The line asymmetry of electron spin resonance

Free Radical Biology & Medicine 40 (2006) 1524 – 1538 www.elsevier.com/locate/freeradbiomed

Original Contribution

The line asymmetry of electron spin resonance spectra as a tool to determine the cis:trans ratio for spin-trapping adducts of chiral pyrrolines N-oxides: The mechanism of formation of hydroxyl radical adducts of EMPO, DEPMPO, and DIPPMPO in the ischemic–reperfused rat liver Marcel Culcasi a , Antal Rockenbauer b , Anne Mercier a , Jean-Louis Clément a , Sylvia Pietri a,⁎ a

Laboratoire Structure et Réactivité des Espèces Paramagnétiques, Sondes Moléculaires en Biologie, CNRS-UMR 6517, Universités dTAix-Marseille I & III, 13397 Marseille cedex 20, France b Chemical Research Center, Institute of Chemistry, Hungarian Academy of Sciences, Budapest, Hungary Received 23 August 2005; revised 13 December 2005; accepted 20 December 2005 Available online 19 January 2006

Abstract Nonstereospecific addition of free radicals to chiral nitrones yields cis/trans diastereoisomeric nitroxides often displaying different electron spin resonance (ESR) characteristics. Glutathione peroxidase–glutathione (GPx-GSH) reaction was applied to reduce the superoxide adducts (nitrone/SOOH) to the corresponding hydroxyl radical (HOS) adducts (nitrone/SOH) of two nitrones increasingly used in biological spin trapping, namely 5-diethoxyphosphoryl-5-methyl-1-pyrroline N-oxide (DEPMPO) and 5-ethoxycarbonyl-5-methyl-1-pyrroline N-oxide, and of 5diisopropoxyphosphoryl-5-methyl-1-pyrroline N-oxide (DIPPMPO), a sterically hindered DEPMPO analogue. The method offered improved conditions to record highly resolved ESR spectra and by accurate simulation of line asymmetry we obtained clear evidence for the existence of previously unrecognized isomer pairs of cis- and trans-[DEPMPO/SOH] and [DIPPMPO/SOH]. Additional nitrone/SOH generation methods were used, i.e. photolysis of hydrogen peroxide and the Fenton reaction. We developed a kinetic model involving first- and second-order decay and a secondary conversion of trans to cis isomer to fully account for the strongly configuration-dependent behavior of nitrone/SOH. In the reductive system and, to a lower extent, in the Fenton or photolytic systems cis-nitrone/SOH was the more stable diastereoisomer. In various biologically relevant milieu, we found that the cis:trans-nitrone/SOH ratio determined right after the spin adduct formation significantly differed upon the GPxGSH vs (Fenton or photolytic) systems of formation. This new mechanistic ESR index consistently showed for all nitrones that nitrone/SOH signals detected in the postischemic effluents of ischemic isolated rat livers are the reduction products of primary nitrone/SOOH. Thus, ESR deconvolution of cis/trans diastereoisomers is of great interest in the study of HOS formation in biological systems. © 2006 Elsevier Inc. All rights reserved. Keywords: Spin trapping; Hydroxyl radical spin adducts; Diastereoisomers; Nitroxides; ESR; Reperfused liver; Kinetics; DEPMPO; DIPPMPO; EMPO; Free radical

Abbreviations: ESR, electron spin resonance; DMPO, 5,5-dimethyl-1-pyrroline N-oxide; EMPO, 5-ethoxycarbonyl-5-methyl-1-pyrroline N-oxide; DEPMPO, 5diethoxyphosphoryl-5-methyl-1-pyrroline N-oxide; DTPA, diethylenetriaminepentaacetic acid; HY, hypoxanthine; XO, xanthine oxidase; SOD, superoxide dismutase; CAT, catalase; GPx, glutathione peroxidase; GSH, glutathione; TEMPO, 2,2,6,6-tetramethyl-1-piperidine N-oxyl; TMPPO, 2-diethoxyphosphoryl-2,5,5trimethylpyrrolidine N-oxyl; DIPPMPO, 5-diisopropoxyphosphory-5-methyl-1-pyrroline N-oxide; DMSO, dimethyl sulfoxide. ⁎ Corresponding author. Fax: +33 0 4 91 28 87 58. E-mail address: [email protected] (S. Pietri). 0891-5849/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.freeradbiomed.2005.12.029

M. Culcasi et al. / Free Radical Biology & Medicine 40 (2006) 1524–1538

Introduction In an electron spin resonance (ESR) spin-trapping experiment where a pyrroline N-oxide with a β-hydrogen is used as the spin trap, a chiral nitroxide is obtained if any free radical except hydrogen is trapped. When the nitrone is itself chiral the two faces of the molecule become diastereotopic and diastereoisomery of the spin adduct can occur yielding cis/trans pairs of nitroxides (Fig. 1) whose ESR spectra may exhibit different coupling constants [1]. In an early separative and identificative approach it has been taken advantage that the β-coupling constants of individual diastereoisomers can be significantly different in a properly chosen solvent as to decompose their ESR spectra [1]. In the search for nitrones with improved properties in biological spin trapping, many chiral derivatives of 5,5-dimethyl-1-pyrroline N-oxide (DMPO), including 5-ethoxycarbonyl-5-methyl-1-pyrroline N-oxide (EMPO) and the phosphorylated 5-diethoxyphosphoryl-5-methyl-1-pyrroline N-oxide (DEPMPO), have been synthesized (Fig. 1). The trapping experiments in aqueous buffers may produce mixtures of diastereoisomers having quite distinct ESR spectra [2–19]. Main interest has been S focused on the characteristics of superoxide (O2 −) trapping by S these nitrones and on the stability of the nitrone/ OOH spin adducts, the protonated species that dominate at physiological S pH. ESR spectra of O2 − and peroxyl adducts of DEPMPO and its derivatives displayed well-resolved cis/trans features [2,3,5,8,9,15,19,20]. We and others [2,3,5,8,9,15,20] have interpreted the remarkable alternating line width effect in the S main signal of the trans-DEPMPO/ OOH isomer by using an exchange model between rotamers around the O-O bond, a physical process which greatly affects the hyperfine couplings of the β-spins, especially 31P. Some reports have claimed the S presence of diastereoisomers for the O2 − adduct in the EMPO family [7,10,11,13,14,17,18] but chemical exchange can also be responsible for the shape of the ESR spectra [4] so that a unique decomposition of the signal can be done only with a great uncertainty due to the strong overlap of the signal components in the absence of 31P coupling. Owing to the dihedral dependence of β-couplings in nitroxides, analysis of aHβ in five-membered ring diastereoisomeric nitroxides should also allow the distinction between cis and trans diastereoisomers, the larger aHβ value corresponding to a pseudo-axial orientation of the C-Hβ

1525

S

bond [21,22]. Thus in most of the reported hydroxyl (HO ), alkyl, and alkoxy radicals adducts in the EMPO family the ESR spectra consisted of clearly resolved mixtures of cis and trans diastereoisomers [4,7,10,11,14,17,18], a typical example S of which is shown in Fig. 2A for EMPO/ OH. Because of the high steric hindrance of the β-phosphorylated group in DEPMPO-type nitrones, an increased stereoselectivity of the trapping reaction with the size of the trapped radical can be predicted. Interestingly, the methoxy radical adducts of many DEPMPO-type nitrones gave ESR signals consisting of a mixture of cis/trans diastereoisomers [5,8], whereas, on the basis of the relatively simple features of the ESR spectra S S of the nitrone/ OH and / Me adducts in buffers, the trapping was often claimed to yield only the nitroxide having the trans configuration relative to the phosphorylated group [2,3,5,8,20]. In disagreement with this hypothesis of a high S stereoselectivity of the addition of HO to β-phosphorylated nitrones were the theoretical calculations of Villamena et al. [12,23] who concluded that in the hydroxyl spin adducts of DEPMPO and 5-diisopropoxyphosphoryl-5-methyl-1-pyrroline N-oxide (DIPPMPO; Fig. 1) the cis isomer is more thermodynamically stabilized than the trans isomer. Khramtsov et al. [24] indirectly demonstrated that methyl radicals were actually trapped on DEPMPO, yielding an equimolecular mixture of cis/trans diastereoisomers despite S simulation of the ESR signal of DEPMPO/ Me giving a satisfactory correlation assuming a single isomer if the minor asymmetry in the experimental spectra is dismissed [5,24]. This result suggests that a similar situation may occur for DEPMPO/ SOH although the method in the above study could not be applied in this case. S Since cis/trans diastereoisomers of nitrone/ OH adducts could be detected in the EMPO family and the methoxy and S O2 − adducts in the DEPMPO family, one can assume the S same situation for the very small and reactive HO when trapped by DEPMPO or DIPPMPO. In the present study, we used improved computer simulation of highly resolved spectra to provide reliable sets of ESR parameters for cis and trans S S diastereoisomers of DEPMPO/ OH and DIPPMPO/ OH, which calculations are based on the analysis of the ESR line asymmetry. The influence of various methods of nitrone/ SOH production on the cis:trans ratio and on the decay of S S DEPMPO/ OH and DIPPMPO/ OH in aqueous buffers was S compared to that of the EMPO/ OH adducts for which cis/

Fig. 1. General spin-trapping reaction on a DMPO-type nitrone having two diastereotopic faces and a chiral center at C-5 (R1 ≠ R2). Upon a nonstereospecific addition of RU (R ≠ H), both cis and trans diastereoisomeric nitroxides are obtained. In the chemical structures of the studied hydroxyl spin adducts (R = OH), R1 = Me and R2 = CO2Et (EMPO), P(O)(OEt)2 (DEPMPO), or P(O)(OiPr)2 (DIPPMPO). The asterisks label the chiral carbons.

1526


from bovine liver), glutathione peroxidase (GPx, from bovine erythrocytes), its cofactor glutathione (GSH), deferoxamine mesylate salt (Desferal), and 2,2,6,6-tetramethyl-1-piperidine N-oxyl (TEMPO) were from Sigma–Aldrich Chimie (SaintQuentin Fallavier, France). 2-Diethoxyphosphoryl-2,5,5-trimethylpyrrolidine N-oxyl (TMPPO) was prepared by the method of Le Moigne et al. [25] and was purified according to a published procedure [26]. Ultra-pure DEPMPO was obtained by a modification [27] of the method of Fréjaville et al. [2]. DIPPMPO [3] and EMPO [4] were synthesized and purified as previously described. Doubly distilled deionized water was used in all experiments and phosphate buffer solutions used in perfusion experiments were filtered through a 0.2-μm Millipore filter prior to use. Free radical generating systems, formation of spin adducts, and kinetics

Fig. 2. Low-field region of the high-resolution ESR spectra of hydroxyl radical adducts of EMPO, DEPMPO, and DIPPMPO in 0.02–0.15 M phosphate buffer (pH 7.0). EMPO/ OH (A) was recorded 2 min after cutoff of a 30-s UV photolysis of H2O2 (3%) in the presence of 0.1 M EMPO; the arrow indicates the low-field line of a minor triplet and the expanded region the resonances of the cis isomer. DEPMPO/ OH (B) and DIPPMPO/ OH (C) spectra were obtained 4–10 min after reduction of the respective superoxide adducts (from 0.15 M nitrone + XO (0.06 units/ml) + HY (0.4 mM)) by GPx (10 units/ml) + GSH (1.2 mM) followed by 3–5 min bubbling with N2 gas. Traces A'–C' represent the respective simulations of spectra A–C from data of Table 1. Spectrometer settings: microwave power, 10 mW; modulation amplitude in mT, 0.01 (A, C), 0.0056 (B); time constant in ms, 10.24 (A, B), 20.48 (C); gain, 4 × 104 (A), 2.5 × 105 (B), 2 × 105 (C); scan rate in mT.s-1, 0.27 (A), 0.54 (B), 0.24 (C).

S

S

S

trans diastereoisomery can be directly established by the recording of separated ESR lines. Finally, the usefulness of the above cis/trans ESR analysis as a mechanistic tool in free radical biology was demonstrated with EMPO, DEPMPO, and DIPPMPO in spin-trapping experiments on the isolated rat liver undergoing warm ischemia and reperfusion. Experimental procedures Chemicals, biochemicals, and solutions Solvents, starting materials, and reagents used in the syntheses of nitrones were all reagent grade. Diethylenetriaminepentaacetic acid (DTPA), dimethyl sulfoxide, (DMSO), FeSO4, KH2PO4, hydrogen peroxide (H2O2), hypoxanthine (HY), xanthine oxidase (XO, from buttermilk), superoxide dismutase (SOD, from bovine erythrocytes), catalase (CAT,

Procedures for the determination of ESR parameters of highly resolved superoxide and hydroxyl radical adducts of nitrones S Nitrone/ OOH adducts were obtained at room temperature in 20 mM KH2PO4 buffer (pH 7.0) by adding 0.04–0.06 units/ml XO to 0.4 mM HY in the presence of 0.10–0.15 M tested nitrone. The enzymatic reaction was stopped by adding a mixture of SOD (10 units/ml) and CAT (20 units/ ml) either 2–3 min (EMPO) or 6–7 min (DEPMPO or S DIPPMPO) after addition of XO. DEPMPO/ OH or S DIPPMPO/ OH spin adducts yielding ESR signals with undetectable paramagnetic contaminants were further S obtained by adding to the nitrone/ OOH adducts a mixture of GPx (10 units/ml) and GSH (1.2 mM). Samples (final volume 0.5 ml) were subsequently bubbled for 3–5 min with S nitrogen gas prior to ESR measurement. The EMPO/ OH spin adduct was obtained by photolyzing for 30 s a 3% solution of H2O2 in the presence of 0.1 M EMPO in 150 mM phosphate buffer (pH 7.0) with a 1000-W UV Xe–Hg Oriel lamp (Newport Corp., CA, USA). Other systems S The effect on nitrone/ OOH ESR signals of adding 1 mM S metal chelator (DTPA or Desferal) to the HY-XO O2 − generator described above was investigated in phosphate buffer. In S addition to reduction of preformed nitrone/ OOH adducts three S nitrone (0.1 M)/ OH generating systems were alternatively tested, i.e., (i) photolysis of 3% H2O2 in 150 mM KH2PO4 buffer alone or containing 1 mM DTPA and two Fenton systems in 20 mM KH2PO4 buffer, pH 7.4, containing 1 mM H2O2 and (ii) either preformed [DTPA (1 mM)-FeSO4 (0.1 or 1 mM)] chelate or (iii) FeSO4 (1 mM) only. All of the reductive and Fenton experiments described above were repeated in DTPA-free Krebs–Henseleit buffer, pH 7.35 (see preparation below). In two additional series of experiments aimed to mimic the physiological conditions in Krebs buffer, an aliquot of rat liver homogenate (40 μl/ml, see preparation below) was S added immediately after SOD + CAT to nitrone/ OOH S adducts (instead of GPx-GSH) or to nitrone/ OH adducts formed by a DTPA-free Fenton reaction.


Kinetics S Nitrone/ OH adducts were produced in 20 mM phosphate buffer or in Krebs–Henseleit buffer either by the abovedescribed GPx-GSH reductive system or by any of the Fenton systems described above. When the Fenton systems were used the free radical generation was stopped by treatment with CAT (20 units/ml) 1 min after addition of FeSO4 or its DTPA chelate. In preliminary control experiments the efficiency of the enzymes to stop free radical formation was checked. Thus addition of 1 mM FeSO4 to 0.1 M DEPMPO or EMPO solutions that incubated with 1 mM H2O2 and CAT (20 units/ ml) for 2 min resulted in no ESR signal within the first 30 min. Also no ESR signal was observed during 30 min following addition of XO to 0.15 M solutions of nitrones that incubated with HY + SOD (10 units/ml) ± CAT (20 units/ml) for 2 min. S The decay of nitrone/ OH adducts was monitored by timely recording of the complete ESR signal over 35–70 min. Decay curves were constructed from spectral simulation data and were fitted to adequate kinetic equations using a simulation program developed by one of us (A.R.; see Results and discussion). The concentrations determined by ESR simulation for 10 μM solutions of TEMPO (or TMPPO) in 20 mM phosphate buffer were compared to those of S simulated nitrone/ OH to estimate absolute radical concentrations. Data are given as means ± SD from n = 3–7 independent experiments. Statistical analysis was performed using the Prism 4 software (GraphPad Software, CA, USA). For EMPO and DEPMPO, a one-way analysis of variance was carried out to test for any differences among the mean S values of cis-nitrone/ OH percentages at 1–2 min following spin adduct formation in the GPx-GSH, photolytic, or Fenton methods. This was followed by the Duncan test. Significance was accepted at the p b 0.05 level. Isolated liver perfusion The study protocol was in compliance with the Guide for the Care and Use of Laboratory Animals published by the National Institute of Health, publication No. 85-23, revised (1985). Livers from Wistar male rats (200–250 g body weight) fed ad libitum with a standard diet were used. Animals were anesthetized with sodium pentobarbital (50 mg/kg intraperitoneally) before the opening of the abdomen wall. The liver was perfused following the procedure described by Desmoulin et al. [28]. The portal vein was cannulated and the liver was perfused by pumped Krebs– Henseleit bicarbonate buffer at 37°C bubbled with a gas mixture of 95% O2 and 5% CO2 in a nonrecirculating mode. Assuming that the liver weight represents about 4% of the body weight, the perfusion rate was initially set at 30 ml/ min. The composition of the perfusate (pH 7.35) was the following (in mM): NaCl (119); NaHCO3 (25); KCl (4.7); MgSO4 (1.2); KH2PO4 (1.2); CaCl2 (1.3). After excision, the liver was removed and transferred into a thermostated chamber. The perfusion was adjusted to a flow rate of 3

1527

ml/min/g of liver wet weight and the effluent was directly collected by a second peristaltic pump. The perfusion protocol was the following: 20-min perfusion with saline buffer followed by 10-min perfusion with either EMPO, DEPMPO, or DIPPMPO (5–15 mM) and a 10-min washout period with saline buffer only, 90 min of total warm ischemia at 37°C in saline buffer, and a 15-min reperfusion in the presence of the nitrone (5–15 mM). At each time indicated at the end of the control period and at reperfusion (0.5 to 15 min), drops of hepatic effluent were collected for 5 s in glass tubes and directly transferred into the ESR flat cell for immediate analysis. Throughout the perfusion protocol the liver bathed into the perfusate that filled the thermostated chamber, except during reperfusion (to avoid any spin adduct dilution). During the normoxic period, lactate dehydrogenase efflux was measured as a general index of cell injury as described in [29] to test for any toxic effect of the nitrones. Preparation of rat liver homogenate Rat liver homogenate was obtained according to previously described procedures [30,31]. After sacrifice, the liver was removed, weighed, and homogenized with a Potter tissue grinder in 0.05 M phosphate buffer (pH 7.35) in the proportion 1:3, wet weight/v. ESR spectroscopy Samples were quickly introduced into a 10-mm quartz flat cell which was fitted within the resonator cavity of a Bruker ESP 300 spectrometer (Karlsruhe, Germany) operating at Xband (9.79 GHz) with a modulation frequency of 100 kHz and a microwave power of 10 mW. The magnetic field strength and microwave frequency were measured with a Bruker ER 035M NMR gaussmeter and a Hewlett–Packard 5350B frequency counter, respectively. To consider the impact of field inhomoS geneity in the measurement of g-factors of nitrone/ OH adducts we added 2 μl of 0.04 mM aqueous TEMPO to the spin adducts (0.5-ml samples from a DTPA-free Fenton reagent added with excess CAT 1 min after addition of FeSO4) in 20 mM KH2PO4 just prior to recording the ESR spectrum. Control experiments showed that (i) mixing the CAT-supplemented Fenton reagent with aqueous TEMPO resulted in no change in the ESR features (intensity and coupling constants) of the nitroxide and (ii) the reference g value for TEMPO in pure water (g = 2.00558 [32]) was not shifted when the spectrum was recorded in phosphate buffer (data not shown). ESR spectra having 4k resolution in the magnetic field axis were acquired at room temperature. Other instrument settings in individual experiments and conditions for spectral acquisition are detailed under Results and discussion. Computer simulation of the ESR spectra was performed with the program of Rockenbauer and Korecz [33] and sets of parameters applied throughout in the determination of relative percentages of diastereoisomers were derived from the simulation of the bestresolved signals.

1528


Results and discussion High-resolution ESR spectra of superoxide and hydroxyl adducts of EMPO, DEPMPO, and DIPPMPO and impact of DTPA on the line width To calculate the best sets of ESR coupling constants for hydroxyl radical adducts of the three tested nitrones these species should be produced with a minimum of contaminating radicals and improved spectral resolution. Among the S methods used to generate EMPO/ OH (see Experimental procedures) UV photolysis of 3% H2O2 yielded the best signals. The line width was ΔH = 0.018 mT (a value of the order of magnitude of the intrinsic line width) and we obtained resolved hyperfine structure for the diastereoisomers not reported in previous studies [4,7,12,14] (Fig. 2A). A very minor triplet (about 1.5% of the total signal) with aN = 1.490 mT was recorded but it did not overlap with the main EMPO/ SOH signal (Fig. 2A). Inclusion of 1 to 8 mM DTPA in the photolytic system had practically no effect neither on the signal shape (ΔH = 0.019 mT at 8 mM DTPA) nor in the S intensity, whereas when EMPO/ OH was obtained by a Fe (II)-driven Fenton’s reaction in nondegassed buffer the presence of 1 mM DTPA resulted in significant increase of both ΔH (from 0.027 to 0.044 mT) and intensity (about three times). In the DTPA-containing Fenton experiments with EMPO we recorded unresolved lines similar to those S previously reported [4,7,12,14]. Since (i) the EMPO/ OH ESR lines did not broaden when DTPA concentration was raised to 8 mM and (ii) bubbling the DTPA-containing Fenton solution with N2 for 5 min restored the line width obtained in DTPA-free milieu (data not shown), we suggest that DTPAinduced signal broadening was due mainly to an increase of dissolved O2 concentration rather than to relaxation effects reported for DTPA-Fe complexes in the case of nitroxides [34]. The metal-ion-catalyzed Fenton reaction is generally considered an important mechanism of biological damage through the S formation of HO radicals and/or other oxidizing species [35]. S The most common source for the production of nitrone/ OH adducts in spin trapping is the Fenton system. Incubation of DEPMPO (0.1 M) with 1 mM H2O2 and 1 mM FeSO4 in 20 mM phosphate buffer yielded a signal consisting solely of a doublet S of quartets assigned to DEPMPO/ OH [2] but we obtained a mixture of spin adducts when 0.1 M DIPPMPO was the spin trap (see Fig. 6, inset). In this latter experiment the main species S was assigned to DIPPMPO/ OH and simulation of the minor S species was consistent with a mixture of two DIPPMPO/ alkyl adducts one of them having an unexpectedly low beta hydrogen splitting (in mT: aN = 1.587 and 1.607, aP = 5.720 and 4.883, and aHβ = 0.24 and 0.24 for the two species, respectively). Chalier and Tordo [3] have suggested the existence of hydrogen S abstraction by HO on the secondary carbon of the iPrO ligands in DIPPMPO as a likely event able to trigger the formation of S DIPPMPO/ alkyl adducts, a pathway not favored in DEPMPO which bears ethoxy groups. In experiments where DIPPMPO S S directly traps HO the DIPPMPO/ alkyl radicals always ap-

peared strongly overlapping the main signal (17–25% of the total signal). For this reason we prefered producing DIPPMPO/ SOH by the reductive GPx-GSH system (see Experimental procedures) since this method afforded very pure signals for hydroxyl radical adducts of DEPMPO and DIPPMPO. Figs. 2B and 2C show the well-resolved spectra obtained after bubbling with N2 the DTPA-free buffer solution. Typical ΔH values for signals of Figs. 2B and 2C were 0.026 and 0.021 S S mT for DEPMPO/ OH and DIPPMPO/ OH, respectively, and 1 mM DTPA did not affect the signal intensity and the line width. On the other hand, the addition of DTPA (1 or 8 mM) to the DEPMPO-supplemented Fenton system again increased the signal intensity by a factor of three and the line width by about 30%. Lowering the iron concentration to 0.1 mM led to a 15% narrowing of the lines and a threefold decrease of the initial concentration. The initial narrow lines could be restored by N2 bubbling. DTPA has been widely used in spin-trapping studies with DMPO due to its ability to inhibit the formation of radical artifacts in the presence of metal ions, including Fe (II) [36]. In spin-trapping experiments using EMPO, DEPMPO, or DIPPMPO we showed that a good signal resolution requires degassing of the solution when DTPA is present since the O2-producing processes are accelerated by the catalytic activity of this chelator in the Haber–Weiss reaction [37]. For the above three nitrones the formation of secondary radicals is a slow process in the presence of traces of metal ions [2–4]. Evidence for the existence of two components in the ESR S S spectra of DEPMPO/ OH and DIPPMPO/ OH The derivative absorption lines of the ESR spectrum of a single radical species recorded when the g-factor is isotropic should be all symmetric with respect to the center of the line. Among the instrumental distortions modifying the ESR line width, the modulation amplitude or relaxation effects (Fig. 3A) do not affect the symmetry of the line shape while various distortions such as phase dispersion (90° phase defines a pure dispersion signal; Fig. 3B) or nonoptimized filtering (Fig. 3C) result in identically asymmetric lines for a single radical [33]. Only unresolved superposition of two or more radical species can yield ESR lines having different asymmetries. Computer simulations of a mixture of two radicals having the same g-factor but different I = 1/2 couplings (Figs. 3D–3F) show that lines exhibit different asymmetry (α ≠ 0) only when the species have different populations or line widths. S Of the eight major lines of the ESR spectra of DEPMPO/ OH S and DIPPMPO/ OH in aqueous solution the second, third, sixth, and seventh lines are composite due to the nearly identical 14N and 1H couplings [2,3]. These lines could be asymmetric even without superimposition of spectra if two components have different widths due to the relaxation (Fig. 3A). Being noncomposite signals, the alternating asymmetry of the lines in the wings and the center of the spectrum, if any, could be caused only by superimposition of two radical species at significantly different concentrations (Fig. 3E). All DEPMPO/


1529

relaxation parameters α–γ’ entering the line width formula as Eq. (1), DHMM V ¼ a þ hM þ gM2 þ h VM Vþ g VMM V;

ð1Þ

where M and M’ represent the magnetic quantum number of N and 31P nuclei, respectively. For the two phosphorylated hydroxyl radical adducts the line shape was satisfactorily reconstructed using 100% Lorentzian curve, one methyl, and two equivalent and two nonequivalent γ-protons. Table 1 reports the calculated ESR parameters of the diastereoisomers S S of DEPMPO/ OH and DIPPMPO/ OH giving the simulated spectra shown in Figs. 2B′ and 2C′, respectively. Due to the high resolution of the spectra the automatic parameter fitting provided hyperfine coupling constants with a precision of at S least 0.01 mT. For DEPMPO/ OH the calculations gave accidentally identical values for the aP splitting in the cis and trans diastereoisomers. S In the case of EMPO/ OH for which the two diastereoisomers have clearly separated signals (Fig. 2A) a similar pattern for the long-range γ-hydrogen splitting was obtained (Fig. 2A′). To demonstrate the major properties of cis and trans isomers of S the three nitrone/ OOH adducts, their spectral data are also included in Table 1 in which simulations were performed assuming the existence of two exchanging rotamers [2,3] for the trans diastereoisomers. On the basis of a deconvolution analysis involving subtraction of consecutive ESR spectra, Dikalov et al. [19] have postulated a five-membered ring nitroxide bearing a β-epoxide ring and a β-hydroxyl group as a minor byproduct of S the decomposition of DEPMPO/ OOH. In our long practice of S O2 − trapping with DEPMPO we have repeatedly found that S when rather persistent DEPMPO/ alkyl adducts are present as contaminants (e.g., when the nitrone was not ultra-pure) they were not affected by further reaction with GPx-GSH (data not shown). Due to the lack of any contaminating signal in the ESR S spectra of DEPMPO/ OH in the reductive experiments (Figs. 2B and 3G) the subsequent kinetic analysis did not take into S account the impact of the DEPMPO/ OOH decomposition postulated in [19].

14

Fig. 3. Computer simulation of the effects of various conditions on the line symmetry of an ESR doublet (A–F) and evidence for line asymmetry in the DEPMPO/ OH (G) and DIPPMPO/ OH (H) spectra. Asymmetry of the ESR signal is visualized by the angle α ≠ 0 between the lines connecting the bottoms of the two ESR derivative peaks and the mirror image (dotted line) of the straight line connecting the tops. Simulations A–C represent a single radical with a coupling of 1.25 mT and varying line widths of 0.06 and 0.07 mT due to different relaxation of the lines (A), or uniform line width of 0.06 mT, but with a 15° of phase dispersion (B) or with excess filtering distortion (C). Simulations D and E represent the superimposition of doublets of two radicals with couplings of 1.2 and 1.35 mT and with uniform line width of 0.06 mT, while the populations ratios are 1:1 or 2:1, respectively. Simulation F is the same as D with extra unresolved couplings of 0.06 mT for one of the radicals. Signals G and H were obtained in nondegassed buffer by the reduction method described in the legend of Fig. 2 using the same parameter settings except modulation amplitude (mT), 0.02 (G), 0.04 (H); time constant, 10.24 ms; gain, 3.2 × 105 (G), 1.25 × 105 (H); scan rate, 0.54 mT.s−1.

S

SOH

S

S

and DIPPMPO/ OH signals recorded using any of the generators described under Experimental procedures showed marked alternating asymmetry of their two central lines, typical examples of which are shown in Figs. 3G and 3H. Even though marked asymmetries of the central lines could be seen in the S spectra of DEPMPO/ OH [2,12,16,19,24,33,38,39] and S DIPPMPO/ OH [3] the presence of two species was pointed out only in two studies [12,19]. A plausible explanation of asymmetry is the cis/trans diastereoisomery because DEPMPO and DIPPMPO have one chiral C-5 atom and the C = N double bond of the nitronyl function defining two diastereotopic faces for free radical addition. S We performed simulation of high-resolution DEPMPO/ OH S and DIPPMPO/ OH ESR spectra shown in Figs. 2B and 2C using automatic parameter adjustment procedures assuming that line asymmetry is due to the presence of two radical adducts. The spectra of each species can be characterized by at least 10 parameters, i.e., the amplitude factors, the g-values, three main hyperfine splitting constants (aN, aP, and aHβ), and five

Assignment of the diastereoisomers

S

Table 1 shows that for all three nitrone/ OH adducts the minor species has the larger aHβ coupling. If the main factor determining the energy order of the two diastereoisomers is steric repulsion then the cis configuration should be the minor species. The stronger steric effect can favor a more equatorial orientation for the C-OH bond which in turn should make the C2-H bond less equatorial, giving a larger aHβ coupling. Recently Villamena et al. [23] have proposed a different picture for S S S EMPO/ OH compared to DEPMPO/ OH and DIPPMPO/ OH. In the former case the more stable isomer was assigned to the trans configuration while in the latter cases it was assigned to the cis configuration. The inversion of the order was explained in their theoretical calculations by the increased stability of the cis isomer by an intramolecular H bond with the P = O group in the DEPMPO and DIPPMPO hydroxyl radical adducts.

1530


Table 1 Simulated ESR parameters of diastereoisomers of hydroxyl and superoxide spin adducts a of nitrones EMPO, DEPMPO, and DIPPMPO b Spin adduct

Generating system

Diastereoisomer

aN/mT

aHβ/mT

aHγ/mT

EMPO/HO

Photolysis of 3% H2O2 c

cis 24% g = 2.00564

1.418

1.527

trans 76% g = 2.00564

1.411

1.280

0.062 (1 H) 0.050 (1 H) 0.029 (Me) 0.007 (2 H) 0.063 (1 H) 0.043 (1 H) 0.021 (Me) 0.013 (2 H) 0.048 (1 H) 0.034 (Me) 0.013 (1 H) 0.010 (2 H) 0.050 (1 H) 0.034 (Me) 0.017 (2 H) 0.008 (1 H) 0.068 (1 H) 0.040 (Me) 0.029 (1 H) 0.013 (2 H) 0.054 (1 H) 0.041 (Me) 0.025 (1 H) 0.013 (2 H)

S

S

S

Reduction of HOO adduct d

DEPMPO/HO

S

S

Reduction of HOO adduct d

DIPPMPO/HO

S

4.729

1.405

1.421

trans 62% g = 2.00568

4.729

1.405

1.273

cis 28% g = 2.00569

4.597

1.418

1.409

trans 72% g = 2.00569

4.704

1.407

1.319

T1 e (37.5%) T2 e (62.5%)

5.408 4.783

1.325 1.327 1.325 1.305 1.314

0.930 1.237 1.015 1.337 0.975

T1 e (36.6%) T2 e (63.4%)

4.044 5.315 4.804

1.339 1.307 1.314

0.990 1.393 0.957

3.994

1.347

1.003

trans 94% g g = 2.00574

S

HY-XO system f

cis 6% g g = 2.00574 trans 86% g = 2.00578

S

HY-XO system f

cis 14% g = 2.00578 trans 85% g = 2.00579

DEPMPO/HOO

DIPPMPO/HOO

aP/mT

cis 38% g = 2.00568

HY-XO system f

EMPO/HOO

Conformer

cis 15% g = 2.00579

T1 e (44.5%) T2 e (55.5%)

0.150 (1 H) 0.090 (1 H) 0.060 (1 H) 0.048 (1 H) 0.040 (1 H) 0.045 (1 H) 0.036 (Me) 0.139 (1 H) 0.089 (1 H) 0.060 (1 H) 0.047 (1 H) 0.045 (1 H) 0.036 (Me) 0.140 (1 H)

a

In DTPA-free phosphate buffer (0.02 M except for photolysis, 0.15 M; pH 7.0) at room temperature. Concentration of nitrones, 0.10–0.15 M. c Spectrum recorded 1–2 min after end of photolysis. d Obtained by incubating the nitrone with a mixture of hypoxanthine (HY; 0.4 mM) and xanthine oxidase (XO; 0.04–0.08 units/ml) for 6 min, then adding a mixture of SOD (10 units/ml), CAT (20 units/ml), GPx (10 units/ml), and GSH (3 mM), and finally bubbling nitrogen gas for 5 min. ESR spectra were recorded 25–30 min after addition of GPx. e T1 and T2 refer to slowly exchanging rotamers; the computed average exchange times are (in ns): 152, 10.5, and 11.0 for the superoxide adducts of EMPO, DEPMPO, and DIPPMPO, respectively. f Same HY-XO generator as described under footnote d except that incubation was only 3 min with EMPO. All g values were determined relative to internal TEMPO (see Experimental Procedures). g Tentative assignement. b

According to the calculated geometries the dihedral angle for the C-2-H bond is reduced for the ground cis isomer, making its aHβ coupling smaller. Consequently, based on the relative value of aHβ coupling, we cannot conclude whether the steric or hydrogen bond interactions play the dominant role to determine the ground configuration. We have, however, additional experimental evidence to enable assigning the major configuration. First, regardless of S the method used to produce the three nitrone/ OH adducts, we

always found that the minor species has the larger aHβ coupling. Second, we found that the reduction of the nitrone/ SOOH into the nitrone/SOH adducts was quantitative, within the experimental precision, suggesting that the major superoxide adduct should yield the major hydroxyl radical adduct. S Since for DEPMPO and DIPPMPO the trans-nitrone/ OOH was considered the main isomer [2–4] the same holds for the S nitrone/ OH adduct. This statement is also supported by our kinetic measurements on the three nitrones that show that the


1531

S

minor isomer is always the more stable (see below). This S indicates that if in the case of EMPO the cis-EMPO/ OH is the minor species then this should also be expected for the phosphorylated adducts (Table 1).

showing that no HO was produced in the course of the reduction. The absolute increase in the cis concentration consecutive to the reduction therefore suggests either a larger first-order stQability for the cis adduct,

Kinetic behavior of the diastereoisomers

d½cis-nitrone= OH=dt ¼ k1;cis d½cis-nitrone= OH

:

S

The kinetic behavior of nitrone/ OH for the three nitrones is rather sensitive to the experimental conditions, in particular the nature of the generating system. Contrary to the well-studied S decay of the O2 − adducts in the DMPO, EMPO, and DEPMPO families [2–6,8,13,14,16–18,20,40–43] scarce data are availS able for nitrone/ OH and the characteristics and the nature decay processes have not been analyzed in detail [3,12,44,45]. Although we found that the decrease of the adduct concentration can be satisfactorily described by the mixed unimolecular and bimolecular processes, d½nitrone=:OH=dt ¼ k d ½nitrone=:OH k d ½nitrone=:OH ; 1

2

2

ð2Þ

the k1 first- and k2 second-order kinetic constants are widely dependent on the experimental conditions, in particular the presence of paramagnetic species such as iron salts. Even the sign of the constants was found negative in certain experiments, S suggesting a secondary source for nitrone/ OH adduct when no S more primary OH radicals are formed in the system. We also found that the duration of primary free radical generation had a great impact on the subsequent decay kinetics. In most S experiments O2 − formation was stopped about 6 min after S addition of XO where a maximum nitrone/ OOH signal intensity was achieved but occasionally shorter (2-min) or longer (12-min) generation times were used to test the impact of diamagnetic products. The supposed secondary radical generation was characterized by an exponentially decreasing initiation rate,

:

:

d½nitrone= OH=dt ¼ I0 eksec S t ;

ð3Þ

where I0 gives the initiation rate at the beginning of the decay, ksec characterizes the exhaustion rate of the secondary radical source, and I0/ksec gives the quantity of integrated secondary radicals. A common feature of all kinetic experiments using either the Fenton reaction or the GPx-GSH system was that the relative concentration of the minor cis adduct always increased and in certain reductive experiments even the absolute cis adduct concentration had a temporary elevation, while the trans component was steadily decreasing. We performed control experiments with DEPMPO to check whether the above trend S could be accounted for by the unstopped formation of HO radicals during the reduction of DEPMPO/OOH by GPx-GSH. As expected [5] inclusion of excess DMSO to the DTPA-free Fenton’s reagent in the presence of the nitrone yielded only the S DEPMPO/ Me adduct (asymmetric ESR spectrum simulated as two diastereoisomers with coupling constants in mT: aN = 1.350 (1.470); aP = 4.970 (4.860); aHβ = 2.107 (2.115)), while in the experiment, where DMSO was added together with GPx-GSH S S to DEPMPO/ OOH, a pure DEPMPO/ OH signal was detected,

:

:

ð4AÞ

:

d½trans-nitrone= OH=dt ¼ k1;trans d½trans-nitrone= OH;

ð4BÞ

and/or a preference for the formation of the cis isomer during the secondary radical formation,

:

d½cis-nitrone= OH=dt ¼ I0cis S eksec S t

ð5AÞ

:

d½trans-nitrone= OH=dt ¼ I0trans S eksec S t8

ð5BÞ

where I0cis could be as large or even larger than I0trans. The enhanced rate of formation of cis compared to trans isomers in the secondary process with respect to the primary process could be related to the well-known disproportionation of nitroxides having a β-hydrogen, yielding a nitrone and a Nhydroxylamine [46]. Assuming that a reduction process can convert the nitrone back to the nitroxide, a trans attack relative to the bulky substituent is more likely and could favor the cis isomer. The complete kinetic model given by Eqs. (2–5) satisfactorily described all the observed kinetics, but due to the large number of parameters a unique determination of all parameters was not possible. We conducted the kinetical analysis on two S independent curves, namely the total and the cis-nitrone/ OH, which allowed us to determine a maximum of four parameters with reasonable confidence. To obtain a rough description for S the effective stability of nitrone/ OH adducts we also calculated the apparent first-order constant (k1app) when both the second order processes and the secondary radical formation were neglected. To derive the kinetic curves from the set of

S

Table 2 Effect of secondary radical formation on the kinetics of DEPMPO/ OH decay in the reductive system Added Incubation t1/2app chelator time with (min) (1 mM) HY-XO (min)

k1app k1 I0/ksec (103 min−1) (103 min−1) (μM)

DTPA 2 DTPA 6 DTPA 12 Desferal 6 None 2 None 6

46.5 33.7 29.1 55.7 95 91

14.9 20.6 23.8 12.4 7.3 7.6

S[DEPMPO/ OH] 0

(μM) 75 49 47 108 130 130

3.5 6.8 34.0 7.1 2.4 6.9

6.1 15.9 35.5 10.6 5.4 15.1

From 0.15 M nitrone; other reagents concentrations: see legend of Table 1. Kinetic parameters: t1/2app, apparent half-life; k1app, apparent first-order rate constant; k1, first-order rate constant; I0/ksec, integrated concentration of secondary radicals. [DEPMPO/ OH]0 represents the concentration of DEPMPO/UOH immediately following addition of GPx-GSH.

S

1532


differential equations we carried out numerical integration keeping only the first term in the Taylor expansion,

:

Fi ðt þ DtÞ ¼ Fi ðtÞ þ dFi =dtðtÞ Dt;

ð6Þ

where Fi stands for the cis, trans, or total concentrations. To minimize the computational error 1024 kinetic points were used, rendering Δt small and making the higher-order terms negligible. We tested the accuracy of this approximation in the case of Eq. (2) by comparing the results of numerical integration with the explicit formula and found a maximum deviation of less than 0.01%. In the automatic parameter fitting the least square deviations between the experimental and the calculated points were minimized. Since a single individual kinetic curve does not allow one to determine more than two parameters, the

S

S

curves for the total and cis adduct concentrations were adjusted simultaneously by using a weight factor equalizing the contribution of the two kinetic curves to the error function. Two alternative models were used in the computations: the enhanced cis concentration was described by larger k1,cis compared to k1,trans or by a I0cis/I0trans ratio significantly exceeding the cis:trans ratio in the primary process. The first model can be used when both the total and the cis concentrations reveal a monotonous decay (actually this was the case for the Fenton reactions), while the second model could also account for the temporary increase of the cis concentration accompanied by monotonous decay of the total adduct concentration. The combination of the two models is also possible, but in this case the number of independent parameters is too large to obtain a unique determination.

Fig. 4. Observed and simulated decay curves of DEPMPO/ OH and EMPO/ OH (A) and their cis diastereoisomeric components (B) when generated by incubating 0.1 M nitrone with FeSO4 (1 mM) and H2O2 (1 mM) in 20 mM phosphate buffer alone (□ or ○, respectively) or containing 1 mM DTPA ( or ●, respectively), pH 7.0. CAT (20 units/ml) was added 1 min after addition of FeSO4 to stop the radical generation. The calculation was based on the kinetic model described in Eqs. (2–5). The ESR settings in DEPMPO (EMPO) experiments were as follows: microwave power, 10 mW; time constant, 10.24 ms; modulation amplitude, 0.020 (0.016) mT; gain, 4 × 104 (4–10 × 104); scan rate, 0.51 (0.29) mT.s−1.

▪


In the GPx-GSH system where a temporary increase of absolute adduct concentration was found in many cases, the second model assuming an enhanced I0cis/I0trans ratio for the secondary process gave an excellent fit for both kinetic curves. This ratio was found close to one, that is the stereoselectivity typical for the primary hydroxyl radical adduct formation is no more valid for the secondary radicals. This is in accordance with the observation that at the end of the decay experiment the cis:trans ratio could be rather close to unity but is never larger than one. We carried out curve fitting when I0cis/I0trans = 1 was taken and only k1, I0 and ksec were adjusted and still we obtained excellent fit for experiments carried out in strongly different conditions. The results are collected in Table 2 which allows assessing the effect of the secondary radical formation on the apparent

S

S

1533

S

stability of DEPMPO/ OH as formed by the reductive system. Increasing the time of HY-XO reaction led to an increase of S S the starting DEPMPO/ OH concentration [DEPMPO/ OH]0 and an even faster rise of the secondary radical formation, explaining the enhanced apparent stability described by k1app. In the absence of DTPA both the apparent and the real k1 S indicate a significantly reduced stability of DEPMPO/ OH, while the secondary radical formation is not affected. The S larger persistency of DEPMPO/ OH in the presence of DTPA could be related to the inhibition of metal-dependent reduction of nitroxides by GSH. In support, Desferal, a potent metal chelator, also inhibited this reaction, although less efficiently (Table 2). The close orders of magnitudes of S I0/ksec and [DEPMPO/ OH]0 support the suggestion that the source of secondary radicals are the diamagnetic products

S

Fig. 5. Observed and simulated decay curves of DEPMPO/ OH and EMPO/ OH (A) and their cis diastereoisomeric components (B) in 20 mM phosphate buffer alone (□ or ○, respectively) or containing 1 mM DTPA ( or ●, respectively), pH 7.0. Nitrone/ OH adducts were obtained by adding GPx-GSH immediately after addition of SOD + CAT to a HY-XO O2 − generator that incubated with either DEPMPO or EMPO (both at 0.15 M) for 5.5 or 6.5 min, respectively. Reagents concentrations were as indicated in the legend of Fig. 2 and instrument settings in DEPMPO (EMPO) experiments were as follows: microwave power, 10 mW; modulation amplitude, 0.06 (0.025) mT; time constant, 20.48 (40.96) ms; gain, 1.25 × 105 (1.6 × 105); scan rate, 0.26 (0.29) mT.s−1.

S

▪

1534


S

S

formed in the course of O2 − generation. In the literature S the decay of DEPMPO/ OH has been analyzed by firstorder [45], pure second-order [3], or mixed-order [12] models affording quite different values for the apparent half-life (t1/2app). The t1/2app values in Table 2 for a long radical generation in the presence of DTPA are close to the data of [45] and shorter than those of the other reports [3,12]. We assume that these differences can reflect important deviations in the rate of secondary radical formation: the larger is this rate the longer is the apparent stability. The short S DEPMPO/ OH lifetime observed in the present study S indicates that the primary O2 − formation was efficiently stopped (see Experimental procedures). S Comparing the major characteristics of nitrone/ OH decay behavior in the reductive and Fenton systems, the main difference was observed in the kinetics of the cis diastereoisomer: while we observed only a monotonous decay in the Fenton system the reduction experiments often revealed a maximum, typical examples of which are shown in Figs. 4 and 5 for EMPO and DEPMPO. In the Fenton experiments the decay of the cis adduct was always found slower, but the kinetics does not allow distinguishing whether the larger cis stability or the enhanced cis proportion in the secondary radical formation is responsible for the increasing cis:trans ratio. The first approach yielded a mean k1,trans/k1,cis value of 1.39 without any characteristic difference among the three tested nitrones. Recent theoretical studies [23] have proposed S that the HO radical adduct in the pyrroline N-oxide series having the larger stability should display a smaller angle between the C-OH bond and the π orbital of the unpaired electron, that is the HCNπ dihedral angle is large and so the aHβ coupling is small. Our data on the cis:trans ratio of S nitrone/ OH adducts support this assumption only for the primary formation of the nitroxides since the species having the smaller aHβ coupling always have the larger population (Table 1) but either decomposes with the larger rate or is less favored in the secondary back reaction.

To assess the stability of nitrone/ OH adducts in different biologically relevant milieu they have been produced either by the reductive system or by the Fenton reaction. Table 3 shows that DTPA improved the stability of all adducts when formed by the reductive system but in the Fenton system its impact was found opposite. Interestingly this latter method provided three S times larger initial concentrations of nitrone/ OH and significantly broader line width. In the Fenton system, mean ΔH values S for DEPMPO/ OH were 0.043 mT in DTPA-free phosphate buffer and 0.067 mT when 1 mM DTPA was included. In this latter case, decreasing the Fe(II) concentration by an order of magnitude reduced the line width to 0.051 mT. The diminished S stability of nitrone/ OH could be explained by oxidation of the nitroxide by the [Fe(III)-DTPA] complex which is formed in the Fenton reaction. This assumption is supported by the threefold increased lifetime when the initial Fe(II) concentration was decreased by an order of magnitude (Table 3). Therefore we can recommend the use of DTPA in the Fenton system if a fast spectra recording is possible and no high resolution is wanted. In the reductive experiment adding DTPA is useful if an extended observation time is required. Such a situation may occur in biological studies because cellular systems possess many components with a high capacity in reducing nitroxides [26,30,31,47]. In this kinetic respect, Table 3 suggests that DIPPMPO may be the best nitrone detector when hydroxyl radical adducts are expected to be formed. Contrary to the reductive system in the photolytic experiments only a minor increase of the cis proportion was observed and the initial cis:trans ratio was enhanced by 10% similar to the Fenton systems (see Table 4). The different characteristics could S be accounted for by the different origins of nitrone/ OH adducts: while it results from a chemical reaction in the reduction S experiments (note that the amount of contaminant HO adduct in the HY-XO system never exceeded 10%), in the Fenton and S photolytic experiments HO are actually trapped. Owing to the larger steric hindrance of the OOH function, stronger stereoselectivity is expected in the reduction experiments.

Table 3 Apparent first-order decay rate constants k1app (103 min−1) and initial concentrations C0 (μM) of hydroxyl radical adducts in biologically relevant milieu Phosphate

S

GPx-GSH reductive system EMPO/ OH DEPMPO/ OH DIPPMPO/ OH

S S

Fenton system c EMPO/ OH DEPMPO/ OH DIPPMPO/ OH

S

S S

Phosphate + 1 mM DTPA

Krebs–Henseleit

k1app

C0

k1app

k1app

C0

62.3 ± 15.0 44.9 ± 9.8 21.7 ± 6.6

20 ± 3 37 ± 6 38 ± 7

26.0 ± 5.7 26.9 ± 4.9 17.4 ± 0.6

18 ± 3 35 ± 3 b 42 ± 7

65.9 ± 9.7 35.6 ± 3.1 20.7 ± 1.2

12 ± 1 46 ± 3 37 ± 8

25.9 ± 5.5 33.4 ± 1.7 23.5 ± 0.7

42 ± 8 87 ± 8 47 ± 1

72.6 ± 2.3 83.1 ± 2.5 45.6 ± 4.3

174 ± 25 317 ± 31 d 292 ± 40

28.1 ± 0.5 29.8 ± 3.1 33.2 ± 2.9

61 ± 11 88 ± 5 62 ± 6

C0

a

Data are means ± SD (n = 3–7) except otherwise stated; pH was 7.0 (phosphate ± metal chelator) or 7.35 (Krebs–Henseleit). Reagents, abbreviations, and concentrations are indicated in the legends of Figs. 2 and 4 and k1app was calculated from the decay curve of the total adduct concentration as described under Results and discussion. a Added 6 min after addition of HY and XO, followed by treatment with SOD and CAT, to the 0.15 M nitrone solution. b Desferal (1 mM) instead of DTPA: k1app = 55.7 and C0 = 10.6 (n = 1). c CAT was added 1 min after addition of Fe(II) to the H2O2-0.1 M nitrone solution. d With 0.1 mM FeSO4: k1app = 30.6 and C0 = 103.4 (n = 1).


1535

S

Table 4 Percentage of the cis diastereoisomer as a function of the nitrone/ OH producing system, the milieu, and time 3% H2O2 photolysis in 0.15 M phosphate buffer (pH 7.0) Nitrone

No DTPA min after end of irradiation

1 mM DTPA min after end of irradiation

1–2

20–25

35–40

1–2

20–25

35–40

EMPO DEPMPO

25.8 ± 0.5 33.7 ± 0.1

30.7 ± 0.1 36.6 ± 0.7

30.4 ± 0.1 35.9 ± 0.1

29.7 ± 3.5 35.4 ± 0.9

33.8 ± 2.3 37.7 ± 1.7

33.6 ± 2.1 36.0 ± 2.1

S

Reduction of nitrone/ OOH by GPx + GSH in 0.02 M phosphate buffer (pH 7.0) Nitrone

EMPO DEPMPO DIPPMPO

No DTPA min after GPx

1 mM DTPA min after GPx

1–2

12–15

25–30

1–2

12–15

25–30

16.2 ± 0.9* 24.7 ± 1.1§ 14.3 ± 0.5

31.4 ± 0.4 34.2 ± 0.4 18.6 ± 0.1

38.4 ± 0.8 35.4 ± 0.4 29.2 ± 0.5

12.6 ± 0.3* 22.0 ± 1.1§ 14.5 ± 0.4

25.0 ± 0.3 33.9 ± 1.4 18.9 ± 0.2

32.8 ± 1.4 35.7 ± 0.6 27.8 ± 0.9

S

Biologically relevant reduction of nitrone/ OOH in Krebs–Henseleit buffer (pH 7.35) Nitrone

EMPO DEPMPO DIPPMPO

Rat liver homogenate a min after aliquot addition

GPx + GSH min after GPx 1–2

15–20

30–35

1–2

15–20

30–35

14.3 ± 0.5* 19.8 ± 0.4§ 10.3 ± 1.5

35.9 ± 0.9 27.2 ± 0.7 20.7 ± 1.7

38.7 ± 0.6 38.6 ± 1.0 26.6 ± 0.9

18.6 ± 0.5 24.8 ± 0.8 14.7 ± 0.5

36.6 ± 0.9 39.9 ± 1.5 22.3 ± 0.6

41.7 ± 1.8 40.5 ± 1.3 28.6 ± 0.8

Fenton system in 0.02 M phosphate buffer (pH 7.0) Nitrone

No DTPA min after CAT

1 mM DTPA min after CAT

1–2

15–20

30–35

1–2

15–20

30–35

EMPO DEPMPO

29.3 ± 0.4 35.0 ± 0.8

38.0 ± 0.9 38.1 ± 0.6

39.4 ± 0.3 40.0 ± 0.4

24.8 ± 0.2 34.9 ± 0.6

32.6 ± 0.2 37.5 ± 1.3

35.6 ± 0.4 39.6 ± 0.9

Fenton system in Krebs–Henseleit buffer (pH 7.35)b Nitrone

EMPO DEPMPO

+Rat liver homogenate a

No addition min after CAT 1–2

15–20

30–35

1–2

15–20

30–35

27.8 32.7

37.3 37.5

39.9 39.4

29.6 31.3

37.7 36.5

38.5 37.2

S

Nitrone concentration was 0.15 M (reductive systems) or 0.1 M (Fenton) and percentages of cis-nitrone/ OH are means ± SD (n = 3–7), except under bn = 1. Generating systems, abbreviations, reagents concentrations, and conditions of ESR acquisition are as indicated in the legend of Table 1 and Experimental procedures. a Rat liver homogenate was prepared as described under Experimental procedures and added as an aliquot of 80 μl/ml solution immediately after addition of CAT. Statistics: one-way analysis of variance (p b 0.5) followed by Duncan test: *p b 0.01 and §p b 0.02 vs the corresponding (i.e., ± DTPA) value in the photolytic or the Fenton (in 0.02 M phosphate buffer) systems, with the * and § symbols refering to EMPO and DEPMPO, respectively.

Influence of the radical generator on the percentage of cis-hydroxyl radical adduct The final rationale of our study was that mechanistic insights into free radical formation in biological systems could be gained from the ESR analysis of diastereoisomeric spin adducts if the cis:trans ratio may range differently according to the nature of the free radical source. To this purpose we have determined in several buffers and conditions the percentages of the cis-nitrone/ SOH adduct using the photolytic, reductive, and Fenton systems. Due to the presence of high levels of radical contaminants in the S DIPPMPO/ OH signals obtained from the Fenton reaction the cis percentages could not be determined reliably and were therefore not considered. When the HY-XO system is used as S O2 − generator in phosphate or Krebs–Henseleit buffer and radical formation is blocked by SOD + CAT, about 10–15% of S the cis-nitrone/ OOH isomer is obtained for DEPMPO and

DIPPMPO (Table 1), in agreement with previous reports [2,3]. Regardless of the presence of DTPA, further reduction with S GPx-GSH resulted in an initial 10–25% of cis-nitrone/ OH, which increased with time as a result of the higher stability of the cis adduct (Table 3). This result strongly suggests that the hydroperoxyl function in the two diastereoisomers is equally accessible for reduction and/or that the rate of the enzymatic S reaction is much faster than the decay of the nitrone/ OH adduct. S Reduction of EMPO/ OOH also yielded a well-identified S mixture of cis/trans-EMPO/ OH, confirming our hypothesis (Table 1) of the existence of cis/trans diastereoisomery in the S parent O2 − adduct. S For a given nitrone, when the source of nitrone/ OH was H2O2 photolysis or the Fenton reaction the initial percentages of S cis-nitrone/ OH were similar and significantly higher (p b 0.01– 0.02, see legend of Table 4) than that measured with the reductive system. In the Fenton reaction it has been shown that

1536


S

HO attack on a target will preferentially occur at the binding sites of the metal ion catalyst, if any, and that strong metal chelators such as DTPA can reverse this site specificity [48]. Some DMPO-type nitrones bind to metal ions, including Fe(II), by their nitronyl function [49] but such an interaction, occurring at the site of radical trapping, cannot induce stereospecificity. S The cis-nitrone/ OH percentages in the Fenton system being not affected by DTPA for DEPMPO and DIPPMPO, we therefore conclude that no significant binding of Fe(II) to their phosphorylated groups had occurred. Rat liver homogenate, which contains many reductants, S including GSH and ascorbate, converted nitrone/ OOH to S nitrone/ OH and in Krebs–Henseleit the percentages and the profiles of increase of the cis component were found similar to those obtained in the GSH-GPx system. Control experiments also showed that rat liver homogenate did not interfer in the Fenton system and we consequently speculated that the two S mechanisms of nitrone/ OH formation should be discriminated in a biological environment on the basis of the percentage of the cis diastereoisomer. Studies on the ischemic-reperfused rat liver In the final part of the present study, we perfused each of the three tested nitrones to measure free radicals generated in isolated rat livers subjected to warm ischemia and reperfusion and our purpose was to use the diastereoisomeric analysis developed above to propose a plausible mechanism accounting for the formation of nitroxide spin adducts in the effluents. Since we expected that some of the metabolizing properties of the perfused liver may resemble those seen in liver homogenates in vitro EMPO was considered a convenient model because the S cis/trans-EMPO/ OH deconvolution is straightforward, and the S two studied mechanisms of nitrone/ OH production are best distinguished by this compound (Table 4). The concentration range of nitrones (5–15 mM) was selected according to our previous results showing the absence of deleterious effects of DMPO and DEPMPO on rat liver energetic metabolism [50]. In the present study, lactate dehydrogenase efflux measured during the normoxic period (1–12 mU/g liver wet weight/min) did not evidence any toxic effect of all perfused nitrones. During the normoxic period the effluents from nitroneperfused livers were diamagnetic (see Fig. 6A for DIPPMPO). Upon reperfusion following 90 min of warm ischemia strong S ESR signals characteristic of pure nitrone/ OH adducts were detected in the hepatic effluents (Figs. 6B, 6D, and 6E). In the S two series of experiments, nitrone/ OH concentration peaked after 4–6 min for EMPO and after 6–9 min for DEPMPO and DIPPMPO, ranging 0.3–1.2, 6.5–12.7, and 5.6–11.8 nmol/ml/g wet weight, respectively. Complete inhibition of the ESR signal in livers receiving excess of SOD during reflow indicated the S S involvement of O2 − in nitrone/ OH formation (Fig. 6C). S − Formation of O2 in the ischemic-reperfused liver has been proposed in several studies but the cellular sources and locations are still controversial [51–53]. Only one ESR study on this topic in which the authors applied an ex situ spin trapping technique has been published [54] but high levels of

Fig. 6. ESR spectra of effluent perfusate from a rat liver perfused with 5 mM DIPPMPO (A–C) or DEPMPO (D) or 15 mM EMPO (E) and typical ESR signal obtained by carrying out a Fenton reaction in the presence of 0.1 M DIPPMPO in Krebs–Henseleit buffer (inset). (A) Preischemic control; (B, D), liver submitted to 90 min of ischemia at 37°C followed by 9 (5) min of reperfusion; (C) same as B in the presence of SOD (10 units/ml) added to the perfusate throughout reperfusion; (E) same as B except that effluent was collected after 1 min reperfusion. Nitrones were individually perfused preischemically during 10 min and throughout reperfusion. The percentages of cis-nitrone/ OH were: 9.4 (B), 21.5 (D), and 16.5 (E). Spectrometer settings for A–D (E): microwave power, 10 mW; modulation amplitude, 0.06 (0.04) mT; time constant, 20.48 (1.28) ms; gain, 105 (2.5 × 105); scan rate, 0.54 (0.07) mT.s−1. ESR acquisition was initiated 49 ± 3 s after effluent sampling (A–E) or 2 min after addition of CAT (Fenton).

S

S

S

contaminating DMPO/ OH and / alkyl signals in their control experiments seriously hampered the observation of any reliable reperfusion-induced spin adduct formation. In our study we used nitrones less susceptible to self-decomposition into S nitroxides than DMPO [36] and whose O2 − adducts do not S decompose into nitrone/ OH in vitro [2–4]. This latter property S of EMPO, DEPMPO, and DIPPMPO implies that nitrone/ OH signals seen in Figs. 6B, 6D, and 6E may uniquely derive from S the reduction of primary O2 − adducts in the liver and not from S direct HO trapping as stated previously [54]. Strongly S supporting this hypothesis were the cis-nitrone/ OH percentages which ranged 16–23 and 19–25% for EMPO and DEPMPO, respectively, closely matching the data of Table 4 for the GPx-GSH or rat liver homogenate systems. The fact that


SOD, which cannot penetrate cells when exogenously perfused, S scavenges O2 − several orders of magnitude faster than nitrones S implies that only extracellular O2 − is trapped primarily and that S nitrone/ OOH adducts can diffuse within cells to undergo reduction by intracellular GPx. In conclusion, the results of the present study, which do not S rule out a central role for HO in reperfusion injury of the liver, highlight the difficulty in detecting hydroperoxides such as S nitrone/ OOH adducts in highly metabolizing biological S systems and confirm that the detection of HO spin adducts cannot be a definitive proof that this highly reactive radical could have a greater affinity for spin traps than for the many other surrounding targets. In this respect, the results with DIPPMPO in the liver are rather illustrative since the lack of any radical contamination in the signal of Fig. 6B as compared to S the Fenton system (Fig. 6, inset) definitively show that no HO was trapped. Hence we believe that this apparent drawback of DIPPMPO may reveal a great advantage in the ESR investigaS tion of biological situations involving HO . Acknowledgments We thank Patrick Bernasconi for his helpful technical assistance on the ESR spectrometer and Paul Tordo, JeanPierre Finet, and Alain Guichard for helpful discussion. A.R. expresses his thanks for the partial financial support of the Hungarian Scientific Research Fund (OTKA T-046953) and the Grant NKFP 1/A/005/2004 “MediChem2.” References [1] Kotake, Y.; Kuwata, K.; Janzen, E. G. Electron spin resonance spectra of diastereoisomeric nitroxyls produced by spin trapping hydroxyalkyl radicals. J. Phys. Chem. 83:3024–3029; 1979. [2] Fréjaville, C.; Karoui, H.; Tuccio, B.; Le Moigne, F.; Culcasi, M.; Pietri, S.; Lauricella, R.; Tordo, P. 5-(Diethoxyphosphoryl)-5-methyl-1-pyrroline N-oxide: a new efficient phosphorylated nitrone for the in vitro and in vivo spin trapping of oxygen-centered radicals. J. Med. Chem. 38:258–265; 1995. [3] Chalier, F.; Tordo, P. 5-Diisopropoxyphosphoryl-5-methyl-1-pyrroline Noxide, DIPPMPO, a crystalline analog of the nitrone DEPMPO: synthesis and spin trapping properties. J. Chem. Soc. [Perkin 2] 2110–2117; 2002. [4] Olive, G.; Mercier, A.; Le Moigne, F.; Rockenbauer, A.; Tordo, P. 2Ethoxycarbonyl-2-methyl-3,4-dihydro-2H-pyrrole-1-oxide: evaluation of the spin trapping properties. Free Radic. Biol. Med. 28:403–408; 2000. [5] Barbati, S.; Clément, J. L.; Olive, G.; Roubaud, V.; Tuccio, B.; Tordo, P. 31P labeled cyclic nitrones: a new class of spin traps for free radicals in biological milieu. In: Minisci, F., ed. Free Radicals in Biology and Environment, NATO ASI Series, A, Life Sciences, Kluwer Acad. Publishers, Dordrecht, 27:39–47; 1997. [6] Roubaud, V.; Mercier, A.; Olive, G.; Le Moigne, F.; Tordo, P. 5(diethoxyphosphorylmethyl)-5-methyl-4,5-dihydro-3H-pyrrole N-oxide: synthesis and evaluation of spin trapping properties. J. Chem. Soc. [Perkin 2] 1827–1830; 1997. [7] Zhang, H.; Joseph, J.; Vasquez-Vivar, J.; Karoui, H.; Nsanzumuhire, C.; Martásek, P.; Tordo, P.; Kalyanaraman, B. Detection of superoxide anion using an isotopically labeled nitrone spin trap: potential biological applications. FEBS Lett. 473:58–62; 2000. [8] Stolze, K.; Udilova, N.; Nohl, H. Spin trapping of lipid radicals with DEPMPO-derived spin traps: detection of superoxide, alkyl and alkoxyl radicals in aqueous and lipid phase. Free Radic. Biol. Med. 29:1005–1014; 2000.

1537

[9] Clément, J. L.; Barbati, S.; Fréjaville, C.; Rockenbauer, A.; Tordo, P. Synthesis and use as spin trap of 5-methyl-5-phosphono-1-pyrroline Noxide (DHPMPO). pH dependence of the EPR parameters of the spin adducts. J. Chem. Soc. [Perkin 2] 1471–1475; 2001. [10] Zhao, H.; Joseph, J.; Zhang, H.; Karoui, H.; Kalyanaraman, B. Synthesis and biological applications of a solid cyclic nitrone spin trap: a relatively superior trap for detecting superoxide anions and glutathiyl radicals. Free Radic. Biol. Med. 31:599–606; 2001. [11] Tsai, P.; Elas, M.; Parasca, A. D.; Barth, E. D.; Mailer, C.; Halpern, H. J.; Rosen, G. M. 5-carboxy-5-methyl-1-pyrroline N-oxide: a spin trap for the hydroxyl radical. J. Chem. Soc. [Perkin 2] 875–880; 2001. [12] Villamena, F. A.; Hadad, C. M.; Zweier, J. L. Kinetic study and theoretical analysis of hydroxyl radical trapping and spin adduct decay of alkoxycarbonyl and dialkoxyphosphoryl nitrones in aqueous media. J. Phys. Chem. A 107:4407–4414; 2003. [13] Tsai, P.; Ichikawa, K.; Mailer, C.; Pou, S.; Halpern, H. J.; Robinson, B. H.; Nielsen, R.; Rosen, G. M. Esters of 5-carboxyl-5-methyl-1-pyrroline Noxide: a family of spin traps for superoxide. J. Org. Chem. 68:7811–7817; 2003. [14] Stolze, K.; Udilova, N.; Rosenau, T.; Hofinger, A.; Nohl, H. Synthesis and characterization of EMPO-derived 5,5-disubstituted 1-pyrroline N-oxides as spin traps forming exceptionally stable superoxide spin adducts. Biol. Chem. 384:493–500; 2003. [15] Nsanzumuhire, C.; Clément, J. L.; Ouari, O.; Karoui, H.; Finet, J. P.; Tordo, P. Synthesis of the cis diastereoisomer of 5-diethoxyphosphoryl-5methyl-3-phenyl-1-pyrroline N-oxide (DEPMPPOc) and ESR study of its superoxide spin adduct. Tetrahedron Lett. 45:6385–6389; 2004. [16] Villamena, F. A.; Rockenbauer, A.; Gallucci, J.; Velayutham, M.; Hadad, C. M.; Zweier, J. L. Spin trapping by 5-carbamoyl-5-methyl-1-pyrroline Noxide (AMPO): theoretical and experimental studies. J. Org. Chem. 69:7994–8004; 2004. [17] Stolze, K.; Udilova, N.; Rosenau, T.; Hofinger, A.; Nohl, H. Spin adduct formation from lipophilic EMPO-derived spin traps with various oxygenand carbon-centered radicals. Biochem. Pharmacol. 69:297–305; 2005. [18] Stolze, K.; Rohr-Udilova, N.; Rosenau, T.; Stadtmüller, R.; Nohl, H. Very stable superoxide radical adducts of 5-ethoxycarbonyl-3,5-dimethylpyrroline N-oxide (3,5-EDPO) and its derivatives. Biochem. Pharmacol. 69:1351–1361; 2005. [19] Dikalov, S.; Jiang, J.; Mason, R. P. Characterization of the high-resolution ESR spectra of superoxide radical adducts of 5-(diethoxyphosphoryl)-5methyl-1-pyrroline N-oxide (DEPMPO) and 5,5-dimethyl-1-pyrroline Noxide (DMPO). Analysis of conformational exchange. Free Radic. Res. 39:825–836; 2005. [20] Xu, Y.-K.; Chen, Z.-W.; Sun, J.; Liu, K.; Chen, W.; Shi, W.; Wang, H.-M.; Liu, Y. Synthesis, crystal structure, and ESR study of a novel phosphorylated lipophilic spin trap. J. Org. Chem. 67:7624–7630; 2002. [21] Rockenbauer, A.; Györ, M.; Hankovsky, H. O.; Hideg, K. ESR study of the conformation of 5- and 6-membered cyclic nitroxides (aminoxyl) radicals. In: Symons, M. C. R., ed. Electron Spin Resonance, vol. 11A, Cambridge: Royal Society of Chemistry; 1988: pp. 145-182 [22] Rockenbauer, A.; Korecz, L.; Hideg, K. Ring pseudorotation in pyrrolidine-N-oxyl radicals: an analysis of 13-C hyperfine structure of EPR spectra. J. Chem. Soc. [Perkin 2] 2140–2155; 1993. [23] Villamena, F. A.; Hadad, C. M.; Zweier, J. L. Theoretical study of the spin trapping of hydroxyl radical by cyclic nitrones: a density functional approach. J. Am. Chem. Soc. 126:1816–1829; 2004. [24] Khramtsov, V.; Berliner, L. J.; Clanton, T. L. NMR spin trapping: detection of free radical reactions using a phosphorus-containing nitrone spin trap. Magn. Reson. Med. 42:228–234; 1999. [25] Le Moigne, F.; Mercier, A.; Tordo, P. β-Phosphorylated cyclic nitroxides. 2. Synthesis of pyrrolidin- and piperidin-2-yl phosphonates and the corresponding stable nitroxides. Tetrahedron Lett. 32:3841–3844; 1991. [26] Mathieu, C.; Mercier, A.; Witt, D.; Debmkowski, L.; Tordo, P. βphosphorylated nitroxides in the pyrrolidine series: reduction by ascorbate. Free Radic. Biol. Med. 22:803–806; 1997. [27] Barbati, S.; Clément, J. L.; Fréjaville, C.; Bouteiller, J. C.; Tordo, P.; Michel, J. C.; Yadan, J. C. 31P-labeled pyrroline N-oxides: synthesis of 5-diethylphosphone-5-methyl-1-pyrroline N-oxide (DEPMPO) by

1538

[28]

[29]

[30]

[31]

[32]

[33] [34]

[35]

[36] [37]

[38]

[39]

[40]

[41]

M. Culcasi et al. / Free Radical Biology & Medicine 40 (2006) 1524–1538 oxidation of diethyl (2-methylpyrrolidin-2-yl)phosphonate. Synthesis 12: 2036–2040; 1999. Desmoulin, F.; Canioni, P.; Crotte, C.; Gérolami, A.; Cozzone, P. J. Hepatic metabolism during acute ethanol administration: a phosphorus-31 nuclear magnetic resonance study on perfused rat liver after normoxic or hypoxic conditions. Hepatology 7:315–323; 1987. Zar, H. A.; Tanigawa, K.; Kim, Y.-M.; Lancaster, J. R. Rat liver postischemic lipid peroxidation and vasoconstriction depend on ischemia time. Free Radic. Biol. Med. 25:255–264; 1998. Couet, W. R.; Brasch, R. C.; Sosnovsky, G.; Tozer, T. N. Factors affecting nitroxide reduction in ascorbate solution and tissue homogenates. Magn. Reson. Imaging 3:83–88; 1985. Iannone, A.; Tomasi, A.; Vannini, V.; Swartz, H. M. Metabolism of nitroxide spin labels in subcellular fractions of rat liver. II. Reduction in the cytosol. Biochim. Biophys. Acta 1034:290–293; 1990. Polnaszek, C. F.; Schreier, S.; Butler, K. W.; Smith, I. C. P. Analysis of the factors determining the EPR spectra of spin probes that partition between aqueous and lipid phases. J. Am. Chem. Soc. 100:8223–8232; 1978. Rockenbauer, A.; Korecz, L. Automatic computer simulations of ESR spectra. Appl. Magn. Reson. 10:29–43; 1996. Zaplatin, A. N.; Baker, K. A.; Kleinhaus, F. W. Effectiveness and toxicity of several DTPA broadening agents for biological ESR spectroscopy. J. Magn. Reson. B 110:249–254; 1996. Burkitt, M. J. Chemical, biological and medical controversies surrounding the Fenton reaction In: Science Reviews, vol. 28, Progress in Reaction Kinetics and Mechanism; 2003: pp. 75–103. Finkelstein, E.; Rosen, G. M.; Rauckman, E. J. Spin trapping of superoxide. Mol. Pharmacol. 16:676–685; 1979. Egan, T. J.; Barthakur, S. R.; Aisen, P. Catalysis of the Haber-Weiss reaction by iron-diethylenetriaminepentaacetate. J. Inorg. Biochem. 48:241–249; 1992. Karoui, H.; Hogg, N.; Fréjaville, C.; Tordo, P.; Kalyanaraman, B. Characterization of sulfur-centered radical intermediates formed during the oxidation of thiols and sulfite by peroxinitrite. ESR-spin trapping and oxygen uptake studies. J. Biol. Chem. 271:6000–6009; 1996. Grossi, L. Evidence of an electron-transfer mechanism in the peroxynitritemediated oxidation of 4-alkylphenols and tyrosine. J. Org. Chem. 68:6349–6353; 2003. Buettner, G. R.; Oberley, L. W. Considerations in the spin trapping of superoxide and hydroxyl radical in aqueous systems using 5,5-dimethyl-1pyrroline-1-oxide. Biochem. Biophys. Res. Commun. 83:69–74; 1978. Tuccio, B.; Lauricella, R.; Fréjaville, C.; Bouteiller, J.-C.; Tordo, P. Decay of hydroperoxyl spin adduct of 5-diethoxyphosphoryl-5-methyl-1-pyrroline N-oxide: an EPR kinetic study. J. Chem. Soc. [Perkin 2] 295–298; 1995.

[42] Villamena, F. A.; Zweier, J. L. Superoxide radical trapping and spin adduct decay of 5-tert-butoxycarbonyl-5-methyl-1-pyrroline N-oxide (BocMPO): kinetics and theoretical analysis. J. Chem. Soc. [Perkin 2] 1340–1344; 2002. [43] Karoui, H.; Clément, J.-L.; Rockenbauer, A.; Siri, D.; Tordo, P. Synthesis and structure of 5,5-diethoxycarbony-1-pyrolline N-oxide (DECPO). Application to superoxide radical trapping. Tetrahedron Lett. 45: 149–152; 2004. [44] Tsai, P.; Pou, S.; Straus, R.; Rosen, G. M. Evaluation of various spin traps for the in vivo in situ detection of hydroxyl radical. J. Chem. Soc. [Perkin 2] 1759–1763; 1999. [45] Khan, N.; Wilmot, C. M.; Rosen, G. M.; Demidenko, E.; Sun, J.; Joseph, J.; O’Hara, J.; Kalyanaraman, B.; Swartz, H. M. Spin traps: in vitro toxicity and stability of radical adducts. Free Radic. Biol. Med. 34:1473–1481; 2003. [46] Perkins, M. J. Spin trapping. Adv. Phys. Org. Chem. 17:1–64; 1980. [47] Pietri, S.; Mercier, A.; Mathieu, C.; Caffaratti, S.; Culcasi, M. Hemodynamic and metabolic effects of the β-phosphorylated nitroxide 2-diethoxyphosphoryl-2,5,5-trimethylpyrrolidinyloxyl during myocardial ischemia and reperfusion. Free Radic. Biol. Med. 34:1167–1177; 2003. [48] Halliwell, B. How to characterize a biological antioxidant. Free Radic. Res. Communs. 9:1–32; 1990. [49] Villamena, F. A.; Dickman, M. H.; Crist, D. L. R. Nitrones as ligands in complexes of Cu(II), Mn(II), Co(II), Ni(II), Fe(II), and Fe(III) with N-tertbutyl-a-(2-pyridyl)nitrone and 2,5,5-trimethyl-1-pyrroline-N-oxide. Inorg. Chem. 37:1446–1453; 1998. [50] Delmas-Beauvieux, M. C.; Pietri, S.; Culcasi, M.; Leducq, N.; Valeins, H.; Liebgott, T.; Diolez, P.; Canioni, P.; Gallis, J. L. Use of spin traps during warm-ischemia-reperfusion in rat liver: comparative effect on energetic metabolism studied using 31P NMR. MAGMA 5:45–52; 1997. [51] McKelvey, T. G.; Höllwarth, M. E.; Granger, D. N.; Engerson, T. D.; Landler, U.; Jones, H. P. Mechanisms of converssion of xanthine dehydrogenase to xanthine oxidase in ischemic rat liver and kidney. Am. J. Physiol. 254:G753–G760; 1988. [52] Jaeschke, H.; Farhood, A. Neutrophil and Kupffer cell-induced oxidant stress and ischemia-reperfusion injury in rat liver. Am. J. Physiol. 260: G355–G362; 1991. [53] Littauer, A.; De Groot, H. Release of reactive oxygen by hepatocytes on reoxygenation: three phases and role of mitochondria. Am. J. Physiol. 262: G1015–G1020; 1992. [54] Ohno, S.; Togashi, H.; Shinzawa, H.; Matsuo, T.; Sugahara, K.; Shao, L.; Kamada, H.; Takahashi, T. Reactive oxygen species originate from xanthine oxidase after reperfusion of the ischemic rat liver and activate nuclear factor kappa B in the liver. Res. Commun. Biochem. Cell Mol. Biol. 3:69–94; 1999.