The involvement of iron in lipid peroxidation. Importance of ferric to ferrous ratios in initiation.

Intense lipid peroxidation of brain synaptosomes initiated with Fenton's reagent (H2O2 + Fe2+) began instantly upon addition of Fe2+ and preceded detectable OH. formation. Although mannitol or Tris partially blocked peroxidation, concentrations required were 10(3)-fold in excess of OH. actually formed, and inhibition by Tris was pH dependent. Lipid peroxidation also was initiated by either Fe2+ or Fe3+ alone, although significant lag phases (minutes) and slowed reaction rates were observed. Lag phases were dramatically reduced or nearly eliminated, and reaction rates were increased by a combination of Fe3+ and Fe2+. In this instance, lipid peroxidation initiated by optimal concentrations of H2O2 and Fe2+ could be mimicked or even surpassed by providing optimal ratios of Fe3+ to Fe2+. Peroxidation observed with Fe3+ alone was dependent upon trace amounts of contaminating Fe2+ in Fe3+ preparations. Optimal ratios of Fe3+:Fe2+ for the rapid initiation of lipid peroxidation were on order of 1:1 to 7:1. No OH. formation could be detected with this system. Although low concentrations of H2O2 or ascorbate increased lipid peroxidation by Fe2+ or Fe3+, respectively, high concentrations of H2O2 or ascorbate (in excess of iron) inhibited lipid peroxidation due to oxidative or reductive maintenance of iron exclusively in Fe2+ or Fe3+ form. Stimulation of lipid peroxidation by low concentrations of H2O2 or ascorbate was due to the oxidative or reductive creation of Fe3+:Fe2+ ratios. The data suggest that the absolute ratio of Fe3+ to Fe2+ was the primary determining factor for the initiation of lipid peroxidation reactions.

Intense lipid peroxidation of brain synaptosomes initiated with Fenton's reagent (H202 + Fez+) began instantly upon addition of Fez+ and preceded detectable OH' formation. Although mannitol or Tris partially blocked peroxidation, concentrations required were 103-fold in excess of OH' actually formed, and inhibition by Tris was pH dependent. Lipid peroxidation also was initiated by either Fez+ or Fe3+ alone, although significant lag phases (minutes) and slowed reaction rates were observed. Lag phases were dramatically reduced or nearly eliminated, and reaction rates were increased by a combination of Fe3+ and Fez+. In this instance, lipid peroxidation initiated by optimal concentrations of HzOz and Fez+ could be mimicked or even surpassed by providing optimal ratios of Fe3+ to Fez+. Peroxidation observed with Fe3+ alone was dependent upon trace amounts of contaminating Fez+ in Fe3+ preparations. Optimal ratios of Fe3+:Fez+ for the rapid initiation of lipid peroxidation were on order of 1:l to 7:l. No OH' formation could be detected with this system.
Although low concentrations of HzOZ or ascorbate increased lipid peroxidation by Fez+ or Fe3+, respectively, high concentrations of HzOz or ascorbate (in excess of iron) inhibited lipid peroxidation due to oxidative or reductive maintenance of iron exclusively in Fez+ or Fe3+ form. Stimulation of lipid peroxidation by low concentrations of HzOz or ascorbate was due to the oxidative or reductive creation of Fe3+:Fez+ ratios. The data suggest that the absolute ratio of Fe3+ to Fez+ was the primary determining factor for the initiation of lipid peroxidation reactions.
The role of iron in initiating lipid peroxidation reactions within biological membranes has been examined in a variety of systems. Ferric iron (Fe3+) or ferric chelates can initiate lipid peroxidation reactions provided that a reducing agent is present to reduce Fe3+ to Fez+. In many cases, superoxide (0,) generated by the reaction of xanthine or hypoxanthine with xanthine oxidase may act as the reductant (Tien et al., 1982). However, other cellular reducing agents such as ascorbate may also reduce Fe3+ via a presumably 0;-independent mechanism (Bucher et al., 1983a). In the former case, initiation of lipid peroxidation both in vitro and in vivo by Fe3+ and O2 has been hypothesized to occur via production of OH' arising through the iron-catalyzed Haber-Weiss reaction * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom all correspondence should be sent. (Halliwell, 1978) Fe3+ + 0; --* Fez+ + O2 (1) 0; + H02 -H202 + O2 (2) Hz02 + Fez+ + OH' + OH-+ Fe3+ (3) Certain iron chelates, particularly those with ADP, are known to accelerate the Haber-Weiss reaction and the formation of OH.. Although some investigators have provided evidence for the role of Haber-Weiss-generated OH' in lipid peroxidation reactions (Lai and Piette, 1977), others question its involvement (Tien et al., 1982). Ferrous iron (Fez+) may also initiate lipid peroxidation reactions provided an oxidant is present to oxidize Fez+ to Fe3+. In most cases, the oxidant is either molecular Oz, The combination of Fez+ and H202, known as Fenton's reagent, is often used to initiate lipid peroxidation reactions. While Fenton's reagent yields large amounts of the very reactive OH', the role of OH' in initiating lipid peroxidation with Fenton's reagent has been questioned (Koppenol and Liebman, 1984;Bors et al., 1979) in that an iron-oxygen complex (Fe4+0 or FeOH3+) derived from OH ' and Fe3+ or Fez+ and H202 could be the oxidizing species. Alternatively, the autoxidation of Fez+ by O2 which is accelerated by iron chelation and can result in OH' formation (via Reactions 1-3) has been proposed by Weiss (1953) to result in the formation of a reactive iron-oxygen complex Fez+ + O2 + Fe3+ . . . 0; (5) which is stabilized in the presence of certain ligands such as phosphate or chelators (Michelson, 1977).
Regardless of the radical generating system used or the nature of the lipid substrate, a consistent observation in studies involving iron has been that, without exception, oxidation of Fez+ or reduction of Fe3+ is necessary for the initiation of lipid peroxidation. Although conflicting data on the involvement of OH ' or 0; in lipid peroxidation is widespread, Fez+ c, Fe3+ interconversion is obligatorily required. Bucher et al. (1983b) originally reported on the requirement of Fe3+ chelates in the initiation of lipid peroxidation by Fez+ chelates. Their wcrk suggested that ratios of Fe3+ chelate:Fe2+ chelate were involved in the initiation of lipid peroxidation reactions, perhaps through the formation of an Fe3+-dioxygen-Fe2+ chelate complex. The importance of iron ratios in supporting lipid peroxidation also has been alluded to in a report by Ernster et al. 1982), although the concept was not tested directly. We have conducted a series of studies concerning the involvement of nonchelated iron in peroxidation of brain synaptosomal lipids. The results of this work extend the findings of Bucher et al. (1983b) and are in excellent agreement with their observations. Our findings support their original hypothesis that optimal Fe3+:Fez+ ratios are required for the initiation of lipid peroxidation reactions. In this regard, various free radical generating systems used to initiate lipid peroxidation serve chiefly to create an Fe3+:Fe2+ ratio within the system. This hypothesis effectively explains observed paradoxical differences in the function of oxidizing or reducing agents in iron-dependent lipid peroxidation.

MATERIALS AND METHODS
Rat brain synaptosomes were freshly prepared as described (Braughler, 1985), except that synaptosomes were washed and suspended in 0.9% NaC1, pH 7.0.
All incubations were carried out at 37 "C in 0.9% NaCl. Since most buffers can trap OH' radicals or interfere with iron conversions (as described in Fig. 5), reactions were unbuffered, and the pH of 0.9% NaCl used in these experiments was carefully adjusted to 7.0 immediately prior to use. Fez+ solutions were prepared as either Fe(NH4),(SO& or FeCl, in degassed H20 purged with either argon or nitrogen. Fe3+ was prepared as FeC13 in H,O. All iron solutions were prepared fresh and used immediately. Under these circumstances, precipitation of Fe3+ as Fe(OH)3 was not a problem.
Lipid peroxy radical formation in brain synaptosomal membranes was followed by monitoring the rate of 0, consumption in incubations using a Clarke-style O2 electrode. Incubations containing 0.5 mg of synaptosomal protein/ml were continuously stirred at 37°C in a 600-p1 sealed chamber in which the 0, electrode was immersed via a sealed port. Additions to the incubation were made via a resealable port on top of the chamber. The O2 amplifier was interfaced with an Apple IIE computer, and the 0 2 content of the chamber and the using "Micro 2" 0, analysis software developed by Instech Labora-instantaneous 0, consumption rate were continuously monitored tories, Horsham, PA. Overall 0 2 consumption rates for various reactions were calculated through a point-by-point analysis. Conjugated diene formation in detergent-dispensed synaptosomes (0.01 mg of synaptosomal protein/ml of 0.9% NaCl containing 1% Lubrol) was monitored continuously at 232 nm using a Gilford response spectrophotometer. Reference samples contained synaptosomes and all other reactants except iron. Conjugated diene formation was corrected for absorbance changes in 232 nm arising from Fez+ to Fe3+ conversions (Graf et al., 1984) by subtracting the absorbance of incubations containing all reactants except synaptosomes.
The formation of thiobarbituric acid-reactive oxidation products (TBAR') during incubations was determined as described by Buege and Aust (1978) with some modification. One hundred-pl reactions containing 0.1 mg of synaptosomal protein/ml were stopped by the addition of 500 pl of ice-cold 0.8 N HC1 containing 12.5% trichloroacetic acid. Four hundred pl of H20 was then added containing 50 p~ desferrioxamine to prevent further iron-catalyzed TBAR formation. Thiobarbituric acid (final concentration 0.67%) was added, and samples were boiled in the presence or absence of 0.05% butylated hydroxytoluene (BHT) for 20 min. Following boiling, samples were cooled, centrifugd at 1500 X g for 15 min, and the absorbance of the supernatant read at 532 nm. Quantitation was based upon a molar extinction coefficient of 1.56 X 10' .
Lipid hydroperoxide was assayed by the iodometric procedure described by Buege and Aust (1978). All reagents were prepared fresh -daily, degassed, and purged with argon. One-ml reactions containing 0.1 mg of synaptosomal protein/ml were stopped with the addition of 5 m of ice-cold ch1oroform:methanol (2:l). Following centrifugation, 3 ml of the organic phase was removed and taken to dryness in a clean test tube under argon at 45 'C. While still under argon at 25 "C, 1 ml of acidchloroform (3:2) was added, followed by 50 pl of 0.18 mM KI in H20. The tube was stoppered and incubated in the dark at 25 "C for exactly 5 min, after which 3 ml of 0.5% Cd(CH3C0& was added and the sample was centrifuged at 1500 X g for 10 min. The coefficient of 2.8 X lo4 was used for quantitation. absorbance of the upper layer was read at 353 nm. A molar extinction OH' formation was estimated by assaying for the hydroxylated ' The abbreviations used are: TBAR, thiobarbituric acid-reactive oxidation products; BHT, butylated hydroxytoluene. products of salicylate in the absence of added synaptosomes as described by Halliwell(l978). Two-ml incubations containing 2.0 mM salicylate and no synaptosomes were stopped by the addition of 80 pl of 12 M HCl, 0.5 g of NaCl, and 4 ml of ice-cold diethyl ether. The sample was mixed, and the ether layer was removed and evaporated to dryness. The residue was dissolved in 0.25 ml of cold distilled HzO, and the following additions were made in order: 0.125 ml of 10% trichloroacetic acid dissolved in 0.5 M HCl; 0.25 ml of 10% sodium tungstate; and 0.25 ml of 0.5% sodium nitrite (prepared fresh). After 5 min at 25 "C, 0.5 ml of 0.5 M KOH was added and the absorbance was read at 510 nm. A molar extinction coefficient of 3.26 X lo3 derived from solutions of 2,3-dihydroxybenzoate carried through the same extraction and assay procedure was used for quantitation.
H202 was assayed by the oxidation of phenol red. One hundred-pl reactions were stopped by the addition of 800 pl of phenol red solution containing 0.56 mM phenol red, horseradish peroxidase (19 units/ ml), 5.5 mM dextrose, 140 mM NaCl, and 10 mM phosphate buffer, pH 7.4. The sample was incubated for 10 min at 25 "C, and 100 pl of 1 N NaOH was added. The absorbance was read at 610 nm. A standard curve was based upon known amounts of H,O,.
O2 consumption curve data were from representative experiments, and 0, consumption rates were based upon mean rate calculations from data accumulation rates of 1 point/s. Other data in figures and tables are the means of triplicate determinations from representative experiments.

RESULTS
The addition of 200 PM Fez+ to an incubation of rat brain synaptosomes containing 100 FM H,O, resulted in the formation of OH', conjugated dienes, and lipid hydroperoxide, as well as a rapid consumption of 0, (Fig. 1). Careful analysis of the reaction time course indicates a burst in 0, consumption as well as conjugated diene and lipid hydroperoxide formation within the first 5 s after Fez+ addition that was associated with minimal detected OH' formation. Despite continuous 0, consumption and conjugated diene formation, lipid hydroperoxide concentrations fell acutely between 5 s and 1 min. The acute fall in lipid hydroperoxide was presum-

FerricFerrous Ratios in Lipid Peroxidation
ably due to iron-catalyzed degradation of lipid hydroperoxide to either alkoxy or peroxy radicals. This would be expected to result in lipid peroxidation chain propagation and branching reactions initiated by alkoxy and peroxy radicals. Since conjugated diene formation continued to increase between 5 s and 1 min before reaching a steady state at 1 min, chain branching was likely. OH ' formation increased sharply during this time period. After 1 min, lipid hydroperoxide formation again increased in the presence of a continued O2 consumption. Although the O2 concentration was near zero by 4 min, some lipid hydroperoxide formation continued despite the presence of only trace amounts of O2 dissolved in the reaction media.
The formation of TBAR in brain synaptosomes was also examined in incubations containing Fez+ and H202 (Table I).
TBAR actually formed during Fez+ + H202-induced peroxidation at 37 "C was measured by including 0.05% BHT during sample boiling to prevent further formation of TBAR presumably from the breakdown of lipid hydroperoxide (Buege and Aust, 1978). Alternatively, boiling samples in the absence of BHT yielded large amounts of TBAR that represented both TBAR formed during the original incubation as well as TBAR produced during the boiling step. Thus, the difference between TBAR assays boiled in the absence and presence of BHT yielded information concerning the production of potential TBAR-forming material during an incubation. TBAR-forming material (-BHT -(+BHT) values) is presumed to represent products of lipid peroxidation that do not produce TBAR under standard incubation conditions at 37 "C but do decompose to TBAR under conditions of the thiobarbituric acid assay (100 "C at acidic pH). As can be seen in Table I, within 5 s after the addition of Fez+ to an incubation of brain synaptosomes containing H202 little TBAR was formed in the sample itself; however, a significant amount of TBARforming material was produced. After 10 min at 37 "C, there was an increase in both TBAR formed during the incubation as well as in the production of TBAR-forming material. No detectable TBAR-forming material or lipid hydroperoxide was present in synaptosomes not exposed to Fez+ (not shown).
Contrasting observations to those observed with Fez+ and H202 were obtained using either Fez+ or Fe3+ alone. The overall peroxidation rate induced by 200 PM Fez+ was quite slow during the first 30 min as evidenced by a slow consump-  tion of O2 and slow lipid hydroperoxide and conjugated diene formation (Fig. 2). Nevertheless, during the first 5 s following the addition of Fez+, there was a small burst in conjugated diene and lipid hydroperoxide formation. Detectable OH' formation was minimal throughout the first 30 min. At approximately 30 min, the rate of O2 consumption increased sharply along with a modest increase in OH' formation. Quite different results were obtained when peroxidation was induced with Fe3+. Following a short lag phase after the addition of 200 I.IM Fe3+, O2 consumption accelerated briefly, but rapidly slowed (Fig. 3). This was in contrast to peroxidation induced by either Fez+ alone or Fez+ and H202 where O2 consumption continued toward completion once the rapid O2 consumption phase began.
TBAR formation during the first 10 min of an incubation with Fez+ was minimal. However, significant amounts of TBAR-forming material were produced (Table I). By 30 and 60 min after the addition of Fez+, TBAR formation had increased markedly. The increase in TBAR observed at 30 min with Fez+ coincided roughly with an increase in the Oz consumption rate seen at around 30 min (Fig. 2). The rate of TBAR formation and production of TBAR-forming material caused by Fe3+ was faster during the first 10 min compared with that observed with Fez+. TBAR with Fe3+ did not increase greatly after 10 min. Based on the results in Fig. 1 with Fez+ and HZO2, the involvement of OH' in peroxidation of brain synaptosomes induced by Fenton's reagent must be questioned. An analysis of rates of H202 reduction by Fez+ and the formation of OH' during an incubation in the absence of synaptosomes (Fig. 4) reveals marked differences in their kinetics. Greater than 90% of the H,Oz was consumed within 5 s following the addition of Fez+, while OH' production during this time was minimal. Under these circumstances either: 1) OH' formation is not stoichiometric with H2OZ reduction, or 2) the high reactivity and short tlh of OH' prohibited its reaction with the trap used for its detection (salicylate); thus only a fraction of OH' formed was detected. In view of this second possibility, it is difficult to perceive how OH' formed through Fenton's reaction could attack an allylic hydrogen buried within the phospholipid structure unless it was generated at that site.
Further studies into the involvement of OH' in lipid peroxidation induced with Fenton's reagent are shown in Table  I1 where the OH ' traps, Tris and mannitol, were included in reactions with Fez+ and HzOz. Mannitol at a concentration 103-fold that of OH' actually detected caused only a modest reduction in H2O2/Fe2+-induced Oz consumption and early conjugated diene formation. Tris has been reported to be an excellent OH' (Tien et al., 1982), and at pH 7.4, 10 mM Tris caused a large reduction in both 0, consumption and conjugated diene formation. At pH 6.5, however, Tris was far less effective. Small effects of each trap were observed on TBAR formation during the 10-min assay. When Fez+ is oxidized either by autoxidation or by Hz02 to Fe3+, the Fe3+ species formed absorbs light strongly in the UV range, and the absolute absorbance of the UV spectrum is indicative of the concentration of Fe3+ present in solution (Graf et al., 1984). Some insight into the actions of mannitol and Tris on Fez/-HzOz-induced peroxidation can be gained by examining their effects on the UV iron spectrum (Fig. 5 ) . As shown in Fig. 5A, the absorbance of a reaction containing 200 p~ Fez+ and 100 p~ HZO2 increased rapidly with time. The initial absorbance scan of 200 pM Fez+ and 100 p M HZOZ at 5 s after the addition of Fez+ was identical to an initial scan of a combination of 100 pM Fez+ and 100 p M Fe3+ (See Fig.  7A), indicating that the reaction of Fez+ with H202 was very rapid and essentially complete within 5 s, in agreement with data shown in Fig. 4. With time, the absorbance of a solution of Fez+ and HzOz continues to increase rapidly, and by 10 min the spectrum is similar to that of a solution containing almost exclusively Fe3+. Mannitol, which caused only a slight reduction in the peroxidation rate, only very slightly slowed the formation of Fe3+ (Fig. 5B). In contrast, 10 mM Tris at pH 7.4, which inhibited peroxidation, caused a striking reduction in Fe3+ formation with time and altered the shape of the UV spectrum as well (Fig. 5C). Interestingly, the initial 5-s spectra with Tris, pH 7.4, suggested that rapid oxidation of Fez+ by HzOz to an Fe3+ species had occurred. Quite different results were obtained with 10 mM Tris at pH 6.5, as Fez+ conversion was not inhibited, but its rate of formation was reduced (Fig.  50). These findings are consistent with the effects of mannitol and Tris on Fe2+/H2OZ-induced peroxidation of brain synaptosomes shown in Table 11.
Taken together, these results suggest that it is perhaps not OH' formation which is important for the initiation of lipid peroxidation by Fez+ and HzOz, but rather it is the conversion of Fez+ to Fe3+ that is in some way responsible. As demonstrated in Fig. 2 and Table I, peroxidation of brain synaptosomes did occur in the presence of Fez+ alone. However, the rate was much slower than for the combination of Fez+ and H,O,. With Fez+ alone, a considerable lag phase was observed before Oz consumption (peroxy radical formation) achieved a rapid rate.
As demonstrated in Fig. 6, the lag phase associated with Fez+-induced peroxidation could be dramatically reduced by adding a combination of Fe3+ and Fez+. Higher Fe3+:Fez+ ratios were associated and shorter lag phases and more rapid rates of lipid peroxidation as assessed by more rapid 0, consump-

TABLE I1
Effects of mannitol and Tris on Oz consumption, conjugated diene, and TBAR formation with Fez+ and H20z Incubations contained 100 gM HzOZ, 0.1 mg of synaptosomal protein/ml (0, consumption and TBAR), or 0.01 mg of synaptosomal protein/ml (conjugated diene; A23Znm), and the concentrations of Tris or mannitol indicated. Reactions were initiated by the rapid addition of 200 p M Fez+. TBAR values were for 10-min incubations; conjugated dienes were determined at 5 s and 10 min; 0 2 consumption rates were based upon overall 0-4min rates.

+BHT
-BHT -BHT -(+BHT)   tion. Similar observations were obtained for conjugated diene formation (not shown). Optimal ratios of Fe3+:Fe2+ were around 1:l to 7:l. As a rule, excess Fe3+ in the presence of some Fez+ resulted in a rapid rate of 0, consumption. The longer lag phases associated with ratios of Fe3+:Fe2+ below 1 were apparently related to the time required for sufficient Fez+ to autoxidize to Fe3+, creating a more suitable Fe3+:Fez+ ratio for peroxidation.
The largest amount of TBAR was formed when the Fe3+:Fe2+ ratios were on the order of 3:l to 7:l ( Table 111).
The production of TBAR-forming material was not affected as much by changes in the Fe3+:Fe2+ ratio.
OH' formation could not be detected within the limits of

TABLE IV Effects of Tris and mannitol on O2 consumption by brain synaptosomal membranes with different Fe3+:Fe2+ combinations
Overall O2 consumption rates were determined as described under "Materials and Methods." Incubations contained 0.1 mg of synaptosomal protein/ml and the concentrations of Fe3+, Fez+, Tris, mannitol, superoxide dismutase, or catalase indicated. Reactions were initiated by the rapid addition of iron.  IV). Neither Tris, pH 6.5, nor mannitol affected peroxidation caused by combinations of Fe3+ and Fez+. Peroxidation by Fe3+:Fez+ combinations was also not affected by superoxide dismutase, catalase, or their combination. These observations suggested that the absolute ratio of Fe3+:Fe2+ was more important than OH' in determining the rate or extent of lipid peroxidation induced by Fenton's reagent (FeZ+/HzO,). This hypothesis would predict that the primary function of HzOz in such a system would be to create a pool of Fe3+ from Fez+. Furthermore, this hypothesis also predicts that generation of a small amount of Fez+ from Fe3+ should accelerate Fe3+-induced peroxidation and that trapping or oxidative-reductive maintenance of either Fe3+ or Fez+ in their respective forms should inhibit lipid peroxidation. Such was the case. As demonstrated in Table V, Oz consumption in the presence of Fe3+ was markedly accelerated in the presence of 25 ~L M ascorbate. Ascorbate reduces Fe3+ to Fez+, and low concentrations have been shown by others to stimulate iron-dependent lipid peroxidation (Bucher et al., 1983a).

TABLE V Effect of ascorbate and H202 on Oz consumption by brain synaptosomal membranes with different Fe3+:Fez+ combinations
Overall 0, consumption rates were determined as described under "Materials and Methods." Incubations contained 0.1 mg of synaptosomal protein/ml and the concentrations of Fe3+, Fez+, ascorbate, or H202 indicated. Reactions were initiated by the rapid addition of iron. Total per cent 0, consumed was determined when 0, consumption had dateaued for 5 min. The rate of O2 consumption in the presence of 200 pM Fe3+ and 25 p~ ascorbate was somewhat faster than for a combination of 175 pM Fe3+ and 25 pM Fe2+. High concentrations of ascorbate were found to inhibit lipid peroxidation, probably by direct antioxidant properties (Bucher et al., 1983a), but also by maintaining iron in its reduced form. Similarly, H202 could either stimulate or inhibit Fez+-dependent lipid peroxidation. Indeed, O2 consumption rates in the presence of 100 ~L M H,O, and 200 p~ Fez+ were no different than in the presence of the combination of 100 pM Fe2+ and 100 pM Fe3+. H20z in excess of Fez+, however, which would be expected to affect complete conversion of Fez+ to Fe3+, inhibited peroxidation. Such an inhibitory effect of H202 has been reported by others Tien et al., 1982).
Low concentrations of ascorbate also increased TBAR formation in brain synaptosomes by Fe3+, whereas higher concentrations were inhibitory ( Table VI). As with 0 2 consumption, the combination of Fe3+ with a low concentration of ascorbate caused slightly more peroxidation as assessed by TBAR formation than did the iron-equivalent combination of Fe3+ and Fez+. Considerably more TBAR-forming material was formed in the absence of ascorbate, however. TBAR formation with the combination of 175 p~ Fe3+ and 25 V M Fez+ was inhibited by 100 p~ H202, probably due to complete conversion of Fez+ to Fe3+. In fact, TBAR formation in the presence of the combination of 175 FM Fe3+, 25 pM Fez+, and 100 p~ H202 was similar to that observed with 200 p~ Fe3+ alone.
The O2 consumption and TBAR formation observed with Fe3+ alone (Fig. 3, Tables I, V, and VI) resulted from contamination of the Fe3+ preparation with small amounts of Fez+, thus providing a ratio of Fe3+:Fe2+. As demonstrated in Tables V and VI, preincubation of Fe3+ preparations for 5 min with Hz02 or excess EDTA to assure complete Fez+ oxidation prior to addition of synaptosomal membranes rendered Fe3+ incapable of initiating peroxidation. Furthermore, preincubation of Fe3+ with EDTA was associated with a small burst in O2 consumption consistent with autoxidation of contaminating Fez+. Finally, freshly prepared solutions of FeC13 were found to contain about 1.0% Fez+ using 2,2',2"-tripyridine as an indicator of ferrous iron (Schade et al., 1954). These findings

Effects of ascorbate and Hz02 on TBAR formation in brain synaptosomes
The production of TBAR (+BHT) and TBAR-forming material (-BHT -(+BHT)) in brain synaptosomes was determined as described under "Materials and Methods." Incubations contained 0.1 mg of synaptosomal protein/ml and the concentrations of Fe3+, Fez+, ascorbate, or HzOz indicated. Reactions were initiated by the rapid addition of iron and were terminated after 10 min. confirm that freshly prepared FeC13 solutions contained some iron in the Fez+ form. TBAR formation by Fez+ during the incubation was stimulated by both high and low concentrations of H202 (Table  VI) were allowed to react prior to the introduction of synaptosomes, more TBAR was formed in a reaction that had contained the low concentration of H202 compared with the higher concentration. Thus, under certain circumstances, excess H202 could directly inhibit lipid peroxidation reactions. H202 at both high or low concentrations inhibited TBAR formation by Fe3+.
These results are consistent with the idea that the ratio of Fe3+:Fe2+ is the important determinant for lipid peroxidation reactions. The observations with Fez+ and H,02 also suggest that the rapidity with which conversions of Fez+ to Fe3+ take place plays a role in lipid peroxidation. An examination of iron spectral changes reveals that although H202 accelerates the early conversion of Fez+ to Fe3+, it reduces subsequent Fe2+ to Fe3+ conversions (Fig. 7, A and B). Similarly, while Fe3+ displays some spectral changes with time, indicative of Fe3+ to Fez+ to Fe3+ interconversions, Hz02 greatly reduced these. Addition of small amounts of ascorbate to Fe3+ to create a pool of Fez+, on the other hand, allows for more extensive Fez+ to Fe3+ conversion associated with more extensive lipid peroxidation. Such changes in the rates of Fez+ to Fe3+ conversions can have profound effects on lipid peroxidation as evidenced by the data in Tables V and VI.

DISCUSSION
Lipid peroxidation of brain synaptosomes induced with Fenton's reagent (Fez+ and H202) was associated with an instantaneous (within 5 s) and intense production of conjugated dienes, lipid hydroperoxide, and TBAR-forming material upon the addition of Fez+. The burst in lipid hydroperoxide and conjugated diene formation was coincident with a rapid loss of Oz and HzOZ from the media. In parallel incubations containing no synaptosomes, little or no OH ' was detected during the first seconds following the addition of Fez+, despite the consumption of 90% of the HZOz. Thus, the kinetics of OH ' formation by Fenton's reagent did not follow the kinetics of lipid peroxidation.
The direct involvement of OH ' in the initiation of lipid peroxidation was further challenged by results obtained with OH ' traps. Both Tris and, to a lesser extent, mannitol were found to interact with iron in a manner which slowed or blocked the formation of Fe3+. These findings suggested that under conditions employed in these studies, OH. traps were not affecting the interaction of OH' with lipids by scavenging OH., per se, but rather were interfering with iron in some manner.
Peroxidation of brain synaptosomes could be induced by Fez+ alone, although peroxidation was preceded by a significant lag phase. This is in agreement with studies by Bucher et al., (1983b) who demonstrated that Fez+ chelates could initiate lipid peroxidation following a short lag phase. Lag phases with Fez+ chelates (Bucher et al., 1983b) were considerably shorter (5-6 min) than those reported here for free Fez+ (30 min). The lag phase associated with Fe*+-induced peroxidation was significantly shortened by adding a combination of Fe3+ and Fez+ t o brain synaptosomal incubations. Again, similar observations have been reported by Bucher et al., (1983b) using iron chelates. In the present study, even in the absence of chelators, lipid peroxidation rates could be achieved that were similar or greater than those obtained with Fenton's reagent (Fez+ + HzOz) simply by using ratios of Fe3+ to Fe2+. Although lag phases may be explained by the presence of endogenous antioxidants in membrane preparations this is not likely since lag periods in this study corresponded primarily to the time required for autoxidation of Fez+ to yield a suitable ratio of Fe3+:Fe2+.
The findings with ascorbate and H202 were consistent with the hypothesis that Fe3+:Fe2+ ratios are important determinants for lipid peroxidation. Reduction of small amounts of Fe3+ by low concentrations of ascorbate markedly enhanced peroxidation by Fe3+, while oxidation of some Fez+ by HzOZ enhanced peroxidation. On the other hand, high concentrations of ascorbate or HzOz in excess of Fe3+ or Fez+, respectively, inhibited lipid peroxidation. In view of the fact that significant lipid peroxidation could be initiated in vitro by the addition of various molar ratios of Fe3+:Fez+, it is apparent that a major portion of the peroxidation induced in systems utilizing either Fenton's reagent or Fe3+ reduction may be directly attributed to the oxidative or reductive generation of an Fe3+:Fe2+ ratio. No detectable OH' was formed by various combinations of Fe3+ and Fez+ nor was peroxidation caused by Fe3+:Fe2+ combinations inhibited by superoxide dismutase, catalase, their combination, or mannitol. These data argue against the involvement of OH' in Fe3+:Fez+ initiation. Nevertheless, it is possible that OH' generated within the phospholipid environment of the membrane or reacting rapidly with Fe3+ to form some ferry1 ion species (Sheldon and Kochi, 1981;Aust et al., 1985;Halliwell and Gutteridge, 1984) could escape traps and participate in Fe3+:Fe2+-induced lipid peroxidation. Thomas (1984a, 1984b) have in fact proposed that it is OH' formed at or on the membrane and not in the media that is responsible for the initiation of lipid peroxidation by iron. Their conclusion is based upon studies demonstrating that while OH ' traps were unable to inhibit lipid peroxidation in a membrane system containing xanthine, xanthine oxidase, and Fe3+, peroxidation was inhibited by iron chelators, catalase, and superoxide dismutase. These findings are not in disagreement with those reported here as they are consistent with chelation of iron or enzymatic elimination of oxidant (H2OZ) or reductant (0;) preventing formation of an iron ratio. Our findings that catalase and superoxide dismutase do not block peroxidation with an Fe3+:Fez+ ratio support this contention. It is unlikely that the initiation of lipid peroxidation by combinations of Fe3+ and Fez+ may be through iron-catalyzed decomposition of lipid hydroperoxide pre-existing in synaptosomal membranes. In that regard, free transition metals are relatively weak compared with chelated metals at decomposing lipid hydroperoxide, and Fez+ is at least an order of magnitude more effective than Fe3+ (O'Brien, 1969;Halliwell and Gutteridge, 1984). Assays of freshly prepared brain synaptosomes used in the present studies revealed no detectable conjugated dienes, lipid hydroperoxide, or TBAR-forming material within the detection limits of the assays. Although this does not rule out the presence of pre-existing lipid hydroperoxide in the synaptosomal membranes used in these studies, it is unlikely that lipid peroxidation initiated by Fe3+:Fe2+ ratios can be attributed to iron-catalyzed lipid hydroperoxide decomposition since reaction rates and lag periods were slower when Fez+ was in predominance than when Fe3+ was in excess. An absence of preexisting lipid hydroperoxide was also supported by studies in which Fe3+ was preincubated with excess EDTA prior to addition of synaptosomes to eliminate contam-inating Fez+. In those studies, Fe3+-EDTA chelate (which did not contain Fez+) was incapable of initiating lipid peroxidation.
In summary, the rapid and extensive peroxidation of rat brain synaptosomal membranes could be initiated upon the addition of various combinations of Fe3+ and Fez+ without need for added chelators, oxidizing or reducing agents. In other preliminary unpublished studies, similar results have been obtained using either phosphotidylcholine liposomes or lipid micelles prepared from purified linoleic, linolenic, or arachidonic acids as the lipid substrate. Thus, it is unlikely that the results presented here are artifacts due to the presence of endogenous metal chelators, oxidizing or reducing substances provided by the synaptosomal preparation. Peroxidation initiated by Fe3+:Fe2+ ratios did not rely upon detectable production of OH' radicals, although OH' radical involvement cannot be completely ruled out at this time. The findings suggest that lipid peroxidation initiated by combinations of iron with oxidizing or reducing agents such as Fenton's reagent or Fe3+ and ascorbate may largely result from the formation of optimal Fe3+:Fe2+ ratios.