Protein Kinase A Increases Type-2 Inositol 1,4,5-Trisphosphate Receptor Activity by Phosphorylation of Serine 937*

Protein kinase A (PKA) phosphorylation of inositol 1,4,5-trisphosphate receptors (InsP3Rs) represents a mechanism for shaping intracellular Ca2+ signals following a concomitant elevation in cAMP. Activation of PKA results in enhanced Ca2+ release in cells that express predominantly InsP3R2. PKA is known to phosphorylate InsP3R2, but the molecular determinants of this effect are not known. We have expressed mouse InsP3R2 in DT40-3KO cells that are devoid of endogenous InsP3R and examined the effects of PKA phosphorylation on this isoform in unambiguous isolation. Activation of PKA increased Ca2+ signals and augmented the single channel open probability of InsP3R2. A PKA phosphorylation site unique to the InsP3R2 was identified at Ser937. The enhancing effects of PKA activation on this isoform required the phosphorylation of Ser937, since replacing this residue with alanine eliminated the positive effects of PKA activation. These results provide a mechanism responsible for the enhanced Ca2+ signaling following PKA activation in cells that express predominantly InsP3R2.

Protein kinase A (PKA) phosphorylation of inositol 1,4,5trisphosphate receptors (InsP 3 Rs) represents a mechanism for shaping intracellular Ca 2؉ signals following a concomitant elevation in cAMP. Activation of PKA results in enhanced Ca 2؉ release in cells that express predominantly InsP 3 R2. PKA is known to phosphorylate InsP 3 R2, but the molecular determinants of this effect are not known. We have expressed mouse InsP 3 R2 in DT40-3KO cells that are devoid of endogenous InsP 3 R and examined the effects of PKA phosphorylation on this isoform in unambiguous isolation. Activation of PKA increased Ca 2؉ signals and augmented the single channel open probability of InsP 3 R2. A PKA phosphorylation site unique to the InsP 3 R2 was identified at Ser 937 . The enhancing effects of PKA activation on this isoform required the phosphorylation of Ser 937 , since replacing this residue with alanine eliminated the positive effects of PKA activation. These results provide a mechanism responsible for the enhanced Ca 2؉ signaling following PKA activation in cells that express predominantly InsP 3 R2.
Hormones, neurotransmitters, and growth factors stimulate the production of InsP 3 3 and Ca 2ϩ signals in virtually all cell types (1). The ubiquitous nature of this mode of signaling dictates that this pathway does not exist in isolation; indeed, a multitude of additional signaling pathways can be activated simultaneously. A prime example of this type of "cross-talk" between independently activated signaling systems results from the parallel activation of cAMP and Ca 2ϩ signaling pathways (2,3). Interactions between these two systems occur in numerous distinct cell types with various physiological consequences (3)(4)(5)(6). Given the central role of InsP 3 R in Ca 2ϩ signaling, a major route of modulating the spatial and temporal features of Ca 2ϩ signals following cAMP production is potentially through PKA phosphorylation of the InsP 3 R isoform(s) expressed in a particular cell type.
There are three InsP 3 R isoforms (InsP 3 R1, InsP 3 R2, and InsP 3 R3) expressed to varying degrees in mammalian cells (7,8). InsP 3 R1 is the major isoform expressed in the nervous system, but it is less abundant compared with other subtypes in non-neuronal tissues (8). Ca 2ϩ release via InsP 3 R2 and InsP 3 R3 predominate in these tissues. InsP 3 R2 is the major InsP 3 R isoform in many cell types, including hepatocytes (7,8), astrocytes (9,10), cardiac myocytes (11), and exocrine acinar cells (8,12). Activation of PKA has been demonstrated to enhance InsP 3induced Ca 2ϩ signaling in hepatocytes (13) and parotid acinar cells (4,14). Although PKA phosphorylation of InsP 3 R2 is a likely causal mechanism underlying these effects, the functional effects of phosphorylation have not been determined in cells unambiguously expressing InsP 3 R2 in isolation. Furthermore, the molecular determinants of PKA phosphorylation of this isoform are not known.
PKA-mediated phosphorylation is an efficient means of transiently and reversibly regulating the activity of the InsP 3 R. InsP 3 R1 was identified as a major substrate of PKA in the brain prior to its identification as the InsP 3 R (15,16). However, until recently, the functional consequences of phosphorylation were unresolved. Initial conflicting results were reported indicating that phosphoregulation of InsP 3 R1 could result in either inhibition or stimulation of receptor activity (16,17). Mutagenic strategies were employed by our laboratory to clarify this discrepancy. These studies unequivocally assigned phosphorylation-dependent enhanced Ca 2ϩ release and InsP 3 R1 activity at the single channel level, through phosphorylation at canonical PKA consensus motifs at Ser 1589 and Ser 1755 . The sites responsible were also shown to be specific to the particular InsP 3 R1 splice variant (18). These data were also corroborated by replacing the relevant serines with glutamates in a strategy designed to construct "phosphomimetic" InsP 3 R1 by mimicking the negative charge added by phosphorylation (19,20). Of particular note, however, although all three isoforms are substrates for PKA, neither of the sites phosphorylated by PKA in InsP 3 R1 are conserved in the other two isoforms (21). Recently, three distinct PKA phosphorylation sites were identified in InsP 3 R3 that were in different regions of the protein when compared with InsP 3 R1 (22). To date, no PKA phosphorylation sites have been identified in InsP 3 R2.
Interactions between Ca 2ϩ and cAMP signaling pathways are evident in exocrine acinar cells of the parotid salivary gland. In these cells, both signals are important mediators of fluid and protein secretion (23). Multiple components of the [Ca 2ϩ ] i signaling pathway in these cells are potential substrates for modulation by PKA. Previous work from this laboratory established that activation of PKA potentiates muscarinic acetylcholine receptor-induced [Ca 2ϩ ] i signaling in mouse and human parotid acinar cells (4,24,25). A likely mechanism to explain this effect is that PKA phosphorylation increases the activity of InsP 3 R expressed in these cells. Consistent with this idea, activation of PKA enhanced InsP 3 -induced Ca 2ϩ release in permeabilized mouse parotid acinar cells and also resulted in the phosphorylation of InsP 3 R2 (4).
Invariably, prior work examining the functional effects of PKA phosphorylation on InsP 3 R2 has been performed using cell types expressing multiple InsP 3 R isoforms. For example, AR4-2J cells are the preferred cell type for examining InsP 3 R2 in relative isolation, because this isoform constitutes more than 85% of the total InsP 3 R population (8). InsP 3 R1, however, contributes up to ϳ12% of the total InsP 3 R in AR4-2J cells. An initial report using InsP 3 -mediated 45 Ca 2ϩ flux suggested that PKA activation increased InsP 3 R activity in AR4-2J cells (21). A similar conclusion was made in a later study, which documented the effects of PKA activation on agonist stimulated Ca 2ϩ signals in AR4-2J cells (26). Any effects of phosphorylation observed in these experiments could plausibly have resulted from phosphorylation of the residual InsP 3 R1.
Although PKA enhances InsP 3 -induced calcium release in cells expressing predominantly InsP 3 R2, including hepatocytes, parotid acinar cells, and AR4-2J cells (4,13,21,26,27), InsP 3 R2 is not phosphorylated at stoichiometric levels by PKA (21). This observation has called into question the physiological significance of PKA phosphorylation of InsP 3 R2 (28). The apparent low levels of InsP 3 R2 phosphorylation are clearly at odds with the augmented Ca 2ϩ release observed in cells expressing predominantly this isoform. The equivocal nature of these findings probably stems from the fact that, to date, all of the studies demonstrating positive effects of PKA activation on Ca 2ϩ release were conducted in cells that also express InsP 3 R1. The purpose of the current experiments was to analyze the functional effects of phosphorylation on InsP 3 R2 expressed in isolation on a null background. We report that InsP 3 R2 activity is increased by PKA phosphorylation under these conditions, and furthermore, we have identified a unique phosphorylation site in InsP 3 R2 at Ser 937 . In total, these results provide a direct mechanism for the cAMP-induced activation of InsP 3 R2 via PKA phosphorylation of InsP 3 R2.

EXPERIMENTAL PROCEDURES
cDNA Expression Constructs-Mouse InsP 3 R2 cDNA, a kind gift of Dr. Katsuhiko Mikoshiba (Riken, Japan), was used as the template for creation of EGFP-tagged subclones (29). All fragments were amplified by PCR with MluI and NotI restriction sites incorporated into the N termini and C termini, respectively. A BssHII site was incorporated into the N terminus of fragment 3 because of an internal MluI site in this fragment. PCR products were ligated with MluI-and NotI-digested pCI-Neo-EGFP (a kind gift from Dr. Sundeep Malik, University of Rochester) to create the mammalian expression vectors. The full-length rat InsP 3 R2 in the expression vector pCMV5 was a kind gift of Dr. Suresh Joseph (Thomas Jefferson University). A Kozak initiation motif was engineered into the receptor DNA using PCR. The modified receptor DNA was then cloned into the pEF6/V5-His Topo TA expression vector (Invitrogen).
Mutagenesis-All point mutations in the InsP 3 R2 fragments were created using QuikChange XL or QuikChange multisite mutagenesis (Stratagene, La Jolla, CA). Point mutations in the full-length InsP 3 R2 expression constructs to allow S937A and S2633A amino acid substitutions were constructed using a twostep QuikChange mutagenesis strategy (30).
DT40 Cell Lines-The cDNA construct for 3ϫ hemagglutinin-tagged human type-3 muscarinic receptor (M3R) and wild type and mutated InsP 3 R2 constructs were linearized with MfeI. Linearized constructs were introduced into DT40-3KO cells, which are devoid of InsP 3 R, by nucleofection using an Amaxa nucleofector as described previously (31)(32)(33). After nucleofection, the cells were incubated in growth medium for 24 h prior to dilution in selection medium containing 2 mg/ml Geneticin. Cells were then seeded into 96-well tissue culture plates at ϳ1000 cells/well and incubated in selection medium for at least 7 days. Wells exhibiting growth after the selection period were picked for expansion.
Phosphorylation of InsP 3 R2 in Intact Cells-COS-7 cells were transfected with InsP 3 R2 expression constructs 36 -40 h prior to labeling with 32 PO 4 Ϫ for 2 h in phosphate-free Dulbecco's modified Eagle's medium. ϳ150 Ci were added to 1 ml of phosphate-free media in each well of a 6-well culture dish. After labeling, cells were washed three times in Tris-buffered saline, treated with forskolin for 15 min, and lysed in lysis buffer (10 mM Tris, 150 mM NaCl, 100 mM NaF, 1 EDTA, 1% Nonidet P-40, pH 7.4) supplemented with protease inhibitor tablets (Roche Applied Science). Lysates were cleared with pansorbin prior to incubation with immunoprecipitation antibody for 2 h. Protein A/G beads were then added for 1 h. The beads were washed three times in lysis buffer and resuspended in Laemmli sample buffer. Samples were separated by PAGE on 5% polyacrylamide, and the gels were dried for phosphorimaging using a phosphor storage screen (Amersham Biosciences) and Amersham Biosciences PhosphorImager.
PKA Phosphorylation of InsP 3 R in Vitro-COS-7 cells were transfected with wild type or mutated InsP 3 R2 expression constructs 36 -40 h before harvest. Cells were washed twice in phosphate-buffered saline prior to suspension in lysis buffer supplemented with protease inhibitor mixture tablets (Roche Applied Science). Cells were dispersed using a cell scraper (Corning Glass) and allowed to incubate on ice for 20 min. Lysates were cleared with pansorbin prior to incubation with immunoprecipitation antibody for 2 h. Protein A/G beads were then added for 1 h. The beads were washed three times in lysis buffer and three times in PKA phosphorylation buffer (120 mM KCl, 50 mM Tris, 0.1% Triton X-100, 0.3 mM MgCl 2 , pH 7.2) and resuspended in phosphorylation buffer. 20 units of purified recombinant PKA (Promega) were added to the samples along with [␥-32 P]ATP (ϳ5 Ci/reaction) and 0.5 M unlabeled ATP. Kinase reactions were incubated for 0 -15 min, and the beads were washed six times in lysis buffer prior to resuspension and SDS-PAGE. Gels were stained with BioSafe Coomassie (Bio-Rad) and subsequently dried down for phosphorimaging. 32 P signals were detected by phosphorimaging as described above.
Phosphorylation of InsP 3 R2 Serine 937 in COS-7 Cells and Mouse Parotid Acinar Cells-Mouse parotid acinar cells were isolated by collagenase digestion, as described previously (4). COS-7 cells or mouse parotid acinar cells were treated with 10 M forskolin and 100 M IBMX for 15 min. Cell lysates were harvested in lysis buffer supplemented with 100 nM okadaic acid. InsP 3 R2 was immunoprecipitated from cell lysates using ␣-InsP 3 R2-CT. Where noted, some immunoprecipitated samples were treated with 10 units of calf intestinal alkaline phosphatase (New England Biolabs, Ipswich, MA) at 37°C for 1 h. Samples were separated on SDS-PAGE, and phosphorylated serine 937 was detected with ␣-Ser(P) 937 by Western blotting. Blots were stripped and reprobed with ␣-InsP 3 R2-CT to verify equal amounts of total immunoprecipitated InsP 3 R2.
Digital Imaging of Intracellular Ca 2ϩ in DT40 Cells-DT40 cells were loaded with 2 M of the Ca 2ϩ -sensitive dye Fura-2 AM at room temperature for 15-30 min. Fura-2-loaded cells were allowed to adhere to a glass coverslip at the bottom of a perfusion chamber. Cells were perfused in HEPES-buffered physiological saline containing 137 mM NaCl, 0.56 mM MgCl 2 , 4.7 mM KCl, 1 mM Na 2 HPO 4 , 10 mM HEPES, 5.5 mM glucose, and 1.26 mM CaCl 2 , pH 7.4. Imaging was performed using an inverted Nikon microscope through a ϫ40 oil immersion objective lens (numerical aperture, 1.3). Fura-2-loaded cells were excited alternately with light at 340 and 380 nm by using a monochrometer-based illumination system (TILL Photonics), and the emission at 510 nm was captured by using a digital frame transfer CCD camera. In experiments where InsP 3 R or M3R were transiently expressed, cDNA encoding HcRed was included to indicate transfected cells. HcRed fluorescence was detected by excitation at 560 nm and observing the emission at Ͼ600 nm.
Single Channel Recordings-Whole cell patch clamp recordings of single InsP 3 R2 channel activity present in the plasma membrane (34,35) were made from DT40-3KO cells stably expressing mouse InsP 3 R2. K ϩ was utilized as the charge carrier in all experiments, and free Ca 2ϩ was clamped at 200 nM to favor activation of InsP 3 R (bath: 140 mM KCl, 10 mM HEPES, 500 M BAPTA, free Ca 2ϩ 250 nM (pH 7.1); pipette: 140 mM KCl, 10 mM HEPES, 100 M BAPTA, 200 nM free Ca 2ϩ , 5 mM Na 2 -ATP unless otherwise noted (pH 7.1)). Borosilicate glass pipettes were pulled and fire-polished to resistances of about 20 megaohms. Following establishment of stable high resistance seals, the membrane patches were ruptured to form the whole cell configuration with resistances Ͼ5 gigaohms and capacitances of Ͼ8 picofarads. Currents were recorded under voltage clamp conditions at Ϫ100 mV using an Axopatch 200B ampli-fier and pClamp 9. Channel recordings were digitized at 20 kHz and filtered at 5 kHz with a Ϫ3 dB, 4-pole Bessel filter. Activity was typically evident essentially immediately following breakthrough with InsP 3 in the pipette. Analyses were performed using the event detection protocol in Clampfit 9. Channel openings were detected by half-threshold crossing criteria. We assumed that the number of channels in any particular cell is represented by the maximum number of discrete stacked events observed during the experiment. The single channel open probability (P o ) was calculated using the multimodal distribution for the open and closed current levels.

RESULTS
Activation of PKA Enhances Ca 2ϩ Signaling in DT40-M3 Cells Expressing Mouse InsP 3 R2-Analyzing subtype-specific regulation of individual InsP 3 R isoforms in native tissue is hampered by the fact that most mammalian cell types express multiple isoforms and that the functional receptors can form heterotetrameric channels (36 -38). Given these limitations, the functional properties of specific homotetrameric receptors must be determined in a defined system. Kurosaki and colleagues (40) developed such a system based on the DT40 chicken B-cell precursor line by creating a null background (DT40-3KO cells) following elimination of all three InsP 3 R isoforms through homologous recombination (39,40). In order to examine the effects of PKA phosphorylation on InsP3R2, we transiently expressed this isoform in a DT40-3KO cell line (DT40-M3) stably expressing the human M3R (33).
DT40-M3 cells transfected with mouse InsP 3 R2 cDNA, identified by expression of HcRed, responded to the muscarinic agonist carbachol (CCh) stimulation in a concentrationdependent manner (Fig. 1A). Treatment with 1 M CCh produced Ca 2ϩ signals with a range of amplitudes, presumably as a consequence of a range of InsP 3 R2 expression levels following transient transfection (Fig. 1B). Cells that responded to 1 M CCh with amplitudes no greater than 0.2 340 nm/380 nm ratio units were used for the purposes of analyzing the effects of raising cAMP. An example of a DT40-M3 cell expressing mouse InsP 3 R2 and treated three times with 1 M CCh is shown in Fig. 1C. Treatment of cells with 20 M forskolin after the first CCh treatment resulted in Ca 2ϩ transients with Ͼ5-fold larger amplitudes as shown in Fig. 1C, indicating that activation of PKA enhances Ca 2ϩ release from InsP 3 R2. This effect was only evident in cells that responded submaximally to 1 M CCh. As previously observed, forskolin induced a rise in [Ca 2ϩ ] i in some cells (27). This effect was evident in ϳ50% of cells from six separate experiments, and these cells were excluded from analysis. The mechanism underlying this effect is presently unknown; however, it is independent of InsP 3 R as it occurs in DT40-3KO cells at approximately the same frequency. The activation of PKA by this treatment had clear enhancing effects on the Ca 2ϩ signal (Fig. 1D).
We also obtained additional evidence that PKA-induced phosphorylation results in increased InsP 3 R2 activity by examining the effects of activating PKA on the single channel activity of InsP 3 R2. Single InsP 3 R2 measurements were conducted using whole cell recordings of plasma membrane-resident InsP 3 R2 in DT40 cells stably expressing InsP 3 R2 (20, 31, 34, 35,41). This configuration allows the monitoring of single InsP 3 R channels during the activation of endogenous PKA following exposure to forskolin (20). Fig. 2A shows an example of channel activity when low concentrations of InsP 3 (100 nM) were included in the recording pipette. No channel activity is observed in this preparation in the absence of InsP 3 or in DT40-3KO cells devoid of InsP 3 R (31). The open probability of the channel was markedly enhanced following exposure to forskolin (Fig. 2, B (diary plot for representative cell) and C (pooled data)). InsP 3 R2 channel activity returned to pre-PKA activation levels following washout of forskolin. Enhanced channel activity was readily evident at threshold [InsP 3 ] but not observed at higher levels of InsP 3 (1 M; Fig. 2C). In total, these data provide clear evidence that activation of PKA results in enhanced Ca 2ϩ release through increased activity of InsP 3 R2.
PKA Phosphorylates Mouse InsP 3 R2 at Serine 937-A likely mechanism for the positive effects of increasing cAMP on the Ca 2ϩ signal is through direct phosphorylation of InsP 3 R2 by PKA. Because this signaling system is a rich source of potential PKA substrates, we cannot, however, discount other effects of PKA on the M3R-induced Ca 2ϩ signals. Potential loci might include effects on InsP 3 levels, Ca 2ϩ clearance, or Ca 2ϩ entry. In addition, cAMP at millimolar concentrations was recently reported to enhance Ca 2ϩ release from InsP 3 R2 by a mechanism that did not require PKA activity (42). Establishing the PKA phosphorylation site(s) and subsequently performing mutagenesis of the putative sites in InsP 3 R2 are required to definitively rule out these and other possible mechanisms for the increased Ca 2ϩ signal. Although PKA has been shown to phosphorylate InsP 3 R2 in a number of studies by independent groups (21,26,28), the site(s) of phosphorylation are not known. Experiments were performed following transient overexpression of InsP 3 R2 in COS-7 cells. This expression system was chosen because of low endogenous levels of InsP 3 R and the high transfection efficiency such that expressed receptor can be readily distinguished from endogenous protein (see supplemental Fig. S1). In addition, expressed receptors rarely form heterotetramers with endogenous InsP 3 R (43). Fig. 3 and the supplemental material confirm that InsP 3 R2 is a substrate for PKA. Both mouse and rat InsP 3 R2 were phosphorylated in intact cells (Fig. 3A and supplemental Fig. S1) and in vitro (Fig.  3B). Fig. 3C shows alignments of mammalian InsP 3 R protein sequences around the known phosphorylation sites in InsP 3 R1 and InsP 3 R3. None of these sites are conserved in InsP 3 R2 sequences, indicating that the phosphorylation evident in Fig.  3, A and B, occurs at a novel PKA phosphorylation site(s).
Optimal PKA phosphorylation sites are serines or threonines preceded by basic residues at the Ϫ2and Ϫ3-positions (44). The primary amino acid sequences of individual mammalian InsP 3 R2 proteins combined harbor greater than 300 serine and threonine residues. Of these potential candidates, ϳ30 have basic residues (arginine or lysine) upstream of the putative target serine or threonine. The single canonical PKA consensus sequence located in InsP 3 R2 (RRPS 2508 ) is probably not accessible to the kinase, because it is located on the luminal side of the putative permeability (P) loop. Given the large number of potential putative PKA phosphorylation sites present in the primary sequence of InsP 3 R2, we developed a subcloning approach to identify novel PKA phosphorylation sites.
Limited trypsin digestion of InsP 3 R1 and InsP 3 R3 has established that InsP 3 R can be divided into five (in the case of InsP 3 R1) or four (in the case of InsP 3 R3) globular domains with intervening solvent-exposed trypsin digestion sites (22,45,46). In order to maintain domain structure as much as possible, InsP 3 R2 subclones were designed to correspond with predicted limited trypsin digestion products. The InsP 3 R2 sequence was therefore divided into five smaller fragment constructs with EGFP as an epitope tag on the N terminus of each fragment. As depicted in Fig. 4A, fragment 1 corresponded with residues 1-343, fragment 2 with residues 344 -919, fragment 3 with residues 920 -1583, fragment 4 with residues 1584 -1883, and fragment 5 with residues 1884 -2701.
Constructs coding for the five fragments were transfected into COS-7 cells and immunoprecipitated with an antibody directed against EGFP. A separate sample was transfected with a construct coding for enhanced yellow fluorescent proteintagged InsP 3 R1 and served as a positive control. Immunoprecipitated proteins were subjected to in vitro PKA kinase assays, as described under "Experimental Procedures." Samples were then resolved by SDS-PAGE, and 32 P incorporation was determined by phosphorimaging. 32 P was incorporated into the enhanced yellow fluorescent protein-InsP 3 R1 samples as well as into samples from cells expressing fragment 3 and fragment 5 (Fig. 4B). Fig. 4C shows a Western blot probed with ␣-GFP antibody of the various fragments. Fragment 4, which contains a putative PKA phosphorylation site (KKDS 1687 ) did not incorporate 32 P under these conditions (Fig. 4B). Because fragment 4 was somewhat weakly expressed compared with the other fragments, these data do not formally rule out the possibility that a functional PKA-phosphorylation site is present in this domain (but see subsequent functional data). Nevertheless, these results indicate that PKA phosphorylation sites in InsP 3 R2 are probably harbored between residues 920 and 1583 and between residues 1884 and 2701 (Fig. 4A). There are 5 serine residues present on fragment 3. Each of these serines was mutated to an alanine in isolation. Mutated fragment 3 was also generated with all 5 serines replaced with alanines. Immunoprecipitated wild type and mutated fragment 3 fusion proteins were subjected to in vitro PKA kinase reactions. The serine-free fragment 3 protein did not incorporate 32 P (Fig. 5A). PKA phosphorylation was also completely eliminated by mutation of Ser 937 alone (Fig. 5A). An arginine precedes Ser 937 at position 935, fulfilling the minimum requirement for a PKA phosphorylation site. Ser 937 is unique to InsP 3 R2, and it is present in all InsP 3 R2 sequences in the NCBI databases, thus making it a promising candidate serine for physiological PKA phosphorylation. Furthermore, the corresponding region in InsP 3 R3 contains two of the three PKA phosphorylation sites (Ser 916 and Ser 934 ), making fragment 3 a likely common region for modulation. Two candidate PKA phosphorylation sites present in the sequence for InsP 3 R2-fragment 5 corresponding to Ser 2508 and Ser 2633 in the full-length sequence were also mutated. Fig. 5B shows the results from in vitro PKA reactions with these mutations along with a fragment 5 fusion truncated at position 2512. Kinase reactions with the truncation mutant or the S2633A mutant failed to result in 32 P incorporation, whereas phosphorylation was unaffected by the S2508A mutation, indicating that the PKA phosphorylation site on fragment 5 occurs exclusively at Ser 2633 .
Serine 2633 is located in a consensus Akt phosphorylation site (RMRAMS 2633 ), and this putative Akt phosphorylation site is present in all three InsP 3 R isoforms. Ser 2633 has also been identified as a bona fide Akt phosphorylation site in InsP 3 R1 by two groups (47,48). This site is, however, unlikely to represent a PKA site in the context of the full-length InsP 3 R2, since mutation of the known PKA phosphorylation sites in InsP 3 R1 and InsP 3 R3 completely eliminated PKA-induced 32 P incorporation in vitro and in intact cells (22,49). Presumably, expression of the truncated receptor renders Ser 2633 more accessible to PKA. This could result from an altered conformation of the truncated protein. Regardless of whether Ser 2633 is a substrate of PKA in the full-length receptor, Ca 2ϩ release was still increased in cells expressing S2633A mutated InsP 3 R2 following PKA activation in DT40-M3 cells (Fig. 5D).
The sequence surrounding Ser 937 shows considerable divergence from InsP 3 R1 and InsP 3 R3, making the region attractive for the design of a phospho-specific antibody. Similar strategies have been used to produce antibodies recognizing phosphorylated residues in InsP 3 R1 and InsP 3 R3 (22,50). We designed an antibody to specifically recognize phosphorylated Ser 937 in InsP 3 R2 (␣-Ser(P) 937 ). This antibody failed to recognize InsP 3 R2 immunoprecipitated from untreated COS-7 cells; Ϫ prior to forskolin treatment, immunoprecipitation, and PAGE. A, phosphor image of a dried gel loaded with samples from mouse InsP 3 R2-expressing cells. Greater 32 P incorporation into a ϳ250 kDa band in the forskolin-treated transfected samples indicated phosphorylation of InsP 3 R2. B, mouse InsP 3 R2 was immunoprecipitated from COS-7 cells and incubated with purified PKA at 37°C in the presence of [␥-32 P]ATP for the indicated times. Results are representative of at least two separate experiments for each condition. C, alignments of InsP 3 R sequences around the PKA phosphorylation sites in InsP 3 R1 and InsP 3 R3. None of the sites present in these isoforms are present in InsP 3 R2. however, the antibody readily reported a band in samples immunoprecipitated from forskolin-treated cells (Fig. 6A). Importantly, no signal was detected in parallel samples that were treated with alkaline phosphatase following immunoprecipitation. These data indicate that antibody recognition requires the presence of a phosphate group. Similarly, no signal was detected in samples immunoprecipitated from forskolin/ IBMX-treated cells expressing S937A mutated InsP 3 R2. These results clearly show that mouse InsP 3 R2 can be phosphorylated at Ser 937 by endogenous PKA, further enforcing the idea that InsP 3 R2 is a physiological substrate of PKA.
These data demonstrating that PKA phosphorylates InsP 3 R2 at Ser 937 were obtained using overexpressed recombinant InsP 3 R2. In order to determine if PKA can phosphorylate endogenous InsP 3 R2, we also probed for phosphorylation of Ser 937 in InsP 3 R2 immunoprecipitated from mouse parotid acinar cells. We have demonstrated previously that forskolin treatment induces the apparent phosphorylation of InsP 3 R2 (4). In these experiments, InsP 3 R2 was detected in samples immunoprecipitated from forskolin-treated cells with an antiphospho-Ser/Thr antibody. This experimental paradigm could not formally rule out the possibility that the apparent detection of phosphorylated InsP 3 R2 was the result of co-immunopre-cipitation of non-phosphorylated InsP 3 R2 originally present in a heterotetrameric complex with phosphorylated InsP 3 R1. As shown in Fig. 6B, the ␣-Ser(P) 937 antibody clearly recognized phosphorylated InsP 3 R2 in samples that were immunoprecipitated from forskolin/IBMX-treated parotid acinar cells but not in untreated cells. These results provide strong evidence that PKA phosphorylates endogenously expressed InsP 3 R2 at Ser 937 .
Replacement of Serine 937 with Alanine Eliminates the Enhancing Effects of PKA on InsP 3 R2-The results described above clearly demonstrate that PKA phosphorylates InsP 3 R2 at Ser 937 . We next sought to determine if this site is responsible for the positive effects of raising cAMP illustrated in Fig. 1. We generated a stable cell line expressing InsP 3 R2-S937A and compared the effects of PKA activation on Ca 2ϩ signals with those of a cell line stably expressing wild type InsP 3 R2 (31). The cell lines were transiently transfected with the M3R to allow repeated stimulations with CCh. In these experiments, we utilized 5,6-dichloro-1-␤-Dribofuranosylbenzylimadazole-3Ј,5Јcyclic monophosphorothioate (cBIMPs), a specific, highly cellpermeable cAMP analogue, to activate PKA. Unlike forskolin, this treatment did not alter the basal Ca 2ϩ levels. Application of cBIMPs resulted in enhanced Ca 2ϩ signals in response to low [CCh] (Fig. 7A). The enhancing effects of cBIMPs were only evident in cells that responded minimally to CCh, indicating that PKA sensitizes InsP 3 R2 to low levels of stimulation. This effect was manifested as either a potentiation of a minimal response or alternatively generation of a response following cBIMPs incubation, which was not previously evident in the absence of PKA activation (see examples in Fig. 7A). Following washout of cBIMPs, the response was again reduced to pre-PKA activation levels. Cells stably expressing InsP 3 R2-S937A and transiently expressing M3 receptors did not significantly differ in their concentration versus response relationship to CCh (EC 50 1.11 Ϯ 0.023 versus 0.89 Ϯ 0.02 M; wild typeversus S937A-expressing cells). These data indicate that mutation of S937A per se did not adversely influence the activity of InsP 3 R. However, in contrast to the InsP 3 R2 wild type cells, treatment of cells stably expressing InsP 3 R2-S937A with cBIMPs failed in nine experimental runs from three separate transfections totaling 595 cells representative of various response patterns to result in enhanced CCh-induced Ca 2ϩ signals (Fig. 7B). These results indicate that the sole effect of cBIMPs on InsP 3 R2 is to induce the phosphorylation of Ser 937 .

DISCUSSION
Cross-talk between the cAMP-and Ca 2ϩ -signaling pathways can allow efficient regulation of the temporal and spatial aspects of cellular Ca 2ϩ signals. This shaping of Ca 2ϩ signals is thought to account for the wide range of physiological effects of intracellular Ca 2ϩ . Understanding the molecular determinants behind this cross-talk is important in gaining a clearer picture of cell function and may provide targets for possible pharmaceutical interventions. Phosphorylation of InsP 3 R by PKA constitutes a means of regulating Ca 2ϩ signaling by directly altering the Ca 2ϩ release event. We have described an InsP 3 R2-specific and receptor-distinct mechanism that may account for enhanced Ca 2ϩ signaling in response to cAMP in cells expressing FIGURE 5. PKA phosphorylates fragment 3 of mouse InsP 3 R2 at Ser 937 and fragment 5 at Ser 2633 . Mutants in InsP 3 R2 fragment 3 were generated corresponding to Ser 937 , Ser 990 , Ser 1190 , Ser 1351 , and Ser 1581 . Another mutant harboring all five mutations was also generated (AAAAA). A, the phosphor image from an in vitro PKA assay with wild type, S937A, S990A, S1190A, S1351A, S1581A, and "AAAAA" mutated fragment 3 fusions. PKA phosphorylated all fragments except S937A and "AAAAA," indicating that Ser 937 is the sole PKA phosphorylation site in fragment 3 of mouse InsP 3 R2. Mutations in fragment 5 corresponding to S2508A, a truncation at 2512, and S2633A were generated. B, phosphor image from an in vitro PKA assay with these samples. PKA phosphorylated the wild type and S2508A mutated fragment 5 fusions but failed to phosphorylate the truncated or S2633A mutated fusions. C, schematic diagram depicting the positions of the Ser 937 and Ser 2633 phosphorylation sites in the context of the full-length mouse InsP 3 R2. D shows that, following PKA activation, Ca 2ϩ signals are still augmented in cells expressing InsP 3 R2 S2633A. InsP 3 R2. Specifically, PKA phosphorylation of a unique serine residue (Ser 937 ) increases the single channel P o and Ca 2ϩ release activity of InsP 3 R2.
Phosphorylation of Ser 937 in InsP 3 R2 clearly leads to enhanced channel activity, but the ultimate molecular mechanism behind this effect is still a major unanswered question. We have recently identified effects of phosphorylation on InsP 3 R1 channel activity by examining changes in single channel properties in phosphomimetic mutations of this isoform (20). Specifically, phosphorylation is thought to increase channel activity primarily by altering the probability of the channel exhibiting a high P o "burst" phase. It is currently unknown whether PKA phosphorylation of InsP 3 R2 exerts similar effects on this isoform. Analysis of an S937E/D phosphomimetic mutation could yield further mechanistic insights about the enhancing effects of PKA on InsP 3 R2.
In addition to remaining mechanistic questions, an understanding of the physiological consequence of InsP 3 R2 phosphorylation is lacking. Questions regarding the physiological significance of InsP 3 R phosphorylation will be greatly aided by the analysis of InsP 3 R2 knock-out mice (10,11,51). Cardiac myocytes, liver hepatocytes, and astrocytes all express InsP 3 R2 predominantly (8,9). Analysis of agonist-induced Ca 2ϩ signaling in astrocytes showed that acetylcholine, glutamate, and bradykinin-induced Ca 2ϩ signals were completely absent in cells from InsP 3 R2-KO mice (10). As such, modulation of InsP 3 R2 by PKA and other mechanism would probably be important for astrocytic Ca 2ϩ signaling. Similarly, activation of endothelin receptors in atrial myocytes produces arrhythmogenic Ca 2ϩ signals that are eliminated in cells from InsP 3 R2-KO mice (11). ␤-Adrenergic activation is known to modulate a range of proteins involved in cardiac Ca 2ϩ handling (52). Enhanced InsP 3 R2 Ca 2ϩ sig-  naling by this pathway may exacerbate the arrhythmogenic potential of endothelin receptor activation. Finally, cAMP production is known to potentiate InsP 3 -induced Ca 2ϩ signaling in hepatocytes, where InsP 3 R2 is the predominant isoform (13,27). Phosphorylation of Ser 937 of InsP 3 R2 in hepatocytes could mediate this effect. Consistent with this idea, Ser 937 was identified by mass spectroscopy as being phosphorylated in a global screen of hepatic phosphoproteins (53). The role of PKA phosphorylation of InsP 3 R2 in tissues such as liver and heart should be the subject of further study.
The effects of PKA phosphorylation on InsP 3 R2 were only evident at low levels of stimulation. InsP 3 R2 localization is correlated with sites of Ca 2ϩ wave initiation in hepatocytes (54,55). Similarly, regions of astrocyte ER that exhibit enhanced Ca 2ϩ release are enriched with InsP 3 R2 (56). These results indicate that InsP 3 R2 might be involved with initiating and propagating Ca 2ϩ waves in these cells. Sensitizing InsP 3 R2 to lower levels of InsP 3 by PKA phosphorylation could lower the threshold for initiating global Ca 2ϩ transients. A hierarchy of Ca 2ϩ signals is produced in response to InsP 3 , ranging from Ca 2ϩ puffs involving a few InsP 3 R to Ca 2ϩ waves that recruit multiple Ca 2ϩ release sites (57)(58)(59). Similarly, phosphorylation of a few InsP 3 R2s in a cluster could increase the probability that an elementary Ca 2ϩ signal will result in a global Ca 2ϩ wave.
In summary, the results presented here add significantly to our knowledge of the functional effects and the molecular determinants of PKA regulation of InsP 3 R2. The physiological importance of InsP 3 R2 is only beginning to be understood. It should be noted, however, that this isoform is expressed to some extent in most mammalian cells (8). This suggests that enhancing InsP 3 R2 activity by PKA will have profound effects on Ca 2ϩ signaling and cell function in many different physiological contexts. Similarly, phosphorylation of InsP 3 R2 by other kinases, including PKC (60), Ca 2ϩ calmodulin-dependent protein kinase II (61), and Src kinase (62), has been reported. These and other kinases are thought to impact Ca 2ϩ signaling by regulating InsP 3 R2 activity, but knowledge of phosphorylation sites is lacking. Application of the fragment-based approach described here should help to identify the molecular determinants behind the effects of kinases other than PKA.