Reactivity of Tetrahydrobiopterin Bound to Nitric-oxide Synthase*

Levels of tetrahydrobiopterin (BH4) bound to nitric-oxide synthase (NOS) were examined during multiple turnovers of the enzyme in the presence of an NADPH-regenerating system. Our findings show that NOS-bound BH4 does not remain in a static state but undergoes redox reactions. Under these experimental conditions, the redox state of BH4 was determined by the balance between calcium/calmodulin (Ca2+/CaM)-dependent oxidation of BH4 mediated by the uncoupled formation of superoxide/hydrogen peroxide on the one hand and by reductive regeneration of BH4 on the other hand. BH4oxidation was appreciably increased in the presence of arginine. Levels of NOS-bound BH4 were also examined under single turnover conditions in the absence of an NADPH-regenerating system and in the presence of added superoxide dismutase and catalase to suppress the accumulation of superoxide and hydrogen peroxide. BH4oxidation was again dependent on Ca2+/CaM. The insensitivity to superoxide dismutase and catalase suggested that the single turnover oxidation of BH4 did not proceed through superoxide/peroxide, although the involvement of these oxidants could not be definitively excluded. The amount of BH4 oxidized was highest in the presence of arginine, and this oxidation significantly exceeded that in the presence ofN G-hydroxy-l-arginine. The findings that single turnover oxidation of BH4 is stimulated by arginine in the presence of Ca2+/CaM and that BH4 is regenerated are consistent with a role for the pterin as an electron donor in product formation; this role remains to be defined.

Nitric-oxide synthase (NOS) 1 catalyzes the overall conversion of arginine to nitric oxide (NO) and citrulline. The reaction proceeds in two steps; arginine is first oxygenated to the intermediate N G -hydroxy-L-arginine (NHA), which is then oxygenated to the products NO and citrulline (reviewed in Ref. 1). The enzyme has attracted much interest because of its involvement in a variety of physiological processes such as vasodilation, cytotoxic activity, and neurotransmission (2,3). NOS is unique as a heme oxygenase in requiring BH 4 , and elucidation of the role of BH 4 is essential in understanding how this enzyme functions.
It is well established that BH 4 has pronounced effects on the structure of NOS. These effects include the ability to shift the heme iron to its high spin state, promotion of arginine binding, and stabilization of the active dimeric form of the enzyme (reviewed in Ref. 4), as well as stabilization of the ferrous heme iron coordination structure (5). Furthermore, binding of the first molecule of BH 4 to an NOS dimer has an anticooperative effect on the binding of the second molecule of BH 4 (6). Studies of the effects of pterin analogues on NOS provide graphic evidence that these structural effects cannot fully explain the requirement for BH 4 . Whereas the structural effects of BH 4 are mimicked by a series of pterin analogues independent of their oxidation state, pterins must be in the tetrahydro state in order to support NO synthesis (5,7,8). This requirement for a tetrahydropterin suggests an additional, redox, role of BH 4 .
A number of recent reports support a redox role of BH 4 in which the pterin provides electrons for the oxygenation reactions catalyzed by NOS. Studies with recombinant BH 4 -free NOS (9) have implicated BH 4 as acting as a stoichiometric reactant in the first oxygenation step (arginine to NHA) of NO synthesis. The role of BH 4 in this reaction is envisioned by Rusche et al. (9) as being analogous to the well established role of BH 4 in the aromatic amino acid hydroxylase systems (10), except that BH 4 remains tightly bound to NOS throughout the catalytic cycle (11). Further support for this proposal is provided by the finding that non-heme iron, which is an integral component of the aromatic amino acid hydroxylases (10), is a stoichiometric component of NOS and enhances its activity (12). Again by analogy with the hydroxylation of the aromatic amino acids (10), the oxidation of BH 4 by NOS would generate equivalent amounts of NHA and quinonoid dihydrobiopterin (qBH 2 ), and the latter compound would need to be recycled to BH 4 by a dihydropterin reductase-like activity in order to maintain the catalytic function of BH 4 . In these same studies, Rusche et al. (9) further proposed a role for BH 4 in the second oxygenation step (NHA to citrulline and NO). It was suggested (9) that BH 4 may be acting as an electron donor in this step or in an indirect fashion by affecting, for example, the spin equilibrium and reduction potential of the heme.
Further support for a redox role of BH 4 is provided by the single turnover studies of NOS (13). It was suggested that BH 4 plays a critical role in the first oxygenation step by supplying an electron for reducing the ferric heme superoxide species to the ferric heme peroxo species. Based largely on studies of cytochrome P-450, the latter species, or a closely related form, is believed to carry out the oxygenation of arginine (1,14). Since only one electron is required for reduction of the ferric heme superoxide species to the ferric heme peroxo species by BH 4 , the oxidized pterin product was formally represented as the trihydrobiopterin radical (BH 3 ⅐ ), although the possibility of qBH 2 being the product could not be excluded. Based on crystallographic studies and the unexpected finding that arginine recognizes the BH 4 -binding site, Raman et al. (15) suggest that this site may stabilize the positively charged pterin cation radical, thereby supporting a scheme in which electrons are shuttled between the protonated forms of BH 4 and BH 3 ⅐ , rather than between BH 4 and qBH 2 . It has recently been reported (16) that the 5-methyl analogue of BH 4 is functionally active with NOS but does not react with oxygen or stimulate phenylalanine hydroxylase. It is suggested (16) on the basis of these studies that the redox chemistry of BH 4 is not the same in NOS as it is in the aromatic amino acid hydroxylases, but it could not be excluded that BH 4 interacts with the heme or another metal group of NOS through an as yet unrecognized redox chemistry.
Since catalytic amounts of BH 4 are sufficient for NOS activity (11), a minimum requirement for any role of BH 4 as a reductant is that the oxidized pterin products be regenerated with each catalytic cycle by a dihydropterin reductase-like activity intrinsic to the enzyme. No recycling of exogenous BH 4 was detected during NOS turnover (11), but this study did not exclude the possibility of recycling of BH 4 tightly bound to the enzyme (17). The reduction of added qBH 2 to BH 4 by NOS, probably by the flavoprotein "diaphorase" activity of the enzyme, has been demonstrated (18), but the significance of this finding to the essential role of BH4 was uncertain. Furthermore, under the standard assay conditions of this latter study, any regenerated BH 4 that remains firmly bound to NOS would not have been detected, since any such regeneration is limited to the number of BH 4 -binding sites; these sites are negligible relative to the observed amounts of qBH 2 reduced.
The present work provides experimental support for a redox role of BH 4 by showing that NOS-bound BH 4 can cycle with its oxidized form(s). Although details of the oxidation of BH 4 remain to be clarified, this work is an important step toward defining the mechanistic significance of this reaction in NO formation. The regeneration of BH 4 from its oxidized products is probably important in maintaining the pterin in its active, tetrahydro, state.

EXPERIMENTAL PROCEDURES
Chemicals-The following materials were obtained from the sources shown in parentheses: N G -nitro-L-arginine (NNA) and L-␣-lysolecithin (Sigma), CaM from bovine brain (Calbiochem), BH 4 (B. Schircks Laboratories, Jona, Switzerland), L-[2,6-3 H]phenylalanine, L- [2,3,4,5-3 H] arginine, and L-[U-14 C]tyrosine (Amersham Pharmacia Biotech), and NHA (Alexis Corp., San Diego, CA), Enzymes-Beef liver catalase was obtained from Roche Molecular Biochemicals. Superoxide dismutase (SOD) from bovine erythrocytes and glucose-6-phosphate dehydrogenase from Leuconostoc mesenteroides were obtained from Sigma. Recombinant rat phenylalanine hydroxylase was prepared by minor modification (19) of the procedure of Citron et al. (20) and had a specific activity of approximately 2 mol/min/mg. Rat brain NOS was purified from transfected human kidney 293 cells (21). NOS isolated in this way showed a single band on an SDS-polyacrylamide gel and typically had a specific activity of 1.0 -1.6 mol/min/mg as determined by citrulline formation over a 10-min incubation at 37°C with the assay of Bredt and Snyder (22). Molar concentrations of NOS are reported for the 160-kDa subunit. The BH 4 content of the purified NOS was typically 0.5 mol/mol subunit as determined by high performance liquid chromatography (HPLC) analysis of oxidized samples (23). Protein was determined by the bicinchoninic acid method (Pierce manual 23225X) with the use of a bovine serum albumin standard. BH 4 Levels during Turnover of NOS-Reaction mixtures used for these studies contained 1 M NADPH and an NADPH-regenerating system. This low concentration of NADPH was used to prevent significant non-enzymic regeneration of BH 4 that occurs at higher concentrations of NADPH (18,24). We have previously shown (18) that NOS has normal activity using this low concentration of NADPH. The standard reaction mixture contained 2 M NOS, 1 M NADPH, 1.5 mM glucose 6-phosphate, 32 g/ml glucose-6-phosphate dehydrogenase, 35 mM sodium Bicine (pH 7.6), 0.9 mM CaCl 2 , and 0.6 mM EDTA. The following components were also added as indicated: 4 M CaM, 800 g/ml SOD, 16 g/ml catalase, 50 M arginine or NHA, and 200 M NNA. In some experiments, reaction mixtures were supplemented with a concentration of BH 4 equal to one-half the concentration of NOS monomer. In this way, native NOS containing approximately 0.5 BH 4 / NOS monomer was reconstituted to approach a value of 1/NOS monomer. 2 Incubations were at 15°C. Samples of 25 l were taken to establish the initial levels of BH 4 . Immediately thereafter, CaM was added, where appropriate, in order to start the reaction, and samples of 25 l were taken at fixed times for BH 4 analysis. As illustrated under "Results," negligible oxidation of BH 4 occurred in the absence of CaM.
BH 4 Levels under Single Turnover Conditions-Two different procedures were followed depending on whether only BH 4 levels were determined or whether BH 4 levels together with NHA were determined.
The standard reaction mixture in which only BH 4 levels were determined contained 1.7 M NOS, 0.9 mM CaCl 2 , 0.6 mM EDTA, 400 units/ml SOD, 8 g/ml catalase, and 50 mM sodium Hepes (pH 7.4). Where indicated, NOS was reconstituted with BH 4 as described above to contain approximately 1 BH 4 /NOS subunit. Immediately prior to the start of the reaction, NADPH was added to the standard reaction mixture (1 NADPH per NOS subunit), and the mixture preincubated for 2.25 min. This procedure, which is fully described in a separate section below, generated a maximum level of NOS-bound flavin semiquinone. Separate experiments showed that BH 4 was stable during this preincubation period. After the preincubation, samples were taken to establish initial BH 4 levels. Immediately thereafter, the indicated additions In the other tube containing the remainder of the reaction mixture, unlabeled arginine was added to achieve the same final concentration of 18.4 M arginine; this tube was used for determination of BH 4 levels. Immediately after samples were taken to establish initial levels of BH 4 and NHA, CaM was added to a final concentration of 4.7 M to each tube, and 25-l samples were removed at the intervals shown above for determination of NHA formation and BH 4 . All additions were made in very small volumes so as not to significantly change the specified concentrations of reaction components.
All incubations were at 15°C. Determination of BH 4 Oxidation-BH 4 was determined by coupling this pterin to the phenylalanine hydroxylase-catalyzed conversion of an equivalent amount of [ 3 H]phenylalanine to [ 3 H]tyrosine. Reproducibility of the original method (18) was improved by the modifications described below. To the 25-l samples taken from reaction mixtures, 5 l of 1.5 M HClO 4 was added immediately to stop the reaction, release BH 4 from NOS, and stabilize the pterin in its reduced form. Within 4 min after adding the acid, 5 l KHCO 3 was added to partially neutralize the mixture to approximately pH 2.5-3.0. The precipitate of KClO 4 was removed by a 10-s centrifugation, and 30 l of the supernatant solution was added to 9.6 l of a solution containing the remaining reaction components. The final reaction mixture contained, in addition to the BH 4 to be assayed, 1 g of phenylalanine hydroxylase, 4. amounts of NOS either unsupplemented or supplemented with BH 4 . The BH 4 content of NOS samples used for calibration was determined independently by the method of Fukushima and Nixon (23).
Single Turnover Oxygenation of Arginine to NHA-The formation of NHA in the absence of added NADPH was measured by the incorporation of 3

H from L-[ 3 H]arginine into NHA (25).
Generation of FSQ by Addition of NADPH-The air-stable NOSbound flavin semiquinone free radical, FMNH ⅐ (FSQ) was generated immediately before use by preincubation of NOS for 2.25 min in air with NADPH (26). One mol of NADPH (two electron eq) per mol of NOS subunit was required to obtain the maximum level of FSQ (reduced at the level of one electron eq), presumably to compensate for reaction of reduced flavin species with oxygen during the titration. Total reduced flavins were measured by the decrease in absorbance at 458 nm and FSQ by the increase in absorbance at 590 nm (27). All absorbance values were determined with a Hewlett-Packard 8453 diode array spectrophotometer. Addition of NADPH typically resulted in an instantaneous decrease in absorbance at 458 nm, followed by a slow increase in absorbance (first order rate constant of 2.6 min Ϫ1 ) over the next 2 min to approach a constant value approximately midway between the minimal and initial absorbance values. Parallel measurements at 590 nm showed a slow increase in absorbance throughout the incubation period (first order rate constant of 3.4 min Ϫ1 ) to approach a constant value. The combined results are interpreted as reflecting a rapid reduction of the flavins (presumably by formation of FADH 2 and subsequently FMNH 2 ) followed by formation of a stable FMNH ⅐ species resulting from the partial oxidation of FMNH 2 by oxygen. These spectral data are consistent with values 0.4 -0.6 FMNH ⅐ per NOS subunit estimated from the amount of NHA formed per NOS subunit (21). Gachhui et al. (27) similarly observed approximately 0.5 FMNH ⅐ per reductase domain of NOS resulting from titration of this domain with NADPH in air. 4 were examined under three conditions of continuous turnover. In these experiments, continuous turnover of neuronal NOS (nNOS) was achieved in the presence of a low concentration (1 M) of NADPH and an NADPH-regenerating system. Clearly, in the absence of arginine, no NO is formed. Under this fully uncoupled condition, electrons are transferred from NADPH to oxygen with the exclusive formation of superoxide and peroxide (6, 28 -30). In the presence of arginine, but no added BH 4 , nNOS is partially coupled, resulting in the formation of both NO and superoxide/ peroxide (28,30). Maximum coupling of NADPH oxidation to NO formation is approached in the presence of saturating concentrations of both arginine and BH 4 (6,28). However, complete coupling of electron transfer from NADPH to NO formation in the presence of arginine and BH 4 would not be expected under our conditions. This is because it appears that some superoxide is formed even in the presence of saturating arginine and BH 4 (31), and the addition of BH 4 in our studies resulted in incomplete saturation 2 of NOS with BH 4 . BH 4 Levels during Fully Uncoupled Turnover of NOS- Fig. 1 illustrates representative results for the levels of NOS-bound BH 4 in the absence of added BH 4 and arginine. In the absence of CaM, the level of BH 4 remains constant (curve 1). Addition of CaM results in approximately 40% oxidation of BH 4 (curve 2). CaM-dependent oxidation of BH 4 is largely prevented by the additional presence of SOD and catalase (curve 3). Furthermore, the CaM-dependent oxidation of BH 4 proceeding over an 80-s period (curve 4, open circles) is partially reversed by addition of SOD and catalase, in combination with NNA (curve 4, closed circles). NNA is a potent inhibitor of heme-catalyzed reduction of oxygen by neuronal NOS and the consequent formation of superoxide/hydrogen peroxide and nitric oxide (28,29,32,33). Regeneration of BH 4 did not occur in the absence of NNA. These combined findings strongly suggest that BH 4 is oxidized by superoxide and/or peroxide generated by NOS during the uncoupled reduction of oxygen by NADPH and that regeneration of BH 4 (possibly from qBH 2 ) proceeds when the concentration of superoxide/peroxide is decreased to a low level in the combined presence of SOD, catalase, and NNA.

Levels of NOS-bound BH
Effects of Supplementation with Arginine on BH 4 Levels during Turnover of NOS- Fig. 2 compares the BH 4 levels of NOS under partially coupled conditions (presence of supplemental arginine) with those described above under fully uncoupled conditions (Fig. 1). Curves 1 and 2 of Fig. 2 confirm the requirement of CaM for BH 4 oxidation. BH 4 levels are largely stabilized by addition of NNA at the beginning of the incubation (curve 3). In the presence of CaM, addition of arginine markedly increased the oxidation of BH 4 from 40% (curve 2) to 80 -90% (curve 4). Oxidation under the latter conditions (arginine and CaM) was greatly attenuated by addition of SOD plus catalase (curve 5), and partially reversed by addition of SOD, catalase, and NNA (curve 6, closed circles versus open circles), indicating that sufficient superoxide/peroxide is generated even in the presence of supplemental arginine to cause extensive oxidation of BH 4 .
Effects of Supplementation with Arginine and BH 4 on BH4 Levels during Turnover of NOS-Since maximum coupling is approached in the combined presence of arginine and BH 4 (6,28), it was therefore of interest to determine the effect of adding BH 4 to the arginine-supplemented reaction mix used in the experiment described in Fig. 2. The results shown in Fig. 3 confirm that little or no oxidation of NOS-bound BH 4 occurs in the absence of CaM (curve 1) and further show that similar stability of BH 4 is observed in the presence of supplemental BH 4 (curve 2). CaM-dependent oxidation is increased marginally by supplementary BH 4 (curve 4 versus curve 3). In the presence of arginine, supplementary BH 4 initially protects BH 4 against oxidation (curve 6 versus curve 5), but this protective effect is lost as the incubation progresses. These results show that even in the presence of arginine and BH 4 , conditions that favor the coupled formation of NO, BH 4 is oxidized ultimately to the same low level as observed with arginine alone.

BH 4 Levels during Single Turnover of NOS-
In the studies reported above, reaction mixtures contained an NADPH-regenerating system that provides a source of electrons for both the superoxide/peroxide-mediated oxidation of BH 4 and, as demonstrated in the presence of SOD plus catalase plus NNA, the regeneration of BH 4 from its oxidized products. The effects noted above therefore represent steady-state changes determined by the amount of oxidation of BH 4 on the one hand and regeneration of BH 4 on the other. Clearly, in order to detect any oxidation of BH 4 that is stoichiometrically coupled with the overall oxygenation of arginine to NO, it is necessary to prevent regeneration of BH 4 from its oxidized products. We therefore examined BH 4 levels of NOS during the single turnover of arginine to NHA and of NHA to citrulline in the absence of an NADPH-regenerating system; SOD/catalase were included in the reaction mixtures to minimize any oxidation of BH 4 by superoxide/peroxide. Fig. 4 illustrates results of experiments to determine the effects of arginine and NHA on BH 4 oxidation. The possible trace amount of BH 4 oxidation in the absence of CaM (curve 1) could reflect the inherent instability of BH 4 under these conditions. Addition of CaM (curve 2) or CaM plus NHA (curve 3) caused a small increase in the oxidation. Maximal BH 4 oxidation was observed in the presence of CaM plus arginine (curve 4). Higher levels of SOD/catalase added to the reaction mixture of curve 2 resulted in similar results (data not shown).
The relationship between BH 4 oxidation and oxygenation of arginine to NHA under single turnover conditions is illustrated in Fig. 5. NHA, the immediate product of arginine oxidation, was the dominant product; citrulline formation was negligible (data not shown). The ratio of BH 4 oxidized to NHA formed was not constant. It may be estimated from the results of Fig. 5 (inset) that this ratio increased with time from 1.0 at 30 s to 1.7 at 180 s. In this experiment, NOS was not supplemented with BH 4 . Variable ratios were also observed in a similar experiment in which NOS was supplemented with 1 mol of BH 4 per mol of NOS subunit. Under this condition the ratio of BH 4 oxidized to NHA formed was 3.1 at 30 s and progressively decreased to 1.8 at 180 s (data not shown).

BH 4 Oxidation during Continuous
Turnover-Under continuous turnover conditions in the presence of an NADPH-regenerating system (absence of arginine and BH 4 ), oxidation of NOS-bound BH 4 is dependent upon the presence of Ca 2ϩ /CaM and is prevented by SOD/catalase and by NNA. Furthermore, BH 4 can be regenerated upon addition of SOD/catalase and NNA. These observations are consistent with the oxidation of NOS-bound BH 4 being mediated by superoxide/peroxide, formed during the uncoupled reduction of oxygen by hememediated transfer of electrons from NADPH (6, 28 -30). The oxidation state of NOS-bound BH 4 appears to be determined by a balance between BH 4 oxidation by superoxide/peroxide on the one hand, and reductive regeneration of BH 4 on the other.
For nNOS, added arginine and BH 4 are known to increase the coupling of oxygen reduction to NO formation at the expense of superoxide/peroxide formation (28,30). We therefore expected that the presence of arginine, especially in the presence of added BH 4 , would decrease the amount of BH 4 oxidation. Instead, arginine appreciably increased BH 4 oxidation (Fig. 2, curve 4 versus curve 2; Fig. 3 curve 5 versus curve 3). The combined presence of arginine and added BH 4 initially delayed BH 4 oxidation (Fig. 3, curve 6), but ultimately BH 4 was oxidized to the same low level as observed in the presence of arginine alone. A likely interpretation of these effects is that arginine facilitates the reduction of heme iron (34). Bearing in mind that some superoxide is produced by nNOS even at sat-

FIG. 3. Effects of supplemental BH 4 and arginine on BH 4 levels during NOS turnover.
Native NOS containing 0.5 BH 4 /NOS subunit was reconstituted with supplemental BH 4 to a value approaching 1 BH 4 /NOS subunit (see "Experimental Procedures"). BH 4 levels are expressed as percentages of the initial total BH 4 content. Thus, although the amount of initial total BH 4 in the BH 4 -supplemented reaction mixtures is twice that of the unsupplemented reaction mixtures, there is also almost twice as much NOS containing bound BH 4 . This representation allows direct approximation of relative BH 4 levels based on the corresponding amount of BH 4 -bound NOS. Conditions are the same as those described in Fig. 1 except for the absence or presence, respectively, of supplemental BH 4 : Ⅺ and f curves 1 and 2, no additions; E and q, curves 3 and 4, plus CaM; ƒ and , curves 5 and 6, plus CaM and arginine.
urating concentrations of arginine and BH 4 (31) and that NOS was not completely saturated 2 with added BH 4 , an increase in heme reduction can lead not only to an increase in (coupled) NO formation but also to an increase in (uncoupled) formation of superoxide/peroxide and resultant increase in BH 4 oxidation. Peroxynitrite can also be formed by reaction of superoxide and NO (35). In the combined presence of arginine and BH 4 (Fig. 3,  curve 6), it is envisioned that the oxygenation at early times is largely coupled to NO formation, thereby limiting the formation of superoxide and peroxynitrite required for BH 4 oxidation. As the incubation proceeds, formation of superoxide/peroxynitrite and oxidation of BH 4 resume, as the level of BH 4 is decreased to a level that is no longer effective in coupling.
It is not entirely clear from these studies whether the increased oxidation of BH 4 in the presence of arginine reflects a donation of electrons by BH 4 for arginine oxygenation. The conditions of continuous turnover of NOS in this experiment, in which BH 4 is in a dynamic redox state, are not optimal for assessing such a role for BH 4 ; this role of BH 4 is examined below in a more rigorous manner under single turnover conditions. It is emphasized again that the arginine-stimulated oxidation of BH 4 under continuous turnover conditions is sensitive to SOD/catalase. If SOD/catalase-sensitive oxidation of BH 4 is an essential step in arginine oxygenation, 3 product formation by NOS would be expected to show a similar sensitivity. No such inhibition of NOS by SOD has been observed (36 -41). 4 Limited reports in the literature on the effects of catalase on NOS are conflicting. A preliminary abstract (42) claimed that NO formation (measured by the conversion of oxyhemoglobin to methemoglobin) catalyzed by NOS purified from porcine brain was catalase-dependent. On the other hand, Mittal (38) reported that NO formation (detected by activation of guanylate cyclase) catalyzed by crude extracts of rat brain was abolished by catalase. In neither of these studies did catalase affect citrulline formation, in agreement with reports by other workers (11,41,43).
Other explanations for the effect of arginine during continuous turnover of NOS are considered as follows.
(i) It is possible that the presence of arginine increases BH 4 oxidation by increasing the formation of peroxynitrite at the expense of superoxide. As explained above, whereas superoxide is the major product of oxygen reduction in the absence of arginine, the presence of arginine allows both NO and superoxide to be formed. The latter two compounds can react rapidly with each other to form peroxynitrite (35). Peroxynitrite is generally considered to be a more potent oxidant under physiological conditions than is superoxide or NO (35). However, preliminary studies 5 on the oxidation of 1 mM BH 4 in phosphate buffer (pH 7.6) showed comparable potencies for superoxide and peroxynitrite; hydrogen peroxide was relatively inactive, exhibiting an oxidation rate barely exceeding that of autoxidation by oxygen. These observations argue against an increased formation of peroxynitrite in the presence of arginine causing the increased oxidation of NOS-bound BH 4 . It should be noted, however, that the rank order of potency determined in free solution may not be identical to that determined for NOS-bound BH 4 .
(ii) It is possible that the increased net oxidation of BH 4 in the presence of arginine can be explained by the nature of the BH 4 oxidation products and their recycling to BH 4 . The nature of the product (BH 3 ⅐ or qBH 2 ) could be affected by whether BH 4 is oxidized by the obligate one electron oxidant, superoxide (44), or by peroxynitrite, which can catalyze one or two electron oxidations (45). Possible ways in which each of these oxidized pterin products could be reduced to BH 4 are discussed below.
BH 4 Oxidation under Single Turnover Conditions-In the single turnover system employed in the current work, it was previously shown that FSQ can provide all the electrons required for oxygenation of both arginine and NHA (21). Our NOS preparations contain only one FSQ per NOS dimer, thereby raising the problem of explaining the mechanism of delivery of the second electron required for arginine oxygenation (21). These considerations suggested, and our preliminary findings did not exclude, the possibility that compounds such as BH 4 might provide the second electron for arginine oxygenation by reduction of the ferric superoxide heme species to the ferric peroxo heme species (21). A similar proposal has recently been made by Bec et al. (13) based on the results of single turnover studies of the oxygenation of arginine to NHA supported by the ferrous oxygen complex of nNOS.
It should be noted that the current single turnover experiments were carried out under conditions that optimized the ability to detect any coupling of the oxidation of BH 4 to arginine oxygenation. Thus, any regeneration of BH 4 from qBH 2 was suppressed by omission of an NADPH-regenerating system from the reaction mixture. Furthermore, SOD and catalase were included in the reaction mixture to minimize any oxidation of BH 4 arising from the accumulation of ROS (superoxide and hydrogen peroxide). SOD and catalase were considered the best scavengers of ROS to be employed in our single turnover studies, for the following reasons. (i) It is generally believed that superoxide is a critical component of ROS and that SOD provides the primary defense mechanism, with catalase and the glutathione peroxidase/glutathione system sharing the task of converting hydrogen peroxide to water and oxygen (46 -48). As explained below, it was not feasible to include both catalase and the glutathione peroxidase/glutathione in our single turnover studies. (ii) SOD and catalase are quite specific for superoxide and peroxides, respectively, and are not known to catalyze reactions unrelated to their intended use. (iii) SOD and catalase resulted in efficient, although possibly not complete, scavenging ROS in our studies, even under the extreme conditions that favor the formation of superoxide and other ROS (Figs. 1 and 2). (iv) As previously explained in the discussion of BH 4 oxidation during continuous turnover, available evidence indicates that neither SOD nor catalase inhibits arginine oxygenation. (v) SOD and catalase do not interfere with the assay of BH 4 since they are removed from the reaction mixture by precipitation with perchloric acid. Furthermore, SOD and catalase do not interfere with any oxidation of BH 4 coupled to arginine oxygenation. This is not so with a number of other widely used antioxidants. For example, the antioxidants ascorbate and thiols such as glutathione regenerate BH 4 from qBH 2 (24,49,50) and could thus lead to an underestimation of the amount of oxidation of BH 4 under single turnover conditions. Similarly, ferricytochrome c could not be used as a scavenger of superoxide, because its oxidation of BH 4 (51) under our experimental conditions could not be excluded; such a reaction would lead to an overestimation of any oxidation of BH 4 coupled to arginine oxygenation. Various nitroxides act as mimics of SOD, and their excellent cell permeability has led to their extensive use as superoxide scavengers in biological systems. These compounds are not an appropriate substitute for SOD in our studies because the rate constants for reaction of superoxide with these nitroxides are generally smaller by 3-5 orders of magnitude than that with SOD (52).
The significantly greater oxidation of NOS-bound BH 4 under single turnover conditions in the presence of arginine (Fig. 4, curve 4) relative to NHA (Fig. 4, curve 3) is indeed consistent with a scheme in which BH 4 provides an electron for arginine oxygenation but is not required for the one electron-requiring oxygenation of NHA (53). A molar ratio for BH 4 oxidized/NHA formed of 0.5 is expected if the two-electron oxidation of BH 4 to qBH 2 supplied only the second electron for NHA formation, and a ratio of 1.0 if BH 4 provided both electrons for NHA formation. The observed molar ratio was not constant but increased from 1.0 to 1.7 as the reaction progressed (Fig. 5). This indicates that NOS-bound BH 4 is able to provide all the electrons for oxygenation of arginine to NHA. However, other pathways for BH 4 oxidation must operate in order to explain the high molar ratios of BH 4 oxidized per NHA formed at later incubation times and the significant oxidation of BH 4 in the absence of arginine (Fig.  4, curve 2). Despite the efficiency of SOD and catalase in scavenging ROS, it is important to note that a contribution of ROS in the single turnover oxidation BH 4 cannot be definitively excluded.
Implications of the Redox Reactions of NOS-bound BH 4 -The cycling between BH 4 and its oxidized product(s) reported here is of major significance in that it provides experimental support for a redox role of NOS-bound BH 4 and provides important insights toward defining the mechanistic details of this redox role of BH 4 . The possible significance of the oxidation of NOSbound BH 4 and regeneration of the oxidized pterin products to BH 4 is discussed separately below.
Our combined results are consistent with, but do not differentiate between, the two mechanistic roles recently proposed for BH 4 oxidation: as a reductant of oxygen (9), analogous to its role in the aromatic amino acid hydroxylases (10), and as an electron donor for reducing the ferric heme superoxide species to the ferric heme peroxo species in the oxygenation of arginine to NHA (13). Two additional roles of BH 4 oxidation that are consistent with our findings should be mentioned. The first is based on the report that the basal activity of NOS (measured in the absence of added BH 4 ) is positively correlated with the amount of bound BH 4 (6,54). Oxidation of NOS-bound BH 4 may therefore act as a switch in shutting off the formation of toxic levels of peroxynitrite, formed by reaction of NO with superoxide, that results in a variety of pathophysiological conditions (55,56). Whereas superoxide could continue to be formed by the BH 4 -depleted enzyme (6,28,57), studies of the injury that occurs during postischemic myocardial reperfusion, for example, indicate that peroxynitrite was the oxidant species responsible (56). The second suggested role for NOS-bound BH 4 oxidation is that the reaction is coupled to the conversion of some group(s) on NOS to the reduced form required for activity (11). A precedent for this model is the reduction by BH 4 of the inactive ferric form of phenylalanine hydroxylase to the active ferrous form (58,59). Our current findings could reflect an analogous Ca 2ϩ /CaM dependent reduction by BH 4 of some group(s) on NOS that must be in the reduced form for catalytic activity.
Regardless of the specific role of electron donation by BH 4 , regeneration of BH 4 from its oxidized product(s) probably plays a critical role in maintaining the pterin in its active, reduced, form and provides experimental support for the suggestion that BH 4 recycling is involved in the higher and more linear rates of NO formation in the presence of a BH 4 -regenerating system such as dihydropteridine reductase or reducing thiols (60 -62). Unless qBH 2 is effectively reduced to BH 4 , it rapidly rearranges to 7,8-dihydrobiopterin (63,64). The latter compound is not reduced by NOS to BH 4 (18), resulting in a net loss of BH 4 from the enzyme.
The present work opens up a number of promising lines of research in clarifying the redox role of BH 4 . For example, neither the products of BH 4 oxidation nor the mechanism of regeneration of BH 4 from these products is understood. qBH 2 is the usual product of enzymic oxidation of BH 4 , and regeneration of BH 4 from qBH 2 is catalyzed by dihydropteridine reductase (65). Regeneration of BH 4 by purified NOS proceeds in the absence of any added reductase, and must therefore be catalyzed by an inherent reductase activity of NOS. NOS does indeed catalyze the reduction of qBH 2 (18) . However, a number of observations cast doubt on the physiological significance of this reduction. For example, very high concentrations of qBH 2 were required to achieve rates that are comparable to those of NO synthesis, and the quinonoid form of 6-methyldihydropterin was more active than qBH 2 , whereas the product of this reduction, 6-methyltetrahydropterin, shows relatively little activation of NOS (18). An alternate proposal is that BH 4  It is noteworthy that, at least under our particular continuous turnover conditions in the presence of arginine, NOS appears to have a limited capacity for regeneration of BH 4 in relation to BH 4 oxidation. Only under conditions that severely inhibit superoxide formation (Fig. 2, curve 6) does BH 4 regeneration become the dominant reaction. However, our current work did not allow quantitation of the relative rates of oxidation of NOS-bound BH 4 and regeneration of BH 4 from its oxidized products. Thus, BH 4 regeneration observed on blocking the oxidation reaction was essentially complete by the time the first sample was taken (Fig. 1, curve 4, and Fig. 2, curve 6), and the observed oxidation of BH 4 in Figs. 1-3 is a composite of both oxidation and reduction reactions. Such quantitations will be invaluable in viewing the significance of the redox reactions of BH 4 in relation to overall NOS activity and to the reduction of qBH 2 by the "diaphorase activity" of NOS (18). A better understanding of these redox reactions might also help clarify why the uncoupled reduction of molecular oxygen is much more prominent with nNOS than with the other isozymes of NOS (reviewed in Ref. 4).