Effectors of HIV-1 protease peptidolytic activity.

High concentrations of salts dramatically affect the interaction of small ligands with HIV-1 protease. For instance, the Km and kcat values for Abz-Thr-Ile-Nle-p-nitro-Phe-Gln-Arg-NH2 (S) increased 120-fold and 3-fold, respectively, as the NaCl concentration in the assay decreased from 4.0 to 0.5 M. The Kd value for the competitive inhibitor amprenavir increased 12-fold over this concentration range of NaCl. The bimolecular rate constant for association of enzyme with amprenavir was independent of NaCl concentration, whereas the dissociation rate constant decreased with increasing NaCl concentration. Polyanionic polymers such as heparin or poly A substituted for NaCl. For example, the value of kcat/Km for S was 0.18 microM(-1) x s(-1) when the enzyme (<10 nM) was assayed in the standard buffer supplemented with 5 mM NaCl. If 0.01% poly A were included, the value of kcat/Km increased to 8.6 microM(-1) x s(-1). A DNA oligomer (23-mer) with an hexachlorofluoresceinyl moiety linked to the 5' end was studied as a model polyanionic polymer. The enzyme bound HF23 (Kd < 1 nM) with concomitant quenching of the hexachlorofluoresceinyl fluorescence. The stoichiometry for binding was 3 mol of enzyme per mol of oligomer. The hydrolytic activity of the enzyme with this oligomer was similar to that observed with poly A or high salt concentration when the molar ratio of oligomer to enzyme was greater than one. The results presented herein demonstrate that polyanionic polymers substitute for salts as effectors of HIV protease.

Metabolic complications associated with antiretroviral therapy.

Mortality rates in the HIV-infected patient population have decreased with the advent of highly active antiretroviral therapy (HAART) for the treatment of AIDS. Due to the chronic nature of HAART, long-term metabolic complications are associated with therapy, such as hyperlipidemia, fat redistribution and diabetes mellitus. Currently, all of these symptoms are classified as the lipodystrophy (LD) syndrome(s). However, hyperlipidemia and fat redistribution occur independently, indicating there may be multiple syndromes associated with HAART. Although fat gain/loss and dyslipidemia occur in protease inhibitor (PI) naïve patients treated with nucleoside reverse transcriptase inhibitors (NRTIs), combination therapies (PI and NRTI) accelerate the syndrome. Recent clinical trials, cell culture and animal studies indicate that these effects are not drug class specific and select PIs, NRTIs and non-nucleoside reverse transcriptase inhibitors (NNRTIs) can be associated with metabolic complications. Moreover, the effects can vary between various members of the same class of antiretroviral agents (i.e. not all PIs cause the same adverse reactions) and may be influenced by duration of infection, genetics and environmental factors. Although HAART increases the risk of metabolic complications, this does not outweigh the benefits of survival. In this review, we summarize the latest clinical and scientific information on these metabolic complications, examine current hypotheses explaining the syndromes and comment on the existing methods available to manage these metabolic side effects.

HIV protease inhibitors block adipogenesis and increase lipolysis in vitro.

AIDS therapies employing HIV protease inhibitors (PIs) are associated with changes in fat metabolism. However, the cellular mechanisms affected by PIs are not clear. Thus, the affects of PIs on adipocyte differentiation were examined in vitro using C3H10T1/2 stem cells. In these cells the PIs, nelfinavir, saquinavir, and ritonavir, reduced triglyceride accumulation, lipogenesis, and expression of the adipose markers, aP2 and LPL. Histological analysis revealed nelfinavir, saquinavir and ritonavir treatment decreased oil red O-staining of cytoplasmic fat droplets. Inhibition occurred in the presence of the RXR agonist LGD1069, indicating the inhibitory effects were not due to an absence of RXR ligand. Moreover, these three PIs increased acute lipolysis in adipocytes. In contrast, two HIV PIs, amprenavir and indinavir, had little effect on lipolysis, lipogenesis, or expression of aP2 and LPL. Although, saquinavir, inhibited ligand-binding to PPARgamma with an IC(50) of 12.7+/-3.2 microM, none of the other PIs bound to the nuclear receptors RXRalpha or PPARgamma, (IC(50)s>20 microM), suggesting that inhibition of adipogenesis is not due to antagonism of ligand binding to RXRalpha or PPARgamma. Taken together, the results suggest that some, but not all, PIs block adipogenesis and stimulate fat catabolism in vitro and this may contribute to the effects of PIs on metabolism in the clinic.

Dietary fat alters HIV protease inhibitor-induced metabolic changes in mice.

Human immunodeficiency virus (HIV) protease inhibitors (PI) may alter lipid metabolism in patients with acquired immunodeficiency syndrome (AIDS). However, the influence of dietary fat on the metabolic effects of PI therapy remains unknown. AKR/J mice were fed high or low fat diets and treated with the PI indinavir (IDV), nelfinavir (NFV), saquinavir (SQV) or amprenavir (APV) by subcutaneous delivery for 2 wk. Serum concentrations of glucose, insulin, triglyceride, free fatty acid, glycerol, pancreatic lipase, bilirubin, alkaline phosphatase, blood urea nitrogen and PI, and interscapular and epididymal fat weights were determined. Some metabolic effects of PI were dependent on diet. IDV- and NFV-treated mice had greater serum glucose concentration and body weight; IDV-treated mice had lower serum insulin; NFV-treated mice had greater interscapular fat mass; and SQV treated mice had lower serum triglyceride concentration than control mice fed the low but not the high fat diet. In contrast, NFV- and IDV-treated mice had greater triglyceride concentration and blood urea nitrogen, and SQV treated mice had greater serum cholesterol than control mice fed the high but not the low fat diet. The serum concentration of SQV was lower in mice fed the high fat compared with the low fat diet. Other effects were not dependent on diet. IDV- and NFV-treated mice had greater fatty acids, and IDV-treated mice had greater pancreatic lipase, bilirubin and alkaline phosphatase than control mice fed either diet. APV treatment had little effect on these serum measurements. Thus, changes in dietary fat can influence some but not all of the effects of PI on metabolism. Furthermore, each PI produces different effects in vivo, indicating that various PI affect distinct metabolic pathways.

Novel inhibitors of HIV protease: design, synthesis and biological evaluation of picomolar inhibitors containing cyclic P1/P2 scaffolds.

A novel series of HIV protease inhibitors containing cyclic P1/P2 scaffolds has been synthesized and evaluated for biological activity. The trans 3,5-dibenzyl-2-oxo pyrrolidinone ring system resulted in a 50 pM enzyme inhibitor against HIV protease in vitro when combined with an indanolamine derived P'-backbone. This compound also shows comparable activity to currently marketed drugs in the MT-4 cell-based antiviral assay.

Stimulation of vitamin A(1) acid signaling by the HIV protease inhibitor indinavir.

HIV protease inhibitors (PIs) are effective drugs for the treatment of AIDS. However, PI therapy is sometimes associated with side-effects including increased plasma lipids and altered body fat distribution, although fat redistribution may occur in some patients not treated with PIs. Overdosage with vitamin A(1) acid (all-trans-retinoic acid, ATRA) or its metabolites may cause similar changes in lipid metabolism. Moreover, the PI indinavir and retinoids have been associated with nail, skin, and hair defects, suggesting that indinavir and retinoids may exert their effects through similar molecular mechanisms. This hypothesis was tested by examining the effects of PIs on retinoid signaling in vitro. Mesenchymal stem cells (C3H10T1/2) were cultured in the presence of various PIs (amprenavir, indinavir, nelfinavir, ritonavir, and saquinavir) and synthetic retinoids, and the metabolic response was assessed by measuring the activity of a retinoid-regulated protein, alkaline phosphatase (ALP). Of the PIs tested, only indinavir stimulated ATRA-dependent ALP activity and altered stem cell morphology; the effects of indinavir occurred in the presence of ATRA, but not in its absence. Moreover, indinavir increased the effects of ATRA on lipid accumulation during fat cell differentiation. AGN 193109 (4-[[5,6-dihydro-5, 5-dimethyl-8-(4-methylphenyl)-2-naphthalenyl]ethynyl]-benzoic acid), a retinoic acid receptor (RAR) antagonist, inhibited the synergistic effects of indinavir and ATRA, indicating that indinavir increases RAR signaling. However, indinavir did not potentiate ALP activity in the presence of the RAR agonist CH55 (3,5-di-tert-butylchalcone 4'-carboxylic acid). Unlike ATRA, CH55 does not bind to cytosolic retinoic acid binding protein (CRABP), suggesting that CRABP may regulate the effects of indinavir on RAR signaling. These observations support the proposal that altered retinoid signaling promotes some of the adverse reactions associated with indinavir therapy, such as altered lipid metabolism.

Identifying and Characterizing HIV Protease Inhibitors.

HIV protease catalyzes the hydrolysis of specific peptide bonds of viral polyproteins, thus processing these polyproteins into their active components. These protein processing reactions are requisite for viral replication. Therefore, the HIV protease is an ideal target for the chemotherapeutic treatment of HIV disease (1-3). HIV protease is an aspartyl protease,and the aspartyl protease inhibitor, pepstatin, was one of the first identified inhibitors of HIV protease. More potent inhibitors have been designed and synthesized since, in fact, four protease inhibitors-saquinavir (Ro-31,8959), ritonavir (ABT-538), indinavir (L-735,524), and Nelfinavir (AG1343)-are effective in clinical trials to treat HIV disease (4-6) and recently were approved by the Food and Drug Administration for the chemotherapeutic treatment of HIV infections. Other protease inhibitors in clinical trials are VX-478 (141W94) and ABT-378. Notwithstanding these early successes, it is difficult to comply with these drug's dosing regiments. Furthermore, viral resistance to individual inhibitors and cross-resistance to multiple inhibitors occur in vivo (7,8). Therefore, a medical need still exists for new HIV protease inhibitors with different resistance profiles.

N-Phenylamidines as selective inhibitors of human neuronal nitric oxide synthase: structure-activity studies and demonstration of in vivo activity.

Selective inhibition of the neuronal isoform of nitric oxide synthase (NOS) compared to the endothelial and inducible isoforms may be required for treatment of neurological disorders caused by excessive production of nitric oxide. Recently, we described N-(3-(aminomethyl)benzyl)acetamidine (13) as a slow, tight-binding inhibitor, highly selective for human inducible nitric oxide synthase (iNOS). Removal of a single methylene bridge between the amidine nitrogen and phenyl ring to give N-(3-(aminomethyl)phenyl)acetamidine (14) dramatically altered the selectivity to give a neuronal selective nitric oxide synthase (nNOS) inhibitor. Part of this large shift in selectivity was due to 14 being a rapidly reversible inhibitor of iNOS in contrast to the essentially irreversible inhibition of iNOS observed with 13. Structure-activity studies revealed that a basic amine functionality tethered to an aromatic ring and a sterically compact amidine are key pharmacophores for this class of NOS inhibitors. Maximal nNOS inhibition potency was achieved with N-(3-(aminomethyl)phenyl)-2-furanylamidine (77) (Ki-nNOS = 0.006 microM; Ki-eNOS = 0.35 microM; Ki-iNOS = 0.16 microM). Finally, alpha-fluoro-N-(3-(aminomethyl)phenyl)acetamidine (74) (Ki-nNOS = 0. 011 microM; Ki-eNOS = 1.1 microM; Ki-iNOS = 0.48 microM) had excellent brain penetration and inhibited nNOS in a rat brain slice assay as well as in the rat brain (cerebellum) in vivo. Thus, N-phenylamidines should be useful in validating the role of nNOS in neurological disorders.

Substituted N-phenylisothioureas: potent inhibitors of human nitric oxide synthase with neuronal isoform selectivity.

S-Ethyl N-phenylisothiourea (4) has been found to be a potent inhibitor of both the human constitutive and inducible isoforms of nitric oxide synthase. A series of substituted N-phenylisothiourea analogues was synthesized to investigate the structure-activity relationship of this class of inhibitor. Each analogue was evaluated for human isoform selectivity. One analogue, S-ethyl N-[4-(trifluoromethyl)phenyl]isothiourea (39), exhibited 115-fold and 29-fold selectivity for the neuronal isoform versus the inducible and endothelial derived constitutive isoforms, respectively. Studies have shown the substituted N-phenylisothiourea 39 binds competitively with L-arginine.

Substrate binding is required for release of product from mammalian protein farnesyltransferase.

Protein farnesyltransferase (FTase) catalyzes the modification by a farnesyl lipid of Ras and several other key proteins involved in cellular regulation. Previous studies on this important enzyme have indicated that product dissociation is the rate-limiting step in catalysis. A detailed examination of this has now been performed, and the results provide surprising insights into the mechanism of the enzyme. Examination of the binding of a farnesylated peptide product to free enzyme revealed a binding affinity of approximately 1 microM. However, analysis of the product release step under single turnover conditions led to the surprising observation that the peptide product did not dissociate from the enzyme unless additional substrate was provided. Once additional substrate was provided, the enzyme released the farnesylated peptide product with rates comparable with that of overall catalysis by FTase. Additionally, stable FTase-farnesylated product complexes were formed using Ras proteins as substrates, and these complexes also require additional substrate for product release. These data have major implications in both our understanding of overall mechanism of this enzyme and in design of inhibitors against this therapeutic target.

1400W is a slow, tight binding, and highly selective inhibitor of inducible nitric-oxide synthase in vitro and in vivo.

N-(3-(Aminomethyl)benzyl)acetamidine (1400W) was a slow, tight binding inhibitor of human inducible nitric- oxide synthase (iNOS). The slow onset of inhibition by 1400W showed saturation kinetics with a maximal rate constant of 0.028 s-1 and a binding constant of 2.0 microM. Inhibition was dependent on the cofactor NADPH. L-Arginine was a competitive inhibitor of 1400W binding with a Ks value of 3.0 microM. Inhibited enzyme did not recover activity after 2 h. Thus, 1400W was either an irreversible inhibitor or an extremely slowly reversible inhibitor of human iNOS with a Kd value

Potent inhibition of human neuronal nitric oxide synthase by N(G)-nitro-L-arginine methyl ester results from contaminating N(G)-nitro-L-arginine.

N(G)-Nitro-L-arginine methyl ester (L-NAME), inhibits the three isozymes of nitric oxide synthase (NOS) in vitro and in vivo. The mechanism of NOS inhibition by L-NAME is uncertain. L-NAME was a time-dependent inhibitor of neuronal NOS (nNOS). Concommitantly, L-NAME was hydrolyzed, non-enzymatically, to N(G)-Nitro-L-arginine (L-NA) during the enzyme assay. The time-dependent inhibition of nNOS by L-NAME was the result of this time-dependent formation of L-NA. Furthermore, N(G)-Nitro-L-arginine methyl amide, which is an isosteric analogue of L-NAME that is much less susceptible to hydrolysis, was a rapidly reversible weak inhibitor of NOS. These data suggested that L-NAME itself was a weak and rapidly reversible inhibitor of nNOS. Most of the inhibition of nNOS by a solution of L-NAME is the result of the formation of L-NA. L-NAME was a substrate for porcine liver esterase.

Human immunodeficiency virus. Mutations in the viral protease that confer resistance to saquinavir increase the dissociation rate constant of the protease-saquinavir complex.

Mutations in the human immunodeficiency virus (HIV) protease (L90M, G48V, and L90M/G48V) arise when HIV is passaged in the presence of the HIV protease inhibitor saquinavir. These mutations yield a virus with less sensitivity to the drug (L90M > G48V >> L90M/G48V). L90M, G48V, and L90M/G48V proteases have 1/20, 1/160, and 1/1000 the affinity for saquinavir compared to WT protease, respectively. Therefore, the affinity of mutant protease for saquinavir decreased as the sensitivity of the virus to saquinavir decreased. Association rate constants for WT and mutant proteases with saquinavir were similar, ranging from 2 to 4 x 10(7) M-1 s-1. In contrast, the dissociation rate constants for WT, L90M, G48V, and L90M/G48V proteases complexed with saquinavir were 0.0014, 0.019, 0.128, and 0. 54 s-1, respectively. This indicated that the reduced affinity for mutant proteases and saquinavir is primarily the result of larger dissociation rate constants. The increased dissociation rate constants may be the result of a decrease in the internal equilibrium between the bound inhibitor with the protease flaps up and the bound inhibitor with the flaps down. Interestingly, the affinity of these mutant proteases for VX-478, ABT-538, AG-1343, or L-735,524 was not reduced as much as that for saquinavir. Finally, the catalytic constants of WT and mutant proteases were determined for eight small peptide substrates that mimic the viral cleavage sites in vivo. WT and L90M proteases had similar catalytic constants for these substrates. In contrast, G48V and L90M/G48V proteases had catalytic efficiency (kcat/Km) values with TLNF-PISP, RKIL-FLDG, and AETF-YVDG that were 1/10 to 1/20 the value of WT protease. The decreased catalytic efficiencies were primarily the result of increased Km values. Thus, mutations in the protease decrease the affinity of the enzyme for saquinavir and the catalytic efficiency with peptide substrates.

Protein farnesyltransferase: kinetics of farnesyl pyrophosphate binding and product release.

Protein farnesyltransferase (FTase) catalyzes the prenylation of Ras and several other key proteins involved in cell regulation. The mechanism of the FTase reaction was elucidated by pre-steady-state and steady-state kinetic analysis. FTase catalyzed the farnesylation of biotinylated peptide substrate (BiopepSH) by farnesyl pyrophosphate (FPP) to an S-farnesylated peptide (BiopepS-C15). The steady-state kinetic mechanism was ordered. FTase bound FPP in a two-step process with an effective dissociation rate constant of 0.013 s-1 and an overall Kd of 2.8 nM. BiopepSH reacted with FTase.FPP irreversibly, with a second-order rate constant of 2.2 x 10(5) M-1 s-1, to form FTase.BiopepS-C15. Because most of the FPP in FTase.FPP was trapped as FTase.BiopepS-C15 at high concentrations of BiopepSH, FPP dissociated slowly from the ternary complex relative to catalysis, so that the commitment to catalysis was high. The maximal rate constant for formation of FTase.BiopepS-C15 (enzyme-bound product) is much larger than kcat (0.06 s-1), indicating that product release is the rate-determining step in the reaction mechanism.

Potent and selective inhibition of human nitric oxide synthases. Inhibition by non-amino acid isothioureas.

S-Ethylisothiourea was a potent competitive inhibitor of human nitric oxide synthase (NOS), with Ki values of 17, 36, and 29 nM for the inducible (i), endothelial (e), and neuronal (n) isozymes, respectively. Unlike some potent inhibitors of NOS, no time dependence was observed. S-Ethylisothiourea was not a detectable substrate for eNOS. S-Ethylisothiourea was also a potent inhibitor of mouse iNOS (Ki value of 5.2 nM), and its binding perturbed the spectrum of iNOS consistent with its altering the environment of the bound heme. The optimum binding of S-ethyl- and S-isopropylisothiourea relative to 70 other analogs suggested that these alkyl substitutions fit into a small hydrophobic pocket. Most isothioureas were 2-6-fold selective for the human iNOS (Ki for iNOS versus Ki for eNOS), with one being 19-fold selective. The cyclized mimics of S-ethylisothiourea, 2-NH2-thiazoline, and 2-NH2-thiazole, were also competitive inhibitors of human NOS. A third structural class of inhibitors, bisisothioureas, were, in general, the most selective in their inhibition of human iNOS. S,S'-(1,3-Phenylenebis(1,2-ethanediyl))bisisothiourea was 190-fold selective (Ki value of 0.047 microM against iNOS versus 9.0 microM against eNOS). These results demonstrate that potent and selective inhibition of human NOS isozymes is achievable.

Potent and selective inhibition of human nitric oxide synthases. Selective inhibition of neuronal nitric oxide synthase by S-methyl-L-thiocitrulline and S-ethyl-L-thiocitrulline.

Potent and selective inhibition of neuronal nitric oxide synthase (nNOS) compared to endothelial NOS (eNOS) and inducible NOS (iNOS) may be useful to treat cerebral ischemia (stroke) and other neurodegenerative diseases. S-Methyl-L-thiocitrulline (Me-TC) and S-ethyl-L-thiocitrulline (Et-TC) inhibited the oxidation of L-arginine and the L-arginine-independent oxidation of NADPH by nNOS from human brain. Me-TC and Et-TC were slow, tight binding inhibitors of nNOS with second-order association rate constants (kon) of 2.6 x 10(5) M-1 s-1 and 1.3 x 10(5) M-1 s-1, respectively. The respective dissociation rate constants (koff) were 3 x 10(-4) s-1 and 0.7 x 10(-4) s-1. Thus, the Kd values calculated from koff/kon were 1.2 and 0.5 nM, respectively. L-Arginine was a competitive inhibitor of Me-TC and Et-TC binding with competition constant (Ks) values of 2.2 and 2.7 microM, respectively. The Km of nNOS for L-arginine was 1.6 microM. The active site concentration of nNOS was estimated by titration with Et-TC. Based on this active site concentration, a kcat of 0.4 s-1 for the oxidation of L-arginine, was calculated. Me-TC and Et-TC were less potent inhibitors of human iNOS (Ki values of 34 and 17 nM, respectively) and human eNOS (Ki values of 11 and 24 nM). Thus, Me-TC and Et-TC were 10- and 50-fold, respectively, more potent inhibitors of nNOS than eNOS. Furthermore, Me-TC was also 17-fold selective for rat nNOS in neuronal tissue compared to rat eNOS in vascular endothelium, suggesting that Me-TC may be selective for nNOS in vivo and therefore, may be therapeutically useful to treat neurodegenerative diseases.

Pharmacokinetics, oral bioavailability, and metabolic disposition in rats of (-)-cis-5-fluoro-1-[2-(hydroxymethyl)-1,3-oxathiolan-5-yl] cytosine, a nucleoside analog active against human immunodeficiency virus and hepatitis B virus.

The pharmacokinetics and metabolism of the potent anti-human immunodeficiency virus and anti-hepatitis B virus compound, (-)-cis-5-fluoro-1-[2-(hydroxymethyl)-1,3-oxathiolan-5-yl] cytosine (FTC), were investigated in male CD rats. Plasma clearance of 10 mg of FTC per kg of body weight was biexponential in rats, with a half-life at alpha phase of 4.7 +/- 1.1 min (mean +/- standard deviation) and a half-life at beta phase of 44 +/- 8.8 min (n = 5). The total body clearance of FTC was 1.8 +/- 0.1 liters/h/kg, and the oral bioavailability was 90% +/- 8%. The volume of distribution at steady state (Vss) was 1.5 +/- 0.1 liters/kg. Increasing the dose to 100 mg/kg slowed clearance to 1.5 +/- 0.2 liters/kg/h, lowered the Vss to 1.2 +/- 0.2 liters/kg, and reduced the oral bioavailability to 65% +/- 15%. FTC in the brains of rats was initially less than 2% of the plasma concentration but increased to 6% by 2 h postdose. Probenecid elevated levels of FTC in plasma as well as in brains but did not alter the brain-to-plasma ratio. The urinary and fecal recoveries of unchanged FTC after a 10-mg/kg intravenous dose were 87% +/- 3% and 5% +/- 1.6%, respectively. After a 10-mg/kg oral dose, respective urinary and fecal recoveries were 70% +/- 2.5% and 25% +/- 1.6%. Two sulfoxides of FTC were observed in the urine, accounting for 0.4% +/- 0.03% and 2.7% +/- 0.2% of the intravenous dose and 0.4% +/- 0.06% and 2.5% +/- 0.3% of the oral dose. Also observed were 5-fluorocytosine, representing 0.4% +/- 0.06% of the intravenous dose and 0.4% +/- 0.07% of the oral dose, and FTC glucuronide, representing 0.7% +/- 0.2% of the oral dose and 0.4% +/- 0.2% of the intravenous dose. Neither deaminated FTC nor 5-fluorouracil was observed in the urine (less than 0.2% of dose). The high oral availability and minimal metabolism of FTC encourage its further preclinical development.

Selective inhibition of constitutive nitric oxide synthase by L-NG-nitroarginine.

L-NG-Nitroarginine (NA) inhibited both the L-arginine oxidation and the L-arginine-independent NADPH oxidation reactions catalyzed by the calcium/calmodulin-dependent constitutive nitric oxide synthase (cNOS) from bovine brain. NA binding did not require calmodulin, calcium, or NADPH. The onset of inhibition was slow with a second-order association rate constant (k(on) of 4.4 x 10(4) M-1 s-1. The dissociation rate constant (k(off) was 6.5 x 10(-4) s-1. The Kd value (k(off)/k(on)) of bovine brain cNOS for NA was 15 nM. L-Arginine was a competitive inhibitor of NA binding with a Ks value of 0.8 microM. The Km for L-arginine in the cNOS reaction was 1.2 microM. The NA binding sites of cNOS were titrated with NA, which enabled a kcat of 0.7 s-1, for the oxidation of L-arginine, to be calculated. Finally, a brain cNOS-(3H)NA complex was isolated. In contrast to the potent and slow onset of NA inhibition of brain cNOS, NA inhibition of inducible mouse macrophage NOS (iNOS) was weaker (Ki = 4.4 microM) and rapidly reversible. Thus, NA was a 300-fold more potent inhibitor of bovine brain cNOS than mouse macrophage iNOS.