A redox-based mechanism for the contractile and relaxing effects of NO in the guinea-pig gall bladder

Abstract

1. The purpose of this study was to determine the effects of sodium nitroprusside (SNP), 2,2′-(hydroxynitrosohydrazino)bis-ethanamine (DETA/NO) and 3-morpholinosydnonimine (SIN-1), NO donors which yield different NO reactive species (NO⁺, NO. and peroxynitrite, respectively), as well as exogenous peroxynitrite, on gall bladder contractility.

2. Under resting tone conditions, SNP induced a dose-dependent contraction with a maximal effect (10.3 ± 0.7 mN, s.e.m.) at 1 mm. Consistent with these findings, SNP caused a concentration-dependent depolarization of gall bladder smooth muscle. The excitatory effects of SNP were dependent on extracellular calcium entry through L-type Ca²⁺ channels. Furthermore, the contraction and depolarization were sensitive to tyrosine kinase blockade, and an associated increase in tyrosine phosphorylation was detected in Western blot studies.

3. DETA/NO induced dose-dependent relaxing effects. These relaxations were sensitive to the guanylyl cyclase inhibitor 1H-[1,2,4]oxidiazolo[4,3-a]quinoxaline-1-one (ODQ, 2 μm) but they were not altered by treatment with the potassium channel blockers tetraethylammoniun (TEA, 5 mm) and 4-aminopyridine (4-AP, 5 mm).

4. When tested in a reducing environment (created by 2.5 mm 1,4-dithiothreitol, DTT), SNP caused a relaxation of gall bladder muscle strips. Similarly, the SNP-induced contraction was converted to a relaxation, and associated hyperpolarization, when DTT was added during the steady state of an SNP-induced response.

5. SIN-1 (0.1 mm), which has been shown to release peroxynitrite, induced relaxing effects that were enhanced by superoxide dismutase (SOD, 50 U ml⁻¹). The relaxations induced by either SIN-1 alone or SIN-1 in the presence of SOD were strengthened by catalase (1000 U ml⁻¹) and abolished by ODQ pretreatment. However, exogenous peroxynitrite induced a concentration-dependent contraction, which was dependent on activation of leukotriene (LT) metabolism and extracellular calcium. The peroxynitrite-induced contraction was abolished in the presence of the peroxynitrite scavenger melatonin. These results suggest that SIN-1 behaves as an NO. rather than a peroxynitrite source.

6. We conclude that, depending on the redox state, NO has opposing effects on the motility of the gall bladder, being a relaxing agent when in NO. form and a contracting agent when in NO⁺ or peroxynitrite redox species form. Knowledge of the contrasting effects of the different redox forms of NO can clarify our understanding of the effects of NO donors on gall bladder and other smooth muscle cell types.

Nitric oxide (NO) has widespread actions as an intra- and intercellular signalling molecule that is involved in diverse physiological and pathophysiological mechanisms, including those of the vascular and gastrointestinal smooth muscle (Moncada et al. 1991; Sanders & Ward, 1992). Since NO synthase (NOS) is expressed by enteric neurons, and NOS inhibitors abolish non-adrenergic non-cholinergic (NANC) relaxation in gastrointestinal smooth muscle, NO is thought to be a mediator of inhibitory neuromuscular transmission in the gut (Sanders & Ward, 1992).

Clinically important compounds that generate NO (i.e. NO donors), such as sodium nitroprusside, are thought to act by releasing NO (Bates et al. 1991; Kowaluk et al. 1992). Although NO and NO donors always cause relaxation of vascular smooth muscle via cyclic guanosine 3′,5′-monophosphate (cGMP) accumulation (Moncada et al. 1991) or direct activation of K⁺ channels (Bolotina et al. 1994), both relaxing and contractile effects on gastrointestinal smooth muscle have been reported in response to NO or NO donors (Sanders & Ward, 1992; Saha et al. 1993; Bartho & Lefebvre, 1995; Hirano et al. 1997; Holzer et al. 1997). These contractile responses can be mediated by indirect neural mechanisms or by direct actions on smooth muscle. For example, in guinea-pig ileal longitudinal muscle, NO induces a contraction as the result of the activation of cholinergic neurons (Bartho & Lefebvre, 1994a), while direct contractions in response to NO and NO donors have been reported in the rat ileum (Bartho & Lefebvre, 1994b) and in the opossum oesophageal longitudinal muscle (Saha et al. 1993; Hirano et al. 1997). The oesophageal contractile response appears to be related to guanylate cyclase activation, cyclo-oxygenase stimulation and tyrosine phosphorylation (Saha et al. 1993; Hirano et al. 1997), whereas in the rat ileum, guanylate cyclase activation is unlikely since Methylene Blue does not interfere with the contractions and 8-bromo-cGMP only has a relaxant effect (Bartho & Lefebvre, 1995).

These contrasting results may derive from the complexity of the biological chemistry of NO, which involves an array of interrelated redox forms, nitrosonium cation (NO⁺), nitric oxide (NO.) and nitroxyl anion (NO-), whose properties and reactivities are different (Stamler et al. 1992b). Thus, the term NO neither adequately identifies its redox form nor delineates the chemical reactivity of nitrogen monoxide in biological systems.

In the gall bladder, previous studies have suggested a role for NO based in the presence of nitric oxide synthase (NOS; Mourelle et al. 1993; Talmage & Mawe, 1993; Salomons et al. 1997) and the effects of NOS inhibitors either in the resting tone or in the response to cholecystokinin (Mourelle et al. 1993; Luman et al. 1998). However, evidence both in support of (McKirdy et al. 1994) and against (Chen et al. 1998) NO as an inhibitory non-adrenergic non-cholinergic neurotransmitter in the gall bladder has been reported in studies involving electrical field stimulation. On the other hand, information regarding the effects of exogenous NO or NO donors on gall bladder contractility is lacking (Mourelle et al. 1993; Salomons et al. 1997; Luman et al. 1998). Most of these are in vivo studies, so the relaxing effects reported can be due to direct effects of NO on the gall bladder smooth muscle or to changes in systemic arterial blood pressure.

In an attempt to further elucidate the role of NO in the control of gall bladder contractility and the functional diversity of NO redox forms, we have studied the effects of three NO donors possessing different chemical properties. The drugs were sodium nitroprusside (SNP), an iron-nitrosyl compound with strong NO⁺ character, 2,2′-(hydroxynitrosohydrazino)bis-ethanamine (DETA/NO), a nitric oxide adduct that releases NO. in a controlled manner up to 20 h, and 3-morpholinosydnonimine (SIN-1), which is unique in that it simultaneously generates superoxide anion and NO., the instantaneous combination of which gives rise to peroxynitrite (Kelm et al. 1997). To test whether SIN-1 effects were due to peroxynitrite release, the effects of this donor were compared with those induced by addition of authentic peroxynitrite. In these preparations, we observed different effects depending on the NO species released. These results indicate that NO donors cannot be used indiscriminately and specific NO donors should be chosen according to the redox form of NO that is generated. The NO⁺ donor SNP had excitatory effects which were mediated through tyrosine phosphorylation. Conversely, the NO. donors had relaxing effects which were mediated through guanylate cyclase activation. Furthermore, peroxynitrite, but not SIN-1, had contractile effects due to leukotriene metabolism and activation of calcium entry through L-type calcium channels.

A preliminary report of part of this work has appeared in abstract form (Pozo et al. 1999).

METHODS

Functional studies

Dissection and contraction recording of guinea-pig gall bladder smooth muscle strips

Gall bladders were isolated from 300-450 g male guinea-pigs following deep halothane anaesthesia and cervical dislocation, and immediately placed in cold Krebs-Henseleit solution (K-HS) (for composition see Chemicals, reagents and solutions) at pH 7.35. Animals were handled in accordance with the guidelines laid down by the Animal Care and Use Committee of the University of Extremadura. The gall bladder was opened by cutting along the longitudinal axis and trimmed of any adherent liver tissue. After washing with the nutrient solution to remove any biliary component, the mucosa was scraped off and the gall bladder was cut into strips along the longitudinal axis, each strip measuring approximately 3 mm × 10 mm. On average, four strips were obtained from each guinea-pig gall bladder. Each strip was placed vertically in a 10 ml organ bath filled with the nutrient solution maintained at 37 °C and gassed with 95 % O₂-5 % CO₂. Isometric contractions were measured using force displacement transducers connected to a MacLab system consisting of a MacLab hardware unit and software application which run on the Macintosh computer. The strips were placed under an initial resting tension equivalent to a 1.5 g load and allowed a 60 min period for equilibration, during which time the nutrient solution was changed every 20 min. The muscle length corresponding to the optimal preload was then determined by increasing the length of each strip in increments of 1 mm until a maximal response to acetylcholine (10 μm) was achieved. The optimal preload muscle length was maintained throughout the duration of the experiments.

Effect of NO donors on gall bladder tone and its modification by drugs and other interventions

After establishing that no desensitization was induced after two consecutive administrations of SNP or peroxynitrite with an intermittent wash period of 30 min, the effects of several antagonists on SNP- or peroxynitrite-induced responses were studied. A known concentration of the antagonist was applied 20 min before the second application of SNP or peroxynitrite.

The responses of gall bladder strips to SNP or peroxynitrite were also tested in preparations bathed in Ca²⁺-free K-HS containing ethylene glycol-bis (β-aminoethyl ether)-N,N,N ',N ′-tetraacetic acid (EGTA) (1 mm) (for composition see Chemicals, reagents and solutions). In this case, single doses of SNP or peroxynitrite were tested immediately after replacing Ca²⁺-free K-HS with standard K-HS to avoid Ca²⁺ store depletion.

To study the relaxing effects of the NO donors, preparations were first contracted with 10 μm bethanechol. When a steady state of contraction was reached, cumulative or single doses of the NO donors were added. In some preparations, after testing the relaxing effects of the NO donors on contracted tissues, the protocol was repeated in the presence of other compounds.

Immunochemical determination of protein tyrosine phosphorylation

To determine the influence of SNP on protein tyrosine phosphorylation the muscle strips were prepared and treated in a manner identical to that used for the contraction studies. The strips were removed from the organ bath immediately before reaching maximal contraction in response to 1 mm SNP either in the presence or in the absence of the tyrosine kinase blockers genistein (100 μm) and tyrphostin-B44 (100 μm) and then clamped with liquid nitrogen-cooled forceps. The muscle strips were later ground using a liquid nitrogen-cooled pestle and mortar, homogenized in lysis solution (LS, for composition see Chemicals, reagents and solutions) using a homogenizer (OMNI International) and then sonicated for 5 s. Lysates were centrifuged at 10 000 g for 15 min. Protein concentration was measured by the Bio-Rad protein assay reagent using bovine serum albumin (BSA) as standard and the volume adjusted so that 1 ml aliquots of lysates contained the same amount of protein (200 μg ml⁻¹).

The level of tyrosine phosphorylation was determined as follows: aliquots were incubated with 4 μg antiphosphotyrosine monoclonal antibody (PY20), 4 μg goat antimouse immunoglobulin G (IgG) and 30 μl protein A-agarose overnight at 4 °C. The immunoprecipitates were washed three times with phosphate-buffered saline (PBS, for composition see Chemicals, reagents and solutions). Antiphosphotyrosine immunoprecipitates were fractionated by SDS-PAGE with a NOVEX system using 10 % polyacrylamide gels. Proteins with molecular masses higher than 30 kDa were transferred to 0.20 μm pore size nitrocellulose membranes. Membranes were blocked overnight at 4 °C using blotto solution (BS, for composition see Chemicals, reagents and solutions) and incubated for 3 h at 25 °C with 1 μg ml⁻¹ antiphosphotyrosine monoclonal antibody (4G₁₀, Upstate Biotechnology Inc., Lake Placid, NY, USA). After incubation with the primary antibody, membranes were washed twice for 10 min with blotto solution and incubated for 1 h at 25 °C with antimouse IgG-horseradish peroxidase conjugate. The membranes were finally washed twice for 10 min with blotto solution and twice for 10 min with washing solution (WS, for composition see Chemicals, reagents and solutions), incubated with enhanced chemiluminescence detection reagents (ECL) for 60 s and exposed to Hyperfilm ECL.

Intracellular recording from smooth muscle

The methods to be used for intracellular electrophysiological recording were similar to those previously described (Zhang et al. 1993). After the gall bladders were cut open from the end of the cystic duct to the base, they were pinned flat, mucosal side up, in a dish lined with Sylgard 184 elastomer (Dow Corning). The mucosal layer and underlying connective tissue were gently removed with forceps under microscopic observation whilst bathed in ice cold, recirculated modified Krebs solution (MKS, for composition see Chemicals, reagents and solutions). The preparations were then pinned out in a Sylgard-lined tissue chamber and placed on the stage of an inverted microscope (Nikon, Diaphot). Smooth muscle bundles were visualized at × 200 with Hoffman modulation contrast optics (Modulation Optics, Inc., Greenvale, NY, USA). The preparations were continuously perfused at a rate of 10-12 ml min⁻¹ with the modified Krebs solution and aerated with 95 % O₂-5 % CO₂. Temperature was maintained between 36 and 37 °C at the recording site.

Glass microelectrodes used for intracellular recording were filled with 2.0 m KCl and had resistances in the range 50-110 MΩ. A negative-capacity compensation amplifier (Axoclamp 2A; Axon Instruments, Foster City, CA, USA) with bridge circuit was used to record membrane potentials. Experimental compounds were applied by addition to the superfusing solution.

Chemicals, reagents and solutions

Drug concentrations are expressed as final bath concentrations of active species. Drugs were obtained from the following sources: acetylcholine chloride, 4-aminopyridine (4-AP), bethanechol chloride, 4-bromophenacyl bromide (4-BPB), 8-bromoguanosine 3′,5′-cyclic monophosphate (8-Br-cGMP), catalase, EGTA, genistein, indomethacin, (6-chloro-9-(4-diethylamino)-1-methylbutyl)amino-2-methoxyacridine (Mepacrine), methoxyverapamil hydrochloride (D-600), nordihydroguaiaretic acid (NDGA), superoxide dismutase (SOD, from bovine erythrocytes), SNP, tetraethylammoniun choride (TEA) tetrodotoxin (TTX) and atropine sulfate monohydrate were from Sigma Chemical Co. (St Louis, MO, USA); 1,4-dithio-d,l-threitol (DTT) was from Bio-Rad Laboratoires (Madrid, Spain); arachidonyltrifluoromethyl ketone (AACOCF₃), DETA/NO, 1,4-dihydro-5-(propoxyphenyl)-7H-1,2,3-triazolo[4,5-d]pyrimidine-7-one (zaprinast), GF109203X, ODQ, SIN-1 hydrochloride and tyrphostin-B44 were from Calbiochem (La Jolla, CA, USA); tetramethylammonium peroxynitrite ((CH₃)₄N⁺ONOO⁻) was from Alexis (San Diego, CA, USA). Other chemicals used were of analytical grade from Panreac (Barcelona, Spain).

Stock solutions of AACOCF₃, 4-BPB, genistein, NDGA, ODQ and tyrphostin were prepared in dimethylsulphoxide (DMSO). Peroxynitrite was prepared according to the manufacturer's instructions. The solutions were diluted so that the final concentration of DMSO in the organ bath was ∼0.1 % v/v. This concentration of solvent did not itself affect the mechanical activity of the tissue. Genistein and tyrphostin-B44 were protected from light exposure.

The composition of the Krebs-Henseleit solution (K-HS) was: 113 mm NaCl, 4.7 mm KCl, 2.5 mm CaCl₂, 1.2 mm KH₂PO₄, 1.2 mm MgSO₄, 25 mm NaHCO₃ and 11.5 mmd-glucose, being equilibrated with 95 % O₂-5 % CO₂, pH 7.35. The Ca ²⁺-free K-HS was prepared by substituting EGTA (1 mm) for CaCl₂. The composition of the modified Krebs solution (MK-S) used in the intracellular recordings was: 121 mm NaCl, 5.9 mm KCl, 2.5 mm CaCl₂, 1.2 mm NaH₂PO₄, 1.2 mm MgCl, 25 mm NaHCO₃ and 8 mmd-glucose, pH 7.35.

The composition of the solutions used in the immunochemical determination of protein tyrosine phosphorylation was as follows. Lysis solution (LS): 50 mm Tris-HCl (pH 7.5), 150 mm NaCl, 1 % Triton X-100, 1 % deoxycholate, 0.1 % (w/v) NaN₃, 1 mm EGTA, 0.4 mm EDTA, 2.5 μg ml⁻¹ aprotinin, 25 μg ml⁻¹ leupeptin, 1 mm phenylmethylsulfonyl fluoride (PMSF) and 0.2 mm Na₃VO₄. Phosphate-buffered saline (PBS): 137 mm NaCl, 2.7 mm KCl, 5.62 mm Na₂HPO₄, 1.09 mm NaH₂PO₄, 1.47 mm KH₂PO₄ and 0.2 mm Na₃VO₄, pH 7.4. Blotto solution (BS): 5 % non-fat dried milk in a solution containing 50 mm Tris-HCl (pH 8.0), 2 mm CaCl₂, 80 mm NaCl, 0.05 % (v/v) Tween 20 and 0.02 % NaN₃. Washing solution (WS): 50 mm Tris-HCl (pH 8.0), 2 mm CaCl₂, 80 mm NaCl, 0.05 % (v/v) Tween 20 and 0.02 % NaN₃.

Statistics

Contractile responses are expressed in absolute values (mN) and/or as a percentage of the maximal response elicited by ACh (10 μm). Each concentration-response curve was analysed to evaluate the concentration producing 50 % of the maximal response (EC₅₀) and the maximum asymptote of the curve (E_max). The relaxing effects are expressed as the percentage of bethanechol response or active tension. The active tension was calculated by subtracting the value of tension achieved after exposure of the tissue to Ca²⁺-free K-HS in the presence of 1 mm EGTA to the tension value just prior the addition of the NO donor. The concentration producing the half-maximal inhibition effect (IC₅₀) was assessed by linear interpolation on the semi-logarithmic concentration-response curve. All values are presented as means ±s.e.m. Statistical evaluation was performed by using Student's t test (two-tailed). P values smaller than 0.05 were considered significant.

RESULTS

Contractile effects of SNP on guinea-pig gall bladder

Under resting tension conditions SNP (10 μm-10 mm) elicited a concentration-dependent contraction of isolated gall bladder muscle strips. Figure 1A demonstrates typical effects of single applications of different concentrations of SNP on the contractility of gall bladder muscle strips. The onset of tension development in response to 1 mm SNP occurred within 28.4 ± 4.5 s of challenge and maximal tension was achieved after 3.2 ± 0.4 min. Tension was maintained for up to 25 min, indicating that SNP can exert prolonged effects on gall bladder tone. The contractile response to SNP was concentration dependent (Fig. 1B and C) with an EC₅₀ of 25.6 ± 1.7 μm and a threshold concentration of 10 μm (n = 6). Lower concentrations of SNP (1 nm to 1 μm) did not elicit detectable changes in gall bladder resting tone (data not shown). At 1 mm, SNP produced maximal contraction with a peak tension of 10.3 ± 0.7 mN (Fig. 1B), which was equivalent to 71.4 ± 3.7 % of the response to 10 μm ACh (n = 6). Concentrations higher than 1 mm induced smaller responses (Fig. 1).

Figure 1. Contractile effects of SNP in guinea-pig gall bladder strips.

A, original traces of records of isometric tension in gall bladder strips showing the contractile effects of different concentrations of SNP (applied as indicated by the arrows). Traces are typical of six such experiments. B and C, log concentration-response curves for the contractile effects of SNP expressed in newtons (b) or as a percentage of the maximal contractile response (C). Data points indicate means of values from six experiments. Error bars indicate s.e.m.

To investigate whether the contractile effect of SNP was due to an action of NO rather than by-products co-released with NO, and taking into account that NO yield from SNP is time and light dependent, we tested SNP (1 mm) that had been exposed to light for 24 h. In four experiments, light-exposed SNP caused a barely detectable increase in the resting tone (6.7 ± 7.2 % increase) (data not shown). These results indicate that the contractile effect of SNP in gall bladder smooth muscle is likely to involve the liberation of a redox form of NO.

Role of endogenous neurotransmitters in SNP-induced contractions

Contractile responses to SNP were reversible following washout, and reproducible contractions could be elicited at 30 min intervals; therefore, various test compounds were applied 20 min before a second application of SNP in order to characterize pharmacologically the contractile response.

The response to 1 mm SNP was unaltered by pretreatment with 1 μm atropine (65.1 ± 6.9 vs 57.2 ± 3.1 % of the response to 10 μm ACh, n = 6) indicating that SNP did not act via direct or indirect stimulation of muscarinic receptors located on the gall bladder smooth muscle. Furthermore, SNP-induced contractions were not susceptible to neural blockade with TTX (1 μm) (74.6 ± 13.6 vs 69.1 ± 5.6 % of the response to 10 μm ACh, n = 6), indicating that the release of neurotransmitters from intrinsic nerves was not involved in the contractile effects of SNP in the gall bladder.

Intracellular pathways stimulated in SNP-induced contraction

Protein tyrosine phosphatases (PTP) are susceptible targets for oxidizing agents that react with sulfhydryl groups in these proteins. As a result of this interaction, inhibition of PTP and an increase in the phosphorylation of tyrosine residues occurs (Monteiro et al. 1991). Since NO⁺ has been shown to react with thiol groups in proteins (Stamler et al. 1992a), changes in phosphorylation of tyrosine residues would be expected. To test whether the SNP-induced contractile effects involve an increase in tyrosine phosphorylation, contractile and Western blot experiments were conducted. The possibility that tyrosine kinase inhibitors would affect the contractile action of SNP was tested. Preincubation of the tissue with 100 μm genistein, a naturally occurring flavonoid which potently inhibits several tyrosine kinases, nearly abolished the contraction evoked by SNP (92.0 ± 5.2 % inhibition, n = 5, P < 0.001, Fig. 2A). Similarly, tyrphostin-B44, a tyrosine kinase inhibitor structurally unrelated to genistein, also reduced the contractile response to SNP (52.7 ± 9.4 % inhibition, n = 5, P < 0.01, Fig. 2D).

Figure 2. Activation of tyrosine phosphorylation by SNP.

A and D, effects of genistein (A) and tyrphostin-B44 (D) on contraction elicited by 1 mm SNP. Bars are means ±s.e.m. of contraction amplitudes expressed as a percentage of 10 μm ACh-elicited contractions for each strip (n = 5). Significant differences versus SNP are indicated as **P < 0.01; ***P < 0.001 (Student's t test for paired data). B and E, Western blot analysis of phosphotyrosyl proteins in control and SNP-treated tissues. Gall bladder strips were recovered from the organ bath and processed as described in Methods. Tissues were either untreated (lanes 1A and 2A) or exposed to 1 mm SNP either in the absence (lanes 1C and 2C) or in the presence of genistein (100 μm, lane 1D) or tyrphostin-B44 (100 μm, lane 2D). The effects of the tyrosine kinase inhibitors on basal phosphorylation were also tested (lanes 1B and 2B). Note that SNP increased protein tyrosine phosphorylation, which was reversed by pretreatment with either genistein (lane 1D) or tyrphostin-B44 (lane 2D). These blottograms are typical of five such experiments. C and F, summary of results of densitometric analysis of Western blot data. Levels of phosphorylation in muscle strips after various treatments were expressed as a percentage of control. Bars are means ±s.e.m. of changes in phosphorylation of 175, 130 and 75 kDa proteins (n = 5).

To test whether SNP induced an increase in tyrosine phosphorylation, we conducted Western blotting analysis of the phosphotyrosyl protein content of gall bladder muscle which had been contracted isometrically with 1 mm SNP for 10 min in the absence or presence of either genistein or tyrphostin-B44 (Fig. 2B and E). The SNP-induced contractions were associated with increased tyrosine phosphorylation of several substrates including proteins with molecular masses of 175, 130 and 75 kDa, which were also visualized in untreated samples (Fig. 2B and E, lanes 1C and 2C). When preparations were treated with SNP in the presence of genistein (100 μm), the phosphoprotein signals on the immunoblot remained at control levels (Fig. 2B, lane 1D; and Fig. 2C). Genistein alone slightly reduced basal tyrosine phosphorylation (Fig. 2B, lane 1B; and Fig. 2C). Tyrphostin-B44, like genistein, decreased both basal and SNP-induced tyrosine phosphorylation (Fig. 2E, lanes 2B and 2D; and Fig. 2F).

It is reported that the contractile actions of agents that increase protein tyrosine phosphorylation in gastric longitudinal muscle are related to arachidonic acid (AA) metabolism (Muramatsu et al. 1988); therefore, we next studied whether SNP actions on the gall bladder were sensitive to inhibitors of phospholipase A₂ (PLA₂). Both mepacrine, a general inhibitor of PLA₂, and 4-BPB, an inhibitor of secreted PLA₂ (sPLA₂), reduced SNP-induced contractions (by 97.6 ± 1.4 and 53.2 ± 7.7 %, respectively, n = 6, Fig. 3A). However, 100 μm AACO₃F, an inhibitor of cytosolic PLA₂ (cPLA₂), had no detectable effect on the SNP-induced contraction (Fig. 3A). To test which metabolic pathway for AA was activated by SNP, we pretreated the tissue with 10 μm indomethacin, an inhibitor of the cyclo-oxygenase pathway, and 100 μm NADG, an inhibitor of the 5-lipoxygenase pathway. As illustrated in Fig. 3B, the contractile response to SNP was significantly inhibited by NADG (by 51.2 ± 11.3 %, n = 6, P < 0.05), but was not significantly altered by indomethacin.

Figure 3. Activation of arachidonic acid metabolism by SNP.

A, effects of inhibitors of PLA₂ on SNP-induced contractions. After testing the effects of 1 mm SNP, strips were treated with the general inhibitor of PLA₂ mepacrine (100 μm), the inhibitor of sPLA₂ 4-BPB (100 μm) and the inhibitor of cPLA₂ AACOCF₃ (100 μm) and SNP was tested again. Note that there was no inhibition after the treatment with the inhibitor of cPLA₂. B, effects of inhibitors of the cyclo-oxygenase and lipoxygenase pathways on SNP-induced contractions. The protocol of the experiments was the same as in A, but the strips were treated with indomethacin (10 μm) or NDGA (100 μm) prior to the second SNP challenge. Note that only NADG pretreatment had significant effects. Bars are means ±s.e.m. of contraction amplitudes (n = 6) expressed as a percentage of 10 μm ACh-elicited contractions for each strip. Significant differences versus SNP are indicated as *P < 0.05; **P < 0.01; ***P < 0.001 (Student's t test for paired data).

Activation of protein kinase C (PKC) induces smooth muscle contraction, so the specific inhibitor of PKC GF109203X (1 μm) was used to test whether the SNP-induced contraction of gall bladder muscle was mediated through activation of this kinase. Pretreatment of the tissue with GF109203X had no effect on the SNP response (54.7 ± 7.6 vs 66.8 ± 9.4 % of the response to 10 μm ACh, n = 6). However, this dose of GF109203X did cause a 45.1 ± 5.6 % reduction in the response of gall bladder strips to cholecystokinin (10 nm) a known activator of PKC (n = 7, P < 0.001).

Calcium sources for SNP-induced contractions

To determine the source of calcium for these contractions, the 1 mm SNP response was assayed in calcium-free medium. Because intracellular calcium stores in the gall bladder muscle are rapidly depleted in the absence of extracellular calcium (Lee et al. 1989), muscle strips were suddenly bathed with zero calcium K-HS containing EGTA (1 mm). Contractions in response to SNP added immediately after replacing the bathing solution with Ca²⁺-free K-HS were reduced by 98 % (80.8 ± 15.5 vs 0.96 ± 0.8 % of the response to 10 μm ACh, n = 6, P < 0.001). To determine which type of Ca²⁺ channels were activated, the effect of methoxyverapamil (D-600), an antagonist of dihydropyridine-sensitive, voltage-operated Ca²⁺ channels (VOCCs), was investigated. The addition of 10 μm D-600 significantly reduced (by 82.5 ± 5.2 %, n = 5, P < 0.01, Fig. 4A) the contractile responses of SNP. This concentration of D-600 completely relaxed preparations that had been contracted with 60 mm KCl (20.6 ± 2.5 vs 0.7 ± 0.1 mm, n = 6, P < 0.01) in 16.7 ± 1.5 min.

Figure 4. SNP induces depolarization in gall bladder smooth muscle cells.

A, representative trace of an isometric tension recording showing the contraction induced by 1 mm SNP and the effect of 10 μm methoxyverapamil (D-600) on such a contractile effect. Inset, means ±s.e.m. of tension increases induced by 1 mm SNP either in the absence or in the presence of 10 μm D-600 (n = 5). Data are expressed as a percentage of 10 μm ACh-elicited contractions for each strip. Significant differences versus SNP are indicated as *P < 0.01 (Student's t test for paired data). B, representative experimental trace of records of membrane potential (V_m) showing that when SNP was added to the bathing solution, as indicated, there was a prolonged membrane depolarization that was associated with an increase in the frequency of spontaneous action potentials. These data are consistent with the observation that SNP caused a contraction of gall bladder strips, and that the contraction involves extracellular Ca²⁺. C, in the presence of genistein, the depolarization elicited by 10 μm SNP was abolished. D, means ±s.e.m. of depolarization induced by 10 μm SNP (n = 8), 10 mm (n = 4) SNP, and 10 μm SNP in the presence of 0.1 mm genistein (n = 8). Genistein caused a significant reduction in the SNP-induced depolarization (*P < 0.001; Student's t test for unpaired data). Resting membrane potentials: B, -55 mV; C, -51 mV.

As the activation of VOCCs should be associated with a depolarization of sarcoplasmic membrane, intracellular recordings were made to test whether SNP altered the membrane potential of guinea-pig gall bladder smooth muscle. As demonstrated in Fig. 4B and D, SNP caused a concentration-dependent depolarization of gall bladder smooth muscle that was associated with an increase in the frequency of spontaneous action potentials. During the depolarization to 10 mm SNP, we typically observed high frequency spikes of very small amplitude. In the presence of genistein (100 μm), which eliminated spontaneous action potentials and caused a slight (1-3 mV) hyperpolarization, the depolarization caused by SNP was nearly abolished (P < 0.001; Fig. 4C and D; n = 8). In an additional experiment, we evaluated the effect of 10 μm SNP in the presence of 10 μm D-600. In the presence of D-600 the depolarizing effect of SNP was blocked, and in four of the seven cells tested, a slight hyperpolarization was observed (mean response to SNP in the presence of D-600: -0.9 ± 1.1 mV; range = -5.3 to 4.2 mV; P < 0.001 compared with the effect of 10 μm SNP alone, Student's t test for unpaired data).

Relaxing effects of NO. donors

In contrast to the effects of SNP, exposure of gall bladder strips to DETA/NO (1 mm) resulted only in relaxation, either under normal resting tension (31.2 ± 4.8 % reduction of the active tension seen at 1 mm), or in precontracted conditions (Fig. 5A). As observed in Fig. 5A (inset), the relaxation induced by DETA/NO of the bethanechol-induced contraction was concentration dependent, reaching the maximal effect (34.4 ± 7.1 % inhibition of the bethanechol response) at the highest concentration tested (1 mm).

Figure 5. Relaxing effects of DETA/NO.

A, representative trace of an isometric tension recording showing the concentration-dependent relaxation induced by DETA/NO. After exposition of the strips to 10 μm bethanechol, cumulative concentrations of DETA/NO were added to the organ bath, as indicated by the dots. The trace is typical of six such experiments. Inset, log concentration-response curve for DETA/NO (n = 6). B, effects of the inhibitor of guanylate cyclase ODQ (2 μm) and the inhibitors of BK and K_v potassium channels TEA (5 mm) and 4-AP (5 mm) on the relaxing effects induced by 1 mm DETA/NO. Bars are means ±s.e.m. of relaxation expressed as a percentage of inhibition of 10 μm bethanechol-elicited contractions for each strip. (n = 5 for each group). Significant differences versus DETA/NO are indicated as ***P < 0.001 (Student's t test for paired data).

Since the relaxing effects of NO in several systems are mediated through guanylate cyclase activation, the guanylate cyclase inhibitor ODQ was used to test whether the relaxing effects of 1 mm DETA/NO were mediated through guanylate cyclase activation. Application of ODQ (2 μm) caused a significant reduction (81.4 ± 2.8 %, n = 5, P < 0.001) in the relaxation induced by DETA/NO (Fig. 5B). In agreement with these results, the cGMP analogue, 8-Br-cGMP was tested. At a concentration of 1 mm, 8-Br-cGMP caused an 78.5 ± 9.1 % reduction of active tension under resting tone conditions. When tested in the presence of 10 μm bethanechol, a dose-dependent relaxation was observed with a threshold concentration of 10 μm and an IC₅₀ of 0.13 ± 0.08 mm (n = 6).

As K⁺ channel activation has been shown to mediate the relaxing effects of NO (Bolotina et al. 1994), we also tested DETA/NO-induced relaxations in the presence of tetraethylammonium (TEA, 5 mm) and 4-aminopyridine (4-AP, 5 mm), blockers of BK and K_v channels, respectively. Neither TEA nor 4-AP had any effect in the relaxation induced by DETA/NO (Fig. 5B).

Since the gall bladder relaxation in response to NO clearly involved the activation of guanylate cyclase, and since SNP is well recognized as a NO donor, ODQ was used to test whether the gall bladder response to SNP represented the net result of contractile and relaxing actions. In the presence of 2 μm ODQ, the contractile response to 1 mm SNP was dramatically increased (43.7 ± 6.3 vs 72.9 ± 6.4 % of the response to 10 μm ACh, n = 7, P < 0.01). In agreement with these results, the specific cGMP phosphodiesterase inhibitor zaprinast (10 μm) caused a reduction in the contractile response to SNP (53.2 ± 4.2 vs 40.3 ± 4.4 % of the response to 10 μm ACh, n = 8, P < 0.05) without having any effect on basal tone. To establish the concentration dependence of the relaxing component of the SNP response, the effects of SNP were tested in gall bladder strips precontracted with bethanechol (10 μm). Under these circumstances, a slight, dose-dependent relaxation was observed in response to SNP, with a threshold concentration of 10 nm, and a peak concentration of 10 μm (18.5 ± 4.8 % reduction of bethanechol response, n = 7, Fig. 6A, inset). Higher concentrations induced less relaxation and in most of the experiments the tension increased in response to addition of 0.1 and 1 mm SNP (Fig. 6A).

Figure 6. Relaxing effects of SNP.

A, representative trace of isometric tension record showing the concentration-dependent relaxation induced by low doses of SNP. After exposure of the strips to 10 μm bethanechol, cumulative concentrations of SNP were added to the organ bath, as indicated by the dots. The trace is typical of seven such different experiments. Inset, log concentration-response curve for SNP (n = 7). Note that 0.1 and 1 mm of SNP induced contractile effects even on the plateau of bethanechol response. B, effects of the inhibitor of guanylate cyclase ODQ (2 μm), and the inhibitors of BK and K_v potassium channels TEA (5 mm) and 4-AP (5 mm) on the relaxing effects induced by 10 μm SNP (dose which induced the maximal relaxing effect). Bars are means ±s.e.m. of relaxation expressed as a percentage of inhibition of 10 μm bethanechol-elicited contractions for each strip (n = 8, 7 and 6 for ODQ, TEA and 4-AP group, respectively). Significant differences versus SNP are indicated as *P < 0.05 (Student's t test for paired data).

Similar to DETA/NO, the relaxing effect of 10 μm SNP (the dose which induced maximal relaxing effects) in precontracted tissues was attenuated by ODQ pretreatment (98.3 ± 22.2 % of inhibition, n = 8, P < 0.05) (Fig. 6B) but was not modified by the treatment with the K⁺ channel blockers TEA (5 mm) and 4-AP (5 mm) (Fig. 6B).

Redox-based mechanism for the effects of SNP

It has been established that SNP mediates its effects through NO⁺ generation (Stamler et al. 1992a), but in the presence of reducing agents such as thiols, SNP can yield large amounts of NO. (Bates et al. 1991). To test whether the manipulation of the redox state of the NO groups influences the effects of SNP on gall bladder preparations, we tested 1 mm SNP in a reducing environment created by the thiol agent DTT (2.5 mm).

When added to the bath, DTT (2.5 mm) induced a progressive increase in the active tension of the gall bladder strips, reaching a maximum value of 12.2 ± 1.4 mN, which was equivalent to 88.2 ± 10.8 % of the response to 10 μm ACh (n = 19, Fig. 7A). This contraction was probably due to the activation of PKC and sPLA₂ since it was significantly reduced in the presence of either 1 μm GF109203X (54.9 ± 13.8 % of reduction, n = 6, P < 0.01) or 100 μm 4-BPB (88.2 ± 10.8 % of reduction, n = 6, P < 0.001). The calcium source used to generate this contraction appears to be the extracellular medium, as DTT-induced contractile responses were abolished when tested in a calcium-free medium (n = 5). In addition, after developing tension in response to DTT, the tension decreased below the resting level (6.9 ± 1.72 mN below basal tension, n = 4) when the strips were exposed to a calcium-free K-HS containing 2.5 mm DTT.

Figure 7. Redox-based mechanism for the effects of SNP.

A, original trace showing the contractile effect of 1 mm SNP under resting tone conditions. DTT treatment (2.5 mm) induced a sustained contraction. When 1 mm SNP was assayed on the plateau of DTT response a relaxation was observed. In this experiment, fresh DTT (2.5 mm) was added to the bath immediately before testing SNP. B, percentage of inhibition induced by ODQ (2 μm), TEA (5 mm) and 4-AP (5 mm) on the relaxing response to 1 mm SNP after DTT treatment (2.5 mm). Bars are means ±s.e.m. of six experiments. Significant differences versus SNP are indicated as **P < 0.01 (Student's t test for paired data). C, original trace showing that the contractile effect of SNP is reverted into a relaxing one when DTT was added to the organ bath. Note that the tension decreases to values below those seen before DDT addition. The traces are typical of six such experiments. D, voltage recording of a gall bladder smooth muscle cell in the continuous presence of 1 mm SNP. Following application of 0.25 mm DTT, a transient hyperpolarization was observed. Resting membrane potential = -50 mV.

In the presence of 2.5 mm DTT, SNP elicited a transient relaxation (29.5 ± 8.5 % reduction of active tension, n = 6, not shown), which was consistent with the generation of NO. The transient nature of the SNP-induced relaxation in the presence of DTT, which reached a steady state in 4.95 ± 1.33 min, was probably due to a loss of the reducing power of DTT. Thus, we tested 1 mm SNP after applying fresh DTT (2.5 mm) once the DTT-induced contraction was developed. In this situation, the relaxation induced by SNP addition was stronger (79.5 ± 8.9 % of reduction of active tension, n = 10, Fig. 7A) and less transitory. It could be argued that the SNP-induced relaxing effects observed in the presence of DTT were due to the contractile state induced by DTT pretreatment. To test this possibility, gall bladder strips were precontracted with SNP (1 mm), and then exposed to DTT (2.5 mm), which converted the contractile response to a relaxing one (70.3 ± 5.9 % of reduction of active tension, n = 6, Fig. 7C). The tension dropped to values below those previously seen in response to SNP (40.4 ± 19.2 % reduction), clearly indicating that the relaxing effects were due to a product released in the reducing environment. This experimental protocol was repeated while recording intracellularly from gall bladder smooth muscle. In the presence of 1 mm SNP, DTT (0.25-2.5 mm; n = 7) caused a hyperpolarization that was associated with a strong relaxation and, in most cases, loss of the impalement. An example of a recording in which the impalement was maintained throughout the hyperpolarization caused by 0.25 mm DTT, in the presence of SNP, is shown in Fig. 7D.

To test for possible non-specific effects of such change in tissue redox state, the actions of DTT on responses to DETA/NO and SIN-1 were also determined. The relaxing responses to these agents were not significantly modified when tested during contractions to DDT vs. bethanechol (DETA/NO: 29.6 ± 1.6 % reduction of DDT-induced contraction, 30.3 ± 5.9 % reduction of bethanechol response, n = 6; SIN-1: 31.7 ± 5.6 % reduction of DDT-induced contraction, 29.8 ± 3.8 % reduction of bethanechol response, n = 6).

Similar to the results obtained with DETA/NO, 2 μm ODQ largely antagonized the relaxing response to SNP in the presence of DTT, whereas either 5 mm TEA or 5 mm 4-AP had no effects (Fig. 7B).

Effects of peroxynitrite

To assess peroxynitrite effects, two approaches were used. Firstly, since SIN-1 decays in oxygen-containing solution, with the simultaneous release of and NO. in a 1:1 stoichiometry (Kelm et al. 1997), SIN-1 (0.1 mm) was used to generate peroxynitrite in situ. This concentration of SIN-1 induced relaxing responses in unstimulated gall bladder (19.6 ± 5.0 % reduction of active tension, n = 5, Fig. 8A), although its effects were more prominent when assessed in strips precontracted with 10 μm bethanechol (34.3 ± 2.8 % reduction of bethanechol response, n = 8, Fig. 8A). The response to bethanechol was not modified after SIN-1 treatment (21.8 ± 1.6 mN before vs 23.5 ± 1.6 mN after SIN-1 treatment), indicating that the effects of the SIN-1 were reversible.

Figure 8. Effects of SIN-1 on gall bladder contractility.

A, original traces showing the reduction in tension induced by 100 μm SIN-1, added as indicated by the lines, either under resting tone conditions or in strips precontracted with bethanechol (10 μm). B, two traces from a single muscle strip showing the response to SIN-1 either in the absence or in the presence of SOD (50 U ml⁻¹). An increase in the relaxing response was observed under SOD treatment, indicating that SOD scavenges the released by SIN-1, which increases the amount of NO. SOD must increase H₂O₂ production, as indicated by the extra relaxation induced by catalase (1000 U ml⁻¹) treatment. C, in these experiments, catalase was tested on the SIN-1 response in the absence of SOD. The relaxing response was strengthened in this situation. This indicates that SIN-induced effects are due, in part, to the generation of H₂O₂. Traces are representative of five to eight experiments.

The formation of peroxynitrite from SIN-1 can be suppressed by the enzyme superoxide dismutase (SOD), which competes with NO binding on the superoxide anion, resulting in the formation of NO. (Gergel et al. 1995). When SIN-1 was tested in the continuous presence of SOD (50 U ml⁻¹), the SIN-1-induced relaxing response was potentiated by 206.2 ± 24.5 % (n = 5, P < 0.01, Fig. 8B), indicating that peroxynitrite has contractile effects that counteract NO.-induced relaxation, or that peroxynitrite is less potent at inducing gall bladder relaxation than NO. Higher concentrations of SOD did not induce more potentiation of the SIN-1-evoked relaxing effect (data not shown).

As a result of SOD activity, the levels of H₂O₂ might be elevated. When H₂O₂ was added to the organ bath, gall bladder strips contracted in a concentration-dependent manner (with an E_max of 9.6 ± 0.6 mN at 2 mm, and an ED₅₀ of 72.4 ± 8.3 μm). Thus, if H₂O₂ were produced in the dismutation of superoxide, the addition of catalase, a hydrogen peroxide scavenger, would enhance the relaxations induced by SIN-1 in the presence of SOD. Our results support this model because catalase (1000 U ml⁻¹) strengthened the relaxing responses by 116.7 ± 35.5 % (n = 5, P < 0.01, Fig. 8B). However, not only did catalase enhance the releasing response to SIN-1 in the presence of SOD, but it also was effective at increasing the relaxation in response to SIN-1 alone by 86.2 ± 12.6 % (n = 11, P < 0.001, Fig. 8C). These results suggest that SIN-1 induces H₂O₂ yield by itself.

The relaxation induced by SIN-1 was practically abolished by 2 μm ODQ (75.3 ± 11.9 % of inhibition). These findings suggest a role for stimulation of soluble guanylate cyclase in the mechanism of relaxation to SIN-1. Consistent with the results observed for NO. donors, when the effects of SIN-1 were tested in the presence of SOD (50 U ml⁻¹) and catalase (1000 U ml⁻¹) to yield NO., the relaxing effects observed were largely reduced by ODQ pretreatment (86.7 ± 11.9 % of inhibition, n = 5, P < 0.001).

To test whether SIN-1-induced relaxing effects were due to peroxynitrite release, rather than NO., authentic peroxynitrite was tested. Addition of peroxynitrite to the organ bath induced a concentration-dependent contraction (Fig. 9A and b), with a maximal response of 18.4 ± 4.5 mN (101.1 ± 12.0 % of the response to 10 μm ACh) at a concentration of 30 μm.

Figure 9. Exogenous peroxynitrite causes concentration-dependent contraction of gall bladder strips.

A, representative trace of an isometric tension recording showing concentration-dependent contraction induced by cumulative addition of peroxynitrite as indicated by the dots. B, concentration-response curve for the contractile effect of peroxynitrite expressed as a percentage of the maximal contractile response. Data are means ±s.e.m. of six experimental values. C, effect of the reactive oxygen species scavengers melatonin (1 μm), DMSO (20 mm), SOD (50 U ml⁻¹) plus catalase (1000 U ml⁻¹) on the response to 10 μm peroxynitrite. Also shown are the effects of the calcium channel blocker D-600 (100 μm), inhibitors for 5-lipoxygenase (NDGA, 100 μm), cyclo-oxygenase (indomethacin, 100 μm) and protein kinase C (GF109203X, 1 μm), and the effect of low Na⁺ medium (25 mm Na⁺). After testing the effect of 10 μm peroxynitrite the drugs were applied 20 min before a second application of peroxynitrite. Bars are means ±s.e.m. of the second contraction, expressed as a percentage of the contraction induced by peroxynitrite in the absence of drugs. (n = 5-7). Statistical differences versus the initial control response to peroxynitrite are indicated as *P < 0.05 or **P < 0.01 (Student's t test for paired data).

In order to identify the reactive oxygen species that is responsible for the contractile response, we studied the effects of 10 μm peroxynitrite before and after the treatment with DMSO (20 mm), a scavenger of the hydroxyl radical (OH.); a cocktail of SOD (50 U ml⁻¹) and catalase (1000 U ml⁻¹) to scavenge and H₂O₂; or the specific peroxynitrite scavenger melatonin (1 μm) (Reiter et al. 1998). Neither DMSO nor the combination of catalase and SOD significantly inhibited the contraction induced by peroxynitrite (DMSO, 85.2 ± 11.2 % of control, P > 0.05; catalase plus SOD, 70.8 ± 12.0 % of control, P > 0.05, n = 6 for each). However, 1 μm melatonin induced a clear and statistically significant reduction (82.1 ± 7.6 %; P < 0.001) in the response to peroxynitrite (Fig. 9C) without affecting the ACh-induced contraction (24.0 ± 3.1 mN before vs 24.8 ± 3.1 mN after melatonin treatment, P > 0.05, n = 4). This result strongly supports the concept that the contraction is mainly evoked by peroxynitrite per se, but not by another reactive oxygen species such as OH., or H₂O₂.

The peroxynitrite-evoked contraction was mainly associated with entry of extracellular Ca²⁺ through L-type channels, because the contraction induced by 10 μm peroxynitrite was almost blocked after treatment with 10 μm methoxyverapamil (92.0 ± 2.8 % of reduction of control response, n = 6, P < 0.001). Another intracellular pathway involved in peroxynitrite effect is the synthesis of leukotrienes from arachidonic acid, since NDGA reduced the response by almost 50 % (P < 0.05, Fig. 9C). In contrast, synthesis of prostaglandins and/or thromboxanes is probably not involved in this response since the cyclo-oxygenase inhibitor, indomethacin (10 μm), had only a residual non-significant effect (15.8 ± 13.2 % of inhibition, P = 0.142). In other systems, the primary target for peroxynitrite is the plasmalemmal Na⁺-Ca²⁺ exchanger (Chesnais et al. 1999). To address this possibility, we studied the effect of reduction of extracellular Na⁺, a manoeuvre that depresses this exchanger. However, in our experimental conditions this manipulation did not inhibit the response; rather, it increased the peroxynitrite-evoked contraction by more than 30 % (137.8 ± 15.9). Activation of PKC by peroxynitrite is not likely to occur since GF109203X (1 μm) pretreatment did not significantly modify the peroxynitrite-induced response (Fig. 9C).

DISCUSSION

The data reported here support the hypothesis that the effects of NO donors depend on the redox-activated form of NO generated. In the guinea-pig gall bladder, the NO⁺ donor SNP and exogenous peroxynitrite cause contraction, whereas NO. donors induce relaxation. Four lines of evidence support this conclusion: (1) SNP, a NO donor that generates NO⁺, induced gall bladder contraction; (2) authentic peroxynitrite induced gall bladder contraction; (3) DETA/NO, a NO donor that generates NO., induced gall bladder relaxation and indeed, SIN-1 in the presence of SOD and catalase, a situation in which it becomes a NO. donor, induced relaxation; and (4) SNP in a reducing environment, a situation in which SNP releases NO., induced gall bladder relaxation.

Numerous reports describe contradictory effects of NO or NO donors on the motility of the gastrointestinal system, where either contractile or relaxing effects have been reported (Sanders & Ward, 1992; Saha et al. 1993; Bartho & Lefebvre, 1995; Holzer et al. 1997; Hirano et al. 1997). Part of this discrepancy may derive from the complexity of the biological chemistry of NO and the heterogeneous states of the preparations studied. Although endogenous NO. is synthesized and released by NOS (Moncada et al. 1991), NO. undergoes a large variety of bioconversions, both in vitro and in vivo, leading to the generation of a number of different compounds. These include the nitrosonium cation, the nitroxyl anion, nitrosothiols, iron-nitrosyl complexes, peroxynitrite and nitrosotyrosin, with different effects (Stamler et al. 1992b). The formation of these NO derivatives as well as their stability and metabolism is dependent on the experimental environment.

Effects of NO⁺ on gall bladder contractility

The NO⁺ donor SNP caused a contraction of guinea-pig gall bladder muscle strips. This contraction was a direct effect of the compound on the smooth muscle, since it was not modified by pretreatment with the neurotoxin TTX, as described for the contractile effects of SNP in oesophageal longitudinal muscle (Saha et al. 1993). In guinea-pig ileal longitudinal muscle, contractile responses to NO have been related to stimulation of cholinergic fibres (Bartho & Lefebvre, 1994a); however, the SNP-induced contraction of gall bladder muscle was insensitive to atropine.

The contractile effects of SNP described in this study appear to be related to the generation of a redox form of NO since SNP had no effect on the gall bladder resting tone after prior exposure to light for 24 h, thereby excluding any by-products co-released with NO as responsible for the contraction. Moreover, the main redox form of NO yielded by SNP is likely to be NO⁺ because the contractile effects of SNP were abolished when it was tested in a reducing environment, a situation in which SNP induces release of NO. (Bates et al. 1991; Lipton et al. 1993) and no contraction was observed when NO. donors were tested.

Intracellular pathways stimulated by NO⁺

Reactions mediated by NO⁺ are generally referred to as nitrosation reactions. Nitrosation in aqueous phase can occur at -S, -N, -O and -C centres in organic molecules but physiological conditions favour thiol nitrosation (Stamler et al. 1992b). S-nitrosation of protein thiols (cysteine residues) may represent an important cellular regulatory mechanism, since S-nitrosated proteins may exhibit new functions or can lose their activities (Stamler et al. 1992a; Gopalakrishna et al. 1993).

Sequence analysis of protein tyrosine phosphatase (PTP) isoforms from several sources revealed the presence of a conserved catalytic domain with an essential cysteine residue. This residue has been shown to be the target for oxidants, which resulted in an inhibition of PTP activities (Monteiro et al. 1991). Similarly, this thiol group could be nitrosated by NO⁺ resulting in a loss of the PTP activity, thereby increasing tyrosine phosphorylation, which depends on the balance between PTP and protein tyrosine kinase (PTK) activity. Given that tyrosine phosphorylation is involved in smooth muscle contraction (Di Salvo et al. 1997; Alcón et al. 2000) inhibition of PTP by SNP might induce contraction (see Fig. 10). The findings reported here are in complete agreement with this model because SNP caused a contraction that was associated with an increase in tyrosine phosphorylation of several substrates. Preincubation with genistein and tyrphostin-B44, two chemically unrelated tyrosine kinase inhibitors, dramatically reduced both SNP-induced contractions and SNP-induced tyrosine phosphorylation. These data demonstrate that protein tyrosine phosphorylation is the key regulatory factor mediating SNP-induced contractions. These observations are consistent with the results of previous studies demonstrating either enhanced protein tyrosine phosphorylation (Hirano et al. 1997) or inhibition of PTP (Peranovich et al. 1995) in response to SNP.

Figure 10. NO⁺, NO. and OONO⁻ signalling pathways.

This illustration shows the inter-relationships between the usual NO redox forms in biological systems and the intracellular pathways through which they might mediate their effects on the contractility of the gall bladder smooth muscle. NO⁺, nitrosonium cation; NO., nitric oxide; OONO⁻, peroxynitrite; R-SH, intracellular thiols; PTP-SH, phosphotyrosine phosphatase; RS-NO, nitrosothiols; -Tyr-P, tyrosine phosphorylated proteins; LT, leukotrienes; DTT, dithiothreitol; sGC, soluble guanylate cyclase; , superoxide anion; SOD, superoxide dismutase; CAT, catalase.

Contractions of gall bladder strips elicited by SNP depend on extracellular calcium entry through voltage-operated calcium channels. This conclusion is based on the following experimental observations: (1) the contraction was abolished in Ca²⁺-free medium containing EGTA (1 mm); (2) SNP induces depolarization of gall bladder smooth muscle; and (3) the use of methoxyverapamil, an L-type calcium blocker, reduced the contractile and depolarizing response to SNP. Some studies, in different cellular models, suggest that tyrosine phosphorylation of rectifying K⁺ channels induces inhibition of such channels, leading to depolarization (Schultheiss & Diener, 1998). This could explain the depolarization observed by us, and the stimulation of calcium influx through L-type channels in response to SNP. On the other hand, it has been shown, in smooth muscle cells, that L-type Ca²⁺ channels are modulated by tyrosine kinase activity. Thus, tonic phosphorylation by tyrosine kinases maintains the Ca²⁺ channels in an available state for activation by depolarization (Liu et al. 1997) and stimulation of tyrosine phosphorylation enhanced peak Ca²⁺ currents (Hatakeyama et al. 1996). Consistent with this, we found that the spontaneous action potentials and the SNP-induced depolarization were abolished in gall bladder smooth muscle in the presence of genistein. Therefore, we cannot rule out that the effects observed by us are due to direct changes in L-type calcium channel activity induced by tyrosine phosphorylation. The small (8 %) methoxyverapamil-resistant component of the SNP-induced contraction could be due to Ca²⁺ entry through other channels or to Ca²⁺ depletion from intracellular stores.

The action of SNP on gall bladder muscle depends on stimulation of the arachidonic acid metabolism, similar to the reported action of agents that cause increases in tyrosine phosphorylation (Hollenberg, 1994). It has been reported that PLA₂ activity may also be regulated by direct changes in tyrosine enzyme phosphorylation (Lin et al. 1993), so that in the presence of SNP an increase in the activity of this enzyme could be expected. The reductions in SNP-induced contractions observed in the presence of the PLA₂ general blocker mepacrine and the sPLA-specific blocker 4BPB support this hypothesis. The AA produced by sPLA₂ in response to SNP may be metabolized to leukotrienes because the lipoxygenase inhibitor NDGA reduced the SNP contractile response, whereas the cyclo-oxygenase inhibitor indomethacin had no effect. In oesophagus, leukotrienes have been shown to contract smooth muscle cells through a PKC-sensitive mechanism (Kim et al. 1998). In our study, the lack of effects reported for the PKC inhibitor, and the total dependence on extracellular calcium, indicate that leukotrienes do not induce the SNP-induced contractile effects by stimulation of phosphoinositide metabolism.

Effects of NO. in gall bladder contractility

It is generally accepted that the relaxing effect of NO. in smooth muscle is mediated by cGMP. Our observation that the relaxation produced by DETA/NO was suppressed by ODQ is consistent with the biological reactions reported for NO. Nitric oxide, through nitrosylation reactions, readily forms complexes with transition metal ions, including those regularly found in metalloproteins. The most likely reaction of NO. to form stable nitrosyl adducts is with haem-containing proteins, such as guanylate cyclase, which in turn stimulates the formation of cGMP ultimately leading to cellular responses such as relaxation (see schematic diagram, Fig. 10) (Lincoln & Cornwell, 1993).

Recently, K⁺ channels have emerged as possible mediators of NO.-evoked relaxations. NO has been reported to activate BK or K_V currents, either directly (Bolotina et al. 1994; Yuan et al. 1996) or through the activation of guanylate cyclase (Carrier et al. 1997). In our study, neither TEA nor 4-AP, blockers of BK and K_V channels, respectively (Nelson & Quayle, 1995), affected the relaxing response to DETA/NO. This observation is consistent with previous reports of the failure of potassium channel blockers to inhibit the action of NO donors (Goyal & He, 1998; Ghisdal et al. 2000).

The current observation that the relaxing response to SNP in the presence of DTT was suppressed by ODQ but unmodified by K⁺ channel blockers, similar to the relaxing response to the NO donors, further supports the view that SNP generates large amounts of NO. in the presence of reducing agents such as thiols (Bates et al. 1991). Indeed, in our preparation a more dramatic relaxation was observed when higher reduction power was achieved by renewing DTT just prior to the addition of SNP. In agreement with this, SNP induced inactivation of PKC and 5-glyceraldehyde-3-phosphate dehydrogenase, related to NO. reactions, only when DTT was present (Gopalakrishna et al. 1993). Moreover, biological thionitrites (RS-NO) generated after the interaction between NO⁺ and RSH can yield NO. in the presence of nucleophiles such as electron-rich bases (as DTT) (Stamler et al. 1992a, b). The NO. that is generated could interact with metalloproteins, such as guanylate cyclase, to induce the relaxing effects (see schematic diagram in Fig. 10). This paradigm is supported by the fact that the contractile effect of SNP, probably due to thionitrosation of tyrosine phosphatases, was converted to a relaxation when DTT was added during the plateau of the SNP contractile response. Under these conditions, NO⁺ involved in nitrosation reactions with biological substrates such as protein phosphatases would yield NO., which, through nitrosylation reactions, would activate guanylate cyclase inducing gall bladder relaxation. Similar to our results, Lipton et al. (1993) were able to change the neuroprotective effect of SNP at the NMDA receptor, mediated by NO⁺ nitrosation of SH groups in the receptor, into neurotoxic effects just by adding either a cysteine- or an ascorbate-enriched reaction mixture to support conversion of NO⁺ to NO.

The relaxation observed with SNP at low doses on the gall bladder strips precontracted with cholecystokinin, and the relaxing component at 1 mm SNP unmasked by ODQ, were probably caused by the NO. that was present in the stock solution as a consequence of the reduction induced by the tissue or by photolysis of SNP. Spontaneous NO. release from SNP at light intensity similar to ambient laboratory light is produced at very low rate, but greater photolysis can result from higher illumination intensities (Bates et al. 1991). In our experimental conditions we increased the illumination of the organ bath, so an increase in the NO. output could be expected. On the other hand, reduction of SNP by reducing components in biological tissues can explain the spontaneous NO. release from SNP in both in vivo and in vitro conditions (Bates et al. 1991; Kowaluk et al. 1992). This could be the reason for the fact that only relaxing effects to SNP have been described in vascular preparations, where plasma membranes have been shown to exhibit substantial catalysis of NO. generation (Kowaluk et al. 1992). Our results clearly demonstrate that the guinea-pig gall bladder does not metabolize SNP to the same extent as vascular smooth muscle, since contractile effects in response to SNP are more evident. In our opinion, only when the dose of SNP is not very high does the spontaneous release of NO. (by photolysis and/or metabolism) predominate over NO⁺ generation, and if the tissue is precontracted the relaxing effects are evident.

Effects of peroxynitrite in gall bladder contractility

In the current study, peroxynitrite consistently produced a contraction of gall bladder muscle strips. These results are in agreement with previous reports that showed positive inotropic effect in cardiac fibres (Chesnais et al. 1999). However, relaxing effects for this redox form of NO have also been reported (Wu et al. 1994; Iesaki et al. 1999).

In pathological situations (e.g. ischaemia-reperfusion, heart failure and inflammation) NO can potentially exert beneficial or deleterious effects. Deleterious effects are generally related to excessive production of NO, often through inducible nitric oxide synthase (iNOS) with concurrent reactive oxygen species production and peroxynitrite formation (Grisham et al. 1999). In the current study, we found that, while NO. caused a relaxation of the gall bladder, peroxynitrite induced a contraction. Indeed, this contractile effect is consistent with the contractile effects reported for oxidants (H₂O₂, monochloramine) that are associated with pathological conditions in the gall bladder (Moummi et al. 1991), suggesting that peroxynitrite can also be a mediator for the motility alterations induced by inflammation in this tissue.

The contractile effect shown in our study was caused directly by peroxynitrite as it was nearly abolished by the pretreatment with the peroxynitrite scavenger melatonin. Although melatonin has been shown to inhibit ileal smooth muscle contraction in response to agonists through activation of apamin-sensitive channels (Reyes-Vazquez et al. 1997), a direct inhibitory effect of melatonin is not likely to be responsible for the reduction in the gall bladder response to peroxynitrite since melatonin did not reduce the effectiveness of the contractile agonist ACh. Furthermore, while OH. and H₂O₂ production from peroxynitrite has been demonstrated (Beckman & Koppenol, 1996), in our preparation, the release of neither OH. nor H₂O₂ accounts for peroxynitrite-induced contraction. This is because the gall bladder contraction was not modified by pretreatment with the OH. scavenger DMSO and the presence of SOD and catalase (to prevent and H₂O₂ output) did not reduce it.

The source for the increase in intracellular calcium responsible for the peroxynitrite-induced contraction seems to be tissue specific. For instance, in thymocytes, peroxynitrite elevated intracellular calcium through the activation of calcium efflux from the mitochondria and the inhibition of the ATP-dependent pump in the endoplasmic reticulum (Virag et al. 1999), whereas in skeletal muscle, peroxynitrite induced an inhibition of sarcoplasmic reticulum Ca²⁺-ATPase activity (Viner et al. 1999). In cardiac fibres the positive inotropic effect of this compound was related to the inhibition of the Na⁺-Ca²⁺ exchanger (Chesnais et al. 1999). In the gall bladder, the increase in calcium necessary for contraction seems to be due to the activation of extracellular calcium entry through L-type calcium channels as it was abolished when peroxynitrite was tested in calcium-free medium or in the presence of methoxyverapamil. This is consistent with results from studies of peroxynitrite-induced responses in cardiac myocytes (Ishida et al. 1996).

Little is known about the intracellular signals mediating peroxynitrite effects. Our work provides some insights into the mediators of peroxynitrite response, showing that an increase in the metabolism of arachidonic acid can mediate, at least in part, peroxynitrite effects. This is consistent with the increase in arachidonic acid release reported for peroxynitrite treatment in PC12 cells (Guidarelli et al. 2000). The arachidonic acid produced was metabolized to leukotrienes, since the inhibition of 5-lypoxygenase, but not cyclo-oxygenase, reduced peroxynitrite-induced gall bladder contraction.

Under in vivo conditions, low amounts of peroxynitrite are formed continuously. This process can be mimicked in experimental systems with the and NO.-releasing compound SIN-1. Peroxynitrite generated in situ from SIN-1 has been shown to affect biological targets in nearly the same manner as bolus addition of authentic peroxynitrite (Chesnais et al. 1999). However, we found converse effects for SIN-1 and peroxynitrite, with peroxynitrite acting as a contractile agent and SIN-1 as a relaxing agent. The relaxing effects of SIN-1 are potentiated by SOD. This observed potentiation was attributed to an increase in the relative rate of NO. production, as it could not be inactivated through its reaction with (Lipton et al. 1993). These results are also consistent with the contractile effect for peroxynitrite. Thus, in the gall bladder, where SOD and catalase activity has been described (Cullen et al. 1999), SIN-1 would release both NO and peroxynitrite, with the NO effects predominant. When peroxynitrite formation is blocked, NO.-mediated relaxation would be more evident. We also report here that catalase increased the potentiation by SOD of the relaxing effect of SIN-1. Hence, not only does the SOD-induced response have a relaxing component due to NO., but it also has a contractile one related to H₂O₂ yield. These results are consistent with a previous report (Gergel et al. 1995) that describes increased cytotoxicity of SIN-1 to HepG2 cells in the presence of SOD that is related to the release of H₂O₂.

The mechanism of the relaxation in response to SIN-1 appears to involve stimulation of the soluble form of guanylate cyclase, since it was inhibited by ODQ, which reinforces the idea that SIN-1 releases mainly NO. in our preparation. In fact, in the presence of SOD and catalase, SIN-1 response is also mediated through stimulation of guanylate cyclase.

The present study clearly demonstrates that there is an inter-relationship between the different redox forms of NO and that depending on the tissue and/or the experimental conditions either endogenous or exogenous NO can be converted into different forms of NO which have, in some cases, converse effects. Further studies are needed to test whether these complex interactions are also related to the cytotoxic effects shown by NO in pathological conditions.