Inhibition of neuronal nitric oxide synthase activity by N1-acetyl-5-methoxykynuramine, a brain metabolite of melatonin

We assessed the effects of melatonin, N(1)-acetyl-N (2)-formyl-5-methoxykynuramine (AFMK) and N(1)-acetyl-5-methoxykynuramine (AMK) on neuronal nitric oxide synthase (nNOS) activity in vitro and in rat striatum in vivo. Melatonin and AMK (10(-11)-10(-3) m), but not AFMK, inhibited nNOS activity in vitro in a dose-response manner. The IC(50) value for AMK (70 microm) was significantly lower than for melatonin (>1 mm). A 20% nNOS inhibition was reached with either 10(-9) m melatonin or 10(-11) m AMK. AMK inhibits nNOS by a non-competitive mechanism through its binding to Ca(2+)-calmodulin (CaCaM). The inhibition of nNOS elicited by melatonin, but not by AMK, was blocked with 0.05 mm norharmane, an indoleamine-2,3-dioxygenase inhibitor. In vivo, the potency of AMK to inhibit nNOS activity was higher than that of melatonin, as a 25% reduction in rat striatal nNOS activity was found after the administration of either 10 mg/kg of AMK or 20 mg/kg of melatonin. Also, in vivo, the administration of norharmane blocked the inhibition of nNOS produced by melatonin administration, but not the inhibition produced by AMK. These data reveal that AMK rather than melatonin is the active metabolite against nNOS, which may be inhibited by physiological levels of AMK in the rat striatum.     

Selective L-nitroargininylaminopyrrolidine and L-nitroargininylaminopiperidine neuronal nitric oxide synthase inhibitors

Selective inhibition of the localized excess production of NO by neuronal nitric oxide synthase (nNOS) has been targeted as a potential means of treating various neurological disorders. Based on observations from the X-ray crystal structures of complexes of nNOS with two nNOS-selective inhibitors, (4S)-N-{4-amino-5-[(2-amino)ethylamino]pentyl}-N'-nitroguanidine (L-Arg(NO2)-L-Dbu-NH2 (1) and 4-N-(Nomega-nitro-L-argininyl)-trans-4-amino-L-proline amide (2), a series of descarboxamide analogues was designed and synthesized (3-7). The most potent compound was aminopyrrolidine analogue 3, which exhibited better potency and selectivity for nNOS than parent compound 2. In addition, 3 provided higher lipophilicity and a lower molecular weight than 2, therefore having better physicochemical properties. Nalpha-Methylated analogues (8-11) also were prepared for increased lipophilicity of the inhibitors, but they had 4- to 5-fold weaker binding affinity compared to their parent compounds.     

Synthesis and biological evaluation of 4,5-dihydro-1H-pyrazole derivatives as potential nNOS/iNOS selective inhibitors. Part 2: Influence of diverse substituents in both the phenyl moiety and the acyl group

In a preliminary article, we reported a series of 4,5-dihydro-1H-pyrazole derivatives as neuronal nitric oxide synthase (nNOS) inhibitors. Here we present the data about the inhibition of inducible nitric oxide synthase (iNOS) of these compounds. In general, we can confirm that these pyrazoles are nNOS selective inhibitors. In addition, taking these compounds as a reference, we have designed and synthesized a series of new derivatives by modification of the heterocycle in 1-position, and by introduction of electron-donating or electron-withdrawing substituents in the aromatic ring. These derivatives have been evaluated as nNOS and iNOS inhibitors in order to identify new compounds with improved activity and selectivity. Compound 3r, with three methoxy electron-donating groups in the phenyl moiety, is the most potent nNOS inhibitor, showing good selectivity nNOS/iNOS.     

Sympathetic control of heart rate in nNOS knockout mice

Inhibition of neuronal nitric oxide synthase (nNOS) in cardiac postganglionic sympathetic neurons leads to enhanced cardiac sympathetic responsiveness in normal animals, as well as in animal models of cardiovascular diseases. We used isolated atria from mice with selective genetic disruption of nNOS (nNOS(-/-)) and their wild-type littermates (WT) to investigate whether sympathetic heart rate (HR) responses were dependent on nNOS. Immunohistochemistry was initially used to determine the presence of nNOS in sympathetic [tyrosine hydroxylase (TH) immunoreactive] nerve terminals in the mouse sinoatrial node (SAN). After this, the effects of postganglionic sympathetic nerve stimulation (1-10 Hz) and bath-applied norepinephrine (NE; 10(-8)-10(-4) mol/l) on HR were examined in atria from nNOS(-/-) and WT mice. In the SAN region of WT mice, TH and nNOS immunoreactivity was virtually never colocalized in nerve fibers. nNOS(-/-) atria showed significantly reduced HR responses to sympathetic nerve activation and NE (P < 0.05). Similarly, the positive chronotropic response to the adenylate cyclase activator forskolin (10(-7)-10(-5) mol/l) was attenuated in nNOS(-/-) atria (P < 0.05). Constitutive NOS inhibition with L-nitroarginine (0.1 mmol/l) did not affect the sympathetic HR responses in nNOS(-/-) and WT atria. The paucity of nNOS in the sympathetic innervation of the mouse SAN, in addition to the attenuated HR responses to neuronal and applied NE, indicates that presynaptic sympathetic neuronal NO does not modulate neuronal NE release and SAN pacemaking in this species. It appears that genetic deletion of nNOS results in the inhibition of adrenergic-adenylate cyclase signaling within SAN myocytes.     

1-(1,3-Benzodioxol-5-ylmethyl)-3-[4-(1H-imidazol-1-yl)phenoxy]-piperidine analogs as potent and selective inhibitors of nitric oxide formation

A new series of 1-(1,3-benzodioxol-5-ylmethyl)-3-[4-(1H-Imidazol-1-yl)phenoxy]-piperidine analogs were designed and identified as potent and selective inhibitors of NO formation based both on the crystal structure of a murine iNOS Delta114 monomer domain/ inhibitor complex and inhibition of the NO formation in human A172 cell assays. Compound 12S showed high potency and high iNOS selectivity versus nNOS and eNOS.     

Effects of long-term inhibition of neuronal nitric oxide synthase (NOS1) in uninephrectomized diabetic rats.

Nitric oxide (NO) has been implicated in the pathogenesis of renal hemodynamic changes in diabetes mellitus (DM). However, the role of NO in the pathophysiology of diabetic nephropathy remains controversial. Renal hemodynamic changes in experimental DM can be acutely normalized by selective inhibition of neuronal NO synthase (nNOS). This observation suggests a nephroprotective potential of nNOS inhibition in DM. To explore this issue we assessed the long-term effects (12 weeks) of selective nNOS inhibition with the specific inhibitor S-methyl-L-thiocitrulline (SMTC) in uninephrectomized control and streptozotocin-diabetic rats. No beneficial effects of SMTC were observed in nondiabetic controls. In contrast, SMTC delayed the development of proteinuria (32+/-8 vs. 53+/-9 mg/24h, week 8, p < 0.05) and glomerulosclerosis (GS, 0.30+/-0.08 vs. 0.57+/-0.05, p < 0.05) in diabetic rats. These effects coincided with early effects of treatment on the glomerular filtration rate, and were associated with lower renal expression of nNOS. Furthermore, SMTC-treated diabetic rats demonstrated reduced weight gain and urinary sodium excretion as compared to vehicle-treated counterparts, despite similar metabolic control and blood pressure. In summary, long-term nNOS inhibition had modest nephroprotective effects in uninephrectomized diabetic rats. These effects may be mediated by renal hemodynamic mechanisms, as well as by lower food (protein) intake.     

Estradiol increases urethral tone through the local inhibition of neuronal nitric oxide synthase expression

Estrogens are known to modulate lower urinary tract (LUT) trophicity and neuronal nitric oxide synthase (nNOS) expression in several organs. The aim of this study was to explore the effects of endogenous and supraestrus levels of 17beta-estradiol (E2) on LUT and urethral nNOS expression and function. LUT function and histology and urethral nNOS expression were studied in adult female mice subjected either to sham surgery, surgical castration, or castration plus chronic E2 supplementation (80 microg.kg(-1).day(-1), i.e., pregnancy level). The micturition pattern was profoundly altered by long-term supraestrus levels of E2 with decreased frequency paralleled by increased residual volumes higher than those of ovariectomized mice. Urethral resistance was increased twofold in E2-treated mice, with no structural changes in urethra, supporting a pure tonic mechanism. Acute nNOS inhibition by 7-nitroindazole decreased frequency and increased residual volumes in ovariectomized mice but had no additive effect on the micturition pattern of long-term supraestrus mice, showing that long-term supraestrus E2 levels and acute inhibition of nNOS activity had similar functional effects. Finally, E2 decreased urethral nNOS expression in ovariectomized mice. Long-term supraestrus levels of E2 increased urethral tone through inhibition of nNOS expression, whereas physiological levels of E2 had no effect.     

Cyclopropyl- and methyl-containing inhibitors of neuronal nitric oxide synthase

Inhibitors of neuronal nitric oxide synthase have been proposed as therapeutics for the treatment of different types of neurological disorders. On the basis of a cis-3,4-pyrrolidine scaffold, a series of trans-cyclopropyl- and methyl-containing nNOS inhibitors have been synthesized. The insertion of a rigid electron-withdrawing cyclopropyl ring decreases the basicity of the adjacent amino group, which resulted in decreased inhibitory activity of these inhibitors compared to the parent compound. Nonetheless, three of them exhibited double-digit nanomolar inhibition with high nNOS selectivity on the basis of in vitro enzyme assays. Crystal structures of nNOS and eNOS with these inhibitors bound provide a basis for detailed structure–activity relationship (SAR) studies. The conclusions from these studies will be used as a guide in the future development of selective NOS inhibitors.     

Inhibition of neuronal nitric-oxide synthase by phosphorylation at Threonine1296 in NG108-15 neuronal cells

We demonstrate that neuronal nitric-oxide synthase (nNOS) is directly inhibited through the phosphorylation of Thr(1296) in NG108-15 neuronal cells. Treatment of NG108-15 cells expressing nNOS with calyculin A, an inhibitor of protein phosphatase 1 and 2A, revealed a dose-dependent inhibition of nNOS enzyme activity with concomitant phosphorylation of Thr(1296) residue. Cells expressing a phosphorylation-deficient mutant in which Thr(1296) was changed to Ala proved resistant to phosphorylation and suppression of NOS activity. Mimicking phosphorylation mutant of nNOS in which Thr(1296) is changed to Asp showed a significant decrease in nNOS enzyme activity, being competitive with NADPH, relative to the wild-type enzyme. These data suggest that phosphorylation of nNOS at Thr(1296) may involve the attenuation of nitric oxide production in neuronal cells through the decrease of NADPH-binding to the enzyme.     

Angiotensin II feedback is a regulator of renocortical renin,COX-2, and nNOS expression

Salt restriction leads to parallel increases of renin, cyclooxygenase-2 (COX-2), and neuronal nitric oxide synthase (nNOS) gene expression in the juxtaglomerular apparatus of rat kidneys. Because the upregulation of these genes is strongly enhanced if salt restriction is combined with inhibition of the renin-angiotensin-aldosterone system, our study aimed to find out whether the juxtaglomerular expressions of renin, COX-2, and nNOS are subject to a common direct negative feedback control by ANG II. For this purpose, male Sprague-Dawley rats were fed a low-salt diet (0.02% wt/wt) with or without additional treatment with the ANG I-converting enzyme (ACE) inhibitor ramipril (10 mg x kg body wt(-1) x day(-1)) for 1 wk, and renocortical renin, COX-2, and nNOS mRNAs were assayed. To narrow down possible indirect effects of the ACE inhibitor that may result from insufficient aldosterone production, the animals received mineralocorticoid substitution with fludrocortisone (6 mg. kg body wt(-1) x day(-1)). Thus mineralocorticoid substitution prevented the fall of systolic blood pressure and of glomerular filtration induced by ramipril in rats on low-salt diet. Although fludrocortisone had no effect on basal renin, COX-2, and nNOS mRNA, it clearly attenuated the threefold increases of both renin and COX-2 mRNA in response to low-salt diet. In rats on low-salt diet, ramipril further increased renin mRNA ninefold, COX-2 mRNA fourfold, and nNOS 2.5-fold in the absence of fludrocortisone. In the presence of fludrocortisone, ramipril increased renin mRNA 10-fold, COX-2 mRNA 2.5-fold, and nNOS mRNA 2.5-fold. These data indicate that mineralocorticoid substitution lowers the overall expression of juxtaglomerular renin and COX-2 during low-salt intake and attenuates a further rise of COX-2 expression by ACE inhibition, but it does not change the stimulatory effect of ACE inhibition on renin and nNOS expression. We conclude that the expression of renin, COX-2, and nNOS in the juxtaglomerular apparatus during low-salt diet is markedly limited by a direct feedback inhibition through ANG II.     

Role of neuronal nitric oxide synthase in mediating renal hemodynamic changes during pregnancy

Renal plasma flow (RPF) and glomerular filtration rate (GFR) are markedly increased during pregnancy. We recently reported that the renal hemodynamic changes observed during pregnancy in rats are associated with enhanced renal protein expression of neuronal nitric oxide synthase (nNOS). The purpose of this study was to determine the role of nNOS in mediating renal hemodynamic changes observed during pregnancy. To achieve this goal, we examined the effects of the nNOS inhibitor 7-nitroindazole (7-NI) on kidney function in normal conscious, chronically instrumented virgin (n = 6) and pregnant rats (n = 9) at day 16 of gestation. Infusion of 7-NI had no effect on RPF (4.7 +/- 0.7 vs. 4.8 +/- 0.9 ml/min), GFR (2.2 +/- 0.2 vs. 2.5 +/- 0.4 ml/min), or mean arterial pressure (MAP; 127 +/- 7 vs. 129 +/- 10 mmHg) in virgin rats. In contrast, 7-NI infused into pregnant rats decreased RPF (8.9 +/- 1.6 vs. 6.5 +/- 1.4 ml/min) and GFR (4.4 +/- 0.7 vs. 3.3 +/- 0.7 ml/min) while having no effect on MAP (123 +/- 4 vs. 123 +/- 3 mmHg). In summary, inhibition of nNOS in pregnant rats at midgestation results in significant decreases in RPF and GFR. nNOS inhibition in virgin rats had no effect on renal hemodynamics. These data suggest that nNOS may play a role in mediating the renal hemodynamic changes that occur during pregnancy.     

PIN inhibits nitric oxide and superoxide production from purified neuronal nitric oxide synthase

A protein inhibitor of neuronal nitric oxide synthase (nNOS) was identified and designated as PIN. PIN was reported to inhibit nNOS activity in cell lysates through disruption of enzyme dimerization. However, there has been lack of direct characterization of the effect of PIN on NO production from purified nNOS. Furthermore, nNOS also generates superoxide (.O(2)(-)) at low levels of L-arginine. It is unknown whether PIN affects .O(2)(-) generation from nNOS. Therefore, we performed direct measurements of the effects of PIN on NO and .O(2)(-) generation from purified nNOS using electron paramagnetic resonance spin trapping techniques. nNOS was isolated by affinity chromatography and a fusion protein CBP-PIN was used to probe the effect of PIN. While the tag CBP did not affect nNOS activity, CBP-PIN caused a dose-dependent inhibition on both NO and L-citrulline production. In the absence of L-arginine, strong .O(2)(-) generation was observed from nNOS, and this was blocked by CBP-PIN in a dose-dependent manner. With low-temperature polyacrylamide gel electrophoresis, neither CBP nor CBP-PIN was found to affect nNOS dimerization. Thus, these results suggested that PIN not only inhibits NO but also .O(2)(-) production from nNOS, and this is through a mechanism other than decomposition of nNOS dimers.     

Renal nerves and nNOS: roles in natriuresis of acute isovolumetric sodium loading in conscious rats

It was hypothesized that renal sympathetic nerve activity (RSNA) and neuronal nitric oxide synthase (nNOS) are involved in the acute inhibition of renin secretion and the natriuresis following slow NaCl loading (NaLoad) and that RSNA participates in the regulation of arterial blood pressure (MABP). This was tested by NaLoad after chronic renal denervation with and without inhibition of nNOS by S-methyl-thiocitrulline (SMTC). In addition, the acute effects of renal denervation on MABP and sodium balance were assessed. Rats were investigated in the conscious, catheterized state, in metabolic cages, and acutely during anesthesia. NaLoad was performed over 2 h by intravenous infusion of hypertonic solution (50 micromol.min(-1).kg body mass(-1)) at constant body volume conditions. SMTC was coinfused in amounts (20 microg.min(-1).kg(-1)) reported to selectively inhibit nNOS. Directly measured MABPs of acutely and chronically denervated rats were less than control (15% and 9%, respectively, P < 0.005). Plasma renin concentration (PRC) was reduced by renal denervation (14.5 +/- 0.2 vs. 19.3 +/- 1.3 mIU/l, P < 0.005) and by nNOS inhibition (12.4 +/- 2.3 vs. 19.6 +/- 1.6 mlU/l, P < 0.005). NaLoad reduced PRC (P < 0.05) and elevated MABP modestly (P < 0.05) and increased sodium excretion six-fold, irrespective of renal denervation and SMTC. The metabolic data demonstrated that renal denervation lowered sodium balance during the first days after denervation (P < 0.001). These data show that renal denervation decreases MABP and renin secretion. However, neither renal denervation nor nNOS inhibition affects either the renin down-regulation or the natriuretic response to acute sodium loading. Acute sodium-driven renin regulation seems independent of RSNA and nNOS under the present conditions.     

Exploring the binding conformations of bulkier dipeptide amide inhibitors in constitutive nitric oxide synthases

A series of L-nitroarginine-based dipeptide inhibitors are highly selective for neuronal nitric oxide synthase (nNOS) over the endothelial isoform (eNOS). Crystal structures of these dipeptides bound to both isoforms revealed two different conformations, curled in nNOS and extended in eNOS, corresponding to higher and lower binding affinity to the two isoforms, respectively. In previous studies we found that the primary reason for selectivity is that Asp597 in nNOS, which is Asn368 in eNOS, provides greater electrostatic stabilization in the inhibitor complex. While this is the case for smaller dipeptide inhibitors, electrostatic stabilization may no longer be the sole determinant for isoform selectivity with bulkier dipeptide inhibitors. Another residue farther away from the active site, Met336 in nNOS (Val106 in eNOS), is in contact with bulkier dipeptide inhibitors. Double mutants were made to exchange the D597/M336 pair in nNOS with N368/V106 in eNOS. Here we report crystal structures and inhibition constants for bulkier dipeptide inhibitors bound to nNOS and eNOS that illustrate the important role played by residues near the entry to the active site in isoform selective inhibition.     

Nitric oxide from neuronal nitric oxide synthase sensitises neurons to hypoxia-induced death via competitive inhibition of cytochrome oxidase

Hypoxia/ischaemia is known to trigger neuronal death, but the role of neuronal nitric oxide synthase (nNOS) in this process is controversial. Nitric oxide (NO) inhibits cytochrome oxidase in competition with oxygen. We tested whether NO derived from nNOS synergises with hypoxia to induce neuronal death by inhibiting mitochondrial cytochrome oxidase. Sixteen hours of hypoxia (2% oxygen) plus deoxyglucose (an inhibitor of glycolysis) caused extensive, excitotoxic death of neurons in rat cerebellar granule cell cultures. Three different nNOS inhibitors (including the selective inhibitor N-4S-4-amino-5-2-aminoethyl-aminopentyl-N'-nitroguanidine) decreased this neuronal death by half, indicating a contribution of nNOS to hypoxic death. The selective nNOS inhibitor did not, however, block neuronal death induced either by added glutamate or by added azide (an uncompetitive inhibitor of cytochrome oxidase), indicating that nNOS does not act downstream of glutamate or cytochrome oxidase. Hypoxia plus deoxyglucose-induced glutamate release and neuronal depolarisation, and the nNOS inhibitor decreased this. Hypoxia inhibited cytochrome oxidase activity in the cultures, but a selective nNOS inhibitor prevented this inhibition, indicating NO from nNOS was inhibiting cytochrome oxidase in competition with oxygen. These data indicate that hypoxia synergises with NO from nNOS to induce neuronal death via cytochrome oxidase inhibition causing neuronal depolarisation. This mechanism might contribute to ischaemia/stroke-induced neuronal death in vivo.     

1,5-Disubstituted indole derivatives as selective human neuronal nitric oxide synthase inhibitors

A series of 1,5-disubstituted indole derivatives was designed, synthesized and evaluated as inhibitors of human nitric oxide synthase. A variety of flexible and restricted basic amine side chain substitutions was explored at the 1-position of the indole ring, while keeping the amidine group fixed at the 5-position. Compounds having N-(1-(2-(1-methylpyrrolidin-2-yl)ethyl)- (12, (R)-12, (S)-12 and 13) and N-(1-(1-methylazepan-4-yl)- side chains (14, 15, (-)-15 and (+)-15) showed increased inhibitory activity for the human nNOS isoform and selectivity over eNOS and iNOS isoforms. The most potent compound of the series for human nNOS (IC(50)=0.02 μM) (S)-12 showed very good selectivity over the eNOS (eNOS/nNOS=96-fold) and iNOS (iNOS/nNOS=850-fold) isoforms.     

Novel inhibitors of neuronal nitric oxide synthase

Selective inhibitors of neuronal nitric oxide synthase (nNOS), which are devoid of any effect on the endothelial isoform (eNOS), may be required for the treatment of some neurological disorders. In our search for novel nNOS inhibitors, we recently described some 1-[(Aryloxy)ethyl]-1H-imidazoles as interesting molecules for their selectivity for nNOS against eNOS. This work reports a new series of 1-[(Aryloxy)alkyl]-1H-imidazoles in which a longer methylene chain is present between the imidazole and the phenol part of molecule. Some of these molecules were found to be more potent nNOS inhibitors than the parent ethylenic compounds, although this increase in potency resulted in a partial loss of selectivity. The most interesting compound was investigated to establish its mechanism of action and was found to interact with the tetrahydrobiopterin (BH(4)) binding site of nNOS, without interference with any other cofactors or substrate binding sites.     

Calcium/calmodulin-dependent protein kinase I inhibits neuronal nitric-oxide synthase activity through serine 741 phosphorylation

We demonstrate here that neuronal nitric-oxide synthase (nNOS) is phosphorylated and inhibited by a constitutively active form of Ca²⁺/calmodulin (CaM)-dependent protein kinase I (CaM-K I1-293). Substitution of Ser⁷⁴¹ to Ala in nNOS blocked the phosphorylation and the inhibitory effect. Mimicking phosphorylation at Ser⁷⁴¹ by Ser to Asp mutation resulted in decreased binding of and activation by CaM, since the mutation was within the CaM-binding domain. CaM-K I1-293 gave phosphorylation of nNOS at Ser⁷⁴¹ in transfected cells, resulting in 60–70% inhibition of nNOS activity. Wild-type CaM-K I also did phosphorylate nNOS at Ser⁷⁴¹ in transfected cells, but either CaM-K II or CaM-K IV did not. These results raise the possibility of a novel cross-talk between nNOS and CaM-K I through the phosphorylation of Ser⁷⁴¹ on nNOS.     

Structural basis for dipeptide amide isoform-selective inhibition of neuronal nitric oxide synthase

Three nitric oxide synthase (NOS) isoforms, eNOS, nNOS and iNOS, generate nitric oxide (NO) crucial to the cardiovascular, nervous and host defense systems, respectively. Development of isoform-selective NOS inhibitors is of considerable therapeutic importance. Crystal structures of nNOS-selective dipeptide inhibitors in complex with both nNOS and eNOS were solved and the inhibitors were found to adopt a curled conformation in nNOS but an extended conformation in eNOS. We hypothesized that a single-residue difference in the active site, Asp597 (nNOS) versus Asn368 (eNOS), is responsible for the favored binding in nNOS. In the D597N nNOS mutant crystal structure, a bound inhibitor switches to the extended conformation and its inhibition of nNOS decreases >200-fold. Therefore, a single-residue difference is responsible for more than two orders of magnitude selectivity in inhibition of nNOS over eNOS by L-N(omega)-nitroarginine-containing dipeptide inhibitors.     

Inhibition of superoxide generation from neuronal nitric oxide synthase by heat shock protein 90: implications in NOS regulation

Besides NO, neuronal NO synthase (nNOS) also produces superoxide (O(2)(-.) at low levels of L-arginine. Recently, heat shock protein 90 (hsp90) was shown to facilitate NO synthesis from eNOS and nNOS. However, the effect of hsp90 on the O(2)(-.) generation from NOS has not been determined yet. The interrelationship between its effects on O(2)(-.) and NO generation from NOS is also unclear. Therefore, we performed electron paramagnetic resonance measurements of O(2)(-.) generation from nNOS to study the effect of hsp90. Purified rat nNOS generated strong O(2)(-.) signals in the absence of L-arginine. In contrast to its effect on NO synthesis, hsp90 dose-dependently inhibited O(2)(-.) generation from nNOS with an IC(50) of 658 nM. This inhibition was not due to O(2)(-.) scavenging because hsp90 did not affect the O(2)(-.) generated by xanthine oxidase. At lower levels of L-arginine where marked O(2)(-.) generation occurred, hsp90 caused a more dramatic enhancement of NO synthesis from nNOS as compared to that under normal L-arginine. Significant O(2)(-.) production was detected from nNOS even at intracellular levels of L-arginine. Adding hsp90 prevented this O(2)(-.) production, leading to enhanced nNOS activity. Thus, these results demonstrated that hsp90 directly inhibited O(2)(-.) generation from nNOS. Inhibition of O(2)(-.) generation may be an important mechanism by which hsp90 enhances NO synthesis from NOS.