Phosphorylation of Inositol 1,4,5-Trisphosphate Receptors by cAMP-dependent Protein Kinase

The ability of cAMP-dependent protein kinase (PKA) to phosphorylate type I, II, and III inositol 1,4,5-trisphosphate (InsP3) receptors was examined. The receptors either were immunopurified from cell lines and then phosphorylated with purified PKA or were phosphorylated in intact cells after activating intracellular cAMP formation. The former studies showed that the type I receptor was a good substrate for PKA (0.65 mol Pi incorporated/mol receptor), whereas type II and III receptors were phosphorylated relatively weakly. The latter studies showed that despite these differences, each of the receptors was phosphorylated in intact cells in response to forskolin or activation of neurohormone receptors. Detailed examination of SH-SY5Y neuroblastoma cells, which express ≥99% type I receptor, revealed that minor increases in cAMP concentration were sufficient to cause maximal phosphorylation. Thus, VIP and pituitary adenylyl cyclase activating peptide (acting through Gs-coupled pituitary adenylyl cyclase activating peptide-I receptors) were potent stimuli of type I receptor phosphorylation, and remarkably, even slight increases in cAMP concentration induced by carbachol (acting through Gq-coupled muscarinic receptors) or other Ca2+mobilizing agents were sufficient to cause phosphorylation. Finally, PKA enhanced InsP3-induced Ca2+ mobilization in a range of permeabilized cell types, irrespective of whether the type I, II, or III receptor was predominant. In summary, these data show that all InsP3 receptors are phosphorylated by PKA, albeit with marked differences in stoichiometry. The ability of both Gs- and Gq-coupled cell surface receptors to effect InsP3 receptor phosphorylation by PKA suggests that this process is widespread in mammalian cells and provides multiple routes by which the cAMP signaling pathway can influence Ca2+ mobilization.

Phosphorylation of Purified InsP 3 Receptors-Receptors were immunoprecipitated from the three cell lines (14,16) and were phosphorylated in a manner similar to that described previously (23). After removal of culture medium, cells were detached with 155 mM NaCl, 10 mM Hepes, 1 mM EDTA, pH 7.4 (HBSE), were centrifuged (500 ϫ g for 2 min), and were disrupted with 12 ml of ice-cold lysis buffer (50 mM Tris, 150 mM NaCl, 1% Triton X-100, 1 mM EDTA, 0.2 mM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, 10 M leupeptin, 10 M pepstatin, 0.2 M soybean trypsin inhibitor, pH 8.0). After 30 min on ice, cells were centrifuged (38,000 ϫ g for 10 min at 4°C). Supernatants were then incubated at 4°C with either CT1, CT2, or CT3 for 1 h and then for a further 1 h with protein A-Sepharose CL-4B. Immune complexes were then isolated by centrifugation (500 ϫ g for 2 min), were washed twice with ice-cold phosphorylation buffer (120 mM KCl, 50 mM Tris, 0.1% Triton X-100, 0.3 mM MgCl 2 , pH 7.2), and were finally resuspended in phosphorylation buffer. Aliquots of washed beads were then placed in 1.5-ml microfuge tubes together with [␥-32 P]ATP (ϳ5 Ci), 0 -5 M nonradioactive ATP, and 20 units of PKA catalytic subunit (final volume, 200 l), were mixed gently, and were then incubated at 30°C. Reactions were stopped by adding 1.3 ml of ice-cold phosphorylation buffer plus 1 mM ATP. Beads were then centrifuged (16,000 ϫ g for 10 s), were washed twice with 1.5 ml of phosphorylation buffer plus ATP, and finally were resuspended in 2ϫ gel loading buffer (14).
Electrophoresis, Immunoblotting, and Autoradiography-To assess the concentration and phosphorylation of purified type I, II, and III InsP 3 receptors and the composition of immunoprecipitates, samples of washed beads were electrophoresed in 4% gels and were either silverstained or immunoblotted as described (14,16), the molecular mass and concentration of receptors being established by comparison to standards of myosin (molecular mass, 205 kDa) and ␤-galactosidase (molecular mass, 116 kDa). Radioactivity associated with electrophoresed InsP 3 receptors was assessed initially by autoradiography of dried gels and then quantitated by excision and scintillation counting of the ϳ240 -280-kDa region.
Phosphorylation Stoichiometry-Based on silver staining, equal amounts of phosphorylated type I, II, and III receptor were electrophoresed, and associated radioactivity was quantitated. The number of moles of phosphate incorporated was then calculated from the specific activity of the [␥-32 P]ATP. The number of moles of InsP 3 receptor loaded onto the gel was determined by measuring [ 3 H]InsP 3 binding to portions of the receptor preparations destined for phosphorylation. Briefly, washed beads in phosphorylation buffer were centrifuged (500 ϫ g for 2 min) and were washed and resuspended in 20 mM Tris, 1 mM EDTA, pH 8.0, and were then incubated with [ 3 H]InsP 3 as described (24). The number of specific binding sites was then determined and was assumed to equal the number of moles of InsP 3 receptor. These experiments also confirmed that the equal amounts of type I, II, and III receptor defined by silver staining bound approximately equal amounts of InsP 3 .
Phosphorylation of InsP 3 Receptors in Intact Cells-This was assessed either directly after labeling cells with 32 P i or with a backphosphorylation procedure (23,25), in which inhibition of PKA-catalyzed transfer of 32 P from [␥-32 P]ATP to purified receptors reveals the degree to which PKA consensus sites are occupied by nonradioactive phosphate in intact cells. In the former procedure, cells were harvested in HBSE, were washed once with phosphate-free, 95% O 2 /5% CO 2gassed minimal essential medium, and were resuspended in the same medium, and 500-l portions were incubated in microfuge tubes with ϳ125 Ci of 32 P i for 30 -45 min at 37°C. Cells were then stimulated, were centrifuged (16,000 ϫ g for 10 s), and were resuspended in 1 ml of ice-cold lysis buffer plus 1 mM Na 3 VO 4 , 100 mM NaF, and 100 nM okadaic acid. InsP 3 receptors were then purified by immunoprecipitation with CT1, CT2, or CT3 and protein A-Sepharose CL-4B and were electrophoresed. In the back-phosphorylation procedure, cells were harvested in HBSE, were washed once in gassed minimal essential medium, and were resuspended in the same medium, and 500-l portions were stimulated. Cells were then centrifuged (16,000 ϫ g for 10 s) and were resuspended in 1 ml of ice-cold lysis buffer plus 1 mM Na 3 VO 4 , 100 mM NaF, and 100 nM okadaic acid. InsP 3 receptors were then purified by immunoprecipitation, were phosphorylated with [␥-32 P]ATP and PKA as described above, and were electrophoresed.
cAMP Measurement-Cells were harvested in HBSE, were washed and resuspended in gassed minimal essential medium (containing 0.25 mM 3-isobutyl-1-methylxanthine (IBMX) in some experiments), and were aliquotted into 100-l portions. After 10 min at 37°C, cells were stimulated for 2 min, and 100 l of ice-cold 1 M trichloroacetic acid was added. After 15 min on ice, samples were centrifuged (16,000 ϫ g for 3 min), and 160 l of supernatant was removed and thoroughly mixed with 40 l of 10 mM EDTA and 200 l of freon/octylamine (1:1). After centrifugation (16,000 ϫ g for 5 min), 100 l of supernatant was neutralized with 50 l of 25 mM NaHCO 3 , and cAMP content was measured by radioimmunoassay. 45 Ca 2ϩ Mobilization and Receptor Phosphorylation in Permeabilized Cells-Cells (1-2 dishes) were harvested in 10 ml of HBSE, were centrifuged (500 ϫ g for 2 min), were resuspended in 10 ml of ice-cold cytosol buffer (120 mM KCl, 2 mM KH 2 PO 4 , 2 mM MgCl 2 , 10 M EGTA, 2 mM ATP, 20 mM Hepes, pH 7.0), were centrifuged again (500 ϫ g for 2 min at 4°C), and were resuspended in 0.8 ml of cytosol buffer. Cells were then permeabilized with nine discharges of a 3 microfarad capacitor (field strength, 3.75 kV/cm) as described (26), were diluted with cytosol buffer to 5 ml, were re-centrifuged, and were finally resuspended at 0.8 -1 mg protein/ml in 1.3 ml of cytosol buffer. Suspensions were then incubated at 25°C without or with 100 units/ml PKA for 10 min and then with ϳ0.3 Ci of 45 Ca 2ϩ /ml for 10 min. Aliquots of cell suspension (90 l) were then added to tubes containing 10 l of either InsP 3 or ionomycin and were incubated at 25°C for 1 or 2 min, respectively. Incubations were terminated by addition of 4 ml of ice-cold cytosol buffer and immediate filtration through Whatman GF/B filters; radioactivity bound to filters was assessed after addition of 4 ml of Ecoscint H and overnight extraction. Radioactivity (the amount of 45 Ca 2ϩ remaining sequestered in the permeabilized cells or " 45 Ca 2ϩ content") is expressed as a percentage of 45 Ca 2ϩ uptake (that remaining sequestered in the absence of stimulus). Importantly, uptake was unaffected by PKA; in the absence and the presence of PKA, respectively, 45 Ca 2ϩ uptake was 37 Ϯ 5 and 38 Ϯ 5 ϫ 10 3 cpm for SH-SY5Y cells, 37 Ϯ 1 and 35 Ϯ 1 ϫ 10 3 cpm for AR4 -2J cells, and 34 Ϯ 2 and 36 Ϯ 2 ϫ 10 3 cpm for RINm5F cells. In some experiments, control or PKA-treated permeabilized cells were centrifuged (16,000 ϫ g for 10 s), and InsP 3 receptors were immunoprecipitated and back-phosphorylated exactly as described for intact cells.

RESULTS
Phosphorylation of Immunoprecipitated Type I, II, and III InsP 3 Receptors-Previous studies have shown that enrichment of type I, II, and III InsP 3 receptors in SH-SY5Y, AR4 -2J, and RINm5F cells, respectively, makes these cells convenient starting points for InsP 3 receptor purification (14,16). Because SH-SY5Y cells contain Ն99% type I receptor, a preparation composed solely of type I receptor can be immunoprecipitated from these cells with antiserum CT1 (14,16). Type II and III receptor preparations immunoprecipitated from AR4 -2J and RINm5F cells with antisera CT2 and CT3 are, in contrast, not homogeneous because they are "contaminated" with traces of co-immunoprecipitating type I receptor, which represents 12 and 4% of total receptor in these cell lines (14,16). Fig. 1A (lanes 1-9) reveals the extent to which the type I, II, and III InsP 3 receptor preparations are phosphorylated by PKA in vitro. Because equal amounts of type I, II, and III receptors were loaded, it is clear that the type I receptor (lanes 1, 4, and 7) is phosphorylated more readily than type II (lanes 2, 5, and 8) or type III (lanes 3, 6, and 9) receptors. The phosphorylation seen is totally PKA-dependent, because none is detected if the kinase is omitted (lanes 10 -12). Fig. 1A also shows that two proteins are phosphorylated in the type II and III preparations, one of which (the upper band) co-migrates with type I receptor.
Immunoblotting was performed to identify the different phosphoproteins ( Fig. 1, B-D). Fig. 1B (lanes 2 and 3) confirms the presence of type I receptor in the type II and III receptor preparations and that the upper phosphorylated band in the type II and III receptor preparations is indeed type I receptor.
Faster migrating type II receptor (Fig. 1C, lane 2) and type III receptor (Fig. 1D, lane 3) correspond in size to the lower of the phosphorylated bands in the type II and III receptor preparations (Fig. 1A). Regarding the efficiency of phosphorylation, it is remarkable that although type II receptor comprises ϳ90% of the type II receptor preparation (14,16), it accounts for only ϳ20% of the total 32 P incorporated (Fig. 1A, lanes 2, 5, and 8), showing that the type II receptor is phosphorylated very poorly. For the type III receptor preparation, however, ϳ90% of 32 P incorporated is type III receptor-associated (Fig. 1A, lanes 3, 6, and 9), confirming that the type III receptor is phosphorylated relatively well. For each receptor, phosphorylation was maximal at 15 min, at which point 0.65 Ϯ 0.06, 0.04 Ϯ 0.01, and 0.14 Ϯ 0.03 mol of P i were incorporated/mol type I, II, and III receptor, respectively (Fig. 2).
InsP 3 Receptor Phosphorylation in Intact SH-SY5Y Cells-Because the type I receptor is a good substrate for PKA in vitro (Figs. 1 and 2), we examined whether it is phosphorylated in intact SH-SY5Y cells in response to cAMP elevation using either a back-phosphorylation procedure (23,25) or 32 P i labeling of intact cells. Fig. 3A shows that vasoactive intestinal peptide (VIP), which stimulates cAMP levels in SH-SY5Y cells (27), causes type I receptor phosphorylation in intact cells because back-phosphorylation was inhibited (lanes 1-5). Pituitary adenylyl cyclase activating peptide (PACAP), which belongs to the same neurohormone family as VIP (28), had a similar effect (lanes 6 -8). Because IC 50 values for VIP and PACAP were 7 and 0.15 nM, respectively (Fig. 3B), and PACAP-I receptors bind PACAP with ϳ100-fold higher affinity than VIP (28,29) and are present in SH-SY5Y cells (29), it is likely that the effects of both PACAP and VIP are mediated by these receptors. Fig. 3C shows that forskolin, which also elevates cAMP levels in SH-SY5Y cells (30,31), similarly inhibited back-phosphorylation (maximally by ϳ60%). This indi-FIG. 2. Stoichiometry of InsP 3 receptor phosphorylation. The number of moles of P i incorporated into type I, II, or III receptors was calculated from the radioactive content of the ϳ240 -280-kDa region of gels and the estimate that 20 and 90% of 32 P i is type II and III receptor-associated in the type II and III receptor preparations. The number of moles of receptor loaded was calculated from the number of InsP 3 binding sites in the preparations. Data shown are the means Ϯ range of two independent experiments.
FIG. 1. PKA-mediated phosphorylation of type I, II, and III InsP 3 receptor preparations. InsP 3 receptors were immunoprecipitated from unstimulated SH-SY5Y, AR4 -2J, and RINm5F cell lysates with CT1, CT2, and CT3, respectively. The type I, II, and III receptor content of these preparations was then quantified by silver staining of electrophoresed samples so that equal amounts of receptor could be loaded onto gels. The positions to which type I, II, and III receptors migrated (approximately 270, 255, and 260 kDa, respectively) are indicated. Data shown are representative of two identical experiments. A, autoradiograph showing phosphorylation that results from incubation of type I, II, or III receptor preparations with [␥-32 P]ATP (5 M) in the presence (lanes 1-9) or the absence (lanes 10 -12) of 100 units/ml PKA. B-D, immunoblots of type I, II, and III receptor preparations probed with CT1 (B), CT1ϩCT2 (C), or CT1ϩCT3 (D). Antibody dilutions were adjusted so that equal amounts of type I, II, and III receptor produced approximately equal amounts of chemiluminescence. inc., incubation; prep., preparation.
cates that the effects of VIP and PACAP are meditated by increases in cAMP alone, a view supported by findings that 1 mM dibutyryl cAMP inhibited back-phosphorylation by 65 Ϯ 4%, whereas 1 mM dibutyryl cGMP inhibited by only 20 Ϯ 4% (mean Ϯ S.E., n ϭ 4). 2 Surprisingly, carbachol also inhibited back-phosphorylation, maximally by ϳ45% (Fig. 3D). This was unexpected because the predominant muscarinic receptor type in SH-SY5Y cells is m3 (32,33), which via phosphoinositidase C generates InsP 3 and diacylglycerol and thus elevates [Ca 2ϩ ] i and protein kinase C activity (33). The effect of carbachol (IC 50 ϭ ϳ1 M) was blocked by atropine and was not additive with a maximal concentration of VIP, suggesting that VIP and carbachol act via the same pathway (Fig. 3D). To probe this effect, cells were exposed to 0.1 M phorbol 12-myristate 13-acetate to activate protein kinases C (30, 33), 0.2 M ionomycin to raise [Ca 2ϩ ] i to a similar extent as 100 M carbachol, 2 and thapsigargin to discharge intracellular Ca 2ϩ stores (26). Although phorbol 12myristate 13-acetate was without effect, both ionomycin and thapsigargin inhibited back-phosphorylation to a similar extent as carbachol (Fig. 3D). Thus, mere elevation of [Ca 2ϩ ] i causes InsP 3 receptor phosphorylation at PKA consensus sequences, suggesting that increases in [Ca 2ϩ ] i that accompany m3 receptor activation (33) mediate the effect of carbachol.
Analysis of 32 P i -labeled intact cells (Fig. 3E) showed that VIP, PACAP, forskolin, and carbachol all increased 32 P incorporation into the type I receptor and that the effect of carbachol was blocked by atropine. Fold increases were relatively small, however, perhaps because the receptor already contains significant amounts of 32 P prior to stimulation (lane 1). 3 Nevertheless, Fig. 3E establishes that the back-phosphorylation procedure reliably reflects events in intact cells and, indeed, that it is the more sensitive way of measuring PKA-mediated effects.
InsP 3 Receptor Phosphorylation in Intact AR4 -2J and RINm5F Cells-Given the limited extent to which type II and III receptors are phosphorylated in vitro (Fig. 1), we examined whether they are phosphorylated in intact cells. Fig. 4A (lanes  1-4) shows that agents likely to raise cAMP levels in AR4 -2J cells (34) inhibit the back-phosphorylation of both type I and II receptors to a similar extent, maximally by ϳ60% (Fig. 4B). Similarly, agents likely to elevate cAMP levels in RINm5F cells (35) inhibit the back-phosphorylation of both type I and III receptors, by far the strongest effect (ϳ75% inhibition) being seen with forskolin (Fig. 4B). Thus, type II and III receptors are phosphorylated in intact cells, a finding confirmed by the fact that the 32 P-content of type II receptors is increased by VIP, forskolin and PACAP in 32 P i -labeled AR4 -2J cells (Fig. 4C,  lanes 1-4) and that forskolin clearly increases the 32 P content of type III receptors in 32 P i -labeled RINm5F cells (lanes 5-8).
Agonist-induced Changes in cAMP Concentration-We next examined cAMP levels in the three cell types to establish the extent to which VIP, PACAP, and forskolin were stimulatory and whether [Ca 2ϩ ] i -elevating agents might also raise cAMP levels, because this could account for the effects of carbachol, ionomycin and thapsigargin on phosphorylation (Fig. 3D).
In SH-SY5Y cells in the absence of IBMX (Fig. 5A, lower panel), VIP and PACAP increased cAMP levels only modestly (by ϳ100%). Thus, IBMX was included to amplify agonist effects and facilitate accurate measurement of potency. In the presence of IBMX (Fig. 5A, upper panel), VIP and PACAP produced ϳ1,000 and 1,200% increases with EC 50 values of 580 and 17 nM, respectively. Interestingly, VIP and PACAP exhibited the same ϳ50-fold potency difference as that seen for phosphorylation (Figs. 3B), but the absolute IC 50 values for inhibition of back-phosphorylation were ϳ100-fold lower than the absolute EC 50 values for stimulation of cAMP formation. Thus, a receptor reserve exists for phosphorylation and submaximal increases in cAMP concentration are sufficient to cause maximal phosphorylation. Carbachol also elevated cAMP levels (Fig. 5A), by ϳ35% in the absence of IBMX and by ϳ130% in the presence of IBMX (EC 50 ϭ 0.27 M). This modest effect was blocked by atropine (Fig. 5A) and was mimicked by ionomycin (0.2 M) and thapsigargin (1 M), which produced ϳ170 and ϳ180% increases, respectively. 2 Thus, the effect of carbachol on cAMP levels appears to result from elevation of [Ca 2ϩ ] i , and InsP 3 receptor phosphorylation correlates with increases in cAMP concentration.
Similarly, in AR4 -2J cells, the extent to which VIP, forskolin and PACAP elevate cAMP levels (Fig. 5B) correlates with the extent to which these agents cause InsP 3 receptor phosphorylation (Fig. 4). In RINm5F cells, only forskolin elevated cAMP levels substantially (Fig. 5B), consistent with it being the only test substance to markedly stimulate phosphorylation (Fig. 4).
Consequences of Receptor Phosphorylation-To establish whether InsP 3 receptor function was modulated by PKA, we examined whether PKA altered InsP 3 -induced Ca 2ϩ mobilization in permeabilized cells. Firstly, however, we sought to establish that PKA could enter and phosphorylate InsP 3 receptors in permeabilized cells. This was found to be the case (Fig.  6A, inset), because back-phosphorylation of type I receptor was much greater in control permeabilized SH-SY5Y cells (lane 1) than in permeabilized cells exposed to PKA (lane 2), and analogous results were obtained for permeabilized AR4 -2J and RINm5F cells. 2 In each cell type, PKA significantly enhanced the potency of InsP 3 by ϳ20% (Fig. 6, A-C) and in SH-SY5Y cells also caused an ϳ10% increase in maximal response (Fig.  6A). These effects on InsP 3 action were truly PKA-dependent, because they were blocked if PKA was denatured by heating (see Fig. 6A legend), a manipulation that also blocked the kinase activity of PKA (Fig. 6A, inset, lane 3). Further, PKA did not enhance ionomycin-induced 45 Ca 2ϩ release (Fig. 6, A-C) or alter Ca 2ϩ uptake (see Experimental Procedures), showing that the PKA-mediated modification of InsP 3 action was not due to a nonspecific change in Ca 2ϩ store characteristics. Finally, combination of the data in Fig. 6 (n ϭ 4) with other independent determinations of InsP 3 potency yielded the following EC 50 values for InsP 3 in the absence or the presence of PKA; for SH-SY5Y cells, 82 Ϯ 4 and 68 Ϯ 3 nM (n ϭ 12), for AR4 -2J cells, 72 Ϯ 3 and 56 Ϯ 2 nM (n ϭ 11), and for RINm5F cells, 400 Ϯ 35 and 301 Ϯ 18 nM (n ϭ 10), respectively, the effect of PKA being significant in all cell types (p Ͻ 0.05, by unpaired t test). That InsP 3 exhibits relatively low potency in RINm5F cells (Fig. 6C) appears to reflect the predominance in this cell type of type III receptor, which has a lower affinity for InsP 3 than type I or II receptors (8, 18, 19, 45).

FIG. 4. InsP 3 receptor phosphorylation in intact AR4 -2J and RINm5F cells.
In panels A and B, InsP 3 receptors were immunoprecipitated from control or stimulated AR4 -2J and RINm5F cells with CT2 and CT3, respectively, and were then exposed to PKA and [␥-32 P]ATP (back-phosphorylated). In panel C, receptors were immunoprecipitated from control or stimulated 32 P i -labeled cells. The arrows mark the migration positions of the type I, II, and III receptors. A, autoradiograph showing back-phosphorylation after 2 min of incubation without stimulus (Ϫ, lanes 1 and 5) or with 1 M VIP (V, lanes 2 and 6), 10 M forskolin (F, lanes 3 and 7), and 0.1 M PACAP (P, lanes 4 and 8); representative of more than three independent experiments. B, quantitation of inhibition of back-phosphorylation. Note that the values shown are derived from the 32 P content of the 240 -280-kDa region of gels and thus are the total of type I plus type II receptor for AR4 -2J cells and type I plus type III receptor for RINm5F cells (means Ϯ S.E., n Յ 3). 0, control. C, autoradiograph showing phosphorylation of InsP 3 receptors in 32 P i -labeled cells incubated as in panel A; representative of two independent experiments.

DISCUSSION
The major findings presented herein are (i) that all of the known InsP 3 receptor types are phosphorylated by PKA in vitro, albeit inefficiently in the case of type III and particularly type II receptors, (ii) that each of the receptors is phosphorylated in intact cells in response to neurohormone receptor activation, (iii) that slight changes in intracellular cAMP concentration cause maximal phosphorylation, and (iv) that PKAdependent phosphorylation enhances Ca 2ϩ mobilization irrespective of which InsP 3 receptor type is predominant.
It is clear from these and previous findings (2,3), that the type I receptor is an excellent substrate for PKA. Previous measurements of phosphorylation stoichiometry in vitro using column chromatography-purified cerebellar type I receptors showed that although relatively high PKA concentrations incorporated 2 mol P i /mol receptor, consistent with the presence of two serines within PKA consensus sequences (2, 3), lower PKA concentrations phosphorylated only serine 1755 and thus incorporated only 1 mol P i /mol receptor (21,36). Recent studies have also indicated that serine 1755 is phosphorylated preferentially in intact cells (37). Our value of 0.65 mol P i /mol SH-SY5Y cell type I receptor is consistent with the view that only one serine, presumably serine 1755, is phosphorylated. It is possible that the incorporation is Ͻ 1 mol/mol in our studies because some of the PKA sites are already occupied by nonradioactive phosphate groups, an argument supported by the fact that the type I receptor is clearly a phosphoprotein in unstimulated cells (Fig. 3E, lane 1). These phosphates remain receptorassociated during immunoprecipitation (Figs. 1-4) but may not survive purification by column chromatography (21,36).
Our work also shows, for the first time, that PKA phosphorylates type II and III receptors, albeit with low efficiency (0.04 and 0.14 mol P i /mol receptor, respectively). This was not an artifact resulting from the incubation of purified receptors and kinases, because both receptors were phosphorylated in intact cells, most notably in response to forskolin. The identity of the sites phosphorylated and the reason why the phosphorylation stoichiometry is so low were not examined in the present study. Regarding the first point, however, although the sites phosphorylated in the type I receptor are not conserved in type II or III receptors, other serines within the PKA consensus sequence (R/K)(R/K)XS (20) are present in the coupling domains of type II and III receptors: in rat and human type II receptors at serine 1687 (8,10), in rat and human type III receptors at serines 934 and 1133, and in rat type III receptor at serine 1460 (9 -11). Thus, it is quite plausible that type II and III receptors are PKA substrates. Regarding the stoichiometry, the low values for type II and III receptors cannot be explained by occupation of phosphorylation sites by nonradioactive phosphate, because the receptors were not heavily phosphorylated in 32 P ilabeled resting cells. Rather, either the conformation of the type II and III receptors or their orientation when tetramerized may make phosphorylation of the consensus sequences unfavorable.
Overall, our studies in intact cells show that modest increases in cAMP concentration result in type I, II, and III InsP 3 receptor phosphorylation. That the type I receptor is phosphorylated in SH-SY5Y cells is, to our knowledge, the first demonstration of this modification in intact neuronal cells and a number of aspects of this work are worthy of comment. Firstly, very modest increases in cAMP concentration are sufficient to maximally stimulate InsP 3 receptor phosphorylation. Thus, a receptor reserve is apparent for VIP and PACAP, which via G s -coupled PACAP-I receptors activate adenylyl cyclase (28,29) to produce Ͼ10-fold increases in cAMP concentration. In contrast, activation of m3 receptors produced much smaller increases in cAMP concentration, and thus, half-maximal values for carbachol-induced phosphorylation and cAMP formation were similar. The mechanism by which carbachol increases cAMP levels is intriguing because m3 receptors are coupled via G q to phosphoinositidase C (33). Previous studies in SH-SY5Y cells (30,31), however, indicate that cAMP elevation can result from activation of Ca 2ϩ -dependent adenylyl cyclases, which are abundant in neuronal cells (38,39). Our observation that ionomycin and thapsigargin elevate cAMP levels supports this view and suggests that carbachol, by elevating [Ca 2ϩ ] i (33), stimulates Ca 2ϩ -activated adenylyl cyclases, which are likely to be present in SH-SY5Y cells (39). Thus, in SH-SY5Y cells and possibly other neuronal cells, cell surface receptors coupled to either G s or G q can increase cAMP levels and thus PKA-dependent type I InsP 3 receptor phosphorylation. Although the effects of this phosphorylation are not currently clear (2,3), these data make it apparent that analyses of Ca 2ϩ mobilization in neuronal cells in response to agonists that act via G q -coupled receptors (e.g. carbachol) should take into account the possibility that type I receptors become phosphorylated at the same time that InsP 3 is activating InsP 3 receptors.
At present, there is little consensus regarding the effects of PKA on InsP 3 receptor function (3). On one hand, PKA enhances the effects of InsP 3 on purified reconstituted cerebellar type I receptors (40) and appears to enhance InsP 3 action in intact hepatocytes (41). On the other hand, PKA inhibits the effects of InsP 3 on cerebellar microsomes (42) and appears to decrease sensitivity to InsP 3 in platelets (43) and trachea (44).
Our studies indicate that PKA causes a modest (ϳ20%) enhancement of InsP 3 potency in SH-SY5Y, AR4 -2J, and RINm5F cells and also enhances by ϳ10% the maximal effect of InsP 3 in SH-SY5Y cells. It is noteworthy that the dual effect of PKA on SH-SY5Y cells parallels the modest effects of PKA on purified cerebellar type I InsP 3 receptor (40).
Thus, PKA enhances InsP 3 action irrespective of the fact that SH-SY5Y, AR4 -2J, and RINm5F cells express different receptor complements. Although it might be expected that InsP 3 potency would be affected in SH-SY5Y and RINm5F cells in which a relatively large proportion of the receptors are phosphorylated, it is somewhat surprising that a potency shift was seen in AR4 -2J cells in which the predominant InsP 3 receptor type is very weakly phosphorylated. It may be significant though that this cell type contains 12% type I receptor (14), which is a excellent substrate for PKA and which forms heterotetramers with type II receptors (14,16,17). Perhaps such heterotetramers are affected by PKA to a similar extent as, or even more than, homotetramers of type I receptor. This is an intriguing possibility, particularly because hepatocytes express type I and II receptors in similar proportion to AR4 -2J cells (13,14,16), and Ca 2ϩ signaling in hepatocytes appears particularly sensitive to PKA (2, 41).
FIG. 6. Effects of PKA on InsP 3 -induced Ca 2؉ mobilization in permeabilized SH-SY5Y, AR4 -2J, and RINm5F cells. Electrically permeabilized cells were preincubated without PKA (open symbols) or with 100 units/ml PKA (closed symbols) or with an equivalent amount of denatured PKA (PKA that was heated at 140°C for 1-3 h) and then with 45 Ca 2ϩ . Cells were then incubated with InsP 3 (E, q) or ionomycin (Ⅺ, f), and the amount of 45 Ca 2ϩ remaining sequestered ( 45 Ca 2ϩ content) was determined. Data shown and quoted below are the means Ϯ S.E. of four independent experiments, and the significance of PKA effects were calculated by unpaired t test. A, SH-SY5Y cells. In the absence and the presence of PKA, respectively, EC 50 values for InsP 3 were 75 Ϯ 2 and 65 Ϯ 2 nM (p Ͻ 0.05), and for ionomycin they were 509 Ϯ 32 and 504 Ϯ 13 nM (not significant). The EC 50 value for InsP 3 in cells exposed to denatured PKA was 74 Ϯ 3 nM and the 45 Ca 2ϩ release curve (not shown for clarity) was superimposable on the curve for cells incubated without PKA. Inset, autoradiograph showing back-phosphorylation of type I receptor immunoprecipitated from permeabilized cells incubated without PKA (lane 1), with PKA (lane 2), or with denatured PKA (lane 3). B, AR4 -2J cells. In the absence and the presence of PKA, respectively, EC 50 values for InsP 3 were 65 Ϯ 4 and 53 Ϯ 3 nM (p Ͻ 0.05), and for ionomycin they were 536 Ϯ 46 and 549 Ϯ 49 nM (not significant). C, RINm5F cells. In the absence and the presence of PKA, respectively, EC 50 values for InsP 3 were 321 Ϯ 31 and 241 Ϯ 8 nM (p Ͻ 0.05), and for ionomycin they were 604 Ϯ 43 and 592 Ϯ 54 nM (not significant).