Oxidation and Nitrosation in the Nitrogen Monoxide/Superoxide System*

Based on the previous report of McCord and co-workers (Crow, J. P., Beckman, J. S., and McCord, J. M. (1995) Biochemistry 34, 3544–3552), the zinc dithiolate active site of alcohol dehydrogenase (ADH) has been studied as a target for cellular oxidants. In the nitrogen monoxide (⋅NO)/superoxide (O 2 ⨪ ) system, an equimolar generation of both radicals under peroxynitrite (PN) formation led to rapid inactivation of ADH activity, whereas hydrogen peroxide and ⋅NO alone reacted too slowly to be of physiological significance. 3-Morpholino sydnonimine inactivated the enzyme with an IC50 value of 250 nm; the corresponding values for PN, hydrogen peroxide, and⋅NO were 500 nm, 50 μm, and 200 μm. When superoxide was generated at low fluxes by xanthine oxidase, it was quite effective in ADH inactivation (IC50 (XO) ≈ 1 milliunit/ml). All inactivations were accompanied by zinc release and disulfide formation, although no strict correlation was observed. From the two zinc thiolate centers, only the zinc Cys2His center released the metal by oxidants. The zinc Cys4 center was also oxidized, but no second zinc atom could be found with 4-(2-pyridylazo)resorcinol (PAR) as a chelating agent except under denaturing conditions. Surprisingly, the oxidative actions of PN were abolished by a 2–3-fold excess of ⋅NO under generation of a nitrosating species, probably dinitrogen trioxide. We conclude that in cellular systems, low fluxes of ⋅NO and O 2 ⨪ generate peroxynitrite at levels effective for zinc thiolate oxidations, facilitated by the nucleophilic nature of the complexed thiolate group. With an excess of ⋅NO, the PN actions are blocked, which may explain the antioxidant properties of⋅NO and the mechanism of cellularS-nitrosations.

been used to study such oxidations, but in most cases, the effective concentrations are nonphysiological and can only partly reflect the physiological steady-state levels of oxidants. We have recently obtained evidence that PN can control vascular tone already at submicromolar concentrations by nitration of a tyrosine residue in prostacyclin synthase under physiological conditions (8). Simultaneous low-level generation of nitrogen monoxide ( ⅐ NO) and superoxide radicals (O 2 . ) in endothelial cells proved to be sufficient for this oxidative modification (9,10), although it has been argued that physiological levels of both radicals are not sufficient for tyrosine nitration (11,12). Meanwhile, literature data show that PN generated from superoxide and nitrogen monoxide can nitrate tyrosine (13)(14)(15). Moreover, this discrepancy found an explanation in a series of studies proving that the sensitivity and selectivity of heme proteins for tyrosine nitrations reside in an autocatalytic role of the metal ion (16,17). By forming transition metal intermediates from cellular levels of PN, neighboring tyrosines can be selectively attacked without affecting other tyrosines randomly.
Following this concept of PN action under physiological conditions, we now extend oxidations by PN also to cysteines. Free and also protein-bound thiols are preferred targets of PN, especially when they are present in their thiolate forms (18 -22). Then a direct two-electron oxidation of the sulfur by the electrophilic PN molecule can take place. It has been reported that the essential ZnCys 2 His center of alcohol dehydrogenase (ADH) can be oxidized by micromolar bolus concentrations of PN (23) and also by ⅐ NO (24) and hypochlorite (25). This can be interpreted as a catalytic action of Zn 2ϩ by converting thiol groups after deprotonation to the more reactive thiolate form. Such zinc thiolate centers may be the subject of a redox regulation by being reversibly oxidized and reduced as part of a signaling mechanism in oxidative stress (26 -28). Therefore, ADH provides a model in which the oxidation of the active site Zn 2ϩ dithiolate center can be studied in the ⅐ NO/O 2 . system by monitoring the enzyme activity. In the present work, we confirm the validity of this model and report on chemically surprising and biologically relevant changes in the oxidizing, nitrating, and nitrosating potential of the system when varying the relative flux rates of ⅐ NO and O 2 . .

EXPERIMENTAL PROCEDURES
Materials-PN was synthesized from ⅐ NO and O 2 . according to the method of Kissner et al. (29). SIN-1 and spermine NONOate were * This work was supported by the Deutsche Forschungsgemeinschaft, the Fonds der Chemischen Industrie, and the Boehringer Ingelheim-Stiftung. 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  ADH Activity Test-The activity of ADH was measured on an Aminco DW-2a UV/Vis spectrophotometer in the dual wavelength mode. A solution of ADH (33 nM) in 0.1 M potassium phosphate buffer, pH 7.4, was placed in a cuvette equipped with a magnetic stirrer, and then NAD ϩ (300 M) was added, and the temperature of the solution was adjusted to 37°C. The reduction of NAD ϩ to NADH was started by addition of ethanol (172 mM) from a syringe through a septum to the stirred solution. The formation of NADH was followed at 340 versus 400 nm, and the specific activity of ADH was calculated for the first 30 s of the measurements in the linear region of the curves by using an extinction coefficient for NADH of ⑀ 340 ϭ 6220 M Ϫ1 cm Ϫ1 (23). The specific activity of freshly prepared ADH was 65 Ϯ 3 mol min Ϫ1 mg Ϫ1 (units/ mg). For all experiments, the same stock solutions of ADH (3.3 M in water) and NAD ϩ (3 mM in buffer) were used, aliquots of which were frozen at Ϫ70°C.
Incubation of ADH with Peroxynitrite, Hydrogen Peroxide, or Potassium Superoxide-To test the effects of oxidants on the activity of ADH, the enzyme (33 nM) without NAD ϩ was preincubated with these oxidants at pH 7.4 and 37°C. PN (0.1-10 M) was added by a Vortex mixer to a solution of ADH and incubated for 5 min at pH 7.4 and 37°C, and then all samples were kept on ice until they were measured. H 2 O 2 (25-1000 M) was incubated with ADH for 1 or 2 h at pH 7.4 and 37°C, and aliquots were taken after defined time intervals. The effects of catalase (5-500 units/ml) and Cu,Zn-SOD (100 and 500 units/ml) on this inactivation by H 2 O 2 (0.5 or 1 mM) were measured in the presence of these proteins during incubation. Potassium superoxide (KO 2 ) (2.5-100 M from a freshly prepared stock solution in Me 2 SO) was incubated for 1 min before the activity test. The final amount of Me 2 SO was kept below 2% to prevent activity loss by this compound. were determined by the cytochrome c/oxy-Hb assay as described previously (34). For any series of measurements, an internal control was carried out, and ADH was incubated under similar conditions as indicated for each experiment, but the oxidants were omitted. All values refer to this internal control as "percentage of control." Before the activity measurement, 10% of a NAD ϩ (3.3 mM) solution was added to the samples containing ADH at a concentration of 33 nM, yielding final concentrations of 300 M NAD ϩ and 30 nM ADH in the activity assay.
Determination of Zinc Release-The zinc release from ADH (10 M) after incubation with different oxidants was measured by using the zinc chelator 4-(2-pyridylazo)resorcinol (PAR) with an extinction coefficient of ⑀ 500 ϭ 90,000 M Ϫ1 cm Ϫ1 at pH 7.4 that was determined from a calibration curve. This value was between those determined at pH 7 (65,000 M Ϫ1 cm Ϫ1 ) (35) and pH 7.8 (96,000 M Ϫ1 cm Ϫ1 ) (23). PAR was present at a concentration of 100 M and was added immediately before the measurement because PAR may slowly extract zinc from ADH (23).
The protein was preincubated with SIN-1, H 2 O 2 , XO/hypoxanthine, and/or spermine NONOate for 60 -150 min in 0.1 M potassium phosphate buffer at pH 7.4 and 37°C. For the time-dependent zinc release, aliquots were taken every 10 -20 min. The protective effect of catalase and Cu,Zn-SOD was also assayed in the above-described oxidant systems. A maximal release of 1.96 zinc atoms/ADH monomer was found after acidic hydrolysis (6 N HCl at 100°C for 12 h), in good agreement with the theoretical value of 2 zinc atoms/subunit. The pH was adjusted to 7.4 before measurements of zinc release by PAR. A value of 1.9 Ϯ 0.05 mol of zinc was measured by inductively coupled plasma mass spectrometry (Spurenanalytisches Labor; Dr. Baumann, Maxhü tte, Germany).
Determination of Thiol Oxidation-The thiol content of ADH (10 M) was measured after oxidation by different oxidants by using the disulfide compound 2,2Ј-dithio-bipyridine (DTBP) at a concentration of 200 M. Free SH groups were determined spectrophotometrically at 343 versus 400 nm by monitoring the release of 2-thiopyridinone (⑀ 343 ϭ 7600 M Ϫ1 cm Ϫ1 ) (36). DTBP was used because its self-hydrolysis is slower compared with Ellman's reagent. However, because DTBP reacts only slowly with the thiol groups of ADH, the protein was denatured anaerobically for 5 min at 90°C in the presence of 10% acetonitrile before the addition of DTBP and the spectroscopic determination. This denaturation procedure allowed a complete reaction of SH groups within 100 s. In untreated control samples, 5.7 Ϯ 0.5 free thiol groups/ ADH subunit could be found, although the theoretical value corresponds to 7 cysteines/monomer (6 are ligands for the 2 zinc atoms as published in the Swiss Protein Database, accession number P00330) (23). Aliquots were taken and measured after time intervals of 10 min or after the total incubation time in the presence of different oxidant concentrations. All values are given as a percentage of the control (5.7 thiols/subunit). In the case of high PN concentrations (up to 10 mM), the expected pH shift of 0.1-0.3 was adjusted by prior addition of 0.1 M HCl to the enzyme solution, yielding a pH of 7.1-7.3 before the PN addition. The reactivity of PAR and DTNB in samples with high initial hydrogen peroxide concentrations was checked by the addition of ZnSO 4 or cysteine, respectively.
Detection of Disulfide Bonds in ADH by Raman Spectroscopy-ADH (1 mM) in 0.2 M potassium phosphate buffer at pH 7.4 was incubated for 2 h at 37°C with SIN-1 (1 mM) or H 2 O 2 (50 mM) or was quickly mixed with PN (10 mM). The solubilized protein was partially precipitated after this treatment. The samples were then buffer-exchanged into 50 mM NaCl, pH 7.0, on Sephadex G-25 fast desalting column (Amersham Biosciences, Inc.) and concentrated to 50 l in Microcon-10 centrifuge tubes (Amicon). 10-l aliquots were spotted onto glass coverslips, dried under vacuum, and stored in argon atmosphere. Raman spectra were taken on a Jobin Yvon LabRAM spectrometer (courtesy of Jobin Yvon; GmbH, Bensheim, Germany) with 633 nm excitation from a heliumneon laser.
Quantitation of 4-Nitrosophenol by UV/Vis Spectroscopy and HPLC-The formation of 4-nitrosophenolate (pK Ϸ 6) in reactions of phenol (5 mM) with SIN-1 (200 M) and/or spermine NONOate (100 M) was monitored spectrophotometrically at 400 versus 500 nm over a time period of 2 h in 0.1 M potassium phosphate buffer at pH 7.4 and 37°C (37). Its yield in these samples was determined by HPLC as described previously (38).

Inhibition of ADH Activity by Nitrogen Monoxide, Bolus Additions, and in Situ
Generation of Peroxynitrite-In ADH, the ZnCys 2 His cluster is essential for enzyme activity (39), allowing correlation of the loss of activity with the oxidation by PN (Fig. 1). By bolus additions of PN, ADH became inactivated with IC 50 values of 0.5 M PN and thus with similar sensitivity as reported previously (23) because such values are dependent on the target concentrations. This inactivation was complete at about 4 M within 1 min, in agreement with the short half-life of PN of about 1 s at pH 7 (40).
Because under cellular conditions PN originates from the fast combination of ⅐ NO and O 2 . (29, 41), we used SIN-1 as a suitable agent for the simultaneous generation of both radicals in equimolar concentrations (11,30). According to the clear-cut results in Fig. 2, the inhibition occurred with an IC 50 value of about 0.25 M and hence perfectly matched the data from bolus additions of PN. The presence of Cu,Zn-SOD abolished the inhibition (Fig. 2, inset), and ⅐ NO alone generated by spermine . by the XO/hypoxanthine system provided a better control (Fig. 4). Such conditions caused inactivation in a time-dependent (Fig. 4, inset) and XO concentration-dependent fashion (Fig. 4). In the presence of Cu,Zn-SOD, almost 40% of the enzyme could be prevented from inactivation, and in the presence of catalase, close to 50% of the enzyme could be prevented from inactivation. The addition of both together resulted in almost 100% restoration of activity. This allows the conclusion that O 2 . , at least under constant generation, is an inactivating agent by itself. H 2 O 2 as an oxidant requires quite high concentrations or long incubation times to exert its effects, and its concentration for half-maximal inhibition after 2 h was around 25-50 M (Fig. 6). As expected, catalase already blocked the inactivation at low levels (Fig. 6, inset), whereas Cu,Zn-SOD did not (data not shown).  . and ⅐ NO production. The initial formation rates in the XO system (0.5 milliunit/ml and 1 mM hypoxanthine) and from spermine NONOate (5 M) were 3 Ϯ 0.2 nM/s for O 2 . and 2.4 Ϯ 0.1 nM/s for ⅐ NO at pH 7.4 and 37°C. Cu,Zn-SOD at concentrations above 100 units/ml could block the inhibition almost completely at this minimum (Fig. 7B). At concentrations of spermine NONOate above those necessary for equalizing O 2 . production, the inhibition of ADH was effectively released at about 15-20 M NONOate, whereas at higher concentrations, the inhibition started again, in agreement with the low oxidative action of ⅐ NO from spermine NONOate alone (see Fig. 3). Such data could be interpreted as a scavenging action of PN by ⅐ NO, which we tested in a second system with SIN-1 as a PN source (Fig. 8).
Using a sufficient concentration of SIN-1 for ADH inhibition (5 M; see Fig. 2), a supplementation of the system with spermine NONOate also released the inhibition with effective concentrations above 10 M. According to literature data, the reaction of PN with ⅐ NO is complicated and leads to a nitrosating intermediate (43). To document its formation, ADH was replaced by phenol and incubated with either SIN-1, spermine NONOate, or both, and the kinetics of 4-nitrosophenol formation were followed spectrophotometrically. In addition, the yields were determined by HPLC. Spermine NONOate 2 (100 M) and SIN-1 (200 M) produced about the same amounts of 4-nitrosophenol, whereas the addition of both compounds together exceeded the individual systems by 250% (Table I) 2 One mol of spermine NONOate will produce 2 mol of ⅐ NO by thermal decomposition.

Oxidations in the ⅐ NO/O 2 . System
previous work on PN action had shown a release of Zn 2ϩ as a consequence of oxidation of the zinc dithiolate center in ADH (23). Because the release of zinc can easily be followed by complexation with the chelator PAR, we have monitored the formation of the Zn(PAR) 2 complex spectrophotometrically between 450 and 550 nm. The time-dependent release of Zn 2ϩ under the influence of SIN-1, H 2 O 2 , and/or spermine NONOate is shown in Fig. 9. It is evident that ⅐ NO is ineffective in zinc release, H 2 O 2 is still a sluggish oxidant, superoxide is quite effective, and SIN-1, at a concentration of 250 and 1000 M, is highly effective. The efficacy of the ⅐ NO/O 2 . system was close to that of SIN-1. These data qualitatively follow the inactivation results and even match quantitatively for SIN-1, if one takes into account the 300-fold higher ADH concentrations in these samples (10 M instead of 30 nM) and the fact that for PN the target concentration directly influences the IC 50 values. The results obtained from additions of Cu,Zn-SOD and/or catalase allowed us to identify PN as the most effective zinc-releasing oxidant (data not shown).
To obtain a more direct comparison between zinc release and activity loss, one has to take into account the presence of a second zinc thiolate center (ZnCys 4 ), which is not involved in catalysis but may have structural significance (23,39). Analysis of the zinc content by inductively coupled plasma mass spectrometry resulted in 2 zinc atoms/ADH subunit, and hydrolysis with hydrochloric acid and subsequent PAR complexation also gave the same value (1.96 zinc atoms/subunit). However, we could not observe the direct extraction of zinc from the ZnCys 4 cluster by PAR as previously reported (23). This second zinc atom, which is not related to the enzyme activity, even required drastic conditions to be released. Based on these values of 2 zinc atoms/subunit, one observes the release of one mol of Zn 2ϩ upon titration with SIN-1 ( Fig. 10; see also Fig. 9). This maximal zinc release was already achieved by the addition of 300 M SIN-1. Higher concentrations of SIN-1 did not result in increased zinc release but rather decreased it (data not shown). H 2 O 2 had to be used at a concentration of 10 mM to inactivate ADH (10 M) completely, but it only caused the release of 0.5 zinc atom/subunit, whereas PN (10 mM) caused the release of 1.4 zinc atoms/ADH monomer (data not shown). Although the two parameters change almost inversely, the inactivation process seems to be somewhat faster than the zinc release. This may either be due to an additional inactivation mechanism (affecting about 20% of the zinc release) or because the zinc release does not exactly follow the oxidation of the two thiolate residues. The main mechanism, however, consists of the oxidation of the thiolate ligands because at the maximum zinc release (250 M SIN-1), no thiol groups are left (data not shown). Further evidence for this assumption provided the time-dependent inactivation of ADH by the thiol group determination compound DTBP, which was accompanied by zinc release (data not shown). Half oxidation of the thiol groups of ADH (10 M) was observed at around 25 M SIN-1, 5 mM H 2 O 2 , or 7.5 milliunits/ml XO. 500 M NONOate caused only 27% of thiol oxidation, thus confirming that the ZnCys 2 His center resists nitrosation and oxidation by ⅐ NO.
Direct Detection of Disulfide Formation in ADH by Raman Spectroscopy-For a final confirmation of the oxidation processes, Raman spectra were recorded since the disulfide stretching modes are relatively strong around 500 -540 cm Ϫ1 (45,46). After treatment with SIN-1, the appearance of two bands could be clearly observed, whereas in the control, these modes were absent (Fig. 11). After treatment with H 2 O 2 and PN, these disulfide stretching bands could also be observed (data not shown). In all samples, no SH frequencies were noticed, which is in line with the fact that a total of 6 of 7 thiol groups are present as thiolate ligands of zinc in the ZnCys 2 His and ZnCys 4 centers. The residual SH group may be too weak in absorption or could have been oxidized in our sample because we had found fewer thiols than theoretically predicted.   (16,17). By analogy, we postulate here that the oxidation of thiol groups by PN can be greatly enhanced by Zn 2ϩ , which converts thiols to the more nucleophilic thiolate anions. ADH is a suitable example because its ZnCys 2 His center is required for activity and can serve as a target for PN. Inactivation of the enzyme occurred with an IC 50 value of 0.25-0.5 M PN, regardless of whether a bolus addition was used or a steady-state was generated. Such concentrations, however, are dependent on the concentration of the target or on the presence of other potential reaction partners; therefore, discussions on a possible physiological relevance have to await additional studies. Our results have some additional implications. The oxidative release of 1 zinc atom/ADH monomer could be shown, although 2 zinc atoms are present in the ADH subunit according to inductively coupled plasma mass spectrometry analysis. The second Zn 2ϩ could be partially released by a large excess of PN (1.4 zinc atoms/subunit) or completely released after protein hydrolysis (1.96 zinc atoms/subunit). According to the literature, in the native state, the second Zn 2ϩ is bound to four other Cys residues (23), but these were not detectable after oxidative release of the first zinc from the active site. It remains a possibility that oxidation of this ZnCys 4 center moved the zinc to another binding site with sufficiently high affinity to resist complexation by PAR. Further evidence for this came from the observation that SIN-1 at a concentration of 25 M already oxidized 50% of the thiols but only released 0.3 zinc atom/ADH subunit. Qualitatively, the same results were found with H 2 O 2 , which oxidized 50% of the thiols at a concentration of 5 mM, but only 0.2 zinc atom/ADH subunit was released under these conditions. Moreover, the direct formation of disulfide bonds could be monitored by Raman spectroscopy. For the oxidation of zinc finger motifs lacking a direct connection to activities, this might be a convenient method for the detection of oxidations at these potential targets for redox regulation.
Except for PN, other reactive oxygen and nitrogen species were examined but gave very different results. H 2 O 2 turned out to be a slow oxidant and certainly per se is not able to cause inactivation under cellular conditions. Also, ⅐ NO could not inactivate ADH, although the formation of S-nitroso compounds by ⅐ NO and even the reaction with zinc fingers have been reported (24). On the other hand, O 2 . , which is not considered a strong oxidant and even has reducing properties, was quite sufficient for inactivation. In all cases, the loss of activity was paralleled by Zn 2ϩ release and thiol oxidation, which allowed us to exclude mechanisms other than oxidation of the It should be pointed out that conflicting data in the literature on the use of SIN-1 as a PN-generating agent can be explained by this finding. If, in cellular systems, trapping of superoxide 3 A. Daiber, unpublished observations. h. An aliquot of each sample was taken before the PAR measurements and diluted 300-fold, and then the activity of ADH (30 nM, final concentration) was determined q, Zn release; OE, activity.
FIG. 11. Direct detection of disulfide bonds by Raman spectroscopy in ADH after incubation with SIN-1. Raman spectra are shown for ADH after treatment with SIN-1 and for the untreated control. The samples were evaporated to dryness before the measurements. The spectrum of the SIN-1 sample shows the typical disulfide stretching bands between 500 and 540 cm Ϫ1 . The thiol stretching band of thiol groups at 2500 -2600 cm Ϫ1 was visible in neither the oxidized sample nor the untreated control.

Oxidations in the ⅐ NO/O 2 . System
by NO-synthase, SODs or antioxidants shifts the ⅐ NO/O 2 . ratios to values greater than 2 or 3, then the actions of PN are eliminated. A limitation in the oxygen supply may also have the same effect. This is likely to happen if SIN-1 concentrations close to or greater than the dioxygen levels are used. SIN-1 should not be used at concentrations greater than 300 M in air-saturated solutions because it could also switch to a ⅐ NO donor under anaerobic conditions. An excess of ⅐ NO could not only eliminate PN actions but even prevail nitrosating conditions. In vivo, Snitrosations could be the consequence, which, in view of the upcoming regulatory properties of S-nitroso derivatives, would be an attractive new property of the ⅐ NO/O 2 . regulatory system.
In summary, the oxidation of the Zn 2ϩ thiolate cluster in ADH not only serves as a suitable model for redox regulations involving zinc finger proteins but also unravels new chemical properties of the ⅐ NO/O 2 . system. One clearly has to take into consideration the participation of metal catalysis in PN action at physiological levels and the fine tuning of the relative rates of ⅐ NO and O 2 . fluxes that lead to sudden changes in the nature of reactive intermediates. As outlined recently, the different sources for O 2 . , their compartmentalization, and the presence of antioxidants and SODs will be as important as the different sources for ⅐ NO and their regulatory factors. Thus, the ⅐ NO/O system turns out to be a major, but not easily handled, player in the cellular signaling network.