Cyclic GMP-dependent Protein Kinase Iα Inhibits Thrombin Receptor-mediated Calcium Mobilization in Vascular Smooth Muscle Cells*

Vascular smooth muscle contractile state is regulated by intracellular calcium levels. Nitric oxide causes vascular relaxation by stimulating production of cyclic GMP, which activates type I cGMP-dependent protein kinase (PKGI) in vascular smooth muscle cells (VSMC), inhibiting agonist-induced intracellular Ca2+ mobilization ([Ca2+]i). The relative roles of the two PKGI isozymes, PKGIα and PKGIβ, in cyclic GMP-mediated inhibition of [Ca2+]i in VSMCs are unclear. Here we have investigated the ability of PKGI isoforms to inhibit [Ca2+]i in response to VSMC activation. Stable Chinese hamster ovary cell lines expressing PKGIα or PKGIβ were created, and the ability of PKGI isoforms to inhibit [Ca2+]i in response to thrombin receptor stimulation was examined. In Chinese hamster ovary cells stably expressing PKGIα or PKGIβ, 8-Br-cGMP activation suppressed [Ca2+]i by thrombin receptor activation peptide (TRAP) by 98 ± 1 versus 42 ± 5%, respectively (p <0.002). Immunoblotting studies of cultured human VSMC cells from multiple sites using PKGIα- and PKGIβ-specific antibodies showed PKGIα is the predominant VSMC PKGI isoform. [Ca2+]i following thrombin receptor stimulation was examined in the absence or presence of cyclic GMP in human coronary VSMC cells (Co403). 8-Br-cGMP significantly inhibited TRAP-induced [Ca2+]i in Co403, causing a 4-fold increase in the EC50 for [Ca2+]i. In the absence of 8-Br-cGMP, suppression of PKGIα levels by RNA interference (RNAi) led to a significantly greater TRAP-stimulated rise in [Ca2+]i as compared with control RNAi-treated Co403 cells. In the presence of 8-Br-cGMP, the suppression of PKGIα expression by RNAi led to the complete loss of cGMP-mediated inhibition of [Ca2+]i. Adenoviral overexpression of PKGIβ in Co403 cells was unable to alter TRAP-stimulated Ca2+ mobilization either before or after suppression of PKGIα expression by RNAi. These results support that PKGIα is the principal cGMP-dependent protein kinase isoform mediating inhibition of VSMC activation by the nitric oxide/cyclic GMP pathway.

The cyclic GMP-dependent protein kinase type I (PKGI) 2 mediates vascular smooth muscle cell (VSMC) relaxation via the nitric oxide/ cyclic GMP pathway. One of the PKGI-dependent mechanisms of VSMC relaxation involves the inhibition of Ca 2ϩ mobilization (reviewed in Refs. 1,2). Smooth muscle contraction begins upon receptor-mediated generation of inositol 1,4,5-triphosphate (IP 3 ), which releases intracellular stores of Ca 2ϩ from the sarcoplasmic reticulum and is followed by an influx of extracellular Ca 2ϩ via voltage-gated Ca 2ϩ channels (1,3,4). A rise in intracellular Ca 2ϩ activates the Ca 2ϩ /calmodulin-dependent myosin light chain kinase, which phosphorylates the myosin light chain, activating the myosin ATPase and actomyosin cross-bridge cycling, leading to an increase in tension (1). The ability of PKGI to oppose agonist-mediated Ca 2ϩ mobilization is well established (5)(6)(7)(8)(9), but the relative roles of the PKGI isoforms PKGI␣ and PKGI␤ are poorly understood.
Many signaling events involved in Ca 2ϩ mobilization are regulated by PKGI. The phosphorylation of the thromboxane receptor by PKGI desensitizes signaling by this receptor in a manner analogous to G-protein-coupled receptor kinases (10,11). Some of these studies were done in human platelets, which express predominantly PKGI␤ (12), suggesting that PKGI␤ can phosphorylate and attenuate signaling by the thromboxane receptor (11). The phosphorylation and activation of the regulator of G-protein signaling 2 (RGS2) by PKGI␣ terminates thrombin receptor signaling by increasing the GTPase activity of G␣q (13). One report suggests PKGI can phosphorylate and inhibit the activation of PLC␤3 (14). Voltage-gated Ca 2ϩ channels regulate Ca 2ϩ entry, and a primary regulator of membrane potential in VSMCs is the large conductance Ca 2ϩ -activated K ϩ channel (BKCa 2ϩ ) (15). The activation of BKCa 2ϩ by Ca 2ϩ sparks (reviewed in Ref. 16) or PKGI (17) hyperpolarizes the cell, leading to decreased Ca 2ϩ entry and cellular relaxation. Ca 2ϩ efflux from IP 3 -sensitive intracellular stores also is inhibited by PKGI␤-mediated phosphorylation of the IP 3 receptor-associated cyclic GMP kinase substrate (IRAG) in a complex of sarcoplasmic reticulum membrane proteins, including PKGI␤, IRAG, and the IP 3 receptor (18 -21). The presence of IRAG is essential in the cyclic GMP-dependent inhibition of Ca 2ϩ signaling in colonic SMCs, suggesting that PKGI␤ is a principal regulator of intracellular Ca 2ϩ levels in these cells (19). However, in VSMCs derived from PKGI knock-out mice blood vessels, the transfection of PKGI␣, but not PKGI␤, decreases noradrenaline-induced Ca 2ϩ mobilization (9).
Our laboratory has shown that PKGI␣ binds to and phosphorylates RGS2 to terminate PAR-1 thrombin receptor signaling (13), and others have shown that the stable expression of PKGI␣ in CHO cells inhibits thrombin-mediated Ca 2ϩ mobilization in the presence of 8-Br-cGMP (22). However, the role of PKGI␤ was not explored in either of these studies. The relative roles of the two PKGI isozymes, PKGI␣ and PKGI␤, in cyclic GMP-mediated inhibition of Ca 2ϩ transients in VSMCs therefore remain unresolved. In this study, we investigated the relative abilities of PKGI isoforms to inhibit a rise in Ca 2ϩ in response to thrombin receptor-activating peptide (TRAP) in cells that lack PKGI (CHO cells) and in human coronary artery VSMCs that express predominantly PKGI␣ (Co403 cells). We provide evidence in these studies that PKGI␣ is the predominant PKGI isoform mediating inhibition of VSMC activation.

EXPERIMENTAL PROCEDURES
Materials-Oligonucleotides and the thrombin receptor-activating peptide SFLLRN were synthesized and purified by the Tufts University Core Facility (Boston, MA). Lysophosphatidic acid and platelet-derived growth factor-BB were from Sigma. The thromboxane analog U46619 was from Calbiochem. The rabbit polyclonal anti-PKGI antiserum was from Stressgen (Victoria, B.C, Canada), and the goat polyclonal anti-PKGI␣ antibody was from Santa Cruz Biotechnology (Santa Cruz, CA). Purified PKGI␣ was from Promega (Madison, WI). The LacZ adenovirus (23) and PKGI␣ and FLAG-PKGI␤ adenoviruses (24) were kind gifts from B. Berk (University of Rochester) and K. Bloch (Harvard Medical School), respectively. The Re-Blot Plus Western blot recycling kit was from Chemicon International (Temecula, CA).
Cell Culture-Chinese hamster ovary cells (CHO-K1) were purchased from the American Type Culture Collection and propagated in F12 medium containing 10% fetal bovine serum, penicillin (100 units/ ml), and streptomycin (100 g/ml) (Invitrogen). Immortalized human coronary artery smooth muscle cells (Co403) were developed in our laboratory by the explant method and characterized by both morphology and by immunohistochemical studies of the expression of smooth muscle cell-specific ␣-actin. The cells were immortalized by retroviral constructs containing the E6 and E7 human papillomavirus proteins as we have reported previously (25). These cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, penicillin (100 units/ml), and streptomycin (100 g/ml) (Invitrogen). CHO cells and Co403 used in this study were from passages 8 -12 and 9 -14, respectively.
Antibodies-Isoform-specific antibodies against PKGI␣ and PKGI␤ were raised in rabbits using glutathione S-transferase-fused antigens containing the PKGI␣-specific amino acids 1-59 or the PKGI␤-specific residues 1-90. Antibodies were immunoaffinity purified from antisera. Briefly, PKGI␣ 2-58, PKGI␤ 2-33, and PKGI␤ 43-74 peptides were synthesized and high performance liquid chromatography-purified by the Tufts University Core Facility and covalently coupled to CNBractivated Sepharose (Amersham Biosciences). The antigen-containing columns were washed with 10 mM Tris, pH 7.5, 100 mM glycine, pH 2.5, and 10 mM Tris, pH 8.8. Whole antisera were diluted 1:10 in 10 mM Tris, pH 7.5, and applied to the appropriate column three times. The columns were washed with 20 bed volumes of 10 mM Tris, pH 7.5, and then with 20 bed volumes of 500 mM NaCl, 10 mM Tris, pH 7.5. Antibodies were eluted with 10 bed volumes of 100 mM glycine, pH 2.5, and fractions were collected in tubes containing 1 bed volume of 1 M Tris, pH 8.0. Fractions containing antibody were determined by UV spectrophotometry and SDS-PAGE followed by Coomassie staining and were pooled and applied to a PD-10 desalting column (Amersham Biosciences) to elute the antibodies into phosphate-buffered saline. The resulting antibodies were concentrated to ϳ1 mg/ml using a 30-kDa centrifugal concentrator (Millipore; Billerica, MA) and preserved in 0.02% sodium azide. A titration of purified PKGI␣ and lysate from CHO cells transfected with PKGI␤ was run on a gel to confirm isoform-specific recognition by immunoaffinity-purified antibodies by immunoblot. The membranes were stripped and reprobed with a common PKGI antibody to determine the relative sensitivities of the isoform-specific antibodies.
PKGI␣ and PKGI␤ Constructs-DNA encoding PKGI␣ and PKGI␤ were amplified by PCR from a human aortic cDNA library (Clontech). The 5Ј-primer for PKGI␣ was 5Ј-GGAATTCCATGAGCGAGCTAG-AGGAAGACTTTGCC-3Ј, and the 5Ј-primer for PKGI␤ was 5Ј-CGA-GAATTCCATGGGCACCTTGCGGGATTTACAGTACG-3Ј. The 3Ј-primer used to amplify both isoforms was 5Ј-CGACTCTAGAGCT-TAGAAGTCTATATCCCATCCTGAGTTGTC-3Ј. The PCR products were cleaved by EcoRI and XbaI and cloned into a pcDNA3.1 vector (Invitrogen) containing a neomycin resistance gene that allows for the selection of stable transfectants in mammalian cells. The PKGI␣ and PKGI␤ inserts were confirmed by sequencing.
CHO Stable Cell Lines-The backbone vectors of the PKGI␣ and PKGI␤ constructs were linearized by cleavage with ScaI and were used to transfect subconfluent CHO cells using Polyfect (Qiagen; Valencia, CA). Stably transfected cell clones were selected in F12 medium containing 500 g/ml of geneticin and supplemented with 10% fetal bovine serum. Two different cell clones expressing equivalent levels of PKGI␣ (CHO-PKGI␣) and PKGI␤ (CHO-PKGI␤) were chosen for experiments. Control cells (CHO-WT) were grown in F12 medium supplemented with 10% fetal bovine serum.
Measurements of Intracellular Ca 2ϩ -Cells were seeded into 4-well LabTek (Nalge Nunc; Rochester, NY) chambered cover glass slides or on coverslips to be mounted into a chamber and grown for 1 day until they reached a density of 50 -70% confluency. CHO and Co403 cells were serum deprived for 32 and 16 h, respectively. Cells were loaded for 30 min with 2.5 M fura-2-AM (Molecular Probes; Carlsbad, CA) premixed with 0.02% Pluronic F127 (Molecular Probes) in a modified Ringer's buffer (140 mM NaCl, 6.6 mM KCl, 10 mM glucose, 2 mM CaCl 2 , 1.8 mM MgSO 4 , and 5 mM HEPES, pH 7.4), washed with phosphate-buffered saline, and treated with modified Ringer's buffer in the presence or absence of 1 or 5 mM 8-Br-cGMP (Sigma) for 20 -40 min. The chamber slides were placed in the light path of a Nikon Eclipse TE2000-U microscope equipped with a xenon light source. Recordings were made with an excitation wavelength of 340 or 380 nm and an emission wavelength of 500 nm; changes in the 340/380 ratio were assessed in response to contractile agonist treatment as reported previously (26,27). Vasoconstrictors were used at a concentration corresponding to the EC 50 unless otherwise indicated.
Cells were seeded in 6-well plates at a density of 1 ϫ 10 5 cells/well and transfected 2 days later with a final total concentration of 100 nM double-stranded RNA using Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol as reported previously (28). Forty-eight hours later, cells were either infected with an adenovirus construct at multiplicity of infection of 500 (LacZ) or 300 (PKGI␣ and FLAG-PKGI␤) or washed and placed in fresh medium. The cells were split into a 12-well plate for immunoblotting and LabTek chamber slides for Ca 2ϩ imaging 72 h after siRNA treatment. After the cells adhered to the plate and chamber slides, they were arrested in serum-free Dulbecco's modified Eagle's medium, and assays were conducted the next day. To determine silencing efficiency, cells were lysed in lysis buffer (250 mM NaCl, 1% Triton X-100, 10% glycerol, 25 mM ␤-glycerol phosphate, 20 mM Tris, pH 7.5), and the protein concentration in the soluble fraction was determined by Bradford assay (Bio-Rad; Hercules, CA). 20 g of each lysate was applied to a 7.5% gel, subjected to SDS-PAGE, and transferred to nitrocellulose. Membranes were immunoblotted using the indicated antibodies.

Effect of PKGI␣ and PKGI␤ on TRAP-induced Ca 2ϩ Transients in
CHO Cells-CHO cell lines stably expressing PKGI␣ or PKGI␤ were created and used to study the effect of cyclic GMP on intracellular Ca 2ϩ mobilization in response to thrombin receptor activation. The CHO cell lines expressed comparable levels of the two PKGI isoforms, PKGI␣ and PKGI␤ (764 and 810 pg/g total protein, respectively) (Fig. 1A). TRAP, a hexapeptide agonist of the PAR-1 thrombin receptor, was used to activate CHO-PKGI␣ and CHO-PKGI␤ cells in the absence or presence of 8-Br-cGMP (Fig. 1). In CHO-WT cells, TRAP induced a rapid Ca 2ϩ transient that returned to resting levels within 2 min of treatment; pretreatment of CHO-WT cells with 8-Br-cGMP had no effect on TRAPinduced Ca 2ϩ mobilization (Fig. 1B). In both CHO-PKGI␣ and CHO-PKGI␤ cells, TRAP elicited a Ca 2ϩ transient similar in magnitude and duration to that of CHO-WT cells (Fig. 1B). Pretreatment of the cells with 8-Br-cGMP completely abrogated the TRAP-induced Ca 2ϩ transient in CHO-PKGI␣ cells but decreased the magnitude of the Ca 2ϩ response in CHO-PKGI␤ cells only partially (Fig. 1B). The TRAP-induced increase in intracellular Ca 2ϩ for all the cell lines tested was not statistically different (Fig. 1C). In three experiments, 8-Br-cGMP inhibited the TRAP response in CHO-PKGI␣ and CHO-PKGI␤ cells by 98 Ϯ 1% (p Ͻ0.001) and 42 Ϯ 5% (p Ͻ0.002), respectively (Fig. 1C).

Effect of PKGI␣ and PKGI␤ on TRAP-induced Ca 2ϩ Transients in
Co403 Cells-Fourteen vascular smooth muscle cell lines from human aorta (two lines), internal mammary (two lines), carotid (one line), iliac (two lines), coronary (three lines), radial (two lines), and pulmonary (one line) arteries were screened for their levels of expression of PKGI␣ and PKGI␤ using purified, isoform-specific antibodies. In thirteen of the human VSMC lines examined, PKGI␣ was the predominant isoform detected (densitometric level from 55 to 96%). In the three human coronary artery SMC lines tested, PKGI␣ accounted for Ͼ95% total PKGI in two of the lines and 65% of the total PKGI in the third. In one cell line (carotid artery), more PKGI␤ than PKGI␣ was detected (35% PKGI␣). Because the expression of PKGI decreases in VSMCs after passaging (29), prior to initiating calcium studies candidate cell lines also were screened for the stability of PKGI isoform expression after passaging. Carotid artery SMCs had variable PKGI expression, resistance to siRNA treatment, and inconsistent Ca 2ϩ responses to vasoconstrictors, making them a poor model in which to test the relative roles of the PKGI isoforms in Ca 2ϩ handling. Three other cell lines (aortic, iliac, and radial artery) presented similar issues. Co403 human coronary artery smooth muscle cells were chosen for further study because they are physiologically relevant, expressed stable levels of PKGI in all passages tested, expressed predominantly PKGI␣ (densitometric levels, 96% PKGI␣, 4% PKGI␤; Fig. 2), were sensitive to siRNA treatment, and respond well to vasoconstrictors in single-cell Ca 2ϩ assays.
Single Co403 cells were screened for a Ca 2ϩ response to TRAP, thromboxane analog (U46619), lysophosphatidic acid, and platelet-derived growth factor in the presence and absence of 8-Br-cGMP pretreatment. All of the G-protein-coupled receptor agonists tested (TRAP, lysophosphatidic acid, and U4) elicited Ca 2ϩ transients that were inhibited by 8-Br-cGMP (46 Ϯ 18, 62 Ϯ 2, and 70 Ϯ 16% inhibition, respec-tively; n ϭ 3; p Ͻ0.01 versus control; Fig. 3A). The receptor tyrosine kinase agonist, platelet-derived growth factor-BB, induced a Ca 2ϩ transient that was not affected by 8-Br-cGMP (Fig. 3A). The response of Co403 cells to TRAP was studied further because these cells respond robustly to TRAP and the role of PKGI␣, but not PKGI␤, has been tested in thrombin receptor activation (13,22). TRAP induced a rapid rise in Ca 2ϩ that gradually returned to resting levels after about 2 min (Fig. 3B). Pretreatment with the PKGI activator 8-Br-cGMP consistently inhibited the magnitude of Ca 2ϩ response at 3 M TRAP (25 Ϯ 6% inhibition, n ϭ 8; p Ͻ0.02 versus control; Fig. 3B). The inhibitory effect of 8-Br-cGMP on TRAP-induced Ca 2ϩ mobilization at different concentrations of TRAP was examined. The EC 50 for TRAP-induced Ca 2ϩ mobilization in the absence of 8-Br-cGMP was 2 M. Pretreatment of Co403 cells with 8-Br-cCMP led to a right shift in the dose-response curve for TRAP activation and a 4-fold increase in the EC 50 for TRAP-induced Ca 2ϩ mobilization to 8 M (Fig. 3C).

Effect of Suppression of PKGI␣ Expression on TRAP-induced Ca 2ϩ Transients in Co403
Cells-PKGI␣ expression in Co403 cells was reduced using siRNA directed against PKGI␣, which in all studies shown suppressed the expression of PKGI␣ by Ͼ85% (Fig. 4A). The effect of PKGI␣ suppression in Co403 cells on TRAP-mediated Ca 2ϩ transients was studied next in the presence and absence of 8-Br-cGMP. In the absence of 8-Br-cGMP, the magnitude of the TRAP-induced Ca 2ϩ transients in Co403 cells was not significantly different between mock treated cells and cells treated with control siRNA (Fig. 4B). The total Ca 2ϩ response following decreased PKGI␣ expression by PKGI␣ siRNA treatment was significantly greater than controls (by 164 Ϯ 19%, n ϭ 3, p Ͻ0.05; Fig. 4C, open bars). The TRAP-induced Ca 2ϩ responses were studied next in the presence of 8-Br-cGMP in mock, control siRNA, and PKGI␣ siRNA-treated cells (Fig. 4). In the presence of 8-Br- cGMP, the total Ca 2ϩ response in PKGI␣ siRNA-treated cells was 215 Ϯ 24% greater than controls (n ϭ 3, p Ͻ0.02; Fig. 4C) and reached a level similar to the level of control cells treated with TRAP in the absence of a cyclic GMP donor (compare Fig. 4). In mock and siRNA control treated cells, the duration of the Ca 2ϩ response was reduced similarly with 8-Br-cGMP pretreatment by ϳ25% (Fig. 4B). By contrast, there was no effect on the duration of TRAP-induced Ca 2ϩ transient in the presence of 8-Br-cGMP in human coronary artery SMCs treated with siRNA against PKGI␣ (Fig. 4B). 8-Br-cGMP inhibited TRAP-induced Ca 2ϩ release in mock, control siRNA, and PKGI␣ siRNA-treated cells by 53 Ϯ 7, 49 Ϯ 13, and 33 Ϯ 8%, respectively (PKGI␣ siRNA versus mock control, p Ͻ0.01, n ϭ 3-6; Fig. 4C). These data show that a decrease in the level of expression of PKGI␣ in human coronary artery SMCs affects TRAP-mediated Ca 2ϩ stimulation by 1) increasing the absolute magnitude of the TRAP response in the absence of 8-Br-cGMP, and 2) reducing the inhibitory effect of 8-Br-GMP on TRAPstimulated cells.
Effect of PKGI␣ or PKGI␤ Overexpression on TRAP-induced Ca 2ϩ Transients in Co403 Cells-Co403 cells express predominantly PKGI␣ (Ͼ95%; Fig. 2). To test further the effect of PKGI isoform expression levels on TRAP-induced Ca 2ϩ mobilization, Co403 cells were infected with adenoviral constructs expressing LacZ (control), PKGI␣, or PKGI␤ (23,24). Western blot analyses confirmed increased expression of each PKGI isoform (Fig. 5A). TRAP-induced Ca 2ϩ transients in Co403 cells overexpressing PKGI␣ or PKGI␤ were studied next. Overexpression of either of the PKGI isoforms in Co403 cells did not significantly alter the magnitude of TRAP-induced Ca 2ϩ transients as compared with controls (Fig. 5B). Pretreatment of Co403 cells overexpressing either PKGI␣ or PKGI␤ with 8-Br-cGMP significantly reduced the total Ca 2ϩ mobilization by TRAP in comparison with control adenovirus-infected cells (by 52 Ϯ 13%, n ϭ 4, p Ͻ0.05 and 57 Ϯ 12%, n ϭ 4, p Ͻ0.03, respectively) (Fig. 5C). The ability of PKGI␤ to inhibit TRAP signaling in VSMCs was examined by simultaneously suppressing the expression of PKGI␣ with siRNA and overexpressing PKGI␤ with an adenoviral construct (Fig. 6). This approach resulted in a final ratio of PKGI␤ to PKGI␣ protein of about 10:1 in comparison with native Co403 cells (Fig. 6A). As in the experiments above, siRNA reduction of PKGI␣ led to an increase in the total Ca 2ϩ mobilization by TRAP (by 140 Ϯ 4%, n ϭ 5, p Ͻ0.003; Fig.  6C), and this response was not altered by the overexpression of PKGI␤ (n ϭ 7; Fig. 6C). siRNA reduction of PKGI␣ again led to a significant decrease in the ability of cyclic GMP to inhibit the TRAP-stimulated rise in Ca 2ϩ (siRNA control versus PKGI␣ siRNA ϩ Ad control, p Ͻ0.003, n ϭ 5), and this also was unaffected by overexpression of PKGI␤ (n ϭ 7; Fig. 6C). The infection of Co403 cells treated with siRNA against PKGI␣ failed to reestablish the expression of PKGI␣. In conclusion, overexpression of PKGI␤ did not alter cGMP-mediated inhibition of Ca 2ϩ mobilization in Co403 cells.

DISCUSSION
Our data support a model in human coronary artery VSMC and in CHO cells in which PKGI␣ plays the predominant role in mediating the inhibitory effects of cyclic GMP on intracellular Ca 2ϩ mobilization in response to thrombin receptor activation. Ca 2ϩ mobilization is determined by a balance between intracellular signaling events that release Ca 2ϩ into and sequester Ca 2ϩ from the cytoplasm. The balance between these processes may differ in CHO and Co403 cells, because PKGI␣ activation completely inhibits Ca 2ϩ signaling in CHO cells, but only partially in native Co403 cells, and PKGI␤ alone has an inhibitory effect in CHO cells, but not in Co403 cells. It is also possible that PKGI isoforms act on distinct, cell-specific targets to oppose TRAP-induced activation in ways that differ between CHO cells and Co403 cells. Alternatively, PKGI isozymes may regulate the same targets in CHO and Co403 cells, but the inhibitory extent or the mechanism of Ca 2ϩ mobilization by TRAP may differ in these cells such that PKGI isoforms have a greater inhibitory effect in CHO cells compared with Co403 cells.
G-protein-coupled receptor activation by TRAP, U4, and lysophosphatidic acid is inhibited by 8-Br-cGMP in human coronary artery SMCs. By contrast, Ca 2ϩ mobilization by the receptor tyrosine kinase agonist, platelet-derived growth factor-BB, is not affected by 8-Br-cGMP in these cells. This suggests that 8-Br-cGMP inhibits a mechanism specific to G-protein-coupled receptor, but not receptor tyrosine kinase, activation in Co403 cells. The data also show that reduction in the level of PKGI␣ in human coronary artery SMCs leads to an increase in the agonist-stimulated Ca 2ϩ response even in the absence of cyclic GMP, supporting that PKGI␣ (but not PKGI␤) has a basal inhibitory role on signaling in these cells. A previous report has shown that basal inhibition of PKGI with the peptide inhibitor DT-2 causes pressurized cerebral arteries to constrict beyond their resting diameter in the absence of a nitric oxide or cyclic GMP donor, which is also consistent with the model that basal activity of PKGI contributes to the resting tone of blood vessels (30). Overexpression of PKGI␤ after suppression of PKGI␣ expression by RNA interference fails to rescue the Ca 2ϩ response of Co403 cells to TRAP, further supporting that the basal activation of PKGI␣, but not PKGI␤, is critical in regulating the extent of receptor-activated Ca 2ϩ mobilization. The suppression of PKGI␣ expression significantly inhibits, but does not abolish, the 8-Br-cGMP- dependent inhibition of Co403 activation by TRAP. This persistent effect of 8-Br-cGMP may be due to residual PKGI␣ expression or to the effects of other cGMP effectors, including cGMP-gated channels, cGMP-regulated phosphodiesterases, or activation of protein kinase A. However, it is unlikely that these effects are due to PKGI␤, because the overexpression of PKGI␤ in the relative absence of PKGI␣ did not significantly alter the 8-Br-cGMP-mediated inhibition of TRAP-induced Ca 2ϩ mobilization.
There are conflicting data about the roles of PKGI isoforms in mediating cyclic GMP-dependent relaxation in VSMC. There are several possible explanations for these discrepancies in the literature. VSMC from different vascular beds may express different subsets of proteins involved in PKGI-dependent vasorelaxation and Ca 2ϩ handling. The relative roles of the PKGI isozymes may also be different with different agonists and the signaling pathways they enlist. Earlier pharmacological studies support that PKGI␣ mediates the cyclic GMP-dependent relaxation of pig coronary arteries (35). In PKGI knock-out VSMCs, the transfection of PKGI␣, but not PKGI␤, decreased noradrenaline-induced Ca 2ϩ mobilization, supporting the importance of PKGI␣ in VSMC (9). PKGI␣ and PKGI␤ act on different targets and therefore may work in series to have an additive inhibitory effect. The leucine zipper domains of PKGI␣ and PKGI␤ differ and are important for subcellular targeting of both kinases (13,18,(31)(32)(33)(34). Our laboratory has shown previously that the leucine zipper domain of PKGI␣ mediates its interaction with two important proteins that regulate the contractile state of VSMCs, RGS2 and myosin phosphatase (13,32,33). The PKGI␣/RGS2 pathway limits upstream Gq signaling by increasing the GTPase activity of Gq (13). Because Co403 cells express predominantly PKGI␣, a PKGI␣-RGS2 mechanism is consistent with the observed 8-Br-cGMPdependent inhibition of G-protein-coupled receptor, but not receptor tyrosine kinase, activation and with our previous studies (13).
Others have shown that the leucine zipper domain of PKGI␤ targets it to IRAG, a protein that regulates the release of Ca 2ϩ from downstream IP 3 -sensitive stores (18,34). However, VSMCs isolated from IRAG null mice have a markedly greater noradrenaline-induced Ca 2ϩ response compared with WT mice in both the absence and the presence of a cyclic GMP donor (20), suggesting that IRAG has a role in Ca 2ϩ signaling independent of cGMP. Furthermore, the mutation of IRAG in these studies decreases the concentration of PKGI␤ and mutant IRAG in the aorta (20) and may also alter the amount of other proteins involved in Ca 2ϩ handling in these cells, which may include PKGI␣ substrates. A PKGI␣/RGS2 mechanism, by acting proximally and limiting the amplification of TRAP signaling, might be required to detect downstream effects of PKGI␤/IRAG, even in the absence of a cyclic GMP donor, in Co403 cells. The overexpression of PKGI␣ in these cells increases the amount of PKGI␣ available to act on its target(s), and this may simply elicit a similar response when PKGI␤ is overexpressed in the presence of endogenous PKGI␣.
In summary, our data show that PKGI␣ plays the predominant role in mediating the inhibitory effects of cyclic GMP on Ca 2ϩ mobilization by TRAP in human coronary artery SMCs and CHO cells. Additional studies of specific isozyme function in conditional mouse models will be necessary to further dissect the physiologic significance of the two PKGI isoforms. We are also currently testing vascular regulation in mice containing a targeted mutation in the leucine zipper amino-terminal domain of PKGI␣, and we are developing SMC-specific PKGI␣ knockout mice. These models will be important tools to further our understanding of the relative roles of PKGI isoforms in blood pressure regulation and VSMC proliferation.