Reduction of Nitrite to Nitric Oxide Catalyzed by Xanthine Oxidoreductase*

Xanthine oxidase (XO) was shown to catalyze the reduction of nitrite to nitric oxide (NO), under anaerobic conditions, in the presence of either NADH or xanthine as reducing substrate. NO production was directly demonstrated by ozone chemiluminescence and showed stoichiometry of approximately 2:1 versus NADH depletion. With xanthine as reducing substrate, the kinetics of NO production were complicated by enzyme inactivation, resulting from NO-induced conversion of XO to its relatively inactive desulfo-form. Steady-state kinetic parameters were determined spectrophotometrically for urate production and NADH oxidation catalyzed by XO and xanthine dehydrogenase in the presence of nitrite under anaerobic conditions. pH optima for anaerobic NO production catalyzed by XO in the presence of nitrite were 7.0 for NADH and < 6.0 for xanthine. Involvement of the molybdenum site of XO in nitrite reduction was shown by the fact that alloxanthine inhibits xanthine oxidation competitively with nitrite. Strong pref-erence for Mo 5 S over Mo 5 O was shown by the relatively very low NADH-nitrite reductase activity shown by desulfo-enzyme. The FAD site of XO was shown not to influence nitrite reduction in the presence of xanthine, although it was clearly involved when NADH was the reducing substrate. Apparent production of NO decreased

Xanthine oxidoreductase (XOR) 1 is a complex flavoprotein comprising two identical subunits of M r 145,000. Each subunit contains one molybdenum, one FAD, and two nonidentical ironsulfur redox centers (1,2). Although it has broad specificities for both reducing and oxidizing substrates, its conventionally accepted role is in purine catabolism, where it catalyzes the oxidation of hypoxanthine to xanthine and xanthine to uric acid. In mammals, XOR exists in two interconvertible forms, xanthine dehydrogenase (XDH; EC 1.1.1.204), which predominates in vivo, and xanthine oxidase (XO; EC 1. 1.3.22). Both forms of the enzyme can reduce molecular oxygen, although only XDH can reduce NAD, which is its preferred electron acceptor. Reduction of oxygen leads to superoxide anion and hydrogen peroxide; it is the potential to generate these reactive oxygen species that has led to widespread interest in the enzyme as a pathogenic agent in many forms of ischemia-reperfusion injury (3). More recently, reactive oxygen species are increasingly cited as intermediates in normal signal transduction pathways (4,5).
Nitric oxide (NO) also is a biological messenger and, over the last decade, has become the focus of intense research activity in its own right. It is implicated in a wide range of physiological processes, including smooth muscle relaxation, inhibition of platelet aggregation, neurotransmission, and immune regulation (6 -9). It is generally accepted that the enzymic source of NO in mammalian tissues is NO synthase, which catalyzes the oxidation of arginine in the presence of NADPH and molecular oxygen (10). We have recently briefly reported that XO is capable of catalyzing the reduction of glyceryl trinitrite, as well as inorganic nitrate and nitrite, to NO under hypoxic conditions in the presence of NADH (11). We now examine the kinetics and stoichiometry of NO production catalyzed by XOR in the presence of inorganic nitrite and both NADH and xanthine as reducing substrates under conditions of different oxygen tensions and provide evidence for the sites of action of the various substrates.

EXPERIMENTAL PROCEDURES
Xanthine Oxidase and Xanthine Dehydrogenase-XOR was purified from bovine milk, essentially as described by Sanders et al. (12) with minor modifications. Fresh milk (1500 -2000 ml) was centrifuged (2000 ϫ g, 30 min) at 4°C to separate the cream. All subsequent operations were carried out at 4°C. The cream was resuspended in 5 volumes of 0.2 M K 2 HPO 4 , containing 1 mM EDTA and 5.0 mM dithiothreitol, with stirring for 90 min. The suspension was centrifuged (3000 ϫ g, 30 min), and the subnatant was collected, filtered through glass wool, and mixed slowly with 15% (by volume) cold butanol (Ϫ20°C). Ammonium sulfate (15 g/100 ml) was added slowly with stirring, which was continued for a further 50 min, before centrifugation (8000 ϫ g, 20 min) and filtration of the supernatant through glass wool. Ammonium sulfate (20 g/100 ml) was then added slowly to the supernatant with mixing, and the suspension was stirred for a further 30 min before centrifuging (9500 ϫ g, 30 min). The brown precipitate was suspended in a small volume of buffer A (25 mM Na/MES, pH 6.0, containing 1 mM EDTA) and dialyzed against the same buffer (3.5 liters) overnight. Any precipitate was removed by centrifugation (10,000 ϫ g, 60 min) and the supernatant was filtered through a 0.22-m membrane, before applying to a column (1 ϫ 9 cm) of heparin immobilized on cross-linked 4% beaded agarose (Sigma, type 1), previously equilibrated in buffer A. The column was washed with 50 mM NaCl in buffer A (20 ml), and XOR was eluted with 0.3 M NaCl in the same buffer.
Further purification was effected by ion exchange chromatography.
Eluate from the heparin column was dialyzed into 50 mM Na/Bicine, pH 8.3, containing 50 mM NaCl, filtered through a 0.22-m syringe filter, and applied to an equilibrated HiTrap Q ion exchange column (Amersham Pharmacia Biotech) for fast protein liquid chromatography. The column was eluted with 50 mM Na/Bicine, pH 8.3, followed by an increasing salt gradient (0.05-1.0 M NaCl) in buffer, when enzymecontaining fractions were pooled and concentrated. Typical yields after heparin chromatography were 15 mg of enzyme protein/liter of milk, with A 280 /A 450 5.0 -5.5. Mono Q chromatography yielded enzyme with A 280 /A 450 ratios of 5.0 -5.2 in yields of approximately 85%. The enzyme-containing eluate from HiTrap Q was dialyzed overnight into 50 mM Na/Bicine, pH 8.3, frozen dropwise in liquid nitrogen, and stored at Ϫ80°C until required.
Following thawing, the enzyme showed Ͼ97% oxidase activity as determined by the method described below and constituted XO for the studies described here. XDH was obtained by incubation of purified XO in 50 mM Na/Bicine, pH 8.3, containing 10 mM dithiothreitol for 40 min at 37°C. Dithiothreitol was subsequently removed by gel filtration on Sephadex G25. XDH prepared by these means contained Ͼ90% dehydrogenase form.
The oxidase content of XOR was determined by measuring the rate of oxidation of xanthine to uric acid spectrophotometrically at 295 nm, in a Cary 100 spectrophotometer, using an absorption coefficient of 9.6 mM Ϫ1 cm Ϫ1 (13). Assays were performed at 25.0 Ϯ 0.2°C in air-saturated 50 mM Na/Bicine buffer, pH 8.3, containing 100 M xanthine. The sum of oxidase and dehydrogenase contents was determined as above but in the presence of 0.5 mM NAD.
The enzyme used contained 35-38% functional active sites, as judged by the activity:flavin ratio, assuming fully active enzyme has an activity:flavin ratio of 208 (14). Concentrations of enzyme were determined from the UV-visible spectrum, by using an absorption coefficient of 36 mM subunit Ϫ1 cm Ϫ1 at 450 nm.
Desulfo-XO-Desulfo-XO was prepared essentially according to the procedure described by Massey and Edmondson (15). XO (ϳ1.5 mg ml Ϫ1 ) was incubated in 50 mM potassium phosphate buffer, pH 7.2, containing 10 mM KCN, for 2 h at 37.0 Ϯ 0.2°C. Before assay of xanthine oxidase activity (determined as above), enzyme was gel-filtered on Sephadex G25. This activity was routinely undetectable. Desulfo-enzyme was resulfurated by modification of the method described by Wahl and Rajagopalan (16) as follows. Enzyme (ϳ1.5 mg ml Ϫ1 ) in 50 mM potassium phosphate buffer, pH 7.2, was allowed to equilibrate for 1 h in an anaerobic cabinet (Ͻ3 ppm O 2 ), prior to gel filtration, in the cabinet, using 50 mM MOPS buffer, pH 7.2, itself preequilibrated in the cabinet for at least 24 h. Anaerobic methyl viologen and Na 2 S were then added to final concentrations of 0.2 and 8.4 mM, respectively, before titration with anaerobic 1 mM sodium dithionite until a faint blue color appeared. Enzyme was incubated at 22°C in the cabinet, and at regular intervals samples were gel-filtered as above (also in the cabinet), taken out, and assayed for xanthine oxidase activity (determined as above).
Deflavo-XO-Deflavo-XO was prepared from XO according to the procedure described by Komai et al. (17). Enzyme so prepared showed xanthine oxidase activity (determined as above) of 16.0 Ϯ 1.0 nmol of urate min Ϫ1 mg Ϫ1 . Reconstitution by the addition of FAD (200 M) (17) yielded enzyme with specific activity of 1060 Ϯ 30 nmol of urate min Ϫ1 mg Ϫ1 .
Iodoniumdiphenyl (IDP)-inhibited Enzyme-IDP-inhibited XO was prepared essentially as described by O'Donnell et al. (18) for DPIinhibited enzyme. XO (16 nmol) was incubated with 200 M IDP in 50 mM potassium phosphate buffer, pH 7.2, at 25.0 Ϯ 0.2°C. Aliquots of xanthine (100 nmol) were added, and after 5 min, a sample of the reaction mixture was assayed for xanthine oxidase activity (determined as above). After the addition of 1.5 mol of xanthine, xanthine oxidase activity had fallen to 8.5 Ϯ 4.0 nmol of urate min Ϫ1 mg Ϫ1 , and this enzyme was used for relevant studies.
Reagents-Dithiothreitol was purchased from Alexis Corp. Oxygenfree nitrogen and compressed air were purchased from British Oxygen Corp. Standard NO in nitrogen was obtained from Air Products Corp. All other reagents, unless otherwise stated, were purchased from Sigma.
Kinetics and Stoichiometry-Steady state kinetic studies under anaerobic conditions were carried out in nitrogen-sparged 50 mM potassium phosphate buffer, pH 7.2, at 37.0 Ϯ 0.2°C in an anaerobic cabinet (Belle Technologies) containing Ͻ3 ppm O 2 , using a Hi-Tech SF61 rapid mixing device. With xanthine as reducing substrate, urate was determined as described above for determination of oxidase content of XOR. With NADH as reducing substrate, NADH utilization was followed at 340 nm using an absorption coefficient of 6.22 mM Ϫ1 cm Ϫ1 (19).
Nitric oxide generation was followed by using an ozone chemilumi-nescence assay in a continuous flow apparatus (Sievers NOA 280). The mean concentration of NO in the ozone chamber, derived from readings taken every second, is displayed as ppb, calibrated by using a known concentration of NO in nitrogen (Air Products). The apparatus was modified to allow a constant stream of nitrogen (containing Ͻ3 ppm O 2 ), mixed, as necessary, with defined concentrations of oxygen to flow over the surface of a stirred reaction mixture (1 ml). In the course of equilibration, displayed concentrations of NO rise to a peak value, usually attained approximately 2 min after initiation of the reaction. This value minus background levels is used, together with the measured gas flow, to calculate the molar production of NO. Reactions were carried out at 37.0 Ϯ 0.2°C in nitrogen-sparged 50 mM potassium phosphate buffer, pH 7.2, containing sodium nitrite and either xanthine or NADH and were initiated by the addition of enzyme. pH profile experiments were also carried out in phosphate buffer. The stoichiometry of NO production with respect to urate production or to NADH depletion was calculated as follows. Ratios of the means of replicate determinations of NO production versus the means of urate production (or NADH depletion) were determined for each concentration of variable substrate. These ratios were then treated as individual data for determination of overall means and S.D. values.

Stoichiometry of NO Production and Enzyme Inactivation under Anaerobic Conditions-XO-catalyzed generation of NO
in the presence of nitrite could be directly demonstrated by ozone chemiluminescence (see "Experimental Procedures"). In the presence of NADH, rates of NO production and NADH depletion both followed Michaelis-Menten kinetics ( Fig. 1), showing a stoichiometry (see "Experimental Procedures") of (2.19 Ϯ 0.15):1, consistent with transfer from NADH of pairs of electrons (12), each of which reduces one molecule of nitrite to NO.
With xanthine as reducing substrate, the stoichiometry of NO versus urate production was considerably less than 2:1, approximating rather 1:1 (data not shown). In view of reports that XOR is inactivated by NO (20 -22), particularly in the presence of xanthine (22), we sought evidence of such inactivation under the experimental conditions of NO generation. As can be seen in Fig. 2, in the presence of xanthine the enzyme is indeed progressively inactivated over the time course of NO production. After 2 min, the point at which the NO reading is commonly taken (see "Experimental Procedures"), inactivation has reached approximately 60%, a value consistent with the FIG. 1. Stoichiometry of anaerobic NO production catalyzed by XO in the presence of NADH and nitrite. XO (0.2 M) was incubated in nitrogen-sparged 50 mM potassium phosphate buffer, pH 7.2, at 37 Ϯ 0.2°C with 100 M NADH and varying concentrations of sodium nitrite. NO production (q) was determined by using a chemiluminescence NO detector (see "Experimental Procedures"). NADH depletion (E) was followed by monitoring absorbance at 340 nm (see "Experimental Procedures") in an anaerobic cabinet (Ͻ3 ppm O 2 ). Michaelis-Menten curves were fitted to the data. Stoichiometry of NO production versus NADH depletion was calculated from the respective V max values. Values are means Ϯ S.D. (n ϭ 3). Where error bars are not shown, their width is less than that of the symbol. approach to 1:1 stoichiometry discussed above.
Ichimori et al. (22) demonstrated inactivation of conventional aerobic XO activity by exogenous NO and attributed this inactivation to generation of desulfo-XO, in which MoϭS is replaced by MoϭO. In the present context, it was clearly of interest to determine whether inactivation observed in the course of NO generation results from desulfuration of XO. A sample of XO (initial activity of 4500 Ϯ 120 nmol of urate min Ϫ1 mg Ϫ1 ) was incubated anaerobically for 20 min in the presence of nitrite and xanthine, as outlined in the legend to Fig. 2. The reaction mixture was gel-filtered (see "Experimental Procedures") to yield 90% inactivated enzyme (450 Ϯ 30 nmol of urate min Ϫ1 mg Ϫ1 ), which could, indeed, be reactivated on sulfuration (see "Experimental Procedures") (final activity of 5430 Ϯ 290 nmol of urate min Ϫ1 mg Ϫ1 ). This is consistent with the conclusion that NO, whether exogenously added (22) or endogenously generated in the presence of xanthine, inactivates XO by transforming it to the desulfo-form. In contrast to the situation observed with xanthine, rates of enzyme inactivation were very much lower in the presence of NADH, leading to less than 5% inactivation after 2 min (Fig. 2).
Steady-state Kinetics of the Reduction of Nitrite by NADH or Xanthine Catalyzed by XO or XDH under Anaerobic Conditions-The kinetics of the reduction of nitrite by NADH catalyzed by XO were followed by observing the rate of NADH depletion at 340 nm. Treating either nitrite or NADH as the varied substrate, the variation of rate with substrate concentration follows Michaelis-Menten kinetics. With xanthine as reducing substrate, the reaction rates were monitored by following urate accumulation at 295 nm. When nitrite is treated as the variable substrate concentration, Michaelis-Menten behavior was again found. However, the variation of rate with xanthine concentration fitted well to Equation 1 for substrate inhibition (Fig. 3), where K m is the Michaelis constant and K i is Substrate inhibition was relieved by increasing nitrite concentration; over the range of 20 -120 mM nitrite, K i varied from 30 to 230 M.
Using XDH, with xanthine as the reducing substrate, the kinetic behavior was identical to that using XO (Fig. 4). However, with NADH as the reducing substrate, Michaelis-Menten kinetics were followed (as with the oxidase), but the parameters were significantly different from those obtained using the oxidase form (Table I).
Because of the substrate inhibition by xanthine, it was not possible to draw any mechanistic conclusions from the values of the apparent steady-state kinetic parameters. However, with NADH as reducing substrate, V max app /V m app for XO was constant throughout the range of substrates employed, consistent with a ping-pong mechanism. Interestingly, with XDH, V max app /V m app increased with increasing substrate concentration for both NADH and nitrite.
All of the above kinetic experiments were carried out with enzyme of 35-38% functionality (see "Experimental Procedures"). NADH-nitrite reductase activity of desulfo-XO (see "Experimental Procedures") was low but significant, contributing rates of less than 3% in those of the standard enzyme preparation.
pH Profiles of Anaerobic NO Production, NADH Oxidation, and Xanthine Oxidation-The pH profiles for NO production in the presence of nitrite and NADH or xanthine catalyzed by XO under anaerobic conditions are shown in Fig. 5. Essentially identical profiles were obtained when NADH depletion or urate production, respectively, were monitored in the presence of nitrite (data not shown).
Involvement of Molybdenum and FAD Sites in the Catalytic Activity of XO-Direct involvement of the molybdenum site in nitrite reduction was demonstrated by inhibition of anaerobic NADH oxidation in the presence of alloxanthine. The well known inhibitor, allopurinol, exerts its effect by acting as a reducing substrate for XOR, thereby being itself oxidized to alloxanthine (also known as oxypurinol), which then binds very tightly to the reduced form of the enzyme (14,23). Consistent with this is the observation (24) that alloxanthine acts as an uncompetitive inhibitor of xanthine, which binds to the completely oxidized form of the enzyme. In the course of its anaerobic reduction, nitrite might be expected to bind also to the reduced form of molybdenum; that this is indeed the case is indicated by the demonstration that alloxanthine inhibited NADH depletion competitively with nitrite (Fig. 6).
It was shown, furthermore, that reduction of nitrite to NO is dependent on the sulfo-form of XO. Conversion of XO to its desulfo-form (see "Experimental Procedures") resulted in apparent loss of its ability to catalyze either aerobic xanthine oxidation or anaerobic NO production in the presence of NADH and nitrite. Both of these activities were recovered, with essentially identical time courses, upon resulfuration (see "Experimental Procedures") (Fig. 7).
Urate production in the presence of xanthine and nitrite catalyzed by XO anaerobically was essentially identical to that catalyzed aerobically by IDP-inhibited XO (Fig. 8), consistent with a lack of involvement of the FAD site in nitrite reduction. With deflavo-enzyme, under aerobic conditions, rates of urate production were somewhat lower (Fig. 8), probably reflecting instability of this form of the enzyme (see "Discussion").
In the presence of NADH and nitrite, neither deflavo-nor IDP-inhibited enzyme (see "Experimental Procedures") showed detectable NO production or NADH depletion (data not shown); a result that is unsurprising, in view of the well established direct interaction of NADH with the FAD site of XO (17).
XO-catalyzed Generation of NO in the Presence of Oxygen-In the presence of NADH and nitrite, XO-catalyzed NO production apparently decreased as oxygen concentration increased, remaining detectable only at higher concentrations of nitrite ( Fig. 9). This phenomenon can be largely attributed to rapid reaction of NO with superoxide to form peroxynitrite (27) as evidenced by the effects of superoxide dismutase. In the presence of 200 units of superoxide dismutase, the rate profile of NO production versus nitrite concentration was identical to that obtained anaerobically in the absence of superoxide dismutase (Fig. 10). DISCUSSION It has been known for over 75 years that XOR has the ability to catalyze the reduction of nitrate to nitrite under anaerobic conditions (28 -31). However, XO-catalyzed further reduction of nitrite to NO has only very recently been reported. We briefly described this phenomenon in the presence of NADH under hypoxic conditions (11); a subsequent communication (32) reported NO generation in both hypoxia and ambient air. The present study examines the kinetics and mechanism of nitrite reduction catalyzed by XOR with both NADH and xanthine as reducing substrates and addresses the effects of oxygen.
Anaerobic NO production catalyzed by XO in the presence of nitrite and NADH could be directly demonstrated by ozone chemiluminescence using a chemiluminescence NO detector. Comparison with rates of NADH utilization showed a stoichiometric ratio of approximately 2:1 versus NO production. This result is predicted on the basis of a redox reaction comprising a two-electron oxidation of NADH to NAD ϩ and a one-electron reduction of nitrite to NO. Involving as it does both spectrophotometric and chemiluminescence NO detector data, this finding serves to validate the latter. Corresponding experiments in the presence of xanthine, rather than NADH, showed relatively less NO production. This initially surprising finding was satisfactorily explained in terms of NO-induced inactivation of XO under conditions of the assay. While it is not clear why such inactivation should be markedly greater in the presence of xanthine rather than NADH, this may reflect the extent to which the enzyme is present in reduced states (22). Our findings complement those recently reported by Ichimori et al. (22) involving inactivation of xanthine-reduced XO by exogenous NO. These authors demonstrated that enzyme inactivation in their system resulted from sulfo-desulfo conversion. We now show not only that XO can be inactivated by endogenously generated NO, but that in this case also the mechanism involves MoϭS to MoϭO conversion.
From the above discussion, it is clear that kinetic experiments in the chemiluminescence NO detector will be more or less subject to complications resulting from NO-induced inactivation of XO. This is because, as a consequence of phase  equilibration and gas flow factors, data in our system are first obtained approximately 2 min after initiation of the reaction. We accordingly preferred spectrophotometric monitoring of NADH depletion or of urate production for determination of steady-state kinetic parameters. Determination of the steady state rate can be made earlier in the time course, where product inactivation is less of a consideration, and this allowed determination of the kinetic parameters shown in Table I.
The substrate steady-state kinetics are not intended to provide mechanistic information as such but rather serve to supply operational parameters with which to estimate the rates of XO-catalyzed NO production that might be expected to occur in vivo (see below). However, one point deserves comment. XDH is seen to catalyze nitrite reduction in the presence of NADH almost 50 times more efficiently than does XO (as judged by V max /K m nitrite ; Table I). With NADH as reducing substrate, the dehydrogenase is more effective than the oxidase form also when oxygen is the oxidizing substrate (12,33). This observation is particularly relevant in vivo, where, at least intracellularly, XDH is the predominant form (34).
Concerning involvement of the various redox sites of XO in nitrite reduction and NO production, the evidence seems to be clear. Inhibition of NADH oxidation by alloxanthine competitively with respect to nitrite implicates molybdenum as the site of nitrite reduction and is fully consistent with interaction of nitrite with the reduced form of XO. We have shown, moreover, that NO production in the presence of nitrite and NADH depends upon the MoϭS group, activity being largely lost on desulfuration to MoϭO and regained concurrently with xanthine oxidase activity on resulfuration. While it is tempting to infer that this dependence implies direct involvement of molybdenum-bound sulfur in the mechanism of nitrite reduction, it could also reflect the midpoint reduction potentials of molybdenum. Porras and Palmer (35) report midpoint potentials, at pH 7.2 and room temperature, for bovine milk XO of Mo(VI)/ Mo(V) (Ϫ300 mV) and Mo(V)/Mo(IV) (Ϫ300 mV). The corresponding values for desulfo-enzyme are quoted as Ϫ335 and Ϫ309 mV. In the context of midpoint potentials for FAD/FADH ⅐ and FADH ⅐ /FADH 2 of Ϫ235 and Ϫ215 mV, respectively (35), transfer of electrons from NADH to molybdenum via the FAD site will be less favorable in the case of desulfo-enzyme.
Possible influence of the FAD center of XO on xanthine oxidation in the presence of nitrite was initially examined by comparing XO and deflavo-XO. While rates of urate production were similar, those with deflavo-enzyme were slightly lower, an observation that might be attributable to instability of this enzyme form, which is prone to aggregation and precipitation (17). Accordingly, use was made of IDP, which, like diphenylene iodonium (18), was shown to inactivate the FAD site of XO, as judged by loss of xanthine oxidase activity (see "Experimental Procedures"). In this case, rates of urate production catalyzed by XO and flavin-inactivated enzyme were essentially identical, confirming the lack of influence of the FAD site in this reaction. In contrast, NADH depletion in the presence of nitrite was totally inhibited by IDP.
The above experiments concerning the mechanisms of NO production were carried out with XO. The similarity in kinetic behavior of XO and XDH in catalyzing the reaction between nitrite and xanthine (Fig. 4), suggests that conclusions drawn on the basis of XO may also be applicable to XDH.
In the presence of oxygen, the situation is complicated by concomitant generation of NO and superoxide anion. As oxygen tensions rise, levels of NO production were seen to fall, becoming undetectable at lower concentrations of nitrite. These effects of oxygen could be completely reversed by the addition of superoxide dismutase, indicating that they can be entirely accounted for by rapid reaction of NO with superoxide, presumably yielding peroxynitrite (36). Rates of superoxide production in the presence of NADH were shown to correspond closely to the observed decrease in detected NO as oxygen tensions rise (data not shown). An obvious issue concerns the physiological role of XORcatalyzed generation of NO. Indeed, the enzyme's function in vivo has long been debated, particularly as its rather specialized location in endothelial, epithelial, and connective tissue cells implies more than a housekeeping role in purine catabolism (37). In the context of vascular endothelium, the capacity of XOR to produce superoxide and hydrogen peroxide has led to its implication in the inflammatory process (38,39). More recently, attention has focused on interactions of superoxide with endogenous NO and their possible deleterious consequences in atherosclerosis (40 -42). Our finding that XOR itself can also catalyze NO production adds a new dimension to these discussions. XOR can be seen as complementary to NO synthase, in that, unlike the latter (10), it is capable of producing NO under anoxic conditions. XOR could, accordingly, promote NO-induced vasodilation in ischemia, when NO synthase activity is low. NO synthase is induced under such conditions (43), prior to assuming the burden of NO production as oxygen tension rises.
In view of our present observations, the subcellular localization of the enzyme assumes particular importance. Until relatively recently, XOR has been generally assumed to be exclusively cytoplasmic (44), where XDH predominates (34). In this context, XDH would certainly be more efficient than XO as a catalyst of NO production, at least with NADH as reducing substrate (Table I). Of late, attention has been directed to extracellular endothelial XOR, arising either endogenously (45,46) or via the circulation. In the presence of serum, the oxidase form, XO, predominates (34). Levels of circulating enzyme are normally low (47) but can rise dramatically in a variety of inflammatory disease states, particularly those involving liver damage (48 -50). XO has been shown to bind to vascular endothelium (41) by mechanisms that probably involve specific interactions with glycosylaminoglycans (51,52), and the enzyme may be expected to be greatly concentrated at the luminal surface. In this context, superoxide dismutase assumes relevance. Mammalian extracellular superoxide dismutase C has strong affinity for heparin (53), and its immediate colocalization with XO could conceivably result in similar rates of NO production in the presence or absence of oxygen (Fig. 10).
In the absence of significant interactions with superoxide dismutase, XO is a potential source of peroxynitrite, with its capacity for endothelial injury (54). Indeed, levels of both XO (55) and anti-XO antibodies (56) have been shown to be elevated in plasma of atherosclerosis patients, and the enzyme has been detected in their vessel walls and plaques (57,58).
Maximal rates of XOR-catalyzed nitrite reduction (Table I) are significant in the context of estimated rates of NO production in the vasculature, which, according to Beckman et al. (59), range up to 8 M min Ϫ1 . However, K m values for nitrite are in the millimolar range (Table I), considerably higher than the micromolar levels to be found in plasma (60). This prompts consideration of the enzyme's role in other physiological environments. In particular, XO could fulfill a bactericidal function in the digestive tract, throughout which it is present in epithelial cells (61). By way of the milk fat globule membrane (62), it also occurs in the neonatal gut. While levels of nitrite in the digestive tract are generally in the micromolar range (63,64), they are potentially much higher in the microenvironment of enteric bacteria, which, at least in anaerobic culture, can excrete millimolar levels of nitrite (65,66), derived from dissimilatory nitrate reductase (67). Under such conditions and in the presence of micromolar levels of purines, XOR could generate NO, a potent bactericidal agent (68 -70), at rates of approximately 1000 nmol min Ϫ1 mg Ϫ1 (Table I). Reported concentrations of XOR in the intestine depend very much on the investigator and the species but, conservatively, are on the order of 10 Ϫ4 mg mg Ϫ1 tissue protein (71). This would yield rates of NO production of about 100 pmol min Ϫ1 g Ϫ1 tissue, a value that can be seen as significant in the context of rates of NO generation by activated macrophages (1 nmol min Ϫ1 mg Ϫ1 macrophage protein (72)). It is noteworthy that the pH optimum for anaerobic XO-catalyzed NO production in the presence of xanthine is pH 6 or less (Fig. 5), much lower than the value (pH 8.8) observed for aerobic xanthine oxidase activity (73) and more appropriate to a role in the digestive tract.
As oxygen tensions rise, so does the potential for the generation of peroxynitrite, also a powerful bactericide (74). It is an attractive concept that, by way of nitrite excretion, enteric bacteria might initiate their own destruction, a concept strengthened by the known affinity of XOR for acidic polysaccharides (51,52), such as occur in many bacterial capsules (75). While much of the above argument is clearly speculative, it is supported by recent ultrastructural studies (61) in the rat digestive tract, in which some bacteria in the cornified layer of epithelial cells showed signs of destruction and/or death and were surrounded by XOR.
In summary, we have shown that, in the presence of NADH or xanthine as reducing substrate, XOR can catalyze the reduction of nitrite to NO. Nitrite reduction takes place at the molybdenum site, is dependent upon the presence of the MoϭS grouping, and, in the presence of xanthine, is unaffected by loss of activity at the FAD site. As oxygen tension increases, the yield of NO decreases as a consequence of its reaction with XOR-generated superoxide. XOR-catalyzed production of NO is potentially important in signaling events in vascular endothelium and as a bactericidal system in the digestive tract.