Ca2+-dependent Protein Kinase-A Modulation of the Plasma Membrane Ca2+-ATPase in Parotid Acinar Cells*

Cross-talk between cAMP and [Ca2+]i signaling pathways represents a general feature that defines the specificity of stimulus-response coupling in a variety of cell types including parotid acinar cells. We have reported recently that cAMP potentiates Ca2+ release from intracellular stores, primarily because of a protein kinase A-mediated phosphorylation of type II inositol 1,4,5-trisphosphate receptors (Bruce, J. I. E., Shuttleworth, T. J. S., Giovannucci, D. R., and Yule, D. I. (2002) J. Biol. Chem. 277, 1340–1348). The aim of the present study was to evaluate the functional and molecular mechanism whereby cAMP regulates Ca2+ clearance pathways in parotid acinar cells. Following an agonist-induced increase in [Ca2+]i the rate of Ca2+clearance, after the removal of the stimulus, was potentiated substantially (∼2-fold) by treatment with forskolin. This effect was prevented completely by inhibition of the plasma membrane Ca2+-ATPase (PMCA) with La3+. PMCA activity, when isolated pharmacologically, was also potentiated (∼2-fold) by forskolin. Ca2+ uptake into the endoplasmic reticulum of streptolysin-O-permeabilized cells by sarco/endoplasmic reticulum Ca2+-ATPase was largely unaffected by treatment with dibutyryl cAMP. Finally, in situ phosphorylation assays demonstrated that PMCA was phosphorylated by treatment with forskolin but only in the presence of carbamylcholine (carbachol). This effect of forskolin was Ca2+-dependent, and protein kinase C-independent, as potentiation of PMCA activity and phosphorylation of PMCA by forskolin also occurred when [Ca2+]i was elevated by the sarco/endoplasmic reticulum Ca2+-ATPase inhibitor cyclopiazonic acid and was attenuated by pre-incubation with the Ca2+ chelator, 1,2-bis(o-aminophenoxy) ethane-N,N,N′,N′-tetraacetic acid (BAPTA). The present study demonstrates that elevated cAMP enhances the rate of Ca2+ clearance because of a complex modulation of PMCA activity that involves a Ca2+-dependent step. Tight regulation of both Ca2+ release and Ca2+ efflux may represent a general feature of the mechanism whereby cAMP improves the fidelity and specificity of Ca2+ signaling.

Intracellular calcium is perhaps the most ubiquitous second messenger system in biology, yet the molecular mechanisms that confer specificity of Ca 2ϩ -dependent processes continue to intrigue investigators. Regulation of both the spatial and temporal properties of intracellular Ca 2ϩ signals is believed to underlie the specificity of stimulus-response coupling in a variety of cell types (1)(2)(3)(4). Recently an accumulation of evidence suggests that specific regulatory control over a variety of Ca 2ϩ signaling pathways can be achieved by the concomitant activation of additional signaling pathways, in particular those that elevate cyclic AMP (5)(6)(7)(8).
Parotid acinar cells represent an excellent model system not only for the study of Ca 2ϩ signaling in general but also to investigate cross-talk between cAMP and Ca 2ϩ signaling (2). This is because there is an abundance of evidence showing that both acetylcholine-evoked fluid secretion and exocytosis are potentiated markedly by cAMP-raising pathways (9 -13). In addition, we have demonstrated recently in parotid acinar cells that raising cAMP potentiates Ca 2ϩ release from intracellular stores, primarily because of a PKA 1 -mediated phosphorylation of type II inositol 1,4,5-trisphosphate receptors (5). This was hypothesized to be the major mechanism for the potentiation of fluid secretion by cAMP-raising agonists in the parotid.
In addition to potentiation of Ca 2ϩ release, it was also observed that the rate of Ca 2ϩ clearance upon removal of CCh was potentiated substantially in the presence of forskolin. This may represent an important additional mechanism by which cAMP tightly controls the spatial and temporal properties of Ca 2ϩ signaling in parotid acinar cells.
The aim of the present study was to systematically determine the molecular mechanisms responsible for the potentiation of Ca 2ϩ clearance by cAMP in mouse parotid acinar cells. The study revealed a novel, complex mechanism by which cAMP can modulate Ca 2ϩ signaling by potentiating PMCA activity in a Ca 2ϩ -dependent manner. Because the PMCA has been suggested to respond to dynamic fluctuations in cytosolic [Ca 2ϩ ] due to its CaM binding properties (14,15), further regulation by cAMP may be important for the fine tuning of Ca 2ϩ signaling. This regulatory control likely contributes to the general mechanism by which cAMP-elevating agonists shape Ca 2ϩ signaling, a process that may have relevance to the regulation of a wide array of specific functions in various cell types.

EXPERIMENTAL PROCEDURES
Isolation of Single Parotid Acinar Cells-Clusters of parotid acinar cells were isolated by collagenase digestion as described previously (5). Following isolation, cells were re-suspended in a HEPES-buffered physiological saline solution containing the following (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 . Aliquots of cell suspensions were loaded with 2 M fura-2/AM for 30 min at room temperature after which they were re-suspended in HEPES-buffered physiological saline solution and kept at 4°C until ready for use.
Digital Imaging of Intracellular Ca 2ϩ ([Ca 2ϩ ] i )-During experiments, cells were allowed to adhere to a glass coverslip, which formed the base of a gravity-fed perfusion chamber. Cells were perfused continually with HEPES-buffered physiological saline solution, and automatic valves were used for switching solutions. [Ca 2ϩ ] imaging experiments were performed using an inverted epifluorescence Nikon microscope with a ϫ40 oil immersion objective lens (numerical aperture, 1.3). Most experiments utilized an imaging system consisting of a DG-4 illumination system (Sutter), a 12-bit progressive interline charged coupled device camera (Sensicam), and Axon Imaging Workbench acquisition software as described previously in more detail (5). This imaging system was used for most standard experiments that required a relatively slow image acquisition rate of 0.1 to 1 Hz and 100to 300-ms exposure times. For the rapid atropine-evoked Ca 2ϩ clearance experiments faster acquisition rates of 10 to 20 Hz and 10-to 20-ms exposure times were required; therefore, an alternative imaging system was used. This contained essentially the same optics and camera but used a TILL polychrome IV monochromator illumination system and TILL VisION acquisition and analysis software (see Ref. 6 for detailed description). All experiments were performed at room temperature.
Ca 2ϩ Uptake into the ER (SERCA Activity)-Measurement of Ca 2ϩ uptake into the ER was isolated by imaging streptolysin (SL-O)-permeabilized parotid acinar cells loaded with the low affinity Ca 2ϩ -sensing fluorescent dye, fura-2FF/AM (10 M) for 60 min at 37°C as described previously (5). Briefly, cells were perfused continually with a Chelex-100 "scrubbed" cytosol-like medium containing the following (in mM): 135 KCl, 1.2 KH 2 PO 4 , 0.5 EGTA, 0.5 N-hydroxyethylethylenediaminetriacetic acid (HEDTA), 0.5 nitriloacetic acid, and 20 HEPES/KOH (pH 7.1). MgCl 2 , CaCl 2 , and MgATP were added accordingly (as calculated by MaxChelator program) to give constant free Mg 2ϩ , Ca 2ϩ , and ATP concentrations of 0.9 mM, 0.06 -1.0 M, and 1 mM, respectively. Permeabilization was achieved by perfusion with an ATP-containing (but Ca 2ϩ -free) cytosol-like medium containing 0.4 international unit of SL-O. Because fura-2FF accumulates in virtually every compartment of the cell, permeabilization could be verified by monitoring the decline in the 360-nm excitation signal (isosbestic point for fura-2FF) as the cytosolic dye leaked out of the cell leaving dye trapped in intracellular organelles (Ͻ20% of pre-permeabilized cells). Image acquisition rate was 0.1 Hz (100-ms exposure). Measurement of organelle [Ca 2ϩ ] was achieved similarly to fura-2 (340-and 380-nm excitation and 510-nm emission) with an image acquisition rate of 0.1 Hz and 300-ms exposure, and fura-2FF 340/380 ratio images were calculated online. Following perfusion of cells with cytosol-like medium devoid of SL-O, Ca 2ϩ , or ATP for 5-10 min, rapid Ca 2ϩ uptake was achieved upon addition of 1 mM Mg-ATP and [Ca 2ϩ ] between 0.06 and 1.0 M. The fura-2FF 340/380 ratio, representing organelle [Ca 2ϩ ], reached a maximum within 3 min presumably because of an equilibrium being established between opposing Ca 2ϩ fluxes across the organelle membrane (steady state). Subsequent addition of 3 M inositol 1,4,5-trisphosphate, 30 M cyclopiazonic acid (CPA), and 1 M FCCP evoked a rapid decrease, slow decrease, or no change in fura-2FF ratio, respectively, confirming that Ca 2ϩ was taken up into the ER. The rate of Ca 2ϩ uptake into the ER for each cell at each ambient [Ca 2ϩ ] was fit to a single exponential decay to yield the time constant (). The mean time constant was compared in the absence and presence of 100 M dibutyryl cyclic AMP (Bt 2 cAMP).
Phosphorylation of PMCA-Parotid acinar cells were isolated from 4 -6 mice as described above, aliquotted appropriately, and treated for 10 min with or without the following test reagents: 1 M CCh, 10 M forskolin, 150 nM phorbol 12-myristate 13-acetate (PMA) or 30 M CPA. In some experiments, cells were pre-incubated with 20 M BAPTA/AM for 30 min at room temperature to buffer any change in [Ca 2ϩ ] i . Cells were then pelleted rapidly by centrifugation and re-suspended in 400 l of ice-cold lysis buffer containing the following (in mM): 50 Tris-HCl (pH 7.4), 250 NaCl, 5 EDTA, 100 NaF, 0.1% Triton X-100, and EDTA-free complete protease inhibitor mixture tablets (Roche Molecular Biochemicals). Cell lysates were then sonicated, left on ice for 30 min, and vortexed every 5 min. Each lysate was incubated with the monoclonal anti-PMCA antibody (ϳ1 g/mg protein; clone 5F10, Affinity Bioreagents) for 1 h, followed by 80 l of protein A-agarose beads (Pierce) for another hour at 4°C, to immunoprecipitate PMCA protein. As a secondary control (blank) an aliquot of cell lysates from untreated cells was incubated with beads without any antibodies. As described previously (5) the beads-protein complex was washed five times in lysis buffer and denatured by boiling in SDS sample buffer (Laemmli) for 5 min. Immunoprecipitated proteins were separated by SDS-polyacrylamide gel electrophoresis (7.5%) and Western blotted using the phospho-(Ser/Thr) PKA substrate antibody (Cell Signaling Technology) to detect phosphorylated PMCA protein. To confirm that approximately equal amounts of protein were loaded into each lane of the gel or whether the above treatments altered PMCA levels, nitrocellulose membranes were incubated in stripping solution (62.5 Tris-HCl (pH 6.7), 2% SDS, and 100 M ␤-mercaptoethanol) for 30 min at 50°C to dissociate any bound antibodies. The membrane was then re-probed subsequently by Western blotting with the anti-PMCA antibody. In some experiments quantification of phosphorylation was achieved by densitometric analysis of visible bands detected by the phospho-(Ser/Thr) PKA substrate antibody. This was performed by imaging nitrocellulose membranes exposed to chemiluminescence reagents (Westpico/Westfemto Super Signal; Pierce) using a 12-bit charged coupled device camera (Sensicam) and LabWorks imaging and analysis software (UVP Bioimaging Systems). This measures the total pixel intensity above background of equal sized areas of interest for each visible band. Duplicates on each membrane were averaged, and to account for variability in band intensities between gels comparisons were made between treated and control conditions using a paired Student's t test.
Data Analysis and Experimental Design-Because of the nature of most experiments an unpaired experimental design was applied (unless otherwise stated in the text), whereby statistical significance was determined between groups of experiments (control and treated) using an unpaired t test or Mann Whitney test. Occasionally, statistical significance was determined, where appropriate, using a paired t test, Wilcoxan test for pairs, or one sample t test. For Ca 2ϩ clearance data the rate of Ca 2ϩ clearance was fit to a single exponential decay using Microcal Origin 5.0 software and quantified by comparing time constants (). 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.

RESULTS
Forskolin Potentiates Ca 2ϩ Clearance-We have shown previously that elevated cAMP potentiates CCh-evoked Ca 2ϩ release (5). In the present study, under similar conditions, the rate of Ca 2ϩ clearance was determined following removal of CCh by fitting the falling phase of the change in fura-2 340/380 ratio to a single exponential decay (Fig. 1B). Time constants () were compared quantitatively in the absence and presence of 10 M forskolin using a paired t test (Fig. 1C). These initial experiments revealed that forskolin caused a 5.3 Ϯ 1.2-fold increase in Ca 2ϩ clearance by reducing from 12.6 Ϯ 2.0 to 2.9 Ϯ 0.5 s ( Fig. 1C; n ϭ 7 experiments, 34 cells). In some cells the rate of decrease in fura-2 340/380 ratio following removal of CCh was initially slow, suggesting that the rate of solution exchange was contributing to the variability. Therefore to increase the accuracy and precision of the data, experiments were repeated by initiating clearance of Ca 2ϩ from the cytosol by adding a 10-fold higher concentration of the muscarinic receptor antagonist, atropine (10 M), in the continued presence of 1 M CCh (Fig. 2). This reduces any error associated with solution exchange, because as soon as just 10% of the perfusate exchanges, the atropine will have essentially displaced all the CCh from the receptors. Because under these conditions atropine will remain bound for a long time, the effect of forskolin was compared with unpaired control experiments. Under these conditions forskolin increased Ca 2ϩ clearance ϳ2.5-fold by reducing from 8.27 Ϯ 1.2 to 3.43 Ϯ 0.52 s (n ϭ 14 experiments, 57 cells; see Fig. 2). Collectively, these experiments demonstrate that elevating intracellular cAMP levels by activation of adenylyl cyclase, with forskolin, potentiates Ca 2ϩ clearance pathways.
The major routes for Ca 2ϩ clearance in non-excitable cells, such as parotid acinar cells, are believed to be Ca 2ϩ uptake into the ER by the SERCA, Ca 2ϩ efflux across the plasma membrane by the PMCA, and Ca 2ϩ uptake into mitochondria (1,3,4). Therefore, to determine the specific molecular loci for the effects of elevated cAMP, on Ca 2ϩ clearance, the following experiments were designed to physically or pharmacologically isolate these different Ca 2ϩ clearance pathways. The rate of Ca 2ϩ clearance during each experimental paradigm varied markedly and thus may not accurately represent the actual rate of each isolated Ca 2ϩ clearance pathway in vivo or under physiological Ca 2ϩ signaling conditions. Nevertheless, the relative rate of Ca 2ϩ clearance following a particular treatment under each condition was always consistent, thereby verifying each experimental protocol.
Effect of Bt 2 -cAMP on Ca 2ϩ Uptake into the ER (SERCA Activity)-One of the major routes for Ca 2ϩ clearance following stimulation is believed to be the re-uptake of Ca 2ϩ into the ER by SERCA (1,3,4). In addition, SERCA activity is potentiated by elevated cAMP in cardiac cells because of PKA-mediated phosphorylation of the inhibitory accessory protein, phospholamban (16 -19). Therefore, a likely locus for the effects of elevated cAMP on Ca 2ϩ clearance in parotid acinar cells is Ca 2ϩ uptake into the ER. To isolate Ca 2ϩ uptake into the ER, parotid acinar cells were permeabilized with SL-O, and Ca 2ϩ uptake was initiated by perfusion with a cytosol-like solution (see "Experimental Procedures") containing 1 mM ATP and varying concentrations of ambient Ca 2ϩ (0.06 -1.0 M). Ca 2ϩ uptake was indicated by an increase in the fura-2FF 340/380 ratio, which reached a steady state between 0.5 and 3.5 min depending on ambient [Ca 2ϩ ] (Fig. 3A). Once Ca 2ϩ uptake reached a steady state, addition of more Ca 2ϩ failed to increase the fura-2FF 340/380 ratio further (data not shown).
Application of inositol 1,4,5-trisphosphate evoked a rapid Ca 2ϩ release confirming that Ca 2ϩ had been taken up into the ER by SERCA Ca 2ϩ pumps (5). Moreover, the SERCA inhibitor CPA and removal of ATP evoked a slow Ca 2ϩ leak, whereas the mitochondrial uncoupler FCCP failed to have any effect (data not shown). The rate of Ca 2ϩ uptake was fit to a single exponential decay, and the mean time constant () was determined for each ambient (loading) [Ca 2ϩ ] in the absence or presence of Bt 2 -cAMP and compared using an unpaired t test (Fig. 3B).
The rate of Ca 2ϩ uptake into the ER increased by ϳ4-fold as the ambient [Ca 2ϩ ] increased from 60 to 200 nM (mean decreased from 28 (20). This suggests that SERCA is much more active and sensitive to small elevations of [Ca 2ϩ ] i above resting levels in parotid than in pancreatic acinar cells. This is consistent with our previous data, which shows that the deactivation of Ca 2ϩ -dependent Cl Ϫ currents was greater in parotid compared with pancreatic acinar cells and that this was due, in part, to a greater SERCA activity (21). Whatever the specific mechanisms that control the Ca 2ϩ -dependent Ca 2ϩ uptake into the ER, the most important observation from the present study was that at all the ambient [Ca 2ϩ ] tested (100, 200, and 600 nM) the rate of Ca 2ϩ uptake into the ER was not affected significantly by treatment with 100 M Bt 2 -cAMP (Fig. 3B).
To further assess the effect of Bt 2 -cAMP on Ca 2ϩ uptake into the ER of parotid acinar cells, a paired experimental design was utilized in the following way. Ca 2ϩ uptake was initiated in the absence of Bt 2 -cAMP by addition of ATP and 0.2 M Ca 2ϩ . Upon reaching steady state, 100 M Bt 2 -cAMP was added, and removal of Ca 2ϩ and ATP caused a slow decline in fura-2FF 340/380 ratio, presumably because of depletion of the ER Ca 2ϩ store (Fig. 3C). Upon reaching a new steady state (10 to 15 min) a second Ca 2ϩ uptake was initiated in the continued presence of Bt 2 -cAMP (Fig. 3C). Using this paired experimental design the second Ca 2ϩ uptake (mean ϭ 9.4 Ϯ 0.5 s) in the absence of Bt 2 -cAMP was ϳ30% slower than the first Ca 2ϩ uptake (mean ϭ 7.2 Ϯ 0.4 s) during time matched control experiments. Nevertheless, under the same conditions the change in rate of Ca 2ϩ uptake following treatment with Bt 2 -cAMP (1.27 Ϯ 0.03-fold increase in ) was not significantly different from time-matched control experiments (1.32 Ϯ 0.04-fold increase in ; see Fig. 3D). These data therefore reinforce the initial observation that elevation of cAMP fails to enhance Ca 2ϩ uptake into the ER of permeabilized parotid acinar cells.
Forskolin Potentiates Ca 2ϩ Clearance When SERCA Is Inhibited-Despite rigorous attempts to isolate Ca 2ϩ uptake into the ER of SL-O-permeabilized parotid acinar cells it remains unlikely that the effects of elevated cAMP on Ca 2ϩ clearance are because of enhanced SERCA activity. However, a simple explanation for the lack of effect of Bt 2 -cAMP could be the permeabilization process itself, which may wash away critical cytosolic factors, such as calmodulin or phospholamban, important for SERCA pump activity (16 -19). To address this further, the effects of elevated cAMP on Ca 2ϩ clearance in intact cells was determined following inhibition of SERCA activity by CPA.
Intact parotid acinar cells were stimulated with 30 M CPA, which slowly raised [Ca 2ϩ ] i primarily because of leak from the ER and activation of capacitative Ca 2ϩ entry (22)(23)(24)(25). The increase in [Ca 2ϩ ] i reached a maximum and then declined slowly to a new steady state, representing a balance of Ca 2ϩ efflux and Ca 2ϩ influx. Removal of external Ca 2ϩ , by chelation with 1 mM EGTA, evoked an immediate clearance of Ca 2ϩ that was primarily because of the PMCA. Similar to previous experiments, the rate of Ca 2ϩ clearance was fit to a single exponential decay (Fig. 4A), and the time constants () were compared between control and forskolin-treated cells using an unpaired experimental design (Fig. 4B). Under these conditions forskolin caused an ϳ2-fold increase in the rate of Ca 2ϩ clearance, as indicated by a decrease in the mean from 28.2 Ϯ 2.1 in control cells (n ϭ 5 experiments; 27 cells) to 16.4 Ϯ 1.9 (n ϭ 5 experiments; 22 cells) in forskolin-treated cells (Fig. 4B). Although the major mechanism for Ca 2ϩ clearance during dynamic Ca 2ϩ signaling is believed to be because of SERCA, these data suggest that potentiation of Ca 2ϩ clearance by elevated cAMP may be because of a pathway other than SERCA, most likely the PMCA.
Effect of Forskolin on PMCA Activity-Because the potentiation of Ca 2ϩ clearance by elevated cAMP appears independent of SERCA activity, another Ca 2ϩ clearance pathway such as Ca 2ϩ efflux across the plasma membrane (PMCA activity) may be the target for the effects of cAMP. To test this hypothesis, experiments similar to those described in Fig. 2 were carried out in the continued presence of 1 mM La 3ϩ , a known inhibitor of PMCA (26 -28). Intact parotid acinar cells were stimulated with 1 M CCh, in the continued presence of La 3ϩ , and Ca 2ϩ clearance was then initiated by the addition of 10 M atropine (Fig. 5A). The effect of forskolin on the rate of clearance under these conditions was compared with unpaired control experiments. In the continued presence of La 3ϩ , forskolin failed to have any significant effect (unpaired t test) on the rate of Ca 2ϩ clearance; mean control ϭ 4.2 Ϯ 0.4 (n ϭ 46 cells, 11 exper- iments) compared with mean forskolin ϭ 5.2 Ϯ 0.9 (n ϭ 34 cells, 11 experiments). Closer examination of the data from these experiments revealed that there was a large inherent variability in the rate of Ca 2ϩ clearance from one cell preparation to another, which could account for the lack of any statistically significant effect of forskolin. However, clearance rate was relatively consistent between experiments from the same cell preparation. Therefore, an average was determined for all experiments from control cells from each cell preparation. The corresponding average was determined from forskolin-treated cells of the same cell preparation and expressed as -fold change of control. Statistical significance was then determined using a one sample t test, which showed that forskolin caused a 2.06 Ϯ 0.31-fold increase in when La 3ϩ was absent (n ϭ 4 for paired analysis, 14 separate experiments; see Fig. 5B) but was ineffective when La 3ϩ was present; 1.11 Ϯ 0.2-fold increase in (n ϭ 4 for paired analysis, 11 separate experiments; see Fig.  5B). Therefore, under conditions where the PMCA pump is inhibited, forskolin failed to have any significant effect on the rate of Ca 2ϩ clearance. This suggests that the potentiation of Ca 2ϩ clearance by elevated cAMP is because of an effect on PMCA activity.
To further address the effect of cAMP on PMCA we attempted to isolate the PMCA activity pharmacologically. Intact parotid acinar cells were stimulated with 100 M CCh to maximally release Ca 2ϩ from the ER, 30 M CPA to prevent reuptake into the ER, 1 M FCCP to prevent mitochondrial uptake, 1 M oligomycin to inhibit the mitochondrial ATP synthase and prevent ATP consumption, and 1 mM La 3ϩ to inhibit Ca 2ϩ efflux (Fig. 6A). This caused a large increase in [Ca 2ϩ ] i and effectively trapped Ca 2ϩ within the cytosol. Subsequent removal of La 3ϩ (to remove the inhibition of PMCA) and chelation of external Ca 2ϩ with 10 mM EGTA (to prevent Ca 2ϩ influx) caused a slow decline in [Ca 2ϩ ] i possibly due, in part, to the rate at which La 3ϩ washes out but mainly to Ca 2ϩ clearance by PMCA. Because this experimental protocol produced a dramatic increase in [Ca 2ϩ ] i that may be detrimental to the cell, the subsequent behavior of the PMCA pump may not reflect its activity under physiological conditions accurately. Therefore, cells were pre-incubated with 2 M BAPTA-AM for 30 min at room temperature in an attempt to attenuate the large increase in [Ca 2ϩ ] i (Fig. 6A). Similar to the above experiments, the resulting rate of Ca 2ϩ clearance was fit to a single exponential decay, and the mean determined as a measure of PMCA activity and compared in the absence and presence of forskolin. It was noted that the apparent rate of Ca 2ϩ clearance under these conditions was significantly slower compared with other experiments. This could be due, in part, to the rate at which La 3ϩ washes out, ATP depletion or local pH changes due to mitochondrial inhibition, or impaired PMCA regulation by other factors such as Ca 2ϩ or PKC. Therefore the rate of Ca 2ϩ clearance under these conditions clearly may not reflect PMCA activity per se under in vivo or more physiological conditions. However, the important observations from these experiments is that under these identical conditions forskolin caused an ϳ2-fold increase in the apparent rate of PMCA activity; mean decreased from 201 Ϯ 35 s (BAPTA control cells, n ϭ 53 cells, 9 experiments) to 110 Ϯ 12 s (forskolin-treated cells, n ϭ 63 cells, 11 experiments). This therefore reinforces the previous data and demonstrates clearly that elevated cAMP enhances Ca 2ϩ clearance during CCh-evoked [Ca 2ϩ ] i signaling most likely because of an ϳ2-fold increase in PMCA activity.
Phosphorylation of PMCA-To determine whether the enhanced Ca 2ϩ clearance correlates with a cAMP-dependent phosphorylation of PMCA, phosphorylation assays were carried out using the phospho-(Ser/Thr) kinase substrate antibody (5,7). Parotid acinar cells were incubated with a variety of reagents, including CCh to increase [Ca 2ϩ ] i , forskolin to elevate cAMP, 150 nM PMA to activate PKC, and 30 M CPA to slowly elevate [Ca 2ϩ ] i independent of PKC activation (see Fig. 7 and Fig. 8). Following incubation, PMCA protein was immunoprecipitated from cell lysates using a PMCA-specific antibody (Affinity Bioreagents), and phosphorylated PMCA protein was detected by Western blotting with phospho-(Ser/Thr) kinase substrate antibody (Fig. 7). This antibody does not distinguish between phosphorylated substrates of PKA and other kinases, such as PKC, protein kinase G, and Ca 2ϩ /calmodulin-dependent kinase (CaMK) (29,30). Therefore, Fig. 7A shows a typical Western blot experiment that demonstrates the relative sensitivity of the phospho-(Ser/Thr) kinase substrate antibody in detecting phosphorylated proteins (indicated by the intensity of visible bands) by incubating cells with agents that activate PKA (forskolin) and other kinases. A variety of proteins of different molecular weight were detected following all treatments but were particularly evident by treatment with forskolin (compare lane 1, untreated cells with lane 4, forskolintreated cells; see Fig. 7A). Nevertheless, some visible bands were detected in untreated cells (Fig. 7A, lane 1), suggesting some proteins may be phosphorylated under basal conditions. Treatment with 150 nM PMA, which activates PKC maximally  1 mM EGTA). B, rate of Ca 2ϩ clearance was fit to a single exponential decay, and the time constants () were compared between control and forskolin-treated cells using an unpaired t test (*, p Ͻ 0.05).

FIG. 5. Forskolin fails to affect the clearance of [Ca 2؉ ] i when PMCA is inhibited.
A, experiments were similar to Fig. 2, except PMCA was inhibited by the presence of 1 mM La 3ϩ throughout. Cells were stimulated with 1 M CCh, and Ca 2ϩ clearance was initiated by addition of 10 M atropine. B, an average was determined for all experiments from control cells and forskolin-treated cells of the same cell preparation and expressed as -fold change of control. Statistical significance was then determined using a one sample t test (*, p Ͻ 0.05). Using unpaired analysis forskolin failed to affect Ca 2ϩ clearance in the presence of La 3ϩ . tion is that the phospho-(Ser/Thr) kinase substrate antibody is more sensitive at detecting PKA-phosphorylated proteins. Interestingly, treatment with 1 M CCh also enhanced the detec-tion of visible bands (Fig. 7A, lane 2 compared with lane 1), whereas treatment with 30 M CPA did not (Fig. 7A, lane 5 compared with lane 1). The effect of CCh could be because of PKC-mediated phosphorylation, because muscarinic receptors are coupled to the activation of PKC, via diacylglycerol. This is supported by the lack of effect of CPA, which increases [Ca 2ϩ ] i but presumably does not activate PKC to the same extent as CCh. An alternative explanation is that the effects of CCh are because of CaMK-mediated phosphorylation. This is because CCh evokes a rapid spike-like increase in [Ca 2ϩ ] i that is more effective in activating CaMK than the more slower increase in [Ca 2ϩ ] i evoked by CPA (32).
To determine whether PMCA is phosphorylated specifically by various treatments, PMCA immunoprecipitates were separated by SDS-PAGE and Western blotted with the phospho-(Ser/Thr) kinase substrate antibody as shown in Fig. 7B and Fig. 8, A and B. Treatment with CCh (lane 2) increased the detection of a visible band (compared with lane 1, untreated) that corresponds to the molecular mass of PMCA (ϳ140 kDa), suggesting that PMCA is phosphorylated by treatment with CCh (Fig. 7B). The combined treatment with CCh and forskolin enhanced this phosphorylation significantly (Fig. 7B, lanes 3 and 4, and quantified in Fig. 8), whereas treatment with forskolin alone (Fig. 7B, lane 5) failed to increase phosphorylation above basal levels. By stripping and re-probing the same nitrocellulose membrane with anti-PMCA, it was confirmed that approximately equal amounts of PMCA protein were loaded into each lane and/or that the above treatments did not affect PMCA protein levels (see bottom panel of Fig. 7B for representative example). The increased band intensity for all phosphorylation experiments (see Fig. 8 and Fig. 9) was therefore because of increased phosphorylation of PMCA. This implies that cAMP either potentiates the CCh-evoked phosphorylation of PMCA or that cAMP-evoked phosphorylation of PMCA can only occur in the presence of CCh. This is consistent with the functional data mentioned previously; forskolin potentiated Ca 2ϩ clearance when [Ca 2ϩ ] i was elevated by CCh but failed to reduce resting [Ca 2ϩ ] i when added to cells alone (5). CChevoked phosphorylation of PMCA could be mediated by PKC, and the effect of forskolin could therefore be because of a potentiation of PKC-mediated phosphorylation. To test this, cells were treated with PMA to directly activate PKC and PMCA protein immunoprecipitated similar to the above experiments. PMA caused a measurable phosphorylation of PMCA Cell lysates were prepared as described under "Experimental Procedures," and protein samples were run on a 7.5% SDS-polyacrylamide gel. The chemiluminescence signal was exposed to film for the visualization of minor phosphorylated proteins. B, parotid acinar cells were incubated with (ϩ) or without (Ϫ) 1 M CCh or 10 M forskolin for 10 min, and samples were prepared as described under "Experimental Procedures." PMCA protein was immunoprecipitated (IP) with anti-PMCA antibody and separated by SDS-PAGE, and phosphorylated PMCA protein was detected by Western blotting (WB) with the phospho-(Ser/Thr) kinase substrate antibody (top panel). The same nitrocellulose membrane was stripped and re-probed with anti-PMCA (␣-PMCA) which demonstrates that approximately equal amounts of protein were loaded into each lane (lower panel). * in B represents a secondary control in which samples were treated identically, but anti-PMCA antibody was omitted. The observed reaction therefore represents nonspecific binding to protein-A beads. but was not affected significantly by co-treatment with forskolin (Fig. 8C). Alternatively, CCh-evoked phosphorylation of PMCA could be mediated by a Ca 2ϩ -dependent step. To examine this cells were treated with CPA rather than CCh, which evokes a slow increase in [Ca 2ϩ ] i because of inhibition of SERCA (see Fig. 4) and presumably does not activate PKC to the same extent as CCh. CPA failed to significantly phosphorylate PMCA above resting levels (Fig. 8C), although in combination with forskolin it evoked a significant phosphorylation of PMCA above resting levels as quantified by densitometric analysis (Fig. 8C). In addition, pre-incubation of cells with 20 M BAPTA-AM for 30 min at room temperature, which completely abolished the CCh-evoked [Ca 2ϩ ] i increase (Fig. 9A), resulted in a significant attenuation of PMCA phosphorylation by forskolin in the presence of CCh (Fig. 9B). It was noted that the CCh-evoked phosphorylation was also attenuated suggesting that this may also have some Ca 2ϩ dependence. Nevertheless, the most important observation from these experiments is that, in support of Fig. 8, the phosphorylation of PMCA by elevated cAMP is essentially Ca 2ϩ -dependent. DISCUSSION The present study demonstrates that activation of signaling pathways that elevate cAMP potentiates Ca 2ϩ clearance from the cytosol of parotid acinar cells. The observation that forskolin failed to affect resting [Ca 2ϩ ] i (see Fig. 1A and Ref. 5) suggests that this potentiation by cAMP is dependent on elevated [Ca 2ϩ ] i. One of the major mechanisms for removing Ca 2ϩ from the cytosol following stimulation in non-excitable cells is the re-uptake of Ca 2ϩ into the ER by SERCA (1, 3, 4). In cardiac cells activation of ␤-adrenergic receptors leading to elevation of cAMP potentiates SERCA activity by PKA-mediated phosphorylation of the inhibitory accessory protein, phospholamban (16 -19). However, the data presented here showed that Bt 2 cAMP failed to affect SERCA activity significantly at a range of ambient [Ca 2ϩ ] in SL-O-permeabilized parotid acinar cells (see Fig. 3B). This could be a consequence of the permeabilization process itself, which may cause the loss of key cytosolic factors (up to 200 kDa) that have important regulatory control over SERCA activity, such as phospholamban or calmodulin. Nevertheless, under conditions where SERCA was inhibited in intact parotid acinar cells, potentiation of Ca 2ϩ clearance by forskolin remained suggesting that the effects of cAMP were on some other Ca 2ϩ clearance pathway. Although FIG. 8. Phosphorylation of PMCA is Ca 2؉ -dependent and PKC-independent. A, parotid acinar cells were treated with (ϩ) or without (Ϫ) 1 M CCh, 10 M forskolin (Forsk), and/or 150 nM PMA for 10 min at room temperature, and protein samples were prepared as described under "Experimental Procedures." As in Fig. 7B, PMCA was immunoprecipitated, and phosphorylated PMCA protein was detected by Western blotting with the phospho-(Ser/Thr) kinase substrate antibody. B, similar experiment to A, except cells were treated with (ϩ) or without (Ϫ) 30 M CPA. C, quantification by densitometric analysis of experiments in A and B from at least three to six separate identical experiments. Quantification was achieved using LabWorks Imaging and Analysis software (UVP Bioimaging Systems) by measuring the total pixel intensity above background of equal sized areas of interest for each visible band. For duplicate bands from each experiment an average value was determined. These values were, in turn, averaged over three to six separate experiments to give the values in the histogram Ϯ S.E. For comparisons between control (Con) and treated conditions statistical significance was determined using a paired Student's t test (*, p Ͻ 0.05). To determine the effect of forskolin on other treatments (CCh, CPA, and PMA) an unpaired Student's t test was used ( ‡, p Ͻ 0.05). this does not rule out an effect of cAMP on SERCA in intact cells completely, it would appear unlikely that elevations in cAMP have any significant direct or indirect effects on SERCA in parotid acinar cells.
Functional isolation of Ca 2ϩ efflux across the plasma membrane by pharmacological inhibition of other Ca 2ϩ transport pathways demonstrates clearly that the elevation of cAMP potentiated Ca 2ϩ efflux markedly. Because the Na ϩ /Ca 2ϩ -exchanger has not been reported to be expressed in parotid acinar cells, the major route for Ca 2ϩ efflux, and therefore the effects of cAMP, appear to be on the PMCA. Another potentially important Ca 2ϩ clearance pathway that could be affected by cAMP in parotid acinar cells is mitochondrial Ca 2ϩ uptake. This is a relatively slow, low affinity, high capacity Ca 2ϩ transport pathway that senses high localized [Ca 2ϩ ] (2,(33)(34)(35)(36)(37)(38). Strategic localization of mitochondria close to Ca 2ϩ release sites has been suggested to be important in "shaping" the spatial, as well as the temporal, properties of [Ca 2ϩ ] i signaling in a variety of cells (33)(34)(35)(36). However, under the conditions of our experiments it remains unlikely that mitochondrial Ca 2ϩ uptake is affected by cAMP. In particular, forskolin failed to affect Ca 2ϩ clearance under the conditions where PMCA was blocked by La 3ϩ , reinforcing the notion that potentiation of Ca 2ϩ clearance is due largely, if not entirely, to an effect of cAMP on PMCA in parotid acinar cells.
The PMCA is a ubiquitously expressed P-type ATPase with a high affinity for Ca 2ϩ and has been suggested to be the major mechanism for the maintenance of resting [Ca 2ϩ ] i (39,40). To date four genes encoding PMCA (PMCA 1-4) have been identified and cloned, with a variety of splice variants that have a specific tissue distribution (15,39). Although the specific PMCA isoform expressed in parotid is unknown, the most likely candidates are PMCA 1 and 4, because these have been shown to be expressed ubiquitously in all tissues, whereas PMCA 2 and 3 are expressed predominantly in excitable cells (41,42). In addition, it has been suggested that only the PMCA 1 isoform contains a PKA phosphorylation site (43). Thus this isoform may be a good candidate to be expressed in parotid acinar cells. The activity of PMCA is affected profoundly by CaM binding and is therefore exquisitely sensitive to dynamic changes in [Ca 2ϩ ] i (15,44,45). In addition, PMCA has also been shown to be regulated by PKC, as well as by acidic phospholipids, oligomerization, and proteolytic cleavage (45). Of particular interest to the present study, PMCA activity has been shown to be increased by PKC-and/or PKA-mediated phosphorylation (26, 31, 46 -52). Moreover, PKC-and/or PKAmediated phosphorylation affects the binding of CaM to PMCA (49 -52). Together, these data indicate a large scope and complexity for the potential regulation of PMCA by signaling pathways that elevate [Ca 2ϩ ] i , in combination with those that activate PKC and/or PKA. PMCA therefore likely represents a potentially important molecular locus for cross-talk between these different signaling pathways.
In the present study phosphorylation assays demonstrated that treatment of cells with CCh caused phosphorylation of PMCA above basal levels. This is consistent with studies in cultured aortic endothelial cells that show that the PMCA is phosphorylated by treatment with a variety of different agonists that are coupled to several signaling pathways (31). The most important observation from the phosphorylation assays is that elevation of cAMP with forskolin evoked a considerable phosphorylation of PMCA in the presence of CCh but failed to have any effect on its own. Physiologically one might expect PMCA activity to be potentiated only when [Ca 2ϩ ] i is elevated above resting levels, and functional data shows that cAMP failed to reduce resting [Ca 2ϩ ] i but potentiated Ca 2ϩ clearance when [Ca 2ϩ ] i was elevated by CCh (5). However, the lack of any measurable PMCA phosphorylation by forskolin alone is at variance with previous studies (31,48,(52)(53)(54). Such differences could be explained partly by the in vitro phosphorylation assays utilized by some of the above studies in which purified PMCA was used. As such this may not reflect the more physiological conditions of the current study accurately. Nevertheless, phosphorylation data from the current study suggests a complex regulation of the PMCA by cAMP/PKA that is dependent on Ca 2ϩ .
It should be noted that the phospho-(Ser/Thr)-PKA substrate antibody can recognize substrates for other kinases including PKC, protein kinase G, and CaMK. Moreover, muscarinic receptors are known to couple to activation of PKC via diacylglycerol, and there is a precedence in the literature for the regulation of the PMCA by activation by PKC (15, 31, 39, 44, 45, 47, 49 -51, 55). Therefore, a reasonable explanation of these data is that CCh-evoked phosphorylation of PMCA is mediated by PKC. The effect of forskolin could then be due either to a potentiation of the PKC-mediated phosphorylation or to an additional independent PKA-mediated phosphorylation. However, our data using PMA show that forskolin failed to increase the direct PKC-mediated phosphorylation of PMCA significantly, arguing against any forskolin-mediated potentiation of a PKC-mediated phosphorylation. Furthermore, our functional data revealed that Ca 2ϩ clearance was potentiated by forskolin when [Ca 2ϩ ] i was elevated initially by CPA (i.e. in the absence of CCh; see Fig. 4). Addition of CPA would presumably not activate PKC or at least not to the same extent as CCh, suggesting that the potentiation of Ca 2ϩ clearance by forskolin is a strictly Ca 2ϩ -dependent phenomenon and is independent of PKC activation. An alternative explanation is that the CChevoked phosphorylation of PMCA is mediated by a Ca 2ϩ -dependent step (as indicated by the reduced phosphorylation of PMCA in the presence of BAPTA; see Fig. 9), for example via CaMK, and that the observed effect of forskolin reflects a potentiation of this CaMK-mediated phosphorylation. However, there are no reports of CaMK-mediated regulation of PMCA, and no CaMK consensus sequences have been identified in the PMCA molecule. Moreover, cAMP has generally been shown to inhibit rather than potentiate the CaMK signaling cascade in other cell types (56,57). Nevertheless, elevation of [Ca 2ϩ ] i with CPA rather than CCh failed to phosphorylate PMCA above resting levels, yet in combination with forskolin, this treatment evoked a significantly greater phosphorylation of PMCA above resting levels. Furthermore, pre-incubation of cells with BAPTA-AM to blunt the CCh-evoked increase in [Ca 2ϩ ] i abolished the phosphorylation of PMCA by forskolin in the presence of CCh. Collectively these data suggest that the phosphorylation of PMCA by elevated cAMP in the presence of CCh is predominantly Ca 2ϩ -dependent, rather than PKC-dependent. One way in which such an independent PKA-mediated phosphorylation could occur only in the presence of an elevation of Ca 2ϩ is via a PKA phosphorylation site whose exposure is [Ca 2ϩ ] i -dependent. Such a hypothesis would be consistent with ideas proposed by Carafoli and colleagues (44, 45, 58 -60) to explain how CaM (and PKC) regulate the PMCA. It has been demonstrated that the PMCA contains an autoinhibitory domain, which also contains a CaM binding region and a PKC phosphorylation site (49), that interacts with the catalytic domain of the pump (59,60). Binding of CaM (or PKCmediated phosphorylation) reduces the binding of the autoinhibitory domain and ultimately increases substrate access and thus the Ca 2ϩ transporting activity of the pump (45). In addition, a PKA phosphorylation site on PMCA has also been identified to be close but distal to the CaM binding domain (53). Therefore, allosteric regulation of PMCA by Ca 2ϩ (61,62) or CaM (and/or PKC-mediated phosphorylation) not only removes the autoinhibition but may also conceivably expose a PKA phosphorylation site, thereby allowing further activation of the pump.
We have demonstrated previously that elevation of cAMP in parotid acinar cells leads to potentiation of Ca 2ϩ release by PKA-mediated phosphorylation of type II inositol 1,4,5trisphosphate receptors (5). The present study now also demonstrates clearly that elevation of cAMP potentiates Ca 2ϩ clearance by PMCA. Initially it may seem counterintuitive to potentiate both Ca 2ϩ release and Ca 2ϩ clearance by a similar mechanism in the same cell. However, this may represent an important general mechanism by which increases in cAMP may "shape" Ca 2ϩ signaling and improve the fidelity of Ca 2ϩdependent processes. This is because the PMCA has been suggested to respond to dynamic fluctuations in cytosolic [Ca 2ϩ ] because of its CaM binding properties (14,15). The binding of CaM to the PMCA can affect its Ca 2ϩ transporting activity profoundly (51,63). The high affinity and slow rates of CaM binding to and from the PMCA would predict that the PMCA responds to dynamic [Ca 2ϩ ] in an integrative manner similar to that described for CaMKII (32). In fact recent mathematical modeling and direct experimental evidence demonstrates that the PMCA exhibits "memory" (64) in that a second exposure to high [Ca 2ϩ ] can differentially affect PMCA activity in a manner that was dependent on both the duration of the first exposure to high [Ca 2ϩ ] and the time between the first and second exposures (64). This effect was demonstrated to be dependent on the affinity and rate of CaM binding to the PMCA. One can predict that such an effect may have very important consequences on the modulation of the frequency and duration of [Ca 2ϩ ] i oscillations. In addition, PKA-mediated phosphorylation of PMCA has been shown previously to increase the PMCA activity by increasing the affinity for CaM binding (52). In the current study, we have demonstrated that cAMP potentiates PMCA activity in a Ca 2ϩ -dependent manner via a PKA-mediated phosphorylation that is also Ca 2ϩ -dependent. This suggests that PKA modulates the Ca 2ϩ -dependent activity of the PMCA in a positive feedback manner, thereby tuning its activity so that it becomes exquisitely sensitive to [Ca 2ϩ ]. Together, these data indicate that PKA-mediated phosphorylation of the PMCA may be an extremely important mechanism by which cAMP shapes [Ca 2ϩ ] i signaling by controlling the Ca 2ϩ spike duration, inter-spike interval, and therefore the frequency of [Ca 2ϩ ] i oscillations.
In addition to shaping the temporal properties of [Ca 2ϩ ] i signals, PKA-mediated regulation of the PMCA in parotid acinar cells may also contribute to the spatial shaping of the [Ca 2ϩ ] i signals. Although the spatial distribution of PMCA in parotid acinar cells is unknown, the PMCA has been shown to be apically located in the structurally and functionally related pancreatic and submandibular acinar cells (65). In addition, functional Ca 2ϩ efflux has also been shown to correlate to cytosolic [Ca 2ϩ ] i spikes (66,67) and is more pronounced across the apical membrane of pancreatic acinar cells (68). If such a mechanism was extrapolated to parotid acinar cells, then PKAmediated regulation of the PMCA may be important for the effective activation of Ca 2ϩ -dependent processes at or close to the apical membrane, such as exocytosis or activation of Ca 2ϩdependent Cl Ϫ channels and therefore fluid secretion (21,69). This may be of particular relevance in the highly invaginated microvillar structure of the apical membrane, whereby the PMCA may be the only means for clearance of local concentrations of Ca 2ϩ because of the absence of ER (70). In conclusion, tight regulation of both Ca 2ϩ release and Ca 2ϩ efflux by cAMP may represent a general feature in defining specificity of stimulus-response coupling in a variety of cell types.