Activation of cGMP-dependent protein kinase by protein kinase C.

The cGMP-dependent protein kinases (PKG) are emerging as important components of mainstream signal transduction pathways. Nitric oxide-induced cGMP formation by stimulation of soluble guanylate cyclase is generally accepted as being the most widespread mechanism underlying PKG activation. In the present study, PKG was found to be a target for phorbol 12-myristate 13-acetate (PMA)-responsive protein kinase C (PKC). PKG1alpha became phosphorylated in HEK-293 cells stimulated with PMA and also in vitro using purified components. PKC-dependent phosphorylation was found to activate PKG as measured by phosphorylation of vasodilator-stimulated phosphoprotein, and by in vitro kinase assays. Although there are 11 potential PKC substrate recognition sites in PKG1alpha, threonine 58 was examined due to its proximity to the pseudosubstrate domain. Antibodies generated against the phosphorylated form of this region were used to demonstrate phosphorylation in response to PMA treatment of the cells with kinetics similar to vasodilator-stimulated phosphoprotein phosphorylation. A phospho-mimetic mutation at this site (T58E) generated a partially activated PKG that was more sensitive to cGMP levels. A phospho-null mutation (T58A) revealed that this residue is important but not sufficient for PKG activation by PKC. Taken together, these findings outline a novel signal transduction pathway that links PKC stimulation with cyclic nucleotide-independent activation of PKG.

mobilization, and ion channel function are likely to be central to many processes (4 -15).
Two genes encode three isoforms of PKG, of which type 1 produces two splice variants. All isoforms are activated by micromolar increases in cellular cGMP levels that bind to tandem cyclic nucleotide binding sites in the regulatory domains (16 -20). Binding of cGMP causes conformational changes that lead to elongation of the enzyme, presumably enabling access of substrates to the active site. As with most protein kinases in this group, PKG activation has been associated with isoformspecific autophosphorylation resulting in different properties of the enzyme, including increased sensitivity to cyclic nucleotides and constitutive activity (21)(22)(23)(24)(25)(26)(27)(28).
Elevation of cellular cGMP levels occurs in response to stimulation of the natriuretic peptide receptors, which have cytosolic guanylyl-cyclase activity, or more commonly by activation of soluble guanylyl-cyclase by nitric oxide (4,9,29,30). Numerous studies have inferred activation of PKG in response to diverse ligands, presumably downstream of nitric oxide production, although rarely have NO levels or PKG activity been measured. The activation of PKG by endothelial nitric-oxide synthase pathways can occur rapidly and is probably the case in endothelial cells stimulated with vascular endothelial growth factor (31,32). However, stimulation of human neutrophils with N-formyl-peptides or lipopolysaccharide also leads to a rapid activation of PKG in the absence of endothelial or neuronal nitric-oxide synthase (8,33,34).
The present work provides evidence that PKG is a substrate for protein kinase C both in vitro and in vivo, and phosphorylation results in PKG activation. This finding highlights a novel signal transduction pathway that provides an alternative cGMP-independent mechanism for the activation of PKG.

EXPERIMENTAL PROCEDURES
Tissue Culture and Reagents-HeLa epithelial cells and HEK-293 fibroblasts harboring the T-antigen were maintained at 37°C, 5% CO 2 in T25 flasks at subconfluent density. All tissue culture media and reagents were from Invitrogen unless otherwise indicated. Cells were grown in RPMI 1640 medium containing 10% fetal bovine serum, 200 M L-glutamine, 10 IU/ml penicillin, 10 g/ml streptomycin. Cells were expanded for an experiment by trypsinization and dilution into either six-well plates or 10-cm dishes as appropriate. The expression constructs encoding the tagged dominant negative regulatory subunit of PKA and the tagged VASP have been described previously (6,35). All reagents and chemicals were from Fisher unless otherwise indicated.
Production of PKG Expression Plasmids-The cDNA encoding fulllength human PKG1␣ and the regulatory region of PKG1␣ fused at the C terminus with a FLAG epitope have been described previously (36). For the present studies, a FLAG epitope (DYKDDDDK) was added to the amino terminus using PCR, and the tagged product was subcloned into the pCDNA3 expression vector. There was no difference noted between wild type and tagged PKG with respect to in vitro kinase assays (see below) or phosphorylation of VASP (data not shown). Generation of point mutations to convert the threonine residue at position 58 of PKG1␣ to alanine or glutamate was accomplished using QuikChange™ according to the manufacturer's instruction using hu-* This work was supported by an investigator grant from the Arthritis Foundation (to D. D. B.) and an American Heart Association grant (to N. O. D.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
ʈ To whom all correspondence should be addressed:  GGGCGCAGGGC for Thr 58 3 Ala and CACATCGGCCCCCGGGAGA-CCCGGGCGCAGGGC for Thr 58 3 Glu. All constructs were sequenced  to verify correct substitution, and the recombinant proteins were tested  for expression by Western blotting with an anti-PKG C-terminal antibody (see below).
Expression and Purification of PKG-PKG1␣ cDNA was transiently transfected into either HEK-293 cells or HeLa cells using Lipo-fectAMINE-Plus™ according to the manufacturer's instructions (Invitrogen). Generally, cells were split the day before transfection to low cell densities (30%), and the following morning they were transfected. This low cell density was essential for experiments involving cotransfection of VASP and PKG, because at higher cell densities a high background phosphorylation of VASP was observed (data not shown). Cells were incubated in the DNA/liposome mixture in half-volume serum-free RPMI 1640, and after 3 h they were supplemented with an equal volume of medium containing 10% serum for a final concentration of 5% serum overnight. For some experiments, 5 h following the addition of DNA, the medium was replaced with serum-free RPMI 1640 before incubation overnight. This method produced ϳ70% transfection of HEK-293 cells and 30% HeLa as measured using enhanced green fluorescent protein vectors (BD Biosciences Clontech; Palo Alto, CA).
In order to purify PKG1␣, the FLAG-tagged fusion protein was transiently expressed in HEK-293 cells. Typically, 1 g of DNA/10-cm dish was used, and three dishes were transfected for each purification. On the morning following transfection, the cells were placed on ice, and the medium was aspirated and replaced with 10 ml of ice-cold PBS. All subsequent steps were performed on ice or at 4°C. The cells were scraped into the PBS and subsequently concentrated by centrifugation. The cell pellets were resuspended (1 ml/10-cm dish) in lysis buffer (50 mM Tris-HCl, pH 8.0, 1 mM EDTA, 1% Nonidet P-40, 150 nM NaCl) containing phosphatase and protease inhibitor cocktails (Invitrogen). Lysis was for 20 min at 4°C followed by clarification by centrifugation at 10,000 ϫ g for 10 min. The expressed PKG was immunopurified in batch using anti-FLAG-M2-agarose beads (Sigma). The extracts were pooled, and 200 l of 50% (v/v) anti-FLAG-M2-agarose (Sigma) was added to the extract and rocked for 1 h at 4°C. The beads were washed three times in lysis buffer and once in PBS. The PKG attached to the beads was then eluted by resuspension in 100 l of PBS containing 400 M FLAG peptide (Calbiochem). After 15-min agitation at 4°C, the beads were pelleted by centrifugation, and glycerol was added (50%, v/v) to the supernatant containing the PKG and stored at Ϫ80°C until needed.
Immunoprecipitation and Western Blotting-For immunoprecipitation, cells grown in six-well plates, or 10-cm dishes were transfected with plasmids encoding FLAG-PKG or FLAG-G1␣R as detailed above. After experimental treatment of the cells, the plates were placed immediately on ice, and the medium was replaced by ice-cold PBS. For 10-cm dishes, the cells were scraped into the PBS and harvested by centrifugation at 4°C, and the pellet was resuspended by repeat pipetting in 1 ml of lysis buffer. The cells grown in six-well plates were washed once in PBS, and 0.5 ml of lysis buffer was added directly to the wells. Cell lysis was by agitation for 30 min at 4°C, and the extracts were clarified by centrifugation at 10,000 ϫ g for 10 min. The supernatants were either used directly for Western blotting by boiling an aliquot for 5 min in PAGE sample buffer, or specific proteins were precipitated by adding 30 l of 50% (v/v) anti-FLAG-M2-agarose beads. Precipitation was performed by rocking for 1 h at 4°C, followed by washing of the beads three times in 1 ml of lysis buffer and once in PBS. Precipitated proteins were eluted from the beads by adding 30 l of PBS containing 400 M FLAG-peptide and incubation for 30 min on ice. As with the lysates, the eluted proteins were boiled in SDS-PAGE sample buffer for 5 min.
Protein samples were separated routinely on 10% polyacrylamide minigels and transferred to supported nitrocellulose. The blots were blocked by incubating at room temperature for 30 min in PTS (PBS containing 5% bovine serum albumin and 0.25% Tween 20). Blots were subsequently probed either for 1 h at room temperature or overnight at 4°C, followed by three 5-min washes in PTS buffer. Blots were then incubated for 1 h in peroxidase-conjugated secondary antibodies, and after extensive washing, the proteins were visualized using chemiluminescence according to the manufacturer's instruction (Pierce).
Phosphorylation of PKG-The phosphorylation of PKG in vitro was assessed using either purified FLAG-PKG immobilized on anti-FLAG-M2-agarose beads or PKG1␣ purified from SF9 cells (Calbiochem). For measurement of PKG phosphorylation in vivo, HEK-293 cells grown in six-well plates were transiently transfected to express FLAG-PKG1␣ as detailed above. Cells were washed twice in phosphate-free Dulbecco's modified Eagle's medium and incubated for 3 h in Dulbecco's modified Eagle's medium containing 1 mCi/ml [ 32 P]PP i (Amersham Biosciences). Following loading of the cells, the medium was removed and replaced with prewarmed Dulbecco's modified Eagle's medium with or without activators or inhibitors as appropriate. After 1 h of incubation, the cells were harvested, and the PKG was immunoprecipitated as detailed above. Phosphorylation of the PKG was assessed by SDS-PAGE and autoradiography.
Measurement of PKG Activation-HEK-293 cells were transfected with wild type FLAG-PKG, and following overnight starvation they were stimulated with 100 M 8-Br-cGMP or 100 nM PMA for 1 h. Cells were harvested, and the transfected PKG was purified by immunoprecipitation as detailed above. Quantitation of the PKG was performed by densitometric analysis of Coomassie Blue TM -stained SDS-polyacrylamide gels in which the FLAG-PKG was compared with a PKG1␣ standard purchased from Calbiochem. Reactions (25 l) contained 5 l of PKG (ϳ50 ng), 100 M BPDEtide, 0.1 M H-89 (Calbiochem), and 10 M 8-Br-cGMP where indicated. Reactions were incubated for 10 min at 30°C in PCR tubes and stopped by placing on ice. Duplicate 12-l aliquots were spotted on numbered P81 phosphocellulose filters and washed five times in 0.75% phosphoric acid, 2 min in acetone. The phosphate transferred to the peptide on the filters was measured by scintillation counting. For each sample, duplicate reactions were performed, and each experiment was repeated at least three times.
Antibody Production-Antibodies were produced in rabbits using standard methods for immunization and serum preparation. The antigens were 10 -15-mer peptides containing an amino-terminal cysteine residue that was used for coupling to keyhole limpet hemocyanin using the Imject maleimide-activated mcKLH kit (Pierce). A peptide corresponding to the C-terminal 10 amino acids of type 1 PKG was used to generate the anti-PKGct antibodies, which recognize the carboxyl terminus of both PKG1␣ and PKG1␤. The peptide sequences used to generate the anti-phospho-Thr 58 antibodies was CAIGPRT*TRAQG-ISAEP (where the asterisk denotes a phosphorylated residue). The sera obtained from inoculated rabbits were tested for antibody titer using a standard enzyme-linked immunosorbent assay against the antigenic peptide.

Activation of PKC by PMA Leads to PKG-dependent VASP
Phosphorylation-VASP is a well characterized substrate for the cyclic nucleotide-dependent protein kinases and can be phosphorylated on several residues by both PKA and PKG (37). Phosphorylation of VASP specifically on serine 157 causes a shift in mobility on SDS-polyacrylamide gels, which is widely used as a measure of kinase activation in vivo. In our studies of VASP phosphorylation, an electrophoretic shift of VASP was observed in HEK-293 cells in the presence of the phorbol ester PMA. In cells treated with 100 nM PMA, phosphorylation of VASP was detected as early as 15 min with a maximum at 1 h (Fig. 1A). The phosphorylation of VASP observed following the addition of PMA was confirmed in Western blots probed with phosphospecific antibodies directed against both serine 157 and serine 239, which are the preferred sites for PKA and PKG, respectively (data not shown). As an analog of diacyl-glycerol, PMA binds to a subset of PKC isoforms, leading to their acti-vation. The importance of PKC in the PMA-stimulated phosphorylation of VASP was examined using the specific PKC inhibitors Gö-6983 and Rö-32-0432 (Fig. 1B). In these experiments, both inhibitors were able to reduce the PMA-induced shift of VASP to basal levels.
The phosphorylation of VASP on serine 157 suggested the involvement of the cyclic nucleotide-dependent protein kinases downstream of PKC activation in these cells. In order to examine the involvement of PKG and PKA, HeLa cells were used, since they lack detectable PKG (Fig. 2). Stimulation of HeLa with 100 M cAMP produced a significant shift in VASP on polyacrylamide gels, as was expected due to the ubiquitous expression of PKA. That PKA was responsible for the cAMPinduced VASP shift in HeLa cells was confirmed by the lack of effect of cAMP in cells transfected to express a mutant regulatory subunit that behaves as a dominant negative for PKA (⌬R1␣). In contrast, VASP did not shift in HeLa treated with either 100 M 8-Br-cGMP or with 100 nM PMA. Although overexposure of Western blots detected a faint band at the size of PKG, the lack of effectiveness of cGMP on VASP phosphorylation confirmed the absence of functional PKG in these cells. Transfection of HeLa to express modest levels of PKG1␣ (0.1 g/well of a six-well plate) conferred VASP responsiveness (mobility shift) in response to both 8-Br-cGMP and PMA. Of interest, co-expression of ⌬R1␣ did not inhibit but notably enhanced the effectiveness of both cGMP and PMA to stimulate a VASP mobility shift. These data indicated that PKG but not PKA was important to VASP phosphorylation in cells treated with PMA.
Phosphorylation of PKG1␣ by PKC in Vivo and in Vitro Leads to Enzyme Activation-Since current literature revealed no precedence for a role of PKG downstream of PKC, experiments were designed to further characterize this relationship. The work with HeLa cells suggested activation of PKG in response to treatment with PMA but did not suggest a mech- anism. To address this issue, HEK-293 cells were transfected to overexpress FLAG-PKG1␣ and treated with 8-Br-cGMP or PMA (Fig. 3). Following stimulation, the PKG was immunoprecipitated from the cells and examined for activity in vitro using BPDEtide as substrate. The PKG from unstimulated cells exhibited an ϳ2-fold increase in activity in the presence of cGMP. In cells treated with 8-Br-cGMP, the washed PKG demonstrated identical enzyme activity in vitro to that obtained from unstimulated cells. This was consistent with previous reports demonstrating rapid inactivation of PKG1␣ upon removal of cGMP. In contrast, the PKG obtained from cells that had been treated with PMA had a much higher basal activity than controls, reaching ϳ75% of maximum activity of control PKG activated by cGMP. The high basal activity of the PKG from PMA-treated cells was further enhanced in the presence of cGMP, producing more activity than PKG from either unstimulated or cGMP-treated cells. These results demonstrated that PKG activation by PMA involved a cGMP-independent modification of PKG1␣, since this stable activation did not occur when the cells were stimulated with cGMP.
Threonine 58 of PKG1␣ Is a Target for PKC Phosphorylation-The apparent independence of cGMP and the relatively stable activation of PKG1␣ in response to PMA treatment of cells suggested that PKG might be a substrate for direct phosphorylation by PKC. To test this notion, FLAG-PKG-expressing HEK-293 cells were loaded with [ 32 P]PP i and stimulated with PMA. The PKG in these cells was subsequently immunoprecipitated, and the phosphorylation state was determined by SDS-PAGE and autoradiography. Compared with untreated cells, incubation of cells with PMA caused a dramatic increase in phosphate incorporation into PKG (Fig. 4A). This phosphorylation of PKG in vivo was not due to autophosphorylation, since similar treatment with 100 M 8-Br-cGMP for 1 h had no effect on PKG phosphorylation. In support of a role for PKC in the phosphorylation of PKG, both PKC inhibitors that were able to block the PMA-stimulated shift in VASP mobility (Gö-6983 and Rö-32-0432), were also effective at blocking PMAstimulated phosphorylation of PKG. Confirmation that PKC could directly phosphorylate PKG was obtained using purified components in kinase assays in vitro (Fig. 4B). In these experiments, incubation of PKG with purified PKC␣ resulted in a dramatic increase in phosphate incorporation, which contrasted with poor incorporation of phosphate in reactions containing cGMP but lacking PKC.
The heterologous phosphorylation of PKG1␣ by PMA-responsive isoforms of PKC is a new phenomenon that prompted further investigation of the target residues. Amino acid sequence analysis of PKG1␣ revealed 11 potential PKC recognition sites (data not shown). One of these sites localized to the inhibitory loop of PKG1␣ (Thr 58 ), adjacent to the pseudosubstrate domain (Fig. 5A). Notably, Thr 58 is the first of tandem threonine residues, the latter of which had previously been reported to be a site for autophosphorylation in PKG1␣ (23, 24, 38). However, in these reports, the numbering system used to identify amino acid residues omitted the amino-terminal methionine such that the studies clearly highlighted Thr 59 as the phosphoacceptor residue, whereas the putative PKC phosphorylation site found here was Thr 58 . In order to examine the phosphorylation of Thr 58 , polyclonal antibodies were produced in rabbits using a peptide antigen that contained phosphorylated Thr 58 . These antibodies demonstrated specificity using the peptide antigen in an enzyme-linked immunosorbent assay but were ineffective when used to probe Western blots of cell lysates from cells treated with PMA (data not shown). Using larger quantities of purified PKG1␣ (Calbiochem), these antibodies were able to specifically identify PKG1␣ that had been phosphorylated by PKC in vitro but did not recognize similar amounts of PKG that was either untreated or incubated in kinase reaction buffer containing cGMP (Fig. 5B). In these experiments, the anti-phospho-Thr 58 antibodies also recognized a faint band corresponding to PKC␣, presumably the autophosphorylated form of this enzyme.
In order to examine the phosphorylation of Thr 58 in vivo, immunoprecipitates from HEK-293 cells transfected to express FLAG-PKG1␣ were used, since the previous experiments indicated a requirement for larger quantities of PKG and the slight cross-reactivity of the anti-phospho-Thr 58 antibodies to autophosphorylated PKC. In these experiments, phosphorylation of Thr 58 was detected between 5 and 15 min following stimulation of the cells with PMA and peaked at 1 h (Fig. 6A). In support of the idea that phosphorylation of Thr 58 in PKG1␣ might contribute to enzyme activation, Thr 58 phosphorylation detected on Western blots slightly preceded the shift in electrophoretic mobility of VASP in parallel experiments. The phosphorylation of Thr 58 by PKC and not by autophosphorylation was further confirmed by the observation that the regulatory region (G1␣R) containing Thr 58 in the absence of catalytic domains could become phosphorylated in response to PMA stimulation of the cells (Fig. 6B).
Phosphorylation of Threonine 58 Is Essential but Not Sufficient for PKG1␣ Activation by PKC-The role of Thr 58 in activation of PKG1␣ by PKC was further addressed by creating phospho-null (T58A) and phospho-mimetic (T58E) mutations in PKG1␣, which were immunopurified from transfected HEK-293 cells (Fig. 7A). In support of the T58E mutation to mimic phosphorylation at this residue, the purified PKG1␣(T58E) mutant was strongly labeled by the anti-phospho-Thr 58 antibodies on Western blots. The activation state of the purified PKG enzymes was assessed with in vitro kinase assays using BPDEtide substrate (Fig. 7B). In these experiments, the T58A mutant had similar enzymatic properties to the wild type enzyme. In contrast, PKG1␣ containing the T58E mutation was partially activated and exhibited a higher level of phosphotransferase activity compared with wild type enzyme both in the presence (20%) and absence (ϳ2-fold) of cGMP.
To examine whether PKG1␣(T58E) also exhibited activation in vivo, this mutant was coexpressed with VASP in HeLa cells. These experiments revealed that the high basal activity of the purified T58E mutant was not detectable in vivo using VASP mobility shift as a readout (Fig. 8). The lack of basal activity of the T58E mutant in vivo was further confirmed using the more sensitive anti-phospho-VASP239 antibodies (data not shown). The previous work describing autophosphorylation of the adjacent residue in PKG1␣ reported that modification of this residue resulted in an increased sensitivity to cGMP compared with unphosphorylated PKG (25,26). In our experiments, when compared with cells expressing wild type PKG1␣, the T58E mutant was found to be much more sensitive to activation by 8-Br-cGMP. In cells expressing PKG1␣(T58E), marked VASP phosphorylation was observed at concentrations of 8-Br-cGMP as low as 1 M (2-fold increase over basal). In contrast, these concentrations of 8-Br-cGMP had a minimal effect on VASP mobility in cells expressing similar levels of wild type PKG. Moreover, the activation state of PKG1␣(T58E) measured by VASP mobility shift in vivo was higher than wild type PKG1␣ at all concentrations of 8-Br-cGMP tested.
Whereas the activity of the T58E mutant was designed to mimic phosphorylation of PKG by PKC, it was of interest to examine the enzymatic properties of the T58A mutant, which is phospho-null with respect to this residue. To address this issue, HEK-293 cells were transfected to express PKG1␣(T58A) and stimulated with PMA for 1 h. Wild type PKG1␣ derived from these cells revealed greater than 2-fold increase in basal activity (in the absence of cGMP) following treatment of the cells with PMA (Fig. 9A). In contrast, the PKG1␣(T58A) mutant had similar basal activity in both untreated cells and those that had been treated with PMA. Moreover, this level of activity exhibited by the PKG1␣(T58A) mutant was identical to wild type PKG derived from unstimulated cells. Similar results were found when immunopurified wild type and PKG1␣ (T58A) were phosphorylated in vitro using purified PKC (Fig. 9B). However, in these studies, the effect of phosphorylation was more pronounced. Wild type PKG, which had been phosphorylated by PKC, was 3-fold more active than nonphosphorylated enzyme. The basal activity of PKG1␣(T58A) following PKC treatment also increased but to a much lesser extent than wild type PKG (ϳ60%). DISCUSSION The present study demonstrates the phosphorylation of PKG1␣ by PMA-responsive isoforms of PKC. Moreover, phosphorylation of PKG1␣ by PKC leads to PKG activation in the absence of cGMP. This is the first report of a cGMP-independent mechanism for PKG activation and highlights a novel signal transduction pathway with widespread implications for a variety of systems.
Phosphorylation and activation of PKG was demonstrated in vitro using purified components, which indicates that this phosphorylation is direct and is cGMP-independent. Using VASP phosphorylation as an indicator, this phenomenon was also shown to occur in vivo in response to PMA stimulation of HEK-293 cells with endogenous levels of PKG and PKC and also using HeLa cells transfected to express modest levels of PKG. PKG1␣ has been shown to autophosphorylate at an extremely slow rate (38). However, the results shown here are unlikely to be due to autophosphorylation, since, unlike PKG1␤, this phenomenon does not lead to activation of PKG1␣ (27). This notion is supported by experiments presented here in which treatment of cells with 8-Br-cGMP did not lead to phosphorylation in vitro or in vivo and did not lead to stable activation in the time frame sufficient for effective activation by PMA/PKC.
To further characterize the mechanism by which PKC phosphorylation can lead to the activation of PKG1␣, we have identified Thr 58 as a PKC target in vivo and in vitro. Mutation of this site to a phospho-null (T58A) indicated an important role for this residue in activation of PKG by PKC, since this mutant was only partially activated by PKC in vitro and not at all in vivo. PKG1␣ harboring a phospho-mimetic T58E mutation was much more sensitive to cGMP stimulation in vivo and was partially active when purified. Similar to work shown here for PKC-dependent phosphorylation of Thr 58 , previous work has demonstrated that autophosphorylation of Thr 59 creates an enzyme that is sensitive to lower levels of cGMP but does not lead to activation of the enzyme (26). The phosphorylation of residues in other Ser/Thr kinases creates an alteration in the shape of the inhibitory loop such that the pseudosubstrate region no longer fits into the catalytic groove. Thus, it is possible that phosphorylation of PKG at Thr 58 or Thr 59 prohibits interaction of the inhibitory domain with the catalytic cleft, thereby destabilizing the inactive form of the enzyme. This hypothesis is supported by experiments shown here in which the T58E mutant was partially active when purified from cells and required lower cGMP levels for activation in vivo.
It has been shown that there are other contact sites between the regulatory and catalytic domains of PKG that stabilize the inactive conformation in addition to the interaction between the catalytic cleft and the pseudosubstrate domain (28,39,40). These sites remain obscure but have been implicated as important to autoinhibition. Because PMA treatment of cells was able to activate PKG, it is plausible that PKC can phosphorylate more than one residue of PKG1␣, and sites other than Thr 58 may therefore destabilize additional contact sites, leading to enzyme activation. However, as shown here, neither phosphorylation site alone is sufficient for activation. That other residues on the PKG1␣ enzyme are targets for PKC phosphorylation in vitro was verified by in vitro kinase assays showing that the T58A mutant is still phosphorylated by PKC (data not shown). The nature of these other site(s) will be potentially useful in generating a mutant that cannot respond to PKC as a useful tool to investigate the physiological responses of PKC mediated by PKG.
Although PMA was used to stimulate PKC in this report, it is unclear which of the PMA-responsive PKC isoforms might be involved in phosphorylation of PKG in vivo. In the present study, it was found that both PKC inhibitors Gö-6983 and Rö-32-0432 were able to block both the PMA-stimulated VASP mobility shift and the phosphorylation of PKG in vivo. These inhibitors exhibit isoform specificity but have in common the ability to block only PKC␣ and PKC␥. Because PKC␣ expression is more widespread, it is likely that this isoform was responsible for observations in HEK-293 and HeLa cells shown here. Since the substrate specificity of different PKC isoforms is relatively conserved, it is also a possibility that PMA-insensitive PKC isoforms might also function in this capacity in cells (41). Identification of the specific PKC isoforms that phosphorylate PKG in vivo should provide some insight into which ligands might stimulate this pathway and what physiological processes might be affected. PKC substrates are often determined by subcellular colocalization with PKC, and many PKC isoforms move to the plasma membrane upon activation (42). It has been reported recently that a relatively small subset of cellular PKG localizes to the plasma membrane (36). If this membrane-bound PKG were specifically targeted by PKC, this could provide both a mechanism of activation of PKG by PKC and also serve to explain the lack of sensitivity of the phospho-Thr 58 antibody on Western blots containing whole cell lysates as observed here.
In summary, we describe here for the first time a mechanism by which PKG can be activated in the absence of cGMP. This phenomenon provides a novel link between PKC-and PKGmediated signal transduction pathways. Future investigations are warranted to determine which upstream pathways and PKC isoforms mediate this process, but certainly this pathway has widespread implications for the regulation of diverse cellular processes.