Available online at www.sciencedirect.com

www.elsevier.com/locate/yniox

Nitric Oxide 18 (2008) 157–167

Differential role of S-nitrosylation and the NO–cGMP–PKGpathway in cardiac contractility

Daniel R. Gonzalez, Ignacio C. Fernandez, Pablo P. Ordenes, Adriana V. Treuer,Gisela Eller, Mauricio P. Boric *

Departamento de Ciencias Fisiologicas, Facultad de Ciencias Biologicas, Pontificia Universidad Catolica de Chile, Alameda 340, P.O. Box 114D,

Santiago, Chile

Received 1 March 2007; revised 5 September 2007Available online 1 October 2007

Abstract

The role of nitric oxide (NO) in cardiac contractility is complex and controversial. Several NO donors have been reported to causepositive or negative inotropism. NO can bind to guanylate cyclase, increasing cGMP production and activating PKG. NO may alsodirectly S-nitrosylate cysteine residues of specific proteins. We used the isolated rat heart preparation to test the hypothesis that the dif-ferential inotropic effects depend on the degree of NO production and the signaling recruited. SNAP (S-nitroso-N-acetylpenicillamine), aNO donor, increased contractility at 0.1, 1 and 10 lM. This effect was independent of phospholamban phosphorylation, was not affectedby PKA inhibition with H-89 (N-[2((p-bromocinnamyl)amino)ethyl]-5-isoquinolinesulfonamide), but it was abolished by the radicalscavenger Tempol (4-hydroxy-[2,2,4,4]-tetramethyl-piperidine-1-oxyl). However, at 100 lM SNAP reduced contractility, effect reversedto positive inotropism by guanylyl cyclase blockade with ODQ (1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one), and abolished by PKGinhibition with KT5823, but not affected by Tempol. SNAP increased tissue cGMP at 100 lM, but not at lower concentrations.Consistently, a cGMP analog also reduced cardiac contractility. Finally, SNAP at 1 lM increased the level of S-nitrosylation of variouscardiac proteins, including the ryanodine receptor. This study demonstrates the biphasic role for NO in cardiac contractility in a givenpreparation; furthermore, the differential effect is clearly ascribed to the signaling pathways involved. We conclude that although NO ishighly diffusible, its output determines the fate of the messenger: low NO concentrations activate redox processes (S-nitrosylation),increasing contractility; while the cGMP–PKG pathway is activated at high NO concentrations, reducing contractility.� 2007 Elsevier Inc. All rights reserved.

Keywords: Nitric oxide; Nitric oxide synthase; NOS; Rat heart; Langendorff; Tempol; SNAP

Nitric oxide (NO) plays a role in almost all aspects ofcardiac function: contractility [1] heart rate [2] and remod-eling [3]. Concerning contractility, the effects of exogenousNO appears to be biphasic [4]. In general, when used athigh concentrations (near 100 lM of a NO donor) theobserved response is a depression of contractility [5–7]but when used at lower concentrations, increases incontractility have been reported [8–11]. A clear cut conclu-sion cannot be drawn from those reports because differentpreparations from different species, as well as different NOdonors and experimental protocols have been used.

1089-8603/$ - see front matter � 2007 Elsevier Inc. All rights reserved.

doi:10.1016/j.niox.2007.09.086

* Corresponding author. Fax: +56 562 222 5515.E-mail address: mboric@bio.puc.cl (M.P. Boric).

Furthermore, the mechanisms underlying NO effects oncardiac contractility still remain controversial and thiscan be associated with the fact that NO biology is extre-mely sensitive to factors such as the chemical nature andthe kinetics of NO delivery of the donors, and the redoxstatus of the preparation used. In a physiological context,a biphasic behavior for NO could reflect the role that thedifferent nitric oxide synthase (NOS) isoforms (NOS1 andNOS3) play in the regulation of contractility [12]; NOS1enhancing contractility and NOS3 regulating it negatively.These contrasting effects have been related to the specificsubcellular localization of the two NOS isoforms [1].

Classically, the synthesis of NO was related to the activa-tion of the enzyme guanylate cyclase, after binding to its

158 D.R. Gonzalez et al. / Nitric Oxide 18 (2008) 157–167

heme group, leading to an increase in the rate of conversionof GTP to cGMP [13]. Among other features, cGMP is ableto activate protein kinase G (PKG). In the heart, targets forPKG include troponin I [14], L-type calcium channel [15]and cyclic nucleotides phosphodiesterases [16] with theirrespective impact on the contractile status of the myocar-dium. In addition, S-nitrosylation of cysteine residues hasemerged as an important feature of NO signaling. Throughthis post-translational modification, NO is able to regulatethe function of enzymes, ion channels and structural proteins[17]. In the heart, S-nitrosylation has been suggested as amechanism by which NO is able to modulate the ryanodinereceptor (RyR2) and the L-type calcium channel, two keyproteins partaking in excitation-contraction coupling [18].

We postulate that low levels of NO can activate the S-nitrosylation pathway, leading to an increase in heart con-tractility, independently of cGMP; whereas higher levels ofNO activate the cGMP–PKG pathway causing a reductionin contractility, indeed overruling the former pathway. Toassess this hypothesis, we used different concentrations ofSNAP, a S-nitrosothiol and a NO donor, in the isolatedrat heart preparation, in the absence or presence of differ-ent blockers of the cGMP pathway. Consistent with thehypothesis, we show that SNAP can activate S-nitrosyla-tion in a redox-sensitive manner at low concentrations,and cGMP synthesis at higher concentrations, with oppo-site roles concerning cardiac contractility. The increase inS-nitrosylation is associated with an increase in contractil-ity and the increase in intracellular cGMP is correlatedwith a decrease in contractility in the isolated heart throughactivation of PKG.

Experimental procedures

Animals

Male Sprague–Dawley rats, weighting 290–300 g. were obtained fromthe Animal Facility of the Faculty of Biological Sciences of the PontificiaUniversidad Catolica de Chile. All protocols were approved by the Insti-tutional Bioethics Committee of the Pontificia Universidad Catolica deChile and conformed to the Guide for the Care and Use of Laboratory Ani-

mals published by the US National Institutes of Health (NIH publicationNo. 85–23, revised 1996).

Reagents

8Br-cGMP, H-89 (N-[2((p-bromocinnamyl)amino)ethyl]-5-isoquino-linesulfonamide), isoproterenol hydrochloride (ISO), ODQ (1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one), SNAP (S-nitroso-N-acetylpeni-cillamine), and Tempol (4-hydroxy-[2,2,4,4-tetramethyl-piperidine-1-oxyl]), were purchased from Calbiochem (La Jolla, CA). Unless statedotherwise, the rest of the reagents were obtained from Merck (Darmstadt,Germany).

Isolated heart preparation

Rats were anesthetized with a mixture of ketamine (90 mg/kg) and xyla-zine (10 mg/kg) and pre-medicated with 1000 UI heparin i.p. Hearts wererapidly excised and perfused through the aorta with Krebs-Henseleit buffer(equilibrated with a gas mixture of 95% O2 and 5% CO2 at 37�), using a peri-staltic pump (Gilson Miniplus 3, France). A polyvinyl chloride balloon con-

nected to a pressure transducer by a polyethylene P-50 cannula (ClayAdams-Becton Dickinson, Sparks, MD), was placed through the left atriumand mitral valve into the left ventricle. The balloon was filled with saline todetermine isovolumetric intraventricular pressure. Perfusion flow wasincreased gradually until reaching 10 ml/min and kept constant throughoutin order to avoid the Gregg effect [19]. The hearts were placed in a heatedchamber and paced at 360 beats/min with platinum electrodes, using a Grassstimulator (pulses of 5 V, 1 ms). Left ventricular pressure (LVP), and coro-nary perfusion pressure (CPP) were measured continuously with pressuretransducers (P23XL, Ohmeda Instruments, Madison, WI, USA), and digi-tized (Chart, ADI Instruments, New South Wales, Australia), to obtainthe rate of change in left ventricular pressure (dP/dt). Minimal diastolicpressure was held constant at 5–10 mm Hg during the experiment.

NOx measurements

The amount of NO and NO2� (NOx) in SNAP solutions was deter-

mined by chemiluminescence as described by Figueroa et al. [20]. SNAPsolutions were freshly prepared in gassed Krebs-Henseleit buffer, in iden-tical conditions as those used in the Langendorff preparation. After30 min, an aliquot of this solution was injected into the reaction chamberof an NO Analyzer (Sievers 280), filled with acetic acid and potassiumiodine. In these conditions, nitrites, but not nitrates are reduced to NO,therefore the content of NO plus nitrite is measured.

Phospholamban phosphorylation

In some experiments, the content of phospholamban, both total pro-tein and the form phosphorylated at ser16 (phospho-phospholamban)were assessed by Western blotting of heart homogenates. Briefly, a portionof the left ventricle (�200 mg) was excised at the specified time of the stim-ulation protocol, and quickly homogenized by an Ultraturrax in 1 mL ofcold Tris–HCl buffer (100 mM pH 7.4) containing antiproteases (5 mMEGTA, 1 lg/mL Aprotinine, 1 mM benzamidine, 10 lg/mL leupeptin,10 lg/mL pepstatin-A, 2 mM PMSF, 200 lg/mL SBTI). The homogenatewas centrifuged 30 s at 100g and the supernatant was mixed with Laemli’sbuffer and resolved in 12% SDS–PAGE. Gels were blotted on nitrocellu-lose and tested sequentially with polyclonal anti phospho Ser16-phospho-lamban (Upstate, Lake Placid, NY) and monoclonal anti phospholambanantibodies (Abcam Plc, Cambridge, UK). The intensity of the signal wasevaluated using the Image J program (NIH public domain software).

S-Nitrosylation assay

Protein S-nitrosylation was assessed by using the biotin switch method[21,22], according to the instructions of a commercially available kit (Nitrog-lo kit, Perkin Elmer, Boston, MA, USA). Briefly, a left ventricular sample(�200 mg) was excised and homogenized, in assay buffer with antiproteasesas indicated above. The homogenate was centrifuged 10 min at 750g at 4 �C,and in the supernatant, the free sulfhydryls groups were blocked with methylmethanethiosulfonate. Nitrosylated proteins present in the cardiac homog-enate were reduced (1.6 mM ascorbic acid) and the newly formed thiols werereacted with biotin-HPDP. After this, proteins were resolved in 5%, 7.5% or12% SDS–PAGE and the biotinylated proteins were visualized using ananti-biotin antibody (Cell Signaling, Beverly, MA), or HRP–streptavidin(Sigma Chemical. St. Louis, MO). Specific S-nitrosylation of cardiac RyRreceptor was evaluated in separate assays, in which proteins were resolvedin 3–8% Tris–acetate gradient gels and after detection of biotin, the mem-brane was stripped and reincubated with a monoclonal antibody againstRyR2 (Affinity Bioreagents, Golden, CO, USA).

cGMP assay

In one series, immediately at the end of treatment with SNAP, heartswere stopped by immersion in ice-cold TCA 10% (w/v), quickly weighedand homogenized with an Ultraturrax in 3 mL of TCA. After centrifuga-

D.R. Gonzalez et al. / Nitric Oxide 18 (2008) 157–167 159

tion at 10,000g for 30 min, the supernatants were collected and extractedfour times with five volumes of ether. The remaining ether was evaporatedand the cGMP content was determined by RIA as described [23].

Experimental protocols

The basic protocol consisted in �30 min equilibration followed by anexperimental period no longer than 60 min. All drugs were applied in theperfusion buffer. Stimulation with SNAP or isoproterenol, consisted in a2-min drug perfusion. Inhibitors (ODQ, KT5823, H-89) and mediators(8Br-cGMP), were applied for at least 15 min prior to stimulation. Heartsreceived a maximum of four stimuli, usually separated by a 15-min recov-ery period.

Data analysis

The hemodynamic values LVPmax, (dP/dt)max, (dP/dt)min, and CPP,were averaged in 1-min periods prior to (baseline) and during the stabi-lized phase observed during drug perfusion (stimulus). Responses arereported either as absolute values (mm Hg/s) or as the percentage ofchange relative to baseline. Data are presented as means ± SEM. Concen-tration-response experiments were analyzed by one way ANOVA, usingthe Newman–Keuls post hoc test. Student’s t-test, either paired orunpaired as appropriate, was applied for experiments comparing singledose effects. Ratio values were submitted to arcos transformation beforestatistical analysis. A value of p < 0.05 was considered to be statisticallysignificant.

Results

Effects of SNAP on cardiac contractility

Perfusing the heart with concentrations of SNAP 0.1,1 and 10 lM produced significant increments in cardiacperformance, as evidenced by parallel changes in left ven-tricular pressure (LVPmax), contractility (dP/dt)max, andventricular relaxation (dP/dt)min, associated with a mod-erate reduction in coronary perfusion pressure (CPP)(Table 1). In contrast, 100 lM SNAP caused the oppo-site, a sustained reduction in cardiac contractility, witha similar degree of coronary dilatation (Table 1). In allcases, a slight delay (10–20 s) occurred between thebeginning of SNAP perfusion and changes in contractil-

Table 1Effect of SNAP on hemodynamic variables of the isolated heart

n LVPmax (mm Hg) (dP/dt)m

Basal 13 60.4 ± 2.0 1783 ±SNAP 0.1 lM 76.9 ± 3.7 ** 2311 ±% Change 28.3 ± 6.5 27.9 ±

Basal 17 63.8 ± 5.8 2046 ±SNAP 1 lM 79.5 ± 5.8 ** 2608 ±% Change 28.8 ± 6.1 30.5 ±

Basal 16 66.8 ± 6.0 2109 ±SNAP 10 lM 74.1 ± 5.2* 2358 ±% Change 13.2 ± 3.8 13.1 ±

Basal 20 62.2 ± 3.8 1737 ±SNAP 100 lM 51.5 ± 3.4** 1453 ±% Change �17.6 ± 2.1 �18.9 ±

*p< 0.05, **p< 0.001 vs. respective basal (paired Student t-test).

ity, which lasted approximately for the duration of thestimulus in most experiments (Fig. 1). A transient maxi-mal (�30–60 s), followed by a less prominent but sus-tained elevation was frequently observed duringapplication of 10 lM SNAP (not shown). Perfusion withvehicle alone (DMSO 1:1000) did not affect any of thehemodynamic variables (not shown). Since all hemody-namic variables responded symmetrically, we used (dP/dt)max for analysis in the rest of this work. Averagingthe net change in (dP/dt)max as a function of SNAP, abiphasic effect of this NO donor was demonstrated inthe same preparation and under similar experimentalconditions (Fig. 2a).

The range of concentrations of NO generated bySNAP was determined by measuring the content ofNOx ðNOþNO2

�Þ in solutions prepared in Krebs-Henseleit solution bubbled with 95% O2– 5% CO2, thesame as when used in the isolated heart. As shown inFig. 2b, the NOx content was directly related to the con-centration of SNAP (containing 35 ± 13, 56 ± 9,196 ± 24 and 1,252 ± 156 nM NOx, at 0.1, 1. 10 and100 lM SNAP, respectively). In addition, we determinedthat NOx content was roughly proportional to the timeelapsed after preparing the solution, therefore, we canestimate that the rate of NO release was �40–50 pmol/ml/min at 100 lM SNAP, and �6–7 pmol/ml/min at10 lM SNAP.

To discern which signaling pathways were involved inthe contrasting effects of SNAP on cardiac contractility,we assessed the concentration of cGMP by RIA in the car-diac tissue exposed to the different concentrations of theNO donor. Baseline cGMP content in control hearts, nottreated with SNAP, was 12.1 ± 2.8 pmol/g wet tissue. After2-min perfusion with SNAP 0.1, 1, 10 and 100 lM thecGMP content was 9.6 ± 2.4, 14.8 ± 3.8, 29.7 ± 12.1 and75.2 ± 13.0 pmol/g wet tissue, respectively, (Fig. 2c). Only100 lM SNAP caused a significant increment of this intra-cellular messenger, although a tendency to increase cGMPwas also observed with 10 lM SNAP.

ax (mm Hg/s) (dP/dt)min (mm Hg/s) CPP (mm Hg)

283 �1076 ± 74 77.2 ± 5.8375 * �1452 ± 158 * 64.2 ± 6.9*

7.7 33.6 ± 9.2 �16.6 ± 5.9

243 �1226 ± 243 71.7 ± 5.2291** �1567 ± 147** 57.4 ± 3.9*

6.7 31.6 ± 5.4 �16.8 ± 0.9

245 �1260 ± 110 73.5 ± 6.1264* �1410 ± 98* 60.3 ± 6.6*

2.7 13.7 ± 3.2 �17.3 ± 2.0

217 �1151 ± 66 80.2 ± 7.4211** �941 ± 69** 69.0 ± 6.5*

2.2 �19.2 ± 2.3 �12.7 ± 2.3

Fig. 1. Effects of SNAP on cardiac contractility. Representative recordings of cardiac responses observed prior to and after perfusion with low and highconcentrations of SNAP. Traces depict maximal left ventricular pressure (LVPmax) and the rate of change in LVP as a function of time (dP/dt) in isolatedrat hearts. Horizontal bars indicate the 2-min period of perfusion with SNAP 1 lM (a and b), or SNAP 100 lM (c and d). Symmetrical changes inmaximal and minimal dP/dt, closely matched variations in LVPmax. Note a slight transient increment in contractility at the beginning of perfusion with100 lM SNAP.

160 D.R. Gonzalez et al. / Nitric Oxide 18 (2008) 157–167

The negative inotropic effects of SNAP are dependent on theguanylate cyclase–cGMP–PKG pathway

The above presented results suggest that the negativeinotropic effect elicited by 100 lM SNAP may be due togeneration and/or accumulation of cGMP as intracellularmediator. To assess this possibility, we performed addi-tional experiments in the presence of 10 lM ODQ, a spe-cific guanylate cyclase blocker at this concentration. After15-min perfusion with ODQ alone, cardiac contractility

was slightly but significantly reduced (net change in (dP/dt)max was �12.6 ± 2.4%, n = 7) (Fig. 3). In the presenceof this blocker, the positive inotropic effect elicited by1 lM SNAP was slightly enhanced as compared with theresponse elicited SNAP in control conditions (net increasein (dP/dt)max of 36.9 ± 5.1 vs. 25.3 ± 3.2%, n = 6) (Fig. 3).In contrast, 10 lM ODQ not only prevented the negativeinotropism induced by 100 lM SNAP, but turned it intoa positive inotropism (Fig. 3), similar to that attained with1–10 lM SNAP in absence of the blocker (Fig. 2a). In

Fig. 2. SNAP increases or decreases cardiac contractility. Negativeinotropism is associated with cGMP production. (a) Biphasic effect ofSNAP on cardiac contractility. Black circles depict the percentage ofchange of (dP/dt)max relative to the control condition (c), attained during2-min perfusion with different concentrations of SNAP. The number ofindependent preparations in each group is shown in parenthesis. (b)Concentration of NO plus nitrite (NOx) in the perfusion media. Krebs-Henseleit solutions with different concentrations of SNAP were analyzedto detect its NO content by chemiluminescence. Under mild reducingconditions, the measured [NO]x value corresponds to authentic NO plusaccumulated nitrite present in the solution. Correcting by the time elapsedsince preparation, the estimated rate of NO release at 10 and 100 lMSNAP was approximately 6.5 ± 0.8 and 41 ± 5 pmol/ml/min, respectively.�p < 0.05 vs. all other concentrations (one way ANOVA, Newman–Keulspost hoc test). (c) Cardiac cGMP content in response to SNAP. Heartswere perfused for 2-min with the indicated concentrations of SNAP, as inthe experiments shown in (a) for hemodynamic determinations. Tissuelevels of cGMP (pmol/mg tissue) were measured by RIA. The last columnrepresents hearts perfused during 15 min with 10 lM ODQ prior toapplication of 100 lM SNAP. *p < 0.05 vs. control (one way ANOVA,Newman–Keuls post hoc test).

Fig. 3. High concentrations of SNAP require guanylate cyclase activationto cause negative inotropism. Contrasting changes in heart contractility,induced by low (1 lM) and high (100 lM) concentrations of SNAP wereassessed in the presence of guanylyl cyclase blocker, ODQ (10 lM).15-min perfusion with ODQ caused a moderate but significant reduction in(dP/dt)max. In these conditions, the positive inotropic effect of 1 lM SNAPwas slightly enhanced, whereas the negative inotropic effect of 100 lMSNAP was reverted to positive inotropism. The number of experiments isshown in parenthesis inside the bars. All conditions were significantlydifferent from baseline. *p < 0.05, **p < 0.0001 vs. respective control.

D.R. Gonzalez et al. / Nitric Oxide 18 (2008) 157–167 161

hearts perfused with 10 lM ODQ, tissue cGMP was almostundetectable despite stimulation with 100 lM SNAP(Fig. 2c), demonstrating that this inhibitor was effectiveat the concentration used.

Considering overall the experiments with and withoutODQ, a patent correspondence between negative inotro-

pism and cGMP accumulation was recognized, as well anassociation between positive inotropism and lack of cGMPaccumulation (Figs. 2 and 3). To further explore the signal-ing downstream of cGMP associated with the negative ino-tropic response induced by high concentrations ofexogenous NO, we used KT5823, an inhibitor of PKG.In the presence of KT5823 (0.5 lM), SNAP (100 lM)was unable to reduce cardiac contractility (Fig. 4). How-ever, during PKG blockade, 100 lM SNAP did not causepositive inotropism, at difference to the response observedduring guanylate cyclase blockade with ODQ (Fig. 3).

To corroborate that the negative inotropic effects ofSNAP are mediated by cGMP, we applied 8Br-cGMP, apermeable analog of cGMP, in increasing concentrationsto explore whether this intracellular mediator could mimicthe response to the NO donor. As shown in Fig. 5a,100 lM 8Br-cGMP induced a decrease in (dP/dt)max com-paratively similar to that elicited by 100 lM SNAP. Thedepressant effect of the cGMP analog was dose-dependent;however, none of the concentrations used caused anincrease in contractility (Fig. 5b).

Taken together, these results indicate that high concen-trations of NO donors, and presumably endogenous NO,reduce cardiac contractility by a pathway involving guany-late cyclase–cGMP and PKG; whereas the positive inotro-pic effect of low concentrations of NO seems independentof a major accumulation of cGMP.

The positive inotropic effects of SNAP are independent of the

cAMP–PKA pathway

We explored the possibility that the positive inotropiceffects induced by low concentrations of SNAP may

Fig. 4. High concentrations of SNAP require PKG to cause negativeinotropism. (a and b) Representative experiments show the change incontractility induced by SNAP 100 lM and lack of effect of the samestimulus after 10-min treatment with the PKG inhibitor, KT5823(0.5 lM). Only the positive section of the dP/dt trace is shown. (c) Thebar graph summarizes the abolishing effect of PKG inhibition withKT5823 on the negative inotropism induced by SNAP 100 lM. *p < 0.05vs. baseline.

Fig. 5. cGMP reduces, but does not increase cardiac contractility. (a)Representative experiment showing the cardiac response to a 2-minperfusion with 8-Br-cGMP (100 lM). Note the prolonged effect of thishigh concentration of the cell-permeable nucleotide analog. Only thepositive section of the dP/dt trace is shown. (b) Summary of cardiaccontractility responses elicited by increasing concentrations of 8-Br-cGMP, expressed as percentage of change in (dP/dt)max relative tobaseline. The number of independent preparations in each group is shownin parenthesis. *p < 0.05 vs. baseline.

162 D.R. Gonzalez et al. / Nitric Oxide 18 (2008) 157–167

involve activation of the classical cAMP–PKA pathway.To achieve this aim, we used two experimental approachescomparing the effects of SNAP 1 lM and a b-adrenergicagonist, isoproterenol (10 nM).

In a first experimental series we studied the effect of H-89, a PKA inhibitor, on the inotropic response induced bySNAP and isoproterenol. Fifteen-minutes perfusion withH-89 (0.5 lM) did not change basal (dP/dt)max, but treat-ment with this PKA inhibitor caused a significant reduc-tion in the inotropic response elicited by 10 nMisoproterenol (from 100.8 ± 18.2% to 86.6 ± 10.6%). Incontrast, H-89 did not affect the response induced by 1lm SNAP (from 20.7 ± 2.4% to 21.2 ± 5.8%) (Fig. 6a–c).

Secondly, we performed additional experiments todetermine the degree of phospholamban (PLB) phosphory-lation in cardiac tissue. After 2-min perfusion with vehicle,10 nM isoproterenol or 1 lM SNAP, hearts were homoge-

nized and submitted to Western blot to detect total PLBand PLB phosphorylated at ser-16. As shown in Fig. 6dand e, treatment with isoproterenol doubled the degree ofphosphorylation of PLB, whereas SNAP had no effect onthis variable. Taken together, these experiments providestrong evidence against a major participation of thecAMP–PKA pathway in the positive inotropic responseelicited by low concentrations of SNAP.

S-Nitrosylation

The previous results suggest that the increase in contrac-tility induced by low concentrations of SNAP involvedmechanisms different from the cGMP–PKG or cAMP–PKA pathways.

Therefore, we performed two experimental approachesto investigate the role of redox-related processes that couldtake place in the heart during perfusion with SNAP, partic-ularly S-nitrosylation, which has emerged as an importantmechanism by which NO regulates several biologicalprocesses.

First, we evaluated the effect of preventing the release oraccumulation of superoxide during the response elicited bySNAP. Superoxide is known to be required under physio-logical conditions for NO-mediated redox reactions[9,24]. For this, we assessed the effect of SNAP 1 lM and

Fig. 6. Positive inotropism induced by low concentrations of SNAP is independent of the cAMP–PKA pathway. (a and b) Cardiac contractility, expressedas (dP/dt)max before and after 2-min stimulation with the b-adrenergic agonist isoproterenol (ISO, 10 nM), or SNAP 1 lM, in control conditions or after20-min perfusion PKA inhibitor H89 (0.5 lM). (c) Bars denote the difference in the net inotropic effect exerted by ISO (10 nM) or SNAP (1 lM) beforeand after treatment with H89 (data from experiments shown in a and b). The number of experiments in each group is shown inside the bars. *p < 0.05 vs.control conditions. (d) Representative Western blot showing phospholamban phosphorylated at ser16 (p-Ser16PLB) and total phospholamban (PLB) inhearts treated with vehicle (control), SNAP 1 lM, or ISO 10 nM. Bands at 7 and 35 KDa represent PLB monomers and pentamers, respectively. (e)Densitometric analysis of p-Ser16-PLB over total PLB ratio. In each experiment, both (monomeric and pentameric) forms of the regulatory protein wereadded for calculation. *p < 0.05 vs. the other two conditions, ANOVA Neuman–Keuls test of arccos transformed values.

Fig. 7. Tempol abolishes the positive inotropism caused by SNAP, but itdoes not affect negative inotropism. Contrasting changes in heartcontractility, induced by low (1 lM) and high (100 lM) concentrationsof SNAP were assessed in the presence of the radical scavenger Tempol. Inmost cases, the hearts were challenged with the pulse of SNAP in controlconditions, then perfused with Tempol (100 lM) for 15 min andre-stimulated with the same concentration of SNAP. Bars depict varia-tions in contractility, expressed as percentage of change of (dP/dt)max

relative to baseline. The number of experiments in each group is shown inparenthesis. *p < 0.05 vs. control.

D.R. Gonzalez et al. / Nitric Oxide 18 (2008) 157–167 163

100 lM in the presence or absence of Tempol, a radicalscavenger, that has been shown to inhibit the nitrosylationreaction in vitro [25,26]. In hearts treated with Tempol(100 lM, 15 min), the increase in contractility induced by1 lM SNAP was abolished, but the radical scavenger didnot alter the negative inotropic effect elicited by 100 lMSNAP or baseline contractility (Fig. 7). These results sug-gest that superoxide blockade specifically preventedredox-related effects but not those mediated by the cGMPcascade.

Secondly, we assessed whether SNAP induced S-nitrosy-lation of proteins in the perfused hearts. Using the biotinswitch method, we analyzed the pattern of proteins con-taining S-nitrosylated cysteines in homogenates of heartstreated or untreated with 1 lM SNAP, concentration thatincreased contractility. As shown in Fig. 8a–e, 2-min perfu-sion with this concentration of SNAP clearly increased thedegree of S-nitrosylation of a broad range of proteins inheart tissue, including the protein band that migrates atthe same location were RyR2 had been previously deter-mined by Western blot (�560 KDa).

To get more insight concerning the mechanisms bywhich NO might increase cardiac contractility, in addi-tional experiments we specifically analyzed the degree ofS-nitrosylation of the RyR2 by determining the biotin overprotein signal ratio in the same membranes. SNAP (1 lM)caused a clear increase in the level of S-nitrosylation,whereas the content of total RyR2 protein was similar in

control and SNAP treated hearts (Fig. 8f and g). The incre-ment in the S-nitrosylation signal at this precise locationwas corroborated using an anti-biotin antibody and alsousing peroxidase-conjugated streptavidin (Fig. 8f).

Fig. 8. Protein S-nitrosylation in the isolated heart. Matched homogenates of hearts treated with vehicle or SNAP 1 lM during 2-min, as in theexperiments to study hemodynamic changes, were analyzed for S-nitrosylation utilizing the biotin switch method. Proteins were resolved in gels at 7.5%(a), or 12% (b), and tested using an anti-biotin antibody. The amount of S-nitrosylated proteins in the 20–60 KDa range was quantified in (c). In (d),proteins were resolved in 5% gel to assess the amount of S-nitrosylation at the migration position corresponding to the RyR2 receptor. The signal intensitywas quantified in (e). In (f), proteins were resolved in 3–8% gradient Tris–acetate gels, blotted and the membrane was tested first to measureS-nytrosylation with anti-biotin antibody (left panel) or HRP-conjugated streptavidin (right panel). The membranes were then stripped and tested withanti-RyR2 antibody. The densitometric analysis of the intensity ratio between the anti-biotin and the RyR2 signal is shown in (g). *p < 0.05 vs. control.

164 D.R. Gonzalez et al. / Nitric Oxide 18 (2008) 157–167

Discussion

This study corroborates the biphasic role for NO in car-diac contractility. We show that in the same preparation,under similar experimental conditions, application of theNO donor SNAP increases or decrease the index of cardiaccontractility (dP/dt)max just depending on the concentra-tion used (Fig. 2). Changes in maximal LVP and relaxationrate (dP/dt)min followed the same pattern (Table 1). Theseresults expand and strengthen previous information gath-ered in different species: guinea pig [5], human [6], rat[8,9,11], cat [7], dog [10] and different preparations:in vivo [10]; isolated heart [9]; muscle strips [7,27], and iso-lated cardiomyocytes [5,8,11,14], indicating that differentNO donors can reduce cardiac contractility at concentra-tions around 100 lM or higher, but they do increase con-tractility at concentrations around 0.1–10 lM. Setting

aside probable differences in the threshold or sensitivityfor the change in contractility, our results rule out the pos-sibility that the differential effect of NO donors is an arti-fact due to differences among species, or cardiacpreparations.

It is known that NO is able to trigger at least two path-ways: cGMP generation and redox-related protein modifi-cations. Our results support the hypothesis that activationof the cGMP–PKG pathway leads to a reduction in cardiaccontractility, whereas the increase in contractility is medi-ated by a cGMP-independent mechanism. First, we foundthat perfusion with low SNAP concentrations (0.1 and1 lM) increased basal contractility in the absence of anydetectable rise in tissue cGMP. On the contrary, 100 lMSNAP reduced contractility, associated with a significantincrement in cardiac cGMP production. Tissue cGMPaccumulation agrees with the measured values of NOx in

D.R. Gonzalez et al. / Nitric Oxide 18 (2008) 157–167 165

SNAP solutions, containing 195 and 1250 nM at 10 and100 lM, respectively, and the reported values of EC50 forguanylate cyclase activation by NO, which range between80 and 250 nM [28]. The involvement of the cGMP–PKGsignaling in the negative inotropic response elicited bySNAP was confirmed by the blockade of guanylate cyclasewith ODQ (Fig. 3) and PKG with KT5823 (Fig. 4). Fur-thermore, the same response could be obtained by directactivation of this pathway with administration of exoge-nous cGMP (Fig. 5).

On the other hand, our results confirm that ODQ treat-ment alone reduces basal contractility (Fig. 3), as previ-ously reported in the rat heart [29]. Thus, some level ofcGMP is required for normal contractility. The finding thatPKG inhibition just abolished, but did not revert the neg-ative inotropism induced by 100 lM SNAP may indicatethat in addition to PKG activation, cGMP modulates theactivity of cyclic nucleotide phosphodiesterases, perhapsaltering the classical cAMP–PKA signaling [30]. Indeed,in isolated rat cardiomyocytes, low increases of cGMPimprove cardiac contractility, associated with increases incAMP [8].

The positive inotropic response induced by low concen-trations of SNAP seems to be independent of modificationsin the classical cAMP–PKA pathway regulating cardiaccontractility. This conclusion is based on the findings thatSNAP-induced increase in contractility was not associatedwith phosphorylation of phospholamban, and was notmodified by PKA inhibition (Fig. 6). The appropriatenessof the method used to assess this signaling pathway wascorroborated by the results observed with beta adrenergicagonist, isoproterenol, used as a positive control. Theseresults are different from those obtained by Vila-Petroffet al. [11] who reported activation of the cAMP pathwayupon treatment of isolated myocytes with SNAP. In thissense, it is possible that in the beating heart, the tonic levelsof PKA activation are not disturbed by the addition ofSNAP.

The positive inotropic response induced by low concen-trations of SNAP appears to occur by redox-mediatedmodifications (S-nitrosylation and/or S-glutathiolation),since it was abolished by the superoxide dismutase mimetic,Tempol. Superoxide ion is known to be required to formsome of the intermediaries of N2O3, the chemical speciesthat is the final donor of NO+, the nitrosylating agent[24,31]. Peroxynitrite is formed as result of the reactionbetween NO and superoxide and behaves as a nitrosylatingagent when both radicals are generated at equal fluxes.Indeed, peroxynitrite increases cardiac contractility in aSOD-sensitive manner [9,27]. On the contrary, Tempoldid not affect the reduction in (dP/dt)max elicited by highSNAP concentrations, indicating that the nitrosylatingpathway is not involved in the negative inotropic response.The concept that low concentrations of SNAP, althoughunable to induce significant guanylate cyclase activationas demonstrated by lack of effect on cardiac cGMP con-tent, are able to trigger a redox-sensitive mechanism affect-

ing key proteins involved in calcium handling can bereasoned by the fact that SNAP is a nitrosothiol, and 0.1and 1 lM of nitrosothiols could be enough to produceS-nitrosylation. Here we showed that 2-min perfusion with1 lM SNAP caused an increase in the degree of nitrosyla-tion of RyR and several other proteins. Thus, redox mod-ifications of SERCA and L-type calcium channels mayoccur at the same time. It is likely that chemical modifica-tion of RyR2 favors Ca2+ release in rat cardiac myocytes,since it has been demonstrated that RyR from canine car-diac muscle is activated by S-nitrosylation [32]. An increasein the activity of SERCA and L-type calcium channels isrequired to increase calcium transients, since the soleincrease in RyR’s open probability is not enough to keepan increase in calcium transients over time [33]. Interest-ingly, it has been described in the rabbit heart that theperoxynitrite donor SIN-1 and authentic NO caused S-glutathiolation in SERCA, increasing the activity of thepump [34]. Also RyR1 from skeletal muscle has beenshown to be both nitrosylated and gluthathiolated inresponse to S-nitrosoglutathione (GSNO), increasing itsactivity [35,36]. We cannot discard the possibility that asimilar reaction may occur in cardiac RyR2.

S-Nitrosylation should have also taken place when per-fusing the heart with the high concentrations of SNAP,despite the response observed was a decrease in contractil-ity. Indeed, using an anti-S-nitroso-cysteine antibody, weconfirmed that the degree of S-nitrosylation also increasedin a broad range of proteins in hearts treated with 100 lMSNAP (data not shown). Although S-nitrosylation pro-gressively activates cardiac RyR2, we cannot discard thatexcessive amounts of the NO donor may cause oxidationof thiols groups leading to irreversible activation and lossof control, as reported in isolated RyR2 [32]. In this lineof thought, a significant increase in S-nitrosylation of car-diac L-type calcium channels has been linked to decrease incalcium current in a model of heart ischemia in mice [37].Alternatively, we think that with a high SNAP concentra-tion, even though redox modifications occur, their effectsare overridden by cGMP–PKG dependent effects. A likelyexplanation for this finding relay on the targets of cGMPand redox modifications. On one hand, RyR, L-type cal-cium channels [38] or SERCA2 [34] are proteins involvedin calcium cycling, which are reported to undergo redoxmodifications, likely providing higher cytosolic calciumtransients. On the other hand, PKG-mediated phosphory-lation of troponin I leads to a decrease in calcium respon-siveness of the myofilaments [14], condition that mayoverride the positive inotropic effect of redox modifica-tions. PKG also can decrease L-type Ca2+ current contrib-uting to the negative inotropic effect [15]. The finding thatthe negative inotropism elicited by 100 lM SNAP wasreverted by ODQ into positive inotropism (Fig. 3), furthersupports the concept that redox modifications do occur athigh concentrations of NO donors, but cGMP accumula-tion prevents and overrules the increase in contractility.Thus, at high production, NO concentrations could be able

166 D.R. Gonzalez et al. / Nitric Oxide 18 (2008) 157–167

to activate both pathways (redox and cGMP), but thedesensitizing effect would predominate because the myofil-ament responsiveness to calcium is located downstream ofthe calcium handling proteins.

In cardiac myocytes, the distribution of NOS isoforms ishighly compartmentalized [39,40], and NO production bythe two NOS isoforms, NOS1 and NOS3 exert oppositeeffects, NOS1 clearly favors contractility, while NOS3decreases the adrenergic response [12,41]. The presentresults with our pharmacological approach are not incom-patible with the importance of compartmentalization forthe in vivo regulation, since NOS1-derived NO must below according to the reported levels of NOS1 expressionin the myocytes [42], however, it has been shown that whenoverexpressed, NOS1 plainly activates the cGMP pathwayand decreases contractility [43]. A word of caution must besaid, since overexpressed NOS1 is also present at theplasma membrane where it can be in close proximity toL-type calcium channels, mimicking the effects of eNOS.Likewise, it has been reported that NOS1 translocate tothe plasmalemma after ischemia-reperfusion, and in thiscondition it can contribute to S-nitrosylate the L-type cal-cium channel [37]. Thus, our results with different concen-trations of SNAP, most likely resemble activation of twodistinct pathways that are associated with activation of dif-ferent NOS isoforms in the heart in different physiologicalor pathophysiological conditions.

In conclusion, we show here that SNAP, a S-nitroso-thiol and NO donor, is able to trigger at least two pathwaysin the heart: S-nitrosylation and cGMP formation, withopposite effects in contractility.

Acknowledgments

We thank Ms Ines Poblete for her technical assistancewith nitric oxide and cGMP measurements and JorgeMiranda for his help in isolated heart experiments.

Supported by Grant 1040816 from Fondecyt (ChileanNational Fund for Science and Technology), and a Schol-arship from Conicyt (Chilean National Council for Scienceand Technology) to Daniel R. Gonzalez.

References

[1] J.M. Hare, Nitric oxide and excitation–contraction coupling, J. Mol.Cell. Cardiol. 35 (2003) 719–729.

[2] S. Chowdhary, D. Harrington, R.S. Bonser, J.H. Coote, J.N.Townend, Chronotropic effects of nitric oxide in the denervatedhuman heart, J. Physiol. 541 (2002) 645–651.

[3] G.W. Booz, Putting the brakes on cardiac hypertrophy: exploiting theNO–cGMP counter-regulatory system, Hypertension 45 (2005) 341–346.

[4] G. Kojda, K. Kottenberg, Regulation of basal myocardial functionby NO, Cardiovasc. Res. 41 (1999) 514–523.

[5] A.J. Brady, J.B. Warren, P.A. Poole-Wilson, T.J. Williams, S.E.Harding, Nitric oxide attenuates cardiac myocyte contraction, Am. J.Physiol. 265 (1993) H176–H182.

[6] M. Flesch, H. Kilter, B. Cremers, O. Lenz, M. Sudkamp, F. Kuhn-Regnier, M. Bohm, Acute effects of nitric oxide and cyclic GMP on

human myocardial contractility, J. Pharmacol. Exp. Ther. 281 (1997)1340–1349.

[7] P. Mohan, D.L. Brutsaert, W.J. Paulus, S.U. Sys, Myocardialcontractile response to nitric oxide and cGMP, Circulation 93(1996) 1223–1229.

[8] G. Kojda, K. Kottenberg, P. Nix, K.D. Schluter, H.M. Piper, E.Noack, Low increase in cGMP induced by organic nitrates andnitrovasodilators improves contractile response of rat ventricularmyocytes, Circ. Res. 78 (1996) 91–101.

[9] N. Paolocci, U.E. Ekelund, T. Isoda, M. Ozaki, K. Vandegaer, D.Georgakopoulos, R.W. Harrison, D.A. Kass, J.M. Hare, cGMP-independent inotropic effects of nitric oxide and peroxynitrite donors:potential role for nitrosylation, Am. J. Physiol. Heart. Circ. Physiol.279 (2000) H1982–H1988.

[10] B. Preckel, G. Kojda, W. Schlack, D. Ebel, K. Kottenberg, E. Noack,V. Thamer, Inotropic effects of glyceryl trinitrate and spontaneousNO donors in the dog heart, Circulation 96 (1997) 2675–2682.

[11] M.G. Vila-Petroff, A. Younes, J. Egan, E.G. Lakatta, S.J. Sollott,Activation of distinct cAMP-dependent and cGMP-dependent path-ways by nitric oxide in cardiac myocytes, Circ. Res. 84 (1999) 1020–1031.

[12] L.A. Barouch, R.W. Harrison, M.W. Skaf, G.O. Rosas, T.P. Cappola,Z.A. Kobeissi, I.A. Hobai, C.A. Lemmon, A.L. Burnett, B. O’Rourke,E.R. Rodriguez, P.L. Huang, J.A. Lima, D.E. Berkowitz, J.M. Hare,Nitric oxide regulates the heart by spatial confinement of nitric oxidesynthase isoforms, Nature 416 (2002) 337–339.

[13] L.J. Ignarro, Nitric oxide. A novel signal transduction mechanism fortranscellular communication, Hypertension 16 (1990) 477–483.

[14] J. Layland, J.M. Li, A.M. Shah, Role of cyclic GMP-dependentprotein kinase in the contractile response to exogenous nitric oxide inrat cardiac myocytes, J. Physiol. 540 (2002) 457–467.

[15] L.H. Jiang, D.J. Gawler, N. Hodson, C.J. Milligan, H.A. Pearson, V.Porter, D. Wray, Regulation of cloned cardiac L-type calciumchannels by cGMP-dependent protein kinase, J. Biol. Chem. 275(2000) 6135–6143.

[16] F. Rochais, G. Vandecasteele, F. Lefebvre, C. Lugnier, H. Lum, J.L.Mazet, D.M. Cooper, R. Fischmeister, Negative feedback exerted bycAMP-dependent protein kinase and cAMP phosphodiesterase onsubsarcolemmal cAMP signals in intact cardiac myocytes: an in vivostudy using adenovirus-mediated expression of CNG channels, J.Biol. Chem. 279 (2004) 52095–52105.

[17] D.T. Hess, A. Matsumoto, S.O. Kim, H.E. Marshall, J.S. Stamler,Protein S-nitrosylation: purview and parameters, Nat. Rev. Mol.Cell. Biol. 6 (2005) 150–166.

[18] J.M. Hare, J.S. Stamler, NO/redox disequilibrium in the failing heartand cardiovascular system, J. Clin. Invest. 115 (2005) 509–517.

[19] Y. Goto, B.K. Slinker, M.M. LeWinter, Effect of coronary hyperemiaon Emax and oxygen consumption in blood-perfused rabbit hearts.Energetic consequences of Gregg’s phenomenon, Circ. Res. 68 (1991)482–492.

[20] X.F. Figueroa, D.R. Gonzalez, A.D. Martinez, W.N. Duran, M.P.Boric, ACh-induced endothelial NO synthase translocation, NOrelease and vasodilatation in the hamster microcirculation in vivo, J.Physiol. 544 (2002) 883–896.

[21] S.R. Jaffrey, H. Erdjument-Bromage, C.D. Ferris, P. Tempst, S.H.Snyder, Protein S-nitrosylation: a physiological signal for neuronalnitric oxide, Nat. Cell Biol. 3 (2001) 193–197.

[22] S.R. Jaffrey, S.H. Snyder, The biotin switch method for the detectionof S-nitrosylated proteins, Sci. STKE 2001 (2001) PL1.

[23] M.P. Boric, H.R. Croxatto, Inhibition of atrial natriuretic peptideexcretory action by bradykinin, Hypertension 26 (1995) 1167–1172.

[24] M.G. Espey, K.M. Miranda, D.D. Thomas, S. Xavier, D. Citrin,M.P. Vitek, D.A. Wink, A chemical perspective on the interplaybetween NO, reactive oxygen species, and reactive nitrogen oxidespecies, Ann. N. Y. Acad. Sci. 962 (2002) 195–206.

[25] I. Koshiishi, T. Takajo, K. Tsuchida, Regulation of S-thiolation andS-nitrosylation in the thiol/nitric oxide system by radical scavengers,Nitric Oxide 16 (2007) 356–361.

D.R. Gonzalez et al. / Nitric Oxide 18 (2008) 157–167 167

[26] A. Schrammel, A.C. Gorren, K. Schmidt, S. Pfeiffer, B. Mayer, S-nitrosation of glutathione by nitric oxide, peroxynitrite, and (*)NO/O(2)(*–), Free Radic. Biol. Med. 34 (2003) 1078–1088.

[27] J.M. Chesnais, R. Fischmeister, P.F. Mery, Positive and negativeinotropic effects of NO donors in atrial and ventricular fibres of thefrog heart, J. Physiol. 518 (Pt. 2) (1999) 449–461.

[28] T.C. Bellamy, J. Wood, J. Garthwaite, On the activation of solubleguanylyl cyclase by nitric oxide, Proc. Natl. Acad. Sci. USA 99 (2002)507–510.

[29] G. Kojda, K. Kottenberg, E. Noack, Inhibition of nitric oxidesynthase and soluble guanylate cyclase induces cardiodepressiveeffects in normal rat hearts, Eur. J. Pharmacol. 334 (1997) 181–190.

[30] R. Fischmeister, L.R. Castro, A. Abi-Gerges, F. Rochais, J. Jurevi-cius, J. Leroy, G. Vandecasteele, Compartmentation of cyclicnucleotide signaling in the heart: the role of cyclic nucleotidephosphodiesterases, Circ. Res. 99 (2006) 816–828.

[31] M.G. Espey, D.D. Thomas, K.M. Miranda, D.A. Wink, Focusing ofnitric oxide mediated nitrosation and oxidative nitrosylation as aconsequence of reaction with superoxide, Proc. Natl. Acad. Sci. USA99 (2002) 11127–11132.

[32] L. Xu, J.P. Eu, G. Meissner, J.S. Stamler, Activation of the cardiaccalcium release channel (ryanodine receptor) by poly-S-nitrosylation,Science 279 (1998) 234–237.

[33] D.A. Eisner, A.W. Trafford, No role for the ryanodine receptor inregulating cardiac contraction? News Physiol. Sci. 15 (2000) 275–279.

[34] T. Adachi, R.M. Weisbrod, D.R. Pimentel, J. Ying, V.S. Sharov, C.Schoneich, R.A. Cohen, S-Glutathiolation by peroxynitrite activatesSERCA during arterial relaxation by nitric oxide, Nat. Med. 10(2004) 1200–1207.

[35] P. Aracena, G. Sanchez, P. Donoso, S.L. Hamilton, C. Hidalgo, S-glutathionylation decreases Mg2+ inhibition and S-nitrosylationenhances Ca2+ activation of RyR1 channels, J. Biol. Chem. 278(2003) 42927–42935.

[36] J. Sun, L. Xu, J.P. Eu, J.S. Stamler, G. Meissner, Nitric oxide, NOC-12, and S-nitrosoglutathione modulate the skeletal muscle calciumrelease channel/ryanodine receptor by different mechanisms. Anallosteric function for O2 in S-nitrosylation of the channel, J. Biol.Chem. 278 (2003) 8184–8189.

[37] J. Sun, E. Picht, K.S. Ginsburg, D.M. Bers, C. Steenbergen, E.Murphy, Hypercontractile female hearts exhibit increased S-nitrosy-lation of the L-type Ca2+ channel alpha1 subunit and reducedischemia/reperfusion injury, Circ. Res. 98 (2006) 403–411.

[38] D.L. Campbell, J.S. Stamler, H.C. Strauss, Redox modulation of L-type calcium channels in ferret ventricular myocytes. Dual mechanismregulation by nitric oxide and S-nitrosothiols, J. Gen. Physiol. 108(1996) 277–293.

[39] O. Feron, C. Dessy, D.J. Opel, M.A. Arstall, R.A. Kelly, T. Michel,Modulation of the endothelial nitric-oxide synthase–caveolin inter-action in cardiac myocytes. Implications for the autonomic regulationof heart rate, J. Biol. Chem. 273 (1998) 30249–30254.

[40] K.Y. Xu, D.L. Huso, T.M. Dawson, D.S. Bredt, L.C. Becker, Nitricoxide synthase in cardiac sarcoplasmic reticulum, Proc. Natl. Acad.Sci. USA 96 (1999) 657–662.

[41] S.A. Khan, K. Lee, K.M. Minhas, D.R. Gonzalez, S.V. Raju, A.D.Tejani, D. Li, D.E. Berkowitz, J.M. Hare, Neuronal nitric oxidesynthase negatively regulates xanthine oxidoreductase inhibition ofcardiac excitation–contraction coupling, Proc. Natl. Acad. Sci. USA101 (2004) 15944–15948.

[42] M.V. Brahmajothi, D.L. Campbell, Heterogeneous basal expressionof nitric oxide synthase and superoxide dismutase isoforms inmammalian heart: implications for mechanisms governing indirectand direct nitric oxide-related effects, Circ. Res. 85 (1999) 575–587.

[43] N. Burkard, A.G. Rokita, S.G. Kaufmann, M. Hallhuber, R. Wu, K.Hu, U. Hofmann, A. Bonz, S. Frantz, E.J. Cartwright, L. Neyses,L.S. Maier, S.K. Maier, T. Renne, K. Schuh, O. Ritter, Conditionalneuronal nitric oxide synthase overexpression impairs myocardialcontractility, Circ. Res. 100 (2007) e32–e44.