Targeted Phosphorylation of Inositol 1,4,5-Trisphosphate Receptors Selectively Inhibits Localized Ca2+ Release and Shapes Oscillatory Ca2+ Signals*

The current study provides biochemical and functional evidence that the targeting of protein kinase A (PKA) to sites of localized Ca2+ release confers rapid, specific phosphoregulation of Ca2+ signaling in pancreatic acinar cells. Regulatory control of Ca2+ release by PKA-dependent phosphorylation of inositol 1,4,5-trisphosphate (InsP3) receptors was investigated by monitoring Ca2+ dynamics in pancreatic acinar cells evoked by the flash photolysis of caged InsP3prior to and following PKA activation. Ca2+ dynamics were imaged with high temporal resolution by digital imaging and electrophysiological methods. The whole cell patch clamp technique was used to introduce caged compounds and to record the activity of a Ca2+-activated Cl− current. Photolysis of low concentrations of caged InsP3 evoked Cl−currents that were inhibited by treatment with dibutryl-cAMP or forskolin. In contrast, PKA activators had no significant inhibitory effect on the activation of Cl− current evoked by uncaging Ca2+ or by the photolytic release of higher concentrations of InsP3. Treatment with Rp-adenosine-3′,5′-cyclic monophoshorothioate, a selective inhibitor of PKA, or with Ht31, a peptide known to disrupt the targeting of PKA, largely abolished forskolin-induced inhibition of Ca2+ release. Further evidence for the targeting of PKA to the sites of Ca2+mobilization was revealed using immunocytochemical methods demonstrating that the RIIβ subunit of PKA was localized to the apical regions of acinar cells and co-immunoprecipitated with the type III but not the type I or type II InsP3 receptors. Finally, we demonstrate that the pattern of signaling evoked by acetylcholine can be converted to one that is more “CCK-like” by raising cAMP levels. Our data provide a simple mechanism by which distinct oscillatory Ca2+ patterns can be shaped.

The current study provides biochemical and functional evidence that the targeting of protein kinase A (PKA) to sites of localized Ca 2؉ release confers rapid, specific phosphoregulation of Ca 2؉ signaling in pancreatic acinar cells. Regulatory control of Ca 2؉ release by PKA-dependent phosphorylation of inositol 1,4,5trisphosphate (InsP 3 ) receptors was investigated by monitoring Ca 2؉ dynamics in pancreatic acinar cells evoked by the flash photolysis of caged InsP 3 prior to and following PKA activation. Ca 2؉ dynamics were imaged with high temporal resolution by digital imaging and electrophysiological methods. The whole cell patch clamp technique was used to introduce caged compounds and to record the activity of a Ca 2؉ -activated Cl ؊ current. Photolysis of low concentrations of caged InsP 3 evoked Cl ؊ currents that were inhibited by treatment with dibutryl-cAMP or forskolin. In contrast, PKA activators had no significant inhibitory effect on the activation of Cl ؊ current evoked by uncaging Ca 2؉ or by the photolytic release of higher concentrations of InsP 3 . Treatment with Rp-adenosine-3,5-cyclic monophoshorothioate, a selective inhibitor of PKA, or with Ht31, a peptide known to disrupt the targeting of PKA, largely abolished forskolin-induced inhibition of Ca 2؉ release. Further evidence for the targeting of PKA to the sites of Ca 2؉ mobilization was revealed using immunocytochemical methods demonstrating that the R II␤ subunit of PKA was localized to the apical regions of acinar cells and co-immunoprecipitated with the type III but not the type I or type II InsP 3 receptors. Finally, we demonstrate that the pattern of signaling evoked by acetylcholine can be converted to one that is more "CCK-like" by raising cAMP levels. Our data provide a simple mechanism by which distinct oscillatory Ca 2؉ patterns can be shaped.
Hormone-, neurotransmitter-, or growth factor-evoked increases in intracellular calcium exert control over a vast array of cellular functions, including secretion and gene expression (1)(2)(3). The regulatory control governing the fidelity and specificity of these processes can be encoded by the amplitude, frequency, or localization of cytosolic calcium signals (⌬[Ca 2ϩ ] c ) (4 -6). 1 It is well established that ⌬[Ca 2ϩ ] c in nonexcitable cells is generally induced by InsP 3 production, following activation of G q proteins coupled to phospholipase C ␤ (7,8). Many cell types express multiple distinct receptors coupled to this general transduction pathway. It is, however, not understood how the selective activation of distinct receptors that utilize the same general intracellular signaling system, such as the G␣ qcoupled formation of InsP 3 , can generate agonist-specific ⌬[Ca 2ϩ ] c (9 -12). Pancreatic acinar cells represent an ideal cell model for investigating this phenomenon because despite the activation of a common Ca 2ϩ release pathway by agonists, very different patterns of ⌬[Ca 2ϩ ] c emerge. Stimulation of cells with acetylcholine (ACh) or cholecystokinin (CCK) results in ⌬[Ca 2ϩ ] c that differ in spike frequency, level of base-line spiking, and local sites of initiation (11)(12)(13)(14).
In a previous study, we showed that the acinar cell type III InsP 3 R is rapidly and selectively phosphorylated by cAMP-dependent kinase A (PKA) following stimulation with physiological levels of CCK but not by ACh (15). This observation is consistent with data in the literature suggesting that CCK but not muscarinic stimulation results in coupling to G␣ s in addition to G␣ q (16). Thus, in this study we have investigated the possibility that selective phosphoregulation of the InsP 3 R by PKA contributes to the shaping of cytosolic Ca 2ϩ signals.

EXPERIMENTAL PROCEDURES
Isolation of Mouse Pancreatic Acinar Cells-Single acinar cells or small two-to three-cell clusters were prepared by a standard collagenase digestion of pancreata from wild-type C57BL/6 mice. Briefly, 25-g mice were sacrificed in accordance with National Institutes of Health policy and established protocol with the Division of Laboratory Animal Medicine, University of Rochester following CO 2 gas asphyxiation and then cervical dislocation. The pancreas was removed, and the capsule was injected with Dulbecco's modified Eagle's medium (Life Technologies, Inc.) containing 30 g/ml collagenase P (Sigma), 1000 units Purified collagenase (Worthington), and 1 mg/ml soybean trypsin inhibitor. Following 20 -30 min of digestion at 37°C, the pancreas was triturated through a 10-ml Falcon pipette tip, passed through a 100-m nylon mesh, and washed with Dulbecco's modified Eagle's medium containing 1% bovine serum albumen (Sigma), and isolated cells and small acini were collected by centrfugation at 100 ϫ g for 3 min. Cell were plated onto poly-L-lysine-coated glass coverslips and allowed to adhere for 5 min prior to application of recording solution perifusate.
Flash Photolysis and Digital Imaging-Photolytic release was performed using a pulsed Xenon arc lamp (T.I.L.L. Photonics) and fiber optic guide fed to a dual port epifluorescence condenser attached to a Nikon TE-200 fluorescence microscope. A high intensity 0.5-ms flash of UV light (360 Ϯ 7.5 nm) was reflected onto the plane of focus with a DM400 dichroic mirror and Nikon 40ϫ oil immersion objective, 1.3 NA. At the high intensity setting about 80 J were discharged. In other experiments, a low intensity (0.1 J) continuous strobe discharge was used to uncage threshold levels of InsP 3 . For simultaneous current recording and direct measurement of ⌬[Ca 2ϩ ] c , 75 M Oregon Green 488 Bapta-2 was added to the patch pipette solution. [Ca 2ϩ ] c imaging was performed using a monochrometer-based illumination system and high speed CCD camera (T.I.L.L. Photonics). Cells were illuminated at 488 Ϯ 15 nm, and fluorescence was collected through a 525 Ϯ 25-nm band pass filter (Chroma). Images were acquired at 79-ms intervals and displayed where F is the recorded fluorescence and F o was obtained from the mean of 15 sequential frames following equilibration with the patch dialysis solution and prior to stimulation. In some experiments, the acetoxymethylester of fura-2 was loaded into acinar cells and used to measure agonist-induced ⌬[Ca 2ϩ ] c . In this case, loaded cells were alternately excited at 340 and 380 Ϯ 15 nm, and emitted light was collected through a 525 Ϯ 25-nm bandpass filter. The ratio of emitted light resultant from excitation at these two wavelengths was utilized as a measure of ⌬[Ca 2ϩ ] c .
SDS-Polyacrylamide Gel Electrophoresis and Immunoblotting-Immunoprecipitations were performed as follows. Samples were sonicated in 0.5 ml of ice-cold lysis buffer (100 mM NaF, 50 mM Tris, 150 mM NaCl, 10 mM EDTA, 1 mM benzamidine, 1% Triton X-100, 1% 2-mercaptoethanol, 10 mg/ml leupeptin, and 10 mg/ml pepstatin, at pH 7.4) and then sonicated. After 30 min on ice, the samples were centrifuged for 30 min at 10,000 ϫ g. The supernatant was assayed for protein concentration. Samples of equal protein concentration were then incubated with antiserum overnight at 4°C. Immunoprecipitation of individual InsP 3 R subtypes was performed with excess antiserum (17). (No InsP 3 R was detectable in the lysate after addition of protein A.) Immobilized protein A beads (Pierce) were added to each sample. As a control, samples were processed with no cellular lysate or with no immunoprecipitating antiserum to ascertain specific bands on the immunoblot. After rotating the samples for 2 h, the samples were microcentrifuged, and the supernatant was discarded. The beads were washed four times with lysis buffer, suspended in SDS-polyacrylamide gel electrophoresis sample buffer, and boiled. The proteins were then separated by SDS-polyacrylamide gel electrophoresis and then transferred to nitrocellulose prior to immunoblotting (Schleicher & Schuell). Immunoreactivity was visualized using peroxidase-conjugated secondary antibodies followed by detection using the ECL detection system exposed on ECL Hyperfilm (Amersham Pharmacia Biotech).
Immunocytochemistry-Subcellular localization was performed using standard immunohistochemical methods in isolated acini fixed with 2% paraformaldehyde as described previously (18). A Noran/Oz laser scanning confocal microscope was used to visualize the distribution of immunofluorescence.

RESULTS AND DISCUSSION
Threshold level InsP 3 -evoked Ca 2ϩ Release Is Inhibited by PKA Activation-To date, no consensus exists in the literature as to the physiological consequence of phosphorylation of individual InsP 3 R types (19 -23). Thus, we investigated the PKAdependent regulation of Ca 2ϩ release in isolated mouse pancreatic acinar cells using the controlled release of photoactivatable (caged) InsP 3 and selective pharmacological activators and inhibitors of PKA function. Whole cell patch clamp methods were used to measure Ca 2ϩ -activated ionic current evoked by the photolysis of caged compounds. This current has been previously established to faithfully report ⌬[Ca 2ϩ ] c , especially below the apical plasma membrane (6,11,24,25).
As shown in Fig. 1 (A and C), high intensity UV light flash photolysis of a reproducible portion of 1-100 M NPE-caged InsP 3 , introduced via diffusional equilibration with the patch pipette solution, could repetitively evoke a Ca 2ϩ -activated current in a concentration-dependent manner. This indicated that prior exposure to Ca 2ϩ had little effect on amplitude or kinetics FIG. 1. PKA activation inhibits Ca 2؉ release. A, application of 60 M dbcAMP reversibly inhibited the Ca 2ϩ -activated ionic current evoked by low doses of InsP 3 in whole cell patch clamped mouse acinar cells. The traces represent recordings from single or two-or three-cell clusters that were voltage clamped at a holding potential (H p ) of Ϫ30 mV at 25°C. No significant changes in current were activated by flash photolysis when H p was 0 mV (the reversal potential of both Ca 2ϩactivated currents) or when caged compounds were omitted from the internal solution. B, the averaged data comparing the percentage of inhibition by dbcAMP of the ionic currents evoked by flash photolysis in cells loaded with different amounts of NPE-caged InsP 3 . The percentage of inhibition was determined by comparing peak current amplitude prior to and following dbcAMP treatment. The number of cells for each data set and statistical significance (p Յ 0.01) is indicated (asterisk). C, repetitive photolytic release of InsP 3 had no effect on peak amplitude or kinetics of current transients induced prior to and following treatment with dbcAMP.
of the whole cell current. A constant interval was maintained between flash discharge (3-5 min) to allow for equilibration of the cytosol with the cage-containing patch pipette. Application of the cell-permeable, hydrolysis-resistant cAMP analogue, dibutyryl cyclic adenosine monophosphate (dbcAMP) (60 M), prior to flash photolysis of caged InsP 3 , rapidly and reversibly reduced or abolished the amplitude of the Ca 2ϩ -activated current. Currents represent the maximum inhibition observed after a 3-5-min exposure to dbcAMP. Only cells that showed recovery to near control current levels within 15 min were included in the analysis. Surprisingly, it was only at the lowest concentrations of NPE-caged InsP 3 tested that this inhibitory effect was revealed as significant (Fig. 1B). Inhibition compared with control was 82.6 Ϯ 9.0% (1 M, n ϭ 7), 36.8 Ϯ 11.1% (3 M, n ϭ 9), 23.4 Ϯ 11.2% (10 M, n ϭ 7), 20.2 Ϯ 6.3% (30 M, n ϭ 8), and 5.8 Ϯ 2.9% (100 M, n ϭ 5). The percentage of inhibition was determined by measuring the peak current in response to a high energy flash discharge prior to and after 3-5-min exposure and following removal of drug. The observation that only the smallest, threshold responses to InsP 3 were inhibited by PKA activation is consistent with a previous study demonstrating that PKA had little effect on Ca 2ϩ release evoked by application of high doses of agonist (26 -28).
It is unlikely that this reduction was mediated by direct inhibitory phosphoregulation of the Ca 2ϩ -activated Cl Ϫ channel because, as shown in Fig. 2, no inhibition of peak current amplitude was observed over a range of currents (32-527 pA, 5 mM NP-EGTA; or 320 -686 pA, 10 mM NP-EGTA) activated by the release of caged Ca 2ϩ . Average peak currents prior to and following treatment with 100 M dbcAMP for 4 min were 217 Ϯ 51 versus 243 Ϯ 50 pA (n ϭ 11, 5 mM NP-EGTA) and 519 Ϯ 73 pA versus 571 Ϯ 84 pA; (n ϭ 5; 10 mM NP-EGTA). On average, flash photolysis of caged Ca 2ϩ (10 mM NP-EGTA) transiently raised [Ca 2ϩ ] c to about 7 M, as measured with the ratiometric, low affinity Ca 2ϩ indicator, benzothiazola coumarin. These lev-els of Ca 2ϩ are generally within a physiological range evoked apically by InsP 3 (1-10 M) (29) and at which the Cl Ϫ current is not likely to be saturated (29 -31). It is generally believed that, unlike CFTR or voltage-activated Cl Ϫ channels, cAMP does not directly modulate the Ca 2ϩ -activated Cl Ϫ channel (32)(33)(34). Furthermore, we observed that PKA activation was found to directly inhibit InsP 3 -evoked ⌬[Ca 2ϩ ] c in a permeablized mouse pancreatic acinar cell preparation, supporting our contention that the reduction in the Ca 2ϩ signal largely resulted from a decreased Ca 2ϩ release rather than an effect on the Ca 2ϩ -activated Cl Ϫ channel or on Ca 2ϩ clearance (data not shown).
A reduction of the current activated by low levels of stimulation indicated that a subset of InsP 3 R that exhibit the highest sensitivity for InsP 3 is preferentially modulated by dbcAMP. Because the pancreatic acinar cell expresses all isoforms of receptor together with heterotetrameric forms (35), it is difficult to categorically identify this high sensitivity receptor. The relative contribution of an InsP 3 R type to the initial signal is presumably determined by several factors including the relative abundance of the particular receptor, the affinity of the receptor for InsP 3 and susceptibility to phosphoregulation by PKA. Because the type II receptor is a poor substrate for PKA phosphorylation (23), it would seem unlikely that this receptor contributes markedly to threshold level Ca 2ϩ release and its modulation by PKA. In contrast, the type I receptor is a good substrate for phosphorylation by PKA and has been reported to have the highest affinity for InsP 3 but is expressed at very low levels in pancreatic acinar cells (17). The type III receptor has been shown to be an efficient substrate for PKA-mediated phosphorylation (15). In addition, the only quantitative study of receptor number in the exocrine pancreas has shown that the type III receptor is the most abundant receptor form (17) and, thus, by mass action would presumably contribute significantly to the initial Ca 2ϩ release. In support of this contention, the type III receptor has also been proposed to serve as an initial trigger for Ca 2ϩ release (36). Furthermore, the open probability of the receptor is modulated very steeply by near resting [Ca 2ϩ ] c (37). It should, however, be noted that the type III receptor has been reported, when studied in isolation, to have the lowest affinity for InsP 3 (23). These data together with the body of work in the literature indicate that the type III is a good candidate for the initial, PKA-modulated Ca 2ϩ signal. A contributing role for the type I receptor alone or in heterologous association with the type III receptor, however, cannot be excluded.
Because Ca 2ϩ spikes induced by low levels of ACh have been shown to initiate at specialized, InsP 3 R-rich sites (trigger zones) tightly associated with the luminal borders and are often confined to the luminal pole (18,29,38,39), we postulated that the effect of dbcAMP on Ca 2ϩ release would also be manifested in this region. To test this hypothesis, we sought to selectively activate Ca 2ϩ release at the acinar cell trigger zone. Thus, rather than using a high intensity flash discharge, we applied a continuous, low level photolytic strobe stimulus (indicated as I in the Fig. 3) to cells loaded with caged InsP 3 to achieve threshold activation levels of InsP 3 . In most cases it was necessary to first determine empirically the concentration of caged InsP 3 that was needed to evoke threshold activation. As shown in Fig. 3, fluorescence digital imaging methods confirmed that this stimulation paradigm preferentially induced Ca 2ϩ rises that were initiated and largely maintained in the luminal pole, characteristic of threshold Ca 2ϩ release. Simultaneous patch clamp measurements revealed the activation of irregular current spikes that mirrored the evoked apical ⌬[Ca 2ϩ ] c . As expected, a subsequent, high intensity UV flash discharge (indicate as II in the Fig. 3) evoked a global Ca 2ϩ rise that initiated in the luminal pole and activated a robust current.
To investigate the inhibitory effect of cAMP on the high sensitivity InsP 3 R subset, cAMP levels were increased by treatment with 10 M forskolin, and the Ca 2ϩ -activated current spike activity was evoked by 15-30 s of a continuous UV strobe. To accurately quantify the complex nature of the Ca 2ϩ release events evoked by continuous strobe prior to and following forskolin application for 3-5 min, the time integral of the current produced was determined. As shown in Fig. 4A, forskolin treatment reversibly inhibited Ca 2ϩ release at the trigger zone. Following wash-off of forskolin, current responses returned to control values within 5-20 min. Forskolin treatment induced nearly 83% inhibition of threshold Ca 2ϩ release (Fig. 4B). Current integrals were, on average, reduced to 16.4 Ϯ 9% of control values by forskolin treatment (p Յ 0.0003, n ϭ 10). Removal of forskolin restored the current integrals over 5-20 min to a level not significantly different from control values (150.7 Ϯ 23%). Next, we tested whether PKA activation was required for the inhibitory effect of cAMP. As shown in Fig. 4 (C and D), inclusion of a specific, competitive PKA inhibitor, Rp-adenosine-3Ј,5Ј-cyclic monophoshorothioate (Rp-cAMPS) (334 M), in the patch pipette solution largely abrogated the forskolin-induced inhibition of spike activity, and the current integrals following forskolin treatment were not significantly different from control values (107 Ϯ 7 and 124 Ϯ 27%, respectively; n ϭ 5). Taken together, these data strongly indicate that intracellular rises in cAMP negatively modulate InsP 3 induced Ca 2ϩ release via PKA-dependent phosphoregulation of a specialized subset of InsP 3 R localized at the acinar cell apical pole.
PKA Is Functionally Targeted to the Type III InsP 3 Receptor-Because forskolin treatment likely induces a spatially uniform rise in cAMP, we suggest that signaling specificity is achieved not in the production of cAMP but in the compartmentalization of its effector, PKA. This co-localization has been shown by others to lead to rapid and efficient phosphorylation of a specific substrate, even on a uniform background of cAMP levels (40 -47). Given the high degree of functional and structural organization exhibited by these polarized cells, we hypothesized that PKA might be targeted to the trigger zone to provide spatiotemporal control over the phosphoregulation of Ca 2ϩ release. PKA targeting is often achieved by the association of the regulatory R subunit dimer with an protein kinase A-anchoring protein bearing a specific subcellular localization signal (46,48). Immunohistochemical methods were performed to visualize the subcellular distribution of PKA in acinar cells using R II␣ and R II␤ subunit-specific monoclonal anitbodies. The R II␣ subunit exhibited diffuse cytosolic distribution (not shown). In contrast, the R II␤ subunit displayed a pattern of labeling that was confined to the apical and granular region of the cells (Fig. 5). This pattern of staining overlaps and is consistent with localization of PKA to the sites of InsP 3 -induced Ca 2ϩ release. It is within this region that the majority of InsP 3 R co-localize with a highly structured sublumenal actin web (18,39,49).
To determine whether R II subunits form molecular associations with type III InsP 3 R, we immunoprecipitated proteins from acinar cell extracts using a monoclonal antibody to the type III InsP 3 R. The enriched proteins were separated, and the Western blots were probed with actin, R II␣ , and R II␤ subunitspecific monoclonal antibodies. As shown in Fig. 6A, actin and both R II subunits immunoprecipitated with the type III InsP 3 R, suggesting that PKA is targeted to its potential kinase substrate, either directly or by a common tethering mechanism, possibly through actin or actin-binding proteins. To demonstrate specificity for the immunoprecipitation of PKA, ly-sates immunoprecipitated with type III InsP 3 R and containing R II␤ (Fig. 6B, lane 1) were probed with a polyclonal antiserum recognizing all heterotrimeric G protein ␤ subunits. Although G␤ subunits were identified in whole pancreatic acinar cell lysates (Fig. 6B, lane 4), no G␤ could be demonstrated to co-immunoprecipitate with type III InsP 3 R antiserum (Fig. 6B,  3). To investigate whether PKA was associated selectively with any particular InsP 3 R type, individual receptors were immunoprecipitated from pancreatic lysates using subtype-specific antiserum (17). Probing of separated proteins with the antiserum used to immunoprecipitate confirmed that individual InsP 3 R types had been immunoprecipitated (Fig.  6C, upper panel). In lane A, specific polyclonal antiserum raised against the C terminus of type I receptor were used (17,18). In lane B, a polyclonal antiserum versus the type II receptor (17,18), and in lane C, a monoclonal antibodies specific for type III receptor were utilized. Each antiserum was utilized in excess and is known to effectively and selectively immunoprecipitate its respective receptor (17,18). In the lower panel, the same immune complexes were probed with antiserum raised against PKA-R II␤ . Only samples immunoprecipitated with type III InsP 3 R contained PKA-R II␤ . These data strongly suggest that PKA is targeted specifically to type III receptors and, thus, is ideally placed to mediate rapid, specific phosphorylation of this receptor.
To test the functional significance of type III InsP 3 R-PKA co-localization, we disrupted PKA targeting in patch clamped acinar cells using a synthetic peptide containing the functional R II binding motif of the thyroid protein kinase A-anchoring protein, Ht31. The Ht31 peptide has been shown to be an efficient and specific inhibitor of R II protein kinase A-anchoring protein interactions (46,48). Fig. 7A shows that inclusion of 30 M Ht31 in the pipette solution could block the forskolin-induced inhibition of the apical Ca 2ϩ release. As shown in Fig.  7B, no statistically significant inhibition was observed following continuous treatment with forskolin for 5 or 10 min or following removal of forskolin, compared with control responses (76.8 Ϯ 15, 49.5 Ϯ 27, and 139 Ϯ 30% of control, respectively; n ϭ 9).
Functional Relevance of PKA Activation to the Shaping of Oscillatory Ca 2ϩ Signatures-Although our experiments provide evidence that the targeted phosphorylation of a subset of InsP 3 R exerts negative regulatory control over local Ca 2ϩ release events, we wondered whether this modulation could account for the distinctive patterns of Ca 2ϩ oscillations induced by CCK or carbachol (CCh). Application of physiological doses of ACh or CCh to pancreatic acinar cells generates sinusoidal [Ca 2ϩ ] c oscillations with frequencies of 4 -6/min that are superimposed on an elevated base line (11,12). In contrast, CCK application induces oscillations consisting of base-line spikes of much longer period (1-2/min) (12,50). To address this relation-ship, we tested whether PKA-dependent phosphorylation could convert CCh-induced oscillations into a pattern that resembled CCK-induced oscillations. Intact acinar cells were treated with low doses of either CCh or CCK to induce [Ca 2ϩ ] c oscillations and the ⌬[Ca 2ϩ ] c monitored with the Ca 2ϩ -sensitive dye, fura-2, prior to and following application of 60 M dbcAMP. As shown in the representative trace in Fig. 8A, introduction of dbcAMP reduced both spike amplitude and the plateau component (defined as the average intra-spike level) and slowed the frequency of [Ca 2ϩ ] c oscillations, suggesting that activation of PKA was sufficient to shift the CCh-induced pattern of oscillations to one that was more "CCK-like". As expected, dbcAMP Numerous factors have been proposed to provide a mechanism underlying agonist-specific [Ca 2ϩ ] c oscillations observed in pancreatic acinar cells (51,52). These include the utilization by CCK receptor signaling of alternative second messengers such as cyclic adenosine diphosphate ribose and nicotinic acid adenine dinucleotide phosphate. Although no data are available regarding the ability of agonists to stimulate levels of these putative messengers, significant modulation of Ca 2ϩ signaling has been reported (25,53). It is clear that if the levels of these molecules are modulated by agonist stimulation, it must be through G␣ q activation, because both CCK and muscarinic signaling are abolished by maneuvers antagonizing G␣ q function. (54,55). Recent reports have also suggested that pancreatic secretagogue receptors interact with differing RGS proteins to play a role in modulating signaling events (56,57). Modulating the kinetics of InsP 3 production at the level of the cell surface receptors would be expected to have significant effects on the generation of [Ca 2ϩ ] c signals. Although the data presented in this study do not preclude any of the aforementioned proposals, selective modulation of the initial ⌬[Ca 2ϩ ] c by PKA appears to contribute significantly to shaping the Ca 2ϩ signal.
We conclude that targeted, PKA-mediated phosphorylation specifically controls the functioning of a discrete subset of intracellular Ca 2ϩ release channels that act as the initial trigger for InsP 3 -induced Ca 2ϩ release. Our data suggest that the type III InsP 3 R, abundant in the apical trigger zone, fulfills this role. We contend that the negative allosteric modulation of the type III InsP 3 R contributes significantly to the characteristic Ca 2ϩ oscillations induced by CCK. Given that most cells have surface receptors that can couple to multiple G protein families (58,59), cross-talk between divergent signaling pathways to modulate Ca 2ϩ signaling at the level of the Ca 2ϩ release channel is likely to be a widespread and important mechanism by which information can be encoded for the selective activation of distinct physiological end points.