Superoxide-dependent Oxidation of Extracellular Reducing Agents by Isolated Neutrophils*

Incubation of stimulated neutrophils with sulfhydryl (RSH) compounds or ascorbic acid (ascorbate) results in rapid superoxide (0;)-dependent oxidation of these reducing agents. Oxidation of RSH compounds to di-sulfides (RSSR) is faster than the rate of 0; production by the neutrophil NADPH-oxidase, whereas about one ascorbate is oxidized per 0;. Ascorbate is oxidized to dehydroascorbate, which is also oxidized but at a slower rate. Oxidation is accompanied by a large increase in oxygen (02) uptake that is blocked by superoxide dismutase. Lactoferrin does not inhibit, indicating that ferric (Fe3+) ions are not required, and Fe3+-lactoferrin does not catalyze RSH or ascorbate oxidation. Two mechanisms contribute to oxidation: 1) 0; oxidizes ascorbate or reduced glutathione and is reduced to hydrogen peroxide (H202), which also oxidizes the reductants. 0; reacts directly with ascorbate, but reduced glutathione oxidation is mediated by the reaction of 0; with manganese (Mn2+). The H202-depend- ent portion of oxidation is mediated by myeloperoxi-dase-catalyzed oxidation of chloride to hypochlorous acid (HOC1) and oxidation of the reductants by HOC1. initiates Mn2+-dependent auto-oxidation

Incubation of stimulated neutrophils with sulfhydryl (RSH) compounds or ascorbic acid (ascorbate) results in rapid superoxide (0;)-dependent oxidation of these reducing agents. Oxidation of RSH compounds to disulfides (RSSR) is faster than the rate of 0; production by the neutrophil NADPH-oxidase, whereas about one ascorbate is oxidized per 0;. Ascorbate is oxidized to dehydroascorbate, which is also oxidized but at a slower rate. Oxidation is accompanied by a large increase in oxygen (02) uptake that is blocked by superoxide dismutase. Lactoferrin does not inhibit, indicating that ferric (Fe3+) ions are not required, and Fe3+lactoferrin does not catalyze RSH or ascorbate oxidation. Two mechanisms contribute to oxidation: 1) 0; oxidizes ascorbate or reduced glutathione and is reduced to hydrogen peroxide (H202), which also oxidizes the reductants. 0; reacts directly with ascorbate, but reduced glutathione oxidation is mediated by the reaction of 0; with manganese (Mn2+). The H202-dependent portion of oxidation is mediated by myeloperoxidase-catalyzed oxidation of chloride to hypochlorous acid (HOC1) and oxidation of the reductants by HOC1. 2) 0; initiates Mn2+-dependent auto-oxidation reactions in which RSH compounds are oxidized and O2 is reduced. Part of this oxidation is due to the RSHoxidase activity of myeloperoxidase. This activity is blocked by superoxide dismutase but does not require 0; production by the NADPH-oxidase, indicating that myeloperoxidase produces 0; when incubated with RSH compounds.
It is proposed that an important role for 0; in the cytotoxic activities of phagocytic leukocytes is to participate in oxidation of reducing agents in phagolysosomes and the extracellular medium. Elimination of these protective agents allows H202 and products of peroxidase/H202/halide systems to exert cytotoxic effects.
Stimulation of phagocytic leukocytes results in activation of an NADPH-oxidase enzyme in the plasma membrane (1)(2)(3)(4)(5) . The enzyme is oriented so as to accept electrons from NADPH in the cytosol, transfer one electron at a time to oxygen (Oz), and release superoxide free radicals (0;) into phagolysosomes or the extracellular medium. The oxidase is specific for O2 and does not catalyze the two-electron reduc-* This research was supported by Grants AI 16795, CA 09346, and DE 04235 from the National Institutes of Health and by the American Lebanese Syrian Associated Charities. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. tion of O2 to hydrogen peroxide (HzO,), so that the enzyme is committed to producing 0;.
Nevertheless, the role of 0; in leukocyte antimicrobial activity is unclear. 0; is short-lived, not highly reactive with most biological materials, and does not readily penetrate bacterial membranes (6)(7)(8). Moreover, most microorganisms contain superoxide dismutase enzymes that would prevent the reaction of 0; with intracellular components (9, 10). Therefore, 0; has little or no antimicrobial activity when microorganisms are incubated with 0,-generating systems. Two mechanisms have been described whereby 0; contributes to antimicrobial and other cytotoxic activities. First, 0, undergoes a rapid spontaneous dismutation reaction that yields H20z. The H202 may react directly with target cells or participate in reactions catalyzed by the hemoprotein peroxidase enzymes of leukocytes. Peroxidase/H,O,/halide systems amplify the toxicity of H20z by producing oxidizing and halogenating agents that are more reactive than H202 (11)(12)(13)(14)(15)(16)(17). In this mechanism, the only role for 0, is as an intermediate in H202 production.
In the second mechanism, some 0; ions dismutate to yield H20z and others act as reducing agents for metal ions. For example, 0; reduces ferric ion (Fe"') to ferrous ion (Fez+), which reacts with HzOz to yield iron-oxygen complexes (6). These complexes may react directly with target cells, release the highly reactive hydroxyl radical (HO), or mimic the activity of peroxidases by oxidizing the halide ion iodide (18)(19)(20)(21).
In this mechanism, 0; is an intermediate in both HzOs production and the formation of a catalyst for HzOz-dependent toxicity.
The hypothesis of this study is that 0; has an additional role, which is to participate in oxidation of substances in the extracellular phase, consisting of the intraphagolysosomal space and the extracellular medium. When a leukocyte encounters a microorganism in uiuo, both cells are bathed in reducing agents such as ascorbate and sulfhydryl (RSH) compounds. Plasma ascorbate is 50 @M in well-nourished individuals (22), and plasma contains 0.6 mM RSH groups, primarily reduced cysteine residues of plasma proteins (23). Reductants may also be found in high amounts at sites of inflammation and infection. For example, damaged cells release intracellular reductants such as ascorbate, GSH,' and NAD(P)H, and certain microorganisms create a reducing environment by reducing disulfide (RSSR) compounds and excreting the products (24).
RSH compounds and ascorbate are scavengers for the antimicrobial oxidizing agents produced by leukocytes. These reductants detoxify products of peroxidase/H2Oz/halide systems, although they react less rapidly with H,Oz. The ability 'The abbreviations and trivial names used are: GSH, reduced glutathione; cyt, c, cytochrome c; GSSG, glutathione disulfide; Nbs, 5-thio-2-nitrobenzoic acid or TNB; Nbs2, 5,5'-dithiobis(2-nitrobenzoic acid) or DTNB; PMA, phorbol 12-myristate 13-acetate. of leukocytes to use H,O, and oxidized halides as antimicrobial agents would be facilitated if leukocytes had a mechanism for eliminating protective reductants from the extracellular phase. Without such a mechanism, leukocytes would have to produce enough H20, or oxidized halides to oxidize both the reductants and the microbial cell components. 0; could participate in two types of reactions that would provide efficient mechanisms for eliminating extracellular reductants. First, free-radicals such as 0; react with readily oxidized substances (RH,) to produce other radicals (RH), and such reactions are self-perpetuating if they produce 0;. e-+ 0, "-f 0; net: e- The equations show that 0; has a catalytic role. 0; initiates a series of reactions that repeats n times, but 0; is not consumed. 0, rather than 0; is the electron-acceptor, so that this is an "auto-oxidation" reaction. The equations also show that production of one 0; ion could result in a large amount of 0, uptake, H,02 production, and RH, oxidation. In principle, one 0; could initiate oxidation of an unlimited amount In the second mechanism, 0; rather than O2 is the electronacceptor and no free-radical chain reaction occurs. 0; oxidizes RH2 to R, and 0; is reduced to H202, which may also oxidize of RHZ.
RH, to R.
These equations show that when 0; acts as an oxidant (electron-acceptor), up to three RH, are oxidized per two 0;. Aims of this study were to determine whether 0; participates in oxidation of extracellular reductants by stimulated neutrophils and to determine whether 0; initiates auto-oxidation or acts as an oxidant.
As described below and in another study on ascorbate oxidation (25), 0; contributes to oxidation. 0;-dependent auto-oxidation is observed with certain RSH compounds, requires Mns+, and is due in part to the RSH-oxidase activity of the neutrophil enzyme myeloperoxidase. On the other hand, 0; acts as an oxidant for GSH and ascorbate. Mn'+ is required for oxidation of GSH but not ascorbate.

EXPERIMENTAL PROCEDURES
Materials-Acetaldehyde was from Eastman Kodak, L-ascorbate from Gallard-Schlesinger, dehydroascorbate and dimethyl sulfoxide (spectrophotometric grade) from Aldrich, PMA from Pharmacia LKB Biotechnologies Inc., and NaCl ("ultrar") from Atomergic Chemetals Corp. Bovine serum albumin, chelators, cyt c (Type VI), lactoferrin, Nbs, and RSH compounds, superoxide dismutase, xanthine oxidase (10 mg/ml), and zymosan were from Sigma. Dehydroascorbate dissolved in 0.154 M NaCl contained less than 1% ascorbate but was only 67% pure based on the yield of ascorbate obtained by reduction or the ability to reduce hypochlorous acid (HOC]) measured a t 291 nm (26). Concentrations of added dehydroascorbate were calculated assuming 67% content by weight. Potassium superoxide (KO,, Pfaltz and Bauer) solutions (27) were prepared by adding 0.5 g of KO, to 4 ml of dimethyl sulfoxide and incubating 1 h a t 37 "C in a glassstoppered tube. Catalase crystals (Boehringer Mannheim) were washed in water and dissolved in 0.14 M NaCl with 15 mM potassium phosphate, p H 7.4 (C1medium) or in 67 mM Na2SO4 with 32 mM potassium phosphate, pH 7.4 (C1"free medium). PMA (5 mg) was dissolved in 0.1 ml of dimethyl sulfoxide and diluted 100-fold into 48 mg/ml bovine serum albumin in water. This solution was diluted 100-fold into CI-or C1"free medium and stored a t -20 "C. Nbs (4 mM) was prepared by adding 20 p l of 2-mercaptoethanol to 74 ml of 2.5 mM Nbs, in CI-or CI--free medium. Lactoferrin (0.1 mM) was incubated 24 h a t 22 "C in 5 mM phosphate, pH 7.6, with ascorbate and ferrous sulfate, each at 0.2 mM (28). After dialysis in 5 mM phosphate, pH 7.6, absorbance of Fe"4actoferrin was 0.446 a t 450 nm, indicating a bound Fe'+ concentration of 0.16 mM. Myeloperoxidase was purified from human leukemic leukocytes, and eosinophil peroxidase was from horse blood (Cleveland Scientific) (29). Neutrophils from human blood were isolated by density gradient centrifugation (14). Zymosan was opsonized in human serum (30). Incubations-Cells were incubated a t 37 "C in 0.5 ml of C1-or C1" free medium with 1 mM glucose and 1 mM MgSO, in closed 15-ml siliconized glass tubes with continuous mixing to maintain aerat.ion. O2 uptake was measured in a stirred chamber a t 37 "C with a Clarketype 0, electrode (Yellow Springs Instruments). Rates of 0, uptake were calculated from the linear portion of plots of 0, concentration versus time in the period from 1 to 5 min after adding PMA (30). Cells were stimulated by adding 10 pl/ml of the aqueous PMA solution to give final concentrations of 80 nM PMA, 4.8 pg/ml albumin (80 nM), and 0.0001% (14 p~) dimethyl sulfoxide. Other incubations were under the same conditions but without glucose, MgSO,, or PMA.
Determinations-To measure 0; production (31), incubation mixtures with cyt c and 5 pg/ml catalase were diluted 7-fold in cold deaerated medium with 5 pg/ml catalase and 15 pg/ml superoxide dismutase. Concentration of reduced cyt c was calculated from the difference (21.1 mM".cm") in absorbance of reduced and oxidized cyt c a t 550 nm (32). Incubations with reductants were stopped by diluting 10-to 15-fold in cold deaerated medium with 5 pg/ml catalase and 15 pg/ml superoxide dismutase and centrifuging. When high catalase or metal ion concentrations were present, 0.1 mM EDTA was included in the diluent. Ascorbate concentration was calculated from the 264 nm absorbance of supernatants assuming a millimolar extinction coefficient of 14.5. Nbs oxidation was determined by calculating Nbs and Nbs, concentrations from absorbance a t 323 and 409 nm (33), assuming millimolar extinction coefficients of 2.05 and 14.05 for Nbs and 18.58 and 0.26 for Nbs,. Oxidation of other RSH compounds was determined by including 0.3 mM Nbsz in the diluent and calculating RSH concentration from the yield of Nbs, measured a t 409 nm (34). Oxidation of dehydroascorbate was determined by diluting 4-fold, centrifuging, incubating 0.5-1111 portions of the supernatants with 0.5 ml of 20 mg/ml cyt c under N, for 4 h, diluting with 8 ml of cold medium, calculating the reduced cyt c concentration from the 550 nm absorbance, and calculating dehydroascorbate concentration assuming reduction of two cyt c per dehydroascorbate.2 Similar cyt c reduction was obtained with dehydroascorbate or its hydrolysis product, diketogulonate (35). When ascorbate was present, ascorbate was determined from absorbance at 264 nm, and values for reduced cyt c were corrected for reduction by ascorbate, using a standard curve of cyt c reduction versus ascorbate concentration. Table I compares oxidation of six RSH compounds and ascorbate by PMA-stimulated neutrophils in CI--free medium. From 25 to 60% of the amount of each reductant was oxidized during 1 h at 37 "C, whereas only about 5% was oxidized without PMA or neutrophils. When the data are corrected for oxidation in the absence of neutrophils, cysteine was the most rapidly oxidized compound and GSH was the slowest. Nevertheless, rates of oxidation differed by no more than 2-to %fold. Rates of oxidation were linear with time up to 45-60 min and roughly proportional to reductant concentration up to 1-2 mM (not shown). The lack of specificity and saturability argues against oxidation at the active sites of RSHand ascorbate-oxidase enzymes.

Requirements for Oxidation-
Oxidation was weakly inhibited by catalase, strongly inhibited by superoxide dismutase, and blocked by the two enzymes (Table I). Similar results were obtained in C1medium and with PMA or opsonized zymosan as the stimulus. Oxidation of all the compounds was faster in C1-medium, but myeloperoxidase-catalyzed oxidation of C1-to HOC1 and the reac-E. L. Thomas, manuscript in preparation.  tion of HOC1 with the reductants was not required for oxidation.
Neutrophil carbohydrate metabolism was required. For example, ascorbate oxidation was inhibited 60% by omitting glucose, 73% by adding 10 mM deoxyglucose to inhibit glycogen utilization, and 83% by adding 10 mM fluoride to inhibit glycolysis. The results suggest that oxidase activation, NADPH production, and NADPH-dependent 0; production were required. However, secretion of lysosomal components would also be inhibited under these conditions.
Oxidation was roughly proportional to the number of neutrophils ( Fig. 1). This figure also compares the amount of oxidation with the amount of 0; produced by the NADPHoxidase. About one ascorbate was oxidized per O;, but RSH oxidation exceeded 0; production. With 1 mM RSH and neutrophils at 0.5-2 X 106/ml, from one to three RSH were oxidized per 0;. Up to eight Nbs were oxidized per 0; when neutrophils were less than 0.5 X 10G/ml. Similar results were obtained with cysteine (not shown). As described below, even higher ratios were obtained when Mn2+ was added or NADPH-oxidase activity was suppressed by lowering the pH. Nevertheless, superoxide dismutase inhibited strongly under all conditions. Role ofH2O2-0xidation in C1"free medium was stimulated by adding azide to inhibit endogenous catalase (Table 11). Oxidation was also stimulated by adding C1-to permit H,O,dependent oxidation of C1-to HOC1. The results indicate that part of the Hz02 was scavenged by intracellular GSH-peroxidase and catalase, and the remainder reacted with extracellular reductants. Azide or C1-increased the participation of H,O, in extracellular oxidation.
The effect of C1-was blocked by adding catalase, confirming that H2OZ was involved in the stimulation by C1-. Similar stimulation was obtained with other halide substrates for myeloperoxidase (Br-, I-) or a pseudohalide substrate (SCN-).
The highest rates of oxidation were obtained with 80 mM Cl-, 10 mM Br-, or 1 mM Ior SCI", consistent with the specificity of myeloperoxidase.

Nbs GSH Ascorbate
Oxidation, percent of control ascorbate oxidation (Fig. 2, right). Fig. 2 also shows that at p H 7.4 the oxidase activity of catalase was nearly undetectable at concentrations up to 5 pg/ml. This catalase concentration was used in all other experiments.
Role of Lysosomal Components-To determine whether stimulated neutrophils secrete an enzyme or factor that catalyzes oxidation, neutrophils were incubated in C1"free medium with PMA or opsonized zymosan and centrifuged, and portions of the supernatants were incubated with Nbs, GSH, or ascorbate. There was no oxidation, indicating that lysosomal components did not catalyze oxidation in the absence of 0; production by the NADPH-oxidase. However, these results do not rule out a role for myeloperoxidase because most of the secreted enzyme is associated with the cell membrane rather than free in the supernatant (15).
Adding purified myeloperoxidase resulted in a large increase in Nbs oxidation, although no C1-or other halide was added (Table 11). In fact, myeloperoxidase was more effective without C1-, indicating that the effect was unrelated to halide oxidation. Myeloperoxidase also promoted Nbs oxidation in the absence of PMA. This oxidation was blocked by superoxide dismutase, although the NADPH-oxidase did not produce detectable amounts of 0; without PMA. The results indicate that myeloperoxidase had Nbs-oxidase activity that was inhibited by superoxide dismutase but which was independent of 0; production by the NADPH-oxidase.
The amount of enzyme added in Table I1 was equal to the amount contained in the cells. That is, if neutrophils (2 x 106/ml) secreted their entire content of myeloperoxidase, the concentration in the medium would be 0.1 p~ (29). Therefore, the RSH-oxidase activity of the endogenous myeloperoxidase was sufficient to contribute to Nbs oxidation. However, the enzyme had only a weak GSH-oxidase activity and no ascorbate-oxidase activity. Stimulated neutrophils oxidized Nbs equally well at p H 5.5 or 7.4, although the rate of 0; production by the NADPHoxidase at pH 5.5 is only 10% that at pH 7.4 (29, 36). Nbs was oxidized much faster than GSH or ascorbate at low pH, and oxidation was inhibited by C1-or azide. These observations indicate that oxidation at low pH was due primarily to the Nbs-oxidase activity of myeloperoxidase. Experiments with the purified enzyme indicated that Nbs-oxidase activity was much greater at low pH. The metal-ion-binding protein lactoferrin is another of the major lysosomal components secreted by neutrophils. Adding 10 p~ lactoferrin to chelate any free Fe3+ had no effect, indicating that Fe3+ was not required (Table 11). Similarly, Fe"-lactoferrin did not stimulate, indicating that FeY+-lactoferrin did not have RSH-or ascorbate-oxidase activity.
Role of Mn2+-Omitting Mg2+ or substituting Ca2+ for M$+ had no effect (Table 11). However, Nbs oxidation was partially inhibited by chelating agents for divalent cations. Chelator concentrations as low as 3 p M gave the maximum amount of inhibition, whereas concentrations up to 0.3 mM did not inhibit 0; production by the NADPH-oxidase. Chelators had no effect on GSH oxidation in Cl--free medium, but inhibited up to 32% when GSH oxidation was stimulated by adding C1-. In contrast, chelators had no effect on ascorbate oxidation with or without C1-. The results suggest that RSH oxidation was promoted by an endogenous divalent cation.
Adding Mn2+ promoted oxidation of RSH compounds but not ascorbate, suggesting that Mn2+ was the endogenous cation (Table 11). MnZ+ did not promote RSH oxidation in the absence of neutrophils but did promote Nbs oxidation by unstimulated neutrophils. Therefore, the small amount of myeloperoxidase released by unstimulated cells was sufficient to catalyze Nbs oxidation provided that Mn2+ was added.
Mn2+ has superoxide dismutase activity (37), which results from oxidation of Mn2+ to Mn3+ by 0; followed by reduction of Mn"+ to Mn2+ by a second 0;. Mn"-catalyzed dismutation of 0; resulted in inhibition of cyt c reduction by stimulated neutrophils or the 0;-generating system xanthine oxidase/ acetaldehyde. Mn2+ at 2 p M inhibited reduction of 0.5 mM cyt c by 50%, indicating that 0; reacted over 100-times faster with Mn2+ than with cyt c. Despite the potent dismutase activity of Mn2+, Mn2+ promoted 0;-dependent RSH oxidation and had little effect on ascorbate oxidation.
Reaction of 0; with Reductants or Mn2+-To determine whether 0; reacts with the reductants or with some other substance in the incubation mixture, concentrations of the reductants and superoxide dismutase were varied. If the reductant competes with superoxide dismutase for reaction with O;, then more enzyme will be required to inhibit oxidation when the reductant concentration is increased. Superoxide dismutase inhibited ascorbate oxidation by 50% (Fig. 3). When ascorbate was increased 4-fold from 0.5 to 2 mM, the amount of enzyme required for half the maximum effect (25% inhibition) increased 4-fold from 3.5 to 14 pg/ml. In contrast, when ascorbate was held constant at 0.5 mM and the number of cells was increased 4-fold, the amount of enzyme required for half the maximum effect remained nearly constant a t 2.7-3.5 pg/ml. Therefore, oxidation resulted from the direct reaction of 0; with ascorbate rather than the reaction of 0; with an enzyme or factor contributed by the cells.
Different results were obtained with RSH compounds. Superoxide dismutase inhibited GSH oxidation by up to 60-70%, and enzyme at 0.1 pg/ml was required for half-maximal inhibition with 2 or 6 mM GSH (Fig. 4.4). Varying the number of cells also had no effect (not shown). Similarly, superoxide dismutase inhibited Nbs oxidation by 80-90%, and half the maximum effect was obtained with 0.2-0.3 pg/ml enzyme regardless of Nbs concentration or the number of cells (Fig.  4B). Instead, more enzyme was required when Mn2+ was added. With 10 or 100 p M Mn2+, superoxide dismutase at 1. imply that in the absence of added Mn2+, sufficient Mn2+ was present to promote oxidation. This low level of Mn2+ was probably contributed by the buffers rather than the cells because increasing the number of cells did not increase the amount of enzyme required for inhibition. Fig. 4 also shows that high concentrations of superoxide dismutase promoted GSH oxidation. This effect may be due to the peroxidase activity of the enzyme (9) and to inactivation of the enzyme and release of Zn2+ and Cu'+. At high concentrations, the enzyme had a weak GSH-oxidase activity that was blocked by chelators. 0, Uptake-Incubation of stimulated neutrophils with reductants resulted in increased rates of 0, uptake that were proportional to reductant concentration and the number of neutrophils. The increase in O2 uptake was blocked by superoxide dismutase (Table 111), although this enzyme only partially inhibited oxidation.
In the absence of reductants, catalase caused nearly a 50% decrease in the observed rate of 0, uptake (Table 111) due to the dismutation of two Hz02 to yield water and one 02. Adding azide to inhibit endogenous catalase caused a small increase in the rate of 0, uptake, and the rate with azide was exactly 2 times the rate with added catalase. Similarly, 0, uptake with reductants and catalase was about 50% slower than uptake with reductants alone or reductants and azide. Therefore, the increase in 0, uptake that accompanied RSH or ascorbate oxidation was the result of increased H,O, production.
Significant rates of O2 uptake were obtained with unstimulated neutrophils in the presence of MnZ+ and RSH compounds, particularly Nbs. This 0, uptake was due to the Mn'+-dependent RSH-oxidase activity of myeloperoxidase. Mn2+ did not promote O2 uptake by neutrophils alone or RSH compounds alone.
Products of Oxidation-Nbs was oxidized to the disulfide (RSSR) compound Nbsz as indicated by the relation between loss of absorbance at 409 nm and the increase at 323 nm. There was no loss of 323 nm absorbance when neutrophils were incubated with Nbs2 in C1"free medium and no increase in O2 uptake when neutrophils were incubated with Nbs2 or other RSSR compounds. Therefore, RSH or RSSR was not oxidized to a higher oxidation state such as the sulfinic or sulfonic acid (RS02H or RS03H).
One-electron oxidation of ascorbate yields monodehydroascorbate, which dismutates to yield ascorbate and dehydroascorbate, whereas two-electron oxidation yields dehydroascorbate directly (38). However, the product of ascorbate oxidation could not be converted to ascorbate by reduction with dithiothreitol, indicating that dehydroascorbate did not accumulate during ascorbate oxidation. Dehydroascorbate would not be expected to accumulate because it undergoes rapid hydrolysis at pH 7.4 to diketogulonate (35). In addition, H202 and HOC1 oxidize dehydroascorbate (35) and diketogdonate.' Table IV shows that ascorbate oxidation by neutrophils resulted in accumulation of diketogulonate, although the yield of this product was less than the amount of ascorbate oxidized. The low yield was due to oxidation of dehydroascorbate and/ or diketogulonate, as indicated by incubating neutrophils with dehydroascorbate and measuring the yield of diketogulonate.   Oxidation was partially inhibited by catalase, indicating that part of the H202 or HOC1 formed during ascorbate oxidation would be consumed in reactions with ascorbate oxidation products. Oxidation was also partially inhibited by superoxide dismutase, suggesting that dehydroascorbate was oxidized by O;, although at a much slower rate than ascorbate. Oxidation of dehydroascorbate was stimulated by Mn2+, but the effect of Mn2+ was suppressed when ascorbate was present. The inability of Mn2+ to stimulate or inhibit oxidation may be due to the chelating activity of ascorbate (39).
Model Systems-Similar oxidation of ascorbate and dehydroascorbate was obtained with stimulated neutrophils or with 0; added as a solution of KO, in dimethyl sulfoxide (Table IV). Oxidation of either compound by KOz was partially inhibited by superoxide dismutase or catalase, and Mn2+ stimulated the oxidation of dehydroascorbate but not ascorbate. Greater oxidation of dehydroascorbate was obtained with KO, than with neutrophils, probably because dehydroascorbate reacts slowly with H202 and thus competes poorly with H,O,-consuming activities in neutrophils. The results indicate that 0; reacts with ascorbate and dehydroascorbate at significant rates and that no enzyme or other factor is required for 0;-dependent oxidation.
Nbs was oxidized when incubated with the hemoproteins myeloperoxidase, eosinophil peroxidase, and catalase (Table  V). Therefore, these metalloproteins had RSH-oxidase activity that was independent of an external source of 0;. Nevertheless, oxidation was inhibited by superoxide dismutase, indicating that 0; was produced when the enzymes were incubated with Nbs and that 0; was an intermediate in Nbs oxidation. Oxidase activity of the peroxidases or catalase was abolished by heating, as was the ability of catalase or superoxide dismutase to inhibit oxidation.
The metalloprotein xanthine oxidase did not have RSHoxidase activity, but Table V shows that Nbs was oxidized when incubated with the 0;-generating system xanthine oxidase/acetaldehyde. These results were obtained at a level of xanthine oxidase that produced 0; at 30-40 nmol-ml". h-I. About 350 p~ Nbs was oxidized, so that about 10 Nbs were oxidized per OF, and this oxidation was inhibited by superoxide dismutase.
Halide ions inhibited the RSH-oxidase activity of the peroxidases but had little effect on the other systems (Table V).

Acet
Complete (Nbs, -C1-)  (Table V), whereas complete inhibition was obtained at enzyme concentrations above 0.4 pM (not shown). A similar dependence on enzyme concentration was observed with other halides or SCN-, and with eosinophil peroxidase. Because the concentration of enzyme secreted by neutrophils at 2 X lo6/ ml would be less than 0.1 p~, halides would not necessarily block the oxidase activity of myeloperoxidase in experiments with neutrophils. Similarly, the peroxidase-inhibitor azide would not block completely. Azide at 20 p~ inhibited by 50%, but complete inhibition was not obtained even at 1 mM azide (Table V) .
Azide caused a 2-to 3-fold increase in the RSH-oxidase activity of catalase (Table V), although azide inhibits catalase activity. The results suggest that the two activities of catalase are distinct and that azide promoted RSH oxidation by preventing the destruction of H,O,.
Chelators for divalent cations caused partial or complete inhibition of the oxidase activity of peroxidases or catalase (Table V), although they do not inhibit peroxidase or catalase activity. Partial inhibition was also obtained with xanthine oxidase/acetaldehyde, although chelators do not inhibit 0; production. Maximum inhibition was obtained with 1-3 p~ chelator. Oxidation was promoted by Mn2+, suggesting that the chelators inhibited by sequestering a small amount of contaminating Mn2+.
Lactoferrin did not inhibit, indicating that Fe3+ was not involved, and Fe3-lactoferrin did not stimulate (Table  V). Free iron as ferric chloride or ferrous ammonium sulfate at concentrations up to 0.1 mM did not catalyze oxidation of 1 mM Nbs, GSH, or ascorbate. Mn2+ did not cause Fe3+, lactoferrin, or Fe3+-lactoferrin to take on oxidase activity, and no oxidation was obtained with Mn2+ alone or Mn2+ plus xanthine oxidase or acetaldehyde.
Myeloperoxidase had greater RSH-oxidase activity with Nbs, cysteine, or dithiothreitol than with GSH or 2-mercaptoethanol and had no ascorbate-oxidase activity (Table V). A similar pattern was observed with eosinophil peroxidase, catalase, or xanthine oxidase/acetaldehyde. Fig. 5 shows that Mn2+ did not alter this specificity and that Mn2+ caused similar stimulation of RSH oxidation by myeloperoxidase or xanthine oxidase/acetaldehyde. This figure also provides an estimate of Mn2+ contamination. By extrapolation from the Myeloperoxidase and xanthine oxidase have dissimilar active sites and enzymatic mechanisms. Therefore, it is unlikely that the specificity of RSH oxidation was due to selective binding of RSH compounds by the enzymes. Instead, the results suggest that xanthine oxidase/acetaldehyde and myeloperoxidase/RSH acted as 0;-generating systems and that differences in rates of oxidation were due to differing interactions of RSH compounds with 0; and Mn2+. Fig. 5 shows that there was little GSH oxidation by xanthine oxidase/acetaldehyde at this low level of enzyme. However, 180 p~ GSH was oxidized when xanthine oxidase was increased to a level (8 pg/ml) that produced 0; at 200 nmol. ml". h-', which is equivalent to neutrophils at 1 x 106/ml. This oxidation was inhibited by catalase but not by superoxide dismutase. Mn2+ caused a 3-fold increase in GSH oxidation to 540 p~, and this oxidation was inhibited 50% by superoxide dismutase. The maximum rate of GSH oxidation was obtained at 10 p~ Mn2+. Therefore, oxidation of three GSH per 0; was observed when the rate of 0; production was high and sufficient Mn2+ was added.

0; as an Oxidant-
The results indicate that neutrophils can use 0; as an oxidant to eliminate GSH or ascorbate'from the extracellular phase. The amount of reductant eliminated in this way is up to 3 times greater than would be obtained if 0, was reduced directly to H202, which explains the utility of releasing 0; rather than H,O,.
Oxidation of extracellular GSH by stimulated neutrophils is consistent with oxidation of GSH by OF, reduction of 0; to Hz02, and oxidation of GSH by H20,. 2H+ + 20; + 2GSH -+ 2H202 + GSSG 2H202 + 4GSH -+ 4H20 + 2GSSG net: 2H' + 20; + 6GSH + 4&O + BGSSG Intermediates such as Mn3+, the thiyl radical (GS), and sulfenic acid (GSOH) may be involved, and the second reaction may be mediated by myeloperoxidase-catalyzed oxidation of C1-to HOCl and the oxidation of GSH by HOCl. Therefore, these equations summarize multistep reactions, although the result is oxidation of up to three GSH per 0;. In the presence of catalase, one GSH would be oxidized per O;, so that catalase inhibits by 67%. Similarly, superoxide dismutase inhibits by 67%. This mechanism results in increased 0, uptake and H202 production relative to what is observed in the absence of GSH. However, 0, is reduced only in the reaction catalyzed by NADPH-oxidase. Therefore, this is not an auto-oxidation reaction.
In the absence of GSH, the rate of 0, uptake is about half the rate of 0; production because spontaneous dismutation of two 0; yields one 0,. When GSH is added, 0, uptake equals 0; production, so that 0, uptake increases 2-fold, although the rate of 0, reduction does not change. With GSH and catalase, one O2 it taken up per two OF, SO that catalase inhibits by 50%. Similarly, superoxide dismutase inhibits 0, uptake by 50%, although either enzyme inhibits GSH oxidation by up to 67%.
Ascorbate oxidation by neutrophils is more complex. 0; acts as an oxidant for ascorbate (AH,), but part of the H2O, produced in this reaction is consumed in the oxidation of AH, by H202 The result is 0;-initiated Mn'+-catalyzed RSH auto-oxidation. In this mechanism, catalase inhibits by 50%, and superoxide dismutase inhibits by a large percentage that depends on the value of n. 0, uptake in the presence of RSH is inhibited 50% by catalase, and superoxide dismutase eliminates all of the additional 0, uptake that is associated with RSH oxidation.
MnZ+ catalyzes the oxidation of GSH by 02, but autooxidation is not observed. One possible explanation is that the reaction of two GS radicals with each other to produce GSSG is faster than the reaction of GS with 0, to produce 0;. Mn2+ catalyzes 0;-dependent auto-oxidation of other RSH compounds by stimulated neutrophils, xanthine oxidase/acetaldehyde, peroxidases, and catalase. These results suggest that the peroxidase/RSH and catalase/RSH combinations act as 0;-generating systems and that Mn2+ catalyzes the attack of 0; on the RSH compounds. The result is that leukocyte peroxidases and catalase have Mnz+-dependent RSH-oxidase activities.
Mn2+-dependent oxidase activities of hemoproteins have been studied extensively (40-45) particularly with horseradish peroxidase as the catalyst and NADH as substrate. These activities have caused confusion in studies on the neutrophil NADPH-oxidase. There are many reports of Mn2+-stimulated NAD(P)H oxidation and 0, uptake by subcellular fractions from neutrophils. In contrast, the "respiratory-burst oxidase" is not stimulated by Mn2+ or inhibited by azide, shows specificity for NADPH, and is present in cells lacking myeloperoxidase (1-5, 46).
The significance of Mn2+-catalyzed reactions in leukocyte functions is unclear because the physiologic concentration of Mn2+ in forms that are active as redox catalysts is unknown. Ascorbate may bind and deactivate Mn2+. Furthermore, the Mn2+-dependent oxidase activities of leukocyte peroxidases are at least partially inhibited by physiologic levels of halides and SCN-. On the other hand, only small amounts of Mn2+ are required, the reactions may be favored at the acid pH of the phagolysosome, and microorganisms may release Mn2+ into phagolysosomes. Certain bacteria accumulate high levels of Mn2+ in forms that react with 0; (47). Therefore Mn2+ may catalyze 0;-dependent oxidation of reductants in phagolysosomes. The Mn2+-dependent oxidase activities of leukocyte peroxidases are of particular interest because these activities could provide an alternative source of 0; and cytotoxic oxidants in cells that are deficient in NADPH-oxidase activity.
Role of 0; in Leukocyte Antimicrobial and Cytotoxic Activities-The results suggest that an important role for 0; in leukocyte function is to participate in oxidation of extracellular reductants. When these reductants are eliminated, H,O, and oxidized halides are free to exert toxicity. 0;-dependent oxidation also results in a large increase in H202 production, and this H,O, could be used to attack target cells or to produce oxidized halides. However, Hz02 production increases only as long as the reductant is present, and most of the H,O, or oxidized halides would be scavenged by the reductant.
Predictions can be made that would provide further tests of the hypothesis. First, superoxide dismutase may block toxic activities of leukocytes when the system under study contains extracellular reductants. Such conditions would be encountered in vivo and in experimental systems containing plasma, serum, interstitial fluid, broken cells, or media conditioned by growing cells. Second, compounds that undergo rapid 0;dependent oxidation may be ineffective in blocking toxicity or may exacerbate toxicity by promoting HzOz production. Third, the combination of superoxide dismutase and a reductant may be much more protective than either agent alone.
Many studies have reported protective effects of superoxide dismutase alone or with catalase. However, the role of extracellular reductants in these effects has not been considered. Results of this study also indicate that protective effects could easily be misinterpreted. Inhibition of leukocyte cytoxicity by superoxide dismutase does not necessarily indicate that 0; reacts with the target cells. Inhibition by superoxide dismutase plus catalase does not necessarily indicate that iron and hydroxyl radicals are involved. Instead, superoxide dismutase may inhibit oxidation of extracellular reductants. When these reductants are protected against 0;-dependent oxidation, they can detoxify H,O, and products of peroxidase/H20,/ halide systems.