A tetrahydrobiopterin radical forms and then becomes reduced during Nomega-hydroxyarginine oxidation by nitric-oxide synthase.

Nitric-oxide synthases are flavoheme enzymes that catalyze two sequential monooxygenase reactions to generate nitric oxide (NO) from l-arginine. We investigated a possible redox role for the enzyme-bound cofactor 6R-tetrahydrobiopterin (H4B) in the second reaction of NO synthesis, which is conversion of N-hydroxy-l-arginine (NOHA) to NO plus citrulline. We used stopped-flow spectroscopy and rapid-freeze EPR spectroscopy to follow heme and biopterin transformations during single-turnover NOHA oxidation reactions catalyzed by the oxygenase domain of inducible nitric-oxide synthase (iNOSoxy). Significant biopterin radical (>0.5 per heme) formed during reactions catalyzed by iNOSoxy that contained either H4B or 5-methyl-H4B. Biopterin radical formation was kinetically linked to conversion of a heme-dioxy intermediate to a heme-NO product complex. The biopterin radical then decayed within a 200-300-ms time period just prior to dissociation of NO from a ferric heme-NO product complex. Measures of final biopterin redox status showed that biopterin radical decay occurred via an enzymatic one-electron reduction process that regenerated H4B (or 5MeH4B). These results provide evidence of a dual redox function for biopterin during the NOHA oxidation reaction. The data suggest that H4B first provides an electron to a heme-dioxy intermediate, and then the H4B radical receives an electron from a downstream reaction intermediate to regenerate H4B. The first one-electron transition enables formation of the heme-based oxidant that reacts with NOHA, while the second one-electron transition is linked to formation of a ferric heme-NO product complex that can release NO from the enzyme. These redox roles are novel and expand our understanding of biopterin function in biology.

nitric oxide (NO) and citrulline (1)(2)(3). The reaction consumes 1.5 mol of NADPH and 2 mol of O 2 for each NO formed from Arg and is catalyzed in two steps (Scheme 1). Arg is first hydroxylated to form water and N -hydroxyl-L-arginine (NOHA) as an enzyme-bound intermediate. NOHA is then further oxidized to generate water, NO, and citrulline. Both reactions take place in the oxygenase domain of NOS, which contains iron protoporphyrin IX (heme), the essential cofactor (6R)-5,6,7,8-tetrahydro-L-biopterin (H 4 B) and the Arg binding sites (4,5). Electrons derived from NADPH are provided to the NOS oxygenase domain by an attached reductase domain that binds FMN, FAD, and NADPH. A calmodulin binding site is located between the oxygenase and reductase domains and serves to regulate electron transfer (6 -8).
The NOS heme iron is ligated by a cysteine thiolate and is responsible for binding O 2 and catalyzing its reductive activation during NO synthesis (9,10). How H 4 B facilitates this process is a topic of current interest. Investigations have typically employed single catalytic turnover reactions with recombinantly expressed NOS oxygenase domains (NOSoxy). During Arg hydroxylation there is a buildup of a H 4 B radical in all three NOS isozymes (11)(12)(13), suggesting that H 4 B performs a reductive role in the reaction. Our previous work established that formation of the H 4 B radical is kinetically coupled to reduction of a heme ferric-superoxy intermediate (Fe III O 2 Ϫ ) that forms in NOS after O 2 binds to its ferrous heme (14). Electron transfer from H 4 B to the Fe III O 2 Ϫ intermediate enables formation of a heme-oxy species that hydroxylates Arg. Moreover, the tempo of Fe III O 2 Ϫ reduction by H 4 B is important and must be maintained for efficient Arg hydroxylation. The tempo is regulated by the protein residues that interact with H 4 B (for example Trp 457 in mouse iNOS) (15) and by the structure of H 4 B itself (16).
Although the redox role for H 4 B during Arg hydroxylation seems established (17), it is still unclear if H 4 B is redox active during the NOHA oxidation reaction. Theoretically, it is not essential that H 4 B provides an electron to the Fe III O 2 Ϫ intermediate in this step (18). Moreover, a model study suggested that NOHA could provide an electron directly to the Fe III O 2 Ϫ species in place of H 4 B (19). Only a small amount of H 4 B radical (ϳ0.03 per heme) was observed during a NOHA single turnover reaction catalyzed by iNOSoxy (11). However, we and others (20 -22) have speculated that this is still consistent with the H 4 B radical forming and then becoming reduced back to H 4 B during the time frame of the reaction. Indeed, we have argued that this behavior may be essential for NOS to generate NO instead of nitroxyl (23). Other experimental evidence indirectly supports a redox role for H 4 B during NOHA oxidation. For example, NO synthesis from NOHA does not occur when enzyme-bound H 4 B is replaced with 6,7-dihydrobiopterin (H 2 B) or with the redox-inactive analog 5-deaza-tetrahydrobiopterin (20,22). In addition, mutational substitution of a NOS Trp residue that interacts with H 4 B caused similar changes in the * This work was supported by National Institutes of Health Grants CA53914 (to D. J. S.) and GM58481 (to R. H.) and by a Fellowship from the American Heart Association, Ohio Valley Affiliate (to C.-C. W.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
To further investigate this possibility, we utilized stoppedflow spectroscopy and rapid-freeze EPR to follow in a timeresolved way the heme and H 4 B transitions that occur during NOHA oxidation as catalyzed by iNOSoxy. We found that a H 4 B radical does form in conjunction with Fe III O 2 Ϫ disappearance, followed by a rapid reduction of the radical, all within the time frame of the single catalytic turnover. Thus, our work suggests that H 4 B reduces the Fe III O 2 Ϫ intermediate in the NOHA reaction and then performs a distinct electron acceptor function at a later point in the reaction. The mechanistic implications of our findings are discussed.

EXPERIMENTAL PROCEDURES
Chemicals-H 4 B, H 2 B, and 5MeH 4 B were obtained from Schirck's Laboratory (Jona, Switzerland). All other chemicals were either from Sigma or Fisher Scientific.
Protein Expression and Purification-Pterin-free mouse iNOSoxy (amino acids 1-498) containing a six-histidine tag at its C-terminal was expressed in Escherichia coli BL21 using the PCWori vector and purified in the absence of H 4 B and Arg as reported previously (14). The enzyme concentration was determined from the 444 nm absorbance of the ferrous-CO complex by using extinction coefficient of 76 mM Ϫ1 cm Ϫ1 . The final concentrated protein was stored in buffer containing 50 mM EPPS, pH 7.5, 10% glycerol, 2 mM ␤-mercaptoethanol, and 0.25 M NaCl.
Ferrous Heme Protein Preparation and Rapid Kinetics-The ferrous iNOSoxy was prepared as described previously (14,16). The rapidscanning stopped-flow spectroscopy and rapid-freeze EPR measurements were done as previously (14). EPR spectra were recorded in a Bruker ER 300 electron paramagnetic resonance spectrometer equipped with an ER 035 NMR gauss meter and a Hewlett-Packard 5352B microwave frequency counter. Temperature control was achieved using Oxford Instruments ESR 900 continuous-flow liquid helium cryostat and ITC4 temperature controller. All spectra were obtained at 150 and 4 K using a microwave power of 2 milliwatts, a frequency of 9.5 GHz, modulation amplitude of 10 G, and modulation frequency of 100 kHz. The Fe II NO complex of iNOSoxy was prepared and the EPR measured as indicated previously (24). The protein concentrations and reaction conditions are described in the text below. All experiments were performed in 10°C and repeated at least three times.
Redox State of Bound Biopterin-We determined the final oxidation state of bound biopterin that remained in the iNOSoxy enzyme following single catalytic turnover reactions, using an established method (25) with modifications. A buffered solution (0.1 ml total) that contained 0.1 mM iNOSoxy, 1 mM H 4 B, and 10 mM Arg (or NOHA) was incubated at room temperature for 20 min to allow for binding. The solution was then passed through two consecutive Micro-spin P-30 columns (Bio-Rad) to remove free H 4 B and substrate. Arg (or NOHA) was then re-added to the solution at a final concentration of 10 mM and the sample was made anaerobic by consecutive cycles of vacuum and purging with N 2 . The iNOSoxy was then reduced to ferrous form by addition of a near stoichiometric amount of dithionite solution. To initiate the single turnover reaction, 0.1 ml of the ferrous iNOSoxy solution was transferred using a gas tight syringe into an open eppendorf tube that contained 20 l of O 2 -saturated buffer (40 mM EPPS, pH 7.5) and 4 l of an iodine solution (0.9% (w/v) iodine in H 2 O). This iodine solution was needed to oxidize any excess dithionite, which if present was found to reduce oxidized biopterin in the reaction samples and thus confound the results. After vigorous mixing for ϳ20 s to allow for completion of the single turnover reaction, 50-l aliquots were transferred either into a vial containing 50 l 0.2 N NaOH and 5 l alkaline iodine solution (0.9% (w/v) iodine and 1.8% (w/v) KI in 0.1 N NaOH) or into a vial containing 50 l 0.2 N HCl and 5 l of acidic iodine solution (0.9% (w/v) iodine and 1.8% (w/v) KI in 0.1 N HCl). Samples then were incubated at room temperature in the dark for at least 1.5 h. The sample in alkaline solution was then neutralized by adding 10 l of 1.0 N HCl, and all remaining iodine in all of the samples was quenched by adding 5 l of freshly prepared 4% (w/v) ascorbic acid solution. The precipitated protein was removed from each sample by centrifuging at 10,000 ϫ g for 10 min. A 5-l aliquot from each sample was then injected into a 10 ϫ 250-mm Microsorb C18 column (Ranin Instrument Inc.) that was equilibrated with 50 mM ammonium phosphate, pH 3.0. A fluorescence detector (Waters 710) was used to detect biopterin products with excitation wavelength set at 360 nm and emission wavelength set at 460 nm. Total biopterin (oxidized ϩ reduced) was determined from the area of the peak eluting at 21.5 min that had undergone acidic workup, while oxidized biopterin was determined from the area of this same peak from the sample that had undergone basic workup (25). The amount of pterin remaining as H 4 B after the reaction was calculated by subtracting the amount of biopterin present in the alkaline solution from the amount of biopterin present in the acidic solution. Any dilution difference between the acidic and basic condition during the work-up was corrected for in the calculation.

H 4 B Radical Formation during NOHA Oxidation-We
looked for H 4 B radical formation during iNOSoxy-catalyzed oxidation of NOHA at 10°C under single turnover conditions. Reactions were initiated by rapid mixing an anaerobic solution containing ϳ300 M ferrous iNOSoxy, 10 mM NOHA, and 2 mM H 4 B with O 2 -saturated buffer (O 2 concentration is ϳ2.2 mM at 0°C (11)) in our rapid-freeze instrument. As shown in Fig. 1, a radical species did form during the NOHA reaction that displayed spectroscopic features that were very similar to the H 4 B radical formed in the Arg hydroxylation reaction (14,15). The radical signal in the NOHA reaction had a g ϭ 2.0, a peak-totrough width of ϳ40 G, and a powder hyperfine structure. In replica reactions run with iNOSoxy that contained 5MeH 4 B in place of H 4 B, we detected a distinct radical signal during the NOHA reaction with properties that were similar to those of the 5MeH 4 B radical formed during the Arg hydroxylation reaction ( Fig. 1) (16). No radical signal was observed in control reactions that contained H 2 B in place of H 4 B (data not shown). The data reveal that biopterin radicals do build up during the NOHA oxidation reaction. Fig. 2 shows the time course of H 4 B or 5MeH 4 B radical buildup and disappearance in the Arg and NOHA oxidation reactions catalyzed by iNOSoxy. Clearly, the biopterin radical was more transient in the NOHA reaction compared with the Arg reaction. Both the H 4 B and 5MeH 4 B radicals reached maximum abundance about 100 ms after initiating the NOHA reaction, and both signals were practically gone within 300 ms (Fig. 2). Fitting the radical signal intensities to a reaction SCHEME 1. Reaction catalyzed by NOS. (14) gave calculated rates of H 4 B radical formation and decay in the NOHA reaction of 31 and 8 s Ϫ1 , respectively, with a calculated yield of 0.55 H 4 B radical formed per enzyme heme (Table I). Similarly, rates of 5MeH 4 B radical formation and decay were calculated to be 35 and 15 s Ϫ1 , respectively, with 0.65 5MeH 4 B radical formed per enzyme heme. Thus, in the NOHA reaction, 5MeH 4 B radical formation was only slightly faster than that of H 4 B, and the 5MeH 4 B radical disappeared about twice as fast (Table I).
Because a H 4 B radical also forms in substrate-free iNOSoxy during single turnover reactions (11), 2 we checked whether the radical signals that we observed during our NOHA reaction might arise from the presence of some substrate-free enzyme. Increasing the NOHA concentration 2-fold did not reduce the radical intensity in our reactions (data not shown), consistent with iNOSoxy being saturated with substrate. In addition, the kinetics of biopterin radical formation and decay seen here in the NOHA reaction are quite distinct from the kinetics observed in reactions with substrate-free iNOSoxy. For example, in a representative substrate-free control reaction, the rates of biopterin radical formation and decay were 9.0 and 1.2 s Ϫ1 , respectively. Together, this argues that the transient biopterin radicals that we observed in the NOHA reaction had formed in iNOSoxy enzyme molecules that were actively catalyzing NOHA oxidation.
Fate of the H 4 B Radical-We next determined whether the H 4 B radical that formed during the NOHA or Arg reactions disappeared via an oxidative or reductive process. iNOSoxy samples were preincubated with H 4 B and substrate (NOHA or Arg) and then passed through a gel filtration column to remove unbound H 4 B just prior to setting up the single turnover reactions. The reactions were initiated by manually mixing at room temperature the filtered, anaerobic ferrous enzyme solutions that contained substrate with an O 2 -saturated solution, and these were then left to react for ϳ20 s. We then determined the reduction state of the biopterin in each enzyme reaction sample. Fig. 3 contains representative HPLC fluorescent traces that were used to determine the amount of oxidized biopterin present in the NOHA or Arg reaction samples. Results from three experiments are listed in Table II. Almost all of the biopterin was in a fully reduced state (91 Ϯ 2% H 4 B) at the end of the NOHA reaction, while at the end of the Arg reaction the biopterin was predominantly in an oxidized state (37 Ϯ 8% H 4 B). 3 As controls, we ran reactions with iNOSoxy that contained H 2 B plus Arg but otherwise underwent identical procedures and with iNOSoxy that contained H 4 B plus Arg that did not undergo the dithionite reduction step. In both cases, the original oxidation state of the biopterin was maintained at the end of the reaction period, indicating that non-enzymatic reduction or oxidation of the biopterin did not occur (data not shown). Together, these data indicate that H 4 B radical disappearance in the NOHA reaction was due to an enzyme-mediated reductive process, while in the Arg reaction radical disappearance was due to an oxidative process.
Heme Transitions during NOHA Oxidation-We utilized rapid-scanning stopped-flow spectroscopy to monitor heme transitions that occurred during the NOHA single-turnover reactions catalyzed by iNOSoxy. As in our previous report (20),

FIG. 2. Kinetics of biopterin radical formation and decay during Arg hydroxylation and NOHA oxidation by H 4 B-or 5MeH 4 Bsaturated iNOSoxy.
A shows the kinetics of biopterin radical formation and decay during Arg hydroxylation for iNOSoxy containing H 4 B (solid circles) or 5MeH 4 B (solid squares). Data are from Ref. 16. B shows the kinetics of biopterin radical formation and decay during NOHA oxidation for iNOSoxy containing H 4 B (solid circles) or 5MeH 4 B (solid squares). Inset, radical buildup versus the log of the reaction time.
Reactions were setup and run as described in Fig. 1 Ϫ intermediate was also rapid and was kinetically associated with the appearance of a heme-NO product complex whose buildup was maximal at 50 -100 ms. B illustrates the formation and disappearance kinetics of the heme-NO product complex as judged by a plot of absorbance at 438 nm versus time. The formation and decay rates estimated in this manner were 30.5 and 2.0 s Ϫ1 , respectively, and thus were similar to the rates derived by the global analysis. NO then subsequently dissociated from the heme-NO product complex to generate ferric enzyme. DISCUSSION We undertook single turnover studies to investigate redox functions of H 4 B in the second reaction of NO synthesis. We found that H 4 B is first oxidized to its radical and then is reduced back to H 4 B during enzymatic conversion of NOHA to NO. The same process occurred in an iNOSoxy reaction that contained 5MeH 4 B in place of H 4 B. Our measures of the amount of pterin radical formed per heme argue that these pterin redox transitions are significant and take place in a majority of enzyme molecules during the NOHA single turnover reactions. Thus, H 4 B appears to undergo two redox transitions during the second step of NO synthesis.
The pterin radicals formed in the NOHA reaction displayed unique kinetic properties. For example, 5MeH 4 B and H 4 B radicals formed at similar rates in the NOHA reactions, whereas in the Arg reactions the rate of radical formation by H 4 B is 4 times slower than 5MeH 4 B (see Table I). Both pterin radicals also disappeared at least 10 times faster in the NOHA reactions compared with the Arg reactions. This is consistent with the pterin radicals being actively reduced during enzyme catalysis in the NOHA reaction, whereas the radicals are left to oxidize following completion of the Arg reaction. Together, these differences establish that the pterin operates under unique circumstances in each reaction of NO synthesis.
Mechanistic Implications-To understand the potential significance and roles of H 4 B redox transitions during NOHA oxidation, it helps to view their timing in relation to the heme transitions that occur during the same reaction. Our stopped-   flow study detected a minimum of three consecutive heme transitions that involve four heme species; the initial ferrous iNOSoxy species binds O 2 to generate an Fe III O 2 Ϫ intermediate, which then appears to convert to an Fe III NO product complex, which subsequently dissociates into ferric iNOSoxy and free NO (Fig. 5).
Our stopped-flow analysis showed that the majority of oxygen binding to ferrous heme occurs within 30 ms after initiating the NOHA reactions. The Fe III O 2 Ϫ intermediate then disappears by 100 ms, and this transition occurs coincident with buildup of the pterin radical. Their tight kinetic relationship is consistent with H 4 B radical formation representing transfer of one electron from H 4 B to the Fe III O 2 Ϫ intermediate, as also occurs in the Arg reaction (14,15). Disappearance of the Fe III O 2 Ϫ species is also coincident with buildup of a heme-NO product complex that we have identified as ferric based on its spectral signature. Thus, H 4 B electron transfer appeared to form a heme-based oxidant that quickly reacts with NOHA and leads to buildup of an immediate heme-NO product complex. Next there occurred a reduction of the H 4 B radical, beginning at around 100 ms and becoming essentially complete by 300 ms. This reductive transition is unique to the NOHA reaction. Unfortunately, reduction of the H 4 B radical was not associated with discernable buildup of a distinct heme spectral species. Our attempts to fit the spectral data so as to incorporate a new intermediate within this time range were not successful. The final transition (from about 300 ms onward) represents dissociation of NO from the ferric heme-NO complex and regenerates the ferric enzyme. Together, our analysis suggests that the NOHA reaction may actually contain at least four consecutive heme transitions, although only three of these transitions are distinguishable by our stopped flow analysis (Fig. 5). Importantly, two of the transitions involve separate one-electron redox transformations by H 4 B.
Our current work reveals that H 4 B may perform a common function in both reactions of NO synthesis, namely, to provide an electron to the Fe III O 2 Ϫ intermediate. However, in the NOHA reaction there also occurs a subsequent reductive transition whereby the H 4 B radical receives an electron. We believe that this function may be essential for NO synthesis. As discussed previously (23), if the H 4 B radical does not take back an electron from the system, then NOHA would undergo a twoelectron oxidation instead of a three-electron oxidation and generate nitroxyl, the one-electron reduced form of NO. This process actually appears to take place in H 4 B-free NOS enzymes that are catalyzing NADPH-driven NOHA oxidation, as judged by buildup of a ferrous heme-NO complex during catalysis (23,26 This remains an open question. Our current data indicate that the electron transfer needs to occur between 100 and 300 ms (in a reaction run at 10°C) to be associated with the observed reduction of the H 4 B radical. As has been discussed previously in detail (12,20,21,23), there are at least three putative intermediates that could serve as the electron donor. Both we (23) and others (12,21)  heme-NO complexes are similar, it is conceivable that oxidation of a putative ferrous heme-NO complex to generate the observed ferric heme-NO complex might occur during the reaction without our detecting it by stopped-flow spectral analysis. Although the NOS ferrous heme-NO complex exhibits a unique EPR signal (24), we did not observe buildup of this particular signal in our stopped-freeze EPR samples that were collected within the appropriate time range (50 -200 ms, data not shown). Thus, at present we have no data to discern which enzyme species reduces the H 4 B radical during the NOHA reaction. This interesting problem requires further study.
Relationship to Other Reports-Our findings differ from two earlier studies that reported little or no H 4 B radical buildup during a NOHA single turnover reaction (11,12). We suspect that this discrepancy may reflect the inherent difficulty in detecting the transient H 4 B radical during the NOHA reaction. In any case, it is useful to consider the details. In one study buildup of 0.03 H 4 B radical per heme was observed (11), and its rate of formation and decay were 15 and 0.2 s Ϫ1 , respectively. These results differ from ours both in magnitude of the radical and its kinetic signature. In our experiments we rapidly mixed ferrous iNOSoxy containing H 4 B and NOHA with oxygenated buffer to initiate the reaction. In contrast, they rapidly mixed an H 4 B-saturated ferrous iNOSoxy with an oxygenated buffer that contained NOHA. This procedure necessitates that substrate bind during the mixing step and so may increase sample heterogeneity. However, when we repeated their experimental method in our laboratory we observed H 4 B radical formation and disappearance rates that matched what we have reported here, despite a somewhat lower yield of H 4 B radical per heme. Thus, the basis for the discrepancy is unclear.
Another study reported no H 4 B radical buildup during NOHA oxidation (12). However, in this case the reactions were run in 50% polyethylene glycol at cryogenic temperatures, initiated by manual mixing of an oxygenated solution, and followed by manual sample collection and freezing. Thus, our different results could be due to the use of a cryosolvent and to the poor time resolution inherent in the mixing and sample collection. Related studies show that this particular cryogenic system bestows unique properties to NOS enzymes, as manifested by differences in spectral properties of the heme-dioxy intermediate, relative stabilities of other heme-oxy species, yields of Arg hydroxylation in single turnover reactions, and yields of H 4 B radical in reactions run with substrate-free enzymes (27)(28)(29). Therefore, meaningful comparisons may not be possible in this case.
Summary-We provide evidence of a new redox function for H 4 B in biology. H 4 B (or 5MeH 4 B) underwent two consecutive one-electron transitions during NOHA oxidation catalyzed by iNOSoxy. The initial transition generated the pterin radical. This was kinetically linked to formation of a heme-based oxidant that quickly reacted with NOHA and so likely represents reduction of the enzyme Fe III O 2 Ϫ intermediate. A subsequent one-electron transfer reformed H 4 B within the enzyme. This step was kinetically linked to appearance of a ferric heme-NO product complex, which upon dissociation released NO. Thus, both redox transitions of H 4 B occurred within the timeframe of the single turnover catalysis and could be ascribed to different steps in the NOHA oxidation reaction mechanism. The apparent dual redox function of H 4 B establishes its common role in O 2 activation by NOS and may also explain how the enzyme generates its free radical product (NO) via a three-electron oxidation of NOHA.