Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

It is generally accepted that increases in cGMP mediate v.s.m. relaxation caused by agents such as SNP, NTG, endothelium-dependent relaxing factor, nitric oxide and ANF (see reviews by Ignarro and Kadowitz, 1985 and Walter, 1989). The suggestion that cGMP mediates v.s.m. relaxation was based on a number of studies that showed that 1) increases in cGMP levels occurred during relaxation of v.s.m. preparations by nitrovasodilators (Diamond and Blisard, 1976; Katsuki et al., 1977) and that these increases were well correlated with relaxation (Kukovetz et al., 1979), 2) nitrovasodilators could activate soluble guanylyl cyclase (Katsuki et al., 1977), 3) inhibitors of guanylyl cyclase, such as methylene blue, decreased cGMP generation by nitrovasodilators and attenuated relaxation (Gruetter et al., 1981), 4) inhibition of cGMP metabolism, by phosphodiesterase inhibitors, potentiated nitrovasodilator-induced relaxation (Kramer and Wells, 1979; Kukovetz et al., 1979) and 5) cGMP analogs such as 8-Br-cGMP could induce relaxation in precontracted v.s.m. preparations (Schultz et al., 1979; Napoli et al., 1980).

The underlying mechanisms by which increases in cGMP levels can lead to relaxation are not well understood. It has been suggested that activation of PKG plays an important role, presumably via phosphorylation of substrates involved in regulating cytoplasmic calcium levels (Lincoln and Cornwell, 1991; Lincoln et al., 1994). Strong evidence implicating PKG as the mediator of the effects of cGMP has been provided by an interesting series of experiments done in isolated aortic smooth muscle cells (Cornwell and Lincoln, 1989). In primary (nonpassaged) rat aorta cells, vasopressin-induced increases in cytosolic Ca⁺⁺ were reduced both by ANF and by 8-Br-cGMP. However, when the study was repeated using cultured cells that had been passaged many times, ANF and 8-Br-cGMP no longer exerted an inhibitory effect on vasopressin-induced increases in Ca⁺⁺ levels. It was noted that levels of PKG were considerably reduced in these passaged cells. When purified PKG was added back to the passaged cells, the Ca⁺⁺-lowering effect of ANF and 8-Br-cGMP was restored, which is consistent with the suggestion that PKG is required for the relaxant effects of cGMP in these cells. Studies using cGMP analogs have provided another line of evidence that supports a role for PKG in smooth muscle relaxation. For example, Francis et al. (1988) investigated the ability of a number of cGMP analogs to relax precontracted porcine coronary artery and guinea pig trachealis. They found that the ability of the analogs to relax these tissues (EC₅₀) correlated well with the K_a values of the analogs for activating purified PKG.

If the actions of cGMP (in terms of v.s.m. relaxation) are, in fact, mediated via activation of PKG, it should be possible to demonstrate such activation directly. To date, there have been only five reports attempting to demonstrate activation of PKG by cGMP-elevating agents in intact v.s.m. preparations (Lincoln and Fisher-Simpson, 1983; Fiscus et al., 1984, 1985; Jiang et al., 1992; Bergh et al., 1995). Only one of these (Fiscus et al., 1985) reported a good correlation between activation of PKG and relaxation, and that was for ANF, a peptide known to increase cGMP levels by activating the particulate form of guanylyl cyclase. In the other studies, which examined the effects of the nitrovasodilator SNP on PKG activity in several vascular preparations, a close correlation between PKG activation and relaxation was not always observed. For example, Bergh et al., (1995) did not find any activation of the enzyme in bovine carotid arteries that were almost completely relaxed by 10 µM SNP. Lincoln and Fisher-Simpson (1983) and Jiang et al., (1992) did find activation of the kinase by SNP in rat aorta and porcine coronary artery, respectively, but only with concentrations of SNP much higher than those required to relax the blood vessels. Finally, Fiscus et al., (1984) also found an increase in PKG activity ratios in rat aorta treated with SNP, they but looked at only one concentration of SNP at a single time-point. The effect of SNP on tension was not monitored in the latter study, and in none of the studies were biochemical parameters measured in the same muscles used for relaxation studies. Thus data directly correlating PKG activation and nitrovasodilator-induced relaxation in intact blood vessels are lacking. The objective of the present study was to provide such data. To this end, experiments were performed in which PKG activity and tension were monitored simultaneously in the same muscle strips after exposure to varying concentrations of two nitrovasodilators, SNP and NTG, for varying times. The results indicate that, at least in rabbit aortic strips, a reasonably good correlation appears to exist between the ability of SNP and NTG to activate PKG and their ability to cause relaxation.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Tissue preparation. New Zealand White rabbits of either sex (2.5-3.0 kg) were anesthetized with sodium pentobarbital (65 mg/kg i.v.) and then exsanguinated in accordance with guidelines established by the University of British Columbia Animal Care Committee. The descending thoracic aortae were excised, placed in Krebs-bicarbonate buffer of the following composition (mM): KCl (4.75), KH₂PO₄ (1.2), MgSO₄ (1.2), CaCl₂ (2.5), NaCl (118), NaHCO₃ (25), D-glucose (11.12) aerated with 95% O₂/5% CO₂ and were carefully trimmed of adhering fat and connective tissue. The aortae were then cut into helical strips ( 7 mm wide and 15-18 mm long). The endothelial layer was removed by gently rubbing a glass rod across the exposed lumen. These strips were then mounted under 2-g preload tension in 20-ml organ baths containing Krebs-bicarbonate buffer maintained at 37°C and aerated with 95% O₂/5% CO₂. Isometric tension was recorded with force-displacement transducers (Grass Model FT03C) connected to a Grass Model 7D polygraph recorder. Strips were allowed to equilibrate for 60 min before experiments were started.

Experimental protocol. After the initial equilibration period, strips were contracted by adding 1.0 µM PE (final bath concentration equivalent to EC₈₀) for 5 min, after which the tissues were washed with buffer for a further 60 min (four washes). Tension was adjusted, as required, to maintain preload tension. The strips were again contracted with 1.0 µM PE for 5 min and then washed as before. After a further 60 min (four washes in between) the strips were again contracted with 1.0 µM PE. When strips had contracted for 5 min (steady-state contraction), NTG was added to the bath in increasing concentrations (0.01 µM-10 µM for 2 min) or for various times (10 µM for various time periods of 30 sec to 10 min). Each strip was treated with only one concentration/time of NTG. A similar protocol was used for the SNP study (0.1-10.0 µM SNP for 2 min and 10 µM SNP for 30 sec to 2 min). All concentrations are final bath concentrations. The strips were then frozen at the relevant time periods by clamping with tongs precooled in liquid nitrogen and were stored for biochemical analyses (cGMP levels and PKG activity ratio). The degree of relaxation was then calculated as a percentage of steady-state contraction induced by 1.0 µM PE. At least one strip from each rabbit was treated as a control (i.e., did not receive any drug).

Estimation of cGMP. Frozen samples (15-50 mg) were placed in a 1-ml capsule that had been precooled by placing on dry ice and contained a chilled metal pestle. Samples were pulverized using a Vari-Mix III dental amalgam mixer (5 sec at medium speed). Then 0.5 ml of ice-cold TCA (6% w/v) was added to the capsule, and the pulverized tissue was homogenized by subjecting the capsule to a 15-sec medium and a 5-sec high-speed agitation. The homogenate was then removed and the capsule washed with 0.5 ml of TCA, which was added to the first homogenate (total volume 1.0 ml). The crude homogenate was then centrifuged at 2000 × g for 15 min at 4°C. The TCA was then extracted four times with 5 ml of water-saturated ether (4°C). The upper ether phase was discarded after each wash. cGMP levels were then determined in samples of the aqueous phase by using a commercially available cGMP radioimmunoassay kit (BIOTRAK, cGMP-Scintillation Proximity Assay). The samples were acetylated, as suggested by Harper and Brooker (1975), to increase the sensitivity of the assay. Tissue cGMP levels were calculated as picomoles of cGMP per gram wet weight of tissue.

PKG and PKA assay. At least 80 mg of frozen tissue was placed in a precooled (dry ice) 1-ml capsule containing a chilled metal pestle and pulverized by using a Vari-Mix III dental amalgam mixer (5 sec at medium speed). The pulverized sample was then homogenized (15 sec at medium and 5 sec at high-speed agitation) in 5 vol of ice-cold buffer of the following composition: 100 mM potassium phosphate (pH 6.8), 1.0 mM 3-isobutyl-1-methylxanthine (IBMX), 10.0 mM EDTA and 10.0 mM 2-mercaptoethanol. The homogenate was transferred to cold centrifuge tubes and then centrifuged at 12,000 × g for 15 min in a Heraeus Contifuge 28RS centrifuge (rotor 3744) at 4°C. The supernatant, containing the soluble fraction, was placed on ice and assayed immediately, in duplicate, for PKG activity as described below.

Column chromatography of PKG and PKA in rabbit aorta extracts. Both PKG and PKA were separated using a Pharmacia MonoQ anion exchange column (HR5/5) coupled to a FPLC system (Pharmacia LKB Biotech, Uppsala, Sweden). Frozen smooth muscle samples were first homogenized in buffer A (5 mM Tris-HCl, pH 7.4, 2 mM EDTA, 1 mM DTT) containing soybean trypsin inhibitor (10 µg/ml), benzamidine (1 mM), leupeptin (2 µg/ml), pepstatin (10 µg/ml) and PMSF (1 mM) using the same protocol as indicated above. The soluble extract was then diluted in homogenization buffer to yield a concentration of 1.2 mg/ml protein. At least 15 mg of protein was then loaded onto the MonoQ column, pre-equilibrated with buffer A, at a flow rate of 0.5 ml/min. The column was developed at 0.5 ml/min with a linear NaCl gradient of 0 to 400 mM in buffer A, and 40 fractions, 0.5 ml each, were collected. All procedures were done at 4°C. The fractions were then assayed as described above for PKG and PKA activity at 30°C. Fractions that showed maximal PKG and PKA activity were then immunoblotted for the presence of PKG.

Western blots. Separation of PKG was accomplished in a Bio-Rad Protean II Electrophoresis unit by SDS PAGE according to the method of Laemmli (1970). Aliquots (200 µl) of fractions from column chromatography of muscle extracts were boiled for 3 min in digestion buffer (2% w/v SDS, 120 mM Tris-HCl, pH 6.8, 10% glycerol, 5% -mercaptoethanol and 0.004% bromophenol blue). Then 50 µl each of purified bovine lung I PKG (0.1 mg/ml), porcine heart PKA (0.1 mg/ml) and PKA catalytic subunit (0.05 mg/ml) were added to 25 µl of digestion buffer and 25 µl of distilled water and boiled for 3 min. Molecular mass prestained standards were treated similarly.

Protein determination. Protein concentrations in tissue fractions and in samples applied to the MonoQ column were determined using a commercially available assay (Bio-Rad) based on the method of Bradford (1976).

Statistical analysis. Values in the drug-treated groups were compared with their respective controls using a Student's t test statistical program (SigmaStat V 1.0, Jandel Scientific, San Mateo, CA). A probability (P) of less than .05 was accepted as the level of significance. In all experiments, at least one strip from each rabbit was treated as control, and mean values were compared on a paired basis with treated strips from the same rabbit. All values are expressed as the mean ± S.E.M.

Materials. The following drugs and chemicals were used in this study: BIOTRAK cGMP scintillation proximity assay kits (Amersham International, Little Chalfont, Buckinghamshire, U.K.), BPDEtide, the PKG substrate (Bachem California, Torrance, CA), (Rp)-8-pCPT-cGMPS (Biolog Life Science Institute, Bremen, Germany),  catalytic subunit of porcine heart PKA and bovine lung I PKG holoenzyme (Biomol Research Laboratories, Inc., Plymouth Meeting, PA), materials used for SDS PAGE and Western blots (Bio-Rad Laboratories, Hercules, CA), [-³²P] adenosine 5-triphosphate (DuPont NEN Research Products, Boston, MA), prestained molecular mass standards (Kinetek Pharmaceutical Corp., Vancouver, B.C., Canada), sodium pentobarbital (MTC Pharmaceuticals, Cambridge, Ontario, Canada) and Nitrostat (Parke-Davis, Scarborough, Ontario, Canada).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

MonoQ chromatography of PKG in rabbit aorta. Several experiments were performed to verify the identity of the PKG activity measured under the standard assay conditions described in "Materials and Methods." In the first of these experiments, PKG and PKA activities in crude soluble fractions from rabbit aorta were resolved by using MonoQ anion exchange chromatography. To demonstrate cGMP-dependent activity, each collected fraction was assayed in either the absence or the presence of 5 µM cGMP, together with PKI to eliminate the phosphotransferase activity of any PKA present in the fractions. As shown in figure 1, most of the PKG activity eluted in a single, large peak around fraction 28, although a smaller peak was also seen at fraction 22. 

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Crude rabbit aorta smooth muscle soluble extracts were fractionated on a MonoQ chromatography column with a linear gradient of 0 to 400 mM NaCl (upper panel). Each fraction was assayed for PKG and PKA phosphotransferase activity at 30°C. PKG activity was measured in the presence (--) and absence (--) of 5 µM cGMP (+ 1 µM PKI). PKA activity was measured in the presence (--) and absence (--) of 5 µM cAMP in the absence of PKI. The amount of protein loaded was at least 15 mg. Results shown here represent data from a single experiment, and very similar results were obtained when it was repeated. In the lower panel, fractions 10 and 28, from MonoQ chromatography of soluble extracts of rabbit aorta that exhibited maximal PKA and PKG activity, respectively (upper panel), were resolved by SDS-PAGE for Western blotting with an antibody raised to the C-terminus of type I PKG as described in "Materials and Methods." Commercially obtained PKA holoenzyme (lane 3), PKA catalytic subunit (lane 4) and type I PKG (lane 5) were included as controls. Lane 1 was loaded with fraction 10 and lane 2 with fraction 28 from the MonoQ column. Positions of proteins of known molecular mass are indicated by their molecular mass values (kDa) at the left of the figure.

Western blotting of PKG in rabbit aorta. Because the kinase activity data from the MonoQ study indicated that PKG activity and PKA activity were present in crude soluble fractions, Western blots were performed using a polyclonal antibody raised against type I PKG. Strong immunoreactivity was found in lane 2 (fig. 1, lower panel), which contained the MonoQ fraction with maximal cGMP-dependent phosphotransferase activity. This band was observed at a molecular mass position of 75 kDa, the same position as a strong immunoreactive band in the lane containing the type I PKG standard (lane 5). This is consistent with known molecular mass values for PKG of 74 to 78 kDa (Butt et al., 1993). The antibody was specific for PKG, because there was no immunoreactivity in the lane containing the MonoQ fraction with maximal PKA activity (lane 1) or in the lanes containing the purified PKA holoenzyme (lane 3) or the PKA catalytic subunit (lane 4). These results indicate that the activity being measured under present assay conditions was due to PKG, not to some other kinase.

View larger version (30K): [in this window] [in a new window]   Fig. 2.   Crude rabbit aorta smooth muscle soluble extracts were fractionated on a MonoQ chromatography column with a linear gradient of 0 to 400 mM NaCl (upper panel). Each fraction was assayed for PKG phosphotransferase activity. PKG activity was measured in the presence (--) and absence (--) of 5 µM cGMP (+ 1 µM PKI). The amount of protein loaded was at least 10 mg. Phosphotransferase activity was measured at 30°C for 10 min as indicated in "Materials and Methods." In the lower panel, various fractions around the peak fraction (fractions 20-33) corresponding to PKG phosphotransferase activity (fig. 2, upper panel) were resolved by SDS-PAGE for Western blotting with an antibody raised to the C-terminus of type I PKG as described in "Materials and Methods." Positions of proteins of known molecular mass are indicated by their molecular mass values (kDa) at the left of the figure.

Inhibition of PKG phosphotransferase activity in vitro. In order to confirm further that the phosphotransferase activity measured in our experiments was due to PKG, we performed preliminary experiments using Rp-cGMP analogs that have been reported to be relatively selective PKG inhibitors (Butt et al., 1994). One of these analogs, (Rp)-8-pCPT-cGMPS, inhibited PKG-mediated phosphorylation of BPDEtide in crude soluble extracts of rabbit aorta over a concentration range of 0.3 to 1000 µM with an IC₅₀ of 6.8 µM (data not shown). Although the concentration required for 50% inhibition of the enzyme in our crude homogenates was higher than that required for inhibition of purified PKG (Butt et al., 1994), these results further support the conclusion that the enzyme activity measured in our assay is due to PKG.

Effect of SNP and NTG on contractility, cGMP levels and PKG activity ratio in rabbit aorta. PE-contracted rabbit aortic strips were exposed to 0.1, 1.0 and 10.0 µM SNP for 2 min. These concentrations of SNP were found to inhibit the PE response by approximately 50%, 85% and 98%, respectively, based on preliminary cumulative SNP concentration-response curves obtained in PE-contracted aorta (data not shown). Two-minute exposures to SNP (0.1, 1.0 and 10.0 µM) induced concentration-dependent relaxations of PE-contracted rabbit aorta while at the same time significantly elevating cGMP levels by approximately 3-, 14- and 57-fold, respectively (table 1). PKG activity ratios were significantly increased by all three concentrations of SNP (1.3-, 1.4- and 1.9-fold with 0.1, 1.0 and 10.0 µM SNP, respectively). These increases correlated reasonably well with the degree of relaxation seen in these same muscles 2 min after administration of SNP (18%, 58% and 71%, respectively).

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Time course for SNP-induced PKG relaxation (--), cGMP elevation (--) and PKG activation (--) in rabbit aortic strips. Strips were contracted with 1 µM PE for 5 min before the addition of 10 µM SNP. Strips were frozen at the times indicated, and cGMP levels and PKG activity ratios were determined as outlined in "Materials and Methods." Each point is the mean (± S.E.M.) of n = 5 to 10. All parameters were significantly different (P < .05) from control values (Student's t test) at all time periods.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Time course for NTG-induced PKG relaxation (--), cGMP elevation (--) and PKG activation (--) in rabbit aortic strips. Strips were contracted with 1 µM PE for 5 min before the addition of 10 µM NTG. Strips were frozen at the times indicated, and cGMP levels and PKG activity ratios were determined as outlined in "Materials and Methods." Each point is the mean (± S.E.M.) of n = 4 to 9. All parameters were significantly different (P < .05) from control values (Student's t test) at all time periods.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

As noted in the introduction, there is good direct evidence that cGMP plays a role in the relaxant effects of nitrovasodilators (see Ignarro and Kadowitz, 1985) and good indirect evidence that PKG plays a role in this process (e.g., Francis et al., 1988; Cornwell and Lincoln, 1989). However, only a few reports have attempted to correlate PKG activation directly with nitrovasodilator-induced relaxation in intact vascular preparations, and these do not provide strong evidence for such a correlation.

The small number of papers in this area probably reflects the difficulties inherent in measuring PKG activation in intact tissues. Until recently, the substrate most commonly used as a phosphate acceptor in PKG assays was histone H2B (see, for example, Lincoln and Fisher-Simpson, 1983; Fiscus et al., 1984, 1985). The H2B peptide has several disadvantages when used as a substrate for the quantitative estimation of PKG activation states in muscle samples. It has been reported to increase the binding of cGMP to PKG (Tse et al., 1981) and to increase basal PKG activity directly (Walton and Gill, 1981). It is an equally good substrate for PKA as for PKG (Colbran et al., 1992) and is also a substrate for other cyclic nucleotide-independent protein kinases (see Fiscus et al., 1984). The presence of substantial levels of cyclic nucleotide-independent protein kinase activity effectively decreases the sensitivity of the assay procedure and necessitates the use of high levels of [-³²P]-ATP in the assays. Partly because of these difficulties, the earlier studies of Fiscus et al. (1984, 1985), which demonstrated activation of PKG by cGMP-elevating agents in v.s.m., were not repeated by other laboratories until quite recently. In 1992, Jiang et al. described an improved PKG assay based largely on the availability of a better substrate, BPDEtide, which is 16-fold more selective for PKG than for PKA and is not a substrate for protein kinase C or for calcium/calmodulin kinase II (Colbran et al., 1992). Using this substrate, Jiang et al. (1992) reported that 10 µM SNP could increase PKG activity ratios in isolated pig coronary arteries. However, activation of PKG was not well correlated with relaxation in that much higher concentrations of SNP were required to activate PKG than were required to relax the arteries in parallel studies. In a more recent report, also using BPDEtide as a substrate, Bergh et al. (1995) were unable to detect activation of PKG in bovine carotid arteries after treatment with a concentration of SNP (10 µM) that could completely relax the arteries. In view of these apparent discrepancies, and because time-dependence and concentration-dependence data were lacking in most of the earlier studies, we decided to investigate the role of PKG in vascular relaxation by simultaneously monitoring the effects of NTG and SNP on cGMP levels, PKG activity ratios and contractility in the same muscle strips. We measured PKG activity as described by Jiang et al. (1992), using BPDEtide as the substrate.

As noted above, the use of BPDEtide as a substrate for the determination of PKG activity offers several advantages over other, previously available substrates in terms of the specificity and sensitivity of the assay. Verification of the specificity of the assay was provided by experiments utilizing MonoQ column chromatography and immunoblotting techniques, as described in "Results." MonoQ column chromatography of crude soluble extracts from rabbit aorta (fig. 1) provided evidence that the phosphotransferase activity measured in our experiments was, in fact, due to PKG. Using BPDEtide as the substrate, and with the PKA inhibitor PKI present in the reaction mixture, we found that neither PKA nor other cyclic nucleotide-independent protein kinases contributed significantly to the kinase activity measured. Western blots performed using fractions from the MonoQ column demonstrated that immunoreactivity to a PKG-specific antibody was restricted to the fractions containing the single, large PKG peak (fig. 1). The identity of the much smaller peak of cGMP-dependent phosphotransferase activity contained in fractions 22 and 23 is not clear. As shown in figure 2, this peak did not exhibit immunoreactivity to a polyclonal antibody raised against the C-terminus of PKG I. This result suggests that this peak does not contain PKG I or PKG I, because both isozymes have homology in the C-terminus (Ruth et al., 1991) and both isozymes should react with the antibody as has been reported by Keilbach et al. (1992). However, it is possible that this peak represents a proteolytic fragment of type I PKG that lacks the recognition site for the antibody but retains phosphotransferase activity. It is also possible that this peak contains type II PKG, which would not be recognized by the type I-specific antibody and which, because it is less electronegative than type I PKG (Francis and Corbin, 1994), would be expected to elute earlier from the MonoQ columns. In any case, this peak contributes only a small amount to the phosphotransferase activity measured in our assays.

The basal PKG activity ratios found in our experiments are considerably lower than those reported using histone H2B as a substrate (Lincoln and Fisher-Simpson, 1983; Fiscus et al., 1984, 1985) and agree very well with those obtained in pig coronary arteries using BPDEtide (Jiang et al., 1992). On the other hand, Bergh et al. (1995), also using BPDEtide, reported basal PKG activity ratios in bovine carotid arteries that are several-fold higher than those found in the present study or those reported by Jiang et al. (1992). Bergh et al. (1995) suggested that the concentration of BPDEtide used in their assay may have been too low to provide a reliable measure of the kinase activity in that tissue. This may explain their inability to detect an activation of the kinase in muscles treated with a full relaxant concentration of SNP (10 µM). The concentration of BPDEtide used in their study (150 µM) was almost an order of magnitude lower than the EC₅₀ for BPDEtide in their preparation (see fig. 5 of Bergh et al., 1995). In the present study, the same concentration of BPDEtide was used, but this concentration is approximately equal to the K_m for the peptide in crude homogenates of rabbit aorta, as determined in preliminary experiments (data not shown). In any case, the assay procedure, as used in our experiments, was sensitive enough to demonstrate activation of the enzyme in rabbit aortic strips with concentrations of SNP and NTG as low as 0.1 µM and at times as early as 10 sec after administration of 10 µM SNP (a time at which relaxation was just beginning to occur). These increases in PKG activity ratios were not due to increases in cyclic nucleotide-independent protein kinase activity; no increases in total PKG activity (measured in the presence of sufficient added cGMP to activate the enzyme completely) were found. If apparent increases in PKG activity ratios were due to activation of a cyclic nucleotide-independent protein kinase, we should have seen increases in total PKG activity as well as in basal or endogenous PKG activity (measured in the absence of added cGMP). The measured increases in each case were due to increases in endogenous PKG activity and are consistent with the increases in tissue levels of cGMP seen in these experiments.

Activation of PKG and relaxation of the arteries by SNP and NTG appeared to be correlated in both a concentration-dependent and a time-dependent manner. The only apparent dissociation seen was with the lowest concentration of NTG (0.01 µM), which relaxed the arteries but did not cause a significant increase in cGMP levels or PKG activity ratios. In the temporal studies, PKG activity ratios were significantly increased, with both drugs, at the earliest time periods at which we could be certain that relaxation was occurring (based on preliminary pilot experiments). If a causal relationship exists between PKG activation and relaxation, then the kinase activity should increase at or before the onset of relaxation. Because we did not attempt to measure the kinase activity before the onset of relaxation, we cannot say whether kinase activation actually preceded relaxation. In the temporal study with NTG, both cGMP levels and PKG activity reached a peak within 1 or 2 min after administration of the drug and began to decline thereafter, even though relaxation was maintained for at least 10 min. This is not inconsistent with a role for the kinase in relaxation if it is assumed that it is acting via phosphorylation of a protein or proteins involved in regulating tension in the muscles. These proteins might still be phosphorylated, and exerting an effect, for some time after the kinase activity had returned toward control. A transient peak in cGMP levels accompanied by sustained relaxation has been previously reported by Keith et al. (1982) in rat aorta treated with high concentrations of NTG.

Thus the results of the present study confirm the earlier reports (Lincoln and Fisher-Simpson, 1983; Fiscus et al., 1984; Jiang et al., 1992) that PKG can be activated by SNP in intact vascular preparations and, in addition, demonstrate activation of the kinase by another vasodilator, NTG. Our studies provide the first data collected in an attempt to directly correlate smooth muscle relaxation caused by these agents with PKG activation in the same preparations. None of the previous studies included concentration-response or temporal studies, both of which are usually considered necessary for establishing a causal relationship in studies of this type. A reasonably good, though not a perfect, correlation between PKG activation and relaxation was observed in the present experiments.

In our opinion, the PKG assay used in our experiments is a more sensitive and reliable procedure than those used previously. However, there are problems inherent in any procedure used to measure PKG activity in an intact tissue. As discussed by Fiscus et al. (1984), one difficulty in trying to estimate the activation state of PKG in an intact tissue is that once the tissue is homogenized, the endogenous cGMP is diluted and begins to dissociate from the kinase, thus inactivating it. The rate of this dissociation is decreased as the temperature is lowered, and for this reason, phosphotransferase activity in our experiments was determined at 0-4°C as suggested by Fiscus et al. (1984). Even with this precaution, some degree of dissociation will presumably occur during the assay procedure, so we are probably underestimating the activation state of the enzyme. As discussed by Fiscus et al. (1984), two distinct binding sites have been reported for the binding of cGMP to PKG, and full activation of the enzyme requires occupancy of both sites. Binding at one of the sites occurs only at high concentrations of cGMP, and dissociation of cGMP from this site is more rapid than from the other. Therefore, it is possible that underestimation of PKG activation may be greater in samples with very high cGMP levels. Although we have no direct evidence for this possibility in the present study, it could explain the fact that we rarely observed activity ratios above 0.5 even in tissues with very high cGMP levels. Underestimation of PKG activity could also account for the apparent dissociation between PKG activation and relaxation seen at the lowest concentration of NTG (0.01 µM). As noted above, 0.01 µM NTG relaxed the aortic strips, but the increase in the PKG activity ratio seen with that concentration of the drug was not statistically significant. We should also consider the possibility that, at lower concentrations of NTG, the relaxant effect may be due to a cGMP-independent mechanism such as a direct action on ion channels. For example, Bolotina et al. (1994) have suggested that NO itself may have a direct effect on calcium-dependent potassium channels in rabbit thoracic aorta. However, in the absence of direct evidence for this possibility in the present study, the apparent dissociation seen with the lowest concentration of NTG may still be explained by a lack of sensitivity of the PKG assay, as discussed above.

In summary, in spite of the technical limitations we have noted, we observed a reasonably good correlation between PKG activation and relaxation of rabbit aorta by SNP and NTG. Thus, although a causal relationship between the two parameters has not been definitely established, our results appear to be consistent with a possible role for activation of PKG in the chain of events leading to vascular relaxation by cGMP-elevating agents.