Endogenous Methylarginines Modulate Superoxide as Well as Nitric Oxide Generation from Neuronal Nitric-oxide Synthase DIFFERENCES IN THE EFFECTS OF MONOMETHYL- AND DIMETHYLARGININES IN THE PRESENCE AND ABSENCE OF TETRAHYDROBIOPTERIN*

The endogenous methylarginines asymmetric dimethylarginine (ADMA) and N G -monomethyl- L -argi-nine ( L -NMMA) regulate nitric oxide (NO) production from neuronal NO synthase (nNOS). Under conditions of L -arginine or tetrahydrobiopterin (BH 4 ) depletion, nNOS also generates superoxide, O 2 .; however, the effects of methylarginines on this O 2 . generation are poorly understood. Therefore, we measured the dose-dependent effects of ADMA and L -NMMA on the rate and amount of O 2 . production from nNOS under conditions of L -arginine and/or BH 4 depletion, using elec- tron paramagnetic resonance spin trapping. In the absence of L -arginine, ADMA (1 (cid:1) M ) inhibited O 2 . generation by (cid:1) 60% from a rate of 56 to 23 nmol/mg/ min, whereas L -NMMA (0.1–100 (cid:1) M ) had no effect. L - Arginine markedly decreased the observed O 2 . adduct formation; however, O 2 . generation from the enzyme still occurs at a low rate (12.1 nmol/mg/min). This O 2 . leak is NOS-derived as it is not seen in the absence of calcium and calmodulin 10 (cid:1) M BH 10 at 37 °C and then stopped with 3 ml of ice-cold stop buffer using 20 m M HEPES 2 m M EDTA, 2 M EGTA, pH 5.5. Separation of L -[ 14 C]citrulline from L -[ 14 C]arginine was performed using the cation exchange resin Dowex AG50WX-8 (0.5 ml, Na (cid:6) form; Sigma). The L -[ 14 C]citrulline in the eluent was then quanti- tated using a liquid scintillation counter.

The biological significance of guanidino-methylated arginine derivatives has been known since the inhibitory actions of N G -monomethyl-L-arginine (L-NMMA) 1 on macrophage-in-duced cytotoxicity were first demonstrated. It was subsequently determined that these effects were mediated through inhibition of NO formation (1). This naturally occurring arginine analogue, together with its structural congener asymmetric dimethylarginine (ADMA), form a pair of L-arginine derivatives with the ability to regulate the L-arginine:NO pathway. In fact, these two compounds, along with N G -nitro-L-arginine methyl ester (L-NAME), have been shown to be potent inhibitors of NOS activity (2)(3)(4)(5). It has been demonstrated that the methylarginine levels in isolated neurons and in the intact brain are sufficient to regulate NO production from neuronal NOS (nNOS) (4).
ADMA and L-NMMA are derived from the proteolysis of methylated arginine residues on various proteins. The methylation is carried out by a group of enzymes referred to as protein-arginine methyl transferases. Subsequent proteolysis of proteins containing methylarginine groups leads to the release of free methylarginine into the cytoplasm where NO production from NOS is inhibited (2,3,5). In addition to inhibition of NO generation, methylarginines may have other important effects on NOS function.
Our laboratory and several others have reported that when cells are depleted of the NOS substrate L-arginine, L-arg, or the cofactor tetrahydrobiopterin, BH 4 , NOS switches from production of NO to the superoxide anion radical, O 2 . (6 -13). In the absence of either of these requisite substrates or cofactors, NOS-mediated NADPH oxidation is uncoupled from NO synthesis and results in the reduction of O 2 to form O 2 . (6,7,11,14). another very important NOS product, and its production may also be regulated by methylarginines. Furthermore, in view of their strong inhibition of NO generation, methylarginines could profoundly modulate the balance of NO and O 2 . generation from the enzyme. Cytosolic L-arg concentrations are generally in the range of 50 to 200 M, and moderate L-arg depletion has been observed in conditions such as wound healing and aging (4,(15)(16)(17)(18)(19)(20). The redox active cofactor BH 4 has been shown to be highly susceptible to oxidative stress. Oxidation of BH 4 has been shown to result in NOS-derived O 2 . generation (9,10). However, little is known regarding the precise effects of methylarginines on NOS-derived O 2 . generation. Although L-NAME has been shown to block O 2 . production from nNOS, studies using L-NMMA have suggested that this endogenous methylarginine does not appear to inhibit O 2 . generation (7,11,13

Materials
Human embryonic kidney 293 cells stably transfected with nNOS were provided by the laboratory of Dr. Valina Dawson, Department of Neuroscience, The Johns Hopkins University School of Medicine. Unless noted otherwise, all chemicals were obtained from Sigma. D-NMMA was obtained from A. G. Scientific (San Diego, CA). Dowex was purchased from Amersham Biosciences. Chromatography columns were purchased from Evergreen Scientific (Los Angeles, CA). Centricon® concentrators were purchased from Amicon (Beverly, MA). The spin trap DEPMPO was purchased from Oxis (Portland, OR). The MGD was synthesized and iron complexes prepared as reported previously (21).

Methods
nNOS Purification-Rat nNOS was purified from stably transfected human kidney 293 cells. These nNOS-transfected cells were grown in minimum essential medium with 10% heat-inactivated fetal calf serum. Cells were then harvested and homogenized in 50 mM Tris-HCl (pH 7.4) containing 1.0 mM EGTA, 1.0 mM diethylenetriaminepentaacetic acid, and 10 mM 2-mercaptoethanol with 1.0 mM phenylmethylsulfonyl fluoride and 2 M leupeptin. After centrifugation at 5000 ϫ g for 10 min at 4°C, the supernatant was loaded on a 2Ј,5Ј-ADP-Sepharose affinity column. After washing the column with 0.45 M NaCl, nNOS was eluted with standard buffer containing 10 mM NADPH and concentrated using Centricon-30 concentrators. Excess NADPH, diethylenetriaminepentaacetic acid, and 2-mercaptoethanol were removed by repetitive washing and 40-fold concentration with 50 mM Tris-HCl (pH 7.4) containing 1.0 mM phenylmethylsulfonyl fluoride and 2 M leupeptin. Concentrated enzyme was then stored at Ϫ80°C in this buffer with the addition of 10% glycerol. Protein content was assayed by the method of Bradford using bovine serum albumin as a standard. The purity of nNOS was Ͼ90% as determined by electrophoresis on 7.5% SDS-polyacrylamide gels (13).
BH 4 Depletion-Human embryonic kidney 293 cells stably transfected with rat nNOS were cultured in T-150 flasks at 37°C, 95% humidity, and 5% CO 2 95% air. After reaching ϳ50% confluency, the cells were treated with 10 mM 2,4-diamino-6-hydroxypyrimidine (DAHP) for 72 h. At the end of this period, the cells had reached confluency and were prepared for nNOS purification as described above. Aliquots of the homogenate were prepared for BH 4 determination.
HPLC Measurements for BH 4 -Human embryonic kidney 293 cells were homogenized in ice-cold 50 mM Tris (pH 7.4) containing ascorbate (1 mg/ml). Ascorbate was added to prevent BH 4 auto-oxidation. 100-l aliquots were loaded into microcentricon tubes with a molecular mass cutoff of 4 kDa. The samples were centrifuged for 30 min at 10,000 ϫ g at 4°C. The wash through was removed and subjected to HPLC analysis for determination of intracellular BH 4 levels. The chromatographic system consisted of a Shimadzu pump with an ESA 7400 autosampler and a Toso Haas (Milford, MA) ODS-80Tm column (4.6 mm ϫ 25 cm inner diameter, 5-m particle size) reverse phase column. Electrochemical detection was carried out with an ESA coulochem 5600 with potential set at 0.4 and 1 V. Fluorescence detection was carried out at an excitation wavelength of 348 nm and an emission wavelength of 444. The mobile phase consisted of sodium acetate (6.8 g), citric acid (1.05 g), and EDTA (20 mg) at pH 5.8 adjusted with acetic acid with the total volume brought to 1 liter using HPLC-grade water. The mobile phase was filtered through a 0.22-M filter, and dithiothreitol (24.7 mg) was added immediately before use. The mobile phase was degassed under helium during the chromatographic analysis with the flow rate set at 1.3 ml/min. EPR Spectroscopy and Spin Trapping-Spin trapping measurements of NO and oxygen radical generation were performed using a Bruker ER 300 spectrometer. The reaction mixture consisted of purified nNOS in 50 mM Tris, pH 7.4, containing 1 mM NADPH, 1 mM Ca 2ϩ , 30 M EDTA, 10 g/ml calmodulin, and 10 M BH 4 in 0.6 ml. For NO measurements, 5 g of nNOS and 100 M L-arg were added to the reaction system with Fe 2ϩ ⅐MGD (0.5 mM Fe 2ϩ and 5.0 mM MGD) used to trap NO, as described previously (22). The samples were loaded into a quartz flat cell and measured at X-band in a TM 110 cavity. Spectra were obtained using the following parameters: microwave power, 20 milliwatt; modulation amplitude, 3.16 G; modulation frequency, 100 kHz. For the detection of O 2 . , 5 g of nNOS was added to the reaction system with 10 mM DEPMPO as the spin trap. Spectra were obtained using the following parameters: microwave power, 20 milliwatt; modulation amplitude, 0.5 G; modulation frequency, 100 kHz. Quantitation of the free radical signals was performed by comparing the double integral of the observed signal with that of a known concentration of TEMPO (2,2,6,6-tetramethyl-piperidine-1-oxyl) free radical in aqueous solution. To quantify rates of O 2 . generation, adduct signals were corrected for trapping efficiency and decay rate as described previously (23,24). Rates of O 2 .
formation were determined from the DEPMPO-OOH signal over the first 20 min of acquisition. Citrulline Conversion Assay-nNOS activity was measured from the conversion of L-[ 14 C]arginine to L-[ 14 C]citrulline as reported previously (4). Briefly, the purified enzyme (5 g/ml) and cofactors were reacted with 5 M L-[ 14 C]arginine (317 mCi/mmol, PerkinElmer Life Sciences) in 50 mM Tris, pH 7.4, containing 1 mM NADPH, 1 mM Ca 2ϩ , 10 M calmodulin, and 10 M BH 4 for 10 min at 37°C and then stopped with 3 ml of ice-cold stop buffer using 20 mM HEPES with 2 mM EDTA, 2 mM EGTA, pH 5.5. Separation of L-[ 14 C]citrulline from L-[ 14 C]arginine was performed using the cation exchange resin Dowex AG50WX-8 (0.5 ml, Na ϩ form; Sigma). The L-[ 14 C]citrulline in the eluent was then quantitated using a liquid scintillation counter.

Effects of Methylarginines on NO Release from nNOS-To
determine the dose-dependent effects of ADMA and L-NMMA on the rate and amount of NO release from nNOS, EPR spin trapping measurements were performed on nNOS using the well characterized NO spin trap Fe⅐MGD. Purified nNOS (5 g) was incubated in the presence of NOS cofactors and substrates (NADPH, calmodulin, tetrahydrobiopterin, calcium, and 15 N-Larg in concentrations as described under "Methods"). We have previously demonstrated that normal neuronal levels of L-arg are ϳ100 M (4). A strong NO signal was observed exhibiting the characteristic doublet spectrum of the 15 NO⅐Fe⅐MGD complex. Isotopically labeled L-arg was used to prove that the nitrogen was derived from the guanidino group of L-arg and also ensured that only the NO from nNOS was measured (25). In the presence of 100 M L-arginine, both ADMA (100 M) and L-NMMA (100 M) significantly inhibited, by 75 and 79%, respectively, the total amount of NO formation observed over 30 min from nNOS ( Fig. 1). NO measurements were then performed using BH 4 -depleted nNOS. Under these conditions, as expected the enzyme generated almost no detectable NO ( Fig.  1) because it has been demonstrated that the pterin cofactor, BH 4 , is required for NO formation from nNOS (10, 14, 26 -30). . production was not further inhibited (Fig. 3).

Effects of Methylarginines on Superoxide Production in the Presence or Absence of L-Arg-
Parallel experiments were carried out using L-NMMA to determine how this endogenous methylarginine modulates NOS-derived O 2 . production. As previously reported, L-NMMA had no effect on O 2 . generation in the absence of L-arginine (Fig.   4). However, with L-arginine present, L-NMMA at 100 M re- sulted in a 44% increase in the rate of O 2 . production (Fig. 5).
Thus, from these studies it appears that the ability of L-NMMA to increase O 2 . is a result of its ability to compete with L-arg, thereby partially blocking the prominent inhibitory action of  ] arg to citrulline conversion assay, and Ͼ90% loss of NOS activity was observed. EPR spin trapping studies were performed using the NO spin trapping complex Fe⅐MGD. As shown in Fig. 1, there was almost no detectable NO generation from the enzyme prepared from the DAHP-treated cells, indicating that BH 4 levels had been depleted sufficiently to inhibit normal NOS activity. The addition of exogenous BH 4 fully restored NOS function, demonstrating that BH 4 effects are reversible and do not otherwise alter the enzyme.
To determine the effects of ADMA and L-NMMA on O 2 . release from BH 4 -depleted nNOS, EPR spin trapping measurements were performed using the spin trap DEPMPO. Purified nNOS (3 g) was incubated in the presence of NOS cofactors (NADPH, calmodulin, and calcium as described above) with or without the substrate L-arg. EPR measurements using the BH 4 -depleted nNOS gave rise to a strong O 2 . adduct signal (Fig. 6). The signal exhibited a rate of O 2 . formation 68% greater than that seen with L-arg depletion using BH 4 -coupled nNOS, suggesting that NOS uncoupling through BH 4 depletion results in an enhanced rate of   (Figs. 6 and 7). Contrary to the previous experiments, ADMA had no effect on BH 4 -depleted nNOS regardless of whether the substrate L-arg was present or not. Because it has been previously reported (28, 30 -33) that BH 4 is required for binding of L-arg or related substrates, this would be expected. Surprisingly, when the experiments were repeated in the presence of L-NMMA, we observed an ϳ2.6-fold increase in the rate of O 2 . generation from BH 4 -depleted nNOS in the absence of L-arg generation was observed (Fig. 9). Using this BH 4 -depleted nNOS, we observed that neither L-arg nor ADMA is capable of inhibiting O 2 . generation. The fact that L-NMMA is able to cause such a dramatic increase in O 2 . production, together with the data demonstrating dose dependence in the presence of L-arg, demonstrates that L-arg and L-NMMA are competing for the same binding site. However, L-NMMA binding appears to facilitate the transfer of electrons to O 2 , whereas L-arg binding has no effect.
To characterize the mechanism of O 2 . production from nNOS and its modulation by methylarginines, further experiments were carried out using various inhibitors, including the flavoprotein inhibitor diphenylene iodonium (10 M), L-NAME (100 M), superoxide dismutase (10 M), and imidazole (5 mM) (Fig. 10)  production. All four inhibitors largely blocked NOS-derived O 2 . generation. In the absence of calcium and calmodulin, no signal was observed.
Results shown represent the mean Ϯ S.E.; n Ն 4. In the presence of L-arg, L-NMMA (NMMA) (0. generation under varying conditions of substrate/cofactor availability. Thus, these data provide the first evidence demonstrating that endogenous methylarginine levels present within cells are sufficient to critically modulate NOS-derived O 2 . and that these effects are dependent on cofactor/substrate availability.

DISCUSSION
Over the last several years, studies have shown that in addition to producing NO, NOS is also capable of producing O 2 . under conditions of L-arg or tetrahydrobiopterin depletion (7)(8)(9)(10)(11)(12). In cells engineered to express nNOS and in neurons that intrinsically contain it, this O 2 . generation has been shown to be an important mechanism of cellular injury (4,11). Although questions remain regarding the severity of these conditions that arise in normal cells, there is evidence that normal cellular oxidation of BH 4 can increase O 2 . release (10). Furthermore, a range of disease conditions favors L-arg and/or BH 4 depletion. These include tissue damage and wound repair, ischemic syndromes, and inflammatory processes (15-17, 34 -38).
Although most prior studies indicate that marked L-arg depletion is required to detect O 2 . production from NOS, in the current study we have observed that in the presence of normal physiological levels of 100 M L-arg, nNOS continues to generate a significant amount of this important oxygen-radical. The level of O 2 . production in the presence of this physiological L-arg concentration remained at ϳ20% of that in the absence of L-arg. The use of the spin trap DEPMPO, which provides a much more stable O 2 . adduct than DMPO, used in most earlier studies, facilitated the detection of this lower level O 2 . production.
This O 2 . generation was calcium-and calmodulin-dependent, indicating that it was specifically derived from the enzyme. It was also largely (Ͼ90%) inhibited by imidazole, indicating that it was primarily derived from the heme of the oxygenase site. The endogenous methylarginine derivatives ADMA and L-NMMA are capable of regulating NO generation from nNOS. We have previously shown that their intrinsic levels in neurons and brain of ϳ10 M protect against excitotoxic injury (4). However, their role on O 2 . release from the enzyme has remained unclear. Therefore, the current studies were carried out to characterize and quantitate the dose-dependent effects of these endogenous methylarginines on the O 2 . generation from nNOS that is triggered by conditions of L-arg or BH 4 depletion. As noted above, L-arg depletion triggers prominent O 2 . generation (Table I). We observed that only ADMA inhibited O 2 .
generation from NOS in the absence of substrate. A significant, ϳ20%, inhibition was seen with 10 nM ADMA, with ϳ45% inhibition at 100 nM levels, whereas ϳ60% inhibition was seen at levels Ն1.0 M. Results obtained using L-NMMA demonstrated that this monomethylarginine had no effect on O 2 . production in the absence of L-arg but, interestingly, appeared to modestly enhance O 2 . generation in the presence of L-arg. This is likely because of its competitive binding to the L-arg binding site of nNOS. These results appear consistent with a prior report that L-NMMA in the presence of L-arg (100 M) dose-dependently reversed the inhibitory properties of L-arg on O 2 . production with levels reaching almost control values in the presence of 1 mM L-NMMA (7). Of note, we observed that the exogenous NOS inhibitor, L-NAME, also inhibits O 2 . production, and ϳ85% inhibition was seen at 500 M concentration. Together, these results indicate that ADMA and L-NAME are able to inhibit the transfer of electrons to O 2 , whereas L-NMMA binding only inhibits NO production while allowing the transfer of electrons from NADPH to O 2 . Although both ADMA and L-NMMA are known to function as competitive inhibitors of L-arg, it appears that only the more bulky dimethyl group impairs the binding and transfer of electrons to molecular oxygen.
Studies were also carried out to examine the effects of methylarginines on NOS-derived O 2 . production following BH 4 depletion. BH 4 -depleted nNOS was purified from cells following prolonged DAHP treatment. DAHP inhibits GTP cyclohyrolase-1, the rate-limiting enzyme in BH 4 synthesis. Following the DAHP treatment, BH 4 levels were undetectable by HPLC with electrochemical detection. Functional studies carried out using both EPR spectroscopy and the sensitive citrulline conversion assay confirmed the effective depletion of BH 4 ; no nitric oxide could be higher than those observed with L-arg depletion (Table I). Thus, with BH 4 depletion, not only the ability of the enzyme to produce NO is lost but also the ability of L-arg to suppress O 2 . generation.
The results obtained using BH 4 -depleted nNOS provided further information regarding the role of methylarginines in regulation of the enzyme. The removal of BH 4  Several studies have been published regarding the crystal structure of the reductase and oxygenase domains of NOS isoforms (41)(42)(43). These structures may provide some insight regarding the relationship between methylarginine binding and the observed effects on O 2 . or NO production. The structures of the L-arg binding site and the vicinity of the heme center are clearly of critical importance. L-arg, as well as other nitro-arginine compounds, have been shown to bind to NOS with the ␣-amino and carboxyl groups pointing up into the L-amino acidspecific binding pocket where key hydrogen bonding interactions occur with Gln-249, Tyr-359, Glu-363, and Asn-368 (29,41,42). The guanidine group is located roughly coplanar to the heme and makes a bifurcated hydrogen bonding interaction to a conserved active site glutamate at Glu-592 (41,44). Studies from endothelial NOS also suggest that N G -substituted groups can be accommodated in a hydrophobic patch defined by Phe-355 and Val-338 (43). These bonded or additional nonbonded contacts could modulate binding at the active site. Through these interactions, the additional methyl groups on ADMA and L-NMMA could alter the positioning between the guanidino nitrogen and the heme. In the case of ADMA, these interactions result in inhibition of electron transfer to O 2 and decreased O 2 . production. However, under conditions of BH 4 depletion, allosteric changes in overall enzyme conformation could occur, which upon L-NMMA binding facilitate O 2 binding to the heme center or electron transfer with its reduction to O 2 . (45). Further structural studies will be needed to establish precisely how methylarginines bind to the enzyme, the effects of BH 4 on this process, and the effects of these interactions on O 2 binding and electron transfer. Prior cellular studies provide evidence suggesting that L-NMMA shifts the balance of NO and O 2 . generation and enhances O 2 . in cells. We have previously observed that although L-NMMA is as effective as ADMA in inhibiting NO generation, it is less effective in preventing excitotoxic neuronal death (4). Another interesting study (46) reports that L-NMMA enhanced O 2 . generation in cardiac myocytes subjected to ischemia-reperfusion injury. These results were previously unexplained; however, ischemia-reperfusion has been shown to result in BH 4 oxidation, so the increased O 2 . production induced by L-NMMA is likely NOS-derived and BH 4 -dependent. Further studies will be needed to characterize the effects of endogenous levels of methylarginines on O 2 . production in cells and tissues as well as the modulatory action of BH 4 on methylarginine-NOS interactions in in vivo models of physiology and disease. Overall, our findings suggested that endogenous L-arg derivatives play an important role in the regulation of NO  injury, neurodegenerative disease, stroke, and postischemic syndromes, methylarginines may play a critical role in modulating the balance of NO and oxygen radical formation as well as oxidant injury in these and other critical disorders.