Nitroxidation, nitration, and oxidation of a BODIPY

ARTICLE IN PRESS

Nitric Oxide xxx (2006) xxx–xxx www.elsevier.com/locate/yniox

Nitroxidation, nitration, and oxidation of a BODIPY fluorophore by RNOS and ROS Adrian C. Nicolescu 1, Qian Li 1, Laurie Brown, Gregory R.J. Thatcher

*

Department of Medicinal Chemistry and Pharmacognosy, College of Pharmacy, University of Illinois at Chicago, 833 S. Wood St., Chicago, IL 60612-7231, USA Received 10 September 2005; revised 11 December 2005

Abstract BODIPY C11 581/591 (BODIPY11) represents a sensitive probe for quantification of relative antioxidant capacity. However, the mechanism of BODIPY11 fluorescence decay in the presence of reactive oxygen species (ROS) and reactive nitrogen oxide species (RNOS) requires clarification. Azo-initiators provide a continuous source of peroxyl radicals that in simple, aerobic, homogeneous, buffered solution simulate lipid peroxyl radical formation. Inhibition of BODIPY11 fluorescence decay was assayed and quantified for several families of antioxidants, including phenols, NO donors, and thiols. Fluorescence decay of BODIPY11 in these systems demonstrated similar patterns of antioxidant activity to those observed in classical oxygen pressure measurements, and provided a readily applied quantification of antioxidant capacity and mechanistic information, which was analyzed by measurement of induction periods, initial rates, and net oxidation. LC/MS analysis confirmed that peroxyl radical-induced irreversible fluorescence decay of the BODIPY11 fluorophore is due to oxidative cleavage of the activated phenyldiene side chain. The behavior of BODIPY11 towards RNOS was more complex, even in these simple systems. Incubation of BODIPY11 with bolus peroxynitrite or a sydnonimine peroxynitrite source produced a variety of novel products, characterized by LC/MS, derived from oxidative cleavage, nitroxidation, and nitration reactions. The ‘‘NO scavenger’’ PTIO reinforced the antioxidant activity of NO, and inhibited BODIPY11 oxidation induced by the sydnonimine. These observations suggest that BODIPY11 is a well-behaved fluorescence probe for peroxidation and antioxidant studies, but that for study of RNOS even co-application of fluorescence decay with LC/MS measurements requires careful analysis and interpretation. 2006 Elsevier Inc. All rights reserved. Keywords: Antioxidant; Fluorescent probe; Nitroxidation; Nitration; Oxidation; Peroxynitrite; ROS; RNOS

Oxidative stress denotes an imbalance between the production of oxidants and the defense systems of an organism [1]. Reactive oxygen species (ROS) and reactive nitrogen oxide species (RNOS) are important oxidants in vivo. Amongst ROS and RNOS, peroxyl radicals resulting from lipid peroxidation radical chain reactions, and peroxynitrite (ONOO) resulting from the diffusion controlled reaction of nitric oxide (NO) and superoxide radical anion, are important cellular components of oxidative stress. Peroxyl radicals are toxicologically relevant species since they are involved not only in the disruption of cell membrane integ*

1

Corresponding author. Fax: +1 312 996 7107. E-mail address: [email protected] (G.R.J. Thatcher). These authors contributed equally to this work.

1089-8603/$ - see front matter 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.niox.2006.01.010

rity, but also in the oxidation of membrane proteins and the inactivation of receptors and membrane-bound enzymes. Both peroxynitrite and organic peroxyl radicals (ROO) are important reactive species likely to play a role in a number of pathophysiological conditions, such as neurodegenerative disorders, aging, reperfusion injury after ischaemia, and atherosclerosis. Oxidation can be initiated in a number of ways, including ionizing radiation, chemical reaction, enzyme activity, and through transition metal catalysis. A number of initiators have been used to mimic, in vitro, the oxidative stress that leads to lipid peroxidation in vivo, including: (a) azoinitiators; (b) xanthine oxidase; (c) lipoxygenase; (d) peroxynitrite; (e) cupric and cuprous ions; and (f) ferric and ferrous ions [2–9]. The water- and lipid-soluble thermal

ARTICLE IN PRESS 2

A.C. Nicolescu et al. / Nitric Oxide xxx (2006) xxx–xxx

radical sources 2,2 0 -azobis(2-methylpropionamidine)dihydrochloride (ABAP)2 and 2,2 0 -azobis(2,4-dimethylvaleronitrile) (AMVN), respectively, provide reliable initiators for peroxidation studies. Antioxidants, which help cells to cope with oxidative stress by inhibiting damage due to ROS, RNOS, and other free radicals, have been linked to disease prevention. Proteins, enzymes, and a variety of small molecules manifest antioxidant activity. Lipid-soluble antioxidants, in particular, play a major role in protecting membranes and lipoproteins. The assessment of classical antioxidant activity requires measurement of absolute or relative peroxidation rates. Several methods have been used for the quantification of peroxidation, including the spectrophotometric monitoring of formation of conjugated dienes [10]; HPLC and GC monitoring of primary (hydroperoxide) and secondary (aldehyde) peroxidation product formation [11]; spectrophotometric and HPLC monitoring of thiobarbituric acid adduct formation [2]; monitoring redox-sensitive dyes [12–21]; and measuring oxygen uptake by a peroxidizing lipid [3,22]. A suitable method for quantifying the inhibitory effect of antioxidants ideally should offer mechanistic insights into the interaction between the antioxidants and the peroxidizing or oxidizing substrate. One method that meets these requirements employs an oxygen pressure transducer for the measurement of the oxygen uptake during the oxidation of lipid substrates [22–24], but this method requires dedicated instrumentation that is not suitable for parallel processing of multiple samples. Spectrofluorometric methods provide for monitoring of lipid peroxidation with high sensitivity and for parallel measurements on multiple samples. The boron dipyrromethene difluoride (BODIPY) fluorophore is insensitive to pH variation, is thermally stable and relatively insensitive to air oxidation, and possesses good spectrofluorometric characteristics (i.e., a high fluorescence quantum yield (0.9); absorbance and emission wavelength maxima >500 nm) [12–21]. A BODIPY fluorescent probe finding increasing use in vitro both in model and cell-based systems is BODIPY C11 581/591 (BODIPY11), which possesses a 4-phenyl-1,3butadienyl group attached to a conjugated pyrrole system

susceptible to oxidation. We have used BODIPY11 previously to measure lipid peroxidation induced by azo-initiators in phospholipid liposomes [2,6]. In this paper, the reaction kinetics and products of the BODIPY11 probe in response to ROS and RNOS generated from azo-initiators, NO, and peroxynitrite are examined. Given the increasing usage of the BODIPY11 fluorescent probe in more complex biological systems, a better understanding of reactivity and products in simple model systems is required. Using a variety of antioxidants, including NO and related species, BODIPY11 was shown to be a simple, useful reporter for measuring relative antioxidant efficiency, however, in the presence of RNOS, behavior was more complex, with products from oxidation, nitroxidation, and nitration of the conjugated diene moiety observed. Experimental procedures All reactions, unless otherwise specified, were performed in 40% (v/v) acetonitrile in 10 mM phosphate-buffered saline (PBS, 50 mM NaCl) pH 7.4 at 37 C. The PBS stock solution was stored over Chelex resin to minimize effects due to trace transition metals. Chemicals All chemicals, unless otherwise stated, were obtained from Sigma (St. Louis, MO), Aldrich Chemicals (Milwaukee, WI), or BDH (Toronto, Canada). 4,4-Difluoro-5-(-4phenyl-1,3-butadienyl)-4-bora-3a,4a-diaza-s-indacene-3undecanoic acid (BODIPY11) was obtained from Molecular Probes (Eugene, OR). Peroxynitrite was synthesized by two routes: (1) from the reaction of i-amyl nitrite (IAN) with aqueous H2O2 followed by exhaustive washing to remove IAN [6,7]; (2) from the reaction of aqueous NaNO2 with an acidic H2O2 solution with rapid alkaline quenching [8]. In both cases, excess H2O2 was removed with MnO2 and UV–Vis spectroscopy was used to quantify peroxynitrite and check for contamination; typically NO2 and IAN are impurities in the respective preparations [9]. SIN-1 was synthesized by minor adaptation of literature procedures [25]. Oxygen pressure transducer measurements

2 Abbreviations used: ABAP, 2,2 0 azobis(2-methylpropionamide)hydrochloride; AMVN, 2,2 0 -azobis(2,4-dimethyl-valeronitrile); AUC, area under the curve; BODIPY11, 4,4-difluoro-5-(4-phenyl-1,3-butadienyl)-4bora-3a,4a-diaza-s-indacene-3-undecanoic acid; DCN, 1-decyl nitrite; DEA/NO, diethylamine NONOate; DETA/NO, diethylenetriamine NONOate; DLPC, dilinoleoyl-phosphatidylcholine; FU, fluorescence units; GSNO, S-nitrosoglutathione; 5HT, serotonin (5-hydroxytryptamine); IAN, i-amyl nitrite; LC/MS, liquid chromatography coupled with electrospray mass spectroscopy; MeCN, acetonitrile; NONOate, diazeniumdiolate salt; PBS, phosphate-buffered saline; PEN, 2-phenoxy1-ethyl nitrite; PTIO, 2-phenyl-4,4,5,5-tetramethylimidazoline-1-oxyl 3oxide; ROO, peroxyl radical; ROS, reactive oxygen species; RNOS, reactive nitrogen oxide species; SDS, sodium dodecyl sulfate; SIN-1, morpholinosydnonimine; SPE/NO, spermine NONOate; a-TcOH, atocopherol; Trolox C, 6-hydroxy-2,5,7,8-tetramethylchromane-2-carboxylic acid.

The dual channel pressure transducer system has been previously described [6,22]. The reaction vessel (10 ml) was shaken in a 37 C bath fitted with a thermostat. Experiments with micelles used 2 ml volume linoleic acid solution in 0.5 M SDS in 10 mM PBS, pH 7.4. An equal volume of 2 ml PBS was used in the reference cell. Reaction was initiated by addition of 50 ll ABAP stock solution in 10 mM PBS, pH 7.4. When the rate of oxygen uptake became constant, a small volume (10 ll) of Trolox C in 0.5 M SDS solution was added. After the rate of oxygen uptake returned to that before the addition of Trolox C, the same volume of tested compound solution in SDS was added and the rate of oxygen uptake monitored. All

ARTICLE IN PRESS A.C. Nicolescu et al. / Nitric Oxide xxx (2006) xxx–xxx

concentrations of substrate, initiator, and inhibitors were calculated using the micellar reaction volume of 2.54 · 104 L corresponding to 2.0 ml of 0.5 M SDS [26]. For experiments with DLPC liposomes (7.7– 10 · 105 mol) containing AMVN (5.0–8.6 · 106 mol), the liposome emulsion prepared in 10 mM PBS (pH 7.4; 2 ml) was added to the sample cell, and antioxidants were added as described above. The oxygen uptake was corrected for nitrogen and oxygen evolution from the azo-initiator and peroxyl radical recombination, respectively, and oxygen consumption by the initiator.

3

for 3 h. The product mixtures from the reactions of BODIPY11 with ABAP or peroxynitrite were loaded onto PrepSep C18 solid-phase extraction cartridges (0.5 ml), washed with water to remove inorganic salts, and eluted with methanol. The neutral loss of a 20 Da fragment was indicative of BODIPY11 derivatives and these signals were further analyzed by MS–MS. For the isotope exchange experiments, an equal volume of D2O was added to the eluate and the mixture was directly injected into the mass spectrometer for the quantification of isotope peaks of BODIPY11 derivatives. The results were compared with simulated isotope distributions.

Fluorescence assay Results Fluorometric measurements were performed on a Spectramax Gemini XS spectrofluorometer (Molecular Devices) using a quartz covered 96-well glass microplate in triplicate (excitation wavelength 540 nm; emission wavelength at 600 nm; cutoff filter at 590 nm). Aliquots (5 ll) of inhibitor and BODIPY11 (200 nM) stock solutions prepared in 40% acetonitrile/60% 10 mM PBS were added to each well (0.2 ml final reaction volume) and incubated at 37 C. The reaction was initiated by addition of small volumes (10 ll) of ABAP and SIN-1 stock solution in PBS or peroxynitrite in 10 mM NaOH (final pH of reactions solutions was assayed to avoid artifacts due to adjuvant induced pH changes), and the fluorescence decay of the probe was monitored to completion. The emission intensity in relative fluorescence units (RFU) was measured with time, and normalized relative to 100% RFU immediately prior to addition of initiator. LC/MS and LC/MS/MS analyses All LC/MS and LC/MS/MS analyses were performed using an Agilent LC/MSD Ion Trap SL mass spectrometer equipped with an Agilent 1100 HPLC system. HPLC method 1: Agilent Zorbax Rx-C8 column (4.6 mm · 250 mm, 5 lm), water containing 0.1% formic acid (A) and methanol (B) as mobile phase, a 1.5 min isocratic elution at 1% A increasing to 3% A in 1.5 min followed by a gradient of 3–20% A over 7 min. HPLC method 2: Supelco ODS-2C18 column (4.6 mm · 250 mm, 5 lm) eluting with a linear gradient of 10–95% acetonitrile over 60 min, the counter solvent was water containing 2.5 mM ammonium acetate. HPLC method 3: Supelco ODS-2C18 column (4.6 mm · 250 mm, 5 lm), water containing 0.3% acetic acid (A) and methanol (B) as mobile phase, a 8 min isocratic elution at 50% A increasing to 90% A in 12 min then keeping over 10 min. The flow rate was 0.8 ml/min split to 0.2 ml/min prior to the mass spectrometer. Method 2 was used for BODIPY derivative analysis; method 3 was used for ferrulic acid. LC/MS analysis was performed for the reactions of BODIPY11 (40 lM) with: bolus peroxynitrite (0.5 and 1 mM) from either synthetic procedure; SIN-1 (1 mM) incubated for 6 h; or ABAP (20 mM) incubated at 37 C

Antioxidants and O2 consumption Lipid peroxidation by definition involves the consumption of oxygen by a readily oxidizable lipid substrate: the measurement of the oxygen uptake provides quantitative mechanistic information. The pressure transducer method is very reliable for kinetic measurements, but labor intensive. SDS micelles of linoleic acid and multilamellar DLPC liposomes were used as lipid peroxidation models. The well-studied antioxidant Trolox C was used to calculate the rate of chain initiation, Ri, employing the expression Ri = n[Inhibitor]/s, where s is the inhibition period produced by a known amount of antioxidant and n represents the number of peroxyl radical equivalents terminated by one equivalent of antioxidant, which in the case of Trolox C is 2 [24,27,28] (Fig. 1). A kinetic chain length (number of substrate molecules oxidized per molecule of radical that initiates the peroxidation chain) greater than three during the inhibition period ensures that the antioxidant reacts mainly with lipid peroxyl radicals rather than radicals derived directly from the azoinitiator. For classical antioxidants that give a clear inhibition period, absolute inhibition rate constants (kinh) can be determined by linear regression from the slopes of the plots of the measured oxygen uptake (DO2) during the inhibition period versus ln (1 t/s), where the slope is kp[L H]/ kinh, where [L H] is the concentration of the lipid substrate and kp is the propagation rate constant for linoleate in lipid bilayers and micelles (36.1 M1 s1 under the experimental conditions) [29,30]. For Trolox C, studied at varied concentrations in SDS micelles, kinh was found 1.66 ± 0.12 · 104 M1 s1 which compares well with the literature [26,31]. In the DLPC liposomal system using the lipid-soluble azo-initiator, AMVN, kinh for Trolox C was found to be 3.08 ± 0.63 · 104 M1 s1 (Figs. 1A and B). The NONOate NO donor, DEA/NO, and the organic nitrites (i-amyl nitrite, IAN; decyl nitrite, DCN; phenoxyethyl nitrite, PEN) inhibited oxygen consumption in lipid peroxidation with defined induction periods (Figs. 1A and B, and Table 1), allowing calculation of inhibition parameters. A relative antioxidant efficiency (RAE) can

ARTICLE IN PRESS 4


Fig. 1. Induction periods observed for classical antioxidants in azo-initiator induced peroxidation of linoleate and oxidation of BODIPY11. (A) Inhibition of oxygen uptake by DEA/NO, IAN, and Trolox C in ABAP (40 mM) induced peroxidation of linoleic acid in SDS micelles. Data for DCN and PEN are omitted for clarity. (B) Inhibition of oxygen uptake by Trolox C in AMVN (7.25 lM) induced lipid peroxidation of DLPC multilamellar liposomes. (C) Inhibition of ABAP-induced fluorescence decay of BODIPY11 by Trolox C in 40% acetonitrile in 10 mM PBS, pH 7.4 at 37 C; inset: linear dependence of induction period on concentration of Trolox C. (D) a-Tocopherol (a-TcOH) inhibition of AMVN-induced fluorescence decay of BODIPY11 in acetonitrile at 37 C; inset: linear dependence of induction period on a-TcOH concentration. Fluorescence units (FU) are normalized to 100% intensity immediately prior to addition of antioxidant or vehicle, with 0% intensity corresponding to complete reaction. Time courses in triplicate (SD 6 5%).

Table 1 Quantitative parameters for inhibition of lipid peroxidation obtained by measurement of oxygen consumption Inhibitor

Concentration (M) a

DEA/NO DEA/NO IANa DCNa PENa a

5

9.16 · 10 5.60 · 108 8.03 · 103 9.18 · 103 2.36 · 103

s (s)

Ri · 108 (Ms1)

Chain length

kinh (M1 s1)

RAE

2460 2430 1920 3360 1110

1.19 3.27 1.18 1.26 1.34

6–21 4–35 12–40 7–23 4–40

1.20 · 104 3.15 · 105 5.27 · 103 3.35 · 103 9.44 · 103

0.72 7.6 0.32 0.20 0.57

In micellar system; otherwise in DLPC liposomal system.

then be defined as the ratio between the rate constant of test inhibitor and the rate constant of Trolox C [32]. Antioxidants and BODIPY11 degradation In simile with the classical oxygen consumption experiment (Figs. 1A and B), in the BODIPY11 degradation assay the antioxidant activity of Trolox C was manifested by an inhibition period, the length of which was linearly proportional to antioxidant concentration (Fig. 1C). The experi-

mental radical generation rate (Rg), calculated using the observed inhibition periods (s) and applying the formula Rg = 2 · [Trolox C]/s, was 3.48 ± 0.62 · 108 Ms1. This is in good agreement with the theoretical value (3.39 · 108 Ms1) calculated from the initial concentration of the azo-compound and using the expression Rg = 2ekd[ABAP], where e (0.43) and kd (1.32 · 106 s1 at 37 C) are the efficacy and the decomposition rate constant for ABAP, respectively [23,33]. After the induction period, when the antioxidant has been consumed, the


fluorescence of BODIPY11 decays with the same rate as that in the absence of the inhibitor. The BODIPY11 fluorescence decay induced by the lipophilic azo-initiator AMVN, studied in acetonitrile, was quenched by a-tocopherol (a-TcOH), with induction periods linearly dependent on a-TcOH concentration (Fig. 1D). The experimental Rg (10.01 ± 2.45 · 108 Ms1) was in good agreement with the calculated value (9.48 · 108 Ms1) using e (0.095) and kd (4.99 · 106 s1 at 37 C) for the decomposition of 0.1 M AMVN [23]. Trolox C (30 lM) also inhibited BODIPY11 fluorescence decay initiated by AMVN in acetonitrile (data not shown) and gave an Rg (11.1 ± 0.6 · 108 Ms1) in agreement with the calculated value. Pseudo-first-order analysis of the rate of BODIPY11 fluorescence decay showed a linear dependence on ABAP concentration in aqueous acetonitrile, yielding a second-order rate constant approximately 2-fold less than in egg phosphatidylcholine liposomes (1 mg/ml) under similar conditions (Fig. 2). The rate dependence on peroxyl radicals rather than ABAP itself is defined by the equation kobs = kb[ROO]ss, where kb is the second-order rate constant for the reaction of BODIPY with the peroxyl radicals generated by ABAP, and [ROO]ss is the steady-state concentration of the tertiary peroxyl radicals. The linear decay of the fluorescence indicates that steady-state conditions are met. [ROO]ss can be calculated using the equation [ROO]ss = (Rg/2kt)1/2, where Rg is the rate of radical generation by ABAP, and 2kt is the second-order rate constant for the self-recombination reaction of two peroxyl radicals (2ROO fi non-radical products). Rg depends on the initial concentration of ABAP and can be determined experimentally using the inhibition period method for different concentrations of a classical antioxidant such as Trolox C. For 30 mM ABAP in 40% acetonitrile/PBS, Rg was 3.48 ± 0.62 · 108 Ms1. The rate constant, 2kt, for neutral

5

tertiary peroxyl radicals in organic solvents at 30 C has been reported to range between 0.1 and 60 · 104 M1s1 [34]. ABAP produces cationic peroxyl radicals that are anticipated to have slower recombination rates due to electrostatic repulsion. Thus, choosing the lower value for 2kt yields kb = 6.2 · 103 M1 s1, which compares to a value of 6.0 · 103 M1 s1 reported for the reaction of BODIPY11 with tertiary peroxyl radicals generated from AMVN in acetonitrile at 37 C [15]. The NO donor, DEA/NO (t1/2 = 2 min at 37 C and pH 7.4), behaved as an apparent, classical antioxidant producing short induction periods for ABAP-induced BODIPY11 oxidation (Fig. 3A). The dependence of s on DEA/NO concentration was not linear. This is expected, as the antioxidant species is NO, not DEA/NO itself. The calculated number of quenched peroxyl radical equivalents (n = [Inhibitor]/(Rgs)) was lower at higher DEA/NO concentration (e.g., 0.41 at 23.7 lM, and 0.15 at 94.0 lM DEA/NO, respectively), because at higher fluxes, the NO in excess of peroxyl radicals is lost through alternative reactions, such as oxidation, or will simply effuse. SPE/NO (t1/2 = 39 min at 37 C) and DETA/NO (t1/2 = 20 h at 37 C) also showed induction periods for inhibition of ABAP-induced BODIPY11 fluorescence decay at higher concentrations (Figs. 3B and C): from these data SPE/ NO (0.15 mM) was calculated to quench 0.48 peroxyl radical equivalents. Thiols have both antioxidant and prooxidant properties and may simply undergo H-atom abstraction or oxidation to terminate chain propagating peroxidation reactions, but the resulting sulfur–oxygen radicals can propagate radical chains [2,6,35]. Cysteine inhibited ABAP-induced fluorescence decay of BODIPY11 with induction times that were not linear in cysteine concentration (Fig. 4A). The calculated number of quenched peroxyl radical equivalents (n) was found to be 0.13. However, BODIPY11 degradation was only retarded in the induction period, compatible with BODIPY11 degradation by electrophilic thiyl radicals (which add readily to dienes) and sulfur–oxygen radicals formed from peroxyl radical scavenging (reactions (3) and (4) lead to loss of BODIPY fluorescence): R + O2 ! ROO

ROO + CysSH ! ROOH + CysS

ð1Þ

ð2Þ

CysS + O2 ! CysSOO ! propagation

CysS + BODIPY–CH@CH–CH@CHR

ð3Þ 0

! BODIPY–CðSCysÞH–CH@CH–HC R0

Fig. 2. Observed rate constants as a function of ABAP concentration, for fluorescence decay of BODIPY11 in aqueous acetonitrile (solid line) and in DLPC liposomes (dashed line); triplicate data showing SD.

ð4Þ

Human serum albumin (HSA) is a major blood plasma protein that contributes to the antioxidant capacity of blood. At concentrations lower than normally present in human blood plasma (i.e., 3–5 g/100 ml or 4.51–7.52 lM), HSA depressed the rate of BODIPY11 fluorescence decay in a concentration-dependent manner (Fig. 4B), acting to retard the BODIPY11 oxidation rate. Only at high concentration (2.26 lM) was an induction period observed.

ARTICLE IN PRESS 6


Fig. 3. Effects of NONOates on ABAP-induced fluorescence decay of BODIPY11 in 40% acetonitrile in 10 mM PBS, pH 7.4, at 37 C: (A) DEA/NO; (B) SPE/NO; (C) DETA/NO. Fluorescence units (FU) are normalized to 100% intensity immediately prior to addition of antioxidant or vehicle, with 0% intensity corresponding to complete reaction. Time courses in triplicate (SD 6 5%).

Fig. 4. Effects of different antioxidants on ABAP-induced fluorescence decay of BODIPY11 in 40% acetonitrile in 10 mM PBS, pH 7.4, at 37 C: (A) cysteine; (B) human serum albumin (HAS); (C) serotonin (5HT); (D) S-nitrosoglutathione (SNOG or GSNO); (E) i-amyl nitrite (IAN). Fluorescence units (FU) are normalized to 100% intensity immediately prior to addition of antioxidant or vehicle, with 0% intensity corresponding to complete reaction. Time courses in triplicate (SD 6 5%).

Serotonin (5HT), an important neurotransmitter, is expected to be a classical phenolic chain-breaking antioxidant, inhibiting BODIPY11 oxidation. 5HT is able to form an N-stabilized phenoxyl radical, but 5HT was observed only to retard the rate of BODIPY11 oxidation, albeit in a concentration-dependent manner (Fig. 4C). In contrast to NONOates, which cleanly and spontaneously generate NO in neutral aqueous solution, S-nitroso and O-nitroso compounds did not show an induction period in inhibition of BODIPY11 oxidation. It has been reported that nitrosothiols may act as effective antioxidants

in vivo; S-nitrosoglutathione (GSNO) was reported to be a 100-fold more potent antioxidant than its thiol parent [36,37]. In azo-induced BOPIPY11 oxidation, GSNO was less effective than cysteine in that an induction period was not observed (Fig. 4D). The formation of antioxidant NO from a nitrosothiol by homolysis is a 1e reduction requiring catalysis by e donors including transition metal ions or photolysis. In the present experiments, the working excitation wavelength (540 nm) and the use of Chelex C to remove adventitious transition metal cations reduce NO release and antioxidant activity of GSNO. In simile with


the nitrosothiol, IAN also retarded BODIPY11 oxidation, albeit with much lower potency than GSNO, compatible with the slow hydrolysis and minor flux of NO produced by organic nitrites (Fig. 4E) [6]. Quantification of antioxidant activity using BODIPY11 The inhibitors used in the peroxyl radical trapping fluorescence assay are categorized in two groups: (i) inhibitors that show definite inhibition periods, and (ii) inhibitors that depress the rate of fluorescence decay without showing inhibition periods. Three semi-quantitative methods were used to quantify the antioxidant effect on ABAP-induced fluorescence decay of BODIPY11. (1) Defined induction periods. The constant rate of peroxyl radical generation in the reaction mixture using an azo-initiator can be measured using Trolox C, which is known to trap two mole equivalents of peroxyl radicals (Rg = 2[Trolox]/sTrolox). The inhibitory capacity is found by multiplying the rate of peroxyl radical generation by the induction period shown by antioxidants (Rgs), which leads to a molar antioxidant capacity (ACs = Rgs/[inhibitor]). In the aqueous acetonitrile system, the rate of peroxyl radical generation from ABAP (30 mM) was found to be 3.70 ± 0.19 · 108 Ms1, thus for DEA/NO, SPE/ NO, and cysteine, ACs can be calculated. For cysteine, ACs = 0.153 ± 0.025 · 103. For NONOates, ACs varies with concentration, since the antioxidant is not the NONOate itself, but NO: the highest values from the collected data are 0.55 · 103 and 0.24 · 103 for SPE/NO and DEA/NO, respectively. (2) Retardation of BODIPY decay. For antioxidants that do not show an inhibition period, the ratio of initial rates for inhibited (v0,inh) and uninhibited, control (v0) reactions yields a measure of relative antioxidant capacity (ACv = 1 v0,inh/v0). The ACv data

7

can be plotted to estimate IC50 values for inhibition of peroxidation by antioxidants, yielding IC50 = 0.2 ± 0.15, 5 ± 2, 80 lM, 4, 0.1, and 2 mM for HSA, serotonin, GSNO, IAN, SPE/NO, and DETA/NO, respectively. (3) Net protection or area-under-the-curve (AUC). The AUC method has been used in a number of studies and approximates the net protection against oxidation with contributions from both chain-breaking antioxidant and retardation activity. The net antioxidant capacity (ACAUC = (AUCinh AUCun)/AUCun) is calculated from the area under the curve of the BODIPY11 fluorescence decay in the presence (AUCinh) and absence (AUCun) of an antioxidant and is dependent on antioxidant concentration (Fig. 5). In this analysis, the phenolic antioxidants Trolox C and 5HT show a linear dependence upon concentration, despite the fact that one is acting as a classical chainbreaking antioxidant and the other as an oxidation retardant. Interestingly, 5HT has a higher antioxidant capacity than Trolox C by this measure. ACAUC values calculated for the thiol antioxidants, cysteine and HSA, showed saturation behavior compatible with the influence of chain propagating and prooxidant activity at higher concentrations. The optimum NONOate chain-breaking antioxidant would release a flux of NO greater than the flux of peroxyl radicals (40 nM1) and maintain this flux over the course of the experiment. Consequently, the short half life of DEA/NO is insufficient to maintain such a flux and therefore saturation behavior is observed.

Peroxynitrite and BODIPY11 degradation BODIPY11 reaction in phosphatidylcholine liposomes with bolus peroxynitrite led to fluorescence decay (Fig. 6A). Similar observations were made in homogeneous

Fig. 5. Antioxidant capacity from AUC measurements of BODIPY fluorescence decay (ACAUC = (AUCinh AUCun)/AUCun). Solid lines (linear fits for Trolox C and 5HT); dashed lines (first-order fits for HSA, cysteine, GSNO); dotted lines (first-order fits for NONOates). For HSA, the concentration is that of cysteine residues.

ARTICLE IN PRESS 8


Fig. 6. Time course of BODIPY11 fluorescence decay in response to peroxynitrite. (A) Repetitive bolus addition (1.31 mM) to egg phospholipid liposomes in PBS. (B) Repetitive bolus addition (shown by arrows) of peroxynitrite (1.61 mM) to methanol (s) and 40% acetonitrile in 10 mM PBS, pH 7.4 (n). (C) SIN-1 effect on BODIPY11 fluorescence in 40% acetonitrile in 10 mM PBS, pH 7.4. Fluorescence units (FU) are normalized to 100% intensity immediately prior to initial addition of peroxynitrite or SIN-1; time courses in triplicate.

solutions of 40% MeCN/PBS and in methanol (Fig. 6B). The diminished BODIPY degradation in methanol was likely the result of the well-known radical scavenging property of alcohols. The fast decomposition of peroxynitrite at neutral pH leads to the production of oxidizing radicals that can add to BODIPY or more rapidly undergo radical–radical reactions followed by hydrolysis to stable NOx anions. The short lifetime of peroxynitrite and the absence of a substrate able to propagate a chain reaction lead to the observed incomplete oxidation of BODIPY11 by bolus peroxynitrite (Figs. 6A and B). The sydnonimine SIN-1 undergoes base-assisted ringopening to generate an intermediate that will release a steady, but non-linear flux of NO with concomitant 1e transfer to an oxidant, usually O2. Thus the simultaneous generation of NO þ O2 in a 1:1 stoichiometry leads to SIN-1 acting as a peroxynitrite donor, except in cases where other oxidants interfere with electron transfer. In aqueous acetonitrile, the decay of BODIPY11 fluorescence was observed to follow the expected kinetic time course for peroxynitrite release from SIN-1 (Fig. 6C).

NO scavengers and BODIPY11 degradation To further explore and compare the response of BODIPY11 to ROS and RNOS and the influence of antioxidants, the effect of NONOate NO donors was assayed in combination with the putative NO trap nitronyl nitroxide, 2-(phenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide (PTIO). In ABAP-initiated oxidation, the induction period of SPE/NO was prolonged by the addition of increasing concentrations of PTIO, thus PTIO did not serve to trap the antioxidant NO released from SPE/NO (Fig. 7A). If PTIO had behaved as a selective trap for NO, the expectation would be for diminution of the antioxidant effect of the NO donor in ABAP-induced oxidation. SPE/NO, alone, was seen to inhibit the oxidation of BODIPY11 by SIN-1 with an induction period at higher concentration (Fig. 7B). In addition, PTIO, alone, efficiently inhibited the decay of BODIPY11 fluorescence induced by SIN-1 (Fig. 7C). The effects of the NO donor and of PTIO on the reaction of SIN-1 with BODIPY11 are anticipated. The maximum flux of ROS and RNOS from peroxynitrite will occur at a stoichiometry of NO=O2 of 1:1; simplistically,

Fig. 7. Effect of NO donors and NO scavengers on ABAP- or SIN-1-induced BODIPY11 fluorescence decay: (A) PTIO effect on SPE/NO inhibition of ABAP induction; (B) SPE/NO effect on SIN-1 induction; (C) PTIO effect on SIN-1 induction. Fluorescence units (FU) are normalized to 100% intensity immediately prior to addition of antioxidant or vehicle, with 0% intensity corresponding to complete reaction. Time courses in triplicate (SD 6 5%).


at other stoichiometries, ROS and RNOS will be trapped by the excess NO or O2 . The observations with PTIO in ABAP-induced oxidation show that PTIO does not act specifically as an NO trap. LC/MS analysis BODIPY11 and its reaction products can be identified in LC/MS/MS analysis by the loss of a characteristic HF fragment (20 Da). Reaction of the watersoluble azo-initiator ABAP (20 mM) with BODIPY11 (40 lM) in 40% MeCN in 10 mM PBS, pH 7.4, gave as major product the carboxylic acid (m/z

9

[M H] = 419) resulting from oxidative cleavage of the phenyldiene moiety (Fig. 8C). This product was previously reported in a 24-h incubation of ABAP in absolute ethanol to be the major product of reaction (>90% by relative intensity) [16]. The minor product of ABAP-induced oxidation, resulting from olefinic bond cleavage adjacent to the phenyl group (m/z 445), was not observed in this study, although several small signals (109 M1 s1 R þ O2 ! ROO k BODIPY

ROO þ BODIPY ! ROO BODIPY k inh

ROO þ inhibitor ! products 2k t

2ROO ! termination products þ O2 Azo-initiated decay of BODIPY11 fluorescence follows first-order kinetics both in the presence, and in the absence of classical chain-breaking antioxidants in simile with the O2 consumption measurements (Fig. 1), indicating steady-state kinetics. Comparison of O2 consumption with BODIPY11 decay demonstrates similar behavior for these classical antioxidants, which compete for peroxyl radicals, regardless of whether the system contains polyunsaturated fatty acid liposomes or is a simple homogenous solution. Depending on the rates of peroxyl radical reaction with the antioxidant inhibitor and BODIPY11, respectively, the profile of fluorescence decay will reflect one of three scenarios: (i) when the rate of peroxyl radical trapping by the inhibitor is much greater than with BODIPY11 (kinh kBODIPY), the rate of BODIPY11 fluorescence decay will be zero order until the inhibitor is consumed; (ii) when the rate of peroxyl radical reaction with BODIPY11 is much greater than that with the inhibitor (kBODIPY kinh), the rate of BODIPY11 fluorescence decay will be first order and the same as the uninhibited reaction, and (iii) when the rate of peroxyl radical reaction with the inhibitor is comparable to that with BODIPY11 (kBODIPY kinh), the rate of

ARTICLE IN PRESS 12


BODIPY fluorescence decay will be first order and retarded compared to the uninhibited reaction. All these types of behavior, predicted for antioxidants that scavenge peroxyl radicals, are observed in this study (Figs. 1, 3, and 4). Moreover, the decay profiles are compatible with the antioxidant mechanisms, elucidated by other methods, for a variety of inhibitors, including NO donors (for which the flux of NO is the determining factor), thiols (which also possess prooxidant activity), and chain-breaking phenolic antioxidants. The measurement of BODIPY11 decay in aqueous acetonitrile may be used reliably to quantify relative antioxidant activity and to classify antioxidant mechanisms (Fig. 5). The antioxidant activity of NO has been well studied [2– 4,41]. The antioxidant potency of NONOates is quantitatively related to the rate of NO release, with a lipid radical chain termination stoichiometry of 0.4–0.5 mol of lipid peroxyl radicals per mole of NO [2]. The nitronyl nitroxide, PTIO, and its derivative, cPTIO, are often used as ‘‘selective’’ NO scavengers [42], and therefore the effect of PTIO on NO antioxidant and SIN-1 prooxidant activity was of interest. PTIO has frequently been employed to trap NO and thence to yield the prooxidant NO2, which would reverse the antioxidant effect of NO. In this work, PTIO did not act as an NO scavenger in co-incubation with SPE/NO and BODIPY11, but instead enhanced the antioxidant activity of NO towards azo-induced BODIPY11 oxidation, compatible with the known antioxidant activity of nitronyl nitroxides (Fig. 7). In reaction with SIN-1, PTIO has previously been proposed to inhibit superoxide formation, thus enhancing NO release, although other mechanisms are now favored [43,44]. In this study, both PTIO and SPE/NO inhibited SIN-1-induced oxidation of BODIPY11, which is explained by inhibition of peroxynitrite formation and scavenging of peroxynitrite and its radical products. The major product derived from peroxyl radical addition to BODIPY was identified by LC/MS analysis as the carboxylate resulting from oxidative cleavage of the olefinic bond adjacent to the pyrrole ring (Figs. 8C and 9) . The identification of this major product is in accord with a previous study in which under different reaction conditions, the minor product of oxidative cleavage of the alternate olefinic bond was also observed [16]. In contrast to the azo-initiated, peroxyl radical-mediated fluorescence decay of BODIPY11, reaction with bolus peroxynitrite elicited a more complex profile of fluorescence decay and of BODIPY11 products (Figs. 6 and 8). The time course of BODIPY11 decay elicited by bolus peroxynitrite confirmed that in a simple homogeneous solution and in the absence of a polyunsaturated lipid substrate, a chain propagation mechanism is not supported (Fig. 6). In aerobic, homogeneous solutions, one decomposition pathway of peroxynitrite will yield NO2 and an oxy-radical (either CO3 ; or HO in the absence of CO2). Whereas, the 1,4-diene moiety of fatty acids readily undergoes H-atom abstraction, the 1,3-diene of BODIPY11 can only undergo radical

addition. Thus, the predicted initial products of reaction of BODIPY11 with peroxynitrite are those of addition of NO2 and of the oxy-radicals; decomposition to final products is driven by reconjugation either by abstraction of H or H+, yielding radical substitution products. An alternative pathway that might lead to some of the same products is electrophilic addition of peroxynitrous acid to BODIPY11 (Fig. 9). Interestingly, simple phenyldiene model compounds were seen to be unreactive towards bolus peroxynitrite, showing that the phenyldiene moiety of BODIPY11 is of enhanced reactivity towards oxidation. The olefinic cleavage product was not the major product observed in reactions with peroxynitrite. The reaction products of BODIPY11 with bolus peroxynitrite were identified by LC/MS as those derived from addition of an oxygen atom and of NO2, as well as the oxidative cleavage product also observed in BODIPY11 peroxidation (Figs. 8 and 9). One candidate structure for the O-addition product (m/z 519) is an oxirane, which is a typical product of olefinic oxidation (Fig. 9B). LC/MS and H/D isotope labeling studies were required to rule out the oxirane in favour of the alternative conjugated enol as the major product of BODIPY nitroxidation. Products of olefinic nitration were identified (m/z 564 and 609), but the expected simple mononitration product (m/z 549) was not observed (Fig. 9). The simple mononitration product would have reduced reactivity towards further radical addition, which suggests that BODIPY11 nitroxidation precedes nitration or that NO2 olefinic addition is reversible. This is also in accord with the observed product distributions at different peroxynitrite concentrations (Fig. 8). There are a number of potential pathways to the enol nitroxidation product (m/z 519), including epoxidation/ring-opening and addition of NO2 (or CO3 ) followed by rearrangement (Fig. 9). Given reasonable alternative pathways, there is no need to invoke simple hydroxyl radical olefinic addition; hydroxyl radical addition is not probable given the known very rapid trapping of peroxynitrite by CO2 [45]. Nitroxidation and nitration of BODIPY11 leads to substitution products that apparently possess the same extended conjugation as the parent, however, their enolate tautomers have broken conjugation, compatible with loss of fluorescence at higher wavelength in these products relative to BODIPY11 itself [43]. The usefulness of bolus peroxynitrite is problematic even in simple model systems because of the very high local concentrations and the influence of mixing. Alternatively, sydnonimines (almost exclusively SIN-1) have been used as a continuous source of peroxynitrite. BODIPY11 was sensitive to SIN-1 showing a decay time course compatible with the known pathway of decomposition of SIN-1 to peroxynitrite, but the BODIPY11 products identified by LC/ MS were not the same as those seen in reactions with bolus peroxynitrite (Figs. 6C and 8). SIN-1 treatment of BODIPY11 gave oxidative olefinic cleavage in simile with BODIPY11 peroxidation, however, BODIPY11-derived aldehyde products were observed for SIN-1, reflecting the relatively


weak oxidizing environment compared to that produced by the azo-initiator. Observation of differences in reaction products from treatment with either bolus peroxynitrite or SIN-1 is not uncommon, for example, SIN-1 does not support significant tyrosine nitration in contrast to bolus peroxynitrite. Drummen et al. reported the detection of the oxidative cleavage product and the BODIPY11-derived enol in cell culture treated with bolus peroxynitrite, but did not detect nitration products, suggesting that NO2 is rapidly scavenged by molecules such as glutathione [16]. In summary, the present study analyzed mechanisms of BODIPY11 fluorescence decay in simple, aerobic, buffered solution, and the reaction products of BODIPY11 with ROS and RNOS (peroxyl radicals and peroxynitrite) involved in biologically relevant oxidative processes. Peroxyl radical-induced fluorescence decay of BODIPY11 in this system demonstrated similar patterns of antioxidant activity to those observed in classical oxygen pressure measurements in micelles and liposomes, and provided both a readily applied quantitation of antioxidant capacity and mechanistic information. No less than six novel nitroxidation and nitration products from the reaction of BODIPY11 with bolus peroxynitrite and a peroxynitrite donor (SIN-1) were identified by LC/MS/MS. The identity of the previously reported nitroxidation product (m/z 519) as the conjugated enol was unambiguously established by H/D isotope exchange. The behavior of BODIPY11 towards RNOS is more complex, even in the simple systems used in this study, yielding a variety of oxidative cleavage and substitution products derived from nitroxidation and nitration reactions. The diverse BODIPY-derived products, dependent upon the identity of the ROS and RNOS, may allow the use of BODIPY11, in combination with LC/MS, to identify the different ROS and RNOS in biological systems, but the complexity of the product distribution must be considered in any analysis. Similarly, the observed behavior of the nitronyl nitroxide PTIO as an antioxidant rather than as an NO scavenger emphasizes the need for caution in analysis of more complex biological systems.

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

Acknowledgments This work was supported in part both by NSERC Grant 245617-01 and NIH Grant CA 102590. Professor Ross Barclay is acknowledged for assistance with oxygen pressure measurements.

[18]

[19]

References [20] [1] B. Halliwell, J.M.C. Gutteridge, Free Radicals in Biology and Medicine, third ed., Oxford University Press, New York, 1999. [2] A.C. Nicolescu, S.I. Zavorin, N.J. Turro, J.N. Reynolds, G.R. Thatcher, Inhibition of lipid peroxidation in synaptosomes and liposomes by nitrates and nitrites, Chem. Res. Toxicol. 15 (2002) 985– 998. [3] V.B. O’Donnell, P.H. Chumley, N. Hogg, A. Bloodsworth, V.M. Darley-Usmar, B.A. Freeman, Nitric oxide inhibition of lipid

[21]

[22]

13

peroxidation: kinetics of reaction with lipid peroxyl radicals and comparison with alpha-tocopherol, Biochemistry 36 (1997) 15216– 15223. H. Rubbo, R. Radi, M. Trujillo, R. Telleri, B. Kalyanaraman, S. Barnes, M. Kirk, B.A. Freeman, Nitric oxide regulation of superoxide and peroxynitrite-dependent lipid peroxidation. Formation of novel nitrogen-containing oxidized lipid derivatives, J. Biol. Chem. 269 (1994) 26066–26075. S.R. Thomas, M.J. Davies, R. Stocker, Oxidation and antioxidation of human low-density lipoprotein and plasma exposed to 3-morpholinosydnonimine and reagent peroxynitrite, Chem. Res. Toxicol. 11 (1998) 484–494. A.C. Nicolescu, J.N. Reynolds, L.R. Barclay, G.R. Thatcher, Organic nitrites and NO: inhibition of lipid peroxidation and radical reactions, Chem. Res. Toxicol. 17 (2004) 185–196. R.M. Uppu, W.A. Pryor, Synthesis of peroxynitrite in a two-phase system using isoamyl nitrite and hydrogen peroxide, Anal. Biochem. 236 (1996) 242–249. J.S. Beckman, T.W. Beckman, J. Chen, P.A. Marshall, B.A. Freeman, Apparent hydroxyl radical production by peroxynitrite: implications for endothelial injury from nitric oxide and superoxide, Proc. Natl. Acad. Sci. USA 87 (1990) 1620–1624. M.N. Hughes, H.G. Niklin, The Chemistry of pernitrites. Part 1. Kinetics of decomposition of pernitrous acid,, J. Chem. Soc. A 12 (1968) 450–452. F.P. Corongiu, S. Banni, Detection of conjugated dienes by second derivative ultraviolet spectrophotometry, Methods Enzymol. 233 (1994) 303–310. J. Liu, H.C. Yeo, S.J. Doniger, B.N. Ames, Assay of aldehydes from lipid peroxidation: gas chromatography-mass spectrometry compared to thiobarbituric acid, Anal. Biochem. 245 (1997) 161–166. Y.M. Naguib, A fluorometric method for measurement of peroxyl radical scavenging activities of lipophilic antioxidants, Anal. Biochem. 265 (1998) 290–298. Y.M. Naguib, A fluorometric method for measurement of oxygen radical-scavenging activity of water-soluble antioxidants, Anal. Biochem. 284 (2000) 93–98. G. Aldini, K.J. Yeum, R.M. Russell, N.I. Krinsky, A method to measure the oxidizability of both the aqueous and lipid compartments of plasma, Free Radic. Biol. Med. 31 (2001) 1043–1050. Y. Yoshida, S. Shimakawa, N. Itoh, E. Niki, Action of DCFH and BODIPY as a probe for radical oxidation in hydrophilic and lipophilic domain, Free Radic. Res. 37 (2003) 861–872. G.P. Drummen, B.M. Gadella, J.A. Post, J.F. Brouwers, Mass spectrometric characterization of the oxidation of the fluorescent lipid peroxidation reporter molecule C11-BODIPY(581/591), Free Radic. Biol. Med. 36 (2004) 1635–1644. G.P. Drummen, M. Makkinje, A.J. Verkleij, J.A. Op den Kamp, J.A. Post, Attenuation of lipid peroxidation by antioxidants in rat-1 fibroblasts: comparison of the lipid peroxidation reporter molecules cis-parinaric acid and C11-BODIPY(581/591) in a biological setting, Biochim. Biophys. Acta 1636 (2004) 136–150. Y. Christova, P.S. James, R. Jones, Lipid diffusion in sperm plasma membranes exposed to peroxidative injury from oxygen free radicals, Mol. Reprod. Dev. 68 (2004) 365–372. J.P. Fabisiak, Y.Y. Tyurina, V.A. Tyurin, J.S. Lazo, V.E. Kagan, Random versus selective membrane phospholipid oxidation in apoptosis: role of phosphatidylserine, Biochemistry 37 (1998) 13781–13790. J.F. Brouwers, B.M. Gadella, In situ detection and localization of lipid peroxidation in individual bovine sperm cells, Free Radic. Biol. Med. 35 (2003) 1382–1391. J.W. Borst, N.V. Visser, O. Kouptsova, A.J. Visser, Oxidation of unsaturated phospholipids in membrane bilayer mixtures is accompanied by membrane fluidity changes, Biochim. Biophys. Acta 1487 (2000) 61–73. D.D.M. Wayner, G.W. Burton, Measurements of individual antioxidants and radical-trapping activity, in: J. Miquel (Ed.), Handbook of

ARTICLE IN PRESS 14

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

A.C. Nicolescu et al. / Nitric Oxide xxx (2006) xxx–xxx Free Radcals and Antioxidants in Biomedicine, CRC Press, Boca Raton, FL, pp. 223–232. L.R. Barclay, M.R. Vinqvist, Membrane peroxidation: inhibiting effects of water-soluble antioxidants on phospholipids of different charge types, Free Radic. Biol. Med. 16 (1994) 779–788. L.R.C. Barclay, K.U. Ingold, Autoxidation of biological molecules. 2. The autoxidation of a model membrane. A comparison of the autoxidation of egg lecithin phosphatidylcholine in water and in chlorobenzene, J. Am. Chem. Soc. 103 (1981) 6478–6485. K. Masuda, Y. Imashiro, T. Kaneko, Mesoionic compounds. I. Synthesis of 3-dialkylamino-sydnonimines, Chem. Pharm. Bull. 18 (1970) 128–132. L.R.C. Barclay, K.A. Baskin, S.J. Locke, T.D. Schaefer, Benzophenone-photosensitized autoxidation of linoleate in solution and sodium dodecyl sulfate micelles, Can. J. Chem. 65 (1987) 2529–2540. G.W. Burton, K.U. Ingold, Autoxidation of biological molecules. 1. The antioxidant activity of vitamin E and related chain-breaking phenolic antioxidants in vitro, J. Am. Chem. Soc. 103 (1981) 6472–6477. T. Doba, G.W. Burton, K.U. Ingold, Antioxidant and co-antioxidant activity of vitamin C. The effect of vitamin C, either alone or in the presence of vitamin E or a water-soluble vitamin E analogue, upon the peroxidation of aqueous multilamellar phospholipid liposomes, Biochim. Biophys. Acta 835 (1985) 298–303. L.R.C. Barclay, S.J. Locke, J.M. MacNeil, J. VanKessel, Quantitative studies of the autoxidation of linoleate monomers sequestered in phosphatidylcholine bilayer. Absolute rate constants in bilayer, Can. J. Chem. 63 (1985) 2633–2638. L.R.C. Barclay, K.A. Baskin, D. Kong, S.J. Locke, Autoxidation of model membrane. The kinetics and mechanism of autoxidation of mixed phospholipid bilayer, Can. J. Chem. 65 (1987) 2541–2549. L. Castle, M.J. Perkins, Inhibition kinetics of chain-breaking phenolic antioxidants in SDS micelles. Evidence that intermicellar diffusion rates may be rate-limiting for hydrophobic inhibitors such as atocopherol, J. Am Chem. Soc. 108 (1986) 6381–6382. W.A. Pryor, J.A. Cornicelli, L.J. Devall, B. Tait, B.K. Trivedi, D.T. Witiak, M. Wu, A rapid screening test to determine the antioxidant potencies of natural and synthetic antioxidants, J. Org. Chem. 58 (1993) 3521–3532. L.R. Barclay, F. Antunes, Y. Egawa, K.L. McAllister, K. Mukai, T. Nishi, M.R. Vinqvist, The efficiency of antioxidants delivered by liposomal transfer, Biochim. Biophys. Acta 1328 (1997) 1–12.

[34] K.U. Ingold, Peroxy radicals, Acc. Chem. Res. 2 (1969) 1–9. [35] H. Moini, L. Packer, N.E. Saris, Antioxidant and prooxidant activities of alpha-lipoic acid and dihydrolipoic acid, Toxicol. Appl. Pharmacol. 182 (2002) 84–90. [36] P. Rauhala, K.P. Mohanakumar, I. Sziraki, A.M. Lin, C.C. Chiueh, S-nitrosothiols and nitric oxide, but not sodium nitroprusside, protect nigrostriatal dopamine neurons against iron-induced oxidative stress in vivo, Synapse 23 (1996) 58–60. [37] P. Rauhala, A.M. Lin, C.C. Chiueh, Neuroprotection by S-nitrosoglutathione of brain dopamine neurons from oxidative stress, FASEB J. 12 (1998) 165–173. [38] V.B. O’Donnell, J.P. Eiserich, P.H. Chumley, M.J. Jablonsky, N.R. Krishna, M. Kirk, S. Barnes, V.M. Darley-Usmar, B.A. Freeman, Nitration of unsaturated fatty acids by nitric oxide-derived reactive nitrogen species peroxynitrite, nitrous acid, nitrogen dioxide, and nitronium ion, Chem. Res. Toxicol. 12 (1999) 83–92. [39] E.H. Pap, G.P. Drummen, V.J. Winter, T.W. Kooij, P. Rijken, K.W. Wirtz, J.A. Op den Kamp, W.J. Hage, J.A. Post, Ratio-fluorescence microscopy of lipid oxidation in living cells using C11-BODIPY(581/ 591), FEBS Lett. 453 (1999) 278–282. [40] G.P. Drummen, L.C. van Liebergen, J.A. Op den Kamp, J.A. Post, C11-BODIPY(581/591), an oxidation-sensitive fluorescent lipid peroxidation probe: (micro)spectroscopic characterization and validation of methodology, Free Radic. Biol. Med. 33 (2002) 473–490. [41] S.P. Goss, N. Hogg, B. Kalyanaraman, The antioxidant effect of spermine NONOate in human low-density lipoprotein, Chem. Res. Toxicol. 8 (1995) 800–806. [42] A. Law, S. Gauthier, R. Quirion, Neuroprotective and neurorescuing effects of isoform-specific nitric oxide synthase inhibitors, nitric oxide scavenger, and antioxidant against beta-amyloid toxicity, Br. J. Pharmacol. 133 (2001) 1114–1124. [43] S. Goldstein, A. Russo, A. Samuni, Reactions of PTIO and carboxyPTIO with NO, NO2, and O2 , J. Biol. Chem. 278 (2003) 50949– 50955. [44] R.J. Singh, N. Hogg, J. Joseph, E. Konorev, B. Kalyanaraman, The peroxynitrite generator, SIN-1, becomes a nitric oxide donor in the presence of electron acceptors, Arch. Biochem. Biophys. 361 (1999) 331–339. [45] S. Goldstein, G. Czapski, J. Lind, G. Merenyi, Carbonate radical ion is the only observable intermediate in the reaction of peroxynitrite with CO2, Chem. Res. Toxicol. 14 (2001) 1273–1276.