Role of regulator of G protein signaling 2 (RGS2) in Ca(2+) oscillations and adaptation of Ca(2+) signaling to reduce excitability of RGS2-/- cells.

Regulators of G protein signaling (RGS) proteins accelerate the GTPase activity of Galpha subunits to determine the duration of the stimulated state and control G protein-coupled receptor-mediated cell signaling. RGS2 is an RGS protein that shows preference toward Galpha(q). To better understand the role of RGS2 in Ca(2+) signaling and Ca(2+) oscillations, we characterized Ca(2+) signaling in cells derived from RGS2(-/-) mice. Deletion of RGS2 modified the kinetic of inositol 1,4,5-trisphosphate (IP(3)) production without affecting the peak level of IP(3), but rather increased the steady-state level of IP(3) at all agonist concentrations. The increased steady-state level of IP(3) led to an increased frequency of [Ca(2+)](i) oscillations. The cells were adapted to deletion of RGS2 by reducing Ca(2+) signaling excitability. Reduced excitability was achieved by adaptation of all transporters to reduce Ca(2+) influx into the cytosol. Thus, IP(3) receptor 1 was down-regulated and IP(3) receptor 3 was up-regulated in RGS2(-/-) cells to reduce the sensitivity for IP(3) to release Ca(2+) from the endoplasmic reticulum to the cytosol. Sarco/endoplasmic reticulum Ca(2+) ATPase 2b was up-regulated to more rapidly remove Ca(2+) from the cytosol of RGS2(-/-) cells. Agonist-stimulated Ca(2+) influx was reduced, and Ca(2+) efflux by plasma membrane Ca(2+) was up-regulated in RGS2(-/-) cells. The result of these adaptive mechanisms was the reduced excitability of Ca(2+) signaling, as reflected by the markedly reduced response of RGS2(-/-) cells to changes in the endoplasmic reticulum Ca(2+) load and to an increase in extracellular Ca(2+). These findings highlight the central role of RGS proteins in [Ca(2+)](i) oscillations and reveal a prominent plasticity and adaptability of the Ca(2+) signaling apparatus.

G protein-coupled receptor (GPCR) 1 -evoked Ca 2ϩ signaling is initiated by biochemical reactions at the receptor complex that generate Ca 2ϩ -releasing second messengers, which lead to Ca 2ϩ fluxes into and out of the cytosol (1,2). The biochemical complex is composed of a receptor, the heterotrimeric G protein G q (and in some cases G i ), and the effector PLC␤. Ligand binding to GPCRs results in activation of G␣ q that, in turn, activates PLC␤. PLC␤ hydrolyzes phosphatidylinositol bisphosphate to generate IP 3 . IP 3 releases Ca 2ϩ from the ER, which is followed by the activation of store-operated Ca 2ϩ channels in the plasma membrane and Ca 2ϩ influx. The increase in Ca 2ϩ leads to activation of the plasma membrane Ca 2ϩ ATPase (PMCA) and sarco/endoplasmic reticulum Ca 2ϩ ATPase (SERCA) pumps to remove Ca 2ϩ from the cytosol. The overall signal is a transient change in [Ca 2ϩ ] i . In the case of an intense stimulation only a single Ca 2ϩ transient is observed, whereas at weak physiological stimulus intensity the transient is repeated to generate [Ca 2ϩ ] i oscillations (1,2).
Because [Ca 2ϩ ] i oscillations control virtually all cell functions from cell birth to cell death (1)(2)(3), the mechanisms for the generation of Ca 2ϩ oscillations and the regulation of their amplitude and frequency are of major interest. Early work supported models in which the regulation of the IP 3 R Ca 2ϩ release channels by cytoplasmic Ca 2ϩ ([Ca 2ϩ ] i ) generates agonist-evoked [Ca 2ϩ ] i oscillations. These models were based on the findings of the bell-shaped dependence of the IP 3 Rs on [Ca 2ϩ ] i (4) and the generation of [Ca 2ϩ ] i oscillations by nonhydrolyzable IP 3 (5). However, the finding that weak agonist stimulation leads to oscillatory changes in IP 3 concentration (6,7) suggested that a primary mechanism of [Ca 2ϩ ] i oscillations is the cyclical activation of PLC␤. Two mechanisms for the regulation of PLC␤ were suggested; one is the direct regulation of PLC␤ activity by [Ca 2ϩ ] i (8), and the second is the regulation of PLC␤ by regulation of the availability of activated G␣ q (9).
[Ca 2ϩ ] i can indeed regulate PLC␤ activity to influence [Ca 2ϩ ] i oscillations (10). However, in this model the receptor does not control any aspect of the oscillations after the initial stimulation, whereas it is well documented that stimulus intensity affects the amplitude and, in particular, the frequency of the oscillation (1,2).
The availability of activated G␣q is regulated by the regulators of G protein signaling (RGS) proteins (11). The off reaction in the G protein cycle is the hydrolysis of GTP and the reassembly of the G␣␤␥ heterotrimer. This reaction is accelerated by RGS proteins in a manner such that the continuous presence of agonist IP 3 level oscillates to drive [Ca 2ϩ ] i oscillations (9). An open question is which RGS protein dominates the regulation of G q -coupled receptors in vivo. This is an important question in view of the promiscuity of RGS proteins toward G␣ subunits in vitro (11) and the expression of multiple RGS proteins in the same cell (12,13). It is generally assumed that RGS2 regulates the activity of G q because in vitro RGS2 shows some preference toward G␣q (11,14), although other RGS proteins can inhibit G␣q activity as well as or better than RGS2 (11). In addition, the role of RGS2 in Ca 2ϩ signaling and [Ca 2ϩ ] i oscillations in vivo is not known. The availability of the RGS2 Ϫ/Ϫ mice (15) allows direct examination of the roles of RGS2 in Ca 2ϩ signaling in vivo. The RGS2 Ϫ/Ϫ mice are viable and fertile, although they are immune compromised (15) and develop cardiac hypertrophy (16), probably as a result of hypertension (17). We show here that deletion of RGS2 increases the apparent affinity for agonist-stimulated [Ca 2ϩ ] i signaling but without changing the dose response for agonist-stimulated peak IP 3 production. Rather, deletion of RGS2 changes the kinetic of IP 3 production to increase its steady-state level and, consequently, the frequency of [Ca 2ϩ ] i oscillations.
Ca 2ϩ signaling complexes are plastic, and the activity of their components adjusts to perturbations brought about by deletion or overexpression of their components. For example, overexpression of PMCA results in compensatory overexpression of SERCA pumps (18,19), and partial deletion of SERCA2 results in up-regulation of PMCA4 (20). Furthermore, critical cellular activities such as Ca 2ϩ -triggered exocytosis (20) adapt to the change in the characteristics of the Ca 2ϩ signal. Deletion of an RGS protein that regulates the accessibility of active G␣q is expected to change PLC␤ activity and IP 3 production and is likely to lead to the adaptation of Ca 2ϩ signaling to reduce excitability. The excitability of Ca 2ϩ signaling is determined by the ability of IP 3 to release Ca 2ϩ from the ER, the contribution of Ca 2ϩ influx to increase [Ca 2ϩ ] i and augment Ca 2ϩ release, and the capacity of SERCA and PMCA to remove Ca 2ϩ from the cytosol. Ca 2ϩ influx is likely to have a central role in determining the excitability of Ca 2ϩ signaling complexes because it determines the duration and frequency of [Ca 2ϩ ] i oscillations (21)(22)(23). Indeed, the regulation of many Ca 2ϩ -dependent cellular activities correlates with the activity of Ca 2ϩ influx (24).
A detailed analysis of Ca 2ϩ signaling in cells from RGS2 Ϫ/Ϫ mice revealed the adaptation of all components of the Ca 2ϩ signaling complex that leads to reduced excitability of Ca 2ϩ signaling in RGS2 Ϫ/Ϫ cells. These findings demonstrate the central role of RGS proteins in [Ca 2ϩ ] i oscillations and the remarkable plasticity of Ca 2ϩ signaling complexes.

EXPERIMENTAL PROCEDURES
Materials and Solutions-Fura2/AM was from Teff Laboratories, anti-PMCA 5F10 antibodies were from Sigma, anti-PLC␤ isoforms were a generous gift from Dr. Paul Sternweis (University of Texas Southwestern Medical Center, Dallas, TX), anti-SJ1 was a generous gift from Dr. Pietro De Camilli (Yale University, New Haven, CT), anti-IP 3 R1 antibodies were a generous gift from Dr. Greg Mignery (Loyola University Chicago, IL), and anti-SERCA2b antibodies were a generous gift from Dr. Frank Wuytack (Katholieke Universiteit Leuven, Belgium). Anti-IP 3 R3 antibodies were from Transduction Laboratories, and anti-IP 3 3-kinase was from Santa Cruz Biotechnology (Santa Cruz, CA). The standard perfusion solution A contained 140 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 1 mM CaCl 2 , 10 mM HEPES (pH 7.4 with NaOH), and 10 mM glucose. When supplemented with 10 mM pyruvate, 1 mg/ml bovine serum albumin, and 0.02% soybean trypsin inhibitor it was named PSA. The Ca 2ϩ -free medium was solution A without CaCl 2 and with 0.2 mM EGTA.
Preparation of Pancreatic Acini-Acini were prepared as described previously (25). In brief, WT and RGS2 Ϫ/Ϫ mice were anesthetized and killed by cervical dislocation. The pancreata were removed and digested with collagenase P in PSA solution. The acini were washed and, as needed, loaded with Fura2, suspended in PSA, and kept on ice until use.
Measurement of [Ca 2ϩ ] i -Acini loaded with Fura2 were plated on polylysine-coated glass coverslips that form the bottom of a perfusion chamber. The acini were perfused with worm (37°C) solution A, and agonists were delivered with the perfusate. Fura2 fluorescence was recorded using a Photon Technology International image equation and analysis system, and the fluorescence signals were calibrated as detailed previously (25).
Mass Measurement of IP 3 -Acini in solution A were stimulated with the indicated carbachol concentration for 2-120 s. The reactions were stopped and the proteins precipitated by the addition of perchloric acid and by incubating the acini for at least 20 min on ice. IP 3 was extracted with a mixture of 0.2 ml Freon and 0.2 ml of tri-n-octylamine, and IP 3 was measured by a standard radioligand assay (26).
Western Blot Analysis-Microsomes were prepared from the brain, pancreas, and submandibular glands of WT and RGS2 Ϫ/Ϫ mice by homogenization in medium composed of 20 mM HEPES, pH7.4, 150 mM NaCl, 20% glycerol, 1.5 mM MgCl 2 , 0.5% Triton X-100, 1 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride and supplemented with a protease inhibitor mixture (Roche Applied Science). Lysates were separated by SDS/PAGE, and proteins were detected by blotting with the desired antibodies.

Deletion of RGS2 Increases Dynamic of IP 3 Production-
Because cells express multiple RGS proteins (12,13), the most direct way to study the role of a specific RGS protein is deletion of the specific genes in mice. RGS2 is the RGS protein that preferentially activates G␣ q (14). Although deletion of RGS2 resulted in several physiological phenotypes (15)(16)(17), the only Ca 2ϩ signaling effect reported is an increased responsiveness to the stimulation of P2 receptors (12). To determine the consequences of the deletion of RGS2 on Ca 2ϩ signaling, we began by analyzing the effect of the deletion of RGS2 on the expression and activity of IP 3 metabolizing enzymes. Preliminary experiments showed the signal/noise ratio obtained with extracts prepared from the pancreas or salivary glands was not adequate to quantify protein expression. A much more reproducible and quantifiable signal was obtained using brain extracts. Comparable quality of results in brain and secretory cells was obtained only for SERCA2b. Therefore, only data obtained with extracts prepared from brain microsomes are presented.
RGS proteins terminate GPCRs signaling by accelerating the GTPase activity of G␣ subunits to inhibit the activation of effector proteins, including PLC␤, by G␣ (11). The cells may adapt by down-regulation of PLC␤ expression. This possibility was tested by Western blot analysis of PLC␤ expression. Fig.  1A shows that the deletion of RGS2 had no discernible effect on the expression of PLC␤1, PLC␤2, or PLC␤3 isoforms. Once generated, IP 3 is metabolized by IP 3 5-phosphatases that remove the phosphate from the 5 position of many phosphoinositols, including IP 3 , and IP 3 3-kinases that phosphorylate IP 3 to 1,3,4,5-tetrakisphosphate (IP 4 ). To date, eight 5-phosphatases that are grouped into three sub-families have been identified (27). We focused on synaptojanin 1 (SJ1), a member of the 5-phosphatase subfamily II. SJ1 is expressed in two major forms, an abundant and ubiquitously expressed SJ170 and a neuronal specific SJ145 (28). Fig. 1B shows that the deletion of RGS2 had no effect on the expression of SJ170. Interestingly, the deletion of RGS2 precipitously reduced the expression of SJ145. The significance of this finding to Ca 2ϩ signaling in neurons is not known at present. However, because SJ1 can act on phosphatidylinositol 4,5-bisphosphate to regulate clathrin-mediated endocytosis of synaptic vesicles (29), it should be of interest to look at this activity and its behavioral consequences in the RGS2 Ϫ/Ϫ mice. IP 3 3-kinases exist as three isoforms, IP3KA, B, and C, which show tissue-specific expression (30). The 74-kDa IP3KB is ubiquitous and is associated with cellular membranes, including the plasma membrane (27,30). Analysis of IP3KB expression showed that deletion of RGS2 had no effect on the expression of this enzyme (Fig. 1C).
The fact that deletion of RGS2 had no effect on PLC␤, SJ170, and IP3KB expression raised the possibility that it might affect the dynamic of IP 3 production. Measurement of the time course of IP 3 production showed that this is indeed the case. Fig. 2A shows that stimulation of WT and RGS2 Ϫ/Ϫ acini with 1 M carbachol slowly increased IP 3 levels that were the same in the two cell types. However, importantly, the steady-state level of IP 3 after 60 and 120 s of stimulation was significantly higher in RGS2 Ϫ/Ϫ acini. The effect of RGS2 deletion on the dynamic of IP 3 production is further demonstrated in Fig. 2B, where the acini were stimulated with 1 mM carbachol. At this concentration IP 3 production was transient, reaching a maximum level after 2-5 s of stimulation and then steadily reducing to stabilize at a new steady-state level. At a high agonist concentration it was possible to demonstrate that the deletion of RGS2 increased the initial rate of IP 3 production and, again, the steadystate levels of IP 3 at the longer stimulation periods. Experiments similar to those depicted in Fig. 2, A and B were used to determine the dose response for the peak and steady-state levels of IP 3 production. Fig. 2C shows that deletion of RGS2 had no effect on the peak level attained at all carbachol concentrations. On the other hand, deletion of RGS2 increased the steady-state level of IP 3 at all carbachol concentrations (Fig. 2D).
The findings that the deletion of RGS2 increased the initial rate ( Fig. 2B) and steady-state level (Fig. 2D) with no effect on the extent of IP 3 production (Fig. 2C) have several implications. The increased rate of IP 3 production indicates that in resting cells RGS2 exerts tonic inhibition on GPCR signaling. This conclusion is in line with previous observations demonstrating the constitutive inhibitory activity of recombinant RGS proteins when applied in vivo and the activation of Ca 2ϩ signaling by scavenging RGS protein antibodies (9). The increased steady-state level of IP 3 can result from the increased rate of IP 3 production and the reduced rate of signal termination. However, it is significant that a more prominent increased steady state is observed at low agonist concentration (Fig. 2D) at which the IP 3 level oscillates (6, 7). As was shown for [Ca 2ϩ ] i oscillations (31), the increased frequency of IP 3 oscillations will be translated to an increased steady-state when IP 3 levels are measured in a cell population. Therefore, it is possible that the increased steady-state level of IP 3 at low agonist concentrations observed in RGS2 Ϫ/Ϫ cells may reflect the increased frequency of IP 3 oscillations. The effect of RGS2 on steady-state levels of IP 3 highlights the importance of RGS proteins in controlling IP 3 production and [Ca 2ϩ ] i signaling in vivo (9).
Deletion of RGS2 Increases Stimulus Intensity-An increased frequency of IP 3 oscillations at a low agonist concentration predicts an increased frequency of [Ca 2ϩ ] i oscillations, and increased steady-state IP 3 levels at all agonist concentrations predict an increased apparent affinity for agonist-stimulated Ca 2ϩ signaling. These predictions were tested by measuring [Ca 2ϩ ] i in WT and RGS2 Ϫ/Ϫ cells stimulated with carbachol and CCK that activate the G q -coupled M3 and CCK receptors, respectively. Figs. 3, A and B show that the low concentrations of 0.05 and 0.1 M carbachol had no effect on Ca 2ϩ signaling in WT cells but triggered [Ca 2ϩ ] i oscillations with increased frequency in RGS2 Ϫ/Ϫ cells. Similarly, Fig. 3, D and E show that 0.5 and 2.5 pM CCK had no effect or induced low frequency [Ca 2ϩ ] i oscillations in WT cells and higher frequency [Ca 2ϩ ] i oscillations in RGS2 Ϫ/Ϫ cells. Carbachol at 1 M and CCK at 10 pM triggered [Ca 2ϩ ] i oscillations in WT cells but a large Ca 2ϩ transient in RGS2 Ϫ/Ϫ cells. To determine the dose response for the agonists, at the end of each stimulation period the cells were exposed to a maximal agonist concentration to discharge the residual agonist-mobilizable Ca 2ϩ pool and calculate the extent of Ca 2ϩ mobilization at the lower agonist concentration. The results of these measurements are shown in Fig. 3, C and F, and indicate that the deletion of RGS2 decreased the EC 50 for carbachol from 4.3 Ϯ 0.5 to 0.48 Ϯ 0.07 M and the EC 50 for CCK from 53 Ϯ 5 to 7.3 Ϯ 2.6 pM. Similar measurements with epinephrine stimulation of parotid duct cells revealed that the deletion of RGS2 also decreased the EC 50 for epinephrine in parotid duct cells (not shown). These findings satisfy the two predictions that resulted from the change in the kinetic of agoniststimulated IP 3

production
The results in Figs. 1-3 indicate that in vivo RGS2 is a key regulator of G q -dependent Ca 2ϩ signaling. The combined ef-

FIG. 2. Kinetic of IP 3 production in WT and RGS2 ؊/؊ cells.
A and B, pancreatic acini from WT and RGS2 Ϫ/Ϫ mice were stimulated with 1 M (A) or 1 mM (B) carbachol for the indicated times between 2 and 120 s, and the level of IP 3 was determined. The single asterisk (*) indicates being statistically different from WT at p Ͻ 0.05, and the double asterisks (**) indicate being statistically different from WT at p Ͻ 0.01. C, pancreatic acini from WT (f) and RGS2 Ϫ/Ϫ cells (E) were stimulated for either 5 or 10 s with the indicated concentrations of carbachol, and IP 3 levels were plotted as a function of carbachol concentration. D, pancreatic acini from WT (f) and RGS2 Ϫ/Ϫ cells (E) were stimulated for 120 s with the indicated concentrations of carbachol, and IP 3 levels were plotted as a function of carbachol concentration. fects of RGS2 deletion on the kinetic of IP 3 production and [Ca 2ϩ ] i oscillations at a low agonist concentration provide the first direct evidence for the importance of RGS2 in controlling IP 3 and, therefore, [Ca 2ϩ ] i oscillations in vivo and further support the conclusion that RGS proteins have a primary role in the generation of agonist-evoked [Ca 2ϩ ] i oscillations (9). Another important observation is that the deletion of RGS2 affected Ca 2ϩ signaling by all GPCRs examined and in both pancreatic acinar cells and parotid gland duct cells, including the M3, CCK, bombesin, and ␣-adrenergic receptors. In vivo RGS proteins display receptor specificity (32), with the receptor recognition domain residing at the N terminus of RGS proteins (33). Interestingly, of all RGS proteins examined, only RGS2 showed similar potency for all receptors (32). This result is consistent with the present finding that Ca 2ϩ signaling was similarly affected by all of the G q -coupled receptors examined. Evidently, if RGS proteins control Ca 2ϩ signaling in a receptorspecific manner, other RGS proteins that participate in Ca 2ϩ signaling will fulfill this function. Indeed, multiple RGS proteins are found in a single cell (12), including in a single pancreatic acinar cell (13), and the results in Fig. 4 provide direct evidence that RGS proteins other than RGS2 can regulate Ca 2ϩ signaling in pancreatic acini.
Termination of Cell Stimulation in RGS2 Ϫ/Ϫ Cells-An effect of RGS2 deletion on the kinetic of IP 3 production and Ca 2ϩ signaling raises the question of how the termination of cell stimulation was affected in RGS2 Ϫ/Ϫ cells and whether other RGS proteins participate in Ca 2ϩ signaling. The finding that IP 3 levels in RGS2 Ϫ/Ϫ cells remained transient rather than showing continuous accumulation even at low agonist concentration and that maximal levels of IP 3 were the same in WT and RGS2 Ϫ/Ϫ cells indicate that not all negative controls of GPCR-dependent Ca 2ϩ signaling were eliminated in RGS2 Ϫ/Ϫ cells. This indication is further demonstrated in Fig. 4, in which the effect of RGS2 deletion on the termination of cell stimulation was examined. In these experiments the cells were stimulated for 3-5 s with 1 mM carbachol to completely discharge the ER Ca 2ϩ pool, and stimulation was rapidly and completely terminated by washing away the carbachol and adding the antagonist atropine. We showed previously that this leads to rapid termination of cell stimulation, hydrolysis of IP 3 , and re-uptake of Ca 2ϩ into the ER to reduce [Ca 2ϩ ] i bake to basal levels (34,35). Although the rate of [Ca 2ϩ ] i reduction after the addition of atropine mostly reflects the rate on Ca 2ϩ uptake into the ER, it can be used to determine whether the stimulated state is actively terminated by a negative regulatory mechanism such as that exerted by RGS proteins. Fig. 4 shows that the addition of atropine at the [Ca 2ϩ ] i zenith increased the rate of [Ca 2ϩ ] i reduction in both WT and RGS2 Ϫ/Ϫ cells, but the rate in RGS2 Ϫ/Ϫ cells was 1.34 Ϯ 0.05-fold (n ϭ 7) slower that in WT cells. The fast rate of [Ca 2ϩ ] i reduction in WT and RGS2 Ϫ/Ϫ cells leads to two important conclusions. First, as expected, RGS2 does regulate the rate of signal termination of G q -mediated Ca 2ϩ signaling in vivo. Second, other RGS proteins must participate in the termination of G q -mediated Ca 2ϩ signaling, because atropine did rapidly terminate Ca 2ϩ signaling in RGS2 Ϫ/Ϫ cells. In fact, the small increase in the rate of [Ca 2ϩ ] i reduction in the RGS2 Ϫ/Ϫ cells suggest that other RGS proteins have a major role in terminating G q -mediated Ca 2ϩ signaling. Although at present we cannot exclude the possibility that other RGS proteins were recruited to the Ca 2ϩ signaling complex as a result of the deletion of RGS2, our findings do indicate that RGS proteins other than RGS2 can communicate with G q in the Ca 2ϩ signaling complex to efficiently terminate Ca 2ϩ signaling. These can be any of the RGS proteins found in single pancreatic acinar cells (13) that stimulate the GTPase activity of G␣ q (11).
Adaptation of ER Ca 2ϩ Uptake and Ca 2ϩ Release in RGS2 Ϫ/Ϫ Cells-Resting [Ca 2ϩ ] i levels were the same in WT (58 Ϯ 5) and RGS2 Ϫ/Ϫ (55 Ϯ 4) cells. By contrast, Fig. 5, a and d show that maximal stimulation of WT and RGS2 Ϫ/Ϫ cells incubated in Ca 2ϩ -containing medium increased [Ca 2ϩ ] i to 465 Ϯ 50 and 630 Ϯ 65 nM (n ϭ 20), respectively. A higher [Ca 2ϩ ] i increase in RGS2 Ϫ/Ϫ cells can be due to increased Ca 2ϩ release from the ER, increased Ca 2ϩ entry, or both. To distinguish the contribution of each pathway, the cells were stimulated in Ca 2ϩ -free medium. The results in Fig. 5, b and e indicate that the higher [Ca 2ϩ ] i increase in RGS2 Ϫ/Ϫ cells was due to higher Ca 2ϩ release from the ER, because in Ca 2ϩ -free medium carbachol increased [Ca 2ϩ ] i to 450 Ϯ 40 and 605 Ϯ 55 nM (n ϭ 5) in WT and RGS2 Ϫ/Ϫ cells, respectively. Furthermore, the addition of external Ca 2ϩ to stimulated cells incubated in Ca 2ϩ -free medium resulted in a lower increase in [Ca 2ϩ ] i in RGS2 Ϫ/Ϫ than in WT cells. Higher Ca 2ϩ release requires higher Ca 2ϩ content in the ER of RGS2 Ϫ/Ϫ cells. The Western blot analysis in Fig. 5g shows that the deletion of RGS2 resulted in a compensatory increase in SERCA2b expression by 1.87 Ϯ 0.13-fold.
In agreement with the results in Fig. 4, the addition of atropine rapidly reduced [Ca 2ϩ ] i to the basal level in WT and RGS2 Ϫ/Ϫ cells. Another consistent difference between the two cell types was that after the addition of atropine, [Ca 2ϩ ] i oscillated for at least 5 min in WT but not in RGS2 Ϫ/Ϫ cells (Fig. 5,  a, b, d, and e), and the oscillations required incubating the cells with 5 mM external Ca 2ϩ during carbachol stimulation and the incubation with atropine (not shown) and active SERCA pumps (Fig. 5, c and f). The requirement for high external Ca 2ϩ during cell stimulation and active SERCA pumps to observe the oscillations suggest that overloading the ER with Ca 2ϩ may have sensitized the Ca 2ϩ release process. The lack of [Ca 2ϩ ] i oscillations in RGS2 Ϫ/Ϫ cells under the same conditions, despite the SERCA pump overexpression and ER Ca 2ϩ overload, suggests that the Ca 2ϩ release mechanism in RGS2 Ϫ/Ϫ cells was modified to reduce their excitability. This possibility was tested directly by measuring the potency of IP 3 to release Ca 2ϩ from the ER of WT and RGS2 Ϫ/Ϫ cells.
Adaptation of IP 3 -mediated Ca 2ϩ Release in RGS2 Ϫ/Ϫ Cells- Fig. 6 shows an analysis of expression of IP 3 Rs and IP 3 -mediated Ca 2ϩ release in WT and RGS2 Ϫ/Ϫ cells. The deletion of RGS2 down-regulated expression of IP 3 R1 by ϳ25%, whereas it up-regulated the expression of IP 3 R3 by ϳ30% (Fig. 6, A and B). This resulted in a reduced potency for IP 3  ity of Ca 2ϩ release in RGS2 Ϫ/Ϫ cells. Thus, it seems that the entire ER Ca 2ϩ homeostasis was modified in RGS2 Ϫ/Ϫ cells. The adaptation of IP 3 -mediated Ca 2ϩ release will function to reduce Ca 2ϩ release by the elevated steady-state levels of IP 3 in the RGS2 Ϫ/Ϫ cells (Figs. 1 and 2), and the up-regulation of SERCA2b will speed up removal of [Ca 2ϩ ] i between the [Ca 2ϩ ] i spikes in RGS2 Ϫ/Ϫ cells.
Adaptation of Ca 2ϩ Influx in RGS2 Ϫ/Ϫ Cells-Another aspect of cellular Ca 2ϩ homeostasis is Ca 2ϩ influx and efflux across the plasma membrane. The standard Ca 2ϩ removal and re-addition protocol was used to assay for Ca 2ϩ influx and showed a lower response in RGS2 Ϫ/Ϫ cells (Fig. 5, a, b, d, and e). However, the reduced response can be due to the higher SERCA2b pump expression in RGS2 Ϫ/Ϫ cells that can rapidly remove the Ca 2ϩ entering the cells and establish a lower [Ca 2ϩ ] i plateau. In addition, increased PMCA activity in RGS2 Ϫ/Ϫ cells can contribute to the lower apparent [Ca 2ϩ ] i increase due to the addition of external Ca 2ϩ . In the protocol in Fig. 5, c and f, we isolated the Ca 2ϩ influx activity. The acini were incubated in Ca 2ϩ -free medium, stimulated with 1 mM carbachol to maximally deplete the stores, and incubated with the SERCA pump inhibitor cyclopiazonic acid (CPA) (25 M). This concentration of CPA was maintained thereafter. After the return of [Ca 2ϩ ] i to the basal level the cells were incubated with atropine to terminate the stimulated state and minimize the contribution of PMCA. The cells were then exposed to 5 mM Ca 2ϩ , which increased [Ca 2ϩ ] i to 275 Ϯ 30 nM in WT cells and to 185 Ϯ 20 nM (n ϭ 7) in RGS2 Ϫ/Ϫ cells. These results clearly indicate that another adaptation of Ca 2ϩ signaling in RGS2 Ϫ/Ϫ cells is the down-regulation of agonist-activated Ca 2ϩ influx.
Adaptation of PMCA Activity in RGS2 Ϫ/Ϫ Cells-Next, we tested whether PMCA protein and activity underwent adaptation in RGS2 Ϫ/Ϫ cells. The Western blot analysis in Fig. 7A shows that PMCA protein expression was unaltered in RGS2 Ϫ/Ϫ cells. Two protocols were used to assay PMCA activity. The first is based on the measurement of [Ca 2ϩ ] i in single cells or acini and is shown in Fig. 7B. We reasoned that the activation of PMCA limits the increase of [Ca 2ϩ ] i in response to the addition of external Ca 2ϩ to the agonist-stimulated cell and that PMCA activity can be recovered by the termination of cell stimulation with an antagonist. PMCA activity can then be read from the rate and extent of the [Ca 2ϩ ] i increase induced by the antagonist. As an additional control, the cells can be restimulated with a second agonist that should result in the reactivation of PMCA, Ca 2ϩ efflux, and a reduction in [Ca 2ϩ ] i . The results of this protocol with WT and RGS2 Ϫ/Ϫ cells are shown in Fig. 7B. Cells incubated in Ca 2ϩ -free medium were stimulated with carbachol, and SERCA pumps were inhibited with CPA. While stimulated, the cells were exposed to 5 mM Ca 2ϩ , and after the stabilization of [Ca 2ϩ ] i , stimulation was terminated by incubating the cells with atropine. The inhibition of SERCA pumps ensured that the Ca 2ϩ increase triggered by atropine is due to the unstimulation of PMCA. Unstimulation of PMCA resulted in a larger increase in [Ca 2ϩ ] i in RGS2 Ϫ/Ϫ cells. In WT cells the addition of atropine increased [Ca 2ϩ ] i by 52 Ϯ 8 nM and in RGS2 Ϫ/Ϫ cells by 96 Ϯ 11 nM (n ϭ 12). To further demonstrate that the atropine-induced increase in [Ca 2ϩ ] i is due to the stimulation of PMCA, the cells were re-stimulated with CCK, which reduced [Ca 2ϩ ] i back to the level set by carbachol stimulation.
The second assay of PMCA activity is based on the measurement of the extracellular Ca 2ϩ of a large number of cells or acini suspended in lightly buffered Ca 2ϩ -free medium, as detailed previously (31,36). The stimulation of pancreatic acini with carbachol resulted in Ca 2ϩ efflux, which was monitored as an increase in Fura2 fluorescence in the extracellular medium. Fig. 7C shows that after an initial delay, extracellular Ca 2ϩ increased at a rate of 33 Ϯ 4 nM/min in WT acini and 48 Ϯ 5 nM/min in RGS2 Ϫ/Ϫ acini, confirming the results obtained in acini in Fig. 7B. The results in Fig. 7 used an established protocol and a new protocol to confirm (36) the activation of PMCA by GPCRs and to show the adaptation of PMCA activity in RGS2 Ϫ/Ϫ cells. Hence, as was found for the ER membrane, Ca 2ϩ homeostasis across the plasma membrane was adapted to account for the deletion of RGS2 Ϫ/Ϫ .
Reduced Excitability in RGS2 Ϫ/Ϫ Cells-To examine the overall effect of adaptation of the ER and plasma membrane Ca 2ϩ homeostasis on Ca 2ϩ signaling under physiological conditions, we tested the response of agonist-evoked [Ca 2ϩ ] i oscillations to increased extracellular Ca 2ϩ in WT and RGS2 Ϫ/Ϫ cells. Extracellular Ca 2ϩ is essential, as in the case of CCK, or obligatory, as in the case of carbachol-evoked [Ca 2ϩ ] i oscillations (2). Fig. 8A shows that in WT pancreatic acini stimulated with 0.25 M carbachol, increasing external Ca 2ϩ from 1 to 3 to 7.5 mM increased the frequency of [Ca 2ϩ ] i oscillations by ϳ1.8 Ϯ 0.2 and 2.7 Ϯ 0.3-fold, respectively (Fig. 8C). The [Ca 2ϩ ] i oscillations evoked by 0.25 M carbachol in RGS2 Ϫ/Ϫ cells incubated with 1 mM CaCl 2 occurred at a frequency 1.9 Ϯ 0.03-fold higher than that measured in WT cells. However, in contrast to the findings in WT cells, increasing external Ca 2ϩ from 1 to 3 mM had no effect on the [Ca 2ϩ ] i oscillations in RGS2 Ϫ/Ϫ cells, and further increasing external Ca 2ϩ to 7.5 mM increased the frequency of the oscillations by only 1.3 Ϯ 0.2-fold. FIG. 7. PMCA activity in WT and RGS2 ؊/؊ cells. The expression of PMCA was analyzed by Western blot (a). To estimate PMCA activity in single cells and acini (b), the cells were incubated in Ca 2ϩ -free medium and treated with 1 mM carbachol (Carb) and 25 M CPA. The cells were then exposed to media containing 5 mM Ca 2ϩ and, after stabilization of [Ca 2ϩ ] i , PMCA activity was uncovered by inhibiting the stimulated state with atropine. The cells were then stimulated with CCK to reactivate PMCA. The average PMCA activity was measured in cell suspension in WT and RGS2 Ϫ/Ϫ cells (c), and the columns show the average of three experiments.