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


Address correspondence and reprint requests to Professor Darío Acuña-Castroviejo, Departamento de Fisiología, Facultad de Medicina, Avenida de Madrid 11, E-18012 Granada, Spain.


We assessed the effects of melatonin, N1-acetyl-N2-formyl-5-methoxykynuramine (AFMK) and N1-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 IC50 value for AMK (70 µm) 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 Ca2+-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.

Abbreviations used



AlexaFluorTM 594-conjugated CaM






bovine serum albumin






d , l-dithiothreitol


5,6,7,8-tetrahydro-l-biopterin dihydrochloride


flavin adenine dinucleotide


5-hydroxyindole acetic acid






3-(N-morpholino)propanesulfonic acid




nicotinamide adenine dinucleotide




neuronal nitric oxide synthase


nitric oxide




phenylmethylsulfonyl fluoride


reactive nitrogen species


reactive oxygen species


sodium dodecyl sulfate–polyacrylamide gel electrophoresis

Melatonin (aMT) exerts a variety of physiological functions in order to help cells adapt to their environment. aMT is produced by several organs and tissues in addition to the pineal gland (Stefulj et al. 2001; Hardeland et al. 2005), and production is found to decrease with age (Lardone et al. 2006). aMT is both a lipo- and hydrophilic molecule, and it readily crosses all physiological barriers including the blood–brain barrier and plasma membrane (Reiter 1991a; Costa et al. 1995), reaching any subcellular compartments including the nucleus and mitochondria (Acuña-Castroviejo et al. 2001). As a typical hormone, aMT exerts some of its functions through protein Gi-linked specific membrane receptors (Dubocovich and Markowska 2005), whereas other effects of the hormone are mediated by nuclear ROR/RZR, retinoic acid receptors (Acuña-Castroviejo et al. 1995; Wiesenberg et al. 1995). Although not classified as aMT receptors, aMT also binds with high affinity to cytosolic proteins including calmodulin (CaM) and calreticulin (Benitez-King et al. 1993; Macías et al. 2003). After binding to Ca2+-calmodulin (CaCaM) aMT blocks CaCaM-dependent multiple cytosolic pathways, including the activation of neuronal nitric oxide synthase (nNOS) (León et al. 2000). aMT also exerts direct, non-receptor-mediated actions, such as the scavenging of a number of free radicals including both reactive oxygen and nitrogen species (ROS and RNS, respectively) (Tan et al. 2002, 2005b; Reiter et al. 2003).

The apparent multiple actions of aMT raised the question as to whether they are caused by either aMT per se or by some physiologically active aMT metabolites. aMT is metabolized to two main products, N1-acetyl-N2-formyl-5-methoxykynuramine (AFMK) and N1-acetyl-5-methoxykynuramine (AMK) (Hirata et al. 1974; Tan et al. 2000b). Although the main substrate for the indoleamine-2,3-dioxygenase (IDO), a cytosolic enzyme (Boasso et al. 2005), is tryptophan, this enzyme also transforms aMT to AFMK (Hayaishi 1976). An additional pathway for the enzymatic conversion of aMT to AFMK involves myeloperoxidase (Silva et al. 2000). But aMT may also be metabolized to AFMK after scavenging free radicals (Hayaishi 1976; Tan et al. 2001, 2002, 2005a). Subsequently, AFMK can be metabolized to AMK through arylamine formamidase and catalase pathways (Tan et al. 2000b). Electrophysiological and biochemical experiments have shown that the neuroprotective properties of aMT involve the inhibition of nNOS, thereby reducing the production of nitric oxide (NO) (León et al. 2000; Escames et al. 2004). Some of these experiments suggested that an aMT metabolite, rather than aMT itself, may be the responsible for nNOS inhibition (Escames et al. 2001, 2004; León et al. 1998). In an attempt to identify more efficient antagonists of nNOS, a series of compounds structurally related to AMK and AMK itself were synthesized (Camacho et al. 2002, 2004). From these studies, the preliminary data shows that AMK displays higher nNOS inhibitory activity than does aMT.

Currently, no information is available concerning the effect of AMK on nNOS activity. To address this question, herein we assessed the effects of both AFMK and AMK on nNOS activity both in vitro and in vivo, and compared them with those of aMT. To help characterize these effects, the enzymatic transformation of aMT to its metabolites was blocked with the use of specific IDO inhibitors, norharmane and 1-methyl-tryptophan. We used rat striatal tissue as the source of endogenous nNOS because the reported electrophysiological and biochemical experiments (above) with aMT and aMT analogs were performed using this tissue. Experiments with CaM were carried out to further assess the possible interaction of aMT metabolites with CaM. Finally, recombinant nNOS was used to further assess the mechanism of nNOS inhibition by AMK.

Materials and methods


l-Arginine, l-citrulline, l-tryptophan, ascorbate, methylene blue, catalase, HEPES, 3-(N-morpholino)propanesulfonic acid (MOPS), d,l-dithiothreitol (DTT), leupeptin, aprotinin, pepstatin, phenylmethylsulfonyl fluoride (PMSF), hypoxantine-9-β-d-ribofuranosid (inosine), EGTA, bovine serum albumin (BSA), Dowex-50 W (50 × 8–200 mesh) resin, flavin adenine dinucleotide (FAD), NADPH, 5,6,7,8-tetrahydro-l-biopterin dihydrochloride (H4-biopterin), bovine brain CaM (98% sodium dodecyl sulfate–polyacrylamide gel electrophoresis, SDS–PAGE), trichloroacetic acid, aMT, 5-methoxytryptamine (MT), 5-hydroxytryptamine (OHT), 5-hydroxytryptophan (HT), 5-hydroxyindole acetic acid (HIAA), N-acetyl-serotonin (NAS), norharmane and 1-methyl tryptophan were obtained from Sigma-Aldrich (Madrid, Spain). [3H]l-arginine (58 Ci/mmol) was obtained from Amersham (Amersham Pharmacia Biotech GmbH, Barcelona, Spain). Tris HCl and calcium chloride were obtained from Merck & Co., Inc., Whitehouse Station, NJ, USA. A fluorescent derivative of CaM, AlexaFluorTM 594-conjugated CaM (Alexa-CaM), was purchased from Molecular Probes (Invitrogen, Barcelona, Spain). Purified rat brain nNOS (98% SDS–PAGE) was obtained from Alexis Biochemicals (San Diego, CA, USA). The following two kynuramines, the main aMT metabolites, were used in the study: AFMK and AMK. The synthesis of these compounds is published elsewhere (Tan et al. 2000a; Entrena et al. 2005).

Experimental design

Male Wistar rats (200–250 g) housed under a 12-h light/dark cycle with free access to food and water were used. All experiments were performed according to the Spanish Government Guide for Animal Care and the European Community Guide for Animal Care. Animals (six per group) were killed by cervical dislocation and the striata of each animal were quickly removed and immediately used to measure either nNOS or IDO activities. Upon removal, tissues were cooled in an ice-cold buffer (25 mm Tris, 0.5 mm DTT, 10 µg/mL leupeptin, 10 µg/mL pepstatin, 10 µg/mL aprotinin and 1 mm PMSF, pH 7.6). Two striata were placed in 1.25 mL of the same buffer and were sonicated six times for 10 s in the cold (4°C). The crude homogenates were centrifuged at 1000 g for 5 min at 4°C, and an aliquot of the supernatant was frozen at −20°C for total protein determination (Lowry et al. 1951). When required, animals were injected i.p. with 100 mg/kg norharmane and 4 h later with 10–20 mg/kg of aMT, AFMK, AMK or vehicle. One hour later, i.e. 5 h after the administration of norharmane, the animals were killed and their striata processed as described.

IDO activity

IDO activity was determined as described elsewhere (Takikawa et al. 1986). Briefly, the reaction was started by the addition of 0.2 mL of fresh supernatant to 0.8 mL of the reaction mixture consisting of 0.4 mm l-tryptophan, 20 mm ascorbate, 10 µm methylene blue and 100 µg of catalase in 50 mm phosphate buffer, pH 6.5. After 3 h of incubation at 37°C, the reaction was blocked by the addition of 0.2 mL of 30% trichloroacetic acid. The tubes were further incubated for 30 min at 50°C, to hydrolyse the N-formylkynurenine produced by the reaction to l-kynurenine. After centrifugation at 13 000 g for 10 min at 4°C and ultrafiltration (cut-off, 10 000 Mr), the quantity of l-kynurenine produced in the reaction mixture was analysed by reversed-phase HPLC. Supernatant (100 μL) was injected onto a 5-mm C18 reverse phase HPLC column (Waters Chromatography, Barcelona, Spain) at a flow rate of 1.0 mL/min with a mobile phase containing 0.1 m ammonium acetate/acetic acid buffer and 5% acetonitrile (pH 4.65). Kynurenine was detected at 360 nm. One unit of activity was defined as 1 nmol kynurenine/h/mg of protein at 37°C.

NNOS activity determination in rat striatum

NOS activity was measured following the method described by Bredt and Snyder (1989), by monitoring the conversion of [3H]l-arginine to [3H]l-citrulline. The final incubation volume was 100 µL and consisted of 10 µL of crude homogenate added to buffer to give a final concentration of 25 mm Tris, 1 mm DTT, 30 µm H4-biopterin, 10 µm FAD, 0.5 mm inosine, 1 mg/mL BSA, 1 mm CaCl2, 10 µm l-arginine and 50 nm[3H]l-arginine, pH 7.6. When required for kinetic studies, concentrations of l-arginine ranging from 0 to 10 µm were added to the incubation medium. The reaction was started by the addition of 10 µL of NADPH (0.75 mm final) and continued for 30 min at 37°C. Control incubations were performed by the omission of NADPH. The reaction was stopped by the addition of 400 µL of cold 0.1 m HEPES containing 10 mm EGTA and 1 mm l-citrulline, pH 5.5. The mixture was decanted into a 2-mL column packed with Dowex-50 W ion change resin (Na+ form) and eluted with 1.2 mL water. [3H]l-Citrulline was quantified in the eluate by liquid scintillation counting. The retention of [3H]l-arginine was greater than 98%. Enzymatic activity was determined by subtracting the control value, which was usually less than 1% of the radioactivity added. The activity of nNOS was expressed as pmoL of [3H]l-citrulline/min/mg protein.

Purified rat brain nNOS activity determination

The activity of nNOS present in the commercial source was also measured using the method described by Bredt and Snyder (1989) and following the protocol reported by León et al. (2000). After resuspension of an nNOS stock solution in 50 mm HEPES buffer, pH 7.4, aliquots (0.0125 U, equivalent to 0.35 mg of protein) were incubated for 15 min at 37°C in the presence of 15.5 mm CaCl2, 30 µm H4-biopterin, 10 µm FAD, 1 mg/mL BSA, 0.5 mm inosine, 10 µm l-arginine, 10 µg/mL CaM, 0.75 mm NADPH and 50 nm[3H]l-arginine, in a total volume of 100 µL. When required, increasing concentrations of l-arginine (0–10 µm), FAD (0.1–10 µm), H4-biopterin (0.3–30 µm), CaM (0–10 µg/mL) and CaCl2 (0–0.350 mm) were also added to the incubation medium. The reaction was started by the addition of NADPH (075 mm final concentration). The other steps in the procedure were the same as described for nNOS activity determination.

Fluorescence assays with calmodulin

A reduction in Alexa-CaM fluorescence was used to test the binding specificity of aMT, AFK and AMK to CaCaM (Molnar et al. 1995; León et al. 2000). For this purpose, aliquots of 1 µm Alexa-CaM either alone or in the presence of different concentrations of aMT, AFMK, AMK and other indole derivatives including MT, OHT, HT, HIAA and NAS, were incubated with 100 µL of MOPS (10 mm), pH 7.2 containing KCl (100 mm) and CaCl2 (1 mm) for 30 min at 22°C. The samples in 100 µL quartz cuvettes were excited at 594 nm, and the emission at 622 nm was collected as relative fluorescence units in a Hitachi FI-200 fluorometer (Hitachi Ltd., Tokyo, Japan).

Statistical analysis

Data are expressed as the means ± SEM. Statistics included a one-way anova and a post hoc test. To assess any significant difference between CaM concentrations, a two-way anova was applied.


Characteristics of striatal nNOS activity inhibition by melatonin metabolites in vitro

In experiments reported elsewhere, we observed that aMT inhibited the activity of the nNOS present in rat striatal homogenate in a dose-dependent manner (León et al. 2000). Here, we found that AFMK, the initial metabolite of the aMT metabolic pathway, did not influence nNOS activity up to a concentration of 1 mm(Fig. 1). However, AMK, the subsequent metabolite of this pathway, inhibited nNOS activity in rat striatum homogenate in a dose-related manner (Fig. 1). The calculated IC50 for AMK was 70 µm, significantly lower than the reported IC50 for aMT (> 1 mm) (León et al. 2000).

Figure 1.

 Dose–response inhibition of neuronal nitric oxide synthase (nNOS) activity by N1-acetyl-N2-formyl-5-methoxykynuramine (AFMK, ▪) and N1-acetyl-5-methoxykynuramine (AMK, bsl00072). Data are expressed as a percentage of control nNOS activity measured in homogenates of rat striatum. Each point is the mean ± SEM of three experiments performed in triplicate. *p < 0.05, **p < 0.01, ***p < 0.001 (vs. ∞).

Urea–PAGE assays of CaCaM in the presence and absence of EGTA demonstrated that aMT binds to CaCaM (León et al. 2000). Herein, we examined the ability of several indole derivatives to modify the Alexa-CaM emission fluorescence intensity. A reduction in CaM fluorescence was detected in the presence of aMT, which was further significantly reduced when AMK was in the incubation medium (Fig. 2a). No changes in CaM fluorescence were detected when CaM was incubated with other indole derivatives including AFMK, MT, OHT, HT, HIAA and NAS. To further investigate the specificity of this binding, Alexa-CaM was incubated with aMT, AFMK and AMK at concentrations ranging from 10−11 to 10−3 m (Fig. 2b). A concentration-dependent inhibition of fluorescence was found in the presence of aMT and AMK, but not with AFMK. The IC50 values for AMK (65 µm) and aMT (>1 mm) suggest that the former binds to CaCaM with higher affinity than does aMT.

Figure 2.

 Binding ability of melatonin (aMT, •), N1-acetyl-N 2-formyl-5-methoxykynuramine (AFMK, bsl00072) and N1-acetyl-5-methoxykynuramine (AMK, ▪) to calmodulin. Each point is the mean ± SEM of three experiments performed in triplicate. (a) AlexaFluorTM 594-conjugated CaM (Alexa-CaM) was incubated in the presence of either vehicle (C) or 1 nm of the following compounds: aMT, AFMK, AMK, 5-hydroxytryptophan (HT), 5-hydroxytryptamine (OHT), 5-methoxytryptamine (MT), 5-hydroxyindole acetic acid (HIAA) or N-acetyl-serotonin (NAS), and the fluorescence inhibition caused by the binding of these compounds to calmodulin was measured. **p < 0.01, ***p < 0.001 (vs. vehicle); ##p < 0.01 (vs. aMT). (b) Dose-dependent (1 pm−1 mm) effects of aMT, AFMK and AMK on the inhibition of Alexa-CaM fluorescence. *p < 0.05, **p < 0.01, ***p < 0.001 (vs.∞); #p < 0.05, ##p < 0.01, ###p < 0.01 (vs. aMT).

Kinetics of the nNOS activity inhibition by AMK

To investigate the mechanism by which AMK inhibits nNOS activity, rat striatal homogenates were incubated with AMK in the presence of increasing concentrations of l-arginine (0–10 µm). Striatal nNOS activity was saturable and proportional to the substrate concentration (Fig. 3a). In the presence of 1 mm AMK, the activity of the enzyme decreased significantly, as assessed by the Lineweaver–Burk double reciprocal analysis of the data (Fig. 3b). The Km values of control (2.8 ± 0.2 µm) and AMK (2.9 ±0.2 µm) were similar, but the Vmax value for AMK (9.9 ± 0.4 pmol/min/mg protein) was significantly lower than that of the control (29.7 ± 2.2 pmol/min/mg protein). The presence of AFMK did not change the activity of the enzyme (data not shown).

Figure 3.

 Experiments with homogenates from rat striatum showing the kinetics of enzyme-substrate reaction in the presence of N1-acetyl-5-methoxykynuramine (AMK). (a) Homogenates from rat striatum were incubated for 30 min at 37°C with increasing concentrations of l-arginine and in either the absence (•) or the presence (bsl00066) of 1 mm AMK. Each point is the mean of three experiments performed in triplicate. (b) Double-reciprocal plot of the data showing that aMK modified the Vmax and not the Km values of the enzyme-substrate reaction. Each point is the mean ± SEM of three experiments performed in triplicate.

We next tested for the possible existence of an interaction with nNOS co-factors. For this purpose, we used a commercially available purified rat nNOS. In a previous report, the kinetic features of this form of nNOS in terms of its Ca2+- and CaM-dependency were analysed (León et al. 2000). Herein, the effect of AMK on nNOS activity was investigated in an incubation medium containing 0.0125 U of purified nNOS, 17.5 mm CaCl2, 10 µm FAD and 30 µm H4-biopterin. In the absence of added CaM, AMK inhibited purified nNOS activity in a dose-dependent manner, with the effect being significant at 1 nm(Fig. 4a). The incorporation of increasing quantities of CaM into the incubation medium resulted in a progressive loss of the ability of AMK to inhibit nNOS. At 10 µg/mL of CaM, only the highest dose of AMK used, i.e. 1 mm, was able to inhibit the enzyme activity. Fixing the CaM concentration at 0.1 µg/mL, the addition of different concentrations of either FAD (0.1–10 µm; Fig. 4b) or H4-biopterin (0.3–30 µm; Fig. 4c) in the incubation medium did not change the enzyme inhibition by AMK.

Figure 4.

 Effect of N1-acetyl-5-methoxykynuramine (AMK) (10−9−10−3 m) on neuronal nitric oxide synthase (nNOS) activity in the presence of increasing concentrations of nNOS co-factors. (a) Calmodulin (CaM; bsl00072, 0 µg/mL; □, 0.1 µg/mL; •, 1 µg/mL; bsl00083, 10 µg/mL), FAD = 10 µm and H4-biopterin = 30 µm. (b) FAD (bsl00072, 0.1 µm; □, 1 µm; ○, 10 µm), CaM = 0.1 µg/mL and H4-biopterin = 30 µm. (c) H4-biopterin (bsl00072, 0.3 µm; □, 3 µm; •, 30 µm), CaM = 0.1 µg/mL and FAD = 10 µm. The data show the percentage of the inhibition of control nNOS activity and are the means of three experiments performed in triplicate. Comparisons between AMK doses: *p < 0.05, **p < 0.01, ***p < 0.001 (vs. ∞). Comparisons between CaM doses: #p < 0.05 (vs. CaM 0.1 and 1 µg/mL); ###p < 0.001 (vs. CaM 10 µg/mL).

Effects of norharmane on striatal nNOS inhibition by AMT and AMK in vitro

A preliminary study to assess the dose of norharmane that blocks the activity of IDO was performed (Fig. 5). As expected (Chiarugi et al. 2000), high doses of norharmane (0.5–4 mm) reduced the nNOS activity in homogenate and purified nNOS samples. Unexpectedly, however, low doses of norharmane (0.05 mm) slightly increased the nNOS activity in homogenates but not in the purified preparation. Nevertheless, a dose of 0.05 mm was used for the following experiments as it completely reduced IDO activity (data not shown). Thus, the dose-dependent inhibition of nNOS activity by aMT and AMK in rat brain homogenates was assessed in the presence of 0.05 mm norharmane (Fig. 6). The inhibition elicited by aMT was blunted in the presence of this dose of norharmane. The presence of norharmane, however, did not significantly influence the inhibition produced by AMK (Fig. 6). Similar experiments were performed in the presence of 1-methyl-tryptophan, another IDO inhibitor, and the results were essentially the same (data not shown).

Figure 5.

 Dose–response effects of norharmane on neuronal nitric oxide synthase (nNOS) activity. Doses ranging from 0.05 mm to 4 mm norharmane were added to the incubation medium containing an aliquot of either rat striatum homogenate or purified nNOS preparation, and the activity of the enzyme was measured. The data show the percentage of inhibition of control nNOS activity and are the means of three experiments performed in triplicate. *p < 0.05, **p < 0.01, ***p < 0.001 (vs. control).

Figure 6.

 Effects of norharmane on melatonin (aMT)- and N1-acetyl-5-methoxykynuramine (AMK)-induced neuronal nitric oxide synthase (nNOS) inhibition in vitro. (a) Dose–response effects of aMT-induced nNOS inhibition. (b) Dose–response effects of AMK-induced nNOS inhibition. In both cases, rat striatal homogenates were incubated with different concentrations of either aMT or AMK in the presence (○, bsl00083, respectively) or absence (•, bsl00072, respectively) of 0.05 mm norharmane. Each point is the mean ± SEM of three experiments performed in triplicate. *p < 0.05, **p < 0.01, ***p < 0.001 (vs. ∞); ##p < 0.01, ###p < 0.001 (vs. aMT alone).

Effects of norharmane on striatal nNOS inhibition by AMT, AFMK and AMK in vivo

Experiments conducted to follow the kinetics of striatal IDO activity after norharmane administration in vivo were assessed. Four hours after the administration of 100 mg/kg norharmane IDO activity in rat striatum was totally blocked, and the inhibition was maintained 1 h later (Fig. 7). Thus, to study the effects of aMT and metabolites on in vivo nNOS activity in the rat striatum, we injected these compounds 4 h after the administration of norharmane, and the animals were killed 1 h later, i.e. 5 h after the administration of norharmane.

Figure 7.

 Effects of norharmane on indoleamine-2,3-dioxygenase (IDO) activity in vivo. Rats were injected i.p. with 100 mg/kg and killed at different times after injection. Data are expressed as a percentage of control IDO activity measured in the homogenates of rat striatum and are the means of three experiments performed in triplicate. **p < 0.01; ***p < 0.00 1 (vs. time 0).

The effects of the administration of different doses of aMT and AMK on the activity of rat striatum nNOS were analysed. The results of these experiments are depicted in Fig. 8. The results showed that 10 mg/kg AMK produced a similar percentage nNOS inhibition as did 20 mg/kg aMT. In the presence of norharmane, the inhibitory effect of aMT administration on nNOS activity was blunted (Fig. 9). Norharmane administration, however, did not significantly modify the inhibition of striatal nNOS activity elicited by AMK administration (Fig. 9). No effect was observed after AFMK administration. An additional interesting finding was that norharmane administration alone somewhat increased rat striatal nNOS activity.

Figure 8.

 Effects of melatonin (aMT) and N1-acetyl-5-methoxykynuramine (AMK) administration on neuronal nitric oxide synthase (nNOS) activity in vivo. Rats were i.p. injected with 10 and 20 mg/kg of either aMT or AMK and the animals were killed 1 h later. Data are expressed as a percentage of control nNOS activity measured in the homogenates of rat striatum and are the means of three experiments performed in triplicate. *p < 0.05, **p < 0.01, ***p < 0.001 (vs. control).

Figure 9.

 Effect of norharmane (NH) on neuronal nitric oxide synthase (nNOS) activity in vivo. Rats were injected i.p. with 100 mg/kg norharmane, and 4 h later animals received an i.p. injection of melatonin (aMT), N1-acetyl-N 2-formyl-5-methoxykynuramine (AFMK) or N1-acetyl-5-methoxykynuramine (AMK). Rats were killed 5 h after norharmane administration. Data are expressed as a percentage of control nNOS activity measured in the homogenates of rat striatum and are the means of three experiments performed in triplicate. *p < 0.05, **p < 0.01 (vs. control).


The data presented in this study demonstrate that AMK inhibits nNOS through an interaction with CaCaM. These experiments also suggest that AMK is the most important nNOS antagonist among the intermediaries of the aMT metabolic pathway studied here, including aMT itself. The calculated IC50 values of aMT and AMK obtained from the in vitro experiments support this conclusion. When IDO, the enzyme that transforms aMT to AFMK, was blocked with the specific IDO antagonist norharmane, aMT was no longer capable of inhibiting nNOS in vitro, whereas the inhibition induced by AMK remained. Also in vivo, AMK displayed significantly higher nNOS antagonist efficacy than did aMT, and the administration of norharmane blunted the ability of aMT but not that of AMK to reduce nNOS activity in the rat striatum. The fact that AFMK does not influence nNOS confirms the importance of a free amino group in position 2 of the kynuramine molecule for nNOS inhibitory activity (Camacho et al. 2002, 2004; Entrena et al. 2005). Thus, the degree of transformation of aMT to AMK would be of importance for the endogenous brain nNOS regulation.

Kinetic studies showed that the inhibitory effect of AMK on nNOS activity measured in rat striatal homogenates was not prevented by increasing concentrations of l-arginine, the natural substrate for the enzyme. Similar results were found using a purified rat brain nNOS as a source. The data suggest that AMK, similar to its precursor aMT (León et al. 2000), behaves as a non-competitive antagonist of nNOS. The inhibition of nNOS activity by AMK was independent of either the FAD or the H4-biopterin concentration in the incubation medium. Increasing concentrations of CaM, however, resulted in a loss of the inhibitory activity of AMK. Changes in fluorescence obtained with Alexa-CaM experiments reflect conformational/structural changes, suggesting interaction with CaM. These experiments indicate that AMK binds with higher affinity to CaCaM than does aMT, confirming the results of the kinetic studies. The lack of changes in fluorescence in the presence of other indole derivatives including AFMK supports the specificity of the binding of AMK to CaCaM. The IC50 values obtained from the fluorescence experiments for AMK and aMT are in the same range as the IC50 values obtained for the inhibition of nNOS by these compounds. In fact, the current results also show that AMK inhibits nNOS activity in a dose-dependent manner. Compared with aMT, which reduced nNOS activity by 20% at 1 nm (León et al. 2000), AMK reduced nNOS to the same level at a concentration of 0.01 nm. Considering 1 nm aMT is the physiological blood concentration at night (Reiter 1991b), the calculated 5% aMT conversion to AFMK and AMK (Ferry et al. 2005) yields the required AMK concentration to inhibit nNOS by at least 20%. Considering that aMT treatment at doses of 10 mg/kg would produce plasma levels of AFMK/AMK of about 300–400 nm (Ferry et al. 2005), and that these molecules easily cross the blood–brain barrier, we can suspect that in our experimental paradigm, after 10 mg/kg AMK administration, AMK reached brain levels compatible with the 25% nNOS activity reduction found in rat striatum. However, the term ‘physiological’ applied to endogenous aMT levels in bodily fluids and cellular compartments has been recently revised because of the existence of both pineal and extrapineal sources of the indoleamine (Reiter et al. 2005). Although pineal production of aMT is responsible for its plasma circadian rhythm, other organs and body fluids including the CSF have aMT levels that are bewteen two and three orders of magnitude greater (Stefulj et al. 2001; Hardeland et al. 2005; Silva et al. 2005). Consequently, levels of AMK may be sufficiently high to exert regulatory effects on nNOS activity in the brain, not only in the striatum.

Either through IDO-dependent conversion or by single-electron transfer reactions when scavenging the hydroxyl radical, aMT is transformed to AFMK (Hayaishi 1976; Hirata et al. 1974; Tan et al. 2000b, 2002). In turn, AMK is formed by enzymatic deformylation of AFMK (Tan et al. 2000b; Silva et al. 2004). AMK exhibits potent antioxidant properties, exceeding those of AFMK (Ressmeyer et al. 2003), thus contributing to the antioxidant potential of aMT and AFMK (Tan et al. 2000b). Under normal conditions of minimal free radical generation, it is assumed that the enzymatic conversion of aMT would predominate over the oxidative pathway. Under such conditions, the inhibition of the enzymatic pathway would significantly decrease its transformation to AMK, reducing the effects of this metabolite. Two sets of experiments were performed to test this hypothesis. When IDO activity was inhibited, aMT did not further inhibit nNOS activity at any dose. However, IDO inhibition did not influence the reduction in nNOS activity induced by AMK. Besides, administration of either aMT or AMK to normal rats decreased striatal nNOS activity in a dose–response manner. In these experiments, the required dose of aMT was double the dose of AMK required to achieve the same inhibition of nNOS. When animals were previously injected with norharmane, the nNOS inhibition elicited by aMT disappeared, although the effect of AMK remained unchanged. Thus, it seems that the conversion of aMT to AMK is essential for its inhibitory nNOS activity, as no effect of AFMK on nNOS activity was detected.

The minor effect of norharmane on NOS activity requires discussion. After exposure of macrophage and mononuclear phagocytes to norharmane, a dose-dependent inhibition not only of IDO but also of iNOS and eNOS was reported (Chiarugi et al. 2000). At the dose used herein, however, norharmane not only did not inhibit NOS activity, but in vitro and in vivo it produced a slight increase in nNOS. No increased activity of nNOS was found when purified nNOS was used. We conclude that the inhibition of IDO by norharmane, which reduces the conversion of endogenous aMT to AMK, should also reduce an inhibition of nNOS by AMK, and thus a slight increased nNOS activity is found under these conditions. It is surmised that the inhibition of nNOS activity by AMK was not a consequence of norharmane. The data also suggest that under normal conditions, endogenous levels of AMK maintain a tonic inhibition of rat striatal nNOS activity. An additional point should be discussed here. To date, we do not know whether the binding sites for aMT and AMK at CaCaM are identical, and thus the β-carboline norharmane, which is an indole derivative, can inhibit the CaM binding of other indoles such as aMT, but not the binding of AMK, which is structurally different from the indoles. However, conformational studies, which support the theory that aMT and AMK match the pharmacophore with conformations of low energy, suggest that both compounds bind CaM at the same site (Entrena et al. 2005).

AFMK was recently identified in CSF samples of patients with viral meningitis, with values two orders of magnitude higher than the plasma levels of aMT (Silva et al. 2005). Although AMK levels were not measured, we can assume than AMK was also increased in the CSF of these patients. It was suggested that, under conditions of brain inflammation, aMT either newly released into the CSF or even partially produced locally by inflammatory cells, may be either enzymatically or non-enzymatically metabolized by ROS (Silva et al. 2005). Besides, inflammation-induced interferon-γ up-regulates IDO activity (Taylor and Feng 1991; Roy et al. 2005), thereby contributing to the transformation of aMT to AFMK and AMK. In these circumstances, in addition to providing the brain with additional antioxidant and anti-inflammatory molecules, i.e. AFMK and AMK (Mayo et al. 2005), aMT oxidation provides the brain with another neuroprotective mechanism, i.e. the inhibition of nNOS elicited by AMK. Because AMK interacts with NO (Guenther et al. 2005), suppression of nNOS activity may be accompanied by the elimination of NO; this would prevent an excess of NO and subsequent NO-dependent excitotoxicity (Garthwaite et al. 1989).

An important consideration is the consequence of manipulation of tryptophan metabolism. Whereas the methoxyindole pathway of tryptophan produces aMT and its metabolites, with neuroprotective activity (Acuña-Castroviejo et al. 1995), the kynurenine pathway produces potentially neurotoxic compounds, including l-kynurenine and quinolinic acid. An imbalance of these metabolic pathways is related to changes in CNS excitability (Muñoz Hoyos et al. 1997; Chiarugi et al. 2000). The IDO-dependent kynurenine pathway of tryptophan metabolism involves, besides neurotoxin production, the formation of the nicotinamide adenine dinucleotide (NAD) (Grant and Kapoor 2003). One consequence of increasing IDO activity in astroglial cells during inflammation is to maintain NAD levels. Recently, it was found that aMT recycles the NAD radical to NADH, whereas aMT is oxidized to AFMK (Tan et al. 2005a). The recycling of NADH by aMT not only might improve the efficiency of NADH in mitochondrial bioenergetics, but may also improve antioxidant defense mechanisms. Thus, the inhibition of the kynurenine pathway to prevent neurotoxin production from tryptophan may significantly reduce cell viability and CNS functions unless alternate precursors for NAD synthesis and either aMT or AMK are available (Grant and Kapoor 2003).

The effects and mechanisms of the action of AMK reported here suggested the use of AMK as a template for the development of new agents with either nNOS or iNOS selective antagonism (Camacho et al. 2002, 2004; Entrena et al. 2005). As a result of their similarity to kynurenines, these compounds were also tested for their possible effects on kynurenine 3-hydroxylase, the induction of which leads to the production of quinolinic acid. However, neither AMK nor other compounds tested affected this enzyme, thus preventing the neurotoxic side-effects that would be produced under either physiological or pharmacological conditions (Entrena et al. 2005). These results also permitted us the design of the pharmacophore for AMK and related compounds, which may be of great interest in the development of nNOS antagonists (Entrena et al. 2005).

In summary, we have shown for the first time that AMK inhibits nNOS activity, a regulatory mechanism that may be important under physiological conditions. By demonstrating that AFMK does not affect nNOS, our results prompt three important conclusions:

  • (i) the inhibitory effects of aMT on nNOS activity mainly depend on its endogenous conversion to AMK;
  • (ii) the aMT/AMK ratio is important for the physiological inhibition of nNOS;
  • (iii) the administration of AMK may have clinical implications in situations such as excitotoxicity, where a reduction of nNOS activity is important.

During brain inflammatory disease, increased aMT metabolism through induced IDO supplies the brain with two powerful antioxidant and anti-inflammatory molecules, AFMK and AMK, the latter of which is responsible for the inhibition of nNOS, thereby avoiding NO-dependent excitotoxicity. The neuroprotective cascade of the AMT family includes the increased NADH levels recycling the additional NAD produced by IDO induction.


This work was supported in part by the Instituto de Salud Carlos III (ISCIII, Spain) through grants PI02-1447, PI03-0817, PI02-1181, SAF02-01688 and G03/137. JL has a ‘Contrato de Investigadores’ (ISCIII, Spain); MIR and VT are predoctoral fellows from ISCIII, and LCL is a postdoctoral fellow from the Ministerio de Educación (Spain).