The Requirement for Iron(II1) in the Initiation of Lipid Peroxidation by Iron(I1) and Hydrogen Peroxide*

THE JOURNAL OF BIOLOGICAL CHEMISTRY Q 1987 by The American Society of Biological Chemists, Inc. Vol. 262, No. 3, Isaue of January 25, pp. lg9&1104,1987 Prinred in U.S.A. The Requirement for Iron(II1) in the Initiation of Lipid Peroxidation by Iron(I1) and Hydrogen Peroxide* (Received for publication, May 15, 1986) Giorgio Minotti# and Steven D. Auste From the Center ~ for the Study of Active Oxygen in Biology and Medicine, Department of Biochemistry, Michigan State University, East Lansing, Michigan 48824-1319 The initiation of lipid peroxidation by Fe2+ and H202 (Fenton s reagent) is often proposed to be mediated by the highly reactive hydroxyl radical. Using Fe2*, H202, and phospholipid liposomes as a model system, we have found that lipid peroxidation, as assessed by malondialdehyde formation, is not initiated by the hydroxyl radical, but rather requires FeS+ and Fe2+. EPR spin trapping with 5,6-dimethyl-l-pyrroline-N-oxide and the bleaching of para-nitrosodimethylaniline confirmed the generation of the hydroxyl radical in this system. Accordingly, catalase and the hydroxyl radical The final step of this sequence, i.e. the oxidation of Fez+ by H202, is also referred to as Fenton s reaction (15). The intrinsic ability of OH to initiate membrane lipid scavengers mannitol and benzoate efficiently inhibited peroxidation has been questioned by some investigators (12, the generation and the detection of hydroxyl radical. 16-18). This criticism is in part a consequence of the reali- However, catalase, mannitol, and benzoate could either zation that OH is a short-lived radical which reacts with stimulate or inhibit lipid peroxidation. These unusual most organic compounds at nearly diffusion-controlled rates effects were found to be consistent with their ability to (19). Such indiscriminate reactivity makes unlikely a migramodulate the extent of Fe2+ oxidation by H20, and tion of OH from the sit&) of generation to the hydrophobic demonstrated that lipid peroxidation depends on the membrane compartments where lipid peroxidation must be FeS+:Fe2+ ratio, maximal initial rates occurring at 1:l. initiated. In agreement, many in vitro lipid peroxidation sys- These studies suggest that the initiation of liposomal tems are not inhibited by catalase or.oh scavengers (12,16- peroxidation by Fe2+ and H20z is mediated by an oxi- 18). Stimulation of lipid peroxidation by catalase or OH dant which requires both FeS+ and Fe2+ and that the scavengers is also reported in the literature (20,21). rate ofthereaction is determined by theabsolute These observations provided a rationale for the hypothesis FeS+:Fe2+ ratio. It is becoming increasingly evident that free radical-mediated peroxidation of cell membranes is a common pathway in the onset of several disease states, varying from drug-associated toxicity (1, 2) to postischemic reperfusion injury (3) and carcinogenesis (4-6). Partially reduced forms of dioxygen, like superoxide (0;) and its dismutation product hydrogen peroxide (HzOz), have been implicated in the initiation of lipid peroxidation (7, 8). However, it has been shown that neither 0; (9) nor HzOz (10) or the uncatalyzed reaction of 0; with HzOz (11) can directly initiate lipid peroxidation. A stringent requirement for iron in several in vitro lipid peroxidation systems suggests that a metal-driven reaction is the source of a more reactive ultimate oxidant (12). The hydroxyl radical ( OH) is frequently proposed as the initiating species (13), * This work was supported in part by National Institutes of Health Grant GM 33443. The costs of publication this of article were defrayed in part by the payment of page charges. This article must therefore beherebymarked advertisement inaccordancewith18 U.S.C. Section 1734 solely to indicate this fact. 4 FellowofAssociazioneItalianaRicercasulCancroandThe Council for International Exchange of Scholars (Fulbright Program). Permanent address: Institute of General Pathology, Catholic University, 00168 Rome, Italy. 5 To whom correspondence should be addressed. 1098 and the iron-catalyzed Haber-Weiss reaction is described as the sequence through which Oz, H202, and iron rapidly produce OH (14). 0; + Fe3+ + O2 + Fez+ (1) 20; + ZH+ + O2 + H202 (2) Fez+ + H202 --., Fe3+ + OH- + OH (3) of alternate initiators, tentatively described as a complex between oxygen and different valence states of iron. In this respect, much experimental attention has been given to the ferry1 ion (FeO: or FeOH3+) (22, 23), the perferryl ion (Fez+- 0, or Fe3+-02) (12, 16, 24), and a ferrous-dioxygen-ferric complex (25). This paper reports our studies of a lipid peroxidation system involving Fez+, HZOZ, and phospholipid liposomes. Several aspects of this system have been investigated the generation of OH, the autoxidation and HzOa-dependent oxidation of Fez+; the time course of MDA formation; the effects of catalase and OH scavengers like mannitol or benzoate. We have evidence that Hz02 simply serves as an oxidant for Fez+ and that both Fe3+ and Fez+ are required for initiation. The absolute Fe3+:Fe2+ ratio determines the initial rate and the extent of lipid peroxidation. Consistently, catalase, mannitol, and benzoate appear to stimulate or inhibit lipid peroxidation by affecting specifically the Fe3+:Fe2+ ratio, rather than the generation of *OH or its interaction with the phospholipids. EXPERIMENTAL PROCEDURES Materiuk-Mannitol, Z-thiobarbituric acid, butylated hydroxytoluene,brij35, Lubrol PX, 1,lO-o-phenanthroline, DMPO, and 4- aminoantipyrine were purchased from Sigma. Benzoate was from J. T. Baker Chemical Co.; FeClz and phenol from Mallinckrodt Chemical Works; H20z and p-nitrosodimethylaniline from Aldrich. Seph- l The abbreviations used are: MDA, malondialdehyde; EPR, electronparamagneticresonance;dmpo, 5,5-dimethyl-1-pyrroline-Noxide; p-nda, para-nitrosodimethylaniline.

Importance of the Ferric:Ferrous Ratio in Lipid Peroxidation 1099 adex G-25 was a product of Pharmacia P-L Biochemicals. Chelex 100 ion-exchange resin was purchased from Bio-Rad and used to remove trace contaminating metals from buffers and reagents. DMPO was vacuum distilled and further filtered through water-washed activated charcoal as described by Floyd and Wiseman (26). All other chemicals were of analytical grade or better and used without further purification. Enzymes-Bovine liver catalase (EC 1.11.1.6) and type VI salt-free horseradish peroxidase (EC.1.11.1.7) were from Sigma. Catalase was separated from the antioxidant thymol by gel chromatography on Sephadex G-25. Following chromatography the enzyme activity was measured by the method of Beers and Sizer (27). Preparation of Microsomes, Lipids, and Liposomes-Male Sprague- Dawley rats (250-280 g) were obtained from Charles River Breeding Laboratories, Inc. (Wilmington, MA). Liver microsomes were isolated following the procedure of Pederson and Aust (28). Lipids were extracted from freshly prepared microsomes as per Folch et at. (29). All steps were performed at 4 "C with argon-purged buffers and solvents in order to minimize the autoxidation of unsaturated lipids. Extracted lipids were stored in argon-saturated CHC13:CH30H (2:l) at -2O'C. The lipid phosphate content was assayed according to the procedure of Bartlett (30). Liposomes were prepared by indirect anaerobic sonication of the extracted microsomal lipids (16). Lipid Peroxidation Assays-Lipid peroxidation was performed by incubating phospholipid liposomes (1 pmol of lipid phosphate/ml) in 50 mm NaC1, ph 7.0, at 37 "C in a Dubnoff metabolic shaker bath under an air atmosphere. Unless otherwise specified the reactions were initiated by the addition of 0.1 mm I3202 and 0.2 mmfec1,. Aliquots of the incubation mixture were removed at regular time intervals and assayed for MDA by the thiobarbituric acid test (31). Butylated hydroxytoluene (0.03 volume of 2% in ethanol) was added to the thiobarbituric acid reagent in order to prevent iron-catalyzed decomposition of lipid hydroperoxides and the formation of MDA during the heating step of the assay (31). All solutions were carefully adjusted to ph 7.0 just prior to use. Stock solutions of FeCl, were prepared daily by adding the solid iron salt to 50 mm NaCl, ph 7.0, exhaustively bubbled with argon. The solution was capped, protected from light, and used within 4 h. EPR Spin Trapping-Unless otherwise specified, 0.1 mm Hz02, 0.2 mm FeCl,, and 10 mm DMPO in 50 mm NaC1, ph 7.0, were mixed in a test tube, transferred to a flat quartz cuvette, and scanned after 2 min in a Varian Century 112 EPR spectrometer at room temperature. The experimental conditions were: 3320 G magnetic field, 15 milliwatts incident power, 9.412 GHz, 100 khz modulation frequency, 1 G modulation amplitude, 100 G scan range, 0.5-9 time constant, and 4-min scan time. Bleaching of p-nda-the bleaching of p-nda was monitored spectrophotometrically as a kinetic index of 'OH formation (32). The reaction was started by the addition of 0.1 mm HZ02 and 0.2 mm FeC1, to a cuvette containing a 50 pm solution of p-nda in 50 mm NaCl, ph 7.0. The cuvette chamber was maintained at 37 'C, and the absorbance change at 440 nm was monitored continuously. The extinction coefficient of p-nda was 34,200 X lo3 M" cm". Assay for Ferrous Iron-The concentration of ferrous iron was measured colorimetrically using phenanthroline (33). The incubation mixtures were constituted as described for the lipid peroxidation assay, with the exception that lipids were omitted. Aliquots (0.5 ml) were removed periodically and mixed vigorously with 1 ml of 15 mm 1,lO-o-phenanthroline. The absorbance of the phenanthroline-ferrous complex was read at 510 nm against a reagent blank. A standard curve with known amounts offec12 in 1,lO-o-phenanthroline was used for calculations. No change in the absorbance at 510 nm of a solution containing 0.2 mm FeC1, and 15 mm 1,lO-o-phenanthroline was observed upon the addition of0.1 mm H,02, indicating that phenanthroline-chelated ferrous was not oxidized by H,O, during the assay. Assay for H20z-Horseradish peroxidase was used to assay for Hz02 (34). The experimental conditions were the same described for the assay of iron. Samples (0.5 ml) of the reaction mixture were mixed with 1-ml aliquots of a solution containing 24.8 mm phenol, 4.3 mm 4-aminoantipyrine, and 19 units/ml horseradish peroxidase in 1 mm potassium phosphate buffer, ph 6.9. The coupled oxidation of phenol and 4-aminoantipyrine by H,O, in the presence of horseradish peroxidase gave rise to a quinone-imine adduct exhibiting a maximal absorbance at 505 nm. A standard curve with known amounts of H,O, was used for calculations. RESULTS AND DISCUSSION The Initiation of Lipid Peroxidation by Fez+ and HzOz-The peroxidation of phospholipid liposomes was initiated by 0.1 mm Hz02 and 0.2 mm Fez+ (Fig. 1). Neither H202 nor Fez+ alone induced lipid peroxidation; this confirms that HzO, is virtually unreactive with phospholipids and that both Fez+ and H,Oz are required for the generation of the initiating species. The absence of significant MDA formation by Fez+ alone also suggests that the extracted microsomal lipids were essentially free of lipid hydroperoxides since no lipid peroxidation was initiated via a ferrous-catalyzed breakdown of preexisting lipid hydroperoxides (17). Fez+-catalyzed hydroperoxide-dependent lipid peroxidation was observed when liposomes were incubated with exogenous organic hydroperoxides. For instance, the addition of 10 or 25 pc~ cumene hydroperoxide resulted in rates of Fez+-dependent lipid peroxidation of 2.16 and 4.6 nmol of MDA/ml/min, respectively. The Generation of. OH and the Effects of. OH Scavengers- EPR spin trapping experiments showed the formation of 'OH by 0.1 mm H202 and 0.2 mm Fez+ (Fig. 2). Using 10 mm DMPO as a spin trap, the typical DMPO-OH adduct having 30 t 25 20. 0 2 4 6 8 1 0 1 2 TIME (min) FIG. 1. Ferrous and hydrogen peroxide-dependent peroxidation of phospholipid liposomes. Reaction mixtures contained phospholipid liposomes (1 Nmolof lipid phosphate/ml) in 50 mm NaCl, ph 7.0, at 37 "C. Peroxidation was initiated by the addition of 0.1 mm H202 and 0.2 mm FeClz (0); 0.1 mm H202 (0); 0.2 mm FeCl, (X). At the specified times, aliquota from the reaction mixture were assayed for MDA content as described under "Experimental Procedures." FIG. 2. EPR spin trapping of hydroxyl radical generated by ferrous and hydrogen peroxide. &action mixtures contained 10 mm DMPO in 50 mm NaCl, ph 7.0. Additions were made aa follows: A, 0.1 mm H202 and 0.2 mm FeCl,; B, 0.1 mm H20z; C, 0.2 mm FeC1,; D, 0.1 mm H202, 0.2 mm FeCl,, and 10 mm mannitol; E, 0.1 mm Hz02, 0.2 mm FeCl,, and 10 mm benzoate. All other conditions were as described under "Experimental Procedures." 1

1100 Importance of the FerricrFerrous Lipid Ratio Peroxidation in a 1:2:2:1 intensity signal pattern, a 2.006 G value, and hyperfine splitting constants AH = AN = 14.9 G wasobserved. Neither Fez+ nor HzOz alone produced DMPO-OH adducts; the relative amplitude of the DMPO-OH adduct was decreased by ' OH scavengers like mannitol (10 mm) or benzoate (10 mm). In order to monitor the time course of 'OH formation the bleaching of p-nda by Fez+ and HZOz was measured spectrophotometrically. Original investigation by Kraljic and Trumbore (32) demonstrated that the chromophore-no group of p-nda was bleached to a colorless -NOz group by pulse radiolysis. Competition experiments with conventional 'OH scavengers revealed that p-nda was bleached specifically by.oh and that thestimated second order rate constant of p- NDA for 'OH was extremely high (1.25 X 10" M" s-'). Bors et al. (22) and Edwards and Quinn (35) subsequently introduced the bleaching of p-nda as a kinetic assay for 'OH in biochemical systems. Hydrogen peroxide (0.1 mm) and 0.2 mm Fez+ catalyzed the irreversible bleaching of 50 I.IM p-nda (Fig. 3). The reaction was rapid and virtually complete in 30 s; again, Fez+ or H2O2 alone could not induce the bleaching of p-nda. Mannitol (10 mm) and benzoate (10 mm) completely inhibited the ' OH-dependent bleaching of p-nda. The ability of mannitol and benzoate to scavenge 'OH was used as a criterion to assess the involvement of this reactive species in the initiation of lipid peroxidation. As shown in Fig. 4, A and B the effects of mannitol and benzoate on Fez+ and HzOz-dependent lipid peroxidation were virtually opposite and did not reflect their common activity as 'OH scavengers. Specifically, 10 mm mannitol produced a short lag phase, followed by the rapid accumulation of MDA. Raising the concentration of mannitol from 10 to 25 or 50 mm simply resulted in the elongation of the lag phase, after which linear rates of lipid peroxidation were observed (Fig. 4A). In contrast, benzoate increased the initial rate of lipid peroxidation; however, the final extent of MDA formation was decreased by benzoate in a concentration-dependent manner (Fig. 4B). In selected experiments liposomes were dispersed in nonionic detergents like Brij 35 or Lubrol PX. Girotti and Thomas (36, 37) have shown that xanthine, xanthine oxidase, and iron-dependent lipid peroxidation of erythrocyte ghosts is inhibited by * OH scavengers only in the presence of nonionic detergents. The authors proposed that the detergents optimize the interception of 'OH by external polar traps, presumably 1 lot -~ ~ 0 30 60 TIME (Sac) FIG. 3. Hydroxyl radical-mediated bleaching of p-dna by ferrous and hydrogen peroxide. Reaction mixtures contained p- NDA (50 p ~ in ) 50 mm NaCl, ph 7.0, at 37 C. The reactions were initiated by the following additions: 0.1 mm H202, 0.2 mm FeC12 (0); 0.1 mm HzOt (0); 0.2 mm FeCl2 (x); 0.1 mm H20Z, 0.2 M FeCL plus 10 mm mannitol (m); 0.1 mm H2O2, 0.2 mm FeClz plus 10 mm benzoate (A). The decrease in absorbance at 440 nm was monitored continuously. TIME (mi") TIME inm1 FIG. 4. The effects of mannitol (A) or benzoate (B) on ferrous and hydrogen peroxide-dependent lipid peroxidation. Reaction mixtures contained phospholipid liposomes (1 rmol of lipid phosphate/ml) in 50 mm NaCl, ph 7.0, at 37 "C. In A peroxidation was initiated by 0.1 mm Hz02 and 0.2 mm FeC12 (0) plus 10 mm (01, 25 mm (X), or 50 mm (I) mannitol. In B peroxidation was initiated by 0.1 mm H2O2, 0.2 mm FeCl, (0) plus 10 mm (O), 25 mm (X), or 50 mm (I) benzoate. 0 I 2 3 4 TIME (min) FIG. 5. The oxidation of ferrous by hydrogen peroxide. FeCl2 (0.2 mm) was incubated in 50 mm NaCl, ph 7.0, at 37 "C in the presence (0) or in the absence (0) of 0.1 mm H202. Aliquots were taken at the specified times and assayed for FeC12 as described under "Experimental Procedures." The arrows indicate the addition of catalase (400 units/ml); the dashed lines indicate the oxidation of FeC12 following the addition of catalase. by unmasking certain iron-binding sites of the phospholipids where 'OH can be formed via a "site-specific" mechanism. The occurrence of site-specific formation of 'OH on the membrane surface would provide a plausible explanation not only for the lack of inhibition by water soluble 'OH traps but also, more importantly, for the mechanism by which 'OH initiates lipid peroxidation despite its short half-lifetime and diffusion-controlled reactivity. In our system the susceptibility of liposomal peroxidation to either mannitol or benzoate was unaffected by the addition of the detergents (data not shown). These scavengers appear, therefore, to increase or decrease the rate and the extent of lipid peroxidation irrespective of the lipid configuration and of their ability to trap freely diffusible or site-specific generated ' OH. The Oxidation of Fez+ by H2O2--In light of these results, the oxidation of Fez+ by HzOz as a potential source of alternate initiator(s) of lipid peroxidation was investigated. The autox- idation of free 0.2 mm Fez+ at neutral ph and an inert medium like NaCl was extremely slow, virtually nonexistent for several minutes (Fig. 5). The lack of any significant redox cycle of Fez+ was consistent with the ineffectiveness of Fez+ itself to generate 'OH, as detected by EPR spin trapping orp-nda bleaching. The addition of 0.1 mm H2O2 promoted a rapid oxidation of 0.2 mm Fez+, 99.3 nmol/ml being oxidized in 30 s; the oxidation of Fez+ was complete in about 4 min. Fig. 5

Importance of the FerricFerrous Ratio in Lipid Peroxidation 1101 also shows that the oxidation of Fez+ by HZOZ was completely prevented by catalase. The addition of catalase at different times of the reaction was similarly effective in preventing further oxidation of Fez+. The oxidation of Fez+ (Fig. 5) closely agreed with the decomposition of H202 (Fig. 6). In the absence of any oxidizable substrate to compete for the 'OH, the H202 decomposition to Fez+ oxidation occurs with a net stoichiometry of 1:2 (38). As shown in Figs. 5 and 6 the oxidation of 200 nmol of Fez+ was paralleled by the decomposition of 100 nmol of H202, with about 50% of the reaction occurring within 30 s. Both mannitol and benzoate were found to affect the oxidation of Fez+ by HzOz, the effects of the two scavengers being especially evident in the early phase of the reaction. Mannitol caused a concentration-dependent decrease in the initial rate of the oxidation of Fez+ by HzOz (Fig. 7A). Conversely, the oxidation of Fez+ by H20z was dramatically stimulated by benzoate, all Fez+ being converted to Fe3+ in 30 s in the presence of 50 mm benzoate (Fig. 7B). In some experiments the oxidation of Fez+ by H202, in the presence or in the absence of mannitol and benzoate, was monitored spectrophotometrically as the increase in the absorbance at 310 nm, which is indicative of the amount of Fe3+ formed (39). The results of these experiments were in total agreement with those obtained by the o-phenanthroline method, confirming that the rate of Fez+ to Fe3+ oxidation was decreased by mannitol and increased by benzoate (data not shown). Mannitol and benzoate affected only the HZOz-mediated oxidation of Fez+; no stimulation of Fez+ autoxidation by 0 I 2 3 4 TIME ( rnin) FIG. 6. The decomposition of hydrogen peroxide by ferrous. Hz02 (0.1 mm) was incubated in 50 mm NaCI, ph 7.0, at 37 "C in the presence (0) or in the absence (0) of 0.2 mm FeCl2. Aliquots of the reaction mixtures were sampled periodically for the content of HzOz as described under "Experimental Procedures." TlYE (tu1 TIME I*) FIG. 7. The effects of mannitol (A) or benzoate (B) on the oxidation of ferrous by hydrogen peroxide. FeCI, (0.2 mm) was incubated in 50 mm NaCI, ph 7.0, at 37 "C. In A the additions were: 0.1 mm H202 (O), 0.1 mm H202 plus 10 mm (O), 25 mm (X), or 50 mm (W) mannitol; 50 mm mannitol (A). In B the additions were: 0.1 mmhzo, (01, 0.1 mm HZ02 plus 10 mm (O), 25 mm (X), or 50 mm (W) benzoate; 50 mm benzoate (A). either mannitol or benzoate was observed (Fig. 7, A and B). Moreover, the inhibition or stimulation of HzOz-dependent Fe2+ oxidation was paralleled by a decrease or an increase in the rate of H202 decomposition, respectively (data not shown). Finally, neither mannitol nor benzoate was found to influence the stability of HZOz at 37 "C(data not shown). We have no explanation for the mechanism(s) by which mannitol and benzoate, or the product(s) of their reaction with 'OH, modify the reactivity of Fez+ with HZOz. However, we believe that the effects of these 'OH scavengers on the redox state of iron provide a clue as to the identification of the reactive species which initiate lipid peroxidation. Bucher et al. (25) showed that the autoxidation of certain ferrous chelates promoted liposomal peroxidation only after a lag phase. Superoxide dismutase, catalase, and.oh scavengers did not extend the lag phase or inhibit the subsequent rate of lipid peroxidation, indicating that the reaction was not initiated by a reduced species of oxygen. The Fez+ autoxidation product required for the initiation of lipid peroxidation was concluded to be Fe3+. As evidence of this, the lag phase was eliminated by the addition of chelated Fe3+, and the initial rates of lipid peroxidation were found to increase linearly with the Fe3+:Fez+ ratio. A Fe3+-Fez+ complex or, more likely, a Fe3+-dioxygen-Fez+ complex was, therefore, proposed to initiate lipid peroxidation. In our system the autoxidation of free Fez+ is too slow for lipid peroxidation to occur for an appreciable period of time. We, therefore, suggest that lipid peroxidation is initiated only when HzOz oxidizes critical amounts of Fez+ to Fe3+. The lag phase in MDA formation observed with mannitol can be ascribed to the inhibition of Fez+ oxidation by H202 and to the subsequent delay in the formation of the Fe3+. Conversely, the increase in the initial rates of MDA formation observed with benzoate may merely reflect the stimulation of Fez+ oxidation, rapidly providing Fe3+ for the initiation of lipid peroxidation. However, the inhibitory effect of benzoate on the final extent of lipid peroxidation implies that no lipid peroxidation occurs when the iron is predominantly or completely in its oxidized form. The Initiation of Lipid Peroxidation by Fez+ and Fe3+-The requirement for Fe3+ and the importance of the Fe3+:Fez+ ratio in the initiation of lipid peroxidation by Fe2+ and HzOz were further investigated in the following experiment. HzOz (0.1 mm) and 0.2 mm Fez+ were incubated for different lengths of time in order to oxidize known amounts of Fez+; then liposomes were added and the initial rates of lipid peroxidation were determined. As shown in Fig. 8, the initial rates of ipid peroxidation were strictly dependent on the extent of Fez+ oxidation. The highest values ofmda were obtained by adding liposomes 30 s after the reaction of Fez+ and HzOz, when the concentration of both Fez+ and Fe3+ was 0.1 mm (see Fig. 5). No lipid peroxidation was observed by incubating liposomes with only Fez+; no lipid peroxidation was observed either by adding liposomes after 4 min, i.e. when all Fez+ had been oxidized to Fe3+ by HzOz (see Fig. 5). These results indicate several important points. First, lipid peroxidation can be initiated only when both Fez+ and Fe3+ are present; neither Fez+ nor Fe3+ alone can promote the peroxidation of polyunsaturated fatty acids. Second, the absolute Fe3+:Fez+ ratio determines the rate of lipid peroxidation, the highest activity being associated with a ratio approaching 1:l. Third, the initiation of lipid peroxidation does not depend on the generation of 'OH. Indeed, the maximal rates of lipid peroxidation were obtained by adding liposomes 30 s after the reaction of Fez+ with H202, when the generation of ' OH, as measured by the bleaching of p-nda, was already

1102 Importance of the Ferric:Ferrous Ratio in Lipid Peroxidation nmol Faw/ml i 5 0 1 0 0 200 200 100 0 nmol FP/ mi FIG. 8. The dependence of lipid peroxidation on the FeS+:Fe8+ ratio. H202 (0.1 mm) and 0.2 mm FeCI2 were incubated in 50 mm NaC1, ph 7.0, at 37 "C for the following lengths of time: 0,5, 10, 15,30, and 60 s; 1.5,2,3, and 4 min. At these times aliquots were taken for the assay of FeC1, and liposomes (1 pmol of lipid phosphate/ ml) were added. The initial rate of lipid peroxidation was determined by sampling for MDA 0, 1, 2, and 4 min following the addition of liposomes. essentially complete (see Fig. 3). Fourth, a mechanism of initiation dependent upon Fez+ and Fe3+ may be consistent with the time course of liposomal peroxidation by 0.1 mm HzOZ and 0.2 mm Fez+. In this respect, one should consider that the complete oxidation of Fez+ by H20z takes about 4 min; within this time span, varying ratios of Fez+ to Fe3+ are formed, each of which is characterized by the ability to promote significant although different rates of peroxidation. This continuous generation of active initiators would explain why lipid peroxidation is linear for about 4 min. Regardless of the participation of Fez+ in lipid hydroperoxide-dependent propagation reactions, this mechanism of initiation would also explain why the rate of lipid peroxidation becomes nonlinear after 4 min, when all the iron has been oxidized by HzOz and no Fez+ is available to combine with Fe3+ to initiate the peroxidation of phospholipids. The Effects of Catalase-The effects of catalase on Fe2+ and HzOz-dependent lipid peroxidation were studied in two different experiments. In the first experiment, lipid peroxidation was initiated by 0.1 mm HzOz and 0.2 mm Fez+; in the second experiment, lipid peroxidation was initiated by adding liposomes 30 s after the reaction of 0.1 mm Fez+ with 0.2 mm HzOZ, when the Fe3+:Fe2+ ratio was 1:1. When the reaction was initiated by 0.1 mm H202 and 0.2 mm Fez+, catalase completely inhibited lipid peroxidation (Fig. 9). In contrast, when lipid peroxidation was started by adding liposomes 30 s after the reaction of Fez+ with H202, the simultaneous addition of catalase produced an impressive increase of lipid peroxidation (Fig. 9). In these experiments, the effects of catalase on 'OH formation were also investigated by EPR spin trapping. In order to reproduce the conditions under which lipid peroxidation was inhibited or stimulated by catalase, DMPO was incubated with 0.1 mm HzOz and 0.2 mm Fez+ or was added 30 s after the reaction of Fez+ with HzOZ. In both cases the simultaneous addition of catalase inhibited the formation of the DMPO-OH adduct (Fig. 10). Importantly, the addition of DMPO 30 s after the reaction of Fez+ and H202 resulted in a DMPO-OH adduct with a signal intensity that was drastically reduced. This was in excellent agreement with the time course of p-nda bleaching, which indicated that after 30 s the generation of 'OH was essentially complete (see Fig. 3). The data demonstrate that the formation of 'OH is invar- 0 2 4 6 8 0 1 2 TIME (mid FIG. 9. The effects of catalase on ferrousandhydrogen peroxide-dependent lipid peroxidation. In A the reaction mixtures contained phospholipid liposomes (1 pmol of lipid phosphate/ ml) in 50 mm NaCI, ph 7.0, at 37 "C. The peroxidation was initiated by the addition of 0.1 mm H202 and 0.2 mm FeC1, and was sampled periodically for MDA formation; in B catalase (400 units/ml) was added. In C liposomes (1 pmol of lipid phosphate/ml) were added 30 s after the reaction of 0.1 mm H202 and 0.2 mm FeC1, in 50 mm NaCl, ph 7.0, at 37 "C; following the addition of liposomes, aliquots were taken at the specified times for MDA assay. In D liposomes were added with catalase (400 units/ml). A A C FIG. 10. The effect of catalase on EPR spin trapping of hydroxyl radical generated by ferrous and hydrogen peroxide. In A the reaction mixture contained 0.1 mm Hz& 0.2 mm FeC12, and 10 mm DMPO in 50 mm NaC1, ph 7.0; in B catalase (400 units/ ml) was added. In C, DMPO (10 mm) was added 30 s after the reaction of 0.1 mm Hz02 with 0.2 mm FeC& in 50 mm NaC1, ph 7.0; in I) DMPO was added with catalase (400 units/ml). iably inhibited by catalase, regardless of the experimental conditions. Thus, the mechanism by which catalase stimulates or inhibits lipid peroxidation must not be mediated by -OH. This mechanism is rather consistent with the function of HzOZ as an oxidant for Fez+ and with the key role of the Fe3+:Fez+ ratio in the initiation of lipid peroxidation. In the first case (0.1 mm HzOz phs 0.2 mm Fez+), catalase prevents completely the oxidation of Fez+ by HzOz (see Fig. 5); in this situation, no Fe3+ is formed and no lipid peroxidation can occur. In the second case (addition of liposomes and catalase I 8 D

Importance of the FerricFerrous Ratio in Lipid Peroridation 1103 after 30 s) catalase prevents the excessive oxidation of Fez+ by H202 and maintains Fez+ and Fe3+ at an optimized 1:l ratio (see Fig. 5). In this latter situation, the prolonged existence of the initiator results in a substantial increase in the formation of MDA. The Effects of Varying the Concentrations of H202-The relationship between Fez+ oxidation and lipid peroxidation was also demonstrated by experiments performed with different concentrations of H20z. The results of these experiments are summarized in Fig. 11, A and B, and compared to those obtained with a standard concentration of 0.1 mm H202. The peroxidation of phospholipid liposomes by 0.05 mm H2O2 and 0.2 mm Fez+ was characterized by the appearance of a short lag phase, followedby the onset of linear rates ofmda formation (Fig. 11A). Conversely, 0.2 mm Fez+ and 0.4 mm H202 promoted a rapid accumulation of MDA; however, the final extent of lipid peroxidation was significantly reduced (Fig. 11A). These contrasting time courses of lipid peroxidation resemble those described with mannitol and benzoate, respectively, and reflect analogous dramatic differences in the redox state of iron. As shown in Fig. 11B, with 0.05 mm H202 the time required to generate a Fe2+:Fe3+ ratio close to 1:l was about 1 min; with 0.4 mm H202 the same ratio was approached in only 5 s. These patterns of lipid peroxidation and Fez+ oxidation would, therefore, reinforce the idea that the extent and the rapidity of Fez+ oxidation determine both the initial rate and the overall amount of MDA formation. Accordingly, when liposomes were added at the times previously specified, i.e. when the Fe3+:Fe2+ ratio was 1:1, and further oxidation of Fez+ was prevented by the simultaneous addition of catalase (see Fig. 11B), the rate and the extent of lipid peroxidation were virtually comparable, regardless of the initial concentration of H2O2 (Fig. 12). The differences in MDA formation shown in Fig. 11A, obtained with 0.05, 0.1, and 0.4 mm HzOz, were therefore solely dependent on the varying lengths of time required to generate the Fe3+ from Fez+ necessary to initiate lipid peroxidation. In conclusion, these studies indicate that the oxidation of Fez+ by H202 generates a strong oxidant which is capable of peroxidizing phospholipid liposomes. The oxidant requires both Fez+ and Fe3+, and optimum activity occurs at approxi- mately 1:l. Factors affecting the Fe3+:Fe2+ ratio may have important consequences on' the rate and the extent of lipid peroxidation. These factors vary from catalase to chemicals, like mannitol and benzoate, that are commonly thought of as 'OH scavengers but also affect the rate of H202-dependent TIME (nin I TIME lmml FIG. 11. The effects of varying the concentration of hydrogen peroxide on (A) the peroxidation of phospholipid liposomes and (B) the oxidation of ferrous. In A the reaction mixtures contained phospholipid liposomes (1 pmol of lipid phosphate/ ml) in 50 mm NaCl, ph 7.0, at 37 "C. Peroxidation was initiated by the addition of0.2mmfec12 and the following concentrations of H202: 0.1 mm (0),0.05 mm (o), 0.4 mm (X). In B FeClz (0.2 mm) was incubated in 50 mm NaCl, ph 7.0, at 37 "C in the presence of 0.1 mm (O), 0.05 mm (O), 0.4 mm (X) H202. The arrows indicate the addition of catalase (400 units/ml); the dashed lines indicate the oxidation of FeC12 following the addition of catalase. 60 r"---l TIME (mid FIG. 12. The dependence of lipid peroxidation on the FeS+:Fea+ ratio in the presence of varying concentrations of hydrogen peroxide. FeCI, (0.2 mm) was incubated in 50 mm NaCl, ph 7.0, at 37 "c in the presence of 0.1 mm (O), 0.05 mm (0), or 0.4 mm H202 (X) for 30, 60, or 5 s, respectively, in order to generate a 1:l Fe3+:Fe2+ ratio; at these times catalase (400 units/ml) and liposomes (1 pmol of lipid phosphate/ml) were added. Following the addition of liposomes, aliquots of the reaction mixtures were withdrawn periodically and assayed for MDA content. Fez+ oxidation. The oxidation of Fez+ by H202 generates.oh as well. However, in this system, 'OH does not appear to participate in the initiation of lipid peroxidation. Acknowledgments-We would like to thank Dr. Alfred Haug for the use of the Varian Century-112 EPR spectrometer, Lee A. Morehouse and Craig E. Thomas for helpful discussions during the preparation of the manuscript, and Teresa L. Vollmer for expert secretarial assistance. REFERENCES 1. Myers, C., McGuire, W. P., Liss, R. H., Ifrim, I., Grotzinger, K., and Young, R. C. (1977) Science 19, 165-167 2. Mimnaugh, E. G., Trush, M. A., and Gram, T. E. (1981) Biochem. Phurmacol. 30, 2797-2804 3. Nayini, N. R., White, B. C., Aust, S. D., Huang, R. R., Indrieri, R. J., Evans, T. A., Bialek, H., Jacobs, W. A., and Komara, D. 0. (1985) J. Free Rod. Biol. Med. 1, 111-116 4. Pietronigro, D. B., Jones, W. B., Kalty, K., and Demopoulos, H. B. (1977) Nature 267,78-79 5. Petkau, A. (1980) Acta Physwl. Scand. SuppL 492,81-99 6. Emerit, I., Khan, S. H., and Cerutti, P. A. (1985) J. Free Rad. Biol. Med. 1, 51-57 7. Kellog, E. W., 111, and Fridovich, I. (1975) J. Bwl. Chem. 250, 8812-8817 8. Kellog, E. W., 111, and Fridovich, I. (1977) J. Biol. Chem. 262, 6721-6728 9. Fong, K.-L., McCay, P. B., Poyer, J. L., Keele, B. B., and Misra, H. (1973) J. BWl. Chem. 248, 7792-7797 10. Tien, M., Svingen, B. A., and Aust, S. D. (1981) Fed. Proc. 40, 179-182 11. Svingen, B. A., O'Neal, F. O., and Aust, S. D. (1978) Photochem. Photobid. 28,803-809 12. Aust, S. D., and Svingen, B. A. (1982) in Free Radicals in Bwbgy (Pryor, W. A., ed) Vol. 5, pp. 1-28, Academic Press, Orlando, FL

1104 Importance of the Ferric:Ferrous Ratio in Lipid Peroxidation 13. Gutteridge, J. M.C., Richmond, R., and Halliwell, B. (1979) 27. Beers, R.F., Jr., and Sizer, I. W. (1952) J. Biol. Chem. 191,133- Biochem. J. 184, 469-472 140 14. McCord, J. M., and Day, E. D. (1978) FEBS Lett. 86,139-142 28. Pederson, T. C., and Aust, S. D. (1970) Biochem. Phurmacol. 19, 15. Fenton, H. J. H. (1894) J. Chem. SOC. 65,899-910 2221-2230 16. Pederson, T. c., Buege, J. A., and Aust, s. D. (1973) J. Bioi. 29. Folch, J., Lees, M., and Sloane Stanley, G.H. (1957) J. Biol. Chem. 248,7134-7141 Chem. 226,497-509 17. Svingen, B. A., Buege, J. A., O'Neal, F. o., and Aust, s. D. (1979) 30. Bartlett, G. R. (1959) J. Bioi. Chm. 234,466-468 J. Biol. Chem. 254,5892-5899 31. Buege, J. A., and Aust, S. D. (1978) Methods Enzymol. 62, 302-18. Morehouse, L. A., Tien, M., Bucher, J. R., and Aust, S. D. (1983) 310 Biochem. Pharmncol. 32,123-127 32. Kraljic, I., and Trumbore, N. C. (1965) J. Am. Chem. SOC. 87, 19. Anbar, M., and Neta, P. (1967) Znt. J. Appl. Radiat. Zsot. 18, 2547-2550 493-523 33. Tien, M., Morehouse, L. A., Bucher, J. R., and Aust, S. D. (1982) 20. Tyler, D. D. (1975) FEBS Lett. 51, 180-183 Arch. Biochem. Biophys. 218,450-458 21. Thomas, C. E., Morehouse, L. A.9 and Aust, S. D. (1985) J. Bid. 34. G ~ M. A., ~ and ~ ill, ~ A. H.,(1984) ~ ~ ~ ~ ~ 105,3- b ~ d ~ Chem. 260,3275-3280 22. Bors, W., Michel, C., and Saran, M. (1979) Eur. J. Biochem. 95, 22 621-627 35. Edwards, J. C., and Quinn, P. J. (1982) J. Lipid Res. 25, 944-23. Koppenol, W. H., and Liebman, J. F. (1984) J. Phys. Chem. 88, 1000 36. 99-101 Girotti, W., A. and Thomas, J. P. (1984) J. Biol. Chem. 259, 24. Hochstein, P., Nordenbrand, K., and Ernster, L. (1964) Biochem. 1744-1752 Biophys. Res. Commun. 14, 323-328 37. Girotti, A. W., and Thomas, P. J. (1984) Biochem. Biophys. Res. 25. Bucher, J. R., Tien, M., and Aust, S. D. (1983) Biochem. Biophys. Commun. 118,474-480 Res. Commun. 111,777-784 38. Walling, C. (1975) Accts. Chem. Res. 8, 125-131 26. Floyd, R. A., and Wiseman, B. B. (1979) Biochim. Biophys. Acta 39. Graf, E., Mahoney, J. R., Bryant, R. G., and Eaton, J. W. (1984) 586,196-207 J. Biol. Chem. 259,3620-3624