Linoleic acid hydroperoxide reacts with hypochlorous acid, generating peroxyl radical intermediates and singlet molecular oxygen

Abstract

The reaction of hypochlorous acid (HOCl) with hydrogen peroxide is known to generate stoichiometric amounts of singlet molecular oxygen [O₂ (¹Δ_g)]. This study shows that HOCl can also react with linoleic acid hydroperoxide (LAOOH), generating O₂ (¹Δ_g) with a yield of 13 ± 2% at physiological pH. Characteristic light emission at 1,270 nm, corresponding to O₂ (¹Δ_g) monomolecular decay, was observed when HOCl was reacted with LAOOH or with liposomes containing phosphatidylcholine hydroperoxides, but not with cumene hydroperoxide or tert-butyl hydroperoxide. The generation of O₂ (¹Δ_g) was confirmed by the acquisition of the spectrum of the light emitted in the near-infrared region showing a band with maximum intensity at 1,270 nm and by the observation of the enhancing effect of deuterium oxide and the quenching effect of sodium azide. Mechanistic studies using ¹⁸O-labeled linoleic acid hydroperoxide (LA¹⁸O¹⁸OH) showed that its reaction with HOCl yields ¹⁸O-labeled O₂ (¹Δ_g) [¹⁸O₂ (¹Δ_g)], demonstrating that the oxygen atoms in O₂ (¹Δ_g) are derived from the hydroperoxide group. Direct analysis of radical intermediates in the reaction of LAOOH with HOCl by continuous-flow electron paramagnetic resonance spectroscopy showed a doublet signal with a g-value of 2.014 and a hyperfine coupling constant from the α-hydrogen of a_H = 4.3 G, indicating the formation of peroxyl radicals. Taken together, our results clearly demonstrate that HOCl reacts with biologically relevant lipid hydroperoxides, generating O₂ (¹Δ_g). In addition, the detection of ¹⁸O₂ (¹Δ_g) and peroxyl radicals strongly supports the involvement of a Russell mechanism in the generation of O₂ (¹Δ_g).

Hypochlorous acid (HOCl) is a potent oxidant generated in neutrophils by the reaction of chloride ion with hydrogen peroxide (Eq. 1) catalyzed by myeloperoxidase (MPO) (1-4). This heme enzyme is stored at high concentrations in the granules of phagocytic cells (neutrophils and monocytes), and, upon stimulation, it is secreted into both the extracellular milieu and the phagocytic vacuole (3). It is believed that the generation of HOCl by this system constitutes an important defense mechanism against microorganisms (3, 4). However, excessive production of HOCl can also lead to host tissue injury (5), contributing to the development of several diseases, such as atherosclerosis (6, 7) and cancer (8).

[1]

At physiological pH, HOCl is in equilibrium with its conjugate base, hypochorite (OCl^-, pK_a 7.4 at 25°C) (9), (Eq. 2), and both forms appear to be responsible for the oxidation and/or halogenation reactions. It has also been demonstrated that, at acidic conditions, HOCl can be in equilibrium with molecular chlorine (Cl₂,pK_a 3.3) (10) through a reaction that requires Cl^- and H⁺ (Eq. 3) (11). In vitro studies suggest that Cl₂ might be the chlorinating agent that mediates the formation of chlorinated products during phagocytosis (11-13).

[2]

[3]

HOCl is a highly reactive species capable of modifying a variety of biomolecules (5). Free amino and thiol groups of amino acids and peptides constitute important targets for HOCl, yielding unstable chloramines and sulfenyl chloride intermediates, respectively (14-16). Chloramine intermediates are also detected in the reaction of HOCl with exocyclic (RNH₂) and heterocyclic (RNHR) amine functions in DNA bases (14, 17). HOCl reacts with aromatic rings, such as in tyrosine, yielding 3-chlorotyrosine and 3,5-dichlorotyrosine (5, 18, 19). These products have been detected in proteins exposed to MPO or stimulated neutrophils (20) and in low-density lipoprotein isolated from atherosclerotic lesions (19).

Another important target for HOCl is lipids. It is known that HOCl adds across the carbon-carbon double bonds in fatty acids and cholesterol, yielding chlorohydrins (6, 21, 22). HOCl has been shown to induce lipid peroxidation (23). Several groups have detected accumulation of lipid-peroxidation products in liposomes and lipoproteins after incubation with HOCl or MPO-H₂O₂-Cl^- system (24-26). However, the mechanism by which HOCl induces lipid peroxidation remains unclear. The Panasenko group (23, 26, 27) postulated that the presence of lipid hydroperoxides is critical for the initiation of lipid peroxidation in these systems. They have proposed that the reaction of HOCl with lipid hydroperoxides yields free radicals able to cause further oxidation of lipid molecules (27).

It is well known that HOCl reacts with hydrogen peroxide, yielding stoichiometric amounts of singlet molecular oxygen [O₂ (¹Δ_g)] (Eq. 4) (28-30). However, little is known about the reaction of HOCl with other biologically relevant hydroperoxides, such as lipid hydroperoxides, which are the primary products of lipid peroxidation and are also generated in lipoxygenase- and ciclooxygenase-catalyzed reactions (31).

[4]

The aim of this study was to investigate whether O₂ (¹Δ_g) can be generated during the reaction of HOCl with lipid hydroperoxides. We have used hydroperoxides derived from linoleic acid (LAOOH), one of the major fatty acids present in biological systems, and liposomes enriched with phosphatidylcholine hydroperoxides (PCOOH). The generation of O₂ (¹Δ_g) was studied by direct spectroscopic detection and characterization of O₂ (¹Δ_g) monomol emission at the near-infrared region. The reaction mechanism was investigated by measuring the formation of ¹⁸O-labeled O₂ (¹Δ_g) [¹⁸O₂ (¹Δ_g)] in the reaction of ¹⁸O-labeled linoleic acid hydroperoxide (LA¹⁸O¹⁸OH) with HOCl by chemical trapping and mass-spectrometry analysis and of radical intermediates by rapid-mixing continuous-flow electron paramagnetic resonance spectroscopy (EPR).

Results

Singlet Oxygen Generation in the Reaction of HOCl with LAOOH. The generation of O₂ (¹Δ_g) in the reaction of HOCl with LAOOH was investigated by monitoring the light emission at 1,270 nm corresponding to O₂ (¹Δ_g) monomolecular decay (¹Δ_g → ³Σ_g^-) (Eq. 5). For comparison, the light emission for the reaction of HOCl with other hydroperoxides, such as H₂O₂, cumene hydroperoxide (CuOOH), and tert-butylhydroperoxide (t-BuOOH) were also studied (Fig. 1). As expected, an intense light emission was observed upon injection of HOCl (final concentration, 1 mM) into a solution of H₂O₂ (1 mM prepared in D₂O), consistent with the stoichiometric generation of O₂ (¹Δ_g) in this reaction (Fig. 1A) (28-30). Similarly, the injection of HOCl (final concentration, 1 mM) into a solution of LAOOH (1 mM solubilized in 0.1 M phosphate buffer, pH 7.4, in D₂O) also produced a rapid increase in light emission whose intensity was ≈30 times lower compared with the reaction with H₂O₂ (Fig. 1B). In contrast, the injection of HOCl into a solution containing t-BuOOH showed a very small light emission (Fig. 1C), and the injection of HOCl into a solution of CuOOH did not yield any light emission (Fig. 1D). These results indicate that the reaction of HOCl with secondary hydroperoxides, such as LAOOH, generates O₂ (¹Δ_g), whereas the reaction with tertiary hydroperoxides does not yield O₂ (¹Δ_g). The small emission signal observed upon reaction of HOCl with t-BuOOH is probably due to contaminant H₂O₂ normally present in commercial t-BuOOH.

[5]

Fig. 1.

Monomol light emission of O₂ (¹Δ_g) generated in the reaction of HOCl with different hydroperoxides. Light emission observed upon injection of 0.17 ml of 10 mM HOCl (final concentrationm, 1 mM) into 1.5 ml of 1 mM H₂O₂ (A), 1 mM LAOOH (B), 1 mM t-BuOOH (C), and 1 mM CuOOH (D).

The amount of O₂ (¹Δ_g) generated in the reaction of HOCl with LAOOH was estimated by the integration of the light-emission signal at 1,270 nm by using DHPNO₂, a water-soluble O₂ (¹Δ_g) generator, as a reference (see Fig. 8, which is published as supporting information on the PNAS web site). When 1 mM of LAOOH was reacted with various concentrations of HOCl, a concentration-dependent increase in the amount of O₂ (¹Δ_g) was observed from 0.2 to 2.0 mM HOCl, reaching a plateau >2.0 mM (see Fig. 9, which is published as supporting information on the PNAS web site). The maximum concentration of O₂ (¹Δ_g) at the plateau was 133 ± 16 μM, which corresponds to a yield of 13 ± 2% of O₂ (¹Δ_g).

Singlet-Oxygen Spectrum in the Near-Infrared Region. The generation of O₂ (¹Δ_g) by the reaction of HOCl with LAOOH was also confirmed by recording the spectrum of the light emitted in the near-infrared region (Fig. 2A). For comparative purposes, the spectrum of the O₂ (¹Δ_g) generated by H₂O₂ with HOCl was also acquired (Fig. 2B). Both spectra showed an emission band with maximum intensity at 1,270 nm, characteristic of O₂ (¹Δ_g) monomolecular decay. Further evidences that the light emitted in the reaction corresponds to O₂ (¹Δ_g) were obtained by testing the effect of nondeuterated vs. deuterated solvent and the effect of NaN₃, a known O₂ (¹Δ_g) quencher (32) (see Fig. 10, which is published as supporting information on the PNAS web site). The intensity of the light emitted in the reaction in 100% D₂O was about eight times higher than in the reaction in 12% D₂O, consistent with the longer lifetime of O₂ (¹Δ_g) in D₂O than in H₂O (33, 34) (Fig. 10A). The quenching effect of 1 mM of NaN₃ (75% inhibition) on the chemiluminescent reaction of HOCl with LAOOH was also observed (Fig. 10B).

Fig. 2.

Monomol light-emission spectrum of O₂ (¹Δ_g) generated in the reaction of HOCl with LAOOH (A) and H₂O₂ (B).

Light-Emission Measurements in the Reaction of HOCl with Linoleic Acid (LA) or Hydroxy Linoleate (LAOH). To assess whether the hydroperoxyl group of LA is essential for the generation of O₂ (¹Δ_g), we have performed experiments with LA and LAOH (see Fig. 11, which is published as supporting information on the PNAS web site). No emission was observed when HOCl was injected into a solution containing LA, whereas a small emission was observed in the reaction of HOCl with LAOH. Analysis by mass spectrometry showed that this emission is due to the presence of 10% residual LAOOH in the solution of LAOH (see Fig. 12, which is published as supporting information on the PNAS web site).

Effect of pH on Singlet-Oxygen Generation in the Reaction of HOCl with LAOOH. The generation of O₂ (¹Δ_g) in the reaction of H₂O₂ with HOCl is known to vary with pH, showing higher yields at alkaline pH (29). To check whether a similar pH profile is also observed in the reaction of LAOOH with HOCl, we have measured the light emission during the injection of HOCl into a solution containing LAOOH dispersed in phosphate buffer at various pHs (Fig. 3). The light emission was very intense at pH 7.4 and pH 8.0, decreasing at pH 7.0 and pH 9.0. Considering that the pK_a of HOCl is 7.4, these results suggest that the generation of O₂ (¹Δ_g) is favored by the presence of the anionic form of HOCl.

Fig. 3.

Effect of pH on O₂ (¹Δ_g) generation in the reaction of LAOOH with HOCl. Reaction was initiated by injection of 0.17 ml of 10 mM HOCl (final concentration, 1 mM) into 1.5 ml of 1 mM LAOOH solution in 0.1 M phosphate buffer prepared in D₂O at the following pHs 6.0, 7.0, 7.4, 8.0, and 9.0.

Singlet-Oxygen Generation upon Reaction of HOCl with Phospholipid Hydroperoxides in Liposomes. To investigate whether HOCl can also interact with lipid hydroperoxides present in membranes, we have done experiments with unilamellar liposomes containing different concentrations of PCOOH. The presence of increasing amounts of PCOOH in the membrane led to an almost linear increase in the intensity of the light emitted upon injection of HOCl (Fig. 4). Interestingly, the estimated yield of O₂ (¹Δ_g) was augmented at higher concentrations of PCOOH in the membrane. This result shows that the HOCl can also react with phospholipid hydroperoxides present in membranes to generate O₂ (¹Δ_g).

Fig. 4.

Monomol light emission of O₂ (¹Δ_g) observed during injection of HOCl into PC liposomes containing different concentrations of PCOOH. (A) Light emission monitored during injection of 0.17 ml of 100 mM HOCl (final concentration 10 mM) into 1.5 ml of 1 mM PC vesicles containing 0, 10, 20, or 50% PCOOH. (B) The amount of O₂ (¹Δ_g) calculated by integration of the area under emission signal. The percentages indicate the yield of O₂ (¹Δ_g).

Detection of ¹⁸O-Labeled Singlet Oxygen in the Reaction of LA¹⁸O¹⁸OH with HOCl. To characterize the mechanism involved in the generation of O₂ (¹Δ_g) by the reaction of HOCl with LAOOH, we used LA¹⁸O¹⁸OH as a mechanistic tool. Singlet oxygen generated by the reaction of HOCl with LAOOH or LA¹⁸O¹⁸OH was chemically trapped with the anthracene derivative, anthracene-9,10-diyldiethyl sulfate (EAS) (Eq. 6) and the corresponding endoperoxides (EAS^xO^xO, x = 16 or 18), were detected by HPLC coupled to tandem MS (HPLC-MS/MS), as reported in ref. 35.

Figure 6.

The reaction of HOCl with LA¹⁸O¹⁸OH showed the generation of a mixture of three endoperoxides, namely, the completely labeled endoperoxide (EAS¹⁸O¹⁸O), an endoperoxide containing ¹⁸O and ¹⁶O atoms (EAS¹⁶O¹⁸O), and an unlabeled endoperoxide (EAS¹⁶O¹⁶O). Fig. 5 shows the typical chromatograms for EAS^xO^xO analysis by UV and mass detection. Analysis of the products by UV at 215 nm showed a peak corresponding to the endoperoxide at 14.5 min and a peak corresponding to EAS at 19 min. The mass chromatograms obtained by selecting ions with m/z 228, 229, and 230 shows the presence of EAS¹⁶O¹⁶O, EAS¹⁶O¹⁸O, and EAS¹⁸O¹⁸O, respectively. The identity of the ions was confirmed by analyzing the mass spectra of the daughter ions derived from each endoperoxide (see Fig. 13, which is published as supporting information on the PNAS web site).

Fig. 5.

Analysis of EAS endoperoxides formed in the reaction of LA¹⁸O¹⁸OH (5 mM) with HOCl (10 mM) in the presence of EAS (8 mM) by HPLC-MS/MS. (A) HPLC chromatogram monitored at UV 215 nm. Mass chromatograms obtained by selecting the ions at m/z 228 (B), m/z 229 (C), and 230 (D).

In other studies, we have observed that oxygen affects the detection of ¹⁸O₂ (¹Δ_g) generated in the reaction of LA¹⁸O¹⁸OH with metal ions (36) or during thermolysis of an ¹⁸O-labeled naphthalene endoperoxide (35). To determine whether oxygen interferes in the detection of ¹⁸O₂ (¹Δ_g) in this reaction, we have conducted experiments in the presence and absence of oxygen. Fig. 6 shows the relative intensities of the endoperoxides formed in the reaction of HOCl with unlabeled hydroperoxide (LA¹⁶O¹⁶OH) or with LA¹⁸O¹⁸OH. As expected, the reaction of HOCl with LA¹⁶O¹⁶OH yielded a prominent ion at m/z 228, corresponding to EAS¹⁶O¹⁶O (Fig. 6A). In contrast, the reaction of HOCl with LA¹⁸O¹⁸OH conducted in the presence of oxygen showed the formation all three endoperoxides, as observed by the presence of ions at m/z 228, 229, and 230 in the proportion of 40:28:32 (Fig. 6B). Removal of the oxygen from the reaction of HOCl with LA¹⁸O¹⁸OH by a repetitive procedure of freezing and thawing under vacuum led to an expressive increase in the amount of EAS¹⁸O¹⁸O and decrease in the amount of EAS¹⁶O¹⁶O and EAS¹⁶O¹⁸O (Fig. 6C). The proportion of EAS¹⁶O¹⁶O, EAS¹⁶O¹⁸O, and EAS¹⁸O¹⁸O detected after removal of oxygen was 21:17:62.

Fig. 6.

Electrospray ionization (ESI)-MS spectrum of the EAS endoperoxides formed by the reaction of LA^xO^xOH (x = 16 or 18) with HOCl. (A) LA¹⁶O¹⁶OH (5 mM) was reacted with 10 mM HOCl in 0.1 M phosphate buffer, pH 7.4. (B) LA¹⁸O¹⁸OH (5 mM) was reacted with 10 mM HOCl in 0.1 M phosphate buffer, pH 7.4 in 85% D₂O, 10% H₂O, and 5% MeOH. (C) Reaction described in B without oxygen.

Detection of Peroxyl Radical by Continuous-Flow EPR. As recently confirmed by our group, O₂ (¹Δ_g) can be generated by the combination of peroxyl radical, by following the mechanism described by Russell (37). To determine whether a similar type of reaction is involved in the generation of O₂ (¹Δ_g) by the reaction of HOCl with LAOOH, we studied whether peroxyl radicals are formed in this reaction. Fig. 7 shows the experimental and simulated EPR spectra acquired for the reaction of HOCl with LAOOH. Continuous infusion of concentrated solutions of LAOOH (final concentration, 14 mM) and HOCl (final concentration, 14 mM) in phosphate buffer, pH 7.4, containing 25% acetonitrile at room temperature produced a broad doublet signal with a distinctive g-value of 2.014 that is characteristic of peroxyl radicals (Fig. 7A) (38-41). The simulated spectrum showed a hyperfine coupling constant of 4.3 G (Fig. 7B), probably due to the interaction of the radical with the allylic hydrogen. A similar coupling constant was observed by Chamulitrat and Mason (41) during the oxidation of arachidonic acid by lipoxygenase. Thus, this result proves that the reaction of HOCl with LAOOH generates peroxyl radicals.

Fig. 7.

EPR spectrum of linoleate peroxyl radicals. (A) Experimental spectrum obtained during the reaction of HOCl with LAOOH. Spectrometer settings were microwave frequency, 9.650 GHz; microwave power, 10 mW; field-modulation frequency, 100 kHz; field-modulation amplitude, 3 G; receiver gain, 1 × 10⁶; time constant, 164 ms; scan rate, 1.2 G s-1; number of scans, 1. (B) Computer simulation of spectrum A using the following values: aH = 4.3 G; line width, 2.8 G; center of the spectrum, 3,423 G.

Discussion

It is well established that HOCl reacts with hydrogen peroxide to yield stoichiometric amounts of O₂ (¹Δ_g) (28-30). Our results show that HOCl can also react with lipid hydroperoxides, such as fatty acid hydroperoxides or phosphatidylcholine hydroperoxides contained in liposomes, to yield O₂ (¹Δ_g) at physiological pH. The formation of O₂ (¹Δ_g) in the reaction of HOCl with LAOOH was clearly demonstrated by the direct detection and characterization of the O₂ (¹Δ_g) monomol emission at 1,270 nm.

The generation of O₂ (¹Δ_g) was also tested with tertiary hydroperoxides, t-BuOOH or Cu-OOH. The reaction of HOCl with these hydroperoxides did not yield O₂ (¹Δ_g), suggesting that the presence of a hydrogen-α at the carbon to which the hydroperoxide is attached is essential for the generation of O₂ (¹Δ_g). The presence of hydrogen-α is known to be critical for the generation of O₂ (¹Δ_g) by the Russell mechanism (Eq. 7) (37). In this mechanism, primary or secondary peroxyl radicals (ROO^•) react by a cyclic mechanism involving a linear tetraoxide intermediate (ROOOOR) that decomposes to generate a ketone (RO), an alcohol (ROH), and oxygen (36, 37, 42). It has been postulated that this reaction may generate either an electronically excited oxygen molecule (Eq. 7a) or an electronically excited ketone (Eq. 7b). Niu and Mendenhall (43) reported that the yields of O₂ (¹Δ_g), in the case of simple alkylhydroperoxides, ranged from 3.9% to 14.0%. In contrast, the yields of excited carbonyls were 10³ to 10⁴ lower, suggesting that the self-reaction of peroxyl radical generates predominantly O₂ (¹Δ_g).

Figure 9.

The mechanism involved in the generation of O₂ (¹Δ_g) by the reaction of HOCl with LAOOH was studied by using ¹⁸O-labeled hydroperoxide. The reaction of LA¹⁸O¹⁸OH with HOCl in the presence of EAS yielded a mixture of endoperoxides containing ¹⁸O and/or ¹⁶O atoms (EAS¹⁶O¹⁶O, EAS¹⁶O¹⁸O, and EAS¹⁸O¹⁸O). Comparison of the relative amounts of EAS¹⁶O¹⁶O/EAS¹⁶O¹⁸O/EAS¹⁸O¹⁸O detected before and after removal of oxygen showed an expressive increase in the amount of EAS¹⁸O¹⁸O and decrease in the amount of both EAS¹⁶O¹⁶O and EAS¹⁸O¹⁶O. These results indicate that oxygen affects the detection of ¹⁸O₂ (¹Δ_g). It can be also concluded that the reaction of HOCl with LA¹⁸O¹⁸OH yields mainly ¹⁸O₂ (¹Δ_g) and that the oxygen atoms in the ¹⁸O₂ (¹Δ_g) molecule are derived from the hydroperoxide moiety.

The influence of oxygen in the detection of ¹⁸O₂ (¹Δ_g) may be explained by two mechanisms. One is an energy-transfer mechanism between ¹⁸O₂ (¹Δ_g) and ¹⁶O₂ (³Σ_g^-), yielding ¹⁶O₂ (¹Δ_g) and ¹⁸O₂ (³Σ_g^-), as recently demonstrated for aqueous systems by Martinez et al. (35). Another mechanism takes into account the generation of ¹⁸O-labeled peroxyl radicals (LA¹⁸O¹⁸O^•) in the reaction of HOCl with LA¹⁸O¹⁸OH as precursors of ¹⁸O₂ (¹Δ_g). As proposed by Chan (44), the labeled oxygen atoms in LA¹⁸O¹⁸O^• can exchange with the surrounding ¹⁶O_2, yielding unlabeled peroxyl radicals (LA¹⁶O¹⁶O^•). Accordingly, the formation of LA¹⁶O¹⁶O^• may explain the formation of ¹⁶O₂ (¹Δ_g) and the formation of O₂ (¹Δ_g) containing a mixture of ¹⁶O and ¹⁸O atoms, in the presence of oxygen.

On the basis of ¹⁸O₂ (¹Δ_g) detection in the reaction of HOCl with LA¹⁸O¹⁸OH, we have studied the possible mechanisms involved in its generation. The generation of ¹⁸O₂ (¹Δ_g) could occur through a mechanism similar to the reaction of HOCl with H₂O₂. Cahil and Taube (45), using ¹⁸O-labeled hydroperoxide (H¹⁸O¹⁸OH), demonstrated that the oxygen atoms in the oxygen molecule generated by the reaction of HOCl with H₂O₂ were completely labeled. The mechanism proposed for this reaction involves a nucleophilic attack of HO₂^- on the chlorine atom of HOCl to form a [HOO-Cl-OH]^- intermediate, which then generates O₂ (¹Δ_g) by a two-electron transfer (Eq. 8) (29). In analogy, we could suggest the possibility of a nucleophilic attack of LA¹⁸O¹⁸O^- on HOCl to yield a [LA¹⁸O¹⁸O-Cl-OH] intermediate that decomposes, generating ¹⁸O₂ (¹Δ_g) (Eq. 9). However this mechanism does not explain the requirement of a hydrogen-α in the hydroperoxide structure for the formation of O₂ (¹Δ_g) and the detection of O₂ (¹Δ_g), containing a mixture of ¹⁶O and ¹⁸O atoms.

[8]

[9]

An alternative mechanism by which the formation of ¹⁸O₂ (¹Δ_g) in the reaction of HOCl with LA¹⁸O¹⁸OH could be explained is the Russell mechanism (Eq. 7) (37). This mechanism requires the generation of ¹⁸O-labeled LA peroxyl radicals (LA¹⁸O¹⁸O^•). Indeed, peroxyl-radical intermediates were directly detected in the reaction of HOCl with LAOOH by continuous-flow EPR.

The detection of peroxyl radicals, the requirement of a hydrogen-α in the hydroperoxide structure and the O₂ (¹Δ_g) yield of 13 ± 2%, which is very close to the estimated yield of a Russell mechanism, strongly points to the involvement of the Russell mechanism in the generation of O₂ (¹Δ_g) by the reaction of HOCl with LAOOH. However, the formation of peroxyl radical by direct one-electron oxidation of LAOOH by HOCl is thermodynamically unfavorable. HOCl is a relatively poor one-electron oxidant, having an estimated one-electron reduction potential at pH 7 (E°′) in the range of 0.17-0.25 V (46) and, therefore, not able to promote the one-electron oxidation of LAOOH to LAOO^•, which has a reduction potential of 1.0 V (47, 48).

Alternatively, peroxyl radicals could be generated by a mechanism involving chlorine-atom transfer between HOCl and LAOOH. This type of reaction is reported to occur in the reaction of HOCl with NO₂^-, yielding a NO₂Cl intermediate (49) that decomposes, generating Cl^• and ^•NO₂ (50). Similar reactions also occur during interaction of HOCl with thiols (RSH) and amines (RNH₂), yielding the corresponding chlorinated products RSCl and RNHCl. These intermediates are relatively unstable and are suggested to decompose thermally or in the presence of metal ions to generate the radical intermediates RS^• and Cl^• (51) or RNH^• and Cl^• (52, 53).

Analogous to the reactions described above, a chlorine-atom transfer from HOCl to LAOOH could yield unstable chlorinated peroxide intermediates (LAOOCl) (Eq. 10) that may undergo homolytic cleavage to generate LAO^• and ^•OCl (Eq. 11) or LAOO^• and Cl^• (Eq. 12). The Cl^• radical is a strong oxidant (E° Cl^•/Cl = 2.2-2.6 V) (54) capable of promoting the direct oxidation of LAOOH to LAOO^•. In the same way, the alkoxyl radical LAO^•, which has a reduction potential of 1.6 V (47, 48), could also oxidize LAOOH to LAOO^•.

[10]

[11]

[12]

Another mechanism by which HOCl could promote the formation of peroxyl radicals is the reaction with O₂^•- or Fe²⁺, generating ^•OH (55-57). The hydroxyl radical is considered to be one of the most powerful oxidants, with E°′ ^•OH, H/H₂O = 2.31 V (58) and, therefore, capable of oxidizing LAOOH to LAOO^•. However, the involvement of contaminant metal ions or ^•OH in our system can be discarded, because experiments done with Chelex-treated buffer and manitol (1-10 mM) did not affect the formation of O₂ (¹Δ_g)by the reaction of HOCl with LAOOH (data not shown).

Overall, our results clearly demonstrated that HOCl reacts with both fatty acid hydroperoxides and phospholipid hydroperoxides present in membranes to generate O₂ (¹Δ_g). Moreover, our study also provided direct evidence for the generation of lipid peroxyl-radical intermediates in the reaction of HOCl with LAOOH. The physiological relevance of these findings remains to be clarified. Nonetheless, the generation of both lipid peroxyl radicals and O₂ (¹Δ_g) in this reaction may be an additional important mechanism that contributes to the microbicidal activity of HOCl during phagocytosis as well as for the propagation of lipid peroxidation in pathologic conditions involving inflammatory processes, such as atherosclerosis and cancer. As recently reviewed, O₂ (¹Δ_g) is a highly reactive species that can oxidize a variety of biomolecules (59) and can modulate cell-signaling cascades and gene expression (60).

The importance of our study is strengthened by a number of evidences indicating a link between inflammatory disorders to lipid peroxidation and the formation of biologically active oxidized lipids (24, 61, 62). It has been demonstrated that MPO functions as a major catalyst for initiation of lipid peroxidation at sites of inflammation (61, 62). Our findings add a pathway that can contribute to the promotion of lipid peroxidation, providing further insights into the potential involvement of O₂ (¹Δ_g) in oxidative reactions mediated by HOCl/lipid hydroperoxides in biological systems.

Conclusions

In summary, the results presented in this article clearly show that the reaction of HOCl with LAOOH or PCOOH generates O₂ (¹Δ_g). The requirement of a hydrogen-α for the generation of O₂ (¹Δ_g), the formation of ¹⁸O₂ (¹Δ_g) in the reaction of HOCl with LA¹⁸O¹⁸OH, and the direct detection of peroxyl radicals by EPR, strongly suggest the involvement of a Russell mechanism. The generation of O₂ (¹Δ_g) by the reaction of HOCl with lipid hydroperoxides may be another important reaction that occurs at sites of inflammation.

Materials and Methods

Materials. LA, egg-yolk phosphatidylcholine (PC), and sodium azide were obtained from Sigma. Deuterium oxide (D₂O, 99.9%) was from Aldrich (Steinheim, Germany). H₂O₂ was purchased from Peróxidos do Brasil (Paraná, Brazil). HOCl stock solution (≈0.2-0.4 M) was prepared by vacuum distillation at 40°C of commercial hypochlorite solution acidified to pH 6 with phosphoric acid. The HOCl concentration was determined spectrophotometrically (ε₂₉₂ = 350 M^-1·cm^-1 at pH 12) (63). LA¹⁸O¹⁸OH and PCOOH were synthesized as described in refs. 36 and 64. The disodium salt of anthracene, EAS, and the endoperoxide of N,N′-di(2,3-dihydroxypropyl)-1,4-naphthalenedipropionamide (DHPNO₂) were synthesized as described by Di Mascio and Sies (65) and Martinez et al. (66), respectively. All of the other solvents were of HPLC grade and were acquired from Merck.

Singlet-Oxygen Monomol Light-Emission Measurements. Monomolecular photoemission of O₂ (¹Δ_g) at 1270 nm was monitored by a photocounting apparatus described in refs. 36 and 67. For details, see Supporting Materials and Methods, which is published as supporting information on the PNAS web site.

Singlet-Oxygen Spectrum in the Infrared Region. The spectrum of the light emitted in the near-infrared region was recorded with the photomultiplier described above. For details, see Supporting Materials and Methods.

Calculation of Singlet-Oxygen Yield. The yield of O₂ (¹Δ_g) was calculated by using DHPNO₂, a water-soluble endoperoxide, as a clean source of O₂ (¹Δ_g) (see Supporting Materials and Methods and Fig. 8).

Preparation of Liposomes Containing Phospholipid Hydroperoxides. Liposomes of defined size (100 nm) were prepared by an extrusion technique (64). For details, see Supporting Materials and Methods.

¹⁸O-Labeled Singlet-Oxygen Detection by Chemical Trapping. Singlet oxygen generated in the reaction of LA¹⁸O¹⁸OH with HOCl was chemically trapped with EAS. EASO₂ was analyzed by HPLC-MS/MS. Details of the method are described in Supporting Materials and Methods.

epr. EPR spectra of transient species formed at room temperature (25 ± 2°C) were obtained with an EMX spectrometer equipped with a ER 4117 D-MVT dielectric-mixing resonator (Bruker, Billerica, MA) (68). The magnetic field was calibrated with 4-hydroxy-2,2,6,6-tetramethylpiperidine-N-oxyl (TEMPOL) which has a g-value of 2.0056 (69). Computer simulations of spectra were performed by using the program winsim (EPR calculations for MS-Windows NT 95, version 0.96 from Public EPR Software Tools (P.E.S.T.) written by Duling (70). For details, see Supporting Materials and Methods.