Thermodynamic analysis of interactions between cofactor and neuronal nitric oxide synthase

The thermodynamics of cofactor binding to the isolated reductase domain (Red) of nNOS and its mutants have been studied by isothermal titration calorimetry. The NADP(+) and 2',5'-ADP binding stoichiometry to Red were both 1:1, consistent with a one-site kinetic model instead of a two-site model. The binding constant (K(D) = 71 nM) and the large heat capacity change (ΔC(p) = -440 cal mol(-1) K(-1)) for 2',5'-ADP were remarkably different from those for NADP(+) (1.7 μM and -140 cal mol(-1) K(-1), respectively). These results indicate that the nicotinamide moiety as well as the adenosine moiety has an important role in binding to nNOS. They also suggest that the thermodynamics of the conformational change in Red caused by cofactor binding are significantly different from the conformational changes that occur in cytochrome c reductase, in which the nicotinamide moiety of the cofactor is not essential for binding. Analysis of the deletion mutant of the autoinhibitory helix (RedΔ40) revealed that the deletion resulted in a decrease in the binding affinity of 2',5'-ADP with more unfavorable enthalpy gain. In the case of RedCaM, which contains a calmodulin (CaM) binding site, the presence of Ca(2+)/CaM caused a 6.7-fold increase in the binding affinity for 2',5'-ADP that was mostly due to the favorable entropy change. These results are consistent with a model in which Ca(2+)/CaM induces a conformational change in NOS to a flexible "open" form from a "closed" form that locked by cofactor binding, and this change facilitates the electron transfer required for catalysis.     

The FAD-shielding residue Phe1395 regulates neuronal nitric-oxide synthase catalysis by controlling NADP+ affinity and a conformational equilibrium within the flavoprotein domain

Phe(1395) stacks parallel to the FAD isoalloxazine ring in neuronal nitric-oxide synthase (nNOS) and is representative of conserved aromatic amino acids found in structurally related flavoproteins. This laboratory previously showed that Phe(1395) was required to obtain the electron transfer properties and calmodulin (CaM) response normally observed in wild-type nNOS. Here we characterized the F1395S mutant of the nNOS flavoprotein domain (nNOSr) regarding its physical properties, NADP(+) binding characteristics, flavin reduction kinetics, steady-state and pre-steady-state cytochrome c reduction kinetics, and ability to shield its FMN cofactor in response to CaM or NADP(H) binding. F1395S nNOSr bound NADP(+) with 65% more of the nicotinamide ring in a productive conformation with FAD for hydride transfer and had an 8-fold slower rate of NADP(+) dissociation. CaM stimulated the rates of NADPH-dependent flavin reduction in wild-type nNOSr but not in the F1395S mutant, which had flavin reduction kinetics similar to those of CaM-free wild-type nNOSr. CaM-free F1395S nNOSr lacked repression of cytochrome c reductase activity that is typically observed in nNOSr. The combined results from pre-steady-state and EPR experiments revealed that this was associated with a lesser degree of FMN shielding in the NADP(+)-bound state as compared with wild type. We conclude that Phe(1395) regulates nNOSr catalysis in two ways. It facilitates NADP(+) release to prevent this step from being rate-limiting, and it enables NADP(H) to properly regulate a conformational equilibrium involving the FMN subdomain that controls reactivity of the FMN cofactor in electron transfer.     

Stopped-flow analysis of substrate binding to neuronal nitric oxide synthase.

The kinetics of binding L-arginine and three alternative substrates (homoarginine, N-methylarginine, and N-hydroxyarginine) to neuronal nitric oxide synthase (nNOS) were characterized by conventional and stopped-flow spectroscopy. Because binding these substrates has only a small effect on the light absorbance spectrum of tetrahydrobiopterin-saturated nNOS, their binding was monitored by following displacement of imidazole, which displays a significant change in Soret absorbance from 427 to 398 nm. Rates of spectral change upon mixing Im-nNOS with increasing amounts of substrates were obtained and found to be monophasic in all cases. For each substrate, a plot of the apparent rate versus substrate concentration showed saturation at the higher concentrations. K(-)(1), k(2), k(-)(2), and the apparent dissociation constant were derived for each substrate from the kinetic data. The dissociation constants mostly agreed with those calculated from equilibrium spectral data obtained by titrating Im-nNOS with each substrate. We conclude that nNOS follows a two-step, reversible mechanism of substrate binding in which there is a rapid equilibrium between Im-nNOS and the substrate S followed by a slower isomerization process to generate nNOS'-S: Im-nNOS + S if Im-nNOS-S if nNOS'-S + Im. All four substrates followed this general mechanism, but differences in their kinetic values were significant and may contribute to their varying capacities to support NO synthesis.     

Catalase activity of oxygenase domain of rat neuronal nitric oxide synthase. Evidence for product formation from L-arginine

Nitric oxide synthases (NOSs) catalyze the formation of nitric oxide from L-arginine. We purified the heme containing, tetrahydrobiopterin-free, oxygenase domain of rat neuronal nitric oxide synthase (nNOSox) overexpressed in Escherichia coli. We found catalase activity in nNOSox. This is significant because H(2)O(2) may also be a product of nitric oxide synthases. We found H(2)O(2) assisted product formation from N-hydroxy-L-arginine and even from L-arginine both in the presence and in absence of tetrahydrobiopterin. We propose how heme moiety of the oxygenase domain alone is sufficient to carry out both steps of the NOS catalysis.     

Expression of neuronal nitric oxide synthase variant, nNOS-mu, in rat brain.

Neuronal nitric oxide synthase (nNOS) is alternatively spliced. An nNOS splice variant form, nNOS-mu, was first found to be selectively expressed in rat skeletal muscle and heart. To date, the expression of nNOS-mu in the brain has not been well characterized. The aim of this study was to determine whether nNOS-mu is expressed in rat brain, and whether nNOS-mu exhibits a specific expression pattern. To analyze the expression of nNOS-mu, we generated a monoclonal antibody that is specific for nNOS-mu. An immunoblot analysis using this antibody showed that nNOS-mu is expressed in the rat brain at a measurable level, which was 10.3% of total nNOSs. In rat brain, the nNOS-mu expression was high in the mesencephalon and the cerebellum. nNOS-mu was immunohistochemically localized in neurites and perikarya of large neurons. In the cerebellum, granule cells showed marked staining, while weak staining was detected in basket and stellate cells. This expression pattern is different from that described for nNOS and suggests that nNOS-mu plays unique roles in different neurons.     

Potentiometric analysis of the flavin cofactors of neuronal nitric oxide synthase

Midpoint reduction potentials for the flavin cofactors in the reductase domain of rat neuronal nitric oxide synthase (nNOS) in calmodulin (CaM)-free and -bound forms have been determined by direct anaerobic titration. In the CaM-free form, the FMN potentials are -49 +/- 5 mV (oxidized/semiquinone) -274 +/- 5 mV (semiquinone/reduced). The corresponding FAD potentials are -232 +/- 7, and -280 +/- 6 mV. The data indicate that each flavin can exist as a blue (neutral) semiquinone. The accumulation of blue semiquinone on the FMN is considerably higher than seen on the FAD due to the much larger separation (225 mV) of its two potentials (cf. 48 mV for FAD). For the CaM-bound form of the protein, the midpoint potentials are essentially identical: there is a small alteration in the FMN oxidized/semiquinone potential (-30 +/- 4 mV); the other three potentials are unaffected. The heme midpoint potentials for nNOS [-239 mV, L-Arg-free; -220 mV, L-Arg-bound; Presta, A., Weber-Main, A. M., Stankovich, M. T., and Stuehr, D. J. (1998) J. Am. Chem. Soc. 120, 9460-9465] are poised such that electron transfer from flavin domain is thermodynamically feasible. Clearly, CaM binding is necessary in eliciting conformational changes that enhance flavin to flavin and flavin to heme electron transfers rather than causing a change in the driving force.     

Regulation of the successive reaction catalyzed by rat neuronal nitric oxide synthase

The rat neuronal nitric oxide synthase (nNOS) catalyzes two monooxygenase reactions successively from L-arginine (L-Arg) to L-citrulline (L-Cit) via N(omega)-hydroxy-L-arginine (OH-Arg) without most of OH-Arg leaving the substrate-binding site. In the steady-state reaction conditions, the amount of OH-Arg produced is about 1/30-1/50 that of L-Cit. We found in this study using nNOS purified from an Escherichia coli expression system that the ratio of the amount of OH-Arg to L-Cit (OH-Arg/L-Cit) increased to about 1 at low concentration of NADPH. In one cycle of the nNOS reaction, the decrease in NADPH concentration was found to reduce the rates of two monooxygenase reactions but had little effect on the rate constant of OH-Arg dissociation from the enzyme. The addition of NADP(+), the competitive inhibitor for NADPH, caused the decrease in the rates of monooxygenase reactions in a single cycle of the reaction and the increase in the ratio of OH-Arg/L-Cit in the steady state. At low CaM concentrations, the ratio of OH-Arg/L-Cit was about the same as that at high CaM. In a single cycle of the nNOS reaction, the rate of monooxygenation was not altered by the CaM concentration but the amount of metabolized L-Arg decreased with the decrease in CaM concentration, showing that the amount of active nNOS was regulated by complex formation between nNOS and CaM. It becomes clear that there are two regulatory mechanisms for the successive reaction of nNOS. One controls the rates of monooxygenations and the other controls the amount of active species of nNOS.     

Expression of neuronal nitric oxide synthase in several cell types of the rat gastric epithelium.

The aim of this study was to identify which cell types of the rat gastric epithelium express neuronal nitric oxide synthase (nNOS) because the results of the previous studies have been very divergent regarding this point. By the combination of immunohistochemical (IHC) and in situ hybridization (ISH) techniques, we detected expression of nNOS in chief and mucosecretory cells of the gastric epithelium. Moreover, some gastric endocrine cells were immunoreactive for nNOS, although they could not be distinguished in sections treated with ISH techniques. The strongest signal for all antibodies in IHC techniques was obtained when microwave (MW) heating was performed before the IHC procedure. Our results indicate that in the gastric epithelium a variety of cell types are able to produce NO. The NO produced by the different cell types (chief, mucous, and endocrine) may form a complex network of paracrine communication with an important role in gastric physiology.     

Thermodynamic analysis of L-arginine and N omega-hydroxy-L-arginine binding to nitric oxide synthase

Isothermal titration calorimetry has been used to determine thermodynamic parameters of substrate binding to the oxygenase domain of neuronal nitric oxide synthase (nNOS(oxy)) in the presence of the cofactor tetrahydrobiopterin. The intermediate N(omega)-hydroxy-L-arginine (NHA) has a larger affinity than L-Arginine (L-Arg) for nNOS(oxy), with K(d)=0.4+/-0.1 microM and 1.7+/-0.3 microM at 25 degrees C, respectively. nNOS(oxy) binds NHA and L-Arg with DeltaH -4.1+/-0.2 and -1.0+/-0.1 kcal/mol and DeltaS=15 and 23 cal/Kmol respectively. NHA binding is more exothermic probably due to formation of an extra hydrogen bond in the active site compared to L-Arg. The changes in heat capacity (DeltaC(p)) are relatively small for binding of both NHA and L-Arg (-53+/-18 and -95+/-23 cal/L mol, respectively), which indicates that hydrophobic interactions contribute little to binding.     

Gene therapy of erectile dysfunction in the rat with penile neuronal nitric oxide synthase

Gene transfer to the penile corpora cavernosa of constructs of the inducible and endothelial nitric oxide synthase (NOS) cDNAs ameliorates erectile dysfunction in aged rats. In this study, we investigated whether the neuronal NOS (nNOS) variant responsible for erection, penile nNOS (PnNOS), can exert a similar effect, and whether the combination of electroporation with a helper-dependent adenovirus (AdV) improves gene transfer. PnNOS and beta-galactosidase cDNAs were cloned in plasmid (pCMV-PnNOS; pCMV-beta-gal) and "gutless" AdV (AdV-CMV-PnNOS; AdV-CMV-beta-gal) vectors, and injected into the penis of adult (beta-gal) or aged (PnNOS) rats, with or without electroporation. Penile erection was measured at different times after PnNOS cDNA injection, by electrical field stimulation of the cavernosal nerve. The expression of beta-galactosidase or PnNOS was estimated in penile tissue by either histochemistry and luminometry or Western blot, and the effects of AdV-CMV-PnNOS on mRNA expression were examined by a DNA microarray. We found that electroporation increased pCMV-beta-gal uptake, and its expression was detectable at 56 days. In the aged rats treated with pCMV-PnNOS and electroporation, the maximal intracavernosal:mean arterial pressure ratios were elevated for 11 and 18 days when compared with those in controls. Electroporation intensified penile uptake of as few as 10(6) viral particles (vp) of AdV-CMV-beta-gal, and with 10(7) vp beta-galactosidase was still detectable at 60 days. Electroporated AdV-CMV-PnNOS (10(7) vp) was effective at 18 days in stimulating the erection of aged rats, without inducing the expression of cytotoxic genes. In conclusion, intracavernosal gene therapy with PnNOS cDNA corrected the aging-related erectile dysfunction for at least 18 days when given by electroporation in a helper-dependent AdV at low viral loads.     

Solution structure of the extended neuronal nitric oxide synthase PDZ domain complexed with an associated peptide.

The PDZ domain of neuronal nitric oxide synthase (nNOS) functions as a scaffold for organizing the signal transduction complex of the enzyme. The NMR structure of a complex composed of the nNOS PDZ domain and an associated peptide suggests that a two-stranded beta-sheet C-terminal to the canonical PDZ domain may mediate its interaction with the PDZ domains of postsynaptic density-95 and alpha-syntrophin. The structure also provides the molecular basis of recognition of Asp-X-Val-COOH peptides by the nNOS PDZ domain. The role of the C-terminal extension in Asp-X-Val-COOH peptide binding is investigated. Additionally, NMR studies further show that the Asp-X-Val-COOH peptide and a C-terminal peptide from a novel cytosolic protein named CAPON bind to the same pocket of the nNOS PDZ domain.     

Gene therapy of erectile dysfunction in the rat with penile neuronal nitric oxide synthase

Gene transfer to the penile corpora cavernosa of constructs of the inducible and endothelial nitric oxide synthase (NOS) cDNAs ameliorates erectile dysfunction in aged rats. In this study, we investigated whether the neuronal NOS (nNOS) variant responsible for erection, penile nNOS (PnNOS), can exert a similar effect, and whether the combination of electroporation with a helper-dependent adenovirus (AdV) improves gene transfer. PnNOS and beta-galactosidase cDNAs were cloned in plasmid (pCMV-PnNOS; pCMV-beta-gal) and "gutless" AdV (AdV-CMV-PnNOS; AdV-CMV-beta-gal) vectors, and injected into the penis of adult (beta-gal) or aged (PnNOS) rats, with or without electroporation. Penile erection was measured at different times after PnNOS cDNA injection, by electrical field stimulation of the cavernosal nerve. The expression of beta-galactosidase or PnNOS was estimated in penile tissue by either histochemistry and luminometry or Western blot, and the effects of AdV-CMV-PnNOS on mRNA expression were examined by a DNA microarray. We found that electroporation increased pCMV-beta-gal uptake, and its expression was detectable at 56 days. In the aged rats treated with pCMV-PnNOS and electroporation, the maximal intracavernosal:mean arterial pressure ratios were elevated for 11 and 18 days when compared with those in controls. Electroporation intensified penile uptake of as few as 10(6) viral particles (vp) of AdV-CMV-beta-gal, and with 10(7) vp beta-galactosidase was still detectable at 60 days. Electroporated AdV-CMV-PnNOS (10(7) vp) was effective at 18 days in stimulating the erection of aged rats, without inducing the expression of cytotoxic genes. In conclusion, intracavernosal gene therapy with PnNOS cDNA corrected the aging-related erectile dysfunction for at least 18 days when given by electroporation in a helper-dependent AdV at low viral loads.     

CO binding to the isolated oxygenase domain of neuronal nitric oxide synthase: effects of inhibitors and mutations at the substrate-binding site

In order to understand the heme distal structure of neuronal nitric oxide synthase (nNOS), we studied the binding affinity of CO for the ferrous wild type enzyme and the Glu592Ala and Tyr588His substrate binding-site mutants (generated in the oxygenase domain, nNOSox) in the presence of substrates and inhibitors. The CO binding affinities (K(d) = 10-21 mM) of nNOSox in the presence of the substrates, L-Arg and NHA, or inhibitors such as NAME and agmatine were more than two-fold lower than in their absence (K(d) = 5 mM). The presence of NIL strongly inhibited CO binding and gave a K(d) of more than 100 mM. These effects were not observed for the Glu592Ala mutant. The trend in CO binding affinities observed for the Tyr588His mutant was similar to that of the wild type enzyme. The presence of the isolated reductase domain did not affect CO binding. We discuss the role of substrate and inhibitor binding in CO complexation as well as in catalysis.     

Crystal structure of the FAD/NADPH-binding domain of rat neuronal nitric-oxide synthase. Comparisons with NADPH-cytochrome P450 oxidoreductase

Nitric-oxide synthase (NOS) is composed of a C-terminal, flavin-containing reductase domain and an N-terminal, heme-containing oxidase domain. The reductase domain, similar to NADPH-cytochrome P450 reductase, can be further divided into two different flavin-containing domains: (a) the N terminus, FMN-containing portion, and (b) the C terminus FAD- and NADPH-binding portion. The crystal structure of the FAD/NADPH-containing domain of rat neuronal nitric-oxide synthase, complexed with NADP(+), has been determined at 1.9 A resolution. The protein is fully capable of reducing ferricyanide, using NADPH as the electron donor. The overall polypeptide fold of the domain is very similar to that of the corresponding module of NADPH-cytochrome P450 oxidoreductase (CYPOR) and consists of three structural subdomains (from N to C termini): (a) the connecting domain, (b) the FAD-binding domain, and (c) the NADPH-binding domain. A comparison of the structure of the neuronal NOS FAD/NADPH domain and CYPOR reveals the strict conservation of the flavin-binding site, including the tightly bound water molecules, the mode of NADP(+) binding, and the aromatic residue that lies at the re-face of the flavin ring, strongly suggesting that the hydride transfer mechanisms in the two enzymes are very similar. In contrast, the putative FMN domain-binding surface of the NOS protein is less positively charged than that of its CYPOR counterpart, indicating a different nature of interactions between the two flavin domains and a different mode of regulation in electron transfer between the two flavins involving the autoinhibitory element and the C-terminal 33 residues, both of which are absent in CYPOR.     

Inhibition of neuronal nitric oxide synthase by N-phenacyl imidazoles.

Nitric oxide (NO) mediates a series of physiological processes, including regulation of vascular tone, macrofage-mediated neurotoxicity, platelet aggregation, learning and long-term potentiation, and neuronal transmission. Although NO mediates several physiological functions, overproduction of NO can be detrimental and play multiple roles in several pathological diseases. Accordingly, more potent inhibitors, more selective for neuronal nitric oxide synthase (nNOS) than endothelial NOS (eNOS) or inducible NOS (iNOS), could be useful in the treatment of cerebral ischemia and other neurodegenerative diseases. We recently described the synthesis of a series of imidazole derivatives. Among them N-(4-nitrophenacyl) imidazole (A) and N-(4-nitrophenacyl)-2-methyl-imidazole (B) were considered selective nNOS inhibitors. In the present study the action mechanism of compounds A and B was analyzed. Spectral changes observed in the presence of compound A indicate that this inhibitor exerts its effect without interaction with heme iron. Moreover compounds A and B, inhibit nNOS "noncompetitively" versus arginine, but "competitively" versus BH(4).     

Role of neuronal and endothelial nitric oxide synthase in nitric oxide generation in the brain following cerebral ischemia.

Nitric oxide (NO) plays an important role in the pathogenesis of neuronal injury during cerebral ischemia. The endothelial and neuronal isoforms of nitric oxide synthase (eNOS, nNOS) generate NO, but NO generation from these two isoforms can have opposing roles in the process of ischemic injury. While increased NO production from nNOS in neurons can cause neuronal injury, endothelial NO production from eNOS can decrease ischemic injury by inducing vasodilation. However, the relative magnitude and time course of NO generation from each isoform during cerebral ischemia has not been previously determined. Therefore, electron paramagnetic resonance spectroscopy was applied to directly detect NO in the brain of mice in the basal state and following global cerebral ischemia induced by cardiac arrest. The relative amount of NO derived from eNOS and nNOS was accessed using transgenic eNOS(-/-) or nNOS(-/-) mice and matched wild-type control mice. NO was trapped using Fe(II)-diethyldithiocarbamate. In wild-type mice, only small NO signals were seen prior to ischemia, but after 10 to 20 min of ischemia the signals increased more than 4-fold. This NO generation was inhibited more than 70% by NOS inhibition. In either nNOS(-/-) or eNOS(-/-) mice before ischemia, NO generation was decreased about 50% compared to that in wild-type mice. Following the onset of ischemia a rapid increase in NO occurred in nNOS(-/-) mice peaking after only 10 min. The production of NO in the eNOS(-/-) mice paralleled that in the wild type with a progressive increase over 20 min, suggesting progressive accumulation of NO from nNOS following the onset of ischemia. NOS activity measurements demonstrated that eNOS(-/-) and nNOS(-/-) brains had 90% and < 10%, respectively, of the activity measured in wild type. Thus, while eNOS contributes only a fraction of total brain NOS activity, during the early minutes of cerebral ischemia prominent NO generation from this isoform occurs, confirming its importance in modulating the process of ischemic injury.     

Low-temperature stabilization and spectroscopic characterization of the dioxygen complex of the ferrous neuronal nitric oxide synthase oxygenase domain.

Nitric oxide (NO), an intercellular messenger and an immuno-cytotoxic agent, is synthesized by the family of nitric oxide synthases (NOS), which are thiolate-ligated, heme-containing monooxygenases that convert L-Arg to L-citrulline and NO in a tetrahydrobiopterin (BH4)-dependent manner, using NADPH as the electron donor. The dioxygen complex of the ferrous enzyme has been proposed to be a key intermediate in the NOS catalytic cycle. In this study, we have generated a stable ferrous-O2 complex of the oxygenase domain of rat neuronal NOS (nNOS) by bubbling O2 through a solution of the dithionite-reduced enzyme at -30 degrees C in a cryogenic solvent containing 50% ethylene glycol. The most stable dioxygen complex is obtained using the oxygenase domain which has been preincubated for an extended length of time at 4 degrees C with BH4/dithiothreitol and NG-methyl-L-arginine, a substrate analogue inhibitor. The O2 complex of the nNOS oxygenase domain thus prepared exhibits UV-visible absorption (maxima at 419 and 553 nm, shoulder at approximately 585 nm) and magnetic circular dichroism spectra that are nearly identical to those of ferrous-O2 cytochrome P450-CAM. Our spectral data are noticeably blue-shifted from those seen at 10 degrees C for a short-lived transient species (lambdamax = 427 nm) for the nNOS oxygenase domain using stopped-flow rapid-scanning spectroscopy [Abu-Soud, H. M., Gachhui, R., Raushel, F. M., and Stuehr, D. J. (1997) J. Biol. Chem. 272, 17349], but somewhat similar to those of a relatively stable O2 adduct of L-Arg-free full-length nNOS (lambdamax = 415-416.5 nm) generated at -30 degrees C [Bec, N., Gorren, A. C. F., Voelder, C., Mayer, B., and Lange, R. (1998) J. Biol. Chem. 273, 13502]. Compared with ferrous-O2 P450-CAM, however, the ferrous-O2 adduct of the nNOS oxygenase domain is considerably more autoxidizable and the O2-CO exchange reaction is noticeably slower. The generation of a stable ferrous-O2 adduct of the nNOS oxygenase domain, as described herein, will facilitate further mechanistic and spectroscopic investigations of this important intermediate.