Inducible Nitric-oxide Synthase Generates Superoxide from the Reductase Domain*

In the absence of l-arginine, the heme center of the oxygenase domain of neuronal nitric-oxide synthase reduces molecular oxygen to superoxide (O⨪2). Our recent work has provided evidence that inducible NOS (iNOS) may also catalyze O⨪2 formation in macrophages. However, there has been a lack of direct evidence of superoxide generation from the purified iNOS, and it was previously hypothesized that significant O⨪2production does not occur. Moreover, the mechanism and enzyme site responsible for O⨪2 generation is unknown. To determine whether iNOS produces O⨪2 and to identify the mechanism of this process, we performed electron paramagnetic resonance measurements on purified iNOS using the spin trap 5,5-dimethyl-1-pyrrolineN-oxide. In the presence of NADPH, prominent O⨪2adduct signals were detected from iNOS. These signals were totally abolished by superoxide dismutase but not affected by catalase. High concentrations of l-arginine decreased this O⨪2formation, whereas its enantiomer d-arginine did not. Pre-incubation of iNOS with the flavoprotein inhibitor diphenyleneiodonium totally blocked these O⨪2 signals. Conversely, pretreatment of the enzyme with the heme blocker cyanide had no effect on O⨪2 generation. Furthermore, strong O⨪2 generation was directly detected from the isolated iNOS reductase domain. Together, these data demonstrate that iNOS does generate O⨪2, and this mainly occurs at the flavin-binding sites of the reductase domain.

Nitric oxide (NO), 1 a gaseous free radical, has been identified as a ubiquitous signaling molecule in biological systems (1). In cells or tissues, NO is produced by a family of NO synthases (NOSs), which utilize L-arginine, oxygen, and NADPH as sub-strates to synthesize NO as well as the coproduct L-citrulline (2,3). Three distinct isoforms of NOS have been cloned: neuronal NOS (nNOS, type I), inducible NOS (iNOS, type II), and endothelial NOS (eNOS, type III) (4). nNOS and eNOS are also referred to as constitutive NOS, while the expression of iNOS requires induction by microbial endotoxins or cytokines. With tightly bound calmodulin, iNOS is fully active at basal Ca 2ϩ levels, whereas constitutive nNOS and eNOS activity depend on the elevation of intracellular Ca 2ϩ . In addition to calmodulin-binding sites, all three NOS isoforms also contain FAD, FMN, and tetrahydrobiopterin (BH 4 ) binding sites and require these cofactors for their enzymatic function (5).
Whereas the O 2 . generation from nNOS was well documented both in vitro and in intact cells, controversy remains regarding whether iNOS is also capable of producing O 2 . . Considering the similarity in amino acid sequence and enzymatic function between these two isoforms, it would be expected that iNOS will also generate O 2 . . However, iNOS was reported to be much less prone to oxidize NADPH than nNOS under conditions of Larginine depletion, and it was presumed that iNOS does not produce significant amounts of O 2 . (9). This highly limited O 2 . -generating capacity from iNOS has been proposed to be critical for its biological functions since O 2 . could in turn react with and scavenge NO, and it was hypothesized that this would perturb iNOS-mediated immune defense actions (9,10

EXPERIMENTAL PROCEDURES
Materials-The NADPH, L-or D-arginine, BH 4 , calmodulin, L-NAME, dithiothreitol, superoxide dismutase (SOD), catalase, and other reagents were purchased from Sigma unless otherwise noted. Diphenyleneiodonium (DPI) and sodium cyanide (NaCN) were from * This work was supported by National Institutes of Health Grants HL-38324 and HL-52315 (to J. L. Z.) and by National Institutes of Health Grant GM-52419 and The Robert A. Welch Foundation Grant AQ-1192 (to B. S. S .M.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Supported by a grant from American Heart Association Maryland Affiliate (MDFW3797).
Alexis Biochemicals (San Diego, CA) and Fisher, respectively. Cell culture materials were obtained from Life Technologies, Inc. 2Ј,5Ј-ADP-Sepharose and S-200 gel were the products of Amersham Pharmacia Biotech. L-[ 14 C]Arginine was purchased from NEN Life Science Products. DMPO was purchased from Aldrich and further purified by double distillation. DEPMPO was prepared as reported (12).
iNOS Purification-The expression and purification of iNOS was based on that of nNOS, as described previously (13). A similar expression system for iNOS has been reported (14). In brief, two plasmids containing mouse iNOS gene (iNOSpCW) and rat calmodulin sequence (pACMIP) were cotransfected into Escherichia coli BL21 cells via electroporation. Under selective pressure (50 g/ml ampicillin and 35 g/ml chloramphenicol), bacteria were grown to A 600 ϭ 0.8-1.2 at 37°C, and protein expression was induced with 0.25 mM isopropyl ␤-D-thiogalactopyranoside. The heme and flavin precursors, ␦-aminolevulinic acid and riboflavin, were also added to final concentrations of 450 and 3 M, respectively. The flasks were moved to room temperature (22-25°C) and shaken in the dark at 250 rpm. The cells were harvested at about 40 h postinduction, and the cell paste was frozen at Ϫ80°C until purification. Harvested cells were resuspended in 30 ml of resuspension buffer (100 mM Tris-HCl, pH 7.4, 1 mM EDTA, 1 mM dithiothreitol, 10% glycerol, and 1 mM phenylmethylsulfonyl fluoride) per liter of initial culture and were lysed by sonication. After centrifugation (150,000 ϫ g for 60 min), the supernatant was applied to 2Ј,5Ј-ADP-Sepharose 4B column equilibrated in buffer B (50 mM Tris-HCl, pH 7.4, 0.1 mM EDTA, 0.1 mM dithiothreitol, and 10% glycerol). The column was washed with at least 10 column volumes of buffer B. The protein was eluted with buffer B containing 500 mM NaCl and 5 mM 2Ј-AMP, and is approximately 50% pure after this step. The iNOS-containing fractions were concentrated using a Centriprep 30 (Amicon). After reconstitution with BH 4 (250 M), this fraction was applied to a S-200 gel filtration column. The iNOS-containing fractions were pooled, concentrated, and stored at Ϫ80°C. iNOS concentration was determined by CO-difference spectrum, assuming an extinction coefficient of 75 mM Ϫ1 cm Ϫ1 . The purity of iNOS was determined by SDS-polyacrylamide gel electrophoresis and visualized with Coomassie Blue staining. iNOS activity was up to 1300 nmol/min/mg at 37°C assayed by monitoring the conversion of L-[ 14 C]arginine to L-[ 14 C]citrulline as described below. iNOS Reductase Domain Construct and Purification-The iNOS reductase domain construct is residues 499 -1144 of the holoenzyme, encompassing the calmodulin binding site. An alanine and six histidine residues were added directly following the initiation methionine (MAH-HHHHH . . . ). At the beginning of the cDNA is an NdeI (CATATG) site, and at the end is a HindIII site (AAGCTT). The construct was made by inserting polymerase chain reaction-amplified DNA, incorporating the above changes, from the iNOSpCW plasmid into NdeI/HindIII restricted pCW vector. The resultant iNOS reductase domain plasmid was coexpressed with the calmodulin plasmid (pACMIP), as in the holoenzyme. The growth was the same as the above except that delta-ALA was not added at induction. The cells were harvested, and the supernatant of cell lysate was loaded on 2Ј,5Ј-ADP-Sepharose 4B column. Fractions containing iNOS reductase domain were pooled and loaded onto a nickel-charged metal chelating column in buffer B plus 5 mM imidazole. The column was washed with about 10 volumes of this buffer, and the iNOS reductase domain was eluted with buffer B containing 100 mM imidazole. 14 C]arginine, 0.5 mM NADPH, 0.5 mM Ca 2ϩ , 10 g/ml calmodulin, 10 M BH 4 , and 3 g/ml purified NOS. After a 5-min incubation at ambient temperature (23°C) or 37°C, the reaction was terminated. L-[ 14 C]Citrulline was separated by passing reaction mixtures through Dowex AG 50W-X8 (Na ϩ form, Bio-Rad) cation exchange columns and quantitated by liquid scintillation counting (15,16).

L-[ 14 C]Arginine to L-[ 14 C]Citrulline Conversion Assay-iNOS-cata-
EPR Spectroscopy and Spin Trapping-Spin trapping measurements of oxygen free radicals were performed in 50 mM Tris-HCl buffer (pH 7.4) containing 0.5 mM NADPH, 0.5 mM Ca 2ϩ , 10 g/ml calmodulin, 7.3 g/ml purified iNOS, and the spin trap DMPO (50 mM) or DEPMPO (25 mM). EPR spectra were recorded in a quartz flat cell at room temperature (23°C) with a Bruker ER 300 spectrometer operating at X-band with a TM 110 cavity using a modulation frequency of 100 kHz, modulation amplitude of 0.5 G, microwave power of 20 mW, and microwave frequency of 9.785 GHz as described (11,17). The microwave frequency and magnetic field were precisely measured using an EIP 575 microwave frequency counter and Bruker ER 035 NMR gauss meter. Quantitation of the free radical signals was performed by comparing the double integral of the observed signal with that of a known concentra-tion of TEMPO free radical in aqueous solution as described previously (17).

RESULTS
Recombinant mouse iNOS was isolated from an E. coli expression system. This system has been proven to be a powerful and efficient way to prepare large quantities of NOS proteins (13,18). Recombinant eNOS or nNOS from this system has exhibited enzymatic properties indistinguishable from native enzymes isolated from mammalian cells (13,19). iNOS was isolated by NADPH affinity chromatography, and the purity of the preparations was further improved by size exclusion chromatography. As shown in Fig. 1, top, purified protein preparations exhibited a prominent major band (Ͼ90% pure) on SDSpolyacrylamide gel electrophoresis with a molecular mass of 130 kDa, which is in agreement with the molecular mass for native iNOS as previously reported (4,5). This recombinant protein possessed strong NOS activity as measured from conversion of L-arginine to L-citrulline (Fig. 1, bottom). The catalytic activity was independent of additional Ca 2ϩ and could be blocked by the NOS inhibitor, L-NAME (1 mM), confirming that it was derived from iNOS.
To determine whether iNOS generates O 2 . , EPR spin-trapping measurements were performed on iNOS using the well characterized spin trap DMPO. In the absence of the enzyme, no signals were observed from solutions containing DMPO and NADPH (Fig. 2, top, trace A). However, after adding purified iNOS (7.3 g/ml), strong EPR signals were seen (Fig. 2, bottom,  N ϭ 14.2 G, a H ϭ 11 . production (21). As shown in the Fig. 3 A, no EPR signals were observed in the reaction system with 25 mM DEPMPO in the absence of iNOS. After adding iNOS, prominent DEPMPO-OOH signals were seen (Fig. 3B). These signals were totally quenched by SOD (200 units/ml, Fig.  3C Fig. 4A, not shown). However, L-arginine at high concentrations (1-5 mM) markedly decreased the O 2 . signals (Fig. 4, B and C). This inhibition was specifically elicited by L-arginine because its enantiomer, D-arginine, at the same concentration (5 mM) had no effects (Fig. 4D)  the control experiments without the iNOS reductase domain, no EPR signal was detected (Fig. 5A). After adding the iNOS reductase domain (7 g/ml), strong O 2 . generation was observed as expected (Fig. 5B). These O 2 . signals were totally abolished by SOD (200 units/ml, Fig. 5C). In parallel to the results obtained from holoenzyme, O 2 . production from the iNOS reductase domain was also prevented by DPI (20 M, Fig. 5D nine, which is also in accordance with our previous observations in macrophages where iNOS-derived oxygen radical generation was stimulated by L-arginine depletion (11). Second, our results indicate that O 2 . synthesis from iNOS mainly occurs at the flavin-binding sites of its reductase domain. This is different from the process of O 2 . generation from nNOS which is thought to occur primarily at the oxygenase domain (5)(6)(7)(8). This may have important implications in the enzymatic and biological function of iNOS. Although derived from separate genes and chromosomes, the three NOS isoforms share 50 -60% identity in their amino acid sequence (4,5). They are all bi-domain enzymes consisting of a C-terminal reductase and N-terminal oxygenase. The reductase domain contains NADPH, FAD, and FMN binding sites and exhibits 58% homology to NADPH-cytochrome P-450 reductase (22). Binding sites for heme, BH 4 , and L-arginine are located at the oxygenase domain. A unified model has been proposed to explain the enzymatic mechanism of NOSs (5). Their catalytic mechanisms involve flavin-mediated electron transport from C-terminal-bound NADPH and flavins to an N-terminal heme center, where oxygen is reduced and incorporated into the guanidino group of L-arginine giving rise to NO and L-citrulline. Calmodulin binds to a consensus sequence in the NOS enzymes and serves to position the two domains allowing the electron transfer from FMN to heme (5,23  from nNOS (7) 1-10 M). The fact that the signal was not affected by 100 M L-arginine indicates that DMPO in the 50 mM concentrations used effectively outcompeted NO for reaction with O 2 . . The loss of signal at high L-arginine concentrations is most likely due to substrate inhibition. A nonspecific effect of the amino acid is less likely because identical concentrations of D-arginine had no effect. Similarly a reaction between arginine and DMPO-OOH is unlikely because D-arginine had no effect. We hypothesize that higher concentrations of L-arginine suppress the O 2 . generation from iNOS either by altering the conformation of the protein and accessibility of the flavin or by rendering the flavin in a more oxidized state due to more rapid electron transfer to the heme.