Formation of nitrating and chlorinating species by reaction of nitrite with hypochlorous acid. A novel mechanism for nitric oxide-mediated protein modification.

Detection of 3-nitrotyrosine has served as an in vivo marker for the production of the cytotoxic species peroxynitrite (ONOO−). We show here that reaction of nitrite (NO−2), the autoxidation product of nitric oxide (·NO), with hypochlorous acid (HOCl) forms reactive intermediate species that are also capable of nitrating phenolic substrates such as tyrosine and 4-hydroxyphenylacetic acid, with maximum yields obtained at physiological pH. Monitoring the reaction of NO−2 with HOCl by continuous flow photodiode array spectrophotometry indicates the formation of a transient species with spectral characteristics similar to those of nitryl chloride (Cl-NO2). Reaction of synthetic Cl-NO2 with N-acetyl-L-tyrosine results in the formation of 3-chlorotyrosine and 3-nitrotyrosine in ratios that are similar to those obtained by the NO−2/HOCl reaction (4:1). Tyrosine residues in bovine serum albumin are also nitrated and chlorinated by NO−2/HOCl and synthetic Cl-NO2. The reaction of N-acetyl-L-tyrosine with NO−2/HOCl or authentic Cl-NO2 also produces dityrosine, suggesting that free radical intermediates are involved in the reaction mechanism. Our data indicate that while chlorination reactions of Cl-NO2 are mediated by direct electrophilic addition to the aromatic ring, a free radical mechanism appears to be operative in nitrations mediated by NO−2/HOCl or Cl-NO2, probably involving the combination of nitrogen dioxide (·NO2) and tyrosyl radical. We propose that NO−2 reacts with HOCl by Cl+ transfer to form both cis- and trans-chlorine nitrite (Cl-ONO) and Cl-NO2 as intermediates that modify tyrosine by either direct reaction or after decomposition to reactive free and solvent-caged Cl· and ·NO2 as reactive species. Formation of Cl-NO2 and/or Cl-ONO in vivo may represent previously unrecognized mediators of inflammation-mediated protein modification and tissue injury, and offers an additional mechanism of tyrosine nitration independent of ONOO−.

Nitrogen monoxide (nitric oxide, ⅐ NO) 1 is produced by a variety of cells through the activity of constitutive and inducible forms of nitric oxide synthase (1). ⅐ NO is an important endogenous mediator in such diverse biochemical and physiological processes as neurotransmission, smooth muscle relaxation, platelet aggregation and adhesion, macrophage-mediated cytotoxicity, and learning and memory (2,3). Although basal levels of free ⅐ NO are normally quite low (nanomolar), local ⅐ NO concentrations have been shown to increase to levels ranging from 4 to 30 M under pathologic conditions (4,5).
Based upon tyrosine nitration assays and the formation of "peroxynitrite-specific" luminescence, stimulated macrophages (18), neutrophils (19), and endothelial cells (20) have been proposed to form significant quantities of ONOO Ϫ in vitro. In fact, the detection of 3-nitrotyrosine (NO 2 -Tyr) in a variety of pathologic conditions in vivo, such as inflammatory lung disease (21), atherosclerosis (22), and rheumatoid arthritis (23), has been attributed to ONOO Ϫ formation. However, in all of these cases direct proof for the production of ONOO Ϫ in biolog-* This work was supported by National Institutes of Health Grant HL47628 and the Arthritis and Rheumatism Council (United Kingdom). 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. ical systems is lacking, even though its formation in vivo is favorably predicted (24).
Under inflammatory conditions, multiple well characterized reactive oxygen species (ROS) are produced from phagocytic cells (25). For instance, stimulated neutrophils and macrophages produce significant levels of superoxide (O 2 . ) and hydrogen peroxide (H 2 O 2 ) as a result of the activation of the respiratory burst oxidase (26). In the case of neutrophils, some of the H 2 O 2 that is produced under these conditions is converted to the strong oxidant hypochlorous acid (HOCl) by the action of myeloperoxidase as shown in Reaction 1.
HOCl produced from activated human neutrophils has been shown to react with amines (taurine, lysine, and arginine) and tyrosine to form N-chloramines and 3-chlorotyrosine (Cl-Tyr), respectively (27,28), where the latter has been proposed to serve as a selective marker of HOCl production in vivo (29). In addition to ROS, macrophages (30) and neutrophils (19) can also simultaneously produce large fluxes of ⅐ NO through the activation of inducible nitric oxide synthase; however, the ability of human neutrophils to produce ⅐ NO is debated (31). Once formed, ⅐ NO can react with several biological targets, primarily thought to involve heme-iron, hyperreactive sulfhydryls and protein radicals (32)(33)(34). ⅐ NO can also react with O 2 in aqueous solution to produce nitrite (NO 2 Ϫ ) via a complex mechanism thought to involve a variety of reactive nitrogen species (RNS) including ⅐ NO 2 and dinitrogen trioxide (N 2 O 3 ) (35). In fact, NO 2 Ϫ has been used as a marker of ⅐ NO production in vitro and in vivo and has been shown to reach concentrations of up to 4 M in synovial fluid from patients with rheumatoid arthritis (36) and as high as 20 M in human airway fluids (37). These RNS produced during an inflammatory response could theoretically react with a number of ROS to form various novel species. Indeed, the interaction of HOCl with ⅐ NO or NO 2 Ϫ has been proposed to form species capable of nitrosylating and nitrating organic substrates (38,39).
The present study was undertaken to examine the potential interactions of ⅐ NO-derived RNS with the inflammatory oxidant HOCl in an attempt to characterize more fully the various species that may be formed under complex physiological inflammatory conditions. Our results indicate that NO 2 Ϫ reacts with HOCl to form an intermediate species, postulated to be nitryl chloride (Cl-NO 2 ) and/or chlorine nitrite (Cl-ONO), that is capable of nitrating, chlorinating, and dimerizing phenolic compounds including tyrosine. We propose that the formation of Cl-NO 2 and/or Cl-ONO by this reaction represents a novel mechanism of inflammation-mediated biological damage, and offers an additional or alternative mechanism of tyrosine nitration independent of ONOO Ϫ formation.
⅐ NO Experiments-⅐ NO (3,000 ppm in O 2 -free N 2 ) was bubbled through continuously stirred 100-ml solutions of phosphate buffer (100 mM KH 2 PO 4 , pH 7.4) at a flow rate of 20 ml/min. Buffer solutions were purged with either air or purified N 2 for 30 min prior to ⅐ NO exposure and were continuously sparged with either gas throughout the experiment to maintain an oxygenated or deoxygenated solution, respectively. At various time points, aliquots of the solution were withdrawn and purged briefly with N 2 to remove residual ⅐ NO, and NO 2 Ϫ production was determined spectrophotometrically using Griess reagent (1% sulfanylamide, 0.1% N-(1-naphthyl)ethylenediamine, and 2.5% H3PO 4 ) (41). In separate experiments, HPA (5 mM) was added to the buffer solutions prior to ⅐ NO exposure. At various time points, 0.5-ml aliquots of the solution were withdrawn from the reaction mixture and immediately reacted with HOCl (1 mM). After 10 min, the reaction mixture was adjusted to pH 10 -11 with 1 M NaOH. NO 2 -HPA was measured spectrophotometrically at 430 nm; ⑀ ϭ 4400 M Ϫ1 cm Ϫ1 (16).
Nitration and Chlorination Reactions-Buffered solutions (100 mM KH 2 PO 4 ) of HPA, NAT, NAP, or MPA (1-5 mM) were adjusted to the desired pH (5.0 -8.5) with either 10% NaOH or 5% H 3 PO 4 prior to experimentation. Where appropriate, NO 2 Ϫ was added at the desired concentration (0.1-3.0 mM) in 5-ml sample volumes. HOCl was then added to the solutions at 25°C as a small drop (Ͻ20 l) while continuously vortexing. Although reactions are nearly instantaneous, incubations were allowed to proceed for 10 min before reduced GSH was added at 1 mM concentration to scavenge any unreacted HOCl. In separate experiments ONOO Ϫ (in 1.2 M NaOH) or NO 2 BF 4 (in sulfolane) were reacted with the various substrates in a similar manner. The products of the reactions involving HPA as substrate were analyzed directly by HPLC. Since nitrated, chlorinated, and dimerized N-acetylated derivatives of tyrosine were not available, reaction mixtures utilizing NAT as substrate were first hydrolyzed (see below) to liberate the free amino acids and their modified products prior to HPLC analysis. Reaction mixtures involving MPA were transferred to a quartz cuvette, and the absorbance spectrum was measured between 280 and 500 nm at pH 7.4.
Reactions of BSA with NO 2 Ϫ /HOCl-NO 2 Ϫ and HOCl (both 500 mM) were loaded into 1-ml syringes that were attached to an automated syringe pump. Teflon tubing from both syringes converged into a single tube and allowed reaction of the two components for a brief period (Ͻ1 s). The reaction effluent was allowed to drop approximately 6 cm immediately into 10-ml solutions of BSA (10 mg/ml) in 100 mM KH 2 PO 4 (pH 7.4), which were continuously stirred. The volume of each drop was calibrated (33 l), and final concentrations of oxidant exposure were calculated. In some cases, NO 2 Ϫ in one of the syringes was replaced with 100 mM KH 2 PO 4 (pH 7.4) so as to allow exposure of BSA to HOCl alone under the same conditions. Following addition of the NO 2 Ϫ /HOCl mixture or HOCl alone, the solutions were stirred for 15 min and were then quenched by the addition of excess GSH. NO 2 -Tyr, Cl-Tyr, and dityrosine formation in the samples were determined by HPLC following acid hydrolysis as described below.
Synthesis of Cl-NO 2 -Caution: the reagents used and products formed in this synthesis are highly irritant and corrosive to the eyes, skin, and mucous membranes. All of the procedures involved in the synthesis of Cl-NO 2 must be performed in a fume hood to ensure proper ventilation, and appropriate eye and skin protection must be worn. At room temperature, Cl-NO 2 exists as a gas (boiling point, Ϫ15°C) and presents a serious inhalation hazard if not handled properly. The procedure for the synthesis of Cl-NO 2 is essentially that previously described (42,43) with slight modification. Sulfuric acid (61 g) was added dropwise to vigorously stirred nitric acid (50 g) at 0°C in a 500-ml three-necked round bottom flask equipped with a dropping funnel. A cold finger receiving flask (150 ml) was attached to the reaction vessel by a short segment of Teflon tubing, and the flask was immersed in a cooling mixture of dry ice/acetone. After 10 min, chlorosulfonic acid (85 g) was slowly added dropwise via the attached funnel into the mixture of nitric and sulfuric acids over a 4-h period. During the entire synthesis procedure, a gentle flow of N 2 gas was delivered through the apparatus to enhance the evolution and collection of gaseous Cl-NO 2 . It is important that the chlorosulfonic acid is added at a slow enough rate such that brown gas does not appear above the reaction mixture. The colorless gaseous Cl-NO 2 evolved from the reaction mixture was carried by the gentle flow of N 2 into the cold finger receiving flask, where it condensed as a pale yellow liquid. The product was purified by passing ozonized air through the liquefied gas to oxidize any nitrosyl chloride (Cl-NO) that may have been present as a contaminant. The product was carefully transferred to sealed glass vials and stored at Ϫ80°C until Nitration and Chlorination of Tyrosine by NO 2 Ϫ /HOCl Reaction used in experiments. The yield of Cl-NO 2 is typically 80 -90% (approximately 50 g), and the purity has been reported to be 98 -99% (42). To confirm the identity of the reaction product, the purified product was diluted in methanol, and the absorbance spectrum was immediately measured. The observed absorbance spectrum of synthetic Cl-NO 2 showed a series of characteristic absorption maxima between 300 and 400 nm, similar to that reported previously (43). Nitryl Chloride Exposures-Cl-NO 2 (10 ml) was placed in a 50-ml sparging flask immersed in a cooling dry ice/acetone bath. A stream of N 2 gas was allowed to flow through a glass tube fitted with a fritted glass fitting, which was submerged into the undiluted Cl-NO 2 , and was bubbled through the liquid at a flow rate of 75 ml/min. Gaseous Cl-NO 2 evolved into the headspace exited through a glass tube connected to the top of the flask into a glass reaction vessel containing a solution of the analyte to be exposed. Cl-NO 2 in N 2 gas was allowed to bubble through the 100 mM phosphate-buffered solutions (25 ml, pH 7.4) of NAT (5 mM), NAP (5 mM), MPA (1 mM), or BSA (10 mg/ml) for various periods of time (0 -120 s). Aliquots (500 l) of the solutions were sampled at various time points and subjected to acid hydrolysis, and the levels of the modified amino acids were determined by HPLC as described below.
Spectral Characterization of NO 2 Ϫ /HOCl Reaction Intermediates-To characterize the intermediate(s) produced by reaction of NO 2 Ϫ and HOCl we have utilized a continuous flow reaction with photodiode array (PDA) spectrophotometric detection. NO 2 Ϫ and HOCl (both at 25 mM in 50 mM KH 2 PO 4 , pH 6.0) were independently pumped into a mixing junction at a flow rate of 0.3 ml/min. Upon mixing, the reaction effluent was immediately directed into the flow cell of a Waters 996 PDA detector, and the absorbance spectrum of the reaction products was continuously monitored over the range 300 -400 nm. In some cases the reaction medium was supplemented with 25% methanol (HPLC grade) in order to compare the absorbance spectra with that of authentic Cl-NO 2 in methanol, as described above.
HPLC Analysis of Reaction Products-All reaction mixtures were analyzed by HPLC, using a 5-m Spherisorb ODS-2 reverse-phase C-18 column. Samples containing HPA were analyzed directly following experiments without sample preparation by isocratic elution from the column with a mobile phase consisting of 100 mM KH 2 PO 4 (pH 3.5)/ methanol (70/30, v/v) and UV detection at 274 nm. Samples including NAT, NAP, or BSA were first hydrolyzed in 6 M HCl at 110°C in sealed glass vials for 4 and 24 h, respectively, to obtain the free amino acids and their modified products. The hydrolysates were then dried using a Centrivap (Labconco) and resuspended in the appropriate mobile phase. Tyrosine, NO 2 -Tyr, and Cl-Tyr were analyzed by isocratic elution from the column with 50 mM KH 2 PO 4 (pH 3.0)/methanol (92/8, v/v) and subsequent UV detection at 274 nm (15). Dityrosine was detected simultaneously by on-line fluorescence detection using a Waters 470 scanning fluorescence detector (excitation, 284 nm; emission, 410 nm) (15). Phenylalanine (Phe) and ortho-, meta-, and para-chlorophenylalanine (o-, m-, and p-Cl-Phe) were separated on the same column cited above, utilizing a mobile phase consisting of 50 mM KH 2 PO 4 (pH 3.0)/ methanol (85/15, v/v) with UV detection at 220 nm. Peaks were identified and quantitated using authentic external standards. Peak identity was determined by adding to the sample the authentic compound to establish a match in the HPLC retention time. Peak identity was confirmed using a Waters 996 PDA detector. A spectral match between the authentic chemical and the sample analyte of greater than 90% constituted positive identification.

Interactions of ⅐ NO, NO 2
Ϫ , and HOCl-Aliquots of HPA solutions purged with ⅐ NO were reacted with HOCl at various time points to determine if a species is formed under these conditions that is capable of nitrating this model phenolic compound. NO 2 -HPA was immediately formed upon HOCl addition, and the yield increased in a manner dependent on the length of time the solution had been purged with ⅐ NO. The extent of NO 2 -HPA formation was significantly lower under deoxygenated (N 2 -sparged) conditions (Fig. 1A), suggesting that autoxidation of ⅐ NO is involved in the reaction with HOCl. ⅐ NO rapidly autoxidizes to NO 2 Ϫ when bubbled through air-saturated solutions, whereas formation of NO 2 Ϫ is significantly diminished under deoxygenated conditions (Fig. 1B). As shown in Fig. 1, A  Ϫ reacts with HOCl to produce a nitrating species. Indeed, the addition of HOCl to solutions containing NO 2 Ϫ and HPA resulted in the immediate (Ͻ1 s) formation of a persistent yellow color indicative of phenolic nitration. The presence of NO 2 -HPA was confirmed by subsequent HPLC analysis and detection by PDA. Nitration of HPA by the reaction of NO 2 Ϫ and HOCl in solution was found to be pH-dependent (Fig. 2). Maximal formation of NO 2 -HPA from this reaction occurred at neutral pH, whereas its formation decreased at increasingly acidic or basic pH values and is independent of ionic strength (10 -200 mM phosphate) (data not shown). The pH profile of NO 2 -HPA formation by the reaction of NO 2 Ϫ and HOCl (Fig. 2) indicates that the reaction involves HOCl and not ClO Ϫ , because of the rapid decrease in nitration at pH Ͼ 7.5 (the pK a of HOCl). Since the pK a of NO 2 Ϫ is approximately 3.4, it is NO 2 Ϫ , and not HNO 2 , that is the reacting species at all of the pH levels we have studied (pH 5.0 -8.5). The decreasing yield of NO 2 -HPA at low pH is similar to that determined for the reaction of tyrosine with ⅐ NO 2 (15,44) or ONOO Ϫ (15,16) and may well be due to the lower reactivity of the phenol relative to the phenolate species.
Reaction of NO 2 Ϫ (1 mM) and HOCl (1 mM) converted approximately 4% of HPA to NO 2 -HPA, similar to the reported yields of NO 2 -HPA obtained from reaction of ONOO Ϫ (1 mM) with this substrate (16). The reaction of NO 2 Ϫ alone with HPA did not yield NO 2 -HPA at any of the pH values studied herein. In the absence of NO 2 Ϫ , HOCl directly converted HPA into Cl-HPA at Ϫ competed with HPA for reaction with HOCl. When initial NO 2 Ϫ and HOCl concentrations were equal (1 mM), the extent of chlorination was approximately 4-fold higher than nitration. Addition of NO 2 Ϫ in excess over HOCl (up to 3-fold) did not significantly increase the yield of NO 2 -HPA, suggesting that the nitrating species is produced by a 1:1 reaction of HOCl with NO 2 Ϫ . Similarly, excess NO 2 Ϫ did not significantly decrease formation of Cl-HPA, suggesting that the product formed by this reaction is an efficient chlorinating agent as well as a nitrating species.
We also used NAT as a substrate for the reaction of NO 2 Ϫ with HOCl to simulate tyrosine residues in proteins. As shown in Table I, qualitatively and quantitatively similar results were obtained. However, whereas HOCl alone caused very small amounts of dityrosine to be produced (Ͻ0.8 M), the combined addition of NO 2 Ϫ and HOCl induced a 15-fold increase in dityrosine formation. Since dityrosine is formed by the combination of two tyrosyl radicals, this result indicated that the reaction between NO 2 Ϫ and HOCl produces a species that is capable of carrying out a one-electron oxidation of tyrosine to form the tyrosyl radical. Reactions of NAT with ONOO Ϫ and NO 2 BF 4 were also studied in order to compare the nitration mechanisms with that of NO 2 Ϫ /HOCl. Although the reaction of ONOO Ϫ (1 mM) with NAT led to NO 2 -Tyr levels that were 2-fold higher than that achieved by NO 2 Ϫ /HOCl, the levels of dityrosine were nearly identical. The NO 2 ϩ species (NO 2 BF 4 ) reacted with NAT to form NO 2 -Tyr but in lower yields than with either ONOO Ϫ or NO 2 Ϫ /HOCl treatments. The reaction of NO 2 BF 4 with NAT also yielded relatively high levels of dityrosine, suggestive of tyrosyl radical intermediates in its mechanism of tyrosine nitration.
Modification of Tyrosine in BSA by NO 2 Ϫ /HOCl-Reaction of HOCl with solutions of BSA containing NO 2 Ϫ resulted in NO 2 -Tyr formation, but more slowly than with pure HPA or NAT as substrate. However, when NO 2 Ϫ and HOCl were allowed to react just before the addition to BSA (using a dual syringe pump), rapid formation of NO 2 -Tyr was observed in a dose-dependent manner (Fig. 4A). A small amount of dityrosine (approximately 1 M) could also be detected in BSA treated in this manner. The species produced by the reaction of NO 2 Ϫ and HOCl also reacted with BSA to produce relatively high levels of Cl-Tyr (Fig. 4B). However, when NO 2 Ϫ was omitted from one of the syringes (replaced by phosphate buffer), significantly higher levels of Cl-Tyr were detected in BSA, again suggesting that NO 2 Ϫ was reacting with HOCl. No detectable levels of these modified tyrosine products were found in acid hydrolysates of control (nonexposed) solutions of BSA.
Characterization of NO 2 Ϫ /HOCl Reaction Product(s)-To determine the identity of the product(s) formed by reaction of NO 2 Ϫ with HOCl we utilized a continuous flow dual pump PDA system. The spectrum of NO 2 Ϫ at pH 6.0 (50 mM KH 2 PO 4 ) under continuous flow through the PDA detector shows a single absorbance maximum at approximately 370 nm (not shown). The addition of HOCl to the continuous flow apparatus via a separate pump leads to the degradation of the NO 2 Ϫ absorption and the concomitant formation of a series of maxima observed between 320 and 420 nm (Fig. 5A). Johnson and Margerum (45) have studied the reaction of NO 2 Ϫ with HOCl and have suggested that nitryl chloride (Cl-NO 2 ) is a product. Authentic Cl-NO 2 was synthesized as described under "Experimental Procedures," and the absorbance spectrum of this species (in methanol, a polar solvent for which Cl-NO 2 is more stable as compared with aqueous conditions) is shown in Fig. 5C for comparison and shows a series of absorption maxima between Ϫ at various concentrations were exposed to HOCl (1 mM) as a bolus addition. The reactions were allowed to proceed for 10 min, at which time excess GSH was added to scavenge residual HOCl. The yields of NO 2 -HPA (q) and Cl-HPA (å) from the reaction were determined directly by HPLC as described under "Experimental Procedures." Data points are expressed as means of three separate experiments. Nitration and Chlorination of Tyrosine by NO 2 Ϫ /HOCl Reaction 320 and 400 nm, characteristic of that for Cl-NO 2 reported previously (43). The addition of methanol to the reaction mixture of NO 2 Ϫ and HOCl caused a hypsochromic shift (20 nm) in the absorption spectrum (Fig. 5B) without affecting the characteristic series of maxima observed in the absence of methanol. Although the absorbance spectra of authentic Cl-NO 2 and that determined for the product of the reaction between NO 2 Ϫ and HOCl show much similarity, the slight differences may be due to the different conditions for which the spectra were obtained (100% methanol for Cl-NO 2 , and 25% methanol for NO 2 Ϫ /HOCl reaction) and to interference of unreacted NO 2 Ϫ in the spectrum of the NO 2 Ϫ /HOCl reaction product. Hence, we can conclude that the product formed by this reaction shows characteristics similar to those of Cl-NO 2 .

Modification of NAT and Tyrosine Residues in BSA by Cl-NO 2 -Since Cl-NO 2 is proposed to be formed by the reaction of NO 2
Ϫ with HOCl (45) and the absorption spectrum of the product formed by this reaction suggested the potential formation of Cl-NO 2 in our studies, NAT and BSA were exposed to synthetic Cl-NO 2 in order to compare its reactivity to the species produced by the NO 2 Ϫ /HOCl reaction. Exposure of a solution of NAT (5 mM) to a stream of gaseous Cl-NO 2 led to the formation of Cl-Tyr, NO 2 -Tyr, and dityrosine to an extent dependent on the duration of exposure (Fig. 6). Formation of Cl-Tyr and dityrosine reached maximum levels at 40 and 50 s respectively, after which the products were decomposed upon further expo-sure to Cl-NO 2 . The loss of Cl-Tyr is likely due to the formation of dichloro-Tyr, as has been shown for the chlorination of tyrosine by Cl 2 gas (46). In contrast, the formation of NO 2 -Tyr continued to increase over the entire exposure period, suggesting that NO 2 -Tyr is stable under Cl-NO 2 reaction conditions.

FIG. 4. Nitration and chlorination of tyrosine residues in BSA exposed to the products of the reaction between NO 2 ؊ and HOCl. Equal concentrations of HOCl and NO 2
Ϫ (or HOCl alone) were mixed together using an automated dual syringe pump, as described under "Experimental Procedures," and allowed to react as a small droplet with a continuously stirred solution of BSA (10 mg/ml). The volume of each droplet added to the solution was calibrated, and the final concentration of the reactants in solution was calculated (0 -3.2 mM). The reactions were allowed to proceed for 15 min, at which time excess GSH was added to quench the reaction. The yields of NO 2 -Tyr (A) and Cl-Tyr (B) in BSA by NO 2 Ϫ /HOCl combined exposure (å) or by HOCl alone (q) were determined using HPLC following acid hydrolysis. All data points are expressed as the mean Ϯ S.D. of at least three separate experiments.

FIG. 5. Continuous flow UV-visible PDA detection and characterization of the product(s) formed by reaction of NO 2 ؊ with HOCl. Solutions of NO 2
Ϫ and HOCl (both 25 mM in 50 mM KH 2 PO 4 , pH 6.0) were independently pumped into a mixing junction and allowed to flow directly into a PDA detector immediately after mixing. A typical absorbance spectrum (300 -400 nm) of the product(s) formed under these conditions is shown in A. Supplementation of the NO 2 Ϫ and HOCl solutions with 25% methanol led to a 20-nm hypsochromic shift in the absorbance spectrum (B). The absorbance spectrum of synthetic nitryl chloride (Cl-NO 2 ) in methanol is shown C.

FIG. 6. Modification of NAT by synthetic Cl-NO 2 .
Solutions of NAT (5 mM, in 100 mM KH 2 PO 4 , pH 7.4) were exposed to gaseous Cl-NO 2 as described under "Experimental Procedures." The yields of NO 2 -Tyr (E), Cl-Tyr (q), and dityrosine (å) were determined by HPLC following acid hydrolysis to liberate the free amino acid and its modified products. All data points are representative of duplicate determinations and are expressed as means Ϯ S.D. of three separate experiments.

Nitration and Chlorination of Tyrosine by NO 2 Ϫ /HOCl Reaction
The ratio of Cl-Tyr to NO 2 -Tyr formed by Cl-NO 2 averaged 5:1 during the early segment of exposure (20 -40 s), and is similar to the ratio of 4:1 obtained by reaction of NO 2 Ϫ /HOCl with NAT. To determine whether the reactions studied with NAT as substrate are relevant to reactions with intact proteins, gaseous Cl-NO 2 was bubbled through solutions of BSA (10 mg/ml). The time-dependent formation of NO 2 -Tyr, Cl-Tyr, and dityrosine in BSA exposed to Cl-NO 2 is summarized in Table II. The profile of modified tyrosines was qualitatively similar to that observed for the reaction of NAT with Cl-NO 2 , except that the yields of the products were lower, especially at the early time points. Whereas the rapid consumption of NAT by Cl-NO 2 was initiated immediately, the initial loss of tyrosine residues in BSA exposed to Cl-NO 2 was less dramatic, potentially because of competitive reactions with other targets in BSA. The slower initial rate of tyrosine loss in BSA paralleled the oxidation of free sulfhydryl groups in BSA. However, a much more rapid loss of tyrosine ensued following complete depletion of free sulfhydryl groups in BSA (data not shown). The data suggest that reaction of Cl-NO 2 with other amino acid residues (i.e., cysteine, methionine, and lysine) and/or the nonspecific oxidation of the peptide backbone are also important.
Mechanistic Characterization of Nitration and Chlorination Reactions-Since dityrosine is a significant product of the reaction between tyrosine and NO 2 Ϫ /HOCl, it is likely that radical species are involved in the reaction mechanisms. To test this hypothesis, we utilized the O-methylated derivative of HPA, MPA, a substrate that cannot form phenolic radicals. Fig. 7 compares the absorbance spectra of the products formed when NO 2 Ϫ /HOCl or NO 2 ϩ was reacted with MPA at pH 7.4. NO 2 ϩ , derived from the nitryl salt NO 2 BF 4 , appears capable of nitrating MPA (pH 7.4) by electrophilic aromatic substitution giving rise to a strong absorbance maximum at 380 nm (Fig. 7, spectrum C). However, the reaction of NO 2 Ϫ /HOCl with MPA failed to give rise to an absorbance in this region (Fig. 7, spectrum B), suggesting that formation of a phenolic radical is an obligatory step in the nitration mechanism involved in this reaction pathway. Reaction of synthetic Cl-NO 2 or HOCl with MPA produced a product(s) with a spectrum nearly identical to that obtained for the NO 2 Ϫ /HOCl reaction (data not shown) and suggested that Cl-NO 2 was also incapable of nitrating MPA, implicating radical intermediates in the nitration mechanism of this species. The increased absorbance between 290 and 340 nm observed in both of these cases could be due to chlorination of MPA, suggesting that intermediate phenoxyl radical formation is not a compulsory step in aromatic chlorination.
NAP was used as a model substrate to examine in more detail the mechanisms of aromatic chlorination by the product Ϫ / HOCl and Cl-NO 2 suggest a less selective mechanism of chlorination more typical of radical reactions. It is noteworthy that the reaction of NAP with NO 2 Ϫ /HOCl or Cl-NO 2 did not form detectable levels of nitrated products. DISCUSSION Although the mechanisms of biomolecular damage and pathology induced by individual inflammatory oxidants are in general well characterized, an understanding of the complex interactions of ROS and RNS that are likely to occur at sites of inflammation is only just beginning to emerge. The studies reported herein show that the interactions of RNS and HOCl may be important under inflammatory conditions in vivo. We have shown that NO 2 Ϫ , the autoxidation product of ⅐ NO in biological fluids, reacts with HOCl to produce a species that can nitrate, chlorinate, and dimerize biologically relevant phenolic compounds such as tyrosine, both free and within protein. The detection of NO 2 -Tyr in a variety of pathologic states (21)(22)(23) has been used to indicate the formation of ONOO Ϫ in vivo. However, reaction of tyrosine with the products of the NO 2 Ϫ / HOCl reaction also forms NO 2 -Tyr. Hence, our results suggest that NO 2 -Tyr should not be regarded as a specific marker of ONOO Ϫ formation, but only as a marker of RNS.
Mechanism of NO 2 Ϫ /HOCl Reaction-It has long been thought (48) that the reaction of NO 2 Ϫ with HOCl represented a TABLE II Modification of tyrosine residues in BSA exposed to synthetic Cl-NO 2 Solutions of BSA (10 mg/ml) were exposed to Cl-NO 2 as described under "Experimental Procedures." At various time points, aliquots (250 l) of the BSA solution were taken and subjected to acid hydrolysis to liberate free amino acids and their modified products. The products were identified and quantitated using HPLC. Values represent the mean Ϯ S.D. of three separate experiments.   However, this type of mechanism does not easily explain the nitration and chlorination reactions observed in our studies.
Our data suggest a more complex mechanism involving the formation of reactive nitrating and chlorinating intermediates. One-electron oxidation of NO 2 Ϫ by HOCl, producing the reactive radical species Cl ⅐ and ⅐ NO 2 is one possible pathway. Since HOCl is a poor one-electron oxidant, having an estimated oneelectron reduction potential (EЈ 0 ) in the range of ϩ0.17 to ϩ0.26 V at pH 7 (38), it is unlikely that a one-electron oxidation mechanism contributes, since the E value for the ⅐ NO 2 /NO 2 Ϫ couple is approximately ϩ1.04 V (49). In contrast, HOCl is a strong two-electron oxidant (EЈ 0 ϭ ϩ1.08 V) (38) and would favor the conversion of NO 2 Ϫ to the nitryl cation (NO 2 ϩ ) or an "NO 2 ϩ -like" species. In addition to a direct two-electron oxidation of NO 2 Ϫ by HOCl, a bimolecular substitution reaction between these two reactants could be involved. In fact, contrary to the reaction mechanism previously reported (48), Johnson and Margerum (45) have suggested that HOCl reacts with NO 2 Ϫ by Cl ϩ transfer, rather than O atom transfer, to yield the intermediate Cl-NO 2 , which then hydrolyzes to NO 3 Ϫ . The absorbance spectrum of the product of the reaction between NO 2 Ϫ and HOCl (Fig. 5B) was found to be similar to that of authentic Cl-NO 2 (Fig. 5C). The spectrum of the product(s) of the NO 2 Ϫ /HOCl reaction is typical of alkyl nitrites (R-ONO) (50) and therefore could also indicate the formation of a Cl-Obonded species. In fact, the transfer of Cl ϩ to the negatively charged oxygen atom in NO 2 Ϫ is likely and would produce the transient intermediate species Cl-ONO. It is possible that both reactions occur (Fig. 9), the extent to which each pathway initially predominates under neutral aqueous conditions is not known. Cl-ONO can exist as both the cis-and trans-rotamers (Fig. 9), where ab initio calculations predict that the energy difference between the two rotamers is approximately 3 kcal/ mol, with the cis rotamer being the more stable (51). An analogy can be drawn between Cl-ONO and HO-ONO (peroxynitrous acid), where the energy difference between cis-and trans-HO-ONO is also calculated to be approximately 3 kcal/mol (52). Once formed, Cl-ONO can readily isomerize to Cl-NO 2 (53). We propose that intermediate Cl-ONO can isomerize in aqueous solution to Cl-NO 2 by at least two mechanisms ( Fig. 9): 1) intramolecular rearrangement of trans-Cl-ONO involving migration of the chlorine atom to the nitrogen atom forming Cl-NO 2 , or 2) unimolecular homolysis of the Cl-O bond in Cl-ONO to form a geminate pair of solvent-caged radicals Cl ⅐ and ⅐ NO 2 , which undergo cage return to either reform Cl-ONO or by recombination to form Cl-NO 2 (Fig. 9). Some of the solvent-caged Cl ⅐ and ⅐ NO 2 can escape as "free" radicals and could potentially explain, in part, the radical mechanisms involved in the nitration reactions we observed in the NO 2 Ϫ /HOCl reaction. Since Cl-NO 2 is predicted to be 10.7 and 13.8 kcal/mol lower in energy than cis-and trans-Cl-ONO (51), respectively, the isomerization of Cl-ONO to Cl-NO 2 is a favorable process that shifts the equilibrium toward Cl-NO 2 . Isomerization of cis Cl-ONO to Cl-NO 2 is probably not likely, because the large size of the chlorine atom, which would presumably preclude the migration of the chlorine atom to the nitrogen atom and, hence, the trans-rotamer of Cl-ONO, is probably the species that isomerizes to Cl-NO 2 , analogous to the decomposition of transperoxynitrous acid (trans-HO-ONO). Whereas the isomerization of trans-HO-ONO leads to nitric acid (HO-NO 2 ), an unreactive end product, isomerization of Cl-ONO produces another highly reactive species (Cl-NO 2 ). Hence, Cl-ONO and the product of isomerization, Cl-NO 2 , may both be reactive oxidants with nitrating and chlorinating activity.
Decomposition Products of Cl-NO 2 as Reactive Intermediates-We have shown that the product(s) of the reaction between NO 2 Ϫ and HOCl, authentic Cl-NO 2 , or the NO 2 ϩ species (NO 2 BF 4 ) react with tyrosine to form NO 2 -Tyr and dityrosine. Although none of these reactants are themselves radicals, formation of dityrosine suggests the involvement of intermediate tyrosyl radicals. The nitration of aromatic compounds by NO 2 ϩ is often thought to be a classical electrophilic aromatic substitution reaction, but there is strong evidence implicating electron transfer reactions and radical intermediates in these pathways (54). This reaction mechanism involves electron transfer from the aromatic to NO 2 ϩ , followed by radical pair collapse, and it would explain the detection of dityrosine in our studies. Hence, we are unable to distinguish between a nitration mechanism involving ⅐ NO 2 or NO 2 ϩ based solely on the formation of dityrosine. However, a divergence in the characteristics of the reaction mechanisms between NO 2 ϩ and the reactive nitrating species formed by the reaction of NO 2 Ϫ with HOCl is evident in their reactions with MPA, the O-methylated derivative of HPA, a substrate incapable of forming phenoxyl radicals. Whereas NO 2 ϩ appears capable of nitrating MPA, both the product(s) of the NO 2 Ϫ /HOCl reaction and synthetic Cl-NO 2 fail to do so. Similarly, the inability of NO 2 Ϫ /HOCl and Cl-NO 2 to nitrate phenylalanine further argues against NO 2 ϩ as the species involved in tyrosine nitration.
There is evidence suggesting that the reaction of Cl-NO 2 with alkenes and aromatic compounds involves homolytic processes yielding free radical intermediates (42), probably involving both Cl ⅐ and ⅐ NO 2 . Collis et al. (43) have found that Cl-NO 2 decomposes at room temperature by homolysis to form Cl 2 and ⅐ NO 2 as shown in Reaction 3, whereby these spontaneous decomposition products may be responsible, at least in part, for the chlorinating and nitrating behavior of Cl-NO 2 in our experiments. We suggest that phenolic nitration mediated by the NO 2 Ϫ /HOCl reaction involves ⅐ NO 2 .
While the nitration reactions we observed appear to be radicalmediated, chlorination of aromatic amino acids such as phenylalanine ( Fig. 8) appears to be executed largely by electrophilic aromatic substitution. In general, chlorination of aromatic compounds by HOCl, tert-butyl hypochlorite, and Cl 2 has been shown to be mediated by an ionic rather than a free radical mechanism (47). The nearly 2-fold increase in the relative formation of the m-Cl-Phe isomer by reactions of phenylalanine with both NO 2 Ϫ /HOCl and Cl-NO 2 (Fig. 8), however, suggests the potential contribution of a less selective mechanism of chlorination, potentially involving Cl ⅐ . An active chlorinating species common to HOCl and Cl-NO 2 appears to be Cl 2 . In fact, the formation of Cl 2 from HOCl and Cl-NO 2 can be rationalized and would explain the similarities in their chlorinating ability. HOCl is in equilibrium with Cl 2 in aqueous solution as shown in Reaction 4. The formation of Cl 2 from Cl-NO 2 has been proposed to occur by 1) the homolysis of two molecules of Cl-NO 2 to form two Cl ⅐ which combine to form Cl 2 (Reaction 3), and 2) the reaction of Cl-NO 2 with H 2 O (43) as shown in Reaction 5.
Although convincing evidence suggests an electrophilic substitution mechanism for these chlorination reactions, the possibility of a mechanism involving the addition of Cl ⅐ to the aromatic ring cannot be excluded for reactions involving Cl-NO 2 or NO 2 Ϫ /HOCl. Direct Reactions of Cl-NO 2 /Cl-ONO with Tyrosine-The mechanisms of chlorination and nitration discussed thus far have primarily involved species derived from the decomposition of either Cl-NO 2 or Cl-ONO. However, as predicted by the stoichiometry of Reactions 3 and 5, these pathways are particularly favored when Cl-NO 2 or Cl-ONO are present at high concentrations. In vivo, however, Cl-NO 2 and Cl-ONO would be expected to be produced at rates that may favor the direct reaction of either species with biological substrates that are present in relative excess. In nonpolar organic solvents Cl-NO 2 has been shown to be an efficient agent for the nitration of aromatic compounds of intermediate reactivity (55). However, an increase either in the reactivity of the aromatic substrate (from benzene to phenol) or in the polarity of the solvent causes a marked decrease in the nitrating efficiency of Cl-NO 2 and a concomitant increase in the yield of chlorinated products (56). In fact, Obermeyer et al. (57) argued against the localization of a positive charge on the "nitryl" group of Cl-NO 2 , where the structural characteristics of Cl-NO 2 contrast those of typical stable nitryl salts (i.e., NO 2 ϩ BF 4 Ϫ ). Hence, reactions involving activated aromatic substrates such as tyrosine coupled with aqueous conditions would increase aromatic chlorination by Cl-NO 2 , suggesting a change from Cl Ϫ NO 2 ϩ character to a species with considerable Cl ϩ NO 2 Ϫ character. Our data suggest that Cl-NO 2 has significant Cl ϩ character in aqueous solution, and it is this functionality of Cl-NO 2 that dictates its reactivity.
We propose that Cl ϩ NO 2 Ϫ can react directly with tyrosine via electron transfer to yield an intermediate radical pair (tyrosyl radical-Cl ⅐ -NO 2 Ϫ ) (Fig. 10). Radical pair collapse of this complex leads to the rapid formation of Cl-Tyr and NO 2 Ϫ (Fig. 10, reaction A), and is the major product formed by this reaction. This electron transfer-mediated reaction mechanism is analogous to the nitration of phenolic substrates by NO 2 ϩ (54). Dissociation of the radical pair complex and subsequent oxidation of NO 2 Ϫ by Cl ⅐ (a strongly oxidizing species, EЈ 0 of Cl ⅐ /Cl Ϫ ϭ ϩ2.2-2.6 V (Ref. 49)) results in the formation of "free" tyrosyl radical and ⅐ NO 2 (Fig. 10, reaction B). Tyrosyl radical and ⅐ NO 2 can rapidly combine to yield NO 2 -Tyr (k ϭ 3 ϫ 10 9 M Ϫ1 s Ϫ1 (Ref. 44)), and dityrosine formation can be envisaged by the combination of two tyrosyl radicals (Fig. 10, reactions C and D). This proposed mechanism predicts that the yields of the various tyrosine modification products will be on the order of Cl-Tyr Ͼ Ͼ NO 2 -Tyr Ͼ dityrosine, consistent with the data presented herein. The proposed reaction pathway also illustrates the dependence of radical intermediates in the nitration of phenolic compounds by Cl-NO 2 , as suggested by our data.
Since Cl-ONO is a potential transient intermediate in the formation of the reactive species Cl-NO 2 (Fig. 9), part of the reactivity of NO 2 Ϫ /HOCl may be attributed to Cl-ONO. Analogous to a proposed mechanism of ONOOH reactivity (9, 58), a vibrationally excited intermediate derived from trans-Cl-ONO (Cl-ONO * ) may be formed during its isomerization to Cl-NO 2 and contribute to nitration and chlorination of tyrosine by direct reaction. The reaction mechanisms we propose for Cl-NO 2 and Cl-ONO are analogous to those recently determined for ONOOH, whereby both direct and indirect reactions with oxidizable substrates can occur (59). A more detailed examina- Ϫ , resulting in an intermediate radical pair (tyrosyl radical-Cl ⅐ ). Radical pair collapse leads to the formation of Cl-Tyr and NO 2 Ϫ (A) as major products. Dissociation of the complex from the solvent cage allows Cl ⅐ to oxidize NO 2 Ϫ to ⅐ NO 2 (B), which can combine with simultaneously formed "free" tyrosyl radical to yield NO 2 -Tyr (C). Dityrosine formation can be envisaged by the combination of two tyrosyl radicals (D). tion of the reaction kinetics and thermodynamic considerations is necessary in order to elucidate which of the proposed mechanisms predominate.
Physiological Relevance and Biological Implications-The activation and accumulation of neutrophils at sites of tissue injury, leading to the formation of HOCl and other ROS/RNS, is an essential feature of inflammation. Our data suggest that the reaction of HOCl with NO 2 Ϫ , derived from ⅐ NO produced by other phagocytes (30), endothelial cells (2), or epithelial cells (37), may be a contributing pathway operative in tissue injury at sites of inflammation. Moreover, since Cl-NO 2 is conceivably formed in vivo and is capable of nitrating tyrosine residues, our findings may imply a role for this reaction pathway where NO 2 -Tyr is detected in cases of acute inflammatory lung injury (21), atherosclerosis (22), and rheumatoid arthritis (23), a phenomena previously ascribed to ONOO Ϫ formation. In fact, increased levels of NO 2 Ϫ have been observed in similar cases (36,60,61) and further suggest the potential involvement of this pathway in vivo. Recent studies have demonstrated that myeloperoxidase, the phagocytic enzyme that catalyzes HOCl formation, is a component of sputum from cystic fibrosis patients (62), as well as other inflammatory lung diseases, and of human atherosclerotic tissue (63), underscoring the potential importance of HOCl in the pathology of each of these cases. We propose, then, that tyrosine nitration by Cl-ONO and/or Cl-NO 2 , formed by the reaction of NO 2 Ϫ with HOCl, represents an important and additional mechanism for inflammation-mediated tyrosine nitration in vivo, independent of ONOO Ϫ formation.
Our findings indicate that NO 2 Ϫ may not be an appropriate marker of ⅐ NO production by neutrophils or at sites of inflammation, since it is potentially removed by reaction with simultaneously produced HOCl. Determination of ⅐ NO production in tissues and fluids of patients with acute and chronic inflammation, as measured by NO 2 Ϫ "accumulation," is likely a gross underestimate. Moreover, NO 2 Ϫ has been shown to modulate the bactericidal activity of HOCl, its mechanism proposedly mediated by direct reaction of these two species (39,41). Our findings suggest, then, that the reaction product, Cl-NO 2 , is a strongly oxidizing species that may serve as an antimicrobial agent in its own right.
An analogous reaction between hypobromous acid, a product formed by oxidation of Br Ϫ catalyzed by eosinophil peroxidase (64), and NO 2 Ϫ forming Br-ONO and/or Br-NO 2 can be envisioned. In fact, Reaction 6 may represent a general mechanism by which hypohalous acids (HOX) react with NO 2 Ϫ to produce species capable of oxidizing biological molecules.
Collectively, these reaction pathways could therefore represent an important host defense mechanism and a novel pathway for inflammation-mediated tissue injury.
Conclusions-We report here that NO 2 Ϫ and HOCl react to form the reactive intermediates Cl-NO 2 and/or Cl-ONO, species that are capable of nitrating, chlorinating, and dimerizing phenolic compounds such as tyrosine. Our data suggest that NO 2 -Tyr is not necessarily a specific marker of ONOO Ϫ formation in vivo and that Cl-NO 2 and Cl-ONO may be important and previously unconsidered oxidants produced at sites of inflammation.