Interactions of peroxynitrite, tetrahydrobiopterin, ascorbic acid, and thiols: implications for uncoupling endothelial nitric-oxide synthase.

Tetrahydrobiopterin (BH4) serves as a critical co-factor for the endothelial nitric-oxide synthase (eNOS). A deficiency of BH4 results in eNOS uncoupling, which is associated with increased superoxide and decreased NO* production. BH4 has been suggested to be a target for oxidation by peroxynitrite (ONOO-), and ascorbate has been shown to preserve BH4 levels and enhance endothelial NO* production; however, the mechanisms underlying these processes remain poorly defined. To gain further insight into these interactions, the reaction of ONOO- with BH4 was studied using electron spin resonance and the spin probe 1-hydroxy-3-carboxy-2,2,5-tetramethyl-pyrrolidine. ONOO- reacted with BH4 6-10 times faster than with ascorbate or thiols. The immediate product of the reaction between ONOO- and BH4 was the trihydrobiopterin radical (BH3.), which was reduced back to BH4 by ascorbate, whereas thiols were not efficient in recycling of BH4. Uncoupling of eNOS caused by peroxynitrite was investigated in cultured bovine aortic endothelial cells (BAECs) by measuring superoxide and NO* using spin probe 1-hydroxy-3-methoxycarbonyl-2,2,5,5-tetramethyl-pyrrolidine and the NO*-spin trap iron-diethyldithiocarbamate. Bolus ONOO-, the ONOO- donor 3-morpholinosydnonimine, and an inhibitor of BH4 synthesis (2,4-diamino-6-hydroxypyrimidine) uncoupled eNOS, increasing superoxide and decreasing NO* production. Exogenous BH4 supplementation restored endothelial NO* production. Treatment of BAECs with both BH4 and ascorbate prior to ONOO- prevented uncoupling of eNOS by ONOO-. This study demonstrates that endothelial BH4 is a crucial target for oxidation by ONOO- and that the BH4 reaction rate constant exceeds those of thiols or ascorbate. We confirmed that ONOO- uncouples eNOS by oxidation of tetrahydrobiopterin and that ascorbate does not fully protect BH4 from oxidation but recycles BH3. radical back to BH4.

by ascorbate, whereas thiols were not efficient in recycling of BH 4 . Uncoupling of eNOS caused by peroxynitrite was investigated in cultured bovine aortic endothelial cells (BAECs) by measuring superoxide and NO ⅐ using spin probe 1-hydroxy-3-methoxycarbonyl-2,2,5,5-tetramethyl-pyrrolidine and the NO ⅐ -spin trap iron-diethyldithiocarbamate. Bolus ONOO ؊ , the ONOO ؊ donor 3-morpholinosydnonimine, and an inhibitor of BH 4 synthesis (2,4-diamino-6-hydroxypyrimidine) uncoupled eNOS, increasing superoxide and decreasing NO ⅐ production. Exogenous BH 4 supplementation restored endothelial NO ⅐ production. Treatment of BAECs with both BH 4 and ascorbate prior to ONOO ؊ prevented uncoupling of eNOS by ONOO ؊ . This study demonstrates that endothelial BH 4 is a crucial target for oxidation by ONOO ؊ and that the BH 4 reaction rate constant exceeds those of thiols or ascorbate. We confirmed that ONOO ؊ uncouples eNOS by oxidation of tetrahydrobiopterin and that ascorbate does not fully protect BH 4 from oxidation but recycles BH 3 ⅐ radical back to BH 4 .
The endothelial nitric-oxide synthase (eNOS) 1 is a dimeric enzyme composed of two catalytic domains: a C-terminal re-ductase domain, which binds NADPH, FMN, and FAD, and an N-terminal oxygenase domain, which binds a prosthetic heme group, 5,6,7,8-terahydrobiopterin (BH 4 ) (Scheme 1), oxygen, and L-arginine (1)(2)(3)(4)(5). The catalytic production of nitric oxide involves flavin-mediated electron transfer from C-terminal bound NADPH to the N-terminal heme center. At the heme site, oxygen is reduced and incorporated into the guanidino group of L-arginine, producing NO ⅐ and L-citrulline (1,2,6). eNOS is only catalytically active in the dimeric form, and the ability to bind BH 4 is dependent on dimer formation. There is evidence that BH 4 promotes dimer formation, although this is controversial (7). BH 4 plays a critical role in allowing electron transfer from the prosthetic heme to L-arginine. In the absence of BH 4 , electron flow from the reductase domain to the oxygenase domain is diverted to molecular oxygen rather than to L-arginine, leading to a condition known as eNOS uncoupling (8,9), which causes production of superoxide rather than nitric oxide.
Superoxide reacts rapidly with NO ⅐ to form the peroxynitrite anion (ONOO Ϫ ), which is a strong biological oxidant (10) known to oxidize lipids, protein, sulfhydryls, and DNA and to cause nitration of tyrosines (11)(12)(13). Recently, it has been suggested that BH 4 is an important target for oxidation by ONOO Ϫ (Scheme 2) (14). Treatment of purified eNOS with ONOO Ϫ significantly decreases the ability of the enzyme to produce NO ⅐ (15). Laursen et al. (14) and others (16,17) demonstrated that ONOO Ϫ is more potent than either superoxide or H 2 O 2 in causing oxidation of BH 4 . These investigators found that ONOO Ϫ dramatically increased vascular superoxide production in vessels from control mice but not in vessels from eNOS-deficient mice, suggesting that eNOS was the source of superoxide (17).
Cellular BH 4 levels also seem to be dependent on ascorbate. Pretreatment of endothelial cells with ascorbate increases NO ⅐ production without affecting NOS expression or L-arginine uptake (18,19). This effect of ascorbate is BH 4 -dependent as in the absence of BH 4 it is not observed (18,19). Whereas it is logical to assume that ascorbate may prevent oxidation of BH 4 , the precise mechanism whereby ascorbate can enhance cellular levels of BH 4 has not been defined.
In the present study, we examined the reaction of ONOO Ϫ with BH 4 , ascorbate, and thiols using electron spin resonance (ESR) and the spin probe 1-hydroxy-3-carboxy-2, 2,5-tetra-methyl-pyrrolidine (CPH). Uncoupling of eNOS by peroxynitrite in cultured bovine aortic endothelial cells (BAECs) was investigated by measuring O 2 . with new cell-permeable spin probe 1-hydroxy-3-methoxycarbonyl-2,2,5,5-tetramethylpyrrolidine (CMH) (20) and by measuring nitric oxide using colloidal Fe(DETC) 2 , which allows detection and quantification of NO ⅐ with high sensitivity and specificity (21,22). We also studied the role of ascorbate on BH 4 oxidation and uncoupling of eNOS and determined whether ascorbate prevents uncoupling of eNOS by scavenging peroxynitrite or if it improves eNOS function by recycling BH 4 .
Preparation of Spin Probe and BH 4 Stock Solutions-Stock solutions of CPH and CMH (10 mM) dissolved in 0.9% NaCl containing 1 mM diethylenetriamine-pentaacetic acid and purged with argon were prepared daily and kept under argon on ice. diethylenetriamine-pentaacetic acid was used to decrease autoxidation of hydroxylamines catalyzed by trace amount of transition metals. CPH and CMH were used in a final concentration of 1 mM. BH 4 was dissolved in argon-purged PBS with diethylenetriamine-pentaacetic acid (0.1 mM) and kept under argon on ice.
Cell Culture and Treatments-BAECs (Cell Systems, Kirkland, WA) were cultured in Medium 199 (Invitrogen), containing 10% fetal calf serum (Hyclone Laboratories, Logan, UT) as previously described (23). Confluent BAECs from passages 4 -7 cultured on 100-mm plates were used for ESR experiments. Cell suspensions were used for treatment with ONOO Ϫ . For this purpose, BAECs were scraped and centrifuged at 1800 rpm for 10 min and resuspended in 0.2 ml of ESR buffer. Peroxynitrite (0.27 mM) was added to the cell suspension as a bolus and vortexed. ESR measurements of O 2 . were made 3 min later. To test whether supplementation with BH 4 could restore eNOS function after ONOO Ϫ , the suspended cells were divided into two Eppendorf tubes, and one portion was incubated with 20 M BH 4 for 4 min at room temperature. Some of these cell suspensions were incubated with L-NAME (1 mM) or polyethylene glycol-superoxide dismutase (50 units/ ml) for 5 min, and then the spin probe CMH was added, and the mixture vortexed. Superoxide production was determined by inhibition with 50 units/ml polyethylene glycol-superoxide dismutase, whereas superoxide generated by uncoupled eNOS was measured as L-NAME (1 mM)-inhibited 3-methoxycarbonyl-proxyl (CM ⅐ ) nitroxide formation. In preliminary experiments, we confirmed that BH 4 (5-10 M) did not interfere with CMH detection of O 2 . -generated by xanthine and xanthine oxidase.
Measurements of Nitric Oxide with Fe(DETC) 2 -NO ⅐ production in BAECs has been measured with colloid solution of Fe(DETC) 2 as previously described (21,22). Due to its high lipophilicity, the formed NO ⅐ -Fe(DETC) 2 complex is exclusively associated with cell membrane and specifically detects NO ⅐ but not nitrite (21,22). After incubation with Fe(DETC) 2 , medium was aspirated, and cells were harvested with a rubber policeman in Krebs-Hepes buffer, resuspended, and aspirated into 1-ml syringes, which were frozen immediately in liquid nitrogen.
ESR Measurements-Oxidation of the spin probes CPH and CMH by reactive oxygen species (ROS) forms stable nitroxide radicals 3-carboxyproxyl (CP ⅐ ) and CM ⅐ , which can be assayed by ESR spectroscopy (20,24,25). The amount of nitroxide formed equals the concentration of the reacted oxidant species. The concentration of nitroxides was determined from the ESR amplitude according to a calibration curve using standard solutions of the 3-carboxyproxyl radical. ROS formation was measured from the kinetics of nitroxide accumulation by following the ESR amplitude of the low field component of ESR spectra. The rate of superoxide radical formation was determined by measuring the superoxide dismutase-inhibited nitroxide generation.
The reactivity of peroxynitrite scavengers with ONOO Ϫ was studied by competition with CPH using both bolus ONOO Ϫ and ONOO Ϫ generated by SIN-1. BH 4 and other ONOO Ϫ scavengers compete with CPH to react with ONOO Ϫ . The reactivity of each scavenger with bolus ONOO Ϫ was determined using the formula, where A 0 represents the ESR amplitude in the absence of ONOO Ϫ scavengers, A is the ESR amplitude in the presence of ONOO Ϫ scavengers, k is the reaction rate constant, and c is concentration. With SIN-1 as the ONOO Ϫ donor, the reactivity of ONOO Ϫ scavengers was calculated using the formula, where V 0 is the rate of nitroxide accumulation in absence of ONOO Ϫ scavengers, and V is the rate in presence of ONOO Ϫ scavengers. ESR samples were placed in a 100-l capillary and measured at room temperature using a field scan with the following ESR settings: microwave frequency, 9.78 GHz; modulation amplitude, 2 G; microwave power, 10 dB; conversion time 164 ms; time constant, 164 ms. Peroxynitrite and ROS production by BAECs were detected by following the low field peak of the nitroxide ESR spectra using time scans with the following ESR settings: microwave frequency, 9.78 GHz; modulation amplitude, 2 G; microwave power, 10 dB; conversion time, 1.3 s; time constant, 5.2 s. The intermediate BH 3 ⅐ radical was measured by direct ESR spectroscopy without a spin trap. The high resolution spectrum of the BH 3 ⅐ radical was detected and quantified using a microwave frequency of 9.78 GHz, modulation amplitude of 0.7 G, microwave power of 10 dB, conversion time of 82 ms, and time constant of 82 ms.
Frozen probes with NO ⅐ -Fe(DETC) 2 have a three-line ESR spectra whose amplitude is proportional to amount of bioactive NO ⅐ produced in cells (21,22,26). Frozen cell samples were measured in a finger Dewar flask filled with liquid nitrogen at 77 K in field scan with the following ESR settings: field sweep, 160 G; microwave frequency, 9.39 GHz; microwave power, 20 milliwatts; modulation amplitude, 3 G; conversion time, 655 ms; time constant, 5242 ms; receiver gain, 1 ϫ 10 4 ; number of scans, 4.
Computer Simulation of ESR Spectra-Computer simulation of the high resolution BH 3 ⅐ ESR spectra was used for calculation of hyperfine coupling constants. Programs for simulation of ESR spectra and spin trap data base are readily available through the Internet (on the World Wide Web at epr.niehs.nih.gov/). Details of this computer simulation program have been described elsewhere (27). Hyperfinecoupling constants are expressed as an average of ESR parameters obtained from computer simulation. The ESR spectrum of BH 3 ⅐ radi-SCHEME 1. Chemical structure of (6R)-5,6,7,8-tetrahydrobiopterin (BH 4 ). SCHEME 2. Mechanism of the reaction between ONOO ؊ and BH 4 and the role of ascorbate (AH ؊ ). ONOO Ϫ oxidizes BH 4 to the intermediate BH 3 ⅐ radical, which can decay to BH 2 or can be converted back to BH 4 by ascorbate.
cal was simulated as a combination of five nitrogens with four protons with following hyperfine coupling constants (a N ϭ 8.05 G, Statistical Analysis-Data are presented as means Ϯ S.E. Analysis with linear regression was done with the software Sigma Plot. For comparison of two groups, a one-tailed t test was employed using Excel software. Statistical significance was assumed when p Ͻ 0.05.

Reaction of ONOO Ϫ with BH 4 and Other Antioxidants-
Previous studies indicated that ONOO Ϫ readily oxidizes BH 4 in cultured cells and vessels. We therefore hypothesized that the reactivity of ONOO Ϫ with BH 4 would exceed that of ONOO Ϫ with other common intracellular antioxidant small molecules. The reactivity of ONOO Ϫ with BH 4 , BH 2 , GSH, cysteine, ascorbate, and Me 2 SO was studied by examining the competitive reaction between these agents and the hydroxylamine CPH. Boluses of ONOO Ϫ (0.27 mM) were added to reaction mixtures containing these potential ONOO Ϫ scavengers and CPH. In the absence of any scavenger, the reaction of ONOO Ϫ with CPH resulted in formation of CP ⅐ that could be detected as a strong ESR signal (Fig. 1A). BH 4 reduced CP ⅐ nitroxide generation in a concentration-dependent fashion, confirming that BH 4 could prevent the reaction of ONOO Ϫ with CPH (Fig. 1A). In contrast, BH 2 , cysteine, GSH, ascorbate, and Me 2 SO exhibited substantially less reactivity with peroxynitrite. For comparison of these data, CP ⅐ formation by ONOO Ϫ was set as 100%, and the effectiveness of the various scavengers was expressed as a percentage of this value. BH 4 strongly inhibited the ESR amplitude by 94%, whereas ascorbate and thiols inhibited ESR signals to a lesser degree (69 and 63%, respectively). Me 2 SO minimally inhibited the reaction of ONOO Ϫ with CPH (24%) (Fig. 1B).
By using the ESR amplitudes for the respective reactions as described under "Experimental Procedures" and in Equation 1, the relative reactivity of the antioxidant scavengers with ONOO Ϫ was calculated ( Fig. 2A). As a separate approach to quantify the reactivity of these antioxidants with peroxynitrite, the slopes of the lines presented in Fig. 2A were compared (Fig.  2B). According to these data, BH 4 reacted with peroxynitrite 10 times faster than ascorbate and 6 times more rapidly than either cysteine or GSH. These data indicate that, in the concentrations employed, neither dihydrobiopterin, ascorbate, nor thiols are able to fully protect BH 4 from oxidation by ONOO Ϫ .
Reactivity of Peroxynitrite Scavengers Studied with ONOO Ϫ Donor SIN-1-SIN-1 generates superoxide and nitric oxide, resulting in constant production of ONOO Ϫ , and therefore serves as a model for physiological ONOO Ϫ production. We therefore studied reactions of various scavengers with ONOO Ϫ generated by SIN-1. The rate of ONOO Ϫ formation by SIN-1 (5 mM) was measured from the kinetics of CP ⅐ nitroxide accumulation by following the ESR amplitude of the low field component of ESR spectra (Fig. 3A, insert). The control sample showed little nitroxide accumulation, whereas SIN-1 resulted in sharp increase in nitroxide generation (Fig. 3A). The accumulation of CP ⅐ nitroxide was strongly inhibited by 0.25 mM BH 4 (Fig. 3A). BH 4 had the highest reactivity with SIN-1generated ONOO Ϫ followed by cysteine, GSH, ascorbate, and Me 2 SO as calculated using formula 2 (Fig. 3B). In keeping with our results with bolus addition of ONOO Ϫ , BH 4 reacted with SIN-1-generated ONOO Ϫ 10 times faster than ascorbate and 6 times faster than GSH (Fig. 4). Table I provides summary data for experiments with both bolus peroxynitrite and SIN-1. For this analysis, the reactivity of BH 4 with ONOO Ϫ (either as a bolus or generated by SIN-1) was set as 100% and compared with reactivities of other antioxidants with ONOO Ϫ . The reactivities for cysteine, ascorbate, and GSH were similar for both systems.
Formation of BH 3 ⅐ Radical and Its Reaction with Ascorbate and Thiols-The above experiments suggest that ascorbate is only marginally effective in scavenging peroxynitrite. In prior studies, however, it has been reported that ascorbate preserves BH 4 content of purified eNOS permitting full catalytic function of the enzyme. It is also known that ascorbate is incapable of reducing BH 2 back to BH 4 (28). These data suggest that ascorbate may act as a "free radical sink" reducing the intermediate BH 3 ⅐ radical (29), which may be formed upon the reaction of BH 4 with peroxynitrite. We therefore performed additional experiments to examine interactions between ascorbate, BH 4 , and ONOO Ϫ . Whereas buffer containing BH 4 yielded no ESR signal (Fig. 5A), the bolus addition of ONOO Ϫ (0.27 mM) to BH 4 resulted in formation of a five-line ESR signal (Fig. 5A). Neither decomposed ONOO Ϫ nor NaOH (the solvent for ONOO Ϫ ) produced an ESR signal when exposed to BH 4 . High resolution ESR spectra revealed additional hyperfine components (Fig.   FIG. 1. Comparison (Fig. 5A). This computer simulation of the high resolution spectra of BH 4 and ONOO Ϫ was further supported by analysis of the ESR spectrum of BH 3 ⅐ radical in D 2 O (data not shown), which was previously reported by Vasquez-Vivar et al. (28).
We next sought to determine whether ascorbate or thiols could reduce the BH 3 ⅐ radical by adding these antioxidants 2-3 s after mixing of ONOO Ϫ with BH 4 . Ascorbate (100 M) inhibited the BH 3 ⅐ radical ESR signal by 32%, whereas 1 mM ascorbate decreased this signal by 79% (Fig. 5B). In contrast, the addition of either cysteine or GSH in concentrations of 1-10 mM only minimally reduced the ESR signal (Fig. 5B). Thus, ascorbate seems to be much more potent than thiol-containing compounds in reducing the BH 3 ⅐ radical. Recovery of Enzymatic Activity of Uncoupled eNOS in ONOO Ϫ -treated Endothelial Cells by BH 4 Supplementation-Next, we performed experiments to determine whether BH 4 was a target of ONOO Ϫ oxidation in vivo. To assess function of eNOS in cultured BAECs, the spin probe CMH was used to detect O 2 . . ROS production by BAECs was measured from the kinetics of CM ⅐ nitroxide accumulation by following the ESR amplitude of the low field component of ESR spectra (Fig. 6A,  inset). Untreated cells demonstrated minimal accumulation of nitroxide radical (Figs. 6A and 7A). In contrast, cells exposed to peroxynitrite robustly oxidized CMH to CM ⅐ , and this signal was inhibited by the addition of superoxide dismutase or by preincubation of cells with the NOS inhibitor L-NAME (Figs. 6B and 7B). SIN-1 also increased BAEC ROS production, and either L-NAME or superoxide dismutase inhibited this effect ( (14). In keeping with these previous findings, exposure of BAECs to xanthine (50 M) and xanthine oxidase (0.5 milliunits/ml), which generates superoxide and hydrogen peroxide, had no effect on subsequent production of O 2 . by endothelial cells (Fig. 6D). Nitroxide accumulation was similar to the control cells and did not show inhibition by L-NAME, indicating that superoxide did not uncouple eNOS (Fig. 6D). Whereas the effects of ONOO Ϫ , SIN-1, and the BH 4 synthesis inhibitor DAHP on eNOS function are consistent with depletion of BH 4 , these agents may have nonspecific effects on endothelial cell NO ⅐ and O 2 . production. We therefore examined whether BH 4 supplementation was capable of restoring eNOS activity. For this purpose, we measured O 2 . by BAECs incubated with exogenous BH 4 after the treatment with ONOO Ϫ , SIN-1, or DAHP (Fig. 7). In preliminary experiments, we con- firmed that BH 4 (5-10 M) did not interfere with CMH detection of O 2 . generated by xanthine and xanthine oxidase.
Treatment of control cells with BH 4 increased O 2 . production, and this was unaffected by L-NAME (Fig. 7A). In contrast to control cells, supplementation with BH 4 concentration-dependently inhibited O 2 . production in cells exposed to bolus ONOO Ϫ , as did L-NAME (Fig. 7B). A similar effect of BH 4 and L-NAME on endothelial O 2 . production was observed in cells that were exposed to either SIN-1 or the inhibitor of BH 4 synthesis DAHP (Fig. 7, C and D). The addition of L-NAME to BH 4 supplemented cells (Fig. 7, C and D) increased the ESR signal similar to the control cells, providing evidence that eNOS function was completely restored. The activity of eNOS was also determined by measuring NO ⅐ production in BAECs using the NO ⅐ -specific spin probe colloid  . 4. Reactivity of BH 4, thiols, and ascorbate with ONOO ؊ generated by SIN-1. Reactivity (k SCAV /k CPH ) of various scavengers with ONOO Ϫ generated by SIN-1 was calculated from the slopes of (V 0 /V) Ϫ 1 shown in Fig. 3B. The BH 4 -ONOO Ϫ reaction yielded the highest slope implying that BH 4 has the highest reactivity with ONOO Ϫ of the reductants studied, followed by cysteine, GSH, ascorbate and Me 2 SO. The calculated S.E. were less than 7%, and ONOO Ϫ reactivity with BH 4 was significantly different from the reactivity with other peroxynitrite scavengers (p Ͻ 0.01).  (Fig. 8). Treatment of control cells with BH 4 slightly reduced NO ⅐ production (Fig. 8). We then treated cells  N ϭ 8.05  G, a N ϭ 2.31 G, a N ϭ 1.79 G, a N ϭ 1.16 G,  a N ϭ 0.93 G, a N ϭ 8.41 G, a N ϭ 9.50 G, a N  ϭ 2.50 G, a N ϭ 1.06 G). B, reactivity of BH 3 ⅐ radical with the antioxidants ascorbate and thiols; the effect of ascorbate or thiols at different concentrations on the ESR signal of BH 3 ⅐ radical is presented as a percentage of ESR amplitude. The antioxidants were added a few seconds after the reaction of BH 4 with bolus ONOO Ϫ . Data confirm high reactivity of BH 3 ⅐ radical with ascorbate but not with thiols.

FIG. 3. Comparison of reactions between various potential ONOO ؊ scavengers and ONOO
FIG. 6. Superoxide production by endothelial cells following exposure to ONOO ؊ or SIN-1. ROS formation was measured in BAECs after treatment with bolus ONOO Ϫ , ONOO Ϫ donor SIN-1, or superoxide generated by xanthine and xanthine oxidase as accumulation of CM ⅐ nitroxide, which was followed by low field component of the ESR spectra shown by the arrow in the inset (A). Superoxide production was determined by inhibition with 50 units/ml polyethylene glycol-superoxide dismutase, whereas superoxide generated by uncoupled eNOS was measured as L-NAME (1 mM)-inhibited CM ⅐ nitroxide formation. A, CM ⅐ accumulation in nontreated control cells. The inhibition of coupled eNOS with L-NAME increased the amount of detected superoxide. B and C, ROS production in BAECs treated by bolus 0.27 mM ONOO Ϫ or 0.5 mM SIN-1. The inhibition of uncoupled eNOS with L-NAME decreased the amount of detected superoxide. D, ROS production in BAECs treated by 50 M xanthine and 0.5 milliunits/ml xanthine oxidase. with 0.5 mM SIN-1 to generate ONOO Ϫ at a rate of 1-1.5 M/min, levels similar to those observed in pathophysiological conditions (30). This significantly decreased BAEC NO ⅐ production (Fig. 8). Following treatment with SIN-1, supplementation with BH 4 completely restored NO ⅐ production to values similar to that observed in control cells. Similar to SIN-1, treatment of cells with the BH 4 synthesis inhibitor DAHP for 24 h also decreased NO ⅐ production (Fig. 8), and BH 4 reversed this effect (Fig. 8). Thus, by measuring O 2 . and NO ⅐ from cultured endothelial cells, we have shown that ONOO Ϫ derived from SIN-1 and DAHP uncouples eNOS and that BH 4 supplementation corrects this. These data confirm that ONOO Ϫ uncoupled eNOS in BAECs by oxidation of BH 4 , because supplementation with BH 4 after ONOO Ϫ treatment fully restored eNOS function.
Ascorbate Prevention of eNOS Uncoupling-Our previous data have shown that BH 4 administration could recover eNOS function after uncoupling by ONOO Ϫ and also that ascorbate recycles BH 4 after its reaction with ONOO Ϫ via reducing the intermediate BH 3 ⅐ radical. It was of interest to determine whether exogenous BH 4 or ascorbate could protect eNOS against ONOO Ϫ in intact endothelial cells. Treatment of control cells with ascorbate did not affect NO ⅐ production (Fig. 9). As in Fig. 8, SIN-1 treatment markedly decreased endothelial cell NO ⅐ production (Fig. 9). This effect of SIN-1 was not altered by pretreatment of cells with either ascorbate or BH 4 (Fig. 9), indicating that saturating cells with ascorbate or BH 4 does not prevent eNOS uncoupling when the cells are subsequently challenged with the ONOO Ϫ donor SIN-1. In contrast, co-incubation of cells with BH 4 and ascorbate during SIN-1 treatment completely prevented the effect of SIN-1 (Fig. 9) on NO ⅐ pro-duction. The effect of co-incubation with either ascorbate or BH 4 alone during SIN-1 treatment was approximately half that of when these agents were used together (Fig. 9). These data as well as data presented in Fig. 5B strongly support the concept that ascorbate protects eNOS from uncoupling by recycling intracellular BH 4 (Scheme 2). DISCUSSION In the present study, we demonstrated that ONOO Ϫ reacts with BH 4 ϳ10 times faster than with ascorbate and 6 times faster than with the thiol-containing compounds glutathione and cysteine. We also showed that this reaction led to formation of the BH 3 ⅐ radical. The BH 3 ⅐ radical was found to have high reactivity with ascorbate but not with thiols. Finally, we demonstrated that peroxynitrite leads to eNOS uncoupling in cultured endothelial cells and that following exposure of endothelial cells to peroxynitrite, eNOS activity could be fully restored by treatment of the cells with tetrahydrobiopterin. Taken together, these data demonstrate that BH 4 is probably a crucial target for ONOO Ϫ and that even in the presence of common cellular antioxidants such as ascorbate and thiols, ONOO Ϫ can lead to BH 4 oxidation.
By examining the competition between oxidation of the spin probe CPH and various potential antioxidants, we were able to compare rate constants of ONOO Ϫ reactions with tetrahydrobiopterin, ascorbate, and thiols. The rate constants of ONOO Ϫ reactions with ascorbate, cysteine, and glutathione have been previously determined to be 236 M Ϫ1 ⅐s Ϫ1 , 10 3 M Ϫ1 ⅐s Ϫ1 , and 5.8⅐10 2 M Ϫ1 ⅐s Ϫ1 (31-33). Table I  times higher than the rate constant of ascorbate and 6 times higher than the rate constants of thiols. Given our current data, it is possible to estimate the rate constant for the reaction between ONOO Ϫ and BH 4 as being 6⅐10 3 M Ϫ1 ⅐s Ϫ1 .
It has recently been shown that ascorbate treatment increases endothelial cell NO ⅐ synthesis and tetrahydrobiopterin levels (18,19,34), although the mechanism whereby this occurs is not well understood. BH 2 is not reduced to BH 4 by ascorbate (28), and these prior studies showed that ascorbate did not affect expression of eNOS or GTP-cyclohydrolase, the rate-limiting enzyme for BH 4 synthesis (18,34). The authors of this prior study (18) indicated that ascorbate stabilized BH 4 within the endothelial cell. Our present data provide further insight into this effect of ascorbate. It is unlikely that ascorbate prevents oxidation of BH 4 by scavenging ONOO Ϫ , since it was found to be 10-fold less reactive with ONOO Ϫ than BH 4 . Our data indicate that the product of the reaction between ONOO Ϫ and BH 4 is the BH 3 ⅐ radical and that this radical is highly reactive with ascorbate. These data are consistent with the recent observation that the BH 3 ⅐ radical is reduced by ascorbate to BH 4 with a rate constant of ϳ1.7⅐10 5 M Ϫ1 s Ϫ1 (29). Thus, an important mechanism whereby ascorbate stabilizes levels of BH 4 seems to involve reduction of the BH 3 ⅐ radical back to BH 4 , rather than prevention of oxidation of BH 4 by oxidants such as ONOO Ϫ . According to our data, BH 3 ⅐ did not exhibit the same reactivity with thiols, a finding in agreement with previously published data on reactivity of the BH 3 ⅐ radical (29). It also seems that ascorbate does not preserve eNOS activity by scavenging ONOO Ϫ or by preventing BH 4 from reacting with ONOO Ϫ but that ascorbate improves eNOS function by recycling BH 4 (Scheme 2, Fig. 5B). This is supported by our experiments with SIN-1-treated BAECs (Fig. 9), where co-incubation with BH 4 and ascorbate fully prevented loss of eNOS function.
Several clinical studies have shown that administration of intraarterial administration of vitamin C can improve endothelium-dependent vasodilatation in the forearms of humans with hypercholesterolemia, diabetes, and cigarette smoking (35)(36)(37).
The effect of vitamin C in these studies has largely been attributed to scavenging of O 2 . (38). Whereas O 2 . scavenging may be a mechanism for improvement in endothelium-dependent vasodilatation in these studies, our current data would indicate that another effect of vitamin C might involve recycling of the BH 3 ⅐ radical to BH 4 . Our results are consistent with the previously reported data showing that low concentrations of ascorbate stimulate nitric-oxide synthase in activated macrophages (39). Thus, recycling of the BH 3 ⅐ radical to BH 4 by ascorbate may play an important role in preserving the activity of not only endothelial but also the inducible and neuronal isoforms of nitric-oxide synthase.
In keeping with the above findings, we observed that ONOO Ϫ generated from SIN-1 led to a condition of eNOS uncoupling in cultured cells. This was reflected by a decrease in NO ⅐ and a concomitant increase in O 2 . production, which could be inhibited by the NOS inhibitor L-NAME.
Recently, it has been suggested that ONOO Ϫ uncouples eNOS by oxidation of the zinc-thiolate complex that comprises the BH 4 binding site. This in turn leads to dissociation of eNOS dimers to monomers (40). In our experiments, however, BH 4 supplementation fully restored NO ⅐ production in ONOO Ϫtreated cells, a finding that seems at odds with the concept that the BH 4 binding site is disrupted by ONOO Ϫ . If the zincbinding site was a primary target for ONOO Ϫ , eNOS would be irreversibly uncoupled, and BH 4 supplementation would seem unlikely to restore endothelial cell NO ⅐ production after exposure to ONOO Ϫ . Our data also demonstrate that the reactivity of BH 4 with ONOO Ϫ substantially exceeds that of thiols with ONOO Ϫ . Given these considerations, it seems unlikely that ONOO Ϫ would react with the zinc-thiolate center of eNOS in preference to BH 4 . It is possible that both the zinc-thiolate center and BH 4 are targets of oxidation by ONOO Ϫ , particularly when high levels of this oxidant are present, but our data would indicate that BH 4 is preferentially oxidized.
Related to the above discussion, the previous studies at low temperature SDS-PAGE show that BH 4 markedly stabilized the dimer of eNOS (41,42) by preventing dissociation of the heme (42). Low temperature SDS-PAGE itself, however, affects dimer formation. Therefore, it is unclear whether cellular eNOS in situ exists in the same monomer/dimer forms as it does in gel in vitro. Data obtained with transformed yeast suggest that eNOS does not require BH 4 for dimer formation (7,43). Nevertheless, it has been previously reported that BH 4 increases the critical temperature for dissociation of eNOS dimer from 30 -40 to 40 -50°C (7). Thus, BH 4 may have some effect on stabilization of the eNOS dimer, and oxidation of BH 4 by ONOO Ϫ is a likely cause of partial dissociation of the eNOS dimer in intact endothelial cells. BH 4 has been postulated to be deficient in various conditions associated with altered endothelial function (17). Depletion of endothelial BH 4 has been shown to stimulate superoxide production from the isolated eNOS enzyme (44) and from eNOS in intact endothelial cells (45). Supplementation with BH 4 enhances NO ⅐ production, improves endothelium-dependent vasodilatation (46), and efficiently couples NADPH oxidation to NO ⅐ synthesis and inhibits superoxide and hydrogen peroxide formation (8,44). We have shown that oxidation of BH 4 by ONOO Ϫ uncouples eNOS and that BH 4 supplementation fully restores eNOS function after uncoupling, providing evidence that BH 4 is a crucial target for peroxynitrite under physiological conditions.
It is now well established that numerous common diseases such as hypercholesterolemia, hypertension, diabetes, and heart failure are associated with a loss of NO ⅐ production by the endothelium, a condition commonly referred to as endothelial dysfunction (47). In many of these conditions, eNOS uncoupling seems to be present, leading to an increase in endothelial cell O 2 . production and a decrease in NO ⅐ production. Our current data, together with other recent publications (14,16,17), strongly suggest that one mechanism leading to eNOS uncoupling is oxidation of BH 4 by ONOO Ϫ and similar oxidants. Of note, BH 4 can be administered orally and is effective in treatment of mild phenylketonuria (48). Based on our current findings, it is possible that BH 4 or more likely a combination of BH 4 and vitamin C may prove useful in correcting endothelial dysfunction in these common disorders.