Selective Guanylyl Cyclase Inhibitor Reverses Nitric Oxide-Induced Vasorelaxation
Effects of a novel soluble guanylyl cyclase inhibitor, 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ), were characterized on guanylyl cyclase activity in cytosolic fraction of COS-7 cells overexpressing the α1 and β1 subunits of the rat soluble enzyme. ODQ was a noncompetitive inhibitor of soluble guanylyl cyclase with respect to Mn2+ or Mn2+-GTP and was a mixed competitive/noncompetitive inhibitor with respect to nitric oxide (NO) donation. ODQ (10 μmol/L) reduced deta nonoate-stimulated cGMP production in COS-7 cells overexpressing soluble guanylyl cyclase and in rat aortic vascular smooth muscle cells. ODQ did not inhibit particulate forms of the enzyme rat guanylyl cyclase-A, -B, or -C, did not block NO synthase, and did not auto-oxidize deta nonoate-donated NO in the presence of cells at physiological pH. Therefore, ODQ is a selective inhibitor of soluble guanylyl cyclase. Using ODQ in isolated aortic ring preparations, we tested the hypothesis that soluble guanylyl cyclase mediates vasorelaxant activity associated with NO. Phenylephrine (100 nmol/L)-precontracted, isolated rat aortas were relaxed in a concentration-dependent manner by deta nonoate (0.01 to 100 μmol/L) and nitroglycerin (0.01 to 300 μmol/L). ODQ (10 μmol/L) attenuated deta nonoate- and nitroglycerin-mediated relaxation of contracted aortas. ODQ had no effect on natriuretic peptide-, 8-bromo-cGMP-, isoproterenol-, or bimak-alim-mediated aortic relaxation. These results support the hypothesis that soluble guanylyl cyclase mediates vasorelaxant activity associated with nitric oxide.
- cyclic GMP
- endothelium-derived relaxing factor
- guanylyl cyclase
- nitric oxide
- nitric oxide synthase
- natriuretic peptides
- 8-bromo-cGMP = 8-bromo-guanosine 3′:5′-monophosphate
- ANP = type A natriuretic peptide
- BNP = type B natriuretic peptide
- CNP = type C natriuretic peptide
- GC-A, -B, -C = guanylyl cyclase-A, -B, -C
- LY83583 = 6-anilino-5,8-quinolinedione
- NO(S) = nitric oxide (synthase)
- ODQ = 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one
- sGC = soluble guanylyl cyclase
Nitric oxide is now known to be the major endothelium-derived relaxing factor.1,2⇓ Several vasodilators, including bradykinin, acetylcholine, and serotonin, are thought to have activity by virtue of their effects to stimulate endothelial cell NOS.1,2⇓ Several studies over the years have suggested that the vasodilatory actions of NO are mediated by increases in cGMP1–3⇓⇓ NO and several precursors of the radical activated soluble forms of guanylyl cyclase,1–4⇓⇓⇓ and membrane permeable analogs of cGMP vasodilated blood vessels.3,5⇓ In addition, in most studies NO-stimulated production of cGMP correlates with the NO-mediated vasodilation. However, NO-induced smooth muscle relaxation has been dissociated from increases in cGMP in some studies.4,6,7⇓⇓ Potential explanations for the latter observation are that NO also stimulates guanylyl cyclase activity in cell types other than vascular smooth muscle (ie, adventitia, fibroblasts, autonomic nerves, and/or endothelium), that current assays for vascular contractile responses are significantly more sensitive than the measurement of cGMP concentrations within cellular compartments associated with vasoactivity, and that NO may mediate vascular effects independent of guanylyl cyclase stimulation.
Several groups have used either methylene blue or LY83583 as sGC inhibitors to focus on the latter possibility by attempting to dissociate NO-mediated smooth muscle relaxation from increases in enzyme activation.1,8,9⇓⇓ Both of these compounds appear to block the vascular relaxation and lower cGMP concentrations associated with vasoactive agents that stimulate NOS or NO donors, including nitroglycerin, sodium nitroprusside, and 3-morpholino-sydnonimine (ie, SIN-1). Therefore, methylene blue and LY83583 have been used to reach the conclusion that sGC mediates the vasodilatory effect of NO. This conclusion is questionable because both methylene blue and LY83583 are not truly selective sGC inhibitors, and their effects on the NO/sGC system are likely completely or partially explained by indirect actions. Methylene blue is now known not only to inhibit NOS10,11⇓ but also to lower NO concentrations by the production of superoxide anion.12,13⇓ LY83583 was first reported to reduce cGMP concentrations in the guinea pig lung by an unknown mechanism not involving a reduction of sGC activity.14 Regardless, several studies over the last decade have used it to form conclusions about the relationship of NO- and sGC-associated cellular events, including vasorelaxation in aortic rings.8,9⇓ An early report demonstrated that LY83583 blocked the production of endothelium-derived relaxing factor,15 and a very recent study has demonstrated that, like methylene blue, LY83583 also can act as an inhibitor of NOS in rat cerebellar homogenates.11 The latter study,11 as well as others,16,17⇓ has revealed more parallels between methylene blue and LY83583, including the production of superoxide anion, to explain the effects of the latter compound to decrease NOS activity or auto-oxidize NO. Therefore, it is impossible to use LY83583 and/or methylene blue to differentiate between direct inhibition of sGC activity (ie, effects on the receptor) or reductions in NOS or NO concentrations (ie, effects on ligand concentrations).
Garthwaite et al18 recently reported the identification of a potent, selective inhibitor of NO-sensitive guanylyl cyclase called ODQ. They demonstrated that ODQ inhibited purified NO-stimulated sGC activity in a noncompetitive manner with respect to the substrate GTP and further determined that ODQ did not inactivate reactive NO, did not produce superoxide anions, and did not block the sGC-independent effect of NO to induce macrophage cytotoxicity.18 ODQ blocked NO-donor stimulation of sGC in intact endothelial cells and had no effect on ANP stimulation of particulate guanylyl cyclase as assessed by measurement of cGMP production by radioimmunoassay.18 ODQ also did not inhibit forskolin-stimulated adenylyl cyclase in endothelial cells.18 Taken together, these initial results suggested that ODQ was potentially a much better reagent for assessing the involvement of sGC in NO effects than either methylene blue or LY83583.
In the present study, we characterized the effects of ODQ on Mn2+-, Mn2+-GTP-, and deta nonoate19-stimulated sGCs in the cytoplasmic fraction from COS-7 cells overexpressing the α1 and β1 subunits of the rat enzyme (COS-7 sGC cells). Furthermore, we also examined the effect of ODQ in cells overexpressing particulate forms of the enzyme GC-A, GC-B, and GC-C, which act as receptors for ANP, CNP, and heat-stable enterotoxin, respectively. Following this characterization, we used ODQ in rat isolated aortic rings to test the hypothesis that sGC mediates vasorelaxant activity associated with NO donation.
Unless specified in the following text, common laboratory reagents used in this study were purchased from either Sigma Chemical Co or Fisher Scientific. All protease inhibitors were from Sigma Chemical Co as well unless otherwise noted.
Cell Culture and Transfection
Cells were maintained in a humidified 95% air-5% CO2 water-jacketed incubator (Nuaire) at 37°C. HEK-293 cells overexpressing stable transfectants of rat GC-A, GC-B, and GC-C (HEK-293 GC-A, -B, and -C cells, respectively) were kindly provided by Dr David L. Garbers (Howard Hughes Medical Institute, University of Texas Southwestern Medical Center [Dallas]) and cultured as described.20–22⇓⇓
COS-7 cells were cultured as previously described23 and transfected with the α1 and β1 subunits of sGC by the DEAE/ dextran method.24 The rat α1 and β1 subunit cDNAs in the mammalian expression vector (pCMV5) were also provided by Dr D.L. Garbers.25 The expression of active sGC requires the cotransfection of both α and β subunits.25–29⇓⇓⇓⇓ Cells were cotransfected with 5 μg of each subunit. Mock-transfected cells were transfected with 10 μg pCMV5. Cells were allowed to grow for 48 hours before experimentation. Expression of both subunits was confirmed by deta nonoate (nonoate) (Cayman Chemical)-stimulated production of cGMP in whole cells or guanylyl cyclase activity in soluble fractions from these cells.
Rat Aortic Vascular Smooth Muscle Cells
Cells were isolated according to the method of Diglio et al.30 Male Sprague-Dawley rats (300 to 400 g) were given a lethal intraperitoneal injection of sodium pentobarbital (50 mg/kg), and a thoracotomy was performed to remove the full-length thoracic aorta. These manipulations were performed in accordance with the Medical College of Wisconsin Animal Care Committee guidelines. Ring explants (1 to 1.5 mm thick) were placed into 24-well tissue culture plates (Falcon) containing Dulbecco’s modified Eagle’s medium (Gibco Life Technologies) supplemented with 10% defined fetal bovine serum (Hyclone Lab). Explants were removed 3 to 4 days later, and the cells remaining on the dish were grown to confluence. Growth medium was the same as that above during explantation with the following supplements: 2 mmol/L l-glutamine, 100 U/mL penicillin, and 70 μmol/L streptomycin (Gibco). The smooth muscle identity of these cultures was verified morphologically (ie, hill and valley growth pattern) and by immunohistochemical staining for both the presence of smooth muscle-specific α-actin and the absence of factor VIII antigen. Immunohistological staining verified that the cells were more than 97% smooth muscle. The cells were used in passages 4 to 6.
Bovine Coronary Artery Endothelial Cells
Endothelial cells were kindly provided by Dr William B. Campbell (Medical College of Wisconsin [Milwaukee]). The isolation and culturing conditions of these cells have been previously described31 and were followed throughout these studies.
Whole-Cell cGMP Production Studies
Experiments testing ODQ effects (Tocris Cookson) on particulate guanylyl cyclases were examined in HEK-293 GC-A, -B, and -C cells overexpressing each of these enzymes. Those testing the effects of ODQ on sGC were performed in both COS-7 sGC cells and rat aortic vascular smooth muscle cell primary cultures. All experiments were performed in two paired 24-well plates (Falcon or Sarstedt). One plate received ODQ treatment and the other ODQ vehicle (ethanol). Before initiation of experiments, confluent plates of cells were rinsed twice with 5 mL PBS (pH 7.4, 37°C). Cells were preincubated at 37°C for 10 minutes with 0.25 mmol/L isobutylmethylxanthine and ODQ vehicle or 10 μmol/L ODQ in HEPES (25 mmol/L)-buffered (pH 7.4) Dulbecco’s modified Eagle’s medium (Medical College of Wisconsin Tissue Culture Facility). Preincubation medium was removed and replaced with fresh buffer of the same components with the appropriate guanylyl cyclase agonist for 5 minutes at 37°C. Four paired experiments were performed per set of ODQ vehicle- and ODQ-treated plates. HEK-293 GC-A, -B, and -C cells were treated with vehicle (water) and increasing concentrations of ANP, CNP, and heat-stable enterotoxin, respectively. COS-7 sGC cells and rat aortic vascular smooth muscle cells were treated with nonoate vehicle (0.01 mol/L NaOH) and increasing concentrations of the NO donor or its by-product, ethylamine (Aldrich Chemical Co). Specifically, vehicle or one of five concentrations of agonist was given per well (ANP, CNP, heat-stable enterotoxin: vehicle, 10−10, 10−9, 10−8, 10−7, or 10−6 mol/L; nonoate/ ethylamine: vehicle, 10−6, 10−5, 10−4, 10−3, or 10−2 mol/L). After the 5-minute incubation period, the supernatants were aspirated, and the guanylyl cyclase reaction was terminated by the addition of 0.5 mL of 1 mol/L perchloric acid. Antiserum for cGMP was graciously supplied by Dr D.L. Garbers.32 Samples containing cGMP were subjected to alumina-Dowex ion-exchange column purification32 and radioimmunoassay as previously described.33,34⇓ Resultant samples were counted on a Packard Cobra II Auto-Gamma Counter.
Guanylyl Cyclase Assay
COS-7 sGC Cells
Two days after transfection, COS-7 sGC cells were harvested and resuspended in 10 mmol/L HEPES (pH 7.4), 150 mmol/L NaCl, 1 mmol/L EDTA, 1.5 μmol/L aprotinin, 10 μmol/L E-64 (Boehringer Mannheim Biochemicals), and 0.1 μmol/L PMSF. After homogenization, cytosolic proteins were isolated by centrifugation (15 000g, 30 minutes, 4°C). Guanylyl cyclase activity was determined in 100-μL reactions containing 5 μg soluble protein,35 40 mmol/L triethanolamine HCl, pH 7.5, 0.25 mmol/L isobutylmethylxanthine, 2 mmol/L DTT, 1 μmol/L PMSF, and 1 μmol/L pepstatin A. In kinetic studies, 5-minute reactions were performed in the presence or absence of ODQ (10 μmol/L) at 37°C containing the varied or fixed concentrations of Mn2+, Mn2+-GTP, and/or nonoate as described in “Results” and in Table 1. The synthesis of cGMP was verified to be linear over the 5-minute course of the experiment. Reactions were terminated, and cGMP was purified and quantified as described above.
For assays quantifying the conversion of radiolabeled GTP to cGMP, the reaction mixture also contained 1 μCi [α32P]GTP (ICN Radiochemicals) and either 600 nmol/L Mn2+ per 100 nmol/L GTP or 1.1 mmol/L Mn2+ per 1 mmol/L GTP. Proteins were preincubated (10 minutes, 37°C) either in the presence of 10 μmol/L ODQ or vehicle (ethanol) before the addition of Mn2+-GTP. Reactions were terminated after 5 minutes by the addition of 0.5 mL 110 mmol/L zinc acetate and 0.5 mL 110 mmol/L sodium bicarbonate. Samples were centrifuged (1000g, 5 minutes), and supernatants were purified over alumina columns.36 Eluates containing [32P]cGMP were subjected to scintillation counting on Packard Instrument Tri-Carb 2100TR.
Male Sprague-Dawley rats (300 to 400 g) were given a lethal intraperitoneal injection of sodium pentobarbital (50 mg/kg). These manipulations were performed in accordance with the Medical College of Wisconsin Animal Care Committee guidelines. For each experiment, three full-length thoracic aortas were obtained, and the fascia was carefully removed from each vessel. Aortas were homogenized on ice in 5 mL buffer containing 25 mmol/L triethanolamine HCl (pH 7.6), 10 mmol/L NaCl, 5 μmol/L aprotinin, 10 μmol/L E-64, and 0.1 μmol/L PMSF. Guanylyl cyclase assays were performed in 100 μL of the above buffer containing 20 μg protein35 and 0.25 mmol/L isobutylmethylxanthine as described above for COS-7 sGC cells. Samples were preincubated with ODQ vehicle or 10 μmol/L ODQ before the addition of nonoate vehicle or 1 mmol/L nonoate to stimulate the enzyme for 5 minutes. Reactions were stopped, samples purified, and cGMP levels determined as described above.
Samples containing 0.1 mg lysate protein35 from transiently transfected COS-7 sGC cells or endothelial cells were preincubated (10 minutes, 37°C) in the presence of 10 μmol/L ODQ, 0.1 mmol/L Nω-nitro-l-arginine,37 0.1 mmol/L methylene blue,10 or respective vehicle treatment. NOS activity was measured by monitoring the conversion of l-[3H]-arginine to l-[3H]-citrulline as previously described.31 Reactions were terminated, and l-[3H]-citrulline was purified as described.31 The levels of l-[3H]-citrulline were determined by scintillation counting (Packard Tri-Carb 2100TR). Data were expressed as femtomoles of l-[3H]-citrulline formed per milligram of lysate protein per minute.
Male Sprague-Dawley rats (300 to 400 g) received a lethal intraperitoneal injection of sodium pentobarbital (50 mg/kg). These manipulations were performed in accordance with the Medical College of Wisconsin Animal Care guidelines. Thoracic aortas were removed and rings prepared in 5-mL tissue baths as previously described.33,38⇓ Contractile responses were monitored using Grass instrumentation as described.33,38⇓
After the preincubation and precontractions, one ring of each pair received ODQ vehicle (ethanol) and the other ring received ODQ (10 μmol/L) 15 minutes before the addition of the α1-selective adrenoceptor agonist phenylephrine (100 nmol/L). The rings were allowed to maximally contract over the next 15 minutes before the addition of vehicle and increasing cumulative concentrations of nonoate (10−8 to 3×10−3 mol/L), ANP (10−10 to 10−7 mol/L), CNP (10−10 to 10−6 mol/L), nitroglycerin (10−8 to 3×10−6 mol/L), isoproterenol (10−8 to 10−3 mol/L), bimakalim (10−10 to 10−5 mol/L), 8-bromo-cGMP (10−8 to 3×10−4 mol/L; ICN Biochemicals), or the nonoate by-product ethylamine (10−8 to 10−4 mol/L) every 4 minutes or until a plateau was reached following each addition of relaxant agonist. Because nonoate must be stored in 0.01 mol/L NaOH to avoid hydrolysis, appropriate dilutions of the NO donor in Krebs’ buffer were made approximately 1 minute before addition to the tissue baths to eliminate the potential for alkalinization of the bathing medium and associated nonspecific effects. Nitroglycerin was manufactured by Parke-Davis, and bimakalim was manufactured by Merck KGaA Co39 and graciously provided by Dr Garrett J. Gross (Medical College of Wisconsin [Milwaukee]).40
In the presence of paired ODQ vehicle- and ODQ-treated aortic rings, we also tested not only for potential desensitization to phenylephrine-induced contractions over time but also for possible time-dependent nonoate-vehicle effects on contractile activity. Two sets of paired rings were contracted with phenylephrine with or without nonoate-vehicle treatment (ie, appropriate dilutions of the nonoate storage solution, 0.01 mol/L NaOH, akin to those above for nonoate treatment) over a 65-minute period every 4 to 8 minutes to approximate the time necessary for complete responsiveness to each nonoate addition during typical experimentation. Nonoate (1 mmol/L) was added to the paired vessels after the completion of these time-dependent experiments to verify the capacity of each ring to respond to the NO donor.
Auto-oxidation of NO
NO measurements were performed using a commercially available Clark-type electrode (Iso-NO, World Precision Instruments) and documented on a strip-chart recorder. The electrode was calibrated daily using acidified potassium nitrite as the NO donor according to the manufacturer’s instructions. All measurements were made under constant stirring. The effects of ODQ or the superoxide generating system xanthine/xanthine oxidase were studied using a small plastic vial containing 2 mL Krebs’ bicarbonate buffer (pH 7.4)38 and 2×105 COS-7 cells or a rat aortic ring. The electrode was allowed to equilibrate after which non-oate (50 μmol/L) was added. After the electrode signal restabilized in the presence of nonoate, ODQ (10 μmol/L), xanthine (10 μmol/L), xanthine oxidase (5 U/mL), or a combination of xanthine (10 μmol/L) and xanthine oxidase (5 U/mL) were added, and changes in NO concentration were recorded.
The effect of ODQ on agonist-stimulated cGMP production in whole cells and agonist- or 8-bromo-cGMP-induced relaxation of precontracted rat aortic rings was assessed by comparison of control (agonist vehicle) to each treatment concentration by Dunnett’s modification of Student’s paired t test. The effect of ODQ over the full concentration-response curve was assessed for each agonist or 8-bromo-cGMP by repeated measures ANOVA.
Km and Vmax values from the kinetic assays of enzyme activity from the cytoplasmic fraction of COS-7 sGC cells were determined by traditional double-reciprocal conversions.41 Statistical comparisons between ODQ-treated and ODQ vehicle-treated preparations for both Km and Vmax values were made using Student’s paired t test.
In all figures and tables, asterisks denote the statistical significance (*P<.05, **P<.01, ***P<.001) of a specific treatment when compared with a matched control group (ie, vehicle treatment).
Effect of ODQ on NO-Donor-Mediated Stimulation of cGMP Production
These experiments were designed to compare the effects on nonoate-mediated cGMP production in both rat aortic vascular smooth muscle cells (Fig 1A) and COS-7 sGC cells (Fig 1B). In both rat aortic vascular smooth muscle cells (Fig 1A) and COS-7 sGC cells (Fig 1B), nonoate (10−5 to 10−3 mol/L) resulted in a concentration-dependent (ANOVA P<.0001) increase in cGMP production with an EC50 value of approximately 30 μmol/L. In aortic cells (Fig 1A), no greater effect was obtained after 10 mmol/L nonoate treatment (n = 2; L.J.O., T.C.H., J.G.D., unpublished data, 1996), similar to the results in COS-7 sGC cells (Fig 1B). Therefore, in both cell types, 1 mmol/L nonoate resulted in a maximal effect on cGMP production. We chose to use a 10-μmol/L concentration of ODQ in these studies on the basis of previous results demonstrating that concentration to be maximally effective.18 Pretreatment with ODQ reduced the ability of nonoate to increase cGMP levels in COS-7 cells (Fig 1B) and completely blocked the nonoate effect in the vascular smooth muscle cells (Fig 1A) (both by ANOVA, P<.0001). We also found that ODQ completely prevented maximal nonoate (1 mmol/L) stimulation of guanylyl cyclase activity in soluble fractions from rat thoracic aorta, consistent with the results in aortic smooth muscle cell cultures (L.J.O., J.G.D., unpublished data, 1996).
Nonoate is known to hydrolyze into equal molar amounts of NO and ethylamine at neutral pH.19 As a control for a by-product effect, ethylamine (10−7 to 10−3 mol/L) was tested for any effect on sGC activity. It was found to be without effect (L.J.O., J.G.D., unpublished data, 1996).
Effect of ODQ on Nonoate- and Nitroglycerin-Mediated Relaxation of Phenylephrine-Precontracted Rat Aortas
These experiments were performed to determine whether ODQ blocked the vasorelaxant activity of two NO donors, nonoate and nitroglycerin. The summary data from several experiments assessing the effect of ODQ on nonoate-induced relaxation are shown in Fig 2. In the absence of ODQ pretreatment, nonoate (10−7 to 10−4 mol/L) resulted in a concentration-dependent relaxation of phenylephrine-precontracted rat aorta (IC50=1 μmol/L; ANOVA, P<.0001). Complete relaxation was observed on reaching a final nonoate concentration of 10−4 mol/L in the absence of ODQ. After ODQ (10 μmol/L) treatment, the IC50 for nonoate (ANOVA, P<.0001) was increased by 30-fold, and the maximal effect of nonoate was attained at 10−3 mol/L and reduced by about 75%. No further relaxation was observed at 3×10−3 mol/L nonoate (n=3). These results are consistent with a mixed competitive/noncompetitive antagonism of NO effects in the rat aorta. The same ODQ concentration also blocked the ability of nitroglycerin to relax phenylephrine-precontracted aorta in a similar manner (n=3, unpublished data, 1996). Neither the nonoate by-product,19 ethylamine (10−7 to 10−3 mol/L), nor the nonoate vehicle (ethanol) had any effect on basal- or phenylephrine-induced rat aortic contractions (L.J.O., J.G.D., unpublished data, 1996; Fig 2).
The nonoate-mediated relaxation in the ODQ-treated vessel also was not simply due to time-dependent loss of contractile activity, nonoate-vehicle effect, or desensitization to phenylephrine. Control experiments were performed whereby phenylephrine was added to paired ODQ-treated and ODQ vehicle-treated aortic rings. The associated phenylephrine-induced contractions were maintained with or without nonoate-vehicle treatment for the approximate 65-minute duration of the experiments in these studies (n=3). In both variations, addition of 1 mmol/L nonoate at the end of this time period resulted in complete relaxation of the ODQ vehicle-treated ring and partial relaxation (≈20% to 25%) of the ODQ-treated vessel (n=3), akin to the results of Fig 2.
Effect of ODQ on sGC Activity in the Cytosolic Fraction From COS-7 sGC Cells
It has been previously shown that maximal rat lung sGC activation is obtained in the presence of excess Mn2+ over GTP.42 For kinetic analysis of an enzyme system requiring an excess concentration of cation over nucleotide, the system can be viewed as having two substrates: a metal-nucleotide chelate, Mn2+-GTP, and the metal ion, Mn2+.43 Chrisman et al42 have previously demonstrated that sGC is such an enzyme. Similar methods to those of the latter study42 were used in the following experiments assessing enzyme activation in the cytoplasmic fraction from COS-7 sGC cells.
These studies were performed to determine whether ODQ was a competitive and/or noncompetitive inhibitor of sGC with respect to NO, Mn2+, and Mn2+-GTP. Km and Vmax values of sGC for Mn2+ and Mn2+-GTP were determined at near-saturating concentrations42 of Mn2+-GTP (2 mmol/L) and Mn2+ (125 μmol/L), respectively, in the presence and absence of ODQ (Table 1). In both variations, ODQ had a statistically significant effect to lower Vmax approximately threefold with no effect on Km (Table 1). The NOS inhibitor, Nω-nitro-l-arginine (0.1 mmol/L) did not alter these kinetic reactions (L.J.O., J.G.D., unpublished data, 1996). These results indicate a noncompetitive inhibition with respect to both metal-nucleotide chelate or Mn2+ independent of the presence of NO. Km and Vmax values of sGC for NO were determined in the presence and absence of ODQ with fixed concentrations of Mn2+ (125 μmol/L) and GTP (12.5 μmol/L)44 (Table 1). As shown in Table 1, ODQ significantly lowered the Vmax and increased the Km (approximately two-fold) of the enzyme for nonoate-donated NO. The latter result argues that ODQ is a mixed competitive/noncompetitive inhibitor with respect to NO donation. In all three variations, the synthesis of cGMP was linear over the 5-minute course of the experiment. Subsequent experiments also demonstrated that NOS activity was not detected in COS-7 sGC cells (see below, Table 2).
Lack of ODQ Effect on NOS Activity in Endothelial Cells and Absence of NOS Activity in COS-7 sGC Cells
The purpose of these experiments was to test for an effect of ODQ on NOS activity in endothelial cells and in COS-7 sGC cells. Table 2 shows that Nω-nitro-l-arginine and methylene blue (0.1 mmol/L each) completely block the conversion of [3H]-l-arginine to [3H]-l-citrulline in bovine coronary artery endothelial cells. ODQ (10 μmol/L), however, does not effect the same reaction, demonstrating that it does not modulate NO generation. NOS activity was not detectable in COS-7 sGC cells (Table 2).
Effect of Xanthine/Xanthine Oxidase and ODQ on the Auto-oxidation of Nonoate-Donated NO
These studies determined whether ODQ could oxidize NO in the presence of COS-7 cells or rat aortic rings at physiological pH. In both of these preparations, the effects of the superoxide generating system xanthine/xanthine oxidase and ODQ were examined. Following treatment with 50 μmol/L nonoate in the presence of xanthine/xanthine oxidase vehicle, the concentration of NO generated was 168±15 nmol/L (n=3) or 266±47 nmol/L (n=2) in the presence of COS-7 cells or aorta, respectively. Combined treatment with xanthine (10 μmol/L) and xanthine oxidase (5 U/mL) lowered the levels of nonoate-donated NO from those obtained in the presence of nonoate and xanthine/xanthine oxidase vehicle (above) to 49.4±9.8% for COS-7 cells (n=3) and 22.1±3.2% (n=2) for aortic rings. In contrast, ODQ (10 μmol/L) did not result in reductions of NO concentrations, which were 105.1±1.3% (n=3) and 107.8±0.2% (n=2) of the control levels reached in the presence of both 50 μmol/L nonoate and xanthine/xanthine oxidase vehicle (above) for COS-7 cells and rat aortic rings, respectively. Xanthine (10 μmol/L) or xanthine oxidase (5 U/mL) alone had no effect on the amount of nonoate-donated NO in the presence of COS-7 cells or aortic rings (L.J.O., E.T.K., J.G.D., unpublished data, 1996).
Lack of ANP-, CNP-, and Heat-Stable Enterotoxin -Stimulated cGMP Production in HEK-293 Cells Expressing Stable Transfectants of GC-A, -B, and -C, Respectively
These experiments were designed to test whether ODQ had any effect on the membrane-associated guanylyl cyclases GC-A, -B, or -C. ANP, CNP, and heat-stable enterotoxin (10−10 to 10−6 mol/L) result in concentrationdependent (ANOVA, P<.0001) stimulation of cGMP production in HEK-293 GC-A (Fig 3A), -B (Fig 3B), and -C (Fig 3C) cells, respectively. ODQ pretreatment had no effect by ANOVA on the ability of ANP, CNP, or heat-stable enterotoxin to stimulate the membrane-associated guanylyl cyclases: GC-A, -B, or -C (Fig 3A through C, respectively), which act as the receptors for these peptides.
Lack of ODQ Effect on Natriuretic Peptide-Induced Relaxation of Phenylephrine-Precontracted Rat Aorta
These studies tested the ability of ODQ to block the effect of two known vasorelaxant peptides, ANP and CNP, which act as agonists of the membrane-associated guanylyl cyclases GC-A and GC-B. Both ANP (10−9 to 10−7 mol/L; IC50=3 nmol/L; Fig 4) and CNP (10−8 to 10−6 mol/L; IC50=30 nmol/L; Fig 4) resulted in concentration-dependent vasorelaxation of precontracted rat aorta (ANOVA, P<.0001) similar to those published previously from this laboratory.33 The vasorelaxant effects of both ANP and CNP were unaffected by pretreatment with ODQ (Fig 4; ANOVA, P<.0001).
Lack of ODQ Effect on 8-Bromo-cGMP-Mediated Relaxation of Phenylephrine-Precontracted Rat Aorta
These experiments tested whether ODQ blocked cGMP-mediated vasorelaxation downstream of guanylyl cyclase. Fig 5 shows that the membrane-permeable cGMP analog 8-bromo-cGMP (10−7 to 3×10−4 mol/L) resulted in a concentration-dependent relaxation of phenylephrine-contracted rat aortas (IC50=30 μmol/L; ANOVA, P<.0001). ODQ pretreatment of a paired aortic ring in each experiment did not affect the ability of 8-bromo-cGMP to relax the precontracted vessel (Fig 5; ANOVA).
Lack of ODQ Effect on Isoproterenol- and Bimakalim-Mediated Relaxation of Phenylephrine-Precontracted Rat Aortas
To further examine the specificity of ODQ for blocking NO-mediated vasorelaxation, we also tested the quinoxalin derivative for any antagonistic effect on two other known vasodilatory agents that signal their cellular responses independent of guanylyl cyclase activation. ODQ (10 μmol/L) was found to have no effect on the vasorelaxation facilitated by the β-selective adrenoceptor agonist isoproterenol45 and the ATP-sensitive K+ channel opener bimakalim39,40⇓ (L.J.O., J.G.D., unpublished data, 1996).
Nonoate resulted in a concentration-dependent stimulation of cGMP production in rat aortic vascular smooth muscle cells and COS-7 sGC cells. The ability of nonoate to increase cGMP synthesis in these cells was significantly attenuated by ODQ pretreatment, consistent with a potential inhibitory effect of the quinoxalin-derivative on sGC.18 More specifically, ODQ completely blocked nonoate effects in the aortic cells and significantly shifted the nonoate concentration-response curve in the COS-7 sGC cells to the right and suppressed the maximal effect. ODQ therefore presented a mixed competitive/noncompetitive antagonism. Nonoate and nitroglycerin, another NO donor, were likewise found to relax phenylephrine-precontracted rat aorta in a concentration-dependent manner. ODQ caused a similar significant rightward shift of the nonoate-response curve and an attenuation of the maximal effect, indicating a competitive/noncompetitive inhibition of the vasorelaxant effect of the NO donor. This result is consistent with the observations on whole-cell cGMP production in this study and a previous study in isolated intact bovine pulmonary arteries.46
These data led to a further characterization of ODQ effects on the NO/sGC/cGMP signaling cascade in both the COS-7 sGC cells and rat aortas. ODQ could have resulted in the observed effects on cGMP concentrations by directly inhibiting sGC or lowering donated-NO concentrations by the generation of superoxide anion. Kinetic studies on sGC activity in the cytosolic fraction from COS-7 sGC cells demonstrated that ODQ is a mixed competitive/noncompetitive inhibitor of the enzyme with respect to nonoate stimulation in the presence of fixed concentrations of Mn2+ and Mn2+-GTP. This result is consistent with the pharmacological activity of ODQ on both nonoate-mediated cGMP production in whole cells and aortic vasorelaxation and with a recent report that ODQ acts in part by binding to the heme-moiety.46 Additionally, we considered whether the noncompetitive effect of ODQ on sGC activity was mediated by competition with metal ion- or nucleotide-substrate. Kinetic studies with a saturating concentration of Mn2+-GTP or Mn2+ revealed that ODQ was a noncompetitive inhibitor with respect to varying concentrations of Mn2+ or Mn2+-GTP, respectively, in the absence of NO donation or NOS activity. Taken together, our kinetic data suggest that ODQ can modulate enzyme activity both by effects on the NO binding site and by interacting with a region other than that associated with NO, metal, or nucleotide binding.47
The results of the present study with ODQ also yield a clearer conclusion than those previously associating sGC with vasorelaxation using LY83583 and/or methylene blue. For over the last decade, both compounds have been touted as sGC inhibitors. They are now known to be nonspecific in their inhibitory effects on the NO/cGMP signal transduction pathway, including the inactivation of reactive NO12,16,17⇓⇓ and the inhibition of endothelium-derived relaxing factor generation15 or NOS activity in the present study, as well as others.10,11⇓ Moreover, LY83583 also may reduce cGMP levels in cells by an unknown mechanism not associated with reductions of sGC activity.14 We found that ODQ does not inhibit NOS activity and does not auto-oxidize NO in the presence of COS-7 cells or aortic rings. The latter result is consistent with the previous report demonstrating that ODQ does not oxidize NO in the absence of cells and does not attenuate NO-mediated macrophage toxicity.18
The effects of ODQ on vasorelaxation and cGMP production are also selective for NO. ODQ did not affect ANP- and CNP-induced aortic vasorelaxation or ANP-, CNP-, and heat-stable enterotoxin-stimulated cGMP production in HEK-293 GC-A, -B, and -C cells, respectively. GC-A, GC-B, and the related heat-stable enterotoxin receptor GC-C would all be potential targets for ODQ, since they are highly conserved in the catalytic region in comparison to sGC.1 In the present study, ODQ also did not attenuate the relaxation evoked by either isoproterenol, a β-selective adrenoceptor agonist,45 or bimakalim, an ATP-sensitive K+ channel opener.39,40⇓ The former is thought to relax vascular smooth muscle by increasing adenylyl cyclase activity45 and the latter by hyperpolarization resembling that of endothelium-derived hyperpolarizing factor.39,40,48⇓⇓
ODQ also does not have any apparent effect on vascular activity of cGMP downstream of cyclase activation (ie, protein kinase or phosphodiesterase stimulation). This conclusion is based on the novel observations that ODQ is unable to block membrane-permeable cGMP analog- and CNP-mediated vasorelaxation in the absence of the phosphodiesterase inhibitor, isobutylmethylxanthine. A vasorelaxant role for GC-B has been previously established in that a monoclonal antibody specific to the extracellular region of GC-B blocks CNP binding, CNP-mediated production of cGMP, and CNP-induced relaxation of phenylephrine-contracted rat aorta.33
Interestingly, ODQ does not fully antagonize nonoate-mediated aortic relaxation but completely ablates the NO-donor-stimulated cGMP production even at maximally efficacious concentrations (ie, 1 mmol/L) in rat aortic vascular smooth muscle cells and in soluble protein fractions from rat aorta, respectively. Similar results using ODQ in bovine pulmonary arteries led to the suggestion of sGC-dependent and sGC-independent mechanisms for NO-mediated vascular relaxation.46 The data from the present study could be interpreted to fit this dichotomy as well. However, perhaps the assay of NO-donor effects on vascular smooth muscle reactivity is more sensitive than the current measurements of intracellular cGMP concentrations, rendering a clear conclusion of two separate path-ways impossible at this time. Future studies may be able to address this issue more judiciously with comparisons of NO-donor activity in isolated vessels from wild-type versus sGC-gene knockout animals, if and when the latter become available.
The results of the present study clearly demonstrate that ODQ can inhibit sGC activation independent of the presence of NO donors or NOS activity and that ODQ neither modulates the vascular effects of membrane permeable cGMP analogs nor auto-oxidizes NO. These observations dissociate the activities of ODQ from effects on NO concentrations and downstream cGMP-mediated cellular events (eg, activation of cyclic nucleotide-sensitive protein kinases or cyclic nucleotide-mediated activation of ion channels), further supporting a direct effect of the quinoxalin derivative to inhibit the purified enzyme.18,47⇓ These data support the hypothesis that sGC mediates vasorelaxant activity associated with NO donors. ODQ will likely become an important tool in attempting to differentiate between sGC-dependent and -independent effects of NO.
This work was supported by American Heart Association, Wisconsin Affiliate, Grant-in-Aid 96-GB-58 (Dr Drewett) and the 1995 Ruth Salta Junior Investigator Award (Dr Drewett) from the National Heart Foundation, a program of the American Health Assistance Foundation. The authors also thank Dr David L. Garbers and the Howard Hughes Medical Institute for providing the cGMP antibody, the HEK-293 cells overexpressing stable transfectants of the membrane-associated guanylyl cyclases (GC-A, -B, and -C), and the cDNA clones of the rat α1 and β1 soluble guanylyl cyclase subunits; Dr William B. Campbell for providing the bovine coronary artery endothelial cells; Dr Garrett J. Gross for providing bimakalim and nitroglycerin; and Dr Begonia Y. Ho for the critical reading of this manuscript.
Cohen RA, Vanhoutte PM. Endothelium-dependent hyperpolarization beyond nitric oxide and cyclic GMP. Circulation. 1995; 92 : 3337 –3349.
Lincoln TM, Cornwell TL. Intracellular cyclic GMP receptor proteins. FASEB J. 1993; 7 : 328 –338.
Wolin MS, Cherry PD, Rodenburg JM, Messina EJ, Kaley G. Methylene blue inhibits vasodilation of skeletal muscle arterioles to ace-tylcholine and nitric oxide via the extracellular generation of superoxide anion. J Pharmacol Exp Ther. 1990; 254 : 872 –876.
Schmidt MJ, Sawyer DB, Truex LL, Marshall WS, Fleich JH. LY83583: an agent that lowers intracellular levels of cyclic guanosine 3′:5′-monophosphate. J Pharmacol Exp Ther. 1985; 232 : 764 –769.
Mulsch A, Busse R, Liebau S, Forstermann U. LY83583 interferes with the release of endothelium-derived relaxing factor and inhibits soluble guanylate cyclase. J Pharmacol Exp Ther. 1988; 247 : 283 –288.
Cherry PD, Omar HA, Farrell KA, Stuart JS, Wolin MS. Superoxide anion inhibits cGMP-associated bovine pulmonary arterial relaxation. Am J Physiol. 1990; 25 : H1056 –H1062.
Garthwaite J, Southam E, Boulton CL, Nielsen EB, Schmidt K, Mayer B. Potent and selective inhibition of nitric oxide-sensitive guanylyl cyclase by 1H-[1,2,4]Oxadiazolo[4,3-α]quinoxalin-1-one. Mol Pharmacol. 1995; 48 : 184 –188.
Potter LR, Garbers DL. Dephosphorylation of the guanylyl cyclase-A receptor causes desensitization. J Biol Chem. 1992; 267 : 14531 –14534.
Chrisman TD, Schulz S, Potter LR, Garbers DL. Seminal plasma factors that cause large elevations in cellular cyclic GMP are C-type natriuretic peptides. J Biol Chem. 1993; 268 : 3698 –3703.
Vaandrager AB, Schulz S, De Jonge HR, Garbers DL. Guanylyl cyclase C is an N-linked glycoprotein receptor that accounts for multiple heat-stable enterotoxin-binding proteins in the intestine. J Biol Chem. 1993; 268 : 2174 –2179.
Thompson DK, Garbers DL. Dominant negative mutations of the guanylyl cyclase-A receptor. J Biol Chem. 1995; 270 : 425 –430.
Kriegler M. DNA transfer. In: Gene Transfer and Expression. New York, NY: Stockton Press; 1990: 99 –100.
Yuen PST, Doolittle LK, Garbers DL. Dominant negative mutants of the nitric oxide-sensitive guanylyl cyclase. J Biol Chem. 1994; 269 : 791 –793.
Nakane M, Arai K, Saheki S, Kuno T, Buechler W, Murad F. Molecular cloning and expression of cDNAs coding for soluble guanylate cyclase from rat lung. J Biol Chem. 1990; 265 : 16841 –16845.
Campbell WB, Gebremedhin D, Pratt PF, Harder DR. Identification of epoxyeicosatrienoic acids as endothelium-derived hyperpolarizing factors. Circ Res. 1996; 78 : 415 –423.
Hansborough JR, Garbers DL. Sodium-dependent activation of sea urchin spermatazoa by speract and monensin. J Biol Chem. 1981; 256 : 2235 –2241.
Drewett JG, Fendly BM, Garbers DL, Lowe DG. Natriuretic peptide receptor-B (guanylyl cyclase-B) mediates C-type natriuretic peptide relaxation of precontracted rat aorta. J Biol Chem. 1995; 270 : 4668 –4674.
Olson LJ, Lowe DG, Drewett JG. Novel natriuretic peptide receptor/ guanylyl cyclase A-selective agonist inhibits angiotensin II- and for- skolin-evoked aldosterone synthesis in a human zona glomerulosa cell-line. Mol Pharmacol. 1996; 50 : 430 –435.
Schulz S, Green CK, Yuen PST, Garbers DL. Guanylyl cyclase is a heat-stable enterotoxin receptor. Cell. 1993; 63 : 941 –948.
Ishii K, Chang B, Kerwin JF Jr, Huang Z-J, Murad F. Nω-nitro-L-arginine: a potent inhibitor of endothelium-derived relaxing factor formation. Eur J Pharmacol. 1990; 76 : 219 –223.
Drewett JG, Trachte GJ, Marchand GR. Atrial natriuretic factor inhibits adrenergic and purinergic neurotransmission in the rabbit isolated vas deferens. J Pharmacol Exp Ther. 1989; 248 : 135 –142.
Chrisman TD, Garbers DL, Parks MA, Hardman JG. Characterization of particulate and soluble guanylyl cyclases from rat lung. J Biol Chem. 1975; 250 : 374 –381.
Cleland WW. Steady-state kinetics. In: Boyer PD, ed. The Enzymes, Vol. 2. 3rd ed. New York, NY: Academic Press; 1970: 1 –65.
Kimura H, Mittal CK, Murad F. Activation of guanylate cyclase from rat liver and other tissues by sodium azide. J Biol Chem. 1975; 250 : 8016 –8022.
Hoffman BB, Lefkowitz RJ. Catecholamines, sympathomimetic drugs, and adrenergic receptor antagonists, In: Hardman JG, Limbird LE, Molinoff PB, Ruddon RW, Gilman AG, eds. Goodman and Gilman’s The Pharmacological Basis of Therapeutics. 9th ed. New York, NY: McGraw-Hill; 1996: 199 –248.
Brunner F, Schmidt K, Nielsen EB, Mayer B. Novel guanylyl cyclase inhibitor potently inhibits cyclic GMP accumulation in endothelial cells and relaxation of bovine pulmonary artery. J Pharmacol Exp Ther. 1996; 277 : 48 –53.
Schrammel A, Behrends S, Schmidt K, Koesling D, Mayer B. Characterization of 1H-[1,2,4]oxadiazolo[4,3-α]quinoxalin-1-one as a hemesite inhibitor of nitric oxide-sensitive guanylyl cyclase. Mol Pharmacol. 1996; 50 : 1 –5.
Standen NB, Quayle JM, Davies NW, Brayden JE, Huang Y, Nelson MT. Hyperpolarizing vasodilators activate ATP-sensitive K+ channels in arterial smooth muscle. Science. 1989; 245 : 177 –180.