Targeting Sod by Gene and Protein Engineering and Inhibition of Free Radical Injury

Although oxygen toxicity of tissues can be decreased by a variety of antioxidants and some enzymes, such as SOD and catalase, their protective effect on tissue injury in various diseases are fairly small predominantly because of their unfavorable in vivo behavior. To minimize oxidative stress in various diseases. such as ischemic myocardial injury, circulatory disturbance and corneal inflammation, we synthesized three types of SOD derivatives by gene and protein engineering technique. One type of SOD (SM-SOD covalently linked with hydrophobic anions) circulates bound to albumin with a half life of 6 h and accumulates in tissues whose local pH is decreased. The other type of SOD (AC-SOD covalently linked with long chain fatty acids via the ?-amino group of lysyl residues) anchors onto membranc/lipid bilaycrs of various cells. The last type of SOD (HB-SOD synthesized by constructing a fusion gene coding human CuZn-type SOD and a C-terminal heparin-binding domain) binds to heparin-like proteoglycans on vascular cndothelial cell surface. Intravenous administration of either SM-SOD or HB-SOD markedly inhibited postischcmic reflow arrhythmias in the rat. When the left anterior descending artery was occluded permanently. about 65 % of animals died within 30 min predominantly due to irreversible ventricular fibrillation; the motality of animals decreased to 15 % by administering SM-SOD either before or after occlusion. Topically administered AC-SOD bound to the corneal epithelial cell surface and polyrnorp%onuclear leukocytes and efficiently dismutated superoxide radicals at their cell surface. Thus,' endotoxin-induced kcratitis was inhibited markedly by topical instillation of AC-SOD. Unmodified SOD itself failed to inhibit the pathologic events occurring in these disease models. Thus, these SOD derivatives permit in vivo studies on the mechanism and the site for oxygen toxicity in various diseases and provide a new strategy for targeting enzymes and bioactive peptides for medical use to appropriate site(s) of their action.     

Synthesis of a cytochrome c derivative with prolonged in vivo half-life and determination of ascorbyl radicals in the circulation of the rat.

Since cytochrome c and acetylated cytochrome c disappear from the circulation with a half-life of 4 min, these proteins cannot be used for in vivo detection of superoxide radicals and related metabolites. To determine superoxide and other radicals in vivo, a cytochrome c derivative (SMAC) was synthesized by linking 1 mol of poly(styrene-co-maleic acid) butyl ester (SM) to cytochrome c, followed by acetylation of its lysyl amino groups. SMAC retained 8 and 80% of cytochrome c activity to react with ascorbyl and superoxide radicals, respectively. However, SMAC did not serve as a substrate for cytochrome c reductase and cytochrome c oxidase. When injected intravenously to the rat, SMAC circulated bound to albumin with a half-life of 130 min. SMAC was rapidly reduced in the circulation of intact animals. Treatment of animals with paraquat markedly enhanced the reduction of the circulating SMAC. We have synthesized an SM-conjugated superoxide dismutase (SOD) derivative (SM-SOD) that circulates bound to albumin with a half-life of 6 h. Kinetic analysis revealed that SM-SOD effectively inhibited the superoxide-dependent reduction of SMAC either in the presence or absence of 0.5 mM albumin. However, the reduction of the circulating SMAC was not inhibited by SM-SOD both in normal and paraquat-treated animals. Plasma samples from both animal groups also reduced cytochrome c and SMAC by an SOD-insensitive mechanism. However, after treatment with ascorbate oxidase, both plasma samples lost their activity to reduce cytochrome c and SMAC. These and other results suggest that ascorbyl radical might principally be responsible for the reduction of circulating SMAC and that plasma levels of ascorbyl radical might increase in paraquat-treated animals.     

Expression of a hybrid Cu/Zn-type superoxide dismutase which has high affinity for heparin-like proteoglycans on vascular endothelial cells

Since plasma levels of enzymes, such as superoxide dismutase (SOD), that scavenge reactive oxygen species are low, surface membranes of endothelial and parenchymal cells of various tissues are often exposed to oxidative stress. To dismutase superoxide radicals efficiently in and around vascular endothelial cells, we constructed a fusion gene encoding a hybrid SOD (HB-SOD) consisting of human Cu/Zn-SOD and a C-terminal basic peptide that binds to heparin-like proteoglycans. The fusion gene was expressed in yeast, and the resulting HB-SOD was highly purified. Upon sodium dodecyl sulfate-polyacrylamide gel electrophoresis, HB-SOD revealed a protein band with an apparent molecular weight of 20,000. HB-SOD bound to endothelial cells of aortic segments by a mechanism which was inhibited by heparin but not by antithrombin III. When injected intravenously to rats, 125I-labeled HB-SOD rapidly disappeared from the circulation; the rate of disappearance was decreased by heparin. Less than 1% of the injected HB-SOD was found in the urine 20 min after administration at which time more than 70% of SOD was excreted in its intact form. Immunohistochemical studies revealed that HB-SOD predominantly bound to heparin-like proteoglycans on endothelial cells of the artery and other tissues. HB-SOD might permit studies on pathophysiological roles of superoxide radicals in and around vascular endothelial cells in vivo.     

Proton release associated with respiratory burst of polymorphonuclear leukocytes

The stimulation of polymorphonuclear leukocytes (PMNs) by phorbol-12-myristate-13-acetate in the presence of sodium fluoride caused the release of protons into the reaction medium concomitant with the generation of superoxide anions. The rates of oxygen consumption and proton release due to the metabolic burst were 16.3 +/- 3.5 and 10.2 +/- 1.1 nmol/min/10(7) cells respectively. When the superoxide anions were trapped with cytochrome c, the proton release was increased (35.8 +/- 0.5 nmol/min/10(7) cells) until the cytochrome c was reduced. Since the protons released from the activated cells would be consumed by the generated superoxide anions in the extracellular medium, the net amount of the protons released was 3-4-fold greater than that observed in the absence of extracellular cytochrome c. The increased proton release may be coupled to increased cellular respiration, since the inhibition of the respiratory burst with deoxyglucose, p-chloromercuribenzoic acid or chlorpromazine decreased the proton release. Amiloride (2 mM) inhibited the proton release by up to 40%. These observations suggest that some mechanisms other than a Na+/H+ antiport and carbon dioxide diffusion could be transporting the H+ generated in the cytosol of the activated PMNs.     

Synthesis of superoxide dismutase derivative that specifically accumulates in renal proximal tubule cells.

Protection of tissues from oxygen toxicity is one of the major prerequisites to aerobic life. Since a wide variety of xenobiotics with prooxidant activity is excreted by the kidney, renal tubule cells should be protected from hazardous oxygen species. Because intravenously injected Cu/Zn-type superoxide dismutase (SOD) is rapidly excreted in the urine in its intact form, effective dismutation of superoxide radicals cannot be achieved in vivo by intravenously administered SOD. To scavenge superoxide radicals and inhibit their toxic effects in and around renal tubule cells, a hexamethylene-diamine (AH)-conjugated SOD (AH-SOD) was synthesized. When injected intravenously into the rat, (125)I-labeled AH-SOD disappeared from the circulation with a half-life of 3 min and accumulated in the kidney. After 30 min of administration, more than 80% of the radioactivity derived from AH-SOD was found to localize in the kidney without being excreted in the urine. Immunohistochemical examination revealed that, 60 min after administration, the major part of AH-SOD localized in renal proximal tubule cells. Kinetic analysis using right-side-out-oriented renal brush border vesicles revealed that AH-SOD bound to their membrane surface by some mechanism which was inhibited by AH but not by heparin and albumin. These results indicated that AH-SOD rapidly underwent renal glomerular filtration, bound to apical plasma membranes of proximal tubule cells, and localized in these cells for a fairly long time without being excreted in the urine. Thus, AH-SOD might permit studies on the role of superoxide radicals in and around renal proximal tubule cells.     

The flavonoids, quercetin and isorhamnetin 3-O-acylglucosides diminish neutrophil oxidative metabolism and lipid peroxidation.

Two natural flavonoids, quercetin and isorhamnetin 3-O-acylglucosides, were examined for their inhibitory influence on the in vitro production and release of reactive oxygen species in polymorphonuclear neutrophils (PMNs). The generation of superoxide radical, hydrogen peroxide and hypochlorous acid were measured by, respectively, cytochrome c reduction, dichlorofluorescin oxidation and taurine chlorination. Membrane lipid oxidation was studied by the thiobarbituric acid method in mouse spleen microsomes. The addition of flavonoids at the concentration range 1-100 microM inhibited PMNs oxidative metabolism and lipid peroxidation in a dose-dependent manner. The results suggest that these flavonoids suppress the oxidative burst of PMNs and protect membranes against lipid peroxidation.     

A comparative study of superoxide dismutase activity in polymorphonuclear leukocytes, monocytes, and alveolar macrophages of the guinea pig.

Superoxide dismutase, an enzyme which catalyzes the dismutation of superoxide radical formed during the univalent reduction of oxygen, was quantitated by observing the inhibition of cytochrome C reduction in three cell fractions in guinea pig peritoneal PMNs and monocytes and compared to alveolar macrophages. No differences were found in the 16,000 × g pellets containing mitochondria, membranes, and granules and representing 96% of total SOD activity in PMNs and monocytes but only 48% total SOD activity in alveolar macrophages. The 100,000 × g microsomal pellet of alveolar macrophages contained 8% of total SOD activity and two-five times more activity than the respective fractions from monocytes and PMNs. However, there was 70 times more SOD in the 100,000 × g supernatant from alveolar macrophages containing 44% of total enzyme activity than in the same fraction of PMNs and monocytes containing less than 2% total SOD activity. SOD activity is mainly located in the 16,000 × g particulate fraction of PMN and monocytes but more equally distributed between the particulate fractions and cytosol of alveolar macrophages.     

Electron spin resonance study on the permeability of superoxide radicals in lipid bilayers and biological membranes.

The permeability of lipid bilayers and biological membranes to superoxide free radicals was examined by using superoxide dismutase (SOD)-loaded lipid vesicles and SOD-loaded erythrocyte ghosts. After exposing SOD lipid vesicles and SOD ghosts to enzymatically produced superoxide radicals and using spin-trapping and electron spin resonance (ESR) techniques, we found that SOD entrapped within erythrocyte ghosts effectively scavenges external O2.- while SOD inside the lipid bilayers has no effect. These results confirm that O2.- is able to cross through a biological plasma membrane but not across a pure lipid bilayer. The data provide instruction as to how and where anti-oxidant therapy is to be approached relative to the site of oxygen free radical production.     

Oxygen activation at the plasma membrane: relation between superoxide and hydroxyl radical production by isolated membranes

Production of reactive oxygen species (hydroxyl radicals, superoxide radicals and hydrogen peroxide) was studied using EPR spin-trapping techniques and specific dyes in isolated plasma membranes from the growing and the non-growing zones of hypocotyls and roots of etiolated soybean seedlings as well as coleoptiles and roots of etiolated maize seedlings. NAD(P)H mediated the production of superoxide in all plasma membrane samples. Hydroxyl radicals were only produced by the membranes of the hypocotyl growing zone when a Fenton catalyst (FeEDTA) was present. By contrast, in membranes from other parts of the seedlings a low rate of spontaneous hydroxyl radical formation was observed due to the presence of small amounts of tightly bound peroxidase. It is concluded that apoplastic hydroxyl radical generation depends fully, or for the most part, on peroxidase localized in the cell wall. In soybean plasma membranes from the growing zone of the hypocotyl pharmacological tests showed that the superoxide production could potentially be attributed to the action of at least two enzymes, an NADPH oxidase and, in the presence of menadione, a quinone reductase.     

Electron transfer facilitated by superoxide dismutase: a model for membrane redox systems?

Membranes, which are an amalgam of proteins and lipids, effect electron transfer through largely unknown mechanisms. Using albumin with bound fatty acids as a model, we have investigated the possible role of these two membrane constituents in electron transfer. In the presence of albumin: fatty acid, there is substantial enhancement of the reduction of ferricytochrome C by ferrous iron. To assess the possible role of free superoxide in cytochrome C reduction, we added mammalian copper/zinc containing superoxide dismutase (Cu/Zn SOD), which catalyzes the transfer of electrons between superoxide anion radicals, forming oxygen and hydrogen peroxide. Surprisingly, in the presence of either albumin or fatty acid free albumin, Cu/Zn SOD actually accelerates electron transfer from ferrous iron to ferricytochrome C. By contrast, neither inactive Cu/Zn SOD nor active manganese SOD facilitates the ferrous iron-dependent reduction of cytochrome C. These results suggest that, in some circumstances, Cu/Zn SOD may transfer electrons to alternative acceptors and that such transfer depends upon the unique reduction/oxidation reaction mechanism of Cu/Zn SOD. If so, this ubiquitous enzyme could be involved in regulating cellular electron transfer reactions as well as acting as a superoxide 'detoxify-ing' agent.     

NAD(P)H oxidation elicits anion superoxide formation in radish plasmalemma vesicles

Radish plasmalemma-enriched fractions show an NAD(P)H-ferricyanide or NAD(P)H-cytochrome c oxidoreductase activity which is not influenced by pH in the 4.5-7.5 range. In addition, at pH 4.5-5.0, NAD(P)H elicits an oxygen consumption (NAD(P)H oxidation) inhibited by catalase or superoxide dismutase (SOD), added either before or after NAD(P)H addition. Ferrous ions stimulate NAD(P)H oxidation, which is again inhibited by SOD and catalase. Hydrogen peroxide does not stimulate NADH oxidation, while it does stimulate Fe2+-induced NADH oxidation. NADH oxidation is unaffected by salicylhydroxamic acid and Mn2+, is stimulated by ferulic acid, and inhibited by KCN, EDTA and ascorbic acid. Moreover, NADH induces the conversion of epinephrine to adrenochrome, indicating that anion superoxide is formed during its oxidation. These results provide evidence that radish plasma membranes contain an NAD(P)H-ferricyanide or cytochrome c oxidoreductase and an NAD(P)H oxidase, active only at pH 4.5-5.0, able to induce the formation of anion superoxide, that is then converted to hydrogen peroxide. Ferrous ions, sparking a Fenton reaction, would stimulate NAD(P)H oxidation.     

Chemical modification of superoxide dismutase. Extension of plasma half life of the enzyme through its reversible binding to the circulating albumin

Protection of organisms from oxidative stress is one of the major prerequisites for aerobic life. Since intravenously injected Cu++/Zn++-type superoxide dismutase (SOD) rapidly undergoes renal glomerular filtration and appears in urine in its intact form, its clinical use as a scavenger for superoxide radicals has been highly limited. To test whether reversible interaction of SOD with plasma albumin might decrease the rate of disappearance of the enzyme from the circulation, the lysyl residues of the human erythrocyte-type enzyme were covalently linked with poly-(styrene-co-maleic acid) butyl ester (SMA) via amide linkage. Affinity chromatographic analysis by an albumin-Sepharose column revealed that the enzyme samples labeled with SMA (SMA-SOD) tightly bound to the column, while unmodified SOD was eluted in the unbound fractions. SMA-SOD bound to the column could be eluted by the buffer solution containing 0.1% sodium dodecylsulfate. In vivo analysis revealed that intravenously administered SMA-SOD circulated bound to albumin with an extremely long half-life (6 h), while unmodified SOD rapidly underwent renal glomerular filtration with a plasma half-life of 4 min. Thus, SMA-SOD may effectively dismutase superoxide radicals in the circulation.     

A potential role for periplasmic superoxide dismutase in blocking the penetration of external superoxide into the cytosol of Gram-negative bacteria

Superoxide is a key component of the antibacterial weaponry of phagocytes. Presumably, for this reason, strains of Salmonella typhimurium express a periplasmic superoxide dismutase (SOD) that is essential for full virulence. Because most anions cannot easily penetrate lipid membranes, it is thought that the phagosomal superoxide either damages an unknown target on the bacterial surface or reacts with nitric oxide to form peroxynitrite (HOONO), a toxic oxidant that can freely enter bacteria. However, in this study, we tested whether superoxide itself could penetrate membranes. Superoxide that was generated at high pH (>7.5) very slowly reduced cytochrome c that was encapsulated inside lipid vesicles. It did so much more quickly at lower pH (<7). Under the latter conditions, more superoxide was protonated and uncharged (HO2*), and the penetrance of superoxide was proportional to the concentration of this species. The permeability coefficient of HO2* was determined to be 9 x 10(-4) cm sec(-1), just slightly lower than that of water and far higher than the value of the anionic form (O2-, <10(-7) cm sec(-1). When Escherichia coli mutants that lack periplasmic SOD were exposed to super-oxide at pH 6.5, cytosolic fumarase B was damaged. Damage was minimal at higher pH or in strains that contained periplasmic SOD. Thus, in the acidic phagolysosome, superoxide may be able to penetrate and attack cytosolic targets of captive bacteria. This process may contribute to the potency of the oxidative burst. One role of periplasmic SOD may be to avert this damage. In contrast, periplasmic SOD was ineffective at lowering the extracellular super-oxide concentration and, therefore, may have little impact upon HOONO formation.     

Oxygen radicals produced by plant plasma membranes: an EPR spin-trap study

Plant plasma membranes are known to produce superoxide radicals, while the production of the hydroxyl radical, previously detected in complex plant tissues, is thought to occur in the cell wall. The mechanism of production of superoxide radicals by plant plasma membranes is, however, under dispute. It is shown, using electron paramagnetic resonance spectroscopy with a 5-diethoxyphosphoryl-5-methyl-1-pyrroline N-oxide spin-trap capable of differentiating between radical species, that isolated purified plasma membranes from maize roots produce hydroxyl radicals besides superoxide radicals. The results argue in favour of superoxide production through an oxygen and diphenylene iodonium-sensitive, NADH-dependent superoxide synthase mechanism, as well as through other unidentified mechanism(s). The hydroxyl radical is produced by an oxygen-insensitive, NADH-stimulated mechanism, which is enhanced in membranes in which the superoxide synthase is incapacitated by substrate removal or inhibition.     

Oxygen radicals are generated by dye-mediated intracellular photooxidations: a role for superoxide in photodynamic effects

Representative thiazines, xanthenes, acridines, and phenazines photosensitized the oxidation of reduced pyridine nucleotides and reduced glutathione when illuminated with low intensity visible light. Photooxidation resulted in oxygen consumption and in superoxide generation, assayed as the superoxide dismutase (SOD)-inhibitable reduction of ferricytochrome c. The major pathway of electron transfer involved dye reduction rather than singlet oxygen-mediated oxidation of the substrate, as demonstrated by the relative insensitivity of the oxidation to inhibition by sodium azide and by the observable bleaching of the dye. Hydrogen peroxide was a stable end product of photooxidation. Photosensitive dyes were photoreduced intracellularly. These dyes were transported across the membranes of Escherichia coli B and stimulated a light- and concentration-dependent increase in the cyanide-insensitive respiration. Dyes reduced intracellularly subsequently diffused out of the cell where they reduced extracellular cytochrome c. The photosensitive dyes examined in this study exhibited a light-dependent bacteriostatic effect on E. coli B grown in nutrient broth, manifested as an increased lag prior to growth. Restoration of growth coincided with increased levels of SOD, and the intracellular level of SOD correlated with the level of illumination, the dye concentration, and the reactivity of the dye to NADH in vitro. The thiazine dye, toluidine blue o, imposed a light- and oxygen-dependent lethality on E. coli B grown in glucose minimal medium. Toxicity was relieved by hydroxyl radical scavengers, and their ability to protect the cells was proportional to their reactivity with the hydroxyl radical. The results indicate that oxygen radicals and related species mediate photodynamic effects in E. coli B.     

Biochemical examination on the increased superoxide dismutase-like plasma substance observed in a state of acute gastric mucosal lesions

When experimental acute gastric mucosal lesions were produced in guinea pig by water-immersion and restraint stress, superoxide dismutase (SOD)-like substance in the plasma increased. On analysis by gel filtration, it was shown that the molecular weight of the increased SOD-like plasma substance was about 130,000, and even after treatment with trypsin, 84% of this substance remained. Since the molecular weight of intracellular SOD is about 40,000, it seems that this substance is similar to extracellular SOD, located on the endothelial cell-surface, as previously reported by Marklund et al. Our results suggest that in the presence of acute gastric mucosal lesions, SOD-like plasma substance is not identical to intracellular SOD, which derived from cell destruction by stress or free radical-induced microvascular damage or by hemolysis. Furthermore, this substance may itself work as a scavenger of free radicals generated under conditions, such as these described in the present experiment.     

Control of oxygen free radical formation from mitochondrial complex I: roles for protein kinase A and pyruvate dehydrogenase kinase

Human NADH CoQ oxidoreductase is composed of a total of 43 subunits and has been demonstrated to be a major site for the production of superoxide by mitochondria. Incubation of rat heart mitochondria with ATP resulted in the phosphorylation of two mitochondrial membrane proteins, one with a M(r) of 6 kDa consistent with the NDUFA1 (MWFE), and one at 18kDa consistent with either NDUFS4 (AQDQ) or NDUFB7 (B18). Phosphorylation of both subunits was enhanced by cAMP derivatives and protein kinase A (PKA) and was inhibited by PKA inhibitors (PKAi). When mitochondrial membranes were incubated with pyruvate dehydrogenase kinase, phosphorylation of an 18kDa protein but not a 6kDa protein was observed. NADH cytochrome c reductase activity was decreased and superoxide production rates with NADH as substrate were increased. On the other hand, with protein kinase A-driven phosphorylation, NADH cytochrome c reductase was increased and superoxide production decreased. Overall there was a 4-fold variation in electron transport rates observable at the extremes of these phosphorylation events. This suggests that electron flow through complex I and the production of oxygen free radicals can be regulated by phosphorylation events. In light of these observations we discuss a potential model for the dual regulation of complex I and the production of oxygen free radicals by both PKA and PDH kinase.     

Effects of static and 50 Hz alternating electric fields on superoxide dismutase activity and TBARS levels in guinea pigs

The toxic oxygen free radicals are extremely reactive and can cause considerable damage to biomolecules, such as RNA, enzymes, membranes, proteins, and lipids, which may in turn lead to various pathological consequences. Lipid peroxidation, evaluated by determination of thiobarbituric acid reactive substances (TBARS) is the free radical-induced oxidation of polyunsaturated fatty acids. Normally, the oxygen free radicals are neutralized by highly efficient systems in the body. These include antioxidant enzymes like superoxide dismutase (SOD). In a healthy subject, there is a balance between free radicals and levels of antioxidants. The aim of this study was to determine lipid peroxidation and SOD levels in plasma, liver, lung and kidney tissues exposed to different intensities, directions and exposure periods of static and 50 Hz alternating electric fields. Electric field intensities ranging from 0.3 kV/m to 1.8 kV/m were applied in vertical or horizontal direction in exposure periods of 1, 3, 5, 7, and 10 days. The increase in SOD and TBARS levels of plasma, liver, lung, and kidney tissues was found to depend significantly on the type of electric field and the exposure period.     

Studies on the superoxide releasing site in plasma membranes of neutrophils with ESR spin-labels

Superoxide (O2-)-generating membranes of pig blood neutrophils were studied by the ESR spin-label method. Neutrophils were spin-labeled with doxylstearic acids, consisting of nitroxide free radicals bonded to the 5, 7, 12, or 16 position of stearic acid (5-, 7-, 12-, or 16-DS), to detect the reduction of their nitroxide radicals at different positions in the membrane. The spin-labeled cells were then stimulated with phorbol myristate acetate (PMA). Stimulation of the labeled cells resulted in a marked decrease in the spin concentration of 5-DS due to the reduction by O2-, but not in those of the other three DS labels. This reduction of 5-DS was completely inhibited by copper salicylate (CS), a hydrophobic and permeable O2(-)-scavenger, but not by superoxide dismutase (SOD). CS was not inhibitory on the respiratory burst, i.e., O2(-)-generating activity of neutrophils. On the contrary, if the spin-labels were present in the extracellular medium, SOD inhibited the reduction of all four DS labels due to O2- released from PMA-stimulated cells. These results suggest that the O2(-)-releasing site is not located at the outer surface of the plasma membrane but in an inner hydrophobic environment a short distance (around 4-5 A) from its outer surface.     

Oxygen reduction in chloroplast thylakoids results in production of hydrogen peroxide inside the membrane

Hydrogen peroxide production in isolated pea thylakoids was studied in the presence of cytochrome c to prevent disproportionation of superoxide radicals outside of the thylakoid membranes. The comparison of cytochrome c reduction with accompanying oxygen uptake revealed that hydrogen peroxide was produced within the thylakoid. The proportion of electrons from water oxidation participating in this hydrogen peroxide production increased with increasing light intensity, and at a light intensity of 630 micromol quanta m(-2) s(-1) it reached 60% of all electrons entering the electron transport chain. Neither the presence of a superoxide dismutase inhibitor, potassium cyanide or sodium azide, in the thylakoid suspension, nor unstacking of the thylakoids appreciably affected the partitioning of electrons to hydrogen peroxide production. Also, osmolarity-induced changes in the thylakoid lumen volume, as well as variation of the lumen pH induced by the presence of Gramicidin D, had negligible effects on such partitioning. The flow of electrons participating in lumen hydrogen peroxide production was found to be near 10% of the total electron flow from water. It is concluded that a considerable amount of hydrogen peroxide is generated inside thylakoid membranes, and a possible mechanism, as well as the significance, of this process are discussed.