Phosphorylation of Inositol 1,4,5-Trisphosphate Receptors in Parotid Acinar Cells A MECHANISM FOR THE SYNERGISTIC EFFECTS OF cAMP ON Ca 2 (cid:1) SIGNALING* inositol phosphate assays and in situ phosphorylation experi-*

Acetylcholine-evoked secretion from the parotid gland is substantially potentiated by cAMP-raising agonists. A potential locus for the action of cAMP is the intracellular signaling pathway resulting in elevated cytosolic calcium levels ([Ca 2 (cid:1) ] i ). This hypothesis was tested in mouse parotid acinar cells. Forskolin dramatically potentiated the carbachol-evoked increase in [Ca 2 (cid:1) ] i , converted oscillatory [Ca 2 (cid:1) ] i changes into a sustained [Ca 2 (cid:1) ] i increase, and caused subthreshold concentrations of carbachol to increase [Ca 2 (cid:1) ] i measurably. This potentiation was found to be independent of Ca 2 (cid:1) entry and inositol 1,4,5-trisphosphate (InsP 3 ) produc- tion, suggesting that cAMP-mediated effects on Ca 2 (cid:1) release was the major underlying mechanism. Consistent with this hypothesis, dibutyryl cAMP dramatically potentiated InsP 3 -evoked Ca 2 (cid:1) release from streptolysin-O-permeabilized cells. Furthermore, type II InsP 3 recep- tors (InsP 3 R) were shown to be directly phosphorylated by a protein kinase A (PKA)-mediated mechanism after treatment with forskolin. in the a paired experimental design was applied, whereby a particular experimental paradigm was repeated in the absence or presence of a test reagent(s) on the same cell. Therefore, statistical significance was determined, where appropriate, using a paired t test, Wilcoxon test for pairs, or one sample t test. Occasionally, statistical significance was determined between groups of experiments where an unpaired t test or Mann-Whitney test was used. For any parameter analyzed from several cells in a particular experiment, an average value was determined. These values were in turn averaged to give the values expressed in the text as means (cid:4) S.E. ( and the combined effect of Bt 2 cAMP and ryanodine ( 4 experiments, 17 on InsP 3 -evoked Ca 2 (cid:1) release. F , quantification of mean data was achieved by expressing InsP 3 -evoked Ca 2 (cid:1) release in the presence of treatment (or 2nd control) relative to that in the absence of treatment (or 1st control) in the same cell. Statistical significance was determined using a paired one sample t test (* p (cid:6) 0.05). This revealed that Ca 2 (cid:1) release was unaltered by repeated stimulation with InsP 3 either over time (0.3 (cid:2) M , 0.94 (cid:4) 0.06-fold change; 3.0 (cid:2) M , 0.96 (cid:4) 0.03-fold change), or following treatment with 500 (cid:2) M ryanodine (0.3 (cid:2) M , 1.02 (cid:4) 0.05-fold change; 3.0 (cid:2) M , 0.93 (cid:4) 0.05-fold change). However, 100 (cid:2) M Bt 2 cAMP dramatically potentiated Ca 2 (cid:1) release evoked by 0.3 (cid:2) M InsP 3 (3.0 (cid:4) 0.5-fold change) but not by 3 (cid:2) M InsP 3 (0.95 (cid:4) 0.06-fold change). The potentiation by Bt 2 cAMP (3.0 (cid:4) 0.5-fold was prevented by Rp-cAMPS (1.1 (cid:4) 0.2-fold but not affected by co-treatment with ryanodine (2.37 0.16-fold

Calcium is a ubiquitous second messenger that is critically important in the regulation of a variety of cellular functions (1)(2)(3). The spatio-temporal "shaping" of Ca 2ϩ signals is thought to play an important role in defining the specificity of stimulusresponse coupling both between cell types and within the same cell (4,5). However, despite intensive investigation, the molec-ular mechanisms that control frequency-and/or amplitudeencoded Ca 2ϩ oscillations, Ca 2ϩ wave propagation, or localized Ca 2ϩ release events remain poorly understood. An emerging body of evidence indicates that, in various systems, specific control over Ca 2ϩ signals may be achieved by cross-talk between second messenger systems that raise [Ca 2ϩ ] i 1 interacting with those that elevate cAMP. Such cross-talk may alter the sensitivity of a variety of Ca 2ϩ transport processes (6,7).
An example of this cross-talk occurs in the salivary gland, where both fluid and exocytotic secretion are controlled by separate neuronal and/or humoral inputs (6,8). Specifically, neuronally released acetylcholine (ACh) activates acinar cell muscarinic receptors, leading to increased [Ca 2ϩ ] i via the phosphoinositide pathway. Elevations in [Ca 2ϩ ] i activate ion channels essential for unidirectional fluid secretion (9), and, in addition, exert regulatory control over the exocytotic machinery required for protein secretion (6). Muscarinic activation of both fluid secretion, and to a lesser extent exocytosis, has been shown to be dramatically potentiated by the concomitant activation of cAMP-raising pathways, such as by co-released vasoactive intestinal peptide, or by sympathetic stimulation of ␤-adrenoreceptors (10 -13). Although cAMP could have direct effects on ion channels (14) and/or exocytotic proteins (15), an alternative hypothesis is that cAMP interacts directly with the Ca 2ϩ signaling machinery to account for the synergistic effects (6,13,16). Because parotid acinar cells have been used extensively to study Ca 2ϩ signaling, this model system is ideally suited to investigate cross-talk between cAMP and Ca 2ϩ signaling.
InsP 3 production, Ca 2ϩ influx, and Ca 2ϩ release from either InsP 3 Rs or RyRs, are all potential targets for cAMP in modulating Ca 2ϩ signaling, however, the literature is equivocal as to the site of any interaction (16 -23). No single study has been successful in unambiguously identifying a specific molecular target that accounts for the synergistic relationship between cAMP and Ca 2ϩ signaling in parotid acinar cells. Therefore, the aim of the present study was to investigate the molecular target(s) for the interaction between cAMP and Ca 2ϩ signaling in mouse parotid acinar cells. This was achieved using a combination of imaging (intact and SL-O-permeabilized cells), inositol phosphate assays and in situ phosphorylation experi-ments. These experimental paradigms revealed that cAMP dramatically potentiated Ca 2ϩ release through PKA-mediated phosphorylation of InsP 3 receptors, likely the type II isoform. This regulatory control likely underlies the synergistic relationship between ACh and cAMP-elevating agonists in parotid acinar cells. These findings have broad implications and may represent a general feature for the regulation of Ca 2ϩ release events that are linked to a vast array of specific functions in all cell types.

EXPERIMENTAL PROCEDURES
Isolation of Single Parotid Acinar Cells-Single and small groups of parotid acinar cells were isolated by collagenase digestion of freshly dissected parotid glands from wild type Swiss Black mice using a technique similar to that described previously for rat parotid (24). Briefly, 25-g mice were killed by cardiac puncture immediately following CO 2 gas asphyxiation. Parotid glands were dissected, minced, and incubated for 60 min at 37°C in Earle's minimum essential medium (Biofluids, Inc., Rockville, MD) containing 2 mM glutamine, 1% bovine serum albumin, and 0.04 mg/ml collagenase P (Roche Molecular Biochemicals, Mannheim, Germany). Minced tissue was dispersed by multiple trituration every 20 min. Cells were resuspended in bovine serum albumin-free Eagle's basal medium (Invitrogen) supplemented with 2 mM glutamine and penicillin/streptomycin and left on ice until ready for use.
Digital Imaging of [Ca 2ϩ ] i -Isolated parotid acinar cells were resuspended in a HEPES-buffered physiological saline solution (HEPES-PSS) containing (in mM) 5.5 glucose, 137 NaCl, 0.56 MgCl 2 , 4.7 KCl, 1 Na 2 HPO 4 , 10 HEPES (pH 7.4), 1.2 CaCl 2 . The cells were then incubated in the above HEPES-PSS containing 2 M fura-2/AM for 30 min at room temperature, after which they were washed once and resuspended in the above HEPES-PSS and kept on ice. Cells were allowed to adhere to a glass coverslip that formed the base of a gravity-fed perfusion chamber and continually perfused with HEPES-PSS. [Ca 2ϩ ] imaging was performed using an inverted epifluorescence Nikon microscope with a 40ϫ oil immersion objective lens (numerical aperture, 1.3). A field of 3-15 fura-2-loaded cells was excited alternately with light at 340 and 380 nm (Ϯ 10 nm bandpass filters, Chroma) using an illumination system (Sutter, DG-4). Fluorescence images (500 Ϯ 45 nm) were captured and digitized at 12-bit resolution using an interline progressive scan CCD camera (Sensicam). Axon Imaging Workbench was used to drive the DG-4 and image acquisition by the camera. Images were acquired every second with an exposure of 200 ms. Background-subtracted and 340/380 ratio images were calculated on-line and stored immediately to hard disk. Images (480 ϫ 640 pixels) were collected with no binning, thereby giving a spatial resolution of 0.225 m/pixel. All experiments were performed at room temperature.
Measurement of Intra-ER Ca 2ϩ ([Ca 2ϩ ] ER )-Isolated parotid acinar cells were incubated in the above attachment media containing 10 M fura-2FF/AM (K D for Ca 2ϩ ϳ13 M; see Ref. 25) for 60 min at 37°C. Permeabilization and subsequent Ca 2ϩ uptake and Ca 2ϩ release experiments were similar to those reported in pancreatic acinar cells (26,27). Briefly, cells were constantly perfused with a Chelex-100 "scrubbed" cytosol-like medium containing (in mM) 135 KCl, 1.2 KH 2 PO 4 , 0.5 EGTA, 0.5 HEDTA, 0.5 nitriloacetic acid, and 20 HEPES/KOH (pH 7.1). MgCl 2 , CaCl 2 , and MgATP were added accordingly to give a constant free Mg 2ϩ , Ca 2ϩ , and ATP concentration of 0.9 mM, 200 nM, and 1 mM, respectively (as calculated by WEBMAXC version 2.1). Permeabilization was achieved by perfusion with an ATP-containing (but Ca 2ϩ -free) cytosol-like medium containing 0.4 IU of streptolysin-O (SL-O). Permeabilization was monitored and verified using the imaging system by exciting loaded cells with light at 360 nm, which is the isosbestic point for fura-2FF (360 Ϯ 25 nm bandpass filter, Chroma). Images were acquired every 10 s (100-ms exposure) and the fluorescence in each cell monitored over time. The emitted fluorescence declined rapidly within 10 min as the plasma membrane permeabilized and all the cytosolic dye leaked out of the cell. The resultant fluorescence (Ͻ20% of pre-permeabilized cells) was caused by dye trapped in intracellular organelles. The cells were subsequently perfused with the cytosol-like medium devoid of SL-O, Ca 2ϩ , or ATP for 10 min. Measurement of [Ca 2ϩ ] ER was achieved by exciting permeabilized cells with light at 340 and 380 nm as with fura-2, except images were acquired every 10 s with an exposure of 300 ms. This was done to prevent photobleaching of the dye because emitted fluorescence was significantly lower than when the dye was trapped in the cytosol. Rapid Ca 2ϩ uptake was achieved upon addition of 1 mM Mg-ATP and 0.2 M Ca 2ϩ , reaching a steady state within 3 min.
Upon loading the stores with Ca 2ϩ , permeabilized cells were stimulated with various concentrations of InsP 3 in the absence or presence of 100 M dibutyryl cyclic AMP (Bt 2 cAMP), 10 M Rp-cAMPS, and/or 500 M ryanodine. Ca 2ϩ release was measured as a decrease in the fura-2FF 340/380 ratio.
Measurement of Inositol Phosphate Production-Total inositol phosphate production in response to forskolin and/or carbachol (CCh) was determined by a similar method to that described previously (28). Briefly, isolated parotid acinar cells were incubated with 5 Ci/ml myo-[ 3 H]inositol for 2 h, followed by three washes in HEPES-PSS. Cells were then incubated with 1 M CCh and/or 10 M forskolin for 5 min in the presence of 10 mM lithium. Total myo-[ 3 H]inositol phosphates from each sample were extracted by solubilization with 0.5 M trichloroacetic acid, eluted on Dowex columns and detected by liquid scintillation. Total inositol phosphate generation was expressed as a percentage of total phosphoinositides, determined by counting the trichloroacetic acid-insoluble fraction by liquid scintillation.
Phosphorylation of Ca 2ϩ Release Channels-Parotid acinar cells were isolated from 4 mice as described above and metabolically labeled by incubating for 2 h with 14 Ci/ml 32 PO 4 Ϫ (PerkinElmer Life Sciences) in a phosphate-free saline solution containing (in mM) 109.7 NaCl, 4.5 KCl, 1.2 MgCl 2 , 5.95 HEPES (free acid), 7.05 NaHEPES (pH 7.4) 1.13 CaCl 2 and 6 glucose. Following incubation, cells were washed three times in the above 32 PO 4 Ϫ -free media and aliquots treated with or without 10 M forskolin for 10 min at 37°C. Cells were then rapidly pelleted by centrifugation and resuspended in ice-cold lysis buffer containing (in mM) 50 Tris-HCl (pH 7.4), 250 NaCl, 5 EDTA, 100 NaF, 1 benzamidine, 1 dithiothreitol, 1% CHAPS, 10 g/ml phenylmethylsulfonyl fluoride, 1 g/ml leupeptin, 0.7 g/ml pepstatin A, 1 g/ml aprotinin, 1 g/ml antipain. Cell lysates were then sonicated, left on ice for 30 min, and vortexed every 5 min. InsP 3 R or RyR protein was immunoprecipitated from the samples by incubating lysates with a mixture of antibodies (ϳ1 g of antibody/mg of protein) raised against all three InsP 3 R types or all three RyR types for 1 h at 4°C, followed by incubation with 80 l of protein A-agarose beads (Pierce) for an additional 1 h at 4°C. Antibodies directed against InsP 3 R were CT 1 and CT 2 (kind gifts from Richard Wojcikiewicz, State University of New York, Syracuse, NY) and type III antibody (Transduction Laboratories). Antibodies directed against RyR were 34C directed against RyR 1 (Developmental Studies Hybridoma Bank, Iowa City, IA), C3-33 directed against RyR 2 (Affinity Bioreagents, Inc.), and anti-rabbit skeletal muscle RyR antibody directed against RyR 3 (Upstate Biotechnology, Lake Placid, NY). As a secondary control, aliquots of cell lysate from cells treated with forskolin were incubated with beads without any antibodies. Following incubation, the beads-protein complexes were washed five times in lysis buffer by repeated centrifugation and resuspension and then boiled in SDS-sample buffer (Laemmli). Immunoprecipitated proteins were separated by SDS-polyacrylamide gel (5%) electrophoresis (SDS-PAGE), after which the gels were vacuum-dried (Bio-Rad) and exposed to a PhosphorImager intensifier screen. Bands of the appropriate molecular weight, which corresponded to phosphorylated proteins that had incorporated 32 PO 4 Ϫ , were detected using a Molecular Dynamics PhosphorImager.
To determine whether phosphorylation of InsP 3 Rs was mediated by PKA, H-89 and Rp-cAMPS were used to inhibit PKA in combination with an additional and complimentary approach using a phospho-PKA substrate antibody (Cell Signaling Technology). This antibody specifically recognizes proteins containing phosphorylated serine or threonine, with an arginine residue at position Ϫ3, but not the corresponding nonphosphorylated motif (29,30). Although this antibody would not discriminate between substrates of PKA, protein kinase C, or cyclic GMP-dependent protein kinase, the combined use of forskolin and appropriate PKA inhibitors was used to implicate a specific role of PKA. Aliquots of isolated parotid acinar cells were treated with or without 10 M forskolin, 50 nM okadaic acid, and/or the PKA inhibitors, H-89 (2 M) and Rp-cAMPS (30 M) for 10 min at 37°C. Cells were then solubilized in lysis buffer similar to the method above. Lysates were then incubated with phospho-PKA substrate antibody (1:100 dilution) to immunoprecipitate phosphorylated proteins. Specific detection of phosphorylated type II InsP 3 Rs was achieved by Western blotting with the CT 2 antibody. Whole cell lysates or immunoprecipitated proteins were denatured in SDS-sample buffer, separated by SDS-PAGE, and transferred to nitrocellulose (Schleicher & Schuell) prior to immunoblotting with the CT 2 antibody. Immunoreactivity was visualized using peroxidaseconjugated secondary antibodies (Bio-Rad), followed by detection by Super Signal detection system (Pierce) exposed on XAR film (Eastman Kodak Co.).
Data Analysis and Experimental Design-In all experiments (unless otherwise stated in the text), a paired experimental design was applied, whereby a particular experimental paradigm was repeated in the absence or presence of a test reagent(s) on the same cell. Therefore, statistical significance was determined, where appropriate, using a paired t test, Wilcoxon test for pairs, or one sample t test. Occasionally, statistical significance was determined between groups of experiments where an unpaired t test or Mann-Whitney test was used. For any parameter analyzed from several cells in a particular experiment, an average value was determined. These values were in turn averaged to give the values expressed in the text as means Ϯ S.E.

Carbachol-evoked [Ca 2ϩ
] i Changes-To investigate the effects of elevated cAMP levels on [Ca 2ϩ ] i signaling in parotid acinar cells, we first characterized the types of [Ca 2ϩ ] i responses evoked by the muscarinic receptor agonist, CCh. Low concentrations of CCh (10 -300 nM) caused an oscillatory increase in [Ca 2ϩ ] i in 86% of cells tested, which was characterized by a large initial spike followed by rapid sinusoidal oscillations superimposed over an elevated base line. These oscillations were complex in nature, and their frequency and amplitude varied markedly between cells. Higher concentrations of CCh (300 nM to 10 M) induced a biphasic increase in [Ca 2ϩ ] i , which was characterized by a large initial spike followed by a sustained elevation. These patterns of [Ca 2ϩ ] i changes are typical of a variety of exocrine cells; however, the oscillation frequency was significantly higher than reported in pancreatic acinar cells (7-11/min in parotid compared with 4 -6/min; see Ref. 31). Despite the complex nature of these CCh-evoked [Ca 2ϩ ] i changes, there was a consistent concentration-dependent increase in the magnitude of the initial spike-like increase in [Ca 2ϩ ] i . This initial [Ca 2ϩ ] i spike was interpreted to reflect Ca 2ϩ release from intracellular stores and was quantitatively compared in the absence and presence of 10 M forskolin using a paired experimental design.
Forskolin Potentiates Carbachol-evoked [Ca 2ϩ ] i Changes-Using the adenylate cyclase activator forskolin, we investigated the effects of elevating intracellular cAMP levels on CCh-evoked [Ca 2ϩ ] i signaling (33). Repeated stimulations with CCh evoked [Ca 2ϩ ] i changes of equal magnitude (Fig. 1). Forskolin (10 M) induced a dramatic and time-dependent potentiation of this CChevoked initial increase in [Ca 2ϩ ] i . Upon removal of forskolin, there was an equivalent time-dependent recovery ( Fig. 1). On average, forskolin increased the CCh-evoked [Ca 2ϩ ] i response by 148.9 Ϯ 8.5% after 3-5 min, which increased further to 177.1 Ϯ 17.4% after 8 -10 min of incubation with forskolin (see Fig. 1). To test whether the potentiation was the result of specific activation of PKA, cells were also pretreated with 2 M H-89, an inhibitor of serine/threonine kinases such as PKA (34) prior to specific activation of PKA by treatment with forskolin. This completely prevented the potentiation of the CCh-evoked initial increase in [Ca 2ϩ ] i (Fig. 2), suggesting that the potentiation was caused by activation of PKA.
In addition to the potentiation of the CCh-evoked initial increase in [Ca 2ϩ ] i, forskolin also converted oscillatory [Ca 2ϩ ] i changes into a sustained increase, suggesting a leftward shift in the CCh concentration-response curve compared with control cells (Fig. 3A). Consistent with this hypothesis, forskolin treatment enabled normally subthreshold concentrations of CCh (3-30 nM) to evoke oscillatory [Ca 2ϩ ] i responses (49 of 75 cells). Interestingly, in six of these cells, [Ca 2ϩ ] i oscillations were confined to the apical region of the cells (Fig. 3B), revealing conversion of a subthreshold response into a measurable threshold response. To identify the molecular site at which this potentiation was manifested, we systematically investigated the effects of forskolin on Ca 2ϩ entry, InsP 3 production, and Ca 2ϩ release channels in parotid acinar cells using a variety of biochemical and functional assays.  4, A and B; 122.1 Ϯ 11.3% higher than control), suggesting an effect on capacitative Ca 2ϩ entry. This most likely reflected an indirect effect, resulting from the enhanced Ca 2ϩ release and store depletion, on Ca 2ϩ entry. However, additional direct effects of cAMP on Ca 2ϩ entry cannot be completely excluded.
Effects of Forskolin on InsP 3 Production-Activation of muscarinic receptors leads to the generation of InsP 3 through Gprotein-coupled activation of phospholipase C (PLC). Thus, the effects of forskolin could conceivably be caused by an interaction of cAMP with this receptor-signaling complex leading to an enhancement of InsP 3 generation. To test this idea we examined the effects of forskolin on inositol phosphate production by measuring [ 3 H]inositol incorporation to assess PLC activity and thus InsP 3 production. CCh (1 M) significantly increased inositol phosphates from 5.9 Ϯ 0.2 to 9.8 Ϯ 0.1% of total phosphoinositides (Fig. 5). In contrast, forskolin (10 M) had no effect on either basal (6.1 Ϯ 0.1%) or CCh-evoked (8.9 Ϯ 0.3%) inositol phosphate turnover (Fig. 5). This indicated that increased cAMP does not directly stimulate the production of inositol phosphates and suggests a likely site of action is the Ca 2ϩ release process itself.
Effects of Ryanodine on Carbachol-evoked [Ca 2ϩ ] i Changes-In nonexcitable cells, InsP 3 Rs are generally thought to be the trigger for Ca 2ϩ release, whereas RyRs may have a role in propagating further release by Ca 2ϩ -induced Ca 2ϩ release (CICR) (35). We pharmacologically separated these two Ca 2ϩ release pathways by inhibiting RyRs with 500 M ryanodine (36). Ryanodine alone failed to significantly affect the initial increase in [Ca 2ϩ ] i evoked by CCh (101.5 Ϯ 7.3%; Fig. 6, A and C), suggesting that this event does not involve RyRs. However, when applied during a train of Ca 2ϩ oscillations, ryanodine dramatically dampened the oscillatory [Ca 2ϩ ] i response (Fig. 6,  A and B). These data therefore imply that RyRs may be important for propagating and maintaining Ca 2ϩ oscillations in parotid acinar cells as is the case in pancreatic acinar cells (32).
When applied in combination with 10 M forskolin, ryanodine significantly reduced the potentiation of the CCh-evoked [Ca 2ϩ ] i response from 177.1 Ϯ 17.4% to 125.2 Ϯ 2.8% of that observed in the absence of forskolin and ryanodine (Fig. 6, B  and C). Nevertheless, the residual potentiation (125.2 Ϯ 2.8%) remained significantly different from control. One possible interpretation of these data is that the potentiation of Ca 2ϩ release by forskolin is caused by a direct effect of PKA on both RyRs and InsP 3 Rs.
Direct Activation of RyR-To address whether the potentiation by forskolin was caused by a direct effect of PKA on RyRs, we attempted to selectively activate RyRs. Caffeine (20 mM, n ϭ 5 experiments, 22 cells) or low concentrations of ryanodine (0.01-10 M, n ϭ 8 experiments, 49 cells) consistently failed to affect resting [Ca 2ϩ ] i either in the absence or presence of forskolin (Fig. 7, A and B). In addition, 10 M forskolin failed to affect resting [Ca 2ϩ ] i in any cell tested either in the presence or absence of ryanodine (data not shown). Because the effects of ryanodine are use-dependent (see Ref. 36, and references therein), two separate approaches were utilized to facilitate activation of RyRs. First, 10 M ryanodine was added during a train of CCh-evoked [Ca 2ϩ ] i oscillations (see Fig. 7C) and continually applied after CCh was removed (similarly to Ref. 37). The second approach was to increase external [Ca 2ϩ ] to 10 mM in an attempt to progressively elevate resting [Ca 2ϩ ] i , thereby increasing the sensitivity of RyRs to ryanodine (see Fig. 7D). However, ryanodine, either alone or in combination with forskolin, failed to affect resting [Ca 2ϩ ] i in any cell tested (see Fig.  7, C (three separate experiments, 15 cells) and D (four separate experiments, 29 cells). In contrast to conventional activators, such as caffeine or ryanodine, 4-chloro-m-cresol (CmC), a compound well documented to specifically activate RyRs (38 -41), caused slow, concentration-dependent and readily reversible increases in resting [Ca 2ϩ ] i (Fig. 7E, three experiments, 14  cells). However, in contrast to CCh-evoked [Ca 2ϩ ] i increases, 10 M forskolin failed to significantly affect the CmC-evoked [Ca 2ϩ ] i increases (Fig. 7F, five experiments, 33 cells). This indicates that RyR-mediated Ca 2ϩ release in mouse parotid acinar cells is not directly affected by raising cAMP with forskolin, and suggests that PKA has a direct effect on InsP 3 Rs. The role of RyRs may simply be to amplify the enhanced Ca 2ϩ release from InsP 3 Rs by CICR.
InsP 3 -evoked Ca 2ϩ Release from SL-O-permeabilized Cells-To directly assess the effects of cAMP on InsP 3 -evoked Ca 2ϩ release, SL-O-permeabilized cells were used. Despite permeabilization, cells retained their polarity as indicated by apically located secretory granules. Following permeabilization, fluorescence became both dramatically reduced (Ͻ20%) and highly punctate, indicative of dye trapped within organelles (26,42). Perfusion of permeabilized cells with a "cytosolic-like" medium devoid of Ca 2ϩ and ATP caused a slow decline in the fura-2FF-340/380 ratio, which stabilized in 5-10 min, indicating Ca 2ϩ store depletion. The subsequent addition of 0.2 M Ca 2ϩ and 1 mM Mg-ATP evoked a rapid increase in fura-2FF-340/380 ratio that reached a steady state within 3 min (Fig. 8,  A-E), reflecting rapid uptake of Ca 2ϩ into the ER. The observed rate was significantly faster than that reported in pancreatic acinar cells (ϳ10 min to reach steady state; see Refs. 26 and 27).
Following Ca 2ϩ uptake, InsP 3 (0.1-10 M) evoked a concentration-dependent stepwise decrease in [Ca 2ϩ ] ER . For each experiment, mean data were fit to a sigmoidal dose-response curve from which a mean half-maximum concentration (EC 50 ) was determined (Fig. 8F, inset; five separate experiments, 41 cells). This revealed a very steep InsP 3 concentration-response relationship for Ca 2ϩ release with an EC 50 of 0.64 Ϯ 0.11 M and maximal release at 10 M InsP 3 , similar to previous reports (22,43,44). InsP 3 -evoked Ca 2ϩ release was also readily reversible, because removal of InsP 3 in the continued presence of Ca 2ϩ and ATP caused a rapid return of [Ca 2ϩ ] ER to near

FIG. 5. InsP 3 generation is not directly affected by forskolin treatment. [ 3 H]Inositol phosphate production, measured by incorporation of [ 3 H]inositol into phospholipid, was assessed as an indirect meas
ure of InsP 3 generation. Inositol phosphate production was expressed as % of total phosphoinositides. 1 M CCh significantly increased inositol phosphates from 5.9 Ϯ 0.2% to 9.8 Ϯ 0.1% of total phosphoinositides. Forskolin alone (6.1 Ϯ 0.1%) or in combination with 1 M CCh (8.9 Ϯ 0.3%) did not significantly affect inositol phosphate production. Statistical significance was determined by unpaired students t test (* p Ͻ 0.05, n ϭ 3).

FIG. 6. Effects of inhibition of RyRs on CCh-evoked [Ca 2؉ ] i responses and the potentiation by forskolin.
A, 500 M ryanodine dampened the oscillatory [Ca 2ϩ ] i response but failed to significantly affect the initial CCh-evoked [Ca 2ϩ ] i increase. B, using a similar experimental paradigm and in combination with 10 M forskolin, ryanodine reduced the potentiation of the initial CCh-evoked [Ca 2ϩ ] i increase by forskolin. The dotted line represents a gap of 7 to 9 min during which no images were acquired. C, quantification of mean data confirmed that ryanodine failed to significantly affect the initial CCh-evoked [Ca 2ϩ ] i increase (101.5 Ϯ 7.3% increase), but significantly reduced the potentiation by forskolin (125.2 Ϯ 2.8% increase compared with 177.1 Ϯ 17.4% in the absence of ryanodine; p Ͻ 0.05 as determined by Mann Whitney test). However the residual potentiation in the presence of both ryanodine and forskolin remained significant (*p Ͻ 0.05 as determined by one sampled t test).
pre-stimulatory levels, suggesting Ca 2ϩ was rapidly taken back up into the ER (Fig. 8, A-F). This allowed the application of a general experimental paradigm consisting of stimulating with a low dose of InsP 3 (0.3 M), which evoked 24.2 Ϯ 0.1% maximal release, followed by a high dose (3 M), which evoked nearly maximal release (93.1 Ϯ 0.2%; Fig. 8F, inset). This paradigm was then repeated in the presence of 100 M Bt 2 cAMP (Fig. 8B) with or without 10 M Rp-cAMPS or 500 M ryanodine (Fig. 8,  C-E). Time-matched controls were performed to account for loss of dye, photobleaching, or possible desensitization (Fig.  8A). Fig. 8F shows the mean data expressed as -fold increase in Ca 2ϩ released by 0.3 and 3.0 M InsP 3 . Time-matched controls revealed that repeated stimulations with InsP 3 were not significantly different (Fig. 8, A and F). In contrast 100 M Bt 2 cAMP dramatically potentiated Ca 2ϩ release evoked by 0.3 M InsP 3 (3.0 Ϯ 0.5-fold increase; Fig. 8, B and F) but did not significantly effect Ca 2ϩ release evoked by 3 M InsP 3 (0.95 Ϯ 0.06-fold; Fig. 8F). This suggested that cAMP increased the sensitivity of InsP 3 -evoked Ca 2ϩ release without affecting the capacity of the stores to release Ca 2ϩ . Consistent with the effects of forskolin in intact cells, Bt 2 cAMP failed to evoke Ca 2ϩ release when applied in the absence of InsP 3 (Fig. 8B). In addition, 10 M Rp-cAMPS, an inhibitor of PKA that competes for the cAMP-binding site (45), completely prevented the potentiation of InsP 3 -evoked Ca 2ϩ release by Bt 2 cAMP (Fig. 8, C  and F). This is further supportive of data from intact cells suggesting that the potentiation is PKA-mediated.
Similar to the effects of ryanodine on CCh-evoked Ca 2ϩ release in intact cells, ryanodine also failed to significantly affect Ca 2ϩ release from permeabilized cells evoked by either 0.3 M (1.02 Ϯ 0.05-fold change) or 3 M InsP 3 (0.93 Ϯ 0.05-fold change; Fig. 8, D and F (four separate experiments, 20 cells)). However, in contrast to intact cells, ryanodine failed to affect the potentiation of InsP 3 -evoked Ca 2ϩ release by Bt 2 cAMP; release stimulated by 0.3 M InsP 3 increased 2.37 Ϯ 0.16-fold, whereas release stimulated by 3 M InsP 3 was unaffected (1.04 Ϯ 0.01-fold; Fig. 8, E and F (four separate experiments, 17  cells). Therefore, the lack of effect of ryanodine in SL-O-permeabilized cells suggests that, under these conditions, RyRs are functionally uncoupled from InsP 3 Rs. Collectively, these data add credence to the notion that InsP 3 Rs, and not RyRs, are directly modulated by PKA.
In Situ Phosphorylation of InsP 3 R and RyRs-To examine whether the potentiation of Ca 2ϩ release correlated with PKAmediated phosphorylation of InsP 3 Rs and/or RyRs, two complimentary approaches were adopted using receptor specific antibodies to immunoprecipitate potentially phosphorylated protein. First, cells were metabolically labeled with 32 PO 4 Ϫ and protein that had incorporated 32 PO 4 Ϫ upon phosphorylation was detected by autoradiography. Following treatment with forskolin and subsequent immunoprecipitation with InsP 3 R-specific antibodies, there was enhanced labeling of bands at the expected molecular weight for InsP 3 Rs compared with control (Fig. 9A). This demonstrates that InsP 3 Rs are directly phosphorylated, presumably by a PKA-mediated process. In contrast, under similar conditions using RyR-specific antibodies, no detectable signal was observed at the appropriate molecular weight for RyRs. Immunoprecipitation of RyR protein from skeletal muscle, cardiac muscle, and brain tissue lysates confirmed the avidity of RyR-specific antibody (data not shown).
Second, an alternative and complimentary approach used the phospho-PKA substrate antibody to immunoprecipitate phosphorylated protein, followed by specific detection of phosphorylated type II InsP 3 R by immunoblot with the CT 2 antibody. The specificity of the phospho-PKA substrate antibody for phosphorylated protein was verified in Fig. 9B. A variety of proteins of different molecular weight were detected upon treatment with forskolin compared with untreated cells even though equal amounts of total protein were added into each lane (Fig. 9B, lane 3 versus lane 1). The phospho-PKA substrate antibody does not distinguish between substrates of PKA, protein kinase C, and cyclic GMP-dependent protein kinase (29,30). To distinguish between activation of these kinases, cells were treated with the PKA inhibitors H-89 (2 M) and Rp-cAMPS (30 M) prior to and in the continued presence of forskolin to specifically activate PKA. This completely prevented the detection of visible bands (Fig. 9B, lane 4), which strongly suggests that forskolin evokes PKA-mediated phosphorylation of proteins. Furthermore, treatment of cells with the protein phosphatase inhibitor okadaic acid (50 nM) did not dramatically increase protein phosphorylation, suggesting that basal protein phosphatase activity was low and not affected by cell lysis (Fig.  9B, lanes 5 and 6). Finally, immunoprecipitation with the phospho-PKA substrate antibody and subsequent immunoblot with CT 2 antibody showed that forskolin dramatically increased the detection of phosphorylated type II InsP 3 R (Fig. 9C, lanes 4 and  5), which was prevented by pretreatment with H-89 and Rp-cAMPS (Fig. 9C, lanes 6 and 7). Taken together, these data provide convincing evidence that forskolin causes PKA-mediated phosphorylation of type II InsP 3 R. DISCUSSION The present study provides functional and molecular evidence defining a specific site at which cAMP-raising agonists exert synergistic regulatory control over fluid secretion and exocytosis in mouse parotid acinar cells stimulated by muscarinic activation of [Ca 2ϩ ] i signaling (10 -13, 16). The signal transduction pathway following muscarinic receptor stimulation and leading to an increase in intracellular Ca 2ϩ is a rich source of sites for possible PKA regulation, which could account for the dramatic PKA-mediated potentiation of the CCh-evoked initial [Ca 2ϩ ] i increase. The effects of raising cAMP on various aspects of this signaling pathway were therefore systematically investigated. Specifically, Ca 2ϩ entry has been suggested to be important not only in replenishing the stores with Ca 2ϩ but also in maintaining agonist-evoked sustained [Ca 2ϩ ] i signals (4,46), and in modulating the frequency of [Ca 2ϩ ] i oscillations by sensitizing Ca 2ϩ release (47). In addition, in hepatocytes (48) and submandibular acinar cells (49), cAMP has been shown to potentiate InsP 3 production. However, in the present study, forskolin treatment was shown to have no appreciable effect on Ca 2ϩ entry or CCh-evoked PLC activity and InsP 3 production. These data led us to conclude that the primary site of action of PKA is on the Ca 2ϩ release process itself. InsP 3 -evoked Ca 2ϩ release from InsP 3 Rs underlies agonistinduced increases in [Ca 2ϩ ] i in a variety of nonexcitable cells including secretory epithelia (4,32). We observed in a proportion of cells that forskolin altered the sensitivity of CCh-evoked [Ca 2ϩ ] i signals in such a way that apically confined [Ca 2ϩ ] i signals were initiated from subthreshold CCh concentrations. In a variety of exocrine cells, the apical region is regarded as the "trigger zone" from which Ca 2ϩ waves are initiated (32,50,51). Conceptually, such a region could be established by the relative abundance of Ca 2ϩ release channels and/or expression of the most sensitive channels. Both of these criteria are satisfied in rat parotid acinar cells; immunocytochemical studies have reported that the extreme apical region is highly enriched in all InsP 3 R types (18,19). Furthermore, of the three InsP 3 Rs expressed, the type II InsP 3 R, which has been reported to be the most sensitive to InsP 3 (52,53), is the most abundant InsP 3 R type in parotid acinar cells (18). Collectively, this supports the idea that cAMP potentiates Ca 2ϩ release from the most sensitive InsP 3 Rs located in the apical region of the cell.
Ca 2ϩ -induced Ca 2ϩ release is important for the generation of both localized "spikes" and global increases in [Ca 2ϩ ] i (4,32), and it is conceivable that modulation of CICR could contribute to the observed cAMP-mediated potentiation. Although InsP 3 Rs themselves possess all the properties to support CICR (54), there is an emerging body of evidence that RyRs are important for CICR in nonexcitable cells (5,35,55). Both InsP 3 Rs and RyRs have been shown to contain putative consensus sequences for phosphorylation by PKA (56,57) and in some cases phosphorylation results in enhanced release (58 -60). However, in mouse parotid acinar cells, we found that inhibition of RyRs with high concentrations of ryanodine failed to significantly affect the CCh-evoked initial increase in [Ca 2ϩ ] i , consistent with the notion that the mechanism underlying this release primarily involves InsP 3 Rs. When applied during a train of Ca 2ϩ oscillations, however, ryanodine dramatically inhibited oscillations, similarly to previous studies (35,61). Of particular interest, ryanodine significantly reduced the potentiation of the CCh-evoked [Ca 2ϩ ] i response by forskolin (Fig. 6). An obvious explanation for this result is the direct phosphorylation of RyRs by PKA. However, we were unable to obtain any functional or biochemical evidence for this phosphorylation in parotid acinar cells. For example, various maneuvers designed to isolate Ca 2ϩ release through RyR were not augmented by raising cAMP levels, and attempts to demonstrate direct phosphorylation of RyRs were unsuccessful. How can the effects of ryanodine on the initial PKA-potentiated release be reconciled with these data? One possibility is that the apparent contribution of RyRs after PKA treatment is simply the result of the enhanced release of Ca 2ϩ by InsP 3 Rs, resulting in an increased [Ca 2ϩ ] in the vicinity of RyR, thereby increasing the possibility of CICR. This implies that RyRs in parotid acinar cells act to amplify Ca 2ϩ signals that originate from the primary trigger source, which is InsP 3 Rs, but only do so when a threshold [Ca 2ϩ ] i (or microdomain of Ca 2ϩ ) is established. This is consistent with studies showing that RyR 3 is the predominant RyR type expressed in mouse parotid acinar cells, albeit in relatively low abundance compared with skeletal muscle (62). Single-channel data indicate that RyR 3 exhibit very low activity at [Ca 2ϩ ] lower than 1 M, but are dramatically activated by [Ca 2ϩ ] above 1 M (63). As such, RyRs would be functionally uncoupled at rest and only become activated by microdomains of Ca 2ϩ created by Ca 2ϩ released from neighboring InsP 3 Rs.
Activation of PKA with Bt 2 cAMP failed to evoke Ca 2ϩ release from SL-O-permeabilized cells (Fig. 8B), a result consistent with the lack of effect of forskolin in intact cells and in agreement with similar studies using imaging of calcium green C18-labeled permeabilized parotid acinar cells (20). In contrast, cAMP has been demonstrated to evoke Ca 2ϩ release from static suspensions of permeabilized cells and microsomal vesicles of rat parotid acinar cells (18,(21)(22)(23), a result that was attributed to activation of RyRs. Of interest in these studies is the fact that cAMP elevation evoked a relatively small Ca 2ϩ release (ϳ35 nM) compared with InsP 3 (ϳ150 nM) (18,21) and that this release occurred under conditions where basal [Ca 2ϩ ] was estimated to be as high as 431 nM (22). At these [Ca 2ϩ ], any RyRs would be expected to be in a sensitized state. The discrepancy between the present study and the aforementioned report underscores the difficulty in extrapolating averaged Ca 2ϩ release from permeabilized cell suspensions to Ca 2ϩ release in intact cells where resting [Ca 2ϩ ] i is likely to be closer to ϳ100 nM. Alternatively, the presence of such a ryanodine-sensitive cAMP-dependent Ca 2ϩ release pathway, not observed in this study, may simply be the result of species differences. As stated previously, mouse parotid acinar cells express predominantly RyR 3 (62), whereas rat parotid acinar cells (revealed by reverse transcription-PCR) express predominantly RyR 1 (21). Because RyR 1 exhibit a lower threshold for activation by Ca 2ϩ (63), any potentiation of Ca 2ϩ release by cAMP would be amplified under conditions of elevated resting [Ca 2ϩ ] i (18,21,22). Such species differences could also explain the lack of effect of caffeine in mouse acinar cells, as RyR 3 are thought to be caffeineinsensitive (64,65).
The lack of effect of ryanodine on InsP 3 -evoked Ca 2ϩ release in SL-O-permeabilized parotid acinar cells suggests that InsP 3 Rs and RyRs are functionally uncoupled under these conditions, presumably as a result of high perfusion rates resulting in an effective infinite volume mimicking the cytoplasm. As such, microdomains of Ca 2ϩ will likely fail to establish at Ϫ , were incubated with (ϩ) or without (Ϫ) 10 M forskolin for 10 min and samples were prepared as detailed under "Experimental Procedures." Treatment with forskolin dramatically increased the intensity of a band representative of a ϳ250 kDa protein, indicative of phosphorylated InsP 3 Rs. B, verification of Phospho-PKA substrate antibody in detecting phosphorylated proteins upon treatment of cells with (ϩ) or without (Ϫ) 10 M forskolin, 2 M H-89, 30 M Rp-cAMPS and/or 50 nM okadaic acid. Cell lysates were prepared as detailed under "Experimental Procedures," and protein samples run on a 7.5% SDS-polyacrylamide gel. Phosphorylated proteins were detected by Western blotting with the Phospho-PKA substrate antibody. C, detection of phosphorylated type II InsP 3 Rs upon treatment of cells with (ϩ) or without (Ϫ) 10 M forskolin, 2 M H-89 or 30 M Rp-cAMPS. Phosphorylated proteins were immunoprecipitated with the Phospho-PKA substrate antibody, run on a 5% SDS-polyacrylamide gel and phosphorylated type II InsP 3 Rs were detected by Western blotting with the CT 2 antibody. * in A and C represent a secondary control whereby samples were treated identically but immunoprecipitating antibodies were omitted and represent nonspecific binding to protein-A beads.
sufficiently high levels to evoke CICR. In addition, many cytosolic components such as mobile buffers, which may be important for CICR, will be lost in permeabilized cells. Thus, this system allowed us to study the functioning of InsP 3 Rs in isolation. The data obtained indicate that the potentiation of CCh-evoked [Ca 2ϩ ] i signals is mediated through InsP 3 Rs, because activation of PKA still revealed potentiation under conditions where RyRs were functionally uncoupled. In support of this, treatment with forskolin caused phosphorylation of type II InsP 3 Rs; the weight of evidence presented suggests a PKAmediated mechanism. There is an emerging consensus that PKA may regulate Ca 2ϩ release events directly at the level of the Ca 2ϩ release channel. Previous studies have shown that InsP 3 Rs contain several consensus sequences for PKA-mediated phosphorylation (56) and can be phosphorylated both in vitro (66,67) and in vivo (68,69). However, there exists no clear consensus as to the physiological consequence of InsP 3 R-type-specific phosphorylation in terms of Ca 2ϩ release. For example, phosphorylation of type I InsP 3 R from cerebellum caused a decrease in Ca 2ϩ release (66,70), whereas others have shown an increase in Ca 2ϩ release (71). In addition, phosphorylation of type III InsP 3 R in pancreatic acinar cells results in reduced Ca 2ϩ release at low doses of InsP 3 (72), whereas phosphorylation of type II InsP 3 R in liver enhances Ca 2ϩ release (58,60). This latter observation is of particular interest to the present study, as the type II InsP 3 Rs are the dominant form expressed in parotid acinar cells (18). Despite that in vitro phosphorylation studies revealed that type II InsP 3 Rs were relatively poor substrates for PKA phosphorylation (69), the present study convincingly demonstrated that forskolin treatment causes PKA-mediated phosphorylation of type II InsP 3 Rs in mouse parotid acinar cells.
In summary, PKA-mediated phosphorylation of type II InsP 3 Rs is a predominant mechanism for the potentiation of Ca 2ϩ signaling in mouse parotid acinar cells. This largely explains the synergistic relationship between cAMP-raising agonists and ACh-evoked secretion in the parotid. These findings support the emerging consensus that phosphoregulation of InsP 3 Rs is an important mechanism underlying the "shaping" of Ca 2ϩ signals, and thus have broad implications for the fidelity of Ca 2ϩ -mediated processes.