Simultaneous study of control cells and cells overexpressing Orai1 GFP ensured that both cell types were treated and investigated under the same conditions, and yet, endogenous SOCE in control cells was inhibited, whereas SOCE in Orai1/STIM1overexpressing cells remained untouched.
Similar results were obtained when SOCE was triggered by TG, as demonstrated in Fig. 2. In HEK293 cells overexpressing GFP, pretreatment with (S)-BEL inhibited SOCE by 84.4 Ϯ 3.7% ( Fig. 2A), whereas in cells overexpressing Orai1 GFP or STIM1 GFP (Fig. 2, B and C, respectively), it had no effect. Again, in GFP-negative cells adjacent to those expressing Orai1 GFP or STIM1 GFP , pretreatment with (S)-BEL resulted in profound inhibition of TG-induced SOCE: 68.5 Ϯ 3.9% for Orai1 GFPnegative and 76.9 Ϯ 4.0% for STIM1 GFP -negative cells (data not shown). Moreover, pretreatment with (S)-BEL had no visible effect on the dynamics of overexpressed STIM1 mCherry and Orai1 GFP , which formed puncta even in the cells in which catalytic activity of iPLA 2 ␤ was inhibited (Fig. 3).

Depletion of Cholesterol with ␤-Methylcyclodextran Impairs Endogenous SOCE in Naïve Cells, but Has No Effect on SOCE in
Cells Overexpressing Orai1 or STIM1-To determine whether overexpression of Orai1/STIM1 impairs only SOCE dependence on functional activity of iPLA 2 ␤ or may change some other features of endogenous SOCE, we have tested SOCE dependence on plasma membrane cholesterol abundance. ␤-Methyl-cyclodextran (␤-MCD) is used widely to deplete plasma membrane cholesterol and was shown to inhibit SOCE in RBL-2H3 cells (32). Fig. 4 shows that incubation of GFP-transfected HEK293 cells with ␤-MCD inhibited TG-induced SOCE (Fig.  4A) in both GFP-positive (by 81.5 Ϯ 5.3%) and GFP-negative cells (by 93.5 Ϯ 1.4%). However, the effect of MCD was abolished completely by overexpression of either Orai1 GFP (Fig. 4B) or STIM1 GFP (Fig. 4C). Importantly, SOCE in GFP-negative adjacent cells was inhibited by 83.9 Ϯ 4.9% (Orai1 GFP , Fig. 4B) and 65.9 Ϯ 4.4% (STIM1 GFP , Fig. 4C). These results indicate that the overexpression of either Orai1 or STIM1 significantly changes the properties of endogenous SOCE and can make it  resistant to inhibitors of different cellular processes that are essential components of endogenous SOCE.
Down-regulation of iPLA 2 ␤ Protein Inhibits Not Only Endogenous SOCE, but also SOCE in STIM1-overexpressing Cells-Our earlier studies showed that in the endogenous SOCE pathway, iPLA 2 ␤ is required downstream from STIM1, so the results of the present study raised a question as to whether overexpression of STIM1 may overcome only an acute catalytic deficiency of iPLA 2 ␤ (caused by (S)-BEL) or may create a shortcut in SOCE pathway that may function even in the longer term absence of the iPLA 2 ␤ protein. To answer that question, we used an siRNA approach to knockdown iPLA 2 ␤ protein and compared its effects on endogenous SOCE in control HEK293 (at physiological levels of STIM1 expression) with its effects on SOCE in cells overexpressing STIM1 GFP .
Transfection of HEK293 cells with siRNA targeting all splice variants of iPLA 2 ␤ caused up to an 88 Ϯ 3% decrease in mRNA levels of iPLA 2 ␤ at 72 hours after transfection, followed by only ϳ30% loss of protein (supplemental Fig. 1). Fig. 5A shows that even such incomplete down-regulation of iPLA 2 ␤ caused a 52 Ϯ 6% reduction in TG-stimulated Ca 2ϩ entry. Importantly, significant (33 Ϯ 5%) inhibition of SOCE was also observed in siRNA-treated cells overexpressing STIM1 GFP (Fig. 5B). These results suggest that even partial deficiency in the iPLA 2 ␤ protein can impair not only endogenous SOCE but also SOCE in the cells in which STIM1 is overexpressed. The reasons for the different sensitivity of SOCE to acute loss of catalytic activity versus prolonged siRNA-induced deficiency in iPLA 2 ␤ protein in STIM1-overexpressing cells are yet to be determined.
Cross-talk between Expression Levels of Plasma Membrane Variant of iPLA 2 ␤ and Orai1-To investigate whether there may be cross-talk between expression of different components of the SOCE mechanism, mRNA levels of endogenous iPLA 2 ␤, Orai1, and/or STIM1 were assessed in resting HEK293 cells after each of these molecules was knocked down using specific siRNA. Fig. 6A shows that siRNA-induced knockdown of both splice variants of iPLA 2 ␤ (plasma membrane-associated iPLA 2 ␤(L), and cytosolic iPLA 2 ␤(S)) caused progressive increase in mRNA levels of both Orai1 and STIM1: when mRNA of iPLA 2 ␤ was down by 88 Ϯ 3%, mRNAs for Orai1 and STIM1 increased 2-fold.
Strikingly, knockdown of Orai1 (Fig. 6B) dramatically boosted expression levels of the long variant of iPLA 2 ␤, with no significant change in mRNA levels of either STIM1, or the short variant of iPLA 2 ␤: 72 h after cell transfection with siRNA specific to Orai1, mRNA levels of iPLA 2 ␤(L) increased Ͼ4-fold. Molecular knockdown of STIM1 (Fig. 6C) did not cause any significant changes in mRNA levels of Orai1 or either variant of iPLA 2 ␤. These results for the first time demonstrate strong

DISCUSSION
This study presented the first evidence for cross-talk between expression levels of the major molecular determinants of endogenous SOCE and demonstrated that overexpression of Orai1 and STIM1 may change SOCE regulation by endogenous mediators. The important new finding is the existence of an inverse correlation between the expression of endogenous Orai1 and the specific plasma membrane variant of iPLA 2 ␤. This further supports the idea of a close functional relationship between these two components of the endogenous SOCE mechanism. Here, we demonstrated that molecular down-regulation of Orai1 leads to a dramatic 4-fold increase in expression of the plasma membrane but not cytosolic variant of iPLA 2 ␤ and not STIM1. It is tempting to hypothesize that upregulation of one specific variant of iPLA 2 ␤ could be a defensive response of the cell to the loss of significant amount of Orai1 channels. Indeed, we have earlier shown (33,34) that mere activation of plasma membrane-associated iPLA 2 ␤ and/or enrichment of the plasma membrane with its lysophospholipid products can activate Orai1 channels and SOCE bypassing Ca 2ϩ store depletion. Boosting the expression or functional activity of endogenous iPLA 2 ␤ may help sustain activity of endogenous Orai1 channels and help prevent a potentially lethal loss of Orai1-mediated Ca 2ϩ entry, thus ensuring cell survival when Orai1 expression is down-regulated.
The absence of any detectable cross-talk between expression levels of Orai1 and STIM1 was rather unexpected and warrants further investigation. One may anticipate at least some interdependence of these two major determinants of SOCE that have been shown to directly associate with each other. Because STIM1 is essential for initiation of the signaling cascade leading to SOCE activation, it was surprising to find that knockdown of STIM1 had no effect on either Orai1 or iPLA 2 ␤ expression. However, it is important to mention that there is a precedent in a neuroblastoma/glioma cell line (NG115) in which STIM1 is absent, and despite the fact that TG fails to produce any significant SOCE, NG115 cells are able to grow and proliferate (22). These data suggest that endogenous activation of Orai1 by iPLA 2 ␤ or some other yet to be determined regulator(s) may be sufficient for resting cell survival. This does not contradict the fact that molecular knock-down of either Orai1 or STIM1 or  iPLA 2 ␤ are all known to dramatically impair SOCE activated by TG or other stimuli. One may speculate that significant depletion of the stores (induced naturally by agonists or experimentally by TG) may require a full activation of the SOCE mechanism to refill empty stores: deficiency in either component of SOCE will significantly affect its maximum capacity and can impair agonist-induced SOCE and related cell function. In contrast, full capacity of all SOCE components may not be needed for sustaining background SOCE in resting cells, and cross-talk of endogenous Orai1 and iPLA 2 ␤ on expression levels may become especially important for cell adaptation and function under resting conditions.
Although direct coupling of ER-resident STIM1 to plasma membrane-resident Orai1 may be rightfully considered as the straightforward mechanism for signal transduction, there is a significant body of evidence for the presence of additional mediators and regulators of the endogenous SOCE mechanism (for review, see Ref. 20). One of them is iPLA 2 ␤, which was found to be functionally required downstream from STIM1 and preceding Orai1. It is important to emphasize that existence of additional steps in SOCE mechanism, and molecular intermediate(s) between STIM1 and Orai1 do not contradict the studies that demonstrated their spatial proximity and eventual conformational coupling. Indeed, endogenous mediators may be simply needed to set up a stage that will allow such interactions. In contrast to overexpression systems, in naïve cells under physiological conditions, there are a limited number of Orai1 channels in the plasma membrane and ER Ca 2ϩ sensors, and endogenous mediators (similar to iPLA 2 ␤, lysophospholipids, and other factors) may facilitate their targeting to specific areas where the ER comes into close proximity with the plasma membrane. Our results suggest that this process may be also dependent on plasma membrane cholesterol content, as its disturbance, which in our studies was achieved by cholesterol depletion by MCD, significantly impaired endogenous SOCE but not in cells overexpressing Orai1 or STIM1. This result is fully in line with an earlier report (35) of the lack of an effect of MCD on the whole-cell Orai1-mediated current in STIM1overexpressing cells. The absence of the MCD effect in this case was suggested to be a result of MCD-induced changes in membrane potential. However, the striking differences in MCD effects on SOCE in intact naïve and overexpressing cells (Fig. 4) found in this study do not seem to favor such an explanation. Indeed, naïve and overexpressing cells in our study have been treated with MCD in the same way and studied simultaneously, so, if membrane potential is a factor in the MCD effects, one would expect the MCD effects to be the same in naïve and overexpressing cells. The dramatically different effects of MCD and (S)-BEL in naïve and overexpressing cells may favor the idea that overexpression of Orai1 in the plasma membrane or STIM1 in the ER can saturate the SOCE pathway and eliminate the need for mediating and coordinating their proper targeting toward each other. The results of our studies, together with some earlier reports (36,37) clearly demonstrate that overexpression of Orai1, and/or STIM1 may significantly alter SOCE sensitivity to pharmacological inhibitors or molecular manipulations that are known to impair endogenous SOCE in naïve cells.
Discovery that molecular down-regulation of iPLA 2 ␤ leads to a 2-fold increase in expression of Orai1 and STIM1, strongly suggests that endogenous up-regulation of Orai1 and STIM1 expression may be needed to compensate for the loss of iPLA 2 ␤ as an endogenous regulator/mediator that is important for signal transduction from STIM1 to Orai1. This idea is further supported by the striking differences we found in SOCE in naïve and overexpression conditions. Indeed, experimental overexpression of Orai1 and STIM1 may create special conditions for their direct interaction/conformational coupling as described in heterologous systems, in which the need for functional activity of iPLA 2 ␤ as a mediator of signal transduction may be reduced significantly or lost completely. Thus, despite the great importance of heterologous studies, overexpression of major SOCE components may obscure or alter some endogenous processes that are important for the native SOCE mechanism.
In summary, results of these studies demonstrated that there is cross-talk between expression levels of endogenous Orai1 and iPLA 2 ␤ and that some properties of endogenous SOCE may change in response to experimental overexpression of STIM1 and Orai1. Our findings highlight the complexity of the endogenous SOCE mechanism and the important role of iPLA 2 ␤ and plasma membrane cholesterol in endogenous signal transduction, which can be obscured by overexpression of Orai1 or STIM1.