Introduction

Top Abstract Introduction Procedures Results Discussion References

NO is an important regulatory substance implicated in a diverse array of functions in mammalian physiology (Bredt and Synder, 1994; Nathan, 1992). It is synthesized by the enzyme NOS, a cytochrome P450-like heme protein that requires tetrahydrobiopterin, FMN and FAD as cofactors to catalyze the NADPH-dependent oxidation of L-arginine to citrulline and NO (Griffith and Stuehr, 1995; Marletta, 1994). NOS has been identified in three isozymic forms, including a constitutive, Ca⁺⁺- and CaM-dependent form (nNOS) found in neurons and GH₃ pituitary cells (Bredt et al., 1991; Wolff and Datto, 1992), a second CaM-dependent form from endothelial cells (eNOS) (Sessa et al., 1992) and a Ca⁺⁺-independent, cytokine inducible isoform (Stuehr et al., 1991).

The overproduction of NO has been implicated in diverse pathological conditions, including postischemic damage in the brain and kidney (Trifiletti, 1992; Yu et al., 1994) as well as in the profound dilatation of septic shock (Kilbourn et al., 1990). Accordingly, it is an important goal to identify inhibitors of NOS, particularly agents that are nontoxic in vivo, cell permeable and isoform selective. Several classes of isoform-selective inhibitors have been identified, including amidines (Garvey et al., 1997; Southan et al., 1995a); S-alkylisothioureas (Garvey et al., 1994; Nakane et al., 1995; Southan et al., 1995b; Szabo et al., 1994); and aminoguanidine (Wolff and Lubeskie, 1995). In contrast to the amidines and S-alkylisothioureas, which appear to be reversible NOS inhibitors, aminoguanidine was identified as an isoform-selective, alternate substrate mechanism-based inactivator. Diaminoguanidine and N^G-amino-L-arginine, compounds that share the structural feature of an hydrazine structure attached to a guanidino carbon, have also been identified as mechanism-based inactivators of NOS (Wolff and Lubeskie, 1996). Although aminoguanidine displays high isoform selectivity, low toxicity (Makita et al., 1992) and cell permeability (Wolff et al., 1997), it exhibits a low potency. We therefore were interested in examining a series of substituted aminoguanidines and aminoisothioureas to explore further the relationship of structure to the mechanism of NOS inhibition and to identify more potent inhibitors that retain isoform selectivity. We report here the results of these studies.

	Experimental Procedures

Top Abstract Introduction Procedures Results Discussion References

Materials. Aminoguanidine and bovine hemoglobin were obtained from Sigma Chemical (St. Louis, MO). SEITU and SIPITU were obtained from Alexis (San Diego, CA). All other reagents were obtained as previously described (Ruetten et al., 1996; Wolff et al., 1997).

Preparation and characterization of 1- and 2-substituted aminoguanidines and N-substituted S-methylisothioureas. N-Substituted S-methylisothioureas or S-methylisosemicarbazides to be used either as test compounds (table 1, 7-9) or as intermediates in the synthesis of substituted aminoguanidines (below) were prepared by methylation of the appropriate thioureas by standard methods (Reid, 1963) previously described (Ruetten et al., 1996; Southan et al., 1995b). Briefly, the appropriate thiourea was refluxed with either methyl iodide or methyl-p-toluenesulfonate in ethanol to give either the iodide or p-toluenesulfonate "tosylate" salts.

Preparation of substituted aminoguanidines. For 2-phenyl-1-aminoguanidine (table 1, compound 6), N-phenyl-S-methylisothiouronium tosylate (1.15 g, 3.5 mmol) was dissolved in ethanol/water (7:1, 8 ml), and aqueous hydrazine solution (0.52 ml, 35%) was added. The mixture was stirred at room temperature overnight. The solvent was removed under vacuum, and the residue was recrystallized from ethanol/ether (m.p. 144-145°C; elemental analysis: theoretical C 52.17%, H 5.59%, N 20.39%; found C 52.22%, H 5.55%, N 17.44%).

Preparation and characterization of NOS isoforms. Ca⁺⁺- and CaM-dependent NOS was prepared from GH₃ cell extracts by adsorption to ADP-agarose and elution with NADPH and was characterized as previously described (Wolff and Datto, 1992). A typical preparation of GH₃ NOS exhibited a specific activity of ~0.6 µmol of citrulline formed/min-mg at saturating concentrations of arginine and cofactors and was stable to storage at 70° for periods up to 4 months. The GH₃ NOS (nNOS) is identical physically, kinetically and immunologically to the bovine brain NOS (Wolff et al., 1992) but routinely contains substoichiometric quantities of bound BH₄ (~0.15 mol/mol) such that it commonly displays a 6- to 10-fold stimulation by the addition of exogenous tetrahydrobiopterin.

Assay of NOS activity by citrulline formation. NOS activity was measured by a modification of the procedure of Bredt and Synder (1990) as previously described (Wolff and Datto, 1992). Standard incubations for the measurement of citrulline formation by either endothelial constitutive NOS (eNOS) or GH₃ constitutive NOS (nNOS) contained 30 mM HEPES, pH 7.4, 1 mM dithiothreitol, 120 nM [³H]arginine (a subsaturating concentration), 1 mM EGTA, 0.85 mM Ca⁺⁺, 6 µM CaM, 100 µM NADPH and 100 µM BH₄. Standard incubations for the measurement of citrulline formation by the interferon--inducible macrophage NOS contained 30 mM HEPES, pH 7.4, 1 mM dithiothreitol, 120 nM [³H]arginine, 1 mM EGTA, 100 µM NADPH and 300 µM BH₄. Incubations were conducted at 30° for 30 min in duplicate, and the mean values were calculated. Variability of values about the mean routinely averaged ±3% of the mean. Routinely, assays were conducted at dilutions of enzyme that provided 5% to 10% of total substrate consumption. At these conditions of measurement, product formation was linear over time.

Assay of the substrate-independent NADPH-oxidase activity of nNOS. Standard reaction mixtures of 1-ml volume contained 50 mM HEPES, pH 7.4, 100 µM NADPH, 6 µM CaM, 1 mM EGTA and (when present) 0.85 mM Ca⁺⁺ (8 µM free Ca⁺⁺). Reactions were conducted in quartz cuvettes, and NADPH consumption was measured from the change in light absorbance at 340 nm compared with an identical reference cuvette without added enzyme.

Preparation of oxyhemoglobin. Bovine hemoglobin (33 mg) was dissolved in 2 ml of 20 mM HEPES, pH 7.4, 130 mM NaCl and was treated with sodium dithionite sufficient to reduce contaminating methemoglobin. The sample was applied to a 2.5 × 60-cm column of Sephadex G-15 equilibrated with HEPES-buffered saline and subjected to gel filtration, and the excluded hemoglobin fractions were collected. The identity of oxyhemoglobin was verified by checking the position of the absorbance maximum (415 nm). The concentration of the solution was calibrated using a molar extinction coefficient _415 nm = 131 mM¹cm¹ as reported by Noack et al. (1992).

Growth of and measurement of NO formation by cytokine-induced murine RAW 264.7 cells. Murine RAW 264.7 cells were grown to confluency in six-welled (9-cm²) polystyrene dishes in 3 ml of Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. NOS was induced by treating the cells with 75 units of murine interferon-/25 µg of lipopolysaccharide for 16 hr. Growth medium was removed and replaced with 4 ml of modified Ham's F-10 containing either 100 µM (physiological conditions) or 1 mM arginine (maximal rate) and 5 µM oxyhemoglobin. The release of NO from the cells was assessed by measuring the formation of methemoglobin as the absorbance difference at 401 nm (absorbance maximum) and 411 nm (isosbestic point) over time as described by Noack et al. (1992). The nanomoles of NO formed was calculated using an extinction coefficient difference for the methemoglobin and oxyhemoglobin forms of 38 mM¹cm¹. When NO formation was measured in the presence of drug or by drug-treated cells, rates were adjusted for interference generated from the slow autoxidation of oxyhemoglobin. These values were linear over time and represented a rate <4% of the rate observed in cytokine-treated RAW cells measured in the absence of inhibitor.

Assay of the NOS-catalyzed formation of NO by measurement of conversion of oxyhemoglobin to methemoglobin. Standard reaction mixtures were constructed in 1-ml disposable polystyrene cuvettes containing 50 mM HEPES, pH 7.4, 100 µM NADPH, 6 µM CaM, 1 mM EGTA, (when present) 0.85 mM Ca⁺⁺, 0.5 µM BH₄, 5 µM oxyhemoglobin and, unless otherwise indicated, 100 µM arginine. Reaction mixtures were added to both a sample and a reference cuvette, and the instrument zeroed at 401 nm. Reactions were initiated by the addition of NOS enzyme source to the sample cuvette, and time-dependent formation of methemoglobin was measured at 401 nm. NO formation was calculated using the known extinction coefficient for methemoglobin.

Miscellaneous procedures. Protein was determined according to the method of Bradford (1976) with bovine serum albumin as the reference standard.

	Results

Top Abstract Introduction Procedures Results Discussion References

Inactivation of NOS activity of nNOS by substituted aminoguanidines and aminoisothioureas. In incubations measuring NO formation by nNOS (fig. 1) it was observed that the concurrent addition of AMITU and Ca⁺⁺ resulted in the time-dependent loss of NO synthetic activity. This loss of activity was not due to the time-dependent accumulation in solution of an NOS inhibitor because the addition of a second aliquot of nNOS was initially as fully active as the first aliquot and underwent an identical time-dependent loss of activity as had been observed when the first aliquot of nNOS had been concurrently exposed to AMITU and Ca⁺⁺.

View larger version (10K): [in this window] [in a new window]   Fig. 1.   Effect of 1-amino-S-methylisothiourea on NO formation by GH3 pituitary nNOS. A, Standard incubations were constructed in 1-ml polystyrene cuvettes containing 50 mM HEPES, pH 7.4, 100 µM NADPH, 6 µM CaM, 1 mM EGTA, 100 µM arginine and 5 µM oxyhemoglobin. The reaction was initiated at zero time with 8 µg of affinity-purified GH3 pituitary nNOS. At 2 min after initiation, the incubation was adjusted to contain 0.85 mM Ca++, either alone () or in combination with 300 µM 1-amino-S-methylisothiourea (). At 9 min after initiation, a second 8-µg portion of nNOS was added to the AMITU-containing incubation. B, A standard incubation was constructed as described for A containing 8 µg of nNOS added at zero time and adjusted to contain 300 µM AMITU at 1 min. At 9 min after initiation, the incubation was adjusted to contain 0.85 mM Ca++. Throughout, NO-generated formation of methemoglobin was measured as the increase in light absorbance at 401 nm.

Inactivation of the substrate-independent NADPH oxidase activity of nNOS by substituted aminoguanidines and aminoisothioureas. The neuronal constitutive isoform of NOS has been shown in the absence of arginine substrate to catalyze the reduction of oxygen to superoxide anion in a Ca⁺⁺-dependent manner using reducing equivalents derived from NADPH (Klatt et al., 1993; Mayer et al., 1991). This substrate-independent, NADPH-oxidase activity of nNOS has been shown to undergo a Ca⁺⁺-dependent inactivation in the presence of aminoguanidine (Wolff and Lubeskie, 1995). Accordingly, we were interested in whether 1-amino-substituted isothioureas and aminoguanidines also shared this property. Ca⁺⁺-dependent NADPH consumption by GH₃ nNOS was measured in the absence of arginine substrate in a control incubation without drug or containing AMITU at concentrations ranging from 3 to 30 µM (fig. 2A). In the absence of AMITU, nNOS displayed an NADPH-oxidase activity that declined slowly over time, autoinactivating at a first-order rate (0.05 min¹) that required 13 min to inactivate 50% of activity. This behavior has been previously described (Wolff and Lubeskie, 1995) and appears to derive from a superoxide-mediated inactivation of an active site component essential to catalytic activity. In the presence of AMITU, the rate of inactivation of NADPH-oxidase activity was enhanced in a concentration-dependent manner. The half-time of inactivation attributable to AMITU was calculated at each concentration of AMITU by measuring the observed rate of inactivation and subtracting the drug-independent autoinactivation rate. The AMITU-dependent half-times of inactivation when plotted in Kitz-Wilson format (fig. 2B) vs. the reciprocal of the AMITU concentration provided a linear plot, consistent with a AMITU-dependent first-order rate of inactivation (Kitz and Wilson, 1962; Silverman, 1988). The AMITU-dependent inactivation exhibited a maximal rate of 0.25 min¹, with a half-maximal rate of inactivation occurring at a concentration of 11 µM AMITU as calculated from the abscissal intercept (1/K_I). Thus, in the presence of saturating AMITU, the inactivation of nNOS is increased 6-fold from its autoinactivation rate (0.30 vs. 0.05 min¹). In an experiment not shown, the effect of diverse substituted aminoguanidines on the NADPH-oxidase activity of GH₃ pituitary nNOS was measured without or with a 3 mM concentration of agent. A control half-time of autoinactivation of 840 sec was observed, whereas aminoguanidine (40 sec), 1-methylaminoguanidine (450 sec), 2-methylaminoguanidine (40 sec), 2-hydroxyaminoguanidine (180 sec), 2-phenylaminoguanidine (300 sec) and 2-ethylaminoguanidine (160 sec) promoted inactivation with half-times of inactivation as indicated. Thus, each of the substituted aminoguanidines (tables 1 and 2) promoted inactivation of nNOS as documented for 1-amino-S-methylisothiourea (fig. 2) and as observed formerly for aminoguanidine (Wolff and Lubeskie, 1995).

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Effect of 1-amino-S-methylisothiourea on the NADPH oxidase activity of GH3 pituitary nNOS. A, Standard incubations were constructed in a quartz cuvette containing 50 mM HEPES, pH 7.4, 6 µM CaM, Ca++ (8 µM free) and 100 µM NADPH without () or with 3 (), 6 (), 10 () or 30 () µM AMITU. Reactions were initiated by the addition of 10 µg of affinity-purified GH3 nNOS, and NADPH consumption was measured as the decrease of light absorbance at 340 nm. B, Half-time of inactivation of NADPH-oxidase activity adjusted for the contribution of autoinactivation (no added drug) was determined from the data in A at each drug concentration. The half-times of inactivation are plotted vs. the reciprocal of the AMITU concentration in Kitz-Wilson format. A KI value of 11 µM AMITU and a kinact max value of 0.25 min1 was determined.

Determination of the NOS isoform selectivity of substituted aminoguanidines and aminoisothioureas. Clearly, our observations indicated that the substituted aminoguanidines and AMITU were potentially interesting inactivators of NOS. We therefore undertook to examine their isoform selectivity by measuring their IC₅₀ values for inhibition of citrulline formation by each of the NOS isoforms using affinity-purified enzyme (table 2). Of the substituted aminoguanidines tested, only 2-ethylaminoguanidine displayed a higher potency for inhibition of the iNOS isoform (1 vs. 5 µM IC₅₀) than the parent compound aminoguanidine. The 2-ethylaminoguanidine also retained the high isoform selectivity of the parent compound, inhibiting iNOS activity at a concentration 23- and 14-fold lower than the eNOS and nNOS isoforms, respectively. Substitution of the hydrogen at the 1 position of aminoguanidine by a methyl group or of hydrogen at the 2 position of aminoguanidine by a methyl, hydroxyl or phenyl substituent reduced both iNOS inhibitory potency and iNOS selectivity. In contrast with the aminoguanidines, AMITU exhibited selectivity for the nNOS isoform with an IC₅₀ value of 3 µM, a value 8- and 34-fold lower than that observed for the iNOS and eNOS isoforms, respectively. Replacement of a hydrogen atom of AMITU at either the 1 or 3 position with a methyl group resulted in both a dramatic loss of inhibitory potency and isoform selectivity. The compounds S-ethylisothiourea and S-isopropylisothiourea, which had been previously described (Garvey et al., 1994; Nakane et al., 1995) were far more potent than AMITU but exhibited poorer isoform selectivity, particularly in discriminating the nNOS and iNOS isoforms.

Determination of the kinetic constants for inactivation of iNOS and nNOS by substituted aminoguanidines and aminoisothioureas. Because our observations (fig. 1) had indicated that AMITU and substituted aminoguanidines were mechanism-based inactivators producing inhibitions of NOS that progressed with time of incubation, measurements of IC₅₀ values conducted in incubations for a fixed time, while providing an estimate of efficacy as NOS inhibitors, did not formally evaluate precisely their efficiency as inactivators. We therefore measured the time and concentration dependence of inhibition of NO formation by the affinity-purified iNOS and nNOS isoforms using a continuous spectrophotometric assay measuring NO-mediated conversion of oxyhemoglobin to methemoglobin. Incubations were conducted at 100 µM arginine, a saturating concentration of substrate comparable to that found in the normal human extracellular fluid (Moncada and Higgs, 1995; Simmons et al., 1996). The NO formation rate by iNOS was measured in incubations without drug or at concentrations of 2-ethylaminoguanidine ranging from 75 to 500 µM (fig. 3). A time- and concentration-dependent loss of NO synthetic rate was observed. The half-time of inactivation was determined at each drug concentration as the time required for the initial rate (first 20 sec) to decline to a rate one-half this initial rate. The half-times of inactivation were plotted in Kitz-Wilson format vs. the reciprocal of the inactivating concentration of 2-ethylaminoguanidine. A maximal inactivation rate of 0.48 min¹ was determined from the ordinal intercept (k_inact max = 0.693/t_1/2), whereas the concentration of 2-ethylaminoguanidine providing the half-maximal inactivation rate (120 µM) was determined from the abscissal intercept (1/K_I). From these data, the apparent (determined at 100 µM arginine) second-order rate constant for inactivation (k_inact max/K_I) could be calculated. The second-order rate constant for inactivation is a formal measure of the efficiency of a mechanism-based inactivator (Bandyapadhyah et al., 1993; Fitzpatrick and Villafranca, 1986; Rando, 1984; Silverman, 1988). Using affinity-purified nNOS and iNOS isoforms and a paradigm identical to that delineated in the legend to figure 3 (except a different range of drug concentrations), the kinetic constants for inactivation of iNOS and nNOS by each of the substituted aminoguanidines and AMITU were determined and are presented in table 3. Of the substituted aminoguanidines examined, 2-ethylaminoguanidine was the most efficient inactivator of the iNOS isoform and the only substituted aminoguanidine more efficient than the parent compound aminoguanidine in inactivating iNOS, a conclusion identical that suggested by the determinations of IC₅₀ values in fixed-time incubations (table 2). Among the aminoguanidines, 2-methylaminoguanidine exhibited the fastest rate of maximal inactivation vs. iNOS but was a less efficient inactivator in that it required high concentrations (K_I = 3.5 mM) to exert its effects. 2-Methylaminoguanidine behaved similarly vs. the nNOS isoform. All of the substituted aminoguanidines inactivated the iNOS isoform more efficiently than the nNOS isoform, with the parent compound aminoguanidine being the most discriminatory based on the ratio of their second-order rate constants for inactivation. In contrast, AMITU was the most efficient inactivator of the nNOS isoform and was the only agent displaying isoform selectivity for the nNOS isoform.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Effect of 2-ethylaminoguanidine on NO formation by affinity-purified murine macrophage iNOS. A, Methemoglobin formed from reaction of oxyhemoglobin with NO was measured in standard incubations containing 100 µM arginine as described in the text without () or with 75 (), 150 (), 300 (), or 500 () µM 2-ethylaminoguanidine. Reactions were initiated in 1-ml disposable polystyrene cuvettes with 2.4 µg of affinity-purified murine macrophage iNOS. B, Half-times of inactivation of NOS activity are plotted in Kitz-Wilson format vs. the reciprocal of the inactivating concentration of 2-ethylaminoguanidine. A KI value of 120 µM 2-ethylaminoguanidine and a kinact max value of 0.48 min1 were determined.

Inactivation by and recovery from substituted aminoguanidines and isothioureas in cytokine-induced RAW 264.7 cells. Among the aminoguanidines examined, 2-ethylaminoguanidine was the most efficient iNOS inactivator. Accordingly, we were interested in examining the effectiveness of 2-ethylaminoguanidine in an intact cellular system containing iNOS. NO formation by cytokine-induced RAW 264.7 cells was measured in confluent six-welled plates as the ability of NO released from cells to convert oxyhemoglobin to methemoglobin measured spectrophotometrically. NO formation was measured in Ham's F-10 medium containing 100 µM arginine (its normal extracellular concentration) without or with 2-ethylaminoguanidine at concentrations ranging from 25 to 200 µM (fig. 4). During the first 2 to 4 min, the NO formation rates were essentially unaltered by 2-ethylaminoguanidine compared with the control without drug. However, at times beyond 4 min, a time- and concentration-dependent loss of NO synthetic capability was observed. The half-times of inactivation were determined at each 2-ethylaminoguanidine concentration and were plotted in Kitz-Wilson format (fig. 4B). A linear Kitz-Wilson plot was obtained with a k_inact max value of 0.09 min¹ and a K_I value of 55 µM 2-ethylaminoguanidine.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Effect of 2-ethylaminoguanidine on time-dependent formation of NO by cytokine-induced RAW 264.7 cells. A, Confluent monolayers of RAW 264.7 cells in six-well plates treated with interferon-/LPS for 18 hr were incubated in modified Ham's F-10 medium containing 100 µM arginine and 5 µM oxyhemoglobin without () or with 25 (), 50 (), 100 (), or 200 () µM 2-ethylaminoguanidine. Cumulative NO formation was assessed by spectrophotometrically measuring methemoglobin formation. B, Half-times of inactivation of NO-forming capability are plotted in Kitz-Wilson format vs. the reciprocal of the 2-ethylaminoguanidine concentration.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Recovery of NO synthetic capability and iNOS activity in cytokine-induced RAW 264.7 cells treated with 2-ethylaminoguanidine. RAW 264.7 cells grown to confluency in six-welled plates were treated with interferon-/LPS for 18 hr, and the medium was replaced with modified Ham's F-10 containing 1 mM arginine and 5 µM oxyhemoglobin. NO formation rate by the cells was measured as 3.7 nmol of NO/min-mg of protein and varied <4% among the six wells. Well 1 was maintained in Ham's F-10 medium containing 10 µM cycloheximide throughout as the untreated control. Wells 2 to 6 were treated with modified Ham's F-10 containing 300 µM 2-ethylaminoguanidine for 1 hr. Immediately after treatment, the medium in the wells was replaced with modified Ham's F-10 medium containing 1 mM arginine, 10 µM cycloheximide and 5 µM oxyhemoglobin, and NO formation (methemoglobin produced over a 10-min incubation) was measured immediately after drug treatment (zero time) or after 0.5, 1, 2 or 3 hr of recovery in Ham's F-10. Immediately after measurement of NO formation, the cells were rinsed and lysed in a buffer containing 0.5% CHAPSO, protease inhibitor cocktail and 30 µM arginine, and portions were used to assess iNOS activity by measuring citrulline formation in standard incubations as described in the text. Values for NO formation rate () and iNOS activity () are expressed as the percentage of the control sample of well 1 never exposed to drug. The control iNOS activity was 3.2 nmol of citrulline formed/min-mg, whereas control NO formation rate was 3.5 nmol/min-mg.

View larger version (19K): [in this window] [in a new window]   Fig. 6.   Effect of S-isopropylisothiourea on time-dependent formation of NO by cytokine-induced RAW 264.7 cells. A, The S-isopropylisothiourea concentration dependence of inhibition of NO formation. Cytokine-induced RAW 264.7 cells grown to confluency in six-welled plates were incubated in modified Ham's F-10 medium containing 100 µM arginine and 5 µM oxyhemoglobin without () or with 0.1 (), 0.3 (), 1 () or 10 () µM SIPITU. NO formation was measured at 2-min intervals by the spectrophotometric determination of methemoglobin formation. B, Standard incubations were constructed in separate wells of a six-welled plate containing confluent cytokine-induced RAW 264.7 cells in modified Ham's F-10 containing 100 µM arginine and 5 µM oxyhemoglobin. At 8 min after initiation (arrow) each incubation was adjusted to contain 1 µM SIPITU, and measurement of cumulative NO formation continued. At 22 min after initiation, one well was adjusted to 5 mM arginine (), whereas medium was removed from the second well and was replaced with fresh, drug-free assay medium (). Cumulative NO formation was measured throughout by the spectrophotometric determination of methemoglobin formation.

	Discussion

Top Abstract Introduction Procedures Results Discussion References

Aminoguanidines derivatized at the 1 and 2 positions have been prepared and compared with the parent compound aminoguanidine with respect to NOS inhibitory properties. Based on measurements of IC₅₀ values conducted at subsaturating arginine concentrations in incubations for fixed times, substitution of hydrogen at the 1 position by methyl or at the 2 position by methyl, hydroxyl or phenyl substituents reduced the potency and isoform selectivity of aminoguanidine (table 2). In contrast, substitution of hydrogen at the 2 position by an ethyl group increased potency and largely retained the isoform selectivity of the parent compound. Each of the derivatized aminoguanidines produced a time- and concentration-dependent inactivation of NOS (table 3) that followed first-order kinetics as indicated by linear Kitz-Wilson plots. For the nNOS isoform, all inactivations occurred only when enzyme was exposed to drug under conditions that support catalytic activity (presence of Ca⁺⁺) constistent with the proposal that these agents, like the parent compound aminoguanidine, serve as alternate substrate mechanism-based inactivators. The efficiency of mechanism-based inactivators are measured based on their second-order rate constants for inactivation, which normalize the maximal rate of inactivation by dividing by the concentration of drug necessary to elicit the half-maximal inactivation rate, and have units of reciprocal concentration reciprocal time, (e.g., mM¹min¹). The kinetic constants presented in table 3 are based on kinetic measurements conducted at 100 µM arginine, a concentration of substrate that saturates both the iNOS and nNOS isoforms and reflects the concentration of arginine normally found in the extracellular fluid. Thus, the conditions of measurement resemble the conditions that prevail in vivo. Because the aminoguanidines are alternate substrates, they compete with arginine for the catalytic site; thus, the K_I values determined are apparent K_I values and are elevated to a degree predictable on the basis of the Cheng-Prusoff relationship (Craig, 1993; Cheng and Prusoff, 1973), such that apparent K_I = K_I(1 + S/K_m). Similar considerations apply to AMITU. AMITU inactivated the substrate-independent (no arginine present) NADPH-oxidase activity of nNOS (fig. 2) with a K_I value of 11 µM but inactivated the NO synthetic capability of nNOS (measured in the presence of 100 µM arginine) with an apparent K_I value of 300 µM (table 3). Because the K_m value of GH₃ pituitary nNOS for arginine is 4 µM (Wolff and Datto, 1992), the 1 + S/K_m term of the Cheng-Prusoff indicates that the apparent K_I value should be elevated 26-fold by the presence of 100 µM arginine. This corresponds closely to the observed degree of elevation (300 vs. 11 µM).

Based on the second-order rate constants (table 3), 2-ethylaminoguanidine was the only derivatized aminoguanidine homolog examined that displayed increased NOS inactivating efficiency vs. the iNOS isoform (4.0 vs. 2.14 mM¹min¹) compared with the parent compound aminoguanidine and retained most of the isoform selectivity of the parent compound based on the ratios of the second-order rate constants for the iNOS compared with the nNOS isoform. On the basis of the second-order rate constants for inactivation (table 3), aminoguanidine is 19-fold and 2-ethylaminoguanidine is 9-fold more efficient in inactivating iNOS compared with nNOS.

An interesting observation was the kinetic behavior of 2-methylaminoguanidine, which exhibited the highest maximal inactivation rate of any of the compounds examined, with rates of 2.48 min¹ (0.04 sec¹) and 2.57 min¹ (0.043 sec¹) being observed for the iNOS and the nNOS isoforms, respectively. This very rapid rate of maximal inactivation presumably reflects a high inactivating efficiency for the suicide intermediate generated by 2-methylaminoguanidine; that is, this intermediate has the highest ratio of inactivations per turnover. However, very high concentrations of 2-methylaminoguanidine are necessary to successfully displace arginine from the substrate site, as reflected in its very high K_I value for the both the iNOS and nNOS isoforms. This is confirmed by the very rapid rate of inactivation of the NADPH oxidase activity of nNOS (half-time of 40 sec at 3 mM drug) by 2-methylaminoguanidine because this inactivation is measured in the absence of protective arginine.

2-Ethylaminoguanidine inactivated the NO synthetic capability of cytokine-induced RAW 264.7 cells in a time- and concentration-dependent manner that provided a linear Kitz-Wilson plot with a maximal inactivation rate of 0.09 min¹ and an apparent K_I value of 55 µM. This maximal inactivation rate was substantially slower than that observed with isolated iNOS (0.48 min¹). The diminished maximal inactivation rate may be attributable to a maximal inactivation rate limited by the rate of cellular drug uptake. Indeed, even at the highest concentration of 2-ethylaminoguanidine examined (fig. 4) no decrease of NO formation rate was observed until a ~4-min lag time had evolved, whereas with the isolated enzyme a diminished rate of NO formation was detectable within the first 20 sec. It seems reasonable to assert that this is due to the "immediate" access of drug to isolated enzyme, whereas in the intact cell an uptake lag is encountered. In previous experiments from our laboratory (Wolff et al., 1997), the parent compound aminoguanidine was observed to inactivate the NO-forming capability of iNOS in cytokine-induced RAW 264.7 cells with an apparent K_I value of 640 µM and a k_inact max value of 0.22 min¹, as measured under identical conditions. Using the calculated second-rate constants for inactivation in the intact cell, 2-ethylaminoguanidine is 4.8-fold more efficient in inactivating iNOS than the parent aminoguanidine. The inactivation produced by 2-ethylaminoguanidine was in part reversible when cells were transferred to drug-free medium. Both NO synthetic capability and iNOS activity measured in lysates recovered over a 3-hr period to an identical degree and with an identical time course. This recovery was identical regardless of the presence or absence (not shown) of 10 µM cycloheximide, a concentration sufficient to suppress protein synthesis by >99% in the RAW 264.7 cell system (Wolff et al., 1997). A similar recovery was previously observed (Wolff et al., 1997) for cells exposed to either of the mechanism-based inactivators aminoguanidine or N^G-methyl-L-arginine but not diphenylene iodonium. These previous studies provided evidence that the recovery was due to a monomeric population of iNOS not capable of converting the alternate substrate to the suicide intermediate. The monomer population of "undamaged" enzyme could serve as a pool of precursor from which catalytically competent dimer could be assembled during the recovery period. This assembly could occur despite the absence of protein synthesis. Given the close structural homology of aminoguanidine and 2-ethylaminoguanidine, it is reasonable to presume that similar considerations apply to its mode of recovery.

AMITU exhibited behavior quite different in character to that observed for the substituted aminoguanidines. With isolated enzyme, AMITU was isoform selective (table 3) for the nNOS isoform. However, AMITU produced no detectable effects on NO formation in cytokine-induced RAW 264.7 cells. Based on the considerations presented above (Cheng-Prusoff relationship), the failure of AMITU treatment to produce any detectable loss of NO-forming activity in the intact cell or iNOS activity in lysates from treated cells could not be attributed to a protection of enzyme by arginine substrate. This was confirmed when cells were treated with AMITU for 3 hr in the absence of protective arginine and no effect on either NO formation or iNOS activity was detectable. These data support the hypothesis that AMITU is cell impermeable. This finding was somewhat unexpected because previous researchers had reported the effectiveness of S-alkylisothioureas in intact cell systems. Indeed, we observed that both SIPITU (fig. 6) and SEITU (not shown) produced a decreased but linear rate of NO formation within minutes of drug exposure and that the effect of the drug could be largely reversed by increased extracellular arginine and completely reversed within minutes by washout. Thus, the derivatization of S-alkylisothioureas with a 1-amino group has profound effects on drug behavior, converting the compounds from ones acting as linear, reversible inhibitors that are readily cell permeable into a compound that acts as an alternate substrate, mechanism-based inactivator excluded from cell entry.