INTRODUCTION

Nitrovasodilators, nitric oxide (NO),¹ and natriuretic peptides relax vascular smooth muscle through the generation of cGMP (for reviews, see Refs. 1 and 2). The mechanisms by which cGMP causes vascular smooth muscle relaxation are not well understood. Studies have shown that the effects of cGMP are likely mediated through the activation of the cGMP-dependent protein kinase (PKG), and that activation of this kinase leads to the reduction of cytoplasmic Ca²⁺ ([Ca⁺2]_i) in vascular smooth muscle cells (for review, see Ref. 3). Reduction of [Ca²⁺]_i by cGMP has been attributed to several mechanisms including activation of Ca²⁺-ATPases to increase Ca²⁺ uptake or extrusion from the cytoplasm, inhibition of IP₃ formation by inhibition of phospholipase C activation, inhibition of G protein coupling to phospholipase C, inhibition of Ca²⁺ release by the sarcoplasmic reticulum (SR), and activation of Ca²⁺-activated K⁺ channels (for review, see Ref. 3, and references cited). All of these mechanisms appear to contribute to the reduction of [Ca⁺²]_i in different smooth muscle tissues.

Several proteins have been reported to be phosphorylated in response to PKG activation either in vitro or in the intact cell which may contribute to the reduction of [Ca²⁺]_i. Although this list is far from complete, several potentially important substrates include the vasodilator-activated phosphoprotein (4, 5), G_i (6), the -subunit of the Ca²⁺-activated K⁺ channel (7), phospholamban (3), and the type I inositol 1,4,5-trisphosphate receptor (8, 9). Studies using confocal laser scanning microscopy to determine the cellular distribution of PKG suggest that the enzyme is found in the SR (10) where both phospholamban and the IP₃ receptor are localized.

The role of phosphorylation of the IP₃ receptor by cGMP has been studied with respect to the inhibition of agonist-evoked Ca²⁺ release from the SR (8, 9, 11). Several forms of the IP₃ receptor have been identified (12), but the protein isolated and characterized from smooth muscle type I (13) is structurally and functionally similar to the protein isolated from the brain (14). The IP₃ receptor is known to be phosphorylated by several kinases in vitro including cAMP-dependent protein kinase (15), protein kinase C, and Ca²⁺-calmodulin-dependent protein kinase II (16), and by tyrosine kinases during T cell activation (17). The role of phosphorylation of the type I IP₃ receptor is not well understood. It is possible that phosphorylation may regulate Ca²⁺ gating or other regulatory features of the protein. Regardless of the role of phosphorylation of the IP₃ receptor, few studies have demonstrated phosphorylation of this protein in the intact cell in response to second messenger-activated pathways.

In this study we show that the phosphorylation of the IP₃ receptor occurs in the intact rat aorta in response to elevation of either cAMP or cGMP using the technique of back phosphorylation to quantitate the stoichiometry of phosphorylation. Furthermore, these results demonstrate that PKG catalyzes the phosphorylation of the IP₃ receptor in response to both cGMP and cAMP elevation.

EXPERIMENTAL PROCEDURES

Sodium nitroprusside (SNP), S-nitoso-N-acetyl-penicillamine (SNAP), 8-para-chlorophenylthio cyclic AMP (8-CPT-cAMP), and protease inhibitors were obtained from Sigma. Forskolin, KT 5823, and KT 5720 were obtained from Calbiochem-Behring Corp. 8-para-Chlorophenylthio cyclic GMP (8-CPT-cGMP) was purchased from Biolog (La Jolla, CA). Caliculin A was purchased from LC Laboratories. Protein A-agarose was from Life Technologies, Inc. (Grand Island, NY). [-[³²P]ATP was purchased from DuPont NEN. The IP₃ receptor antibody was a gift from Dr. Alan Saltzman (Rhone Poulenc Rorer).

Male Sprague-Dawley rats (250-280 g) were sacrificed by CO₂ asphyxiation, and the aorta were rapidly excised anteriorly below the aortic arch and posteriorly above the bifurcation into the external iliac. Loose fat and connective tissue were carefully stripped, the aortae opened, and the endothelium removed by gentle scraping. The aortic strips were immersed in glass vials with 5 ml of Krebs-Ringer bicarbonate (KRB) buffer containing 117 mM NaCl, 4.7 mM KCl, 1.1 mM MgSO₄, 1.2 mM KH₂PO₄, 25 mM NaHCo₃, 1.5 mM CaCl₂, and 5 mM glucose aerated with 95% O₂, 5% CO₂ at 37 °C and shaken in a water bath to ensure proper diffusion. The strips were equilibrated for 60 min and incubated with different agents for varying lengths of time. The strips were rapidly removed from the medium, blotted once, and frozen in liquid N₂ and stored at 80 °C.

The pulverized aortae were homogenized with 1 ml of 50 mM Tris-Cl, pH 7.7, 0.3 M sucrose, 1 mM EDTA, 100 mM sodium fluoride, 1 mM sodium orthovanadate, 50 mM tetrasodium pyrophosphate, 1 mM phenylmethylsulfonyl fluoride, 10 µg/ml pepstain A, 10 µg/ml leupeptin, 5 µg/ml aprotinin A, 10 nM caliculin A (Buffer A) for 1 min. The homogenate was centrifuged at 5000 × g for 5 min. The supernatant was centrifuged at 100,000 × g for 1 h. The microsomal pellet was stored at 80 °C until used. The microsomal pellet was solubilized with 0.5 ml of Buffer B (Buffer A, in which sucrose was replaced with 1% Triton X-100) on ice using a glass Teflon homogenizer and incubated for 1 h on ice. The samples were centrifuged 14,000 × g for 15 min. The supernatants were removed and protein was estimated. The supernatants were precleared by mixing with 30 µl of Protein A-agarose prebound to 5 µg of rabbit IgG for 1 h at 4 °C. The samples were pulse centrifuged and 2 µl of IP₃ receptor antibody (antisera raised in rabbit to the peptide TFRREADPDDHYQSG which corresponds to amino acid residues 1927-1942 from the type I IP₃ receptor) were added to each supernatant and rotated for 4 h at 4 °C. After 4 h, 30 µl of prewashed Protein A-agarose were added to each sample and rotated overnight at 4 °C. The samples were pulse centrifuged and the supernatants were discarded. The pellets were washed three times with Buffer B and twice with 40 mM Tris-Cl, pH 7.4, 0.1% Triton X-100. The immunoprecipitated IP₃ receptor was separated on SDS-PAGE, transferred to nitrocellulose membrane, and identified by Western blot analysis using IP₃ receptor specific antibodies and Enhanced Chemiluminecent detection system (Amersham Life Sciences).

For the back phosphorylation experiments, the immunoprecipitated samples were phosphorylated using PKG by adding 30 µl of phosphorylation buffer containing 40 mM Tris-Cl, pH 7.4, 0.1% Triton X-100, 10 mM magnesium acetate, 100 µM -[³²P]ATP (5 µCi/sample), and 1 µM cGMP to the immobilized protein preparation. Phosphorylation was initiated by adding 5 µl of PKG (40 nM final concentration) diluted in 20 mM potassium phosphate, pH 7.0, 2 mM EDTA, 150 mM NaCl, 15 mM 2-mercaptoethanol (PEM buffer) containing 1 mg/ml bovine serum albumin and 1 µM cGMP. The samples were incubated at 30 °C for 10 min and the reaction was stopped by adding 10 µl of stopping buffer (312.5 mM Tris-Cl, pH 6.9, 0.5 M sucrose, 15% SDS, 10 mM EDTA, 2.5 M 2-mercaptoethanol, and 0.1% bromphenol blue). The samples were heated at 95 °C for 5 min, and the denatured proteins were resolved on 7.5% SDS-polyacrylamide gel according to the procedure of Laemmli (18). Gels were stained with Coomassie Blue, destained, dried, and subjected to autoradiography at 80 °C. The band corresponding to the IP₃ receptor was cut out of the gel and the radioactivity determined by Cerenkov counting. The autoradiogram was analyzed by densitometry and the amount of ³²P incorporated into the IP₃ receptor band was quantitated.

For the determination of the stoichiometry of IP₃ receptor phosphorylation in intact aortae, the amount of ³²P incorporated in vitro was related to the endogenous level of phosphorylation (19). These calculations were possible since our previous studies defined both the stoichiometry (i.e. 1 mol of phosphate/mol protein) and site (i.e. serine 1755) of phosphorylation of the type I IP₃ receptor by PKG (8). The amount of ³²P incorporated in the IP₃ receptor band from control aortic strips with no agents added was taken as 100%. The amount of ³²P incorporated into the IP₃ receptor band in the treated aortic strips was calculated as a percent of the control value. The decrease in ³²P incorporation in the treated samples compared to control was the moles of phosphate incorporated in the intact cell when PKG is activated. The results are plotted as percent increase in endogenous phosphorylation of the IP₃ receptor.

Cyclic GMP and AMP were determined by radioimmunoassay (20) and protein was determined by Bradford assay (21) using bovine serum albumin as a standard. Bovine lung PKG was purified to apparent homogeneity by affinity chromatography on 8-hexylamine cAMP-agarose as described (22).

RESULTS

In order to study the phosphorylation of the type I IP₃ receptor by PKG in the intact smooth muscle cell, we utilized the technique of back phosphorylation originally described by Forn and Greengard (19). Because both the site and stoichiometry of PKG-dependent IP₃ receptor phosphorylation is known, it was possible to quantitate the extent of phosphorylation by PKG in the intact cell. To do this, it was necessary to isolate the IP₃ receptor from intact aortic tissue in order to both visualize and quantitate phosphate incorporation. Hence, the IP₃ receptor from rat aorta treated with different agents was immunoprecipitated using antisera against an N-terminal peptide from the type I IP₃ receptor, back phosphorylated in vitro using PKG and [-³²P]ATP, and resolved by SDS-PAGE. The data shown in Fig. 1 demonstrate that the IP₃ receptor was immunoprecipitated by the antibody and not by the preimmune serum. The band above 199 kDa was identified as the IP₃ receptor by Western blot analysis using IP₃ receptor specific antibodies.

Treatment of intact rat aorta with NO donor drugs which activate soluble guanylate cyclase resulted in the stoichiometric phosphorylation of the IP₃ receptor in the intact tissue. As shown in Fig. 2, SNP (1 µM) produced a time-dependent increase in the phosphorylation of the IP₃ receptor to a stoichiometry of approximately 0.5 mol of phosphate/mol of receptor above baseline. The endogenous phosphorylation was maximal at 2 min. SNP increased the levels of cGMP from 0.35 ± 0.14 to 5.32 ± 1.34 pmol/mg of protein, while the cAMP level was unchanged (1.93 ± 0.92 to 2.33 ± 0.82 pmol/mg of protein). Phosphorylation of IP₃ receptor was also increased in a concentration-dependent manner when rat aortas were treated with a different NO donor and cGMP elevating agent, SNAP. As shown in Fig. 3, the half-maximally effective concentration of SNAP for stimulating IP₃ receptor phosphorylation was approximately 0.25 µM. This is a concentration of SNAP similar to that which many laboratories have demonstrated produces half-maximal relaxation of contracted rat aortic strips. These results demonstrate that elevation of cGMP levels using concentrations of NO-donor drugs that produce vascular relaxation produce stoichiometric endogenous phosphorylation of the type I IP₃ receptor in intact rat aorta.

To study further the phosphorylation of the type I IP₃ receptor, the effects of activating the cAMP-dependent protein kinase signaling pathway was examined. As shown in Fig. 4, SNP and forskolin increased the endogenous phosphorylation of the IP₃ receptor. On the other hand, angiotensin II (0.1 µM), an activator of both Ca²⁺ and protein kinase C pathways in vascular smooth muscle, neither elevated cyclic nucleotides (2.97 and 0.43 pmol/mg of protein of cAMP and cGMP, respectively) nor increased endogenous phosphorylation of serine 1755 of the IP₃ receptor. Forskolin, however, increased cAMP levels (to 8.22 ± 2.23 pmol/mg of protein) but not cGMP levels (0.52 ± 0.12 pmol/mg of protein) and produced an increase in the phosphorylation of serine 1755 on the type I IP₃ receptor to levels similar to that of SNP in intact tissue. Angiotensin had no significant effect on blocking the SNP-induced IP₃ receptor phosphorylation.

The results shown above suggest that agents which increase either cAMP or cGMP levels lead to the phosphorylation of the IP₃ receptor in intact rat aortas. To determine the kinases involved in this phosphorylation, our first approach was to examine the effects of cyclic nucleotide analogs which are selective PKG and PKA activators. Studies have demonstrated that the 8-para-chlorophenylthio analogs of cyclic nucleotides are selective activators of their respective purified protein kinases and are highly resistant to phosphodiesterase hydrolysis making them useful tools for studying the activation of their respective protein kinases in the intact tissue. As shown in Fig. 5, treatment of aortae for 10 min with 8-CPT-cGMP, a selective PKG activator, increased the endogenous phosphorylation of the IP₃ receptor to 0.5 mol of phosphate/mol of receptor above control in a concentration dependent manner. On the other hand, 8-CPT-cAMP, a selective PKA activator did not increase the phosphorylation of the IP₃ receptor. The phosphorylation in response to 8-CPT-cGMP was rapid and time-dependent with phosphorylation reaching maximal levels between 5 and 10 min of incubation with the analog (Fig. 6). These results suggest that PKG may mediate the phosphorylation of the IP₃ receptor to both cAMP and cGMP.

A further analysis of the role of PKG in the IP₃ receptor phosphorylation was examined using selective inhibitors of PKA and PKG. The structurally related derivatives of staurosporine isolated from Nocardiopsis, KT 5823 and KT 5720, have differing selectivities for inhibiting purified PKG and PKA (23, 24). KT 5823 is selective for PKG (K_i = 234 nM), whereas KT 5720 is selective for PKA (K_i = 56 nM). As shown in Fig. 7, a 20-min preincubation of rat aortae with increasing concentrations of KT 5823 inhibited SNP-dependent phosphorylation of the IP₃ receptor with the half-maximally effective concentration of the inhibitor being approximately 350 nM. KT 5720 was significantly less potent with a half-maximally effective concentration of approximately 1 µM. Interestingly, KT 5823 more selectively inhibited forskolin-dependent phosphorylation of the IP₃ receptor than did KT 5720 with a half-maximally effective concentration of approximately 300 nM compared with approximately 1 µM KT 5720 (Fig. 7). Indeed, the effects of KT 5823 were similar for inhibiting both forskolin- and SNP-dependent phosphorylation of the IP₃ receptor, suggesting that PKG mediates the phosphorylation of this protein in response to elevations in either cAMP or cGMP.

DISCUSSION

The data reported in this article demonstrate that both cGMP and cAMP elevating agents increased in a stoichiometric fashion the endogenous phosphorylation of the type I IP₃ receptor in intact rat aorta. Using the purified protein, our laboratory has shown that PKG catalyzes the phosphorylation of the type I IP₃ receptor isolated from rat cerebellum on serine 1755. The sequence surrounding this site contains the canonical site for phosphorylation by both PKA and PKG (i.e. RRXS) and in addition contains an aromatic residue in the +4 position to the phosphorylatable serine. Colbran et al. (25) have reported that aromatic resides at this position enhance the selectivity for PKG-mediated phosphorylation compared with PKA-mediated phosphorylation. Nevertheless, studies on the phosphorylation of the purified cerebellum IP₃ receptor by Ferris et al. (15) indicate that PKA was capable of phosphorylating both serine 1755 and serine 1589, and we have confirmed this study in our own laboratory (8). Despite these studies, there is no information available on whether or not PKG or PKA catalyze the phosphorylation of the type I IP₃ receptor in the intact cell.

In order to begin to examine the role of PKG in the phosphorylation of this protein in vivo, we used the technique of back phosphorylation to investigate endogenous phosphorylation. This method is quite useful for determining the endogenous stoichiometry of phosphorylation if the site of phosphorylation for the kinase is known and the protein is readily isolated from tissues and cells. Such approaches have recently been used to study the phosphorylation of phospholamban in rat aorta (26), voltage-sensitive sodium channels in neurons (27), and L-type (1) calcium channels in rat skeletal muscle cells (28). The results reported here demonstrate that NO-generating vasodilator drugs stimulate time- and concentration-dependent increases in the phosphorylation of the IP₃ receptor from rat aorta. Furthermore, the concentrations that stimulate IP₃ receptor phosphorylation are similar to those reported by several laboratories including our own that produce relaxation of vascular muscle strips. However, it is important to point out that it is unclear what the role of phosphorylation of this protein is in the relaxation process.

Of particular interest in this study is the fact that elevations of cAMP using forskolin produced phosphorylation of the same residue as elevations in cGMP. However, treatment of rat aortas with 8-CPT-cGMP, but not 8-CPT-cAMP, resulted in the phosphorylation of the IP₃ receptor. Because these cyclic nucleotide analogs are relatively selective for activating their respective kinases, these results do not support a role for PKA in catalyzing the phosphorylation of serine 1755 on the protein. Additional evidence which indicates that PKG, but not PKA, mediates the phosphorylation of serine 1755 on the IP₃ receptor was obtained using the KT compounds. Pretreatment of aorta with KT 5823, a selective inhibitor of PKG, inhibited IP₃ receptor phosphorylation to either SNP or forskolin more potently than did KT 5720, a selective inhibitor of PKA. Inhibition of IP₃ receptor phosphorylation by KT 5823 is well within the range of potency of this compound for inhibiting PKG activity. However, KT 5720 at higher concentrations also inhibits SNP- and forskolin-stimulated IP₃ receptor phosphorylation thereby demonstrating that this compound is capable of inhibiting PKG in the intact cell. These results highlight one important characteristic of protein kinase inhibitors, namely that none of the currently available inhibitors are completely specific for a particular protein kinase. It is therefore important to perform proper dose-response curves when using these compounds to facilitate the interpretation of the data in these types of experiments. Taken together, the data using cyclic nucleotide analogs and protein kinase inhibitors suggest that PKG, but not PKA, mediates the phosphorylation of serine 1755 of the type I IP₃ receptor in intact rat aorta.

The data described here also lend support to the concept that cyclic nucleotide-dependent protein kinases are not necessarily specifically activated by their respective nucleotides. The concept of ``cross-activation'' of PKG by cAMP was originally put forth by Francis et al. (29) and Lincoln et al. (30) for relaxation of vascular smooth muscle. Because the affinity of PKG for cAMP is relatively high (K_a = 2 µM) and the levels of cAMP usually exceed cGMP in tissues, it is likely based on theoretical considerations alone that cAMP activates PKG. Both Francis et al. (29) and our laboratory have confirmed these concepts experimentally in intact smooth muscle tissue and cells using different approaches. Furthermore, cross-activation of PKG by cAMP has been reported by Jiang et al. (31) in swine coronary arteries treated with forskolin. More recently, Cornwell et al. (32) have shown that the growth inhibitory actions of NO in adult rat aortic smooth muscle cells is due in part at least to the activation of PKA by cGMP. These findings clearly indicate that each cyclic nucleotide is capable of activating both cyclic nucleotide-dependent protein kinases in the intact cell.

Because PKA is expressed in rat aortic tissue, and PKA is capable of catalyzing the phosphorylation of the IP₃ receptor on serine 1755 in vitro at least, a question arises as to why PKA does not catalyze phosphorylation of the IP₃ receptor in the intact rat aorta. One possible explanation was provided by Cornwell et al. (10) who demonstrated that PKG, but not PKA, was found to be associated with the SR in rat arotic smooth muscle cells. The localization of PKG with substrate proteins such as the IP₃ receptor in the rat aortic smooth muscle cell SR could facilitate phosphorylation of these proteins. It would appear based on the above discussion that studies using cyclic nucleotide analogs in intact tissues should be interpreted cautiously since their effectiveness depends on several factors such as permeability, rate of hydrolysis, bio-accumulation, and interaction with several receptor proteins.

The role of phosphorylation of the IP₃ receptor is still not understood. Supattapone et al. (33) first demonstrated that PKA-mediated phosphorylation of the IP₃ receptor resulted in diminished potency of IP₃ in releasing Ca²⁺ from brain membrane fractions. Quinton and Dean (34) reported that PKA-dependent phosphorylation of platelet membranes substantially reduced the potency of IP₃ in releasing Ca²⁺ from this preparation. More recently, Cavallini et al. (35) demonstrated that prostacyclin and nitroprusside inhibited IP₃-evoked Ca²⁺ release in intact platelets. These authors suggested that cAMP and cGMP mediate a similar type of IP₃ receptor desensitization, perhaps as a result of PKG-mediated phosphorylation. There are also reports, however, which demonstrate that PKA-mediated phosphorylation of the IP₃ receptor increases the potency of IP₃ in releasing Ca²⁺ in platelets (36) and hepatocytes (37, 38, 39, 40). This variation of results may be due to the capacity of PKA to catalyze the phosphorylation of additional sites on the type I IP₃ receptor protein (i.e. serine 1589), or the tissue specific expression of different IP₃ receptor proteins.

In a recent report by Pfeifer et al. (6), the role of phosphorylation of the IP₃ receptor by PKG was questioned. In this study, the authors used Chinese hamster ovary cells transfected with cDNAs encoding PKG and concluded that PKG did not catalyze the phosphorylation of this protein. Because the type I IP₃ receptor is not uniformly expressed in cultured cells (41), and overexpressed PKG may not always be strategically localized with substrates, it is sometimes difficult to relate the findings from artificial heterologous cell systems with the physiological situation. It is also naive to assume that PKG acts to lower [Ca²⁺]_i by a singular mechanism in cells. At present, it is too early to speculate on the role of PKG-dependent phosphorylation of the IP₃ receptor protein. Studies using the purified receptor in reconstituted systems as well as vascular smooth muscle microsomes may help to elucidate the function of PKG-mediated phosphorylation of the receptor. Since the IP₃ receptor is phosphorylated by several other kinases and because of its complex nature, the regulation by phosphorylation in the intact cells may be different from the regulation of purified receptor in reconstituted systems.