Dityrosine formation outcompetes tyrosine nitration at low steady-state concentrations of peroxynitrite. Implications for tyrosine modification by nitric oxide/superoxide in vivo.

Formation of peroxynitrite from NO and O-(*2) is considered an important trigger for cellular tyrosine nitration under pathophysiological conditions. However, this view has been questioned by a recent report indicating that NO and O-(*2) generated simultaneously from (Z)-1-(N-[3-aminopropyl]-N-[4-(3-aminopropylammonio)butyl]-amino) diazen-1-ium-1,2-diolate] (SPER/NO) and hypoxanthine/xanthine oxidase, respectively, exhibit much lower nitrating efficiency than authentic peroxynitrite (Pfeiffer, S. and Mayer, B. (1998) J. Biol. Chem. 273, 27280-27285). The present study extends those earlier findings to several alternative NO/O-(*2)-generating systems and provides evidence that the apparent lack of tyrosine nitration by NO/O-(*2) is due to a pronounced decrease of nitration efficiency at low steady-state concentrations of authentic peroxynitrite. The decrease in the yields of 3-nitrotyrosine was accompanied by an increase in the recovery of dityrosine, showing that dimerization of tyrosine radicals outcompetes the nitration reaction at low peroxynitrite concentrations. The observed inverse dependence on peroxynitrite concentration of dityrosine formation and tyrosine nitration is predicted by a kinetic model assuming that radical formation by peroxynitrous acid homolysis results in the generation of tyrosyl radicals that either dimerize to yield dityrosine or combine with (*)NO(2) radical to form 3-nitrotyrosine. The present results demonstrate that very high fluxes (>2 microM/s) of NO/O-(*2) are required to render peroxynitrite an efficient trigger of tyrosine nitration and that dityrosine is a major product of tyrosine modification caused by low steady-state concentrations of peroxynitrite.

crease of nitration efficiency at low steady-state concentrations of authentic peroxynitrite. The decrease in the yields of 3-nitrotyrosine was accompanied by an increase in the recovery of dityrosine, showing that dimerization of tyrosine radicals outcompetes the nitration reaction at low peroxynitrite concentrations. The observed inverse dependence on peroxynitrite concentration of dityrosine formation and tyrosine nitration is predicted by a kinetic model assuming that radical formation by peroxynitrous acid homolysis results in the generation of tyrosyl radicals that either dimerize to yield dityrosine or combine with ⅐ NO 2 radical to form 3-nitrotyrosine. The present results demonstrate that very high fluxes (>2 M/s) of NO/O 2 . are required to render peroxynitrite an efficient trigger of tyrosine nitration and that dityrosine is a major product of tyrosine modification caused by low steady-state concentrations of peroxynitrite.
Tyrosine nitration is a well established protein modification occurring in vivo in a number of inflammatory diseases associated with oxidative stress and increased activity of NO synthases (1,2). Nitration of specific tyrosine residues has been reported to affect protein structure and function (3), suggesting that 3-nitrotyrosine formation may not only be a disease marker but could be causally involved in the pathogenesis of certain disease states.
Peroxynitrite, formed in a nearly diffusion-controlled reaction from NO and O 2 . , is considered a potent pathophysiologi-cally relevant cytotoxin. Besides oxidation reactions resulting in dysfunction of various biomolecules, nitration of free and protein-bound tyrosine to yield 3-nitrotyrosine is a well established reaction of peroxynitrite that may contribute to NO cytotoxicity (1). The nitration reaction has been extensively studied in vitro by bolus addition of synthetic peroxynitrite to tyrosine-containing samples including purified proteins, cells, and tissues (3)(4)(5)(6) . In situ, 3-nitrotyrosine was most frequently visualized with monoclonal or polyclonal antibodies (2), but the identity of the product has been confirmed by several laboratories using sophisticated gas chromatography/mass spectroscopy and HPLC 1 methods (7,8).
Thus, there is general agreement that (i) authentic peroxynitrite is a potent nitrating agent that converts free and proteinbound tyrosine to the corresponding 3-nitro derivative, and that (ii) 3-nitrotyrosine does occur in vivo. The conclusion that peroxynitrite is the main cause for in vivo nitration may thus seem obvious, but is not supported by experimental data. In fact, several recent studies have identified alternative pathways of tyrosine nitration (9), and we found that nitration by simultaneously generated NO and O 2 . is much less efficient than the reaction triggered by authentic peroxynitrite (10). The interpretation of the latter results has been disputed, and a number of points have been raised questioning their validity. One point was related to the possibility that urate formed in the XO reaction might have scavenged peroxynitrite and thus prevented tyrosine nitration in long term ( Solutions-All solutions were prepared freshly each day. Water was from a Milli-Q reagent water system from Millipore (Vienna, Austria; resistance Ն18 megaohms ϫ cm Ϫ1 ). SPER/NO and DEA/NO were prepared as 10-fold stock solutions in 10 mM NaOH. DHR was dissolved to 10 mM in acetonitrile and kept in the dark until use. Alkaline solutions of peroxynitrite were prepared from acidified NO 2 Ϫ and H 2 O 2 as described (14). The solutions were diluted with H 2 O to 10 mM (pH ϳ12.8) and further diluted in 10 mM NaOH to 10- Oxidation of DHR-Oxidation of DHR was monitored at 501 nm as described (10,18). The amount of oxidized DHR was calculated using an extinction coefficient of 78.78 mM Ϫ1 cm Ϫ1 . For measurements, 200-l aliquots were taken every 10 -20 min from 3-ml samples. Total incubation time was 3 h. SPER/NO (1 mM) or hypoxanthine/XO (28 milliunits/ ml) alone led to DHR oxidation rates of Ͻ0.06 and Ͻ0.3 M ϫ min Ϫ1 , respectively.
Peroxynitrite Infusion-The infusion experiments were performed with a Merck-Hitachi HPLC pump (655A-11) provided with Peek capillaries (internal diameter, 0.25 mm) under constant stirring of the tyrosine-containing solutions at ambient temperature. Peroxynitrite (2 ml of a 0.1 mM stock solution) was infused at increasing rates (0.1, 0.2, 0.4, 0.5, and 0.8 ml/min) into 18 ml of 0.1 M K 2 HPO 4 /KH 2 PO 4 buffer (pH 7.4) containing 1 mM tyrosine, followed by the determination of 3-nitrotyrosine as described below.
Determination of 3-Nitrotyrosine and Dityrosine-HPLC analysis of 3-nitrotyrosine was performed on a C 18 reversed phase column with 0.1 M KH 2 PO 4 /H 3 PO 4 buffer (pH 3) containing 6% (v/v) methanol at 0.7 ml/min and detection at 274 nm, as described (19). In some experiments 3-nitrotyrosine was detected with a dual-channel electrochemical detector (ESA, Coulochem II, Chelmsford, MA) set to 600 mV and 850 mV (20). Oxidation of 3-nitrotyrosine was followed at 850 mV. A guard cell placed between the solvent delivery system and injector was set to 1000 mV. Calibration curves were recorded daily with authentic 3-nitrotyrosine (2 nM-0.5 M and 60 nM--5 M for electrochemical and UV-visible detection, respectively). HPLC analysis of dityrosine was performed on a C 18 reversed phase column with 50 mM KH 2 PO 4 /H 3 PO 4 buffer (pH 3) containing 1% (v/v) methanol at 0.7 ml/min and fluorescence detection (Hitachi fluorescence spectrophotometer F 1050; excitation 285 nm, emission 410 nm) as described (4). Calibration curves were recorded with authentic dityrosine (50 nM-5 M).
Kinetic Experiments-The rate of peroxynitrite decay was determined by stopped-flow absorbance spectroscopy at 302 nm (Bio-Sequential SX-17MV stopped-flow spectrofluorimeter, Applied Photophysics, Leatherhead, UK) at 22°C. Reservoir 1 contained peroxynitrite (0.2 mM) in 0.01 M NaOH, and reservoir 2, the buffer solution (0.2 M K 2 HPO 4 /KH 2 PO 4 buffer (pH 7.4), containing 2 mM tyrosine). A k av value of 0.27 Ϯ 0.05 s Ϫ1 (mean Ϯ S.D.; n ϭ 9) was calculated from the initial rates of first order peroxynitrite decay. The peroxynitrite steady-state concentrations obtained in the infusion experiments were calculated by dividing infusion rates (nM s Ϫ1 ) by 0.27 s Ϫ1 .

RESULTS
The formation of 3-nitrotyrosine was measured in the presence of four different NO/O 2 . -generating systems. As shown in   (28). Allantoin (0.1 mM) had no effect on tyrosine nitration mediated by authentic peroxynitrite (data not shown). In the presence of uricase, the amount of detectable peroxynitrite was approximately doubled, accompanied by a 6-fold increase in 3-nitrotyrosine formation. The corresponding nitrating efficiency was 0.54%. These results suggested that accumulation of urate does indeed contribute to the low nitrating efficiency of the applied NO/O 2 .
-generating system, but the nitration yield was still about 10-fold lower than that obtained with 70 M authentic peroxynitrite (3.41 Ϯ 0.63 M, corresponding to 4.9%) and about 5-fold lower than the nitration triggered with SPER/NO alone. It was conceivable that this difference was caused by residual urate, because uricase did not completely consume the accumulated urate under our experimental conditions (data not shown). Therefore, two urate-free NO/O 2 .
-generating systems were additionally tested.
Acetaldehyde is known to function as an alternative substrate of XO albeit at much lower turnover numbers (29) Incubation of DEA/NO (0.1 mM) with 1 mM tyrosine in the absence and presence of 0.1 mM FMN led to formation of 86.1 Ϯ 29.5 nM and 678 Ϯ 90.8 nM 3-nitrotyrosine, respectively. The effect of FMN was concentration-dependent; maximal effects were obtained with Ն0.1 mM, the apparent EC 50 was 57.5 Ϯ 7.0 M (Fig. 1B). FMN did not significantly increase DEA/NOmediated nitration in dark conditions (data not shown). Due to a strong interference of FMN with the DHR assay, 3 it was not possible to measure apparent peroxynitrite formation by this system, but the DEA/NO-FMN system allowed us for the first time to demonstrate a stimulation of NO-mediated nitration by co-generation of O 2 . , indicating that the in situ generation of peroxynitrite does lead to tyrosine nitration under certain experimental conditions. Intriguing data were obtained when the nitrating efficiencies of NO/O 2 . -generating systems were studied in the presence of bicarbonate (CO 2 . ). CO 2 . is known to react rapidly with peroxynitrite to yield the potent nitrating adduct nitrosoperoxycarbonate (ONO 2 CO 2 Ϫ ) (33). Therefore, depending on the buffer concentrations of CO 2 (34), tyrosine nitration by authentic peroxynitrite is increased 2-to 4-fold upon the addition of 0.25-50 mM bicarbonate (5,(35)(36)(37). The data obtained with four NO/O 2 . -generating systems tested for tyrosine nitration with and without 25 mM bicarbonate (Fig. 2) clearly demonstrated that CO 2 had no effect whatsoever on nitration by NO  We considered several possibilities to explain the poor nitrating efficiency of NO/O 2 . . Unfortunately, however, most hypotheses, including the proposal of a distinct chemical species that is formed from NO/O 2 . in situ (10), are in conflict with the known theoretical background of NO/O 2 . and/or peroxynitrite chemistry. One remaining possibility was that tyrosine nitration required a certain threshold steady-state level of peroxynitrite to become significant. This would explain the observed differences between bolus addition and continuous generation of peroxynitrite. We have addressed this issue using two experimental approaches. First we studied the nitrating efficiency of increasing peroxynitrite concentrations (5-1,000 M) added as a bolus to buffer solutions containing 1 mM tyrosine. As expected, the total amount of 3-nitrotyrosine gradually increased with increasing concentrations of added peroxynitrite (46.23 Ϯ 1.59 M at 1 mM; data not shown). It was surprising, however, to find that the nitrating efficiency of peroxynitrite increased from 1.4 Ϯ 0.3 to 5.4 Ϯ 0.4% when the peroxynitrite concentration was increased from 5 M to 100 M and leveled off at higher concentrations (Fig. 3A). In another set of experiments 2 ml of a 0.1 mM stock solution of peroxynitrite was infused at increasing rates (8.33 nM s Ϫ1 , 16.67 nM s Ϫ1 , 33.33 nM s Ϫ1 , 41.67 nM s Ϫ1 , and 66.67 nM s Ϫ1 ) into tyrosine-containing buffer solutions (10 M peroxynitrite final in each case). The respective steady-state concentrations of peroxynitrite were calculated from the rate of first order decomposition measured by stoppedflow absorbance spectroscopy under identical conditions (k av ϭ 0.27 Ϯ 0.05 s Ϫ1 ; data not shown). Fig. 3B shows that the nitrating efficiency of infused peroxynitrite increased about 3-fold (from 0.22 Ϯ 0.05 to 0.64 Ϯ 0.08%) when the steady-state concentrations were increased from 30.7 to 247 nM.
Since dityrosine is another product of the reaction between tyrosine and peroxynitrite (4,35,36), we speculated that the tyrosine dimerization reaction may be predominant at low peroxynitrite concentrations. To test this hypothesis, we measured dityrosine formation from 1 mM tyrosine treated with increasing concentrations of peroxynitrite. As shown in Fig. 4, a max-imal yield of 17.0 Ϯ 3.8% dityrosine was obtained with the lowest peroxynitrite concentration that has been tested (1 M) and decreased down to less than 1% at Ն1 mM peroxynitrite. The replot of the 3-nitrotyrosine data (open symbols in Fig. 4) demonstrates that dityrosine is indeed the major product of tyrosine reacting with low concentrations of peroxynitrite. efficiency of the combined system was still significantly lower than that of authentic peroxynitrite, these results suggested that peroxynitrite formed from NO/O 2 . may indeed be capable of triggering nitration under certain conditions. It is conceivable that tyrosine nitration has been quenched by XO that was used for O 2 . generation in the other experimental set-ups. Accordingly, the protein-free DEA/NO-FMN system apparently allowed the detection of the minor nitration reaction triggered by peroxynitrite at low steady-state concentrations.
The most interesting finding of this study was the observation that dityrosine formation almost completely outcompeted nitration at low concentrations of peroxynitrite. As a mechanistic explanation of these surprising results, we propose the scheme depicted in Fig. 5. Accordingly, the key event of both reactions, nitration and dityrosine formation, would be the generation of tyrosyl radicals by ⅐ NO 2 formed in the course of homolytic cleavage of ONOOH (Equations 1 and 2, path a in Fig. 5). The tyrosyl radicals could either react with ⅐ NO 2 to yield 3-nitrotyrosine (Equation 3, path b) or dimerize to give dityrosine (Equation 4, path c). A major competing reaction would be the dimerization of ⅐ NO 2 yielding N 2 O 4 (Equation 5, path d).
Homolysis of ONOOH (Equation 1) has been questioned based on thermodynamical calculations (39), but recent evidence suggests that about 30% of ONOOH does indeed yield free ⅐ NO 2 and ⅐ OH, whereas the residual 70% undergoes rearrangement to nitric acid without escape of free radicals (22,23,40). Tyrosyl radical formation by ⅐ NO 2 and subsequent combination of ⅐Tyr and ⅐ NO 2 has been reported to occur with second order rate constants of 3.2 ϫ 10 5 and 3 ϫ 10 9 M Ϫ1 s Ϫ1 , respectively (24). Rate constants of 9 ϫ 10 8 and 2.25 ϫ 10 8 M Ϫ1 s Ϫ1 , respectively, were reported for the two major competing reactions, i.e. the dimerization of ⅐ NO 2 (24) and the combination of two ⅐ Tyr radicals to yield dityrosine (25).
Together with the rate of peroxynitrite decomposition determined by stopped-flow spectroscopy under our experimental conditions (0.27 s Ϫ1 ), the published rate constants were used for the kinetic simulation of peroxynitrite reacting with excess free tyrosine, assuming 30% homolysis of ONOOH. Fig. 6 shows that the yields of 3-nitrotyrosine and dityrosine predicted by the model for tyrosine reacting with peroxynitrite at concentrations ranging from 1 M to 2 mM are similar in shape to the measured data illustrated in Fig. 4. In agreement with our observations, the model predicts an inverse dependence on peroxynitrite concentration of tyrosine nitration and dimerization. At low peroxynitrite concentrations, dimerization of ⅐Tyr radicals (filled symbols) is the predominant pathway, whereas nitration (open symbols) and ⅐ NO 2 dimerization (not shown), which both follow second order kinetics with respect to ⅐ NO 2 , become the predominant reactions at high peroxynitrite (and thus ⅐ NO 2 ) concentrations. The measured yields of dityrosine agreed well with the predictions of the model, but the measured 3-nitrotyrosine levels were Ն2-fold below the theoretical expectation over the complete range of peroxynitrite concentrations. This quantitative mismatch suggests that reactions not considered in the kinetic simulation compete with tyrosine nitration. These reactions may involve ⅐ OH radicals, as it was shown previously that ⅐ OH radical scavengers significantly enhance peroxynitrite-triggered tyrosine nitration (4). Therefore, it is likely that the reactions of ⅐ OH radicals with ⅐ NO 2 to yield HNO 3 and with ⅐ Tyr radical, resulting in the formation of 3-hydroxytyrosine (dopa) (36), compete with the nitration reaction. Since the rate constants of the reactions triggered by ⅐ OH are not known, it was not possible to account for them in the kinetic model. Nonetheless, we think that, despite some quantitative uncertainties, the proposed model provides a simple and reliable mechanistic explanation for the insignificant nitration efficiency of peroxynitrite generated in situ.
What are the implications of the present study for the effects of peroxynitrite generated from NO/O 2 . in vivo? Obviously, the oxidative chemistry of peroxynitrite, including dityrosine formation, would be expected to be predominant at the relatively low NO/O 2 . fluxes that are likely to occur in most in vivo conditions. As a specific marker of oxidation, dityrosine has been detected in human atherosclerotic plaques (41,42) in the brain of elderly humans (43) or patients affected with Alzheimer's disease (44), in age-related nuclear cataract (45) and other pathologies thought to be associated with oxidative stress. Formation of dityrosine has been attributed mainly to the activation of the myeloperoxidase/H 2 O 2 system of neutrophils and macrophages (46), but other peroxidases (47) and peroxynitrite (36, 48) have been recognized as additional sources of dityrosine. The present results agree with previous studies suggesting that dityrosine formation together with increased NO synthase expression may be a useful marker for peroxynitrite formation in tissues (49,50).
With respect to tyrosine nitration, it seems unlikely that the high NO (58), and reports with transgenic mice and SOD knockout mutants showing that both Cu,Zn and Mn-SOD are protective against stroke (59). Our data render it likely that the molecular mechanisms underlying these pathologies are related to protein oxidation and/or cross-linking rather than nitration. It is conceivable that latter reaction is triggered by peroxynitrite-independent pathways involving myeloperoxidase (60 -62) or other peroxidases (63). As a further alternative, trapping of tyrosyl radicals by NO and subsequent peroxidase-mediated oxidation of nitrosotyrosine could result in the formation of 3-nitrotyrosine (64). The latter mechanism would imply that several pathways have to be activated at the same time to cause significant nitration. In inflammatory tissues, for example, induction of macrophage NO synthase together with the activation of neutrophil NADPH oxidase and secretion of myeloperoxidase would constitute a highly efficient nitrating system operating through several pathways. Further studies should clarify which of these pathways or which combinations of them are responsible for tyrosine nitration in human disease.