INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

GUANOSINE 3',5'-CYCLIC MONOPHOSPHATE (cGMP) has been established as an important modulator of vascular smooth muscle tone (13, 15, 21, 22, 29), but the mechanisms of cGMP pathways have been studied much less than those of adenosine 3',5'-cyclic monophosphate (cAMP) pathways. Agents that stimulate guanylate cyclase activity, such as nitrovasodilators (e.g., nitric oxide and nitroglycerin) and atrial natriuretic factor (ANF), elevate vascular smooth muscle cGMP levels and induce vasorelaxation (29). These increases in cGMP activate cGMP-dependent protein kinase (PKG), which leads to the lowering of cytosolic calcium (9, 23). The intracellular level of cGMP is determined by the dynamic balance between the activities of guanylyl cyclases, which synthesize cGMP, and phosphodiesterases (PDE), which hydrolyze cGMP into the inactive 5'-GMP. One such PDE is the cGMP-binding, cGMP-specific PDE (PDE5; see Ref. 16).

First detected as a cGMP-binding protein and characterized over a decade ago (14, 25), PDE5 was demonstrated to be a putative receptor for cGMP in a variety of rat, bovine, and human tissues (16). Purified bovine lung PDE5 has been shown to be a chimeric protein containing catalytic sites for the degradation of cGMP and multiple allosteric cGMP-binding sites for potential regulation (36). The location and structure of PDE5 domains were further elucidated by obtaining its cDNA sequence (28). The PDE5 is rapidly and specifically phosphorylated at a single serine (serine-92) by physiological concentrations of PKG or cAMP-dependent protein kinase (PKA), and this occurs only when cGMP is bound to the allosteric binding sites of PDE5 (37). The relative specificity of PKG over PKA for phosphorylating PDE5 is attributed in part to the existence of a phenylalanine located four residues COOH-terminal to the phosphorylation site of PDE5 (5). The selectivity for PKG as the catalyst for this phosphorylation reaction, the stringent requirement for cGMP for this reaction to occur, and the existence of a unique PKG-specific sequence for phosphorylation of PDE5 suggest that phosphorylation may be an important regulator of this enzyme.

The present study investigates the effects of cGMP on two of its intracellular receptors in intact vascular smooth muscle cells (VSMC). These two receptors are PKG and PDE5; the latter is also a putative substrate for PKG in these cells. Due to the importance of cGMP in the control of VSMC relaxation, it is imperative that the role of its intracellular receptors, such as PKG and PDE5, and their interactions be explored to develop a more complete understanding of the role of cGMP in vascular physiology. Furthermore, the role that PDE5 plays in regulating the magnitude and persistence of the cGMP response and the effects of phosphorylation on this enzyme are unknown. The PDE5 is an excellent in vitro substrate for PKG (37), and it is one of very few proteins of defined function that could be a physiological substrate for this kinase (1, 11, 13, 17, 19, 38). This study investigates the possibility that the PDE5 is phosphorylated in vivo and examines the potential role of this phosphorylation.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Subculture and labeling of VSMC. VSMC cultures were derived from the thoracic aorta of 6- to 8-wk-old 250 g Sprague-Dawley male rats using the method of Smith and Brock (35). Briefly, the excised aortas were cleaned of fat and digested in 1 ml/aorta of 2 mg/ml collagenase type II in Hanks' balanced salt solution (HBSS) for 30 min at 37°C. After separating the media from the adventitia and endothelia, the medial layer was further digested in 1.5 ml/aorta of 2 mg/ml collagenase type II and 0.25 mg/ml elastase in HBSS for 2 h. The digestion was halted by diluting the cells with 10 ml/aorta of 20% fetal bovine serum (FBS) in Dulbecco's modified Eagle's media (DMEM). The primary aortic cells were then washed in DMEM, counted with a hemocytometer, and plated in uncoated 100-mm polystyrene petri dishes at a density of 1 × 10⁴ cells/cm² in DMEM with 10% FBS. Cell incubations were performed at 37°C in humidified 95% air-5% CO₂. Confluent monolayers of cells were obtained every 3 days and subcultured at a 1:3 split based on surface area. At this time, each 100-mm dish contained ~1 mg total cellular protein. Early passage cells in this study were subcultured no more than three times. Late passage cells were subcultured >20 times. The purity of the VSMC monolayers was confirmed by indirect immunofluorescent microscopy utilizing a monoclonal antibody specific for smooth muscle -actin (34).

Before the cells were labeled, the medium was removed from the monolayers and the cells were washed with HBSS two times and then covered with 2 ml of DMEM without phosphate containing 0.25 µCi/ml [³²P]orthophosphoric acid for 4 h at 37°C in 95% air-5% CO₂. After the cells were labeled, the medium was removed, and the monolayers were washed three times with 5 ml DMEM. Monolayers were then incubated in the presence or absence of effector agents diluted in DMEM for various time points at 37°C. ANF has been previously shown to relax rat aortic strips in an endothelium-independent manner (11).

Immunoprecipitation of PDE5. VSMC monolayers in 100-mm dishes were treated and washed as described above and then extracted in 0.5 ml of lysis buffer consisting of 50 mM tris(hydroxymethyl)aminomethane (Tris, pH 7.4), 50 mM NaCl, 30 mM sodium pyrophosphate, 50 mM NaF, 1 mM sodium vanadate, 10 mM EDTA, 1.2 mM para-aminobenzamidine, 0.1 mM leupeptin, 75 µM pepstatin A, 0.1 mg/ml aprotinin, and 1 µM microcystin. After addition of lysis buffer, the dishes were thawed, scraped, transferred to microfuge tubes, freeze-thawed three times, briefly sonicated (3 pulses of 6 s), centrifuged 30 min at 8,500 g, and supernatants were collected. Supernatants were incubated for 3-4 h at 4°C with 50 µg rabbit anti-bovine lung PDE5 immunoglobulin G raised in this laboratory (28) against purified bovine lung PDE5. After incubation with 20 µl of 100 mg/ml Sepharose-protein A for 1 h at 4°C, the antigen-antibody complex was sedimented and washed four times in a wash buffer consisting of 50 mM Tris (pH 7.5), 200 mM NaCl, 5 mM EDTA, 0.1% Triton X-100, and 0.05% sodium dodecyl sulfate (SDS). Alternatively, for PDE activation studies, washes were performed in the above wash buffer without Triton and SDS, and the Sepharose-protein A complex was suspended in 20 µl of wash buffer without detergent and used directly in PDE activity assays. Pellets were solubilized in 50 µl Laemmli buffer (20), and equal amounts of protein (30 µg/sample) were run on 8% SDS-polyacrylamide gels and stained with 0.1% Coomassie blue R-250. For autoradiographs, gels were dried onto cellulose acetate and exposed to autoradiographic film at 70°C for various times. The band migrating with the 93-kDa purified bovine lung PDE5 standard (16) was excised and counted in nonaqueous scintillant.

Determination of cyclic nucleotide levels. Cyclic nucleotide levels were determined using a protein kinase activation assay (7). The type I PKA used was partially purified through a DEAE-cellulose chromatography step (30). Cell monolayers were flash-frozen in liquid N₂ after addition of 1 ml 10 mM potassium phosphate, pH 6.8, 1 mM EDTA, and 25 mM 2-mercaptoethanol (KPEM) per dish. The dishes were stored at 70°C until assayed. Monolayers were thawed, scraped, added to microfuge tubes, and boiled at 95°C for 5 min. Tubes were spun at 10,000 g for 30 min, and supernatants were collected. Samples were diluted 1:10 in 10 mM potassium phosphate (KP; pH 6.8) buffer with 0.9 mg/ml bovine serum albumin (BSA), and 20 µl were added to 50 µl stock reaction mixture consisting of 40 mM Tris · HCl (pH 7.4), 20 mM magnesium acetate, 130 µM kemptide (LRRASLG), and 0.2 mM [-³²P]ATP. In some studies, 0.2 mM 3-isobutyl-1-methylxanthine (IBMX) was added to the reaction mixture. Reactions were initiated by the addition of 10 µl PKA diluted to 0.4 nM with KPEM and 0.9 mg/ml BSA. After incubation at 4°C for 16-20 h, 50-µl aliquots were spotted onto phosphocellulose paper (Whatman P-81) and placed immediately into 75 mM phosphoric acid. The papers were then washed five times manually for 1 min each, rinsed for 1 min in ethanol, dried, and counted in nonaqueous scintillant (31).

The assay for cGMP levels was performed similarly to that for cAMP (7). The PKG used was partially purified as previously described (24). Samples (10 µl) diluted 1:10 in KP buffer with 0.9 mg/ml BSA were added to 20 µl stock reaction mixture as above, except that 150 µM heptapeptide substrate (RKRSRAE), which is relatively specific for PKG, was substituted for kemptide. PKA inhibitor-(524) (15 µM) was also added to the reaction mixture. Reactions were initiated with the addition of 10 µl PKG (10 nM), incubated, and halted as described for the cAMP assay. All incubations were performed in duplicate, and each experiment was repeated three or more times. Cyclic nucleotide concentrations (pmol/mg protein) were determined by comparison with a standard curve of cyclic nucleotide-activated kinase activities (pmol · min¹ · ml¹) that was performed concurrently with each experiment. Protein in each sample was measured by the method of Bradford (3) and used to standardize for each experiment.

In vitro phosphorylation. Cell extracts or unlabeled immunoprecipitates (10 µl) were added to 20 µl phosphorylation reaction mixture consisting of 1 µM microcystin, 200 µM IBMX, 100 µM [-³²P]ATP, 24 mM MgCl₂, and 0.9 mg/ml BSA in the presence or absence of 300 µM cGMP. Either PKG or PKA can be used to phosphorylate PDE5 at serine-92 in vitro (37). Reactions were initiated by addition of 0.42 µg purified type I PKG. In a single set of experiments (Fig. 1), 0.14 µg catalytic subunit of PKA was used. In this case, 5 µl of a 1:75 dilution of 2.1 mg/ml purified catalytic subunit of PKA in a dilution buffer consisting of 50 mM potassium phosphate (pH 6.8), 0.1 mM dithiothreitol, and 0.9 mg/ml BSA was added to the aliquots of cell extracts or immunoprecipitates. Incubation proceeded for 10 min at 30°C. Reactions were halted by adding 5 µl of 10% SDS-2-mercaptoethanol (1:1) and boiling for 5 min at 95°C. Equal amounts of protein were then resolved on 8% SDS-polyacrylamide gels. Coomassie-stained gels were dried onto cellulose and exposed to autoradiographic film. Excised bands corresponding to the 93-kDa PDE5 marker protein were counted in aqueous scintillant.

View larger version (64K): [in this window] [in a new window]   Fig. 1.   Immunoprecipitation of cGMP-binding, cGMP-specific phosphodiesterase (PDE5) from rat vascular smooth muscle cell (VSMC) extracts phosphorylated in vitro. Supernatant fractions (S) and PDE5 immunoprecipitates (IP) from VSMC extracts were phosphorylated in vitro in the presence (+cGMP) or absence (cGMP) of exogenous purified catalytic subunit of cAMP-dependent protein kinase (PKA) as described in EXPERIMENTAL PROCEDURES. As controls for the specificity of the immunoprecipitation, some fractions were phosphorylated in the presence of cGMP and then treated with preimmune sera (PIS). Control (C) consists of total phosphorylated extract.

Chromatography of bovine aorta extract. Bovine aortas were obtained fresh from a local slaughterhouse and transported to the laboratory on ice. The medial layer was separated manually from the adventitia and minced in a meat grinder. The tissue (100 g) was added to 500 ml of buffer containing KPEM. The tissue was homogenized in a Waring blender (3 × 30 s) and centrifuged at 10,000 g for 30 min at 4°C. The supernatant was filtered through glass wool and applied to a DEAE-Sephacel column (0.9 × 7 cm) equilibrated in KPEM buffer. The column was washed with 100 ml KPEM and developed with a 50-ml NaCl linear gradient (0-300 mM); 1-ml fractions were collected. The DEAE-Sephacel fractions containing PDE catalytic activity were then applied to a blue Sepharose column (3 × 15 cm), eluted with 0.35 M potassium thiocyanate in KPEM (12, 36), and dialyzed overnight against 0.1 M phosphate-buffered saline (pH 7.4).

Immunoblotting. After stimulation of the cells, the media were removed, and the cells were flash-frozen on liquid N₂ in 0.5 ml lysis buffer as mentioned above. Crude supernatant fractions and PDE5 control immunoprecipitates solubilized (1:1) in Laemmli buffer (20) were prepared. Equal amounts of protein per sample (20-200 µg) were resolved on SDS-polyacrylamide gels and transferred to polyvinylidene difluoride (PVDF) membranes. Transferred gels were stained with Coomassie R-250 to determine completeness of transfer. The PVDF membranes were then incubated overnight with a blocking buffer of 3% BSA in 50 mM Tris (pH 7.5), 150 mM NaCl, and 0.1% sodium azide at 4°C with constant rocking. Membranes were incubated with rabbit anti-bovine lung PDE5 (1:10,000 dilution in blocking buffer) or rabbit anti-bovine lung PKG (1:15,000 dilution in blocking buffer) for 1 h at room temperature. As a control, membranes were incubated in either rabbit preimmune serum or in the absence of primary antibody. Membranes were then washed three times for 20 min each in blocking buffer with 0.02% Nonidet P-40. Blots were incubated for 30 min at room temperature with horseradish peroxidase-labeled goat anti-rabbit secondary antibody (Organon Teknika-Cappel) diluted 1:3,000 in blocking buffer. Membranes were washed again as described, and immunoreactivity was visualized by means of chemiluminescence (Amersham) as detected on autoradiographic film.

cGMP-binding assay. cGMP-binding activity of bovine aorta extracts was measured as described by Francis et al. (14). Briefly, fractions were incubated in KPEM containing 2 µM [³H]cGMP, 10 µM 8-I--phenyl-1,N²-etheno-cGMP [a potent competitive inhibitor of [³H]cGMP binding to PKG (32)], 2 µM cAMP, and 1 mg/ml histone VIII-S at 4°C for 1 h. Samples were then filtered onto 0.45 µm nitrocellulose paper, washed three times in KPEM, dried, and counted in nonaqueous scintillant.

PDE activity. PDE5 was assayed for PDE catalytic activity as described (28). Immunoprecipitates of PDE5 were washed in immunoprecipitation wash buffer without detergent and incubated in a reaction mixture consisting of 100 mM 3-(N-morpholino)propanesulfonic acid buffer (pH 7.5), 10 mM ethylene glycol-bis(-aminoethyl ether)-N,N,N',N'-tetraacetic acid, 0.1 M magnesium acetate, 0.9 mg/ml BSA, 20 µM cGMP, and [³H]cGMP for 15 min at 30°C. Samples were then boiled for 3 min and chilled for 3 min before the addition of 10 µl of 10 mg/ml Crotalus atrox snake venom protein containing 5'-nucleotidase activity. Samples were incubated for 10 min at 30°C, and an equal volume of 20 mM Tris · HCl (pH 7.5) was added. Fractions were then chromatographed on DEAE-Sephacel A-25 columns, and the effluent was counted in aqueous scintillant. The presence of equal amounts of PDE5 was determined by Western blotting. Some immunoprecipitates were phosphorylated in vitro as described above before assaying for PDE activity. The presence of in vitro phosphorylated PDE5 was monitored by SDS-polyacrylamide gel electrophoresis and autoradiography.

Materials. Materials used were from the following sources: Sprague-Dawley rats (Harlan), collagenase II and elastase (Worthington), FBS/DMEM (GIBCO), mouse anti-smooth muscle -actin (Boehringer-Mannheim), [³²P]orthophosphate and enhanced chemiluminescence kit (Amersham), DEAE and phosphocellose P-81 paper (Whatman), heptapeptide substrates for PKA and PKG and PKA inhibitor peptide-(524) (Peninsula Laboratories), [-³²P]ATP (New England Nuclear), [³H]cGMP (Amersham), microcystin (Alexis), PVDF Immobilon (Millipore), Sepharose-protein A and blue Sepharose (Pharmacia), and peroxidase-conjugated goat anti-rabbit immunoglobulin G (Cappel). Atriopeptin II (ANF) and all other reagents were obtained from Sigma.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

Identification of PDE5 in bovine and rat aorta. Previously, we identified PDE5 in rat (14) and bovine (36) lung and purified the enzyme to homogeneity. Polyclonal rabbit antisera were then raised to the purified bovine enzyme, and the antibody was characterized for specificity to bovine lung PDE5 and recombinant PDE5 in crude extracts (28). To determine the presence of PDE5 in vascular smooth muscle, crude homogenates of bovine aorta were partially purified as described for bovine lung (36). Fractionation by centrifugation was followed by DEAE chromatography. cGMP-PDE catalytic activity was determined, and the peak fraction of PDE activity was assayed for PDE5 by Western blot analysis. DEAE fractions with maximal PDE activity were pooled and further purified by blue Sepharose chromatography and again probed by Western blot. Thus the polyclonal antibody raised against purified bovine lung PDE5 cross-reacted with a 93-kDa band in bovine aorta. Similar concentrations of PDE5 appeared to be present in both bovine and rat aorta as in bovine lung (data not shown). The PDE5 also binds cGMP at allosteric sites in the NH₂-terminal portion of the protein (16, 28), and the distribution of cGMP-binding activity in both bovine and rat aorta extracts chromatographed on blue Sepharose correlated directly with the 93-kDa protein band that is recognized by rabbit anti-bovine PDE5 (not shown).

Purified lung PDE5 has been shown to be phosphorylated in vitro in a cGMP-dependent manner by either PKG or the catalytic subunit of PKA. When rat aortic extracts were incubated in the presence of purified catalytic subunit, cGMP, and [-³²P]ATP (as described in EXPERIMENTAL PROCEDURES) and then immunoprecipitated with anti-PDE5, a 93-kDa band was preferentially phosphorylated (Fig. 1). The experiment was also performed using PKG, and no differences in the results were obtained (not shown). No detectable phosphorylation of the 93-kDa band was observed in those extracts phosphorylated in the absence of cGMP. This suggested that aortic smooth muscle PDE5 binds cGMP and is phosphorylated in vitro in a manner similar to that of lung PDE5.

Phosphorylation of PDE5 in intact rat VSMC. To determine if PDE5 could be phosphorylated in vivo, a VSMC monolayer model for rat aortic PDE5 was used. Intracellular ATP pools were radiolabeled with [³²P]orthophosphate for 4 h in these monolayer cells before treating the cells with agents known to elevate cyclic nucleotide levels. PDE5 was then immunoprecipitated from crude extracts and resolved by SDS-polyacrylamide gel electrophoresis. In those cells treated with 0.01 µM ANF or higher, a phosphorylated 93-kDa band was immunoprecipitated by the anti-PDE5 antiserum (Fig. 2A). The intensity of this radiolabeled band was maximal at 60-75 s after addition of ANF but subsided by 90 s (Fig. 2B). Very little phosphorylation of the 93-kDa band was observed in those cells treated with ANF for 3 min or longer (not shown). No significant phosphorylation of immunoprecipitated PDE5 was detected in unstimulated cells. Taken together, these data suggested that ANF, a signaling agent that specifically elevates cellular cGMP, produces a concentration- and time-dependent phosphorylation of PDE5. Isoproterenol, which elevated VSMC cAMP levels (not shown), failed to stimulate the phosphorylation of PDE5 at the isoproterenol concentrations (0.1 nM-0.1 mM) used from 0.5 to 5 min.

View larger version (56K): [in this window] [in a new window]   Fig. 2.   Time course and concentration dependence of ANF stimulation of phosphorylation of PDE5 in intact rat VSMC. A: 32Pi-radiolabeled cells were stimulated with indicated concentrations of atrial natriuretic factor (ANF) for 1 min, and the PDE5 was immunoprecipitated and analyzed as described in EXPERIMENTAL PROCEDURES. B: early passage radiolabeled VSMC were stimulated with 0.1 µM ANF for 15-90 s before harvesting the cells for immunoprecipitating the PDE5. Cells used were from the same passage. No significant differences were detectable in the total amount of protein loaded (by Bradford assay) or by Western blot.

To estimate whether or not a significant proportion of PDE5 was being phosphorylated in the intact VSMC, immunoprecipitates from intact nonradiolabeled VSMC were tested for their ability to be phosphorylated in vitro. Immunoprecipitates of PDE5 cells that had not been treated with ANF were phosphorylated in vitro by adding purified bovine type I PKG (0.42 µg) and 125 µM cGMP (not shown). A similar treatment of PDE5 in the immunoprecipitates after exposure of cells to ANF resulted in ~30-50% less incorporation of radioactivity into the excised bands of immunoprecipitated PDE5 than occurred in the controls. This suggested that ANF stimulation of the cells produced a significant stoichiometric amount of endogenous phosphorylation of PDE5.

VSMC cyclic nucleotide levels. The cGMP and cAMP levels in the VSMC were also determined in control cells and in cells that had been treated with ANF or isoproterenol. The concentration of ANF (0.1 µM) that maximally stimulated PDE5 phosphorylation at 1 min (Fig. 2A) also elicited the maximal increase in cGMP levels at 1 min of stimulation (Fig. 3A, inset). No significant increase in cAMP levels was observed with 0.1 µM ANF treatment. Conversely, levels of cAMP were maximally elevated by 30 s in VSMC treated with 100 µM isoproterenol (Fig. 3B), but there was no significant increase in cGMP. In the presence of 0.2 mM IBMX, cyclic nucleotide levels remained maximal beyond 5 min of stimulation. In the absence of IBMX, ANF-stimulated increases in the cGMP levels were maximally elevated at 1 min but then decreased from 3 to 5 min and returned to less than half-maximal by 1 h (data not shown). The maximal cGMP level resulting from ANF treatment correlated with the time course of the effects of ANF on PDE5 phosphorylation. However, it should be noted that the phosphorylation waned after 1 min, whereas the increase in the cGMP level persisted for 5 min.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Effect of ANF and isoproterenol on cGMP and cAMP levels in rat VSMC. Early passage VSMC were assayed for changes in cGMP and cAMP in response to treatment with 0.1 µM ANF (A) or 100 µM isoproterenol (B) over time as described in EXPERIMENTAL PROCEDURES. Data are expressed per mg extract protein and are representative of 3 separate experiments.

PDE activity in VSMC immunoprecipitates. The catalytic activity of PDE5 in VSMC immunoprecipitates was assayed to compare the changes in PDE activity with the change in cGMP levels and phosphorylation of PDE5 in VSMC treated with 0.1 µM ANF. Using early passage cells, PDE5 was immunoprecipitated, washed in a detergent-free buffer, and assayed for PDE catalytic activity. PDE activity was maximal at 1 min (Fig. 4). This coincided with the time at which the cGMP level was maximal (Fig. 3) and immunoprecipitated PDE5 demonstrated the highest degree of phosphorylation (Fig. 2). PDE activity remained elevated beyond 3 min compared with that observed for unstimulated cells and gradually returned to baseline levels by 1 h (Fig. 4). It should be noted that the phosphorylation of PDE5 declined after 90 s (Fig. 2). Whether or not this apparent dissociation between phosphorylation and activation of PDE5 is due to maintenance of PDE activity by some other cellular process that is initiated by phosphorylation is not known. Immunoprecipitated PDE5 from unstimulated VSMC demonstrated a 2.6-fold increased PDE activity upon phosphorylation in vitro with PKG in the presence of cGMP, as described in EXPERIMENTAL PROCEDURES (not shown). In the absence of PKG or ATP, there was no increase in PDE activity when the PDE5 immunoprecipitate was incubated in the phosphorylation mixture. The fact that IBMX did not completely inhibit PDE activity suggested that the assay contained a small, nonspecific background, which, if subtracted, would make the effect of phosphorylation on PDE5 catalytic activity much larger (7.6-fold) than observed.

View larger version (64K): [in this window] [in a new window]   Fig. 4.   Time course of ANF effect on PDE5 activity in intact rat VSMC. Primary cultures of VSMC were stimulated with 0.1 µM ANF from 1 to 60 min, and PDE5 was then immunoprecipitated as described in EXPERIMENTAL PROCEDURES. Supernatants from the immunoprecipitations were boiled, and the cGMP concentration was measured. Immunoprecipitates were assayed for phosphodiesterase catalytic activity as described in EXPERIMENTAL PROCEDURES. Bars represent SE of triplicates assayed. PDE, phosphodiesterase.

PDE5 activity in passaged VSMC. It has been previously reported that PKG expression is diminished after multiple passages of cultured VSMC (8, 10). In the present study, VSMC of various passages were preloaded with [³²P]orthophosphate for 4 h and subsequently stimulated with 0.1 µM ANF before immunoprecipitating PDE5. No detectable PDE5 phosphorylation was observed in unstimulated VSMC, but, in cells treated with ANF, phosphorylation of PDE5 was maximally demonstrated in first-passage cells with some PDE5 phosphorylation occurring up to the fourth passage (Fig. 5, top). Significantly, ANF had little effect on the extent of PDE5 phosphorylation in late passage (passage 6 and beyond) VSMC. However, Western blotting of these same samples of PDE5 immunoprecipitates demonstrated the presence of approximately equivalent amounts of PDE5 in cells from each passage (Fig. 5, middle). Indeed, immunoprecipitated PDE5 from late passage cells could be phosphorylated in vitro by the addition of PKG and cGMP (not shown). As reported by others (8, 10), we detected no PKG by Western blot in the VSMC after passages 3 or 4 (Fig. 5, bottom). This suggested that the absence of ANF-stimulated PDE5 phosphorylation was not due to the loss of PDE5 protein in the mid- to late passage VSMC. Rather, the lack of PDE5 phosphorylation may be due in part to the absence of a protein kinase, such as PKG, in these late passage cells.

View larger version (66K): [in this window] [in a new window]   Fig. 5.   Effect of number of passages of VSMC on phosphorylation (top) and immunoprecipitation (bottom) of PDE5. 32P-radiolabeled cells from passages 1, 3, 4, and 12 were stimulated with 0.1 µM ANF for 1 min and then lysed. Extracts from these cells were then treated with antibody to either PDE5 (BPDE) or cGMP-dependent protein kinase (cGK). The resulting immunoprecipitates were then subjected to SDS-polyacrylamide gel electrophoresis and Western blotting to determine the presence of PDE5 and PKG in each passage sample.

The effects of ANF on PDE catalytic activity of PDE5 were also diminished in late passage cells compared with early passage VSMC. When VSMC were treated with 0.1 µM ANF before immunoprecipitation of PDE5, the early passage immunoprecipitated PDE5 demonstrated large increases in PDE activity (Fig. 6). The later passage cells exhibited no ANF effect on PDE activity. However, the immunoprecipitated PDE5 from untreated VSMC across all passages assayed (not shown) could be phosphorylated in vitro with PKG and cGMP to produce increased PDE activity compared with nonphosphorylated immunoprecipitates. This suggested that the components required for stimulation of PDE activity in vitro by addition of PKG and cGMP are present throughout all passages of VSMC, whereas some component(s) that is necessary for the effects of ANF on PDE activity in intact VSMC, such as PKG, is present only in early passage VSMC. The data also imply that, even though PKA phosphorylates PDE5 in vitro, the level of activated PKA is insufficient to catalyze this phosphorylation reaction under the intracellular conditions in these late passage cells.

View larger version (23K): [in this window] [in a new window]   Fig. 6.   cGMP phosphodiesterase activity in various passages of VSMC. Cells were incubated in the presence or absence of 0.1 µM ANF for 1 min, PDE5 was immunoprecipitated, and immunoprecipitates were assayed for PDE catalytic activity as described in EXPERIMENTAL PROCEDURES. Bars represent SE of triplicates assayed.

To confirm that PDE5 is phosphorylated by PKG in intact VSMC, we exploited the fact that cGMP binding to PDE5 is required for phosphorylation. These experiments employed the use of 8-bromoguanosine 3',5'-cyclic monophosphate (8-BrcGMP), a specific activator of PKG that binds poorly to PDE5 (16). VSMC were preincubated with 1 µM 8-BrcGMP for 30 min before stimulating the [³²P]orthophosphate-labeled cells with 0.1 µM ANF. Phosphorylation of immunoprecipitated PDE5 was augmented in those cells treated with 8-BrcGMP before ANF stimulation (Fig. 7). No phosphorylation of PDE5 was detected in those cells treated with 8-BrcGMP alone, which was consistent with previous studies demonstrating that 8-BrcGMP does not bind to PDE5 (13, 14). Binding of cGMP to the allosteric cGMP binding sites of PDE5 is required for phosphorylation of this enzyme in vitro. This suggested that a specific activator of PKG, 8-BrcGMP, can augment the phosphorylation of PDE5 in the intact VSMC in response to ANF treatment. However, because 8-BrcGMP does not bind to PDE5, a cGMP-elevating agent such as ANF is also required to produce the cGMP-bound form of PDE5, which is necessary for phosphorylation of this enzyme. No detectable phosphorylation of PDE5 was observed in any immunoprecipitate from late passage cells, including those that had been treated with ANF. In late passage cells treated with 1 µM 8-BrcGMP and ANF, there was still no detectable PDE5 phosphorylation (data not shown).

View larger version (48K): [in this window] [in a new window]   Fig. 7.   Immunoprecipitation of phosphorylated PDE5 from early and late passage intact rat VSMC. 32P-radiolabeled cells were incubated with 0.1 µM ANF for 1 min and/or 1 µM 8-BrcGMP (an analogue specific for PKG) for 30 min before immunoprecipitation. Data are representative of 3 experiments.

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

The PDE5, classified as PDE5A (2), has allosteric binding (noncatalytic) sites that are highly specific for cGMP. When cGMP is bound to these sites in vitro, the PDE5 is phosphorylated by PKG and with lesser efficiency by PKA. However, before this study, investigations of the phosphorylation of PDE5 had not been performed in the intact cell, and studies using the purified enzyme have not revealed a change in the catalytic activity of the PDE upon phosphorylation. The present study suggests that PDE5 is phosphorylated in smooth muscle cells in response to ANF-induced elevation of cGMP levels. Under these conditions, the elevation of cGMP would be predicted to activate PKG, and cGMP binding to PDE5 would expose the phosphorylation site for phosphorylation by PKG. The results provide the first evidence that PDE5 can be phosphorylated in the intact cell in response to hormonally induced increases in cGMP. The results also demonstrate that PDE5 is a specific cellular PKG substrate, the state of phosphorylation of which can be related to physiological changes in cGMP.

Tissues with a high content of smooth muscle have been shown to contain high concentrations of PKG. However, other cellular receptors for cGMP, such as PDE5, exist in significant quantities in vascular smooth muscle. A dynamic homeostasis of cGMP levels may exist in these cells whereby the effects of agents that increase intracellular cGMP pools are balanced by a negative feedback mechanism associated with increased PDE activity of PDEs such as PDE5. A similar mechanism for the cAMP cascade has already been established (6, 33). However, for this latter cascade, it has not been found that cAMP binds directly to a PDE as well as to the protein kinase, which is the case for the cGMP cascade. Although aortic smooth muscle contains several PDEs (27), the PDE5 is unique in that cGMP binding to allosteric binding sites exposes a phosphorylation site on the enzyme, thus providing for phosphorylation of PDE5 by the cGMP-activated PKG. This dual control of phosphorylation of PDE5 through the effects of cGMP on both the PKG and the PDE5 provides for precise coordination between two major intracellular receptors for cGMP.

Several lines of evidence point to the phosphorylation of PDE5 by PKG in the intact VSMC. Agents such as ANF, which are associated with cGMP elevation and PKG activation, stimulate PDE5 phosphorylation in the intact cell. After PDE5 phosphorylation, cGMP levels progressively decrease. This ANF-mediated PDE5 phosphorylation is enhanced by preincubating VSMC with 8-BrcGMP, a specific activator of PKG. PDE5 phosphorylation does not occur in the presence of 8-BrcGMP alone, presumably because cGMP binding to PDE5 is required for phosphorylation of the PDE5 to occur. Under these conditions, PDE5 phosphorylation may be augmented when PKG is activated by low levels of 8-BrcGMP in the presence of cGMP levels that are sufficient to expose phosphorylation sites on the PDE5. Furthermore, late passage cells, which have no measurable PKG, fail to demonstrate phosphorylation of PDE5 after ANF treatment. However, the PDE5 from these same cells can be immunoprecipitated and phosphorylated in vitro by PKG.

The phosphorylation of PDE5 by PKG is associated with an increase in the rate of cGMP hydrolysis by the immunoprecipitated PDE5. However, it should be emphasized that the association between the increase in activity of the PDE5 and phosphorylation of the enzyme is correlative. Other more direct approaches should be used in the future. We have been unable to demonstrate activation of the purified enzyme by phosphorylation. Perhaps some factor(s) present in the crude system is required to elicit the activity change. Burns and Pyne (4) have reported that phosphorylation of partially purified PDE5 from guinea pig lung by added catalytic subunit of PKA in the absence of cGMP causes enzyme activation.

Unlike other studies involving only in vitro characteristics of PDE5, this study investigates a physiological role for PDE5 in vascular smooth muscle. Thus, PDE5, a protein with known physiological function, is one of the first demonstrated specific substrates of PKG in intact cells. The implication of PDE5 in the regulation of cGMP levels in VSMC suggests a possible mechanism for a negative feedback loop via an enzyme that degrades the very signal that activates it, cGMP. This would imply that cGMP levels, or possibly smooth muscle contraction/relaxation, are determined by regulation of both the synthesis (guanylyl cyclase) and the degradation (PDE5) of the cyclic nucleotide. Due to the importance of PKG in the control of vascular smooth muscle relaxation (15, 18, 22, 26, 29), understanding the role of other receptors for cGMP, such as PDE5, is central to understanding the role of cGMP in both the physiology and pathology of cardiovascular function.