Ca2+-Calmodulin-dependent Protein Kinase II Potentiates Store-operated Ca2+ Current*

A rise in intracellular Ca2+ (Ca2+i) mediates various cellular functions ranging from fertilization to gene expression. A ubiquitous Ca2+ influx pathway that contributes significantly to the generation of Ca2+i signals, especially in non-excitable cells, is store-operated Ca2+ entry (SOCE). Consequently, the modulation of SOCE current affects Ca2+i dynamics and thus the ensuing cellular response. Therefore, it is important to define the mechanisms that regulate SOCE. Here we show that a rise in Ca2+i potentiates SOCE. This potentiation is mediated by Ca2+-calmodulin-dependent protein kinase II (CaMKII), because inhibition of endogenous CaMKII activity abrogates Ca2+i-mediated SOCE potentiation and expression of constitutively active CaMKII potentiates SOCE current independently of Ca2+i. Moreover, we present evidence that CaMKII potentiates SOCE by altering SOCE channel gating. The regulation of SOCE by CaMKII defines a novel modulatory mechanism of SOCE with important physiological consequences.

A rise in intracellular Ca 2؉ (Ca 2؉ i ) mediates various cellular functions ranging from fertilization to gene expression. A ubiquitous Ca 2؉ influx pathway that contributes significantly to the generation of Ca 2؉ i signals, especially in non-excitable cells, is store-operated Ca 2؉ entry (SOCE). Consequently, the modulation of SOCE current affects Ca 2؉ i dynamics and thus the ensuing cellular response. Therefore, it is important to define the mechanisms that regulate SOCE. Here we show that a rise in Ca 2؉ i potentiates SOCE. This potentiation is mediated by Ca 2؉ -calmodulin-dependent protein kinase II (CaMKII), because inhibition of endogenous CaMKII activity abrogates Ca 2؉ i -mediated SOCE potentiation and expression of constitutively active CaMKII potentiates SOCE current independently of Ca 2؉ i . Moreover, we present evidence that CaMKII potentiates SOCE by altering SOCE channel gating. The regulation of SOCE by CaMKII defines a novel modulatory mechanism of SOCE with important physiological consequences.
A rise in intracellular Ca 2ϩ (Ca 2ϩ i ) 1 mediates various cellular responses ranging from exocytosis to cell death, often in the same cell. This is possible because of the specificity encoded in the spatial, temporal, and amplitude features of Ca 2ϩ signals, which are generated because of either Ca 2ϩ release from intracellular stores or Ca 2ϩ influx from the extracellular space. The interplay between these two Ca 2ϩ sources results in specific Ca 2ϩ i dynamics, which lead to a particular cellular response. A ubiquitous Ca 2ϩ influx pathway activated downstream of phospholipase-coupled receptors is store-operated Ca 2ϩ entry (SOCE) (1). SOCE is important for various cellular functions including store refilling (2), regulation of exocytosis (3), sperm capacitation (4), and T-cell activation (5,6). Because SOCE contributes significantly to the generation of Ca 2ϩ i dynamics, it is important in specifying the ensuing cellular response. SOCE is defined as Ca 2ϩ influx activated in response to intracellular Ca 2ϩ store depletion. Although, store depletion is the primary required signal for SOCE activation, additional distinct mechanisms modulate SOCE activity (1). Characterizing such SOCE modulators is crucial to understanding the mechanisms by which SOCE contributes to defining specific cellular responses. In this paper, we show that both Ca 2ϩ i and Ca 2ϩ -calmodulindependent protein kinase II (CaMKII) positively modulate SOCE current (I SOC ).
Both intracellular and extracellular Ca 2ϩ have been shown to modulate SOCE activity in a complex fashion. This is best characterized for the Ca 2ϩ release-activated Ca 2ϩ current (I CRAC ) (7), which is the first described SOCE current. Hallmarks of I CRAC include inward rectification, an ionic selectivity sequence of Ca 2ϩ Ͼ Ba 2ϩ Ͼ Mn 2ϩ Ͼ Ͼ Na ϩ , and a small unitary Ca 2ϩ conductance estimated to be ϳ20 femtoSiemens (8,9). Intracellular Ca 2ϩ has been shown to negatively regulate I SOC by two mechanisms. 1) Fast Ca 2ϩ -dependent inactivation is due to inhibition of I SOC , arguably following Ca 2ϩ binding to an intracellular site (8,10). 2) Slow Ca 2ϩ -dependent inactivation is due in part to store refilling but also has a store-independent component that is poorly understood (11). In contrast, extracellular Ca 2ϩ (Ca 2ϩ o ) potentiates I SOC through a process termed Ca 2ϩ -dependent potentiation (CDP). CDP is attributed to a positive effect of extracellular Ca 2ϩ on I SOC (12, 13) and probably results from Ca 2ϩ binding to an extracellular site on the channel because it can be replicated by Ni 2ϩ (12). Ni 2ϩ does not permeate SOCE channels but potentiates the Ca 2ϩ current through SOCE channels (12). Furthermore, CDP alters channel gating and not permeation (12), because current inactivation (which is dependent on permeation (8)) remains unchanged during CDP.
In this paper, we describe a novel positive regulation of I SOC by intracellular Ca 2ϩ through CaMKII in Xenopus oocytes. Xenopus oocyte SOCE current has similar characteristics to I CRAC including high Ca 2ϩ selectivity, inward rectification (14), and as shown here extracellular Ca 2ϩ -dependent potentiation. Allowing a Ca 2ϩ i rise during SOCE activation results in larger I SOC . Furthermore, expression of a constitutively active CaMKII (CaMKII ca ) also leads to enhancement of I SOC . CaMKII potentiates I SOC by dramatically increasing the levels of CDP. Because CDP affects SOCE channel gating, our data argue that CaMKII potentiates SOCE by altering gating. CaMKII-mediated potentiation of I SOC has important physiological consequences because store depletion is physiologically accompanied by a rise in Ca 2ϩ i . Because CaMKII is able to decode Ca 2ϩ i dynamics into different levels of enzyme activity (15), CaMKII-mediated potentiation of I SOC provides a mechanism for Ca 2ϩ to regulate its own entry into the cell depending on the levels and kinetics of Ca 2ϩ release following receptor activation.

EXPERIMENTAL PROCEDURES
Oocyte and Electrophysiological Methods-Xenopus laevis oocytes were prepared as described previously (16). SOCE was activated by the depletion of intracellular Ca 2ϩ stores with either ionomycin (10 M) or thapsigargin (1 M for Ն3 h unless otherwise indicated). Oocytes were injected with 7 nmol of BAPTA to buffer the Ca 2ϩ i rise and block the endogenous Ca 2ϩ -activated Cl Ϫ current (I Cl,Ca ) that would otherwise mask the SOCE current. Assuming an oocyte volume of 1 l, this would result in a final concentration of ϳ7 mM BAPTA that completely blocks I Cl,Ca . Oocytes were voltage-clamped with two microelectrodes by the use of a GeneClamp 500 (Axon Instruments). Electrodes were filled with 3 M KCl and had resistances of 0.5-2 megaohms. Voltage stimulation and data acquisition were controlled using a pClamp8 (Axon Instruments). Current data were filtered at 10 kHz, digitized, and analyzed using Clampfit9.0 (Axon Instruments) and Origin® software (Microcal Software Inc.). I SOC was typically measured in 30 Ca 2ϩ solution (in mM: 55 NaCl, 30 CaCl 2 , 10 Hepes, pH 7.4), and Ca 2ϩ -free solution in I SOC -recording experiments was 70 Mg 2ϩ (in mM: 70 MgCl 2 , 10 Hepes, pH 7.4). I SOC time course is plotted as the mean current 30 -35 ms after the voltage step, and the leak current at the beginning of the experiment was subtracted.
Ca 2ϩ -activated Cl Ϫ currents were recorded in Ringer solution (in mM: 123 NaCl, 2.5 KCl, 1.8 CaCl 2 , 18 MgCl 2 , 10 Hepes, pH 7.4) or Ca 2ϩ -free Ringer solution that had the same composition as Ringer solution with the exception that CaCl 2 was omitted, MgCl 2 increased to 5 mM, and 0.1 mM EGTA was added.
cRNA for injection into oocytes was transcribed in vitro using the mMessage mMachine SP6 transcription kit (Ambion). Both pGEM3-5HT1c and pSP-CaMKII(T286D) were linearized with EcoRI and transcribed with SP6 polymerase.
Ca 2ϩ Imaging-Xenopus oocytes were injected with ϳ7.6 M Ca 2ϩ -Green-1 coupled to 70 kDa of dextran and voltage-clamped as described above. The dye was allowed at least 30 min to equilibrate within the oocyte. Imaging experiments were performed in Ringer and Ca 2ϩ -free Ringer solutions. Confocal Ca 2ϩ imaging was performed using an Olympus Fluoview confocal scanning system fitted to IX70 microscope using a ϫ10 (0.3 numerical aperture) objective. Images (256 ϫ 256 pixels) were collected and analyzed using Olympus Fluoview software.

Intracellular Ca 2ϩ Potentiates SOCE-Physiological activation of SOCE is invariably preceded by a rise in Ca 2ϩ
i because of Ca 2ϩ release from stores. Although it is clear that Ca 2ϩ i is not required for SOCE activation, it is not known whether this Ca 2ϩ i modulates I SOC . To address this issue, we activated SOCE under conditions that either allow an intracellular Ca 2ϩ rise or not. To activate SOCE in the absence of a Ca 2ϩ i rise, we injected cells with BAPTA and then depleted Ca 2ϩ stores with ionomycin (BAPTA-Ion). This results in the activation of a characteristic SOCE current as previously described (Fig. 1B, filled squares) (14,17). To allow a Ca 2ϩ i rise during SOCE activation, Ca 2ϩ stores were depleted with ionomycin followed by repetitive hyperpolarization to induce Ca 2ϩ influx (Ion-BAPTA). The cell was then injected with BAPTA, and I SOC was recorded (Fig. 1B, open circles). Treating cells according to the Ion-BAPTA protocol resulted in a significant (p ϭ 1.6 ϫ 10 Ϫ7 ) potentiation of I SOC levels compared with control oocyte (BAPTA-Ion) (Fig. 1, B and C). Therefore, allowing a Ca 2ϩ i rise during SOCE activation leads to current potentiation.
The Ca 2ϩ -mediated potentiation of SOCE described above was obtained under conditions designed to maximize Ca 2ϩ i rise by hyperpolarization-induced Ca 2ϩ influx. To determine whether such Ca 2ϩ influx is required for I SOC potentiation, a similar experiment was performed without hyperpolarization using the ion-BAPTA protocol, that is Ca 2ϩ stores were simply depleted in the absence of BAPTA injection. In this case, I SOC was also potentiated (Fig. 1D), showing that hyperpolarization per se is not required for the observed Ca 2ϩ i -mediated I SOC potentiation.
As discussed above, SOCE has been shown to be potentiated by extracellular Ca 2ϩ , which affects SOCE channel gating in a process referred to as CDP (12). To avoid confusion with CDP, we will refer to the intracellular Ca 2ϩ effect on I SOC as Ca 2ϩ imediated potentiation (CMP). CMP provides a fitting feedback FIG. 1. Intracellular Ca 2؉ rise potentiates I SOC . A, voltage protocols (VP) used to measure I SOC (#1) and to induce Ca 2ϩ influx (#2). For VP#1, the cell was stepped to Ϫ140 mV followed by a ramp from Ϫ140 to 50 mV from a holding potential of Ϫ20 mV repeated once every 20 s. For VP#2, the cell was stepped to ϩ40, Ϫ140, and ϩ40 mV sequentially from a holding potential of Ϫ40 mV once every 30 s. B, time course of I SOC activation following the BAPTA-Ion (B-I) or Ion-BAPTA (I-B) protocols. For the BAPTA-Ion protocol, oocytes were injected with BAPTA (7 nmol/oocyte) followed by ionomycin (10 M) treatment as indicated by the line. I SOC was measured as the average current 30 -35 ms after stepping to Ϫ140 mV in VP#1. For the Ion-BAPTA protocol, cells were treated with ionomycin and repetitively subjected to VP#2 for 10 min before I SOC recording (open circles). La 3ϩ (100 M) was added at the end of the experiment to block I SOC . C, summary of I SOC levels following the BAPTA-Ion (n ϭ 14) and Ion-BAPTA (n ϭ 12) protocols. The data are plotted as mean Ϯ S.E. (p ϭ 1.6 ϫ 10 Ϫ7 ). D, Ca 2ϩ i rise potentiates I SOC independently of hyperpolarization-induced Ca 2ϩ influx. Cells (in 5 mM Ca 2ϩ ) were exposed to ionomycin for 5 min and then injected with BAPTA (Ion-BAPTA), and I SOC was measured as described in B, BAPTA-Ion treatment was as described in B. The means are significantly different (p ϭ 0.0044; n ϭ 5). mechanism between Ca 2ϩ release and SOCE. This is especially relevant because as shown by Parekh et al. (18), SOCE activation is an all or nothing process (18). That is even partial depletion of intracellular Ca 2ϩ stores results in full activation of I SOC . To determine the relationship between store Ca 2ϩ load and I SOC magnitude in Xenopus oocytes, SOCE was measured by Ca 2ϩ imaging in voltage-clamped cells, which allows simultaneous recording of Ca 2ϩ i and SOCE in the same cell ( Fig. 2) (19). Membrane potential was stepped repetitively to ϩ40 mV (where Ca 2ϩ influx is minimal) and Ϫ140 mV (to induce Ca 2ϩ influx through SOCE). Under these conditions, the difference in Ca 2ϩ fluorescence at Ϫ140 and ϩ40 mV (Fig. 2, open triangles) provides a reliable measure of SOCE and correlates well with I SOC (19). To induce partial store depletion, cells were treated with thapsigargin (1 M) for 1 h (complete store depletion requires Ն3 h), which results in SOCE activation as indicated by the increased fluorescence at Ϫ40 V (Fig. 2, open circles). Injection of IP 3 leads to further Ca 2ϩ release (Fig. 2, squares & circles), confirming that thapsigargin induced partial store depletion. However, the additional store depletion following IP 3 injection does not augment SOCE (Fig. 2, open triangles), indicating that SOCE magnitude does not correlate well with the extent of store depletion. This shows that increased store depletion, at least in the range tested, does not translate into a larger SOCE, arguing that SOCE is modulated by other signaling pathways after store depletion. Fig. 1 shows that one such modulatory pathway is the Ca 2ϩ i levels during SOCE activation.
Correlation between Receptor-induced Ca 2ϩ Mobilization and SOCE-To assess the physiological relevance of CMP, we sought to determine whether it could be observed following IP 3 -linked receptor stimulation. We were interested in inducing different levels of Ca 2ϩ i rise while minimizing experimental manipulation of Ca 2ϩ i (such as BAPTA injection or ionomycin treatment). Our approach was to express the G-protein-coupled serotonin receptor (5HT1c) and monitor changes in the endogenous Ca 2ϩ -activated Cl Ϫ currents as markers of Ca 2ϩ i (19). 5HT1c stimulation with serotonin (5HT) results in IP 3 production through phospholipase-␤ activation (20). We have previously shown that Ca 2ϩ -activated Cl Ϫ currents (I Cl1 and I Cl1T ) faithfully report Ca 2ϩ i changes below the plasma membrane in terms of amplitude and kinetics (19). I Cl1 activates in response to Ca 2ϩ release from internal stores, whereas I Cl1T responds to Ca 2ϩ influx from the extracellular space. During Ca 2ϩ release, I Cl1 activates as a sustained outward current upon depolarization (ϩ40 mV, Fig. 3A, upper trace). Ca 2ϩ release results in store depletion and SOCE activation. Ca 2ϩ flowing through SOCE channels activates I Cl1T as a transient current, only when the depolarization pulse is preceded by a hyperpolarization step, which induces Ca 2ϩ influx (Fig. 3A, lower trace). I Cl1T is transient because the Ca 2ϩ that enters through SOCE channels during the Ϫ140-mV pulse dissipates rapidly during the subsequent 40-mV pulse, leading to I Cl1T current decay (Fig,  3A, lower trace) (19) (for a more detailed description of the relationship between Ca 2ϩ signals and Ca 2ϩ -activated Cl Ϫ currents, see Refs. 19, 21, and 22). Thus, monitoring I Cl1 and I Cl1T allows the real-time determination of Ca 2ϩ i (I Cl1 ) and SOCE (I Cl1T ) levels following 5HT1c activation.
To generate graded Ca 2ϩ release responses and determine the effect on SOCE, we injected cells with different amounts of 5HT1c cRNA (10 or 30 ng) and allowed them to express for 2 days. Cells were then stimulated with 5HT (10 M) (Fig. 3, B and C), leading to Ca 2ϩ release, which activates I Cl1 (Fig. 3, B and C, squares) followed by I Cl1T due to SOCE activation (Fig.  3, B and C, circles). I Cl1T activates to a certain threshold and gradually returns to base line as a result of Ca 2ϩ store refilling and SOCE inactivation (Fig. 3, B and C, circles). Cells injected with 10 ng of 5HT1c cRNA produced a smaller I Cl1 (Fig. 3B) than those injected with 30 ng of the receptor (Fig. 3C). Furthermore, I Cl1T activated to a lower threshold and inactivated more rapidly in cells injected with 10 ng of 5HT1c receptor cRNA (Fig. 3, compare B with C, circles). The responses of the Cl Ϫ currents summarized in Fig. 3D show that both Ca 2ϩ i (as indicated by I Cl1 ) and SOCE (as indicated by I Cl1T ) were significantly enhanced at high 5HT1c (30 ng) expression. Therefore, enhanced Ca 2ϩ release correlates with larger SOCE. A simple explanation for these data is that receptor stimulation in cells expressing high levels of 5HT1c leads to a more dramatic Ca 2ϩ store depletion and larger SOCE. However, as shown in Fig. 2, the extent of Ca 2ϩ store depletion does not linearly correlate with SOCE magnitude. Therefore, SOCE potentiation as reported by I Cl1T in cells expressing more 5HT1c receptors is probably because of some mechanism other than the extent of store depletion. CMP of I SOC provides such a mechanism. High 5HT1c expression produces increased Ca 2ϩ i as indicated by I Cl1 , which would be expected to potentiate I SOC (I Cl1T ) through CMP. Note that store depletion provides a required signal for SOCE activation, but the extent of store depletion per se does not modulate SOCE magnitude. Rather, the data in Figs. 2 and 3 argue that it is the magnitude of the Ca 2ϩ i rise that modulates SOCE through CMP. Clearly the magnitude of Ca 2ϩ i rise can correlate with the extent of store depletion. These data reveal a subtle but important distinction in the mechanism of SOCE modulation and are consistent with FIG. 2. Partial store depletion fully activates SOCE. Oocyte was loaded with Ca 2ϩ -Green-1-dextran (7 M) and incubated in thapsigargin (Thaps) (1 M) for 1 h to partially deplete intracellular Ca 2ϩ stores. SOCE was measured by clamping the cell using voltage protocol number 2 (Fig. 1A) as described by Machaca and Hartzell (19). Ca 2ϩ -Green-1-dextran fluorescence at ϩ40 mV (F Ca(ϩ40) ; close squares) provides basal Ca 2ϩ levels because there is minimal Ca 2ϩ influx at this voltage (19). In contrast, Ca 2ϩ influx through SOCE channels is induced at Ϫ140 mV, resulting in increased Ca 2ϩ -Green-1-dextran fluorescence (F Ca(Ϫ140) ; open circles). Therefore, the difference between CG1 fluorescence at Ϫ140 and ϩ40 mV provides a measure of SOCE. Partial store depletion with a short thapsigargin incubation (1 h) activates SOCE. Injection of IP 3 (ϳ2 M) leads to further Ca 2ϩ release from stores as indicated by increased Ca 2ϩ -Green-1-dextran fluorescence at both voltages (squares and circles). However, the additional Ca 2ϩ release leading to further store depletion does not enhance SOCE (open triangles). Switching the cell to Ca 2ϩ -free solution results in a loss of the Ca 2ϩ -Green-1-dextran fluorescence signal at Ϫ140 mV, confirming that the signal is because of Ca 2ϩ influx. These data are representative of four similar experiments. the notion that CMP is a physiologically relevant modulator of SOCE.
The positive correlation between Ca 2ϩ i and SOCE described above is not limited to Xenopus oocytes. Gailly et al. (23) described a similar relationship in Chinese hamster ovary cells (23), arguing that CMP is a widespread mechanism of SOCE regulation.
CaMKII Potentiates SOCE-Ca 2ϩ i -mediated potentiation of I SOC could be either direct or indirect through the activation of Ca 2ϩ -dependent downstream effectors. A direct Ca 2ϩ -mediated potentiation of I SOC is unlikely because Ca 2ϩ has been shown to inactivate I SOC , probably through a direct effect on SOCE channels (fast Ca 2ϩ -dependent inactivation) (8,10). This finding suggests that CMP is the result of activation of Ca 2ϩ -dependent effectors, which in turn act on SOCE. A primary candidate for such an effector pathway is the Ca 2ϩ -CaM-activated protein kinase pathway. The most widespread Ca 2ϩ -CaM-dependent kinase is CaMKII, which is expressed in Xenopus oocytes (24). If Ca 2ϩ potentiates I SOC through CaMKII, it is expected that ectopic activation of CaMKII would lead to increased I SOC levels.
We used a constitutively active CaMKII mutant (CaMKII ca ) to determine whether CaMKII activation potentiates I SOC . The CaMKII holoenzyme is a multisubunit complex that is kept inactive by an autoinhibitory domain. The binding of Ca 2ϩ -CaM relieves autoinhibition and stimulates autophosphorylation at Thr-286, rendering the enzyme Ca 2ϩ -CaM-independent (25). CaMKII ca is a T286D mutation that mimics the effects of autophosphorylation by replacing Thr-286 with Asp, resulting in a Ca 2ϩ -CaM-independent and thus constitutively active CaMKII (26). We injected oocytes with CaMKII ca and measured I SOC (Fig. 4A). Expression of CaMKII ca results in an ϳ3-fold increase in I SOC (p ϭ 1.1 ϫ 10 Ϫ7 ) (Fig. 4, A and B), consistent with CMP acting through CaMKII. In addition, CaMKII ca -mediated I SOC potentiation is expected to be independent of Ca 2ϩ i because CaMKII ca is Ca 2ϩ -CaM-independent. This prediction is confirmed in CaMKII ca -expressing cells, as I SOC is potentiated whether store depletion is preceded by a rise in Ca 2ϩ i (Ion-BAPTA) or not (BAPTA-Ion) (Fig. 4C). This finding further suggests that Ca 2ϩ -mediated potentiation is primarily through CaMKII activation, because no additive effect on I SOC is observed in CaMKII ca -expressing cells subjected to the Ion-BAPTA protocol (Fig. 4C).
To directly confirm CaMKII ca expression, we measured CaMKII-specific activity in lysates from control and CaMKII cainjected cells. Cells expressing CaMKII ca had higher levels of CaMKII activity (Fig. 4D), showing that CaMKII ca was expressed and functional in these cells. Because specific activity was measured in the absence of Ca 2ϩ -CaM (Fig. 4D), CaMKII activity data confirm the Ca 2ϩ -CaM independence of CaMKII ca .
It is important to note that the expression of CaMKII ca by itself is not sufficient to activate I SOC (Fig. 4A, circles). In CaMKII ca -expressing cells, store depletion is still required to activate SOCE because no I SOC is detected before ionomycin treatment (Fig. 4A, circles). However, I SOC levels reach a significantly larger maximal amplitude in CaMKII ca -expressing cells compared with control cells (Fig. 4, A and B). Therefore, CaMKII is not a component of the SOCE activation pathway induced in response to store depletion but rather positively modulates SOCE activity following SOCE activation.
If CMP is acting through CaMKII, CaMKII should be activated in cells where a Ca 2ϩ i rise is induced (Ion-BAPTA). We were unable to detect such an increase in CaMKII activity in oocytes treated according to the Ion-BAPTA protocol. This result argues that either Ca 2ϩ and CaMKII potentiate I SOC by separate mechanisms or that CaMKII activation is transient and/or spatially localized, resulting in small changes in CaMKII activity that are difficult to detect in whole cell lysates. Furthermore, basal CaMKII activity was quite variable in oocytes donated from different females, making it difficult to reliably measure a small increase in CaMKII activity in different batches of cells. However, a Ca 2ϩ i rise has been shown to activate endogenous CaMKII in Xenopus oocytes using an in vivo CaMKII-specific kinase assay (27).
Nonetheless, if CMP is mediated by CaMKII, blocking endogenous CaMKII should inhibit CMP. Therefore, we blocked endogenous CaMKII and determined the effect on SOCE. Oocytes were injected with AIP, a specific inhibitory peptide of CaMKII that mimics the autoinhibitory domain and blocks CaMKII activity (28). Allowing a Ca 2ϩ i rise in control oocytes (Ion-BAPTA) potentiates I SOC by ϳ2-fold but not in cells injected with the CaMKII inhibitor AIP (Fig. 5A). Furthermore, AIP added to the CaMKII assay blocks kinase activity in a dose-dependent manner (Fig. 5B). These data show that inhib-iting endogenous CaMKII activity blocks CMP, supporting the conclusion that Ca 2ϩ i potentiates SOCE through CaMKII activation.
Mechanism of CaMKII Action on SOCE-Because SOCE is activated in response to store depletion, CaMKII could potentiate I SOC by targeting either the coupling mechanism between Ca 2ϩ stores and SOCE or the SOCE channel. To differentiate between these possibilities and obtain a better understanding of the mechanism of action of CaMKII, we wanted to study the effects of CaMKII independently of store depletion. This was accomplished by irreversibly depleting Ca 2ϩ stores with thapsigargin before CaMKII ca expression (Fig. 6). Oocytes were incubated in thapsigargin (1 M) for 3 h to fully deplete Ca 2ϩ store followed by CaMKII ca cRNA injection and incubation in nominally Ca 2ϩ -free (50 M free Ca 2ϩ ) solution. Control cells were treated with thapsigargin alone. Under these conditions, store depletion is complete before CaMKII ca expression and store Ca 2ϩ load remains low throughout the experiment because thapsigargin irreversibly inhibits the ER Ca 2ϩ -ATPase. This allows us to study the effects of CaMKII on I SOC after store depletion and determine whether CaMKII is affecting the coupling mechanism or SOCE channel gating or permeation. In this experiment, SOCE was activated in the absence of the conducting ion (Ca 2ϩ ) and therefore no SOCE current was observed at the beginning of the experiment (Fig. 6A). Switching control (thapsigargin-treated) oocytes to a Ca 2ϩ -containing solution produced an initial I SOC that was further enhanced over time, eventually saturating (Max I SOC ) within ϳ6 min (Fig. 6A, squares). This behavior is the result of classical CDP where extracellular Ca 2ϩ exerts a positive effect on SOCE channel gating (12, 13). CDP can be estimated as the ratio of Oocytes were injected with a constitutively active CaMKII (CaMKII ca , 1 ng/oocyte) and allowed to express for 12-16 h. SOCE was measured using voltage protocol number 1 in Fig. 1A with the exception that the voltage was stepped to Ϫ120 mV instead of Ϫ140 mV. A, time course of I SOC activation in control and CaMKII caexpressing cells. B, normalized I SOC levels (n as indicated; p ϭ 1.1 ϫ 10 Ϫ7 ). C, normalized I SOC levels from control (Con.) and CaMKII ca -injected cells treated according to the BAPTA-Ion or Ion-BAPTA protocols as described in Fig. 1. The asterisks above the bars indicate the significantly different groups (n as indicated; p Ͻ 0.0164). D, basal CaMKII kinase activity from control and CaMKII ca -injected oocytes measured using an in vitro kinase assay without the addition of Ca 2ϩ -CaM (n ϭ 5; p ϭ 0.0469). maximal I SOC /initial I SOC (12) and is 3.11 Ϯ 0.23 in thapsigargin-treated cells (Fig. 6B).
Surprisingly, exposing CaMKII ca -expressing cells to Ca 2ϩcontaining solution results in an initial I SOC with a similar amplitude to initial I SOC in control cells (Fig. 6A, open circles). However, I SOC in CaMKII ca -expressing cells gradually increased to significantly higher levels than in control cells (Fig.  6A, open circles). This shows that CaMKII ca expression does

FIG. 5. Inhibition of endogenous CaMKII blocks Ca 2؉
i -mediated SOCE potentiation. Control and AIP-injected (Inj.) (10 M) oocytes were treated according to the BAPTA-Ion and Ion-BAPTA protocols as indicated. A, normalized I SOC levels in the different treatment groups. The asterisk indicates the only significantly different group (p Ͻ 2.1 ϫ 10 Ϫ4 ). B, basal CaMKII activity in oocyte lysate without AIP (Con) and with 10 and 40 M AIP as indicated (n ϭ 6).
FIG. 6. CaMKII potentiates I SOC by increasing the levels of CDP. Cells were incubated with thapsigargin (Thaps) (1 M) in nominally Ca 2ϩ -free medium (50 M) for 3 h to fully deplete Ca 2ϩ stores. A subset of cells was then injected with 1 ng of CaMKII ca RNA (Thaps-CaMK ca ) and incubated in nominally Ca 2ϩ -free medium for 12-16 h. A, I SOC recorded from a representative control (Thaps) and CaMKII ca -injected oocyte (Thaps-CaMK ca ). Cells were incubated in Ca 2ϩ -free solution (70 Mg 2ϩ ) before switching to Ca 2ϩ -containing solution (30 Ca 2ϩ ) as indicated by the line. SOCE was measured using voltage protocol number 1 in Fig. 1A with the exception that the voltage was stepped to Ϫ120 mV instead of Ϫ140 mV. The addition of La 3ϩ to block I SOC is also indicated. B, normalized I SOC levels showing initial I SOC (indicated by the asterisk in A) and maximal I SOC at the end of the experiment (indicated by the open square and filled circle in the Thaps and Thaps-CaMK ca groups, respectively). The average levels of CDP calculated as maximal I SOC /initial I SOC are also shown. Maximal I SOC and CDP are significantly different between the two groups (p Ͻ 0.00132). C and D, current traces in response to a step voltage (Ϫ20 to Ϫ120 mV for 500 ms) (C) and a current ramp (Ϫ140 to ϩ50 mV) (D) obtained at different time points during the experiment as indicated in A. E and F, superimposed current traces of maximal I SOC from Thaps and Thaps-CaMKII ca -treated cells in response to a step voltage pulse (E) and a voltage ramp (F). not affect initial I SOC levels but results in a significantly (p ϭ 1.46 ϫ 10 Ϫ5 ) larger maximal I SOC (Fig. 6B). It follows that CDP, calculated as the ratio of maximal/initial I SOC , was also augmented (5.72 Ϯ 0.61) in CaMKII ca -expressing cells (Fig.  6B). Therefore, CaMKII potentiates I SOC by increasing the levels of CDP without affecting initial I SOC levels, even after prolonged expression of CaMKII ca .
These data provide important insights into the mechanism of action of CaMKII. The whole cell SOCE current is defined by the following equation: I SOC ϭ NP o i, where N is the number of active channels, P o is the probability of opening, and i is the single channel conductance. Therefore, CaMKII can potentiate I SOC by increasing either N, i, or P o . If CaMKII was affecting the coupling mechanism, an increase in the number of active channels (N) is expected. The fact that CaMKII ca expression does not enhance initial I SOC provides evidence against an increase in channel number (N) or single channel conductance (i) (Fig. 6, A and B). This is because initial I SOC is due to current flowing through all of the open channels after Ca 2ϩ addition. If either N or i was augmented by CaMKII, initial I SOC would be larger in CaMKII ca -expressing cells. This hypothesis argues that CaMKII potentiates I SOC by enhancing the probability of opening (P o ) of SOCE channels.
The observed increase in CDP levels in CaMKII ca -expressing cells strongly supports the conclusion that CaMKII potentiates I SOC by increasing P o . This is because CDP has been shown to be the result of extracellular Ca 2ϩ exerting a positive effect on SOCE channel gating (P o ) (12,13).
Representative current traces in response to a step pulse to Ϫ120 mV (Fig. 6C) and a voltage ramp from Ϫ140 to 50 mV (Fig. 6D) obtained from control (Thaps) and CaMKII ca -injected cells (Thaps-CaMK ca ) are shown. The time points during the experiments at which the traces were obtained are indicated in Fig. 6A (Thaps, open symbols; Thaps-CaMK ca , filled symbols). Note the similar levels of initial I SOC in both treatments (Fig.  6C, open and filled stars). Initial I SOC traces from CaMKII cainjected cells show a time-dependent increase in current amplitude (Fig. 6C, filled star) that is due to CDP occurring during the 500-ms duration of the voltage pulse. Zweifach and Lewis (8) have shown that the extent of Ca 2ϩ -dependent inactivation of I CRAC following a brief hyperpolarization depends on the single channel current (i) (8). That is, increased unitary current (i) results in a faster rate of I CRAC inactivation. Assuming that the same relationship holds for Xenopus oocyte I SOC , if CaMKII increases unitary conductance (i), we expect a faster I SOC inactivation rate shortly after the hyperpolarization pulse. However, the rate of I SOC inactivation shortly after (100 ms) hyperpolarization is similar between control and CaMKII caexpressing cells (Fig. 6E). In fact, at steady-state, control cells exhibit a more marked I SOC inactivation (Fig. 6E). These inactivation kinetics argue that CaMKII does not enhance I SOC unitary conductance. Superimposed current-voltage relationships from control and CaMKII ca -injected cells were similar (Fig. 6F), indicating that CaMKII does not affect the voltage dependence of I SOC .
These results argue that CaMKII ca potentiates I SOC by targeting SOCE channel gating and not the coupling mechanism between Ca 2ϩ stores and SOCE. Three pieces of evidences support the conclusion that CaMKII potentiates I SOC by affecting channel gating (P o ) and not channel number (N) or unitary conductance (i): 1) similar levels of initial I SOC in control and CaMKII ca -expressing cells (Fig. 6, A and B) arguing against an effect of CaMKII on N or i. 2) enhanced CDP in CaMKII ca -expressing cells (Fig. 6B), suggesting that CaMKII enhances channel P o . 3) similar I SOC inactivation rates shortly after hyperpolarization in control and CaMKII ca -expressing cells (Fig. 6E) arguing against an increase in unitary conductance (i).

Ca 2ϩ
i has been shown to negatively regulate I SOC by inducing channel inactivation either directly or through store refilling (8,10,11). Here we show that Ca 2ϩ i can in addition have a potentiating effect on SOCE. The Ca 2ϩ i effect on SOCE is probably the result of CaMKII activation, because Ca 2ϩ i , because inhibition of endogenous CaMKII activity blocks CMP and expression of CaMKII ca is sufficient to potentiate I SOC independently of Ca 2ϩ i . Although CaMKII potentiates I SOC , it is not sufficient by itself to activate SOCE independently of store depletion (Fig. 4A). This observation is consistent with the fact that a Ca 2ϩ i rise is not required for SOCE activation (1). Therefore, the CaMKII pathway modulates SOCE activity but is not an essential component of the SOCE activation pathway in response to store depletion. For CaMKII to exert its effects, SOCE has to be activated by store depletion.
Using pharmacological inhibitors, a role for CaMKII in skeletal muscle SOCE (29) and myosin light chain kinase (a Ca 2ϩ -CaM-dependent kinase related to CaMKII (25)) in endothelial SOCE (30) have been postulated. This finding argues that CaMKII modulation of I SOC is a widespread mechanism that is not cell type-specific. Nonetheless, a previous report by Matifat et al. (27) suggests a negative effect of CaMKII on SOCE in Xenopus oocytes. However, it is not clear from this study that the reported CaMKII effect was due to SOCE modulation because the Ca 2ϩ -activated Cl Ϫ current was used as an indicator of SOCE, making it impossible to differentiate between an effect of CaMKII on the Cl Ϫ current or SOCE. In contrast, we have directly measured the SOCE current while modulating CaMKII activity and show that CaMKII activation potentiates I SOC . This finding argues that the effects of CaMKII reported by Matifat et al. (27) are due to modulation of the Cl Ϫ currents or other Ca 2ϩ influx pathways in the oocyte.
CaMKII provides an excellent modulator of SOCE activity because of its ability to decode spatial and temporal information encoded in Ca 2ϩ i signals into different levels of kinase activity. This capacity is attributable to spatial and structural features of CaMKII. CaMKII localizes to specific subcellular compartments such as the nucleus and cytoskeleton (31,32), and its activity increases exponentially based on the number and frequency of Ca 2ϩ i oscillation (15,33). These exceptional attributes allow CaMKII to provide specificity to Ca 2ϩ i signals by differentially activating effectors based on the kinetics of Ca 2ϩ i signals. CaMKII plays important roles in regulating various cellular functions such as gene expression, cell cycle progression, and learning and memory (25,34). The data presented here show that SOCE can now be added to the list of cellular functions modulated by CaMKII.
In a similar fashion to the CaMKII effect on SOCE described here, CaMKII has been shown to augment both L-type (Ca v 1.x) (35) and T-type (Ca v 3.x) (36) Ca 2ϩ currents. CaMKII regulation of SOCE described in this paper is reminiscent of Ca 2ϩ -dependent facilitation of inward L-type Ca 2ϩ current. As Ca 2ϩ i increases, L-type Ca 2ϩ current is enhanced in a process termed facilitation (37). This facilitation is mediated by CaMKII, which increases the open probability of L-type channels (35). Our data strongly argue that CaMKII potentiates I SOC in a similar fashion by altering channel gating (see Fig. 6). This similarity in the mechanism of action of CaMKII on both voltage-gated and store-operated Ca 2ϩ channels argues for a common cross-talk between Ca 2ϩ i kinetics and Ca 2ϩ influx pathways. It is attractive to postulate that CaMKII provides a mechanism for phospholipase-linked receptors to differentially modulate SOCE activity. That is, the spatiotemporal features of the Ca 2ϩ i signals downstream of receptor activation differentially activate CaMKII and thus SOCE.
In addition to the CaMKII-mediated regulation of SOCE described here, SOCE has been shown to be modulated by protein kinase C (38,39). Both protein kinase C and CaMKII are downstream kinases that can be induced following activation of phospholipase-linked receptors. Therefore, the interplay between various modulatory mechanisms in a specific cellular context combines to generate different levels of Ca 2ϩ influx through SOCE. This Ca 2ϩ influx affects Ca 2ϩ i kinetics and thus Ca 2ϩ -dependent cellular responses. Consequently, CaMKII regulation of SOCE is likely to have broad and important physiological consequences.