Cytochrome-Free Mitochondria of an Anaerobic Protozoan—Blastocystis hominis

ABSTRACT. Isolated mitochondria of the anaerobic protozoan Blastocystis hominis were subjected to spectral analysis, color, catalase, and peroxidase tests and found to be completely negative for cytochrome enzymes, catalase, and peroxide. Based on the absence of cytochrome enzymes, the possible evolution of B. hominis mitochondria from anaerobic bacteria is postulated.     

Studies on the sub-cellular location of particulate nitrate and nitrite reductase,glutamic dehydrogenase and other enzymes in barley roots

Summary The distribution of nitrate and nitrite reductase, glutamic dehydrogenase, cytochrome oxidase, fumarase, peroxidase and catalase in particular fractions of barley roots, separated by differential and density gradient centrifugation, has been determined. Evidence obtained suggests that there are three separate groups of particles, one, the mitochondria, containing cytochrome oxidase, fumarase and glutamic dehydrogenase, one containing catalase, and one containing nitrate and nitrite reductase. The results show that, under certain conditions, the high osmotic pressures obtained in sucrose density gradients may cause artefacts due to the release of enzymes, especially nitrite reductase, from the particles.     

Selective cytochemical localization of peroxidase,cytochrome oxidase and catalase in rat liver with 3,3′-diaminobenzidine

Summary In rat liver, three different enzymes with peroxidatic activity are demonstrated with modifications of the DAB-technique: peroxidase in the endoplasmic reticulum of Kupffer cells, catalase in peroxisomes and cytochrome oxidase in mitochondria. The major problem of the DAB-methods is their limited specifity so that often in tissues incubated for one enzyme the other two proteins are also stained simultaneously. We have studied the conditions for selective staining of each of these three enzymes in rat liver fixed either by perfusion with glutaraldehyde or by immersion in a modified Karnovsky's glutaraldehyde-formaldehyde fixative. The observations indicate that in perfusion fixed material selective staining can be obtained by reduction of the incubation time (5 min) and the use of optimal conditions for each enzyme. In livers fixed by immersion the distribution of the staining is patchy and irregular and usually longer incubation times (15–30 min) are required. Selective staining of peroxidase in Kupffer cells was obtained by brief incubation at room temperature in a medium containing 2.5 mM DAB in cacodylate buffer pH 6.5 and 0.02% H₂O₂. The exclusive staining for cytochrome oxidase in cristae of mitochondria was achieved after short incubation in 2.5 mM DAB in phosphate buffer pH 7.2 containing 0.05% cytochrome c. For selective demonstration of catalase in peroxisomes the tissue was incubated in 5 mM DAB in Teorell-Stenhagen (or glycine-NaOH) butffer at pH 10.5 and 0.15% H₂O₂. The prolongation of the incubation time in peroxidase medium caused marked staining of both mitochondria and peroxisomes. In the cytochrome oxidase medium longer incubations led to slight staining of peroxisomes. The catalase medium was quite selective for this enzyme so that even after incubation for 120 min only peroxisomes stained.     

Selective cytochemical localization of peroxidase, cytochrome oxidase and catalase in rat liver with 3,3'-diaminobenzidine

In rat liver, three different enzymes with peroxidatic activity are demonstrated with modifications of the DAB-technique: peroxidase in the endoplasmic reticulum of Kupffer cells, catalase in peroxisomes and cytochrome oxidase in mitochondria. The major problem of the DAB-methods is their limited specificity so that often in tissues incubated for one enzyme the other two proteins are also stained simultaneously. We have studied the conditions for selective staining of each of these three enzymes in rat liver fixed either by perfusion with glutaraldehyde or by immersion in a modified Karnovsky's glutaraldehyde-formaldehyde fixative. The observations indicate that in perfusion fixed material selective staining can be obtained by reduction of the incubation time (5 min) and the use of optimal conditions for each enzyme. In livers fixed by immersion the distribution of the staining is patchy and irregular and usually longer incubation times (15-30 min) are required. Selective staining of peroxidase in Kupffer cells was obtained by brief incubation at room temperature in a medium containing 2.5 mM DAB in cacodylte buffer pH 6.5 and 0.02% H2O2. The exclusive staining for cytochrome oxidase in cristae of mitochondria was achieved after short incubation in 2.5 mM DAB in phosphate buffer pH 7.2 containing 0.05% cytochrome c. For selective demonstration of catalase in peroxisomes the tissue was incubated in 5 mM DAB in Teorell-Stenhagen (or glycine-NaOH) buffer at pH 10.5 and 0.15% H2O2. The prolongation of the incubation time in peroxidase medium caused marked staining of both mitochondria and peroxisomes. In the cytochrome oxidase medium longer incubations led to slight staining of peroxisomes. The catalase medium was quite selective for this enzyme so that even after incubation for 120 min only peroxisomes stained.     

Localization and Characterization of Peroxidases in the Mitochondria of Chilling-Acclimated Maize Seedlings

We present evidence of two peroxidases in maize (Zea mays L.) mitochondria. One of these uses guaiacol and the other uses cytochrome c as the electron donor. Treatments of fresh mitochondria with protease(s) indicate that ascorbate and glutathione peroxidases are likely bound to the mitochondria as cytosolic contaminants, whereas guaiacol and cytochrome peroxidases are localized within the mitochondria. These two mitochondrial peroxidases are distinct from contaminant peroxidases and mitochondrial electron transport enzymes. Cytochrome peroxidase is present within the mitochondrial membranes, whereas guaiacol peroxidase is loosely bound to the mitochondrial envelope. Unlike other cellular guaiacol peroxidases, mitochondrial guaiacol peroxidase is not glycosylated. Digestion of lysed mitochondria with trypsin activated mitochondrial guaiacol peroxidase but inhibited cytochrome peroxidase. Isoelectric focusing gel analysis indicated guaiacol peroxidase as a major isozyme (isoelectric point 6.8) that is also activated by trypsin. No change in the mobility of guaiacol peroxidase due to trypsin treatment on native polyacrylamide gel electrophoresis was observed. Although both peroxidases are induced by chilling acclimation treatments (14[deg]C), only cytochrome peroxidase is also induced by chilling (4[deg]C). Because chilling induces oxidative stress in the maize seedlings and the mitochondria are a target for oxidative stress injury, we suggest that mitochondrial peroxidases play a role similar to catalase in protecting mitochondria from oxidative damage.     

Changes in enzyme activities and distributions during glucose de-repression and respiratory adaptation of anaerobically grown Saccharomyces carlsbergensis

1. During anaerobic glucose de-repression the respiration rate of whole cells of Saccharomyces carlsbergensis remained constant and was insensitive to antimycin A but was inhibited by 30% by KCN. Aeration of cells for 1 h led to increased respiration rate which was inhibited by 80% by antimycin A or KCN. 2. Homogenates were prepared from sphaeroplasts of anaerobically grown, glucose de-repressed cells and the distribution of marker enzymes was investigated after zonal centrifugation on sucrose gradients containing MgCl(2). These homogenates contained no detectable cytochrome c oxidase or catalase activity. The complex density distributions of NADH- and NADPH-cytochrome c oxidoreductases and adenosine triphosphatase(s) [ATPase(s)] were very different from those of anaerobically grown, glucose-repressed cells. 3. The specific activity of total ATPase was lowered and sensitivity to oligomycin decreased from 58 to 7% during de-repression. 4. Cytochrome c oxidase and catalase activities were detectable in homogenates of cells after 10min aeration. Zonal centrifugation indicated complex, broad sedimentable distributions of all enzyme activities assayed; the peaks of activity were at 1.27g/ml. 5. Centrifugation of homogenates of cells adapted for 30min and 3 h indicated a shift of density of the major sedimentable peak from 1.25g/ml (30min) to 1.235g/ml (3 h). After 30min adaptation a minor zone of oligomycin-sensitive ATPase and 15% of the total cytochrome c oxidase activities were detected at rho=1.12g/l; these particles together with those of higher density containing cytochrome c oxidase, ATPase and NADH-cytochrome c oxidoreductase activities were all sedimented at 10(5)g-min. 6. Electron microscopy indicated that the mitochondria-like structures of anaerobically grown, glucose-de-repressed cells were similar to those of repressed cells. After 10min of respiratory adaptation highly organized mitochondria were evident which resembled the condensed forms of mitochondria of aerobically grown, glucose-de-repressed cells. High-density zonal fractions of homogenates of cells after adaptation also contained numerous electron-dense vesicles 0.05-0.2mum in diameter. 7. The possibility that the ;promitochondria' of anaerobically grown cells may not be the direct structural precursors of fully functional mitochondria is discussed.     

Differential effect of catecholamines and MPP(+) on membrane permeability in brain mitochondria and cell viability in PC12 cells.

The present study examined the effect of dopamine, 6-hydroxydopamine (6-OHDA), and MPP(+) on the membrane permeability transition in brain mitochondria and on viability in PC12 cells. Dopamine and 6-hydroxydopamine induced the swelling and membrane potential change in mitochondria, which was inhibited by addition of antioxidant enzymes, SOD and catalase. In contrast, antioxidant enzymes did not reduce the effect of MPP(+) on mitochondrial swelling and membrane potential. Catecholamines enhanced the Ca(2+) uptake and release by mitochondria, and the addition of MPP(+) induced Ca(2+) release. Catecholamines induced a thiol oxidation in mitochondria that was decreased by antioxidant enzymes. MPP(+) showed a little effect on the cytochrome c release from mitochondria and did not induce thiol oxidation. Catecholamines and MPP(+) induced a cell death, including apoptosis, in PC12 cells that was inhibited by addition of antioxidant enzymes. The result suggests that the oxidation of dopamine and 6-hydroxydopamine could modulate the membrane permeability in brain mitochondria and induce PC12 cell death, which may be ascribed to oxidative stress. MPP(+) appears to exert a toxic effect on neuronal cells by the action, which is different from catecholamines.     

Energization of mitochondrial inner membranes caused by l-malate

It was found that 0.06 μg antimycin A/mg mitochondrial protein, an amount sufficient to inhibit electron transfer between cytochromes b and c₁ completely, fully reversed the oxidation of cytochrome a caused by L-malate in anaerobic mitochondria. The effect of L-malate on cytochrome a was insensitive to oligomycin, but all the uncouplers and detergents tested reversed the oxidation of cytochrome a caused by L-malate in anaerobic mitochondria. It was also found that addition of L-malate to anaerobic mitochondria, like addition of ATP, decreased the fluorescence of 1-anilinonaphthalene-8-sulphonate, and that subsequent addition of uncouplers reversed this effect. The effect of L-malate on the fluorescence of the dye was insensitive to oligomycin. The present findings suggest that addition of L-malate may cause energization of the mitochondrial inner membranes and that the oxidation of cytochrome a caused by L-malate in anaerobic mitochondria may result from an L-malate-induced, energy-linked reversal of electron transfer in site II.     

Enzyme activities of isolated hepatic peroxisomes from genetically lean and obese male mice.

Hepatic peroxisomes have been isolated on isopycnic sucrose gradients from white mice [HA(ICR)] and lean and obese (C57BL/6J) mice. Nearly all of the catalase activity was in the peroxisomal fraction. The activity for β-oxidation of palmitoyl-CoA was about threefold higher per milligram of protein in the isolated peroxisomal fraction or per gram of liver from the obese mouse compared to its lean littermates. Glycerol-3-phosphate dehydrogenase activity also was higher in the peroxisomes and cytoplasm of the obese mouse. The matrix enzymes of the organelles, catalase and urate oxidase of the peroxisome and glutamate dehydrogenase of the mitochondria, had similar activities per gram of liver from either lean or obese mice. Membrane components, NADPH: cytochrome c reductase of the microsomes and β-hydroxybutyrate dehydrogenase of the mitochondria, had lower activities in the obese mouse in inverse proportion to the larger size of the liver.     

Peroxisome development in the metanephric kidney of mouse.

The relationship of enzymatic activity to organelle development and organelle number during differentiation of the metanephric kidney in the mouse was approached from several experimental directions. Biochemical analyses of marker enzymes for peroxisomes (catalase and D-amino acid oxidase), mitochondria (cytochrome oxidase) and lysosomes (acid phosphatase) were performed on kidneys at ages from 17 days prenatal to adult. These data were correlated with a morphometric analysis of populations of peroxisomes and mitochondria in differentiating cells of the proximal tubule. Postnatal development of the metanephric kidney was found to be accompanied by a rapid increase in both the specific activity of catalase and the number of peroxisomes per 100 mu2 in the proximal tubule during the first 4 weeks of postnatal growth. Elaboration of the endoplasmic reticulum (ER) was seen to parallel the increase in number of peroxisomes to which segments of ER were often in close apposition. Extensive interactions between segments of ER and peroxisomes were readily visible in 0.5-mu sections viewed in the high voltage electron microscope. In contrast to peroxisomes, neither mitochondria nor lysosomes followed a similar pattern of net organelle increase, suggesting that a defined population density of mitochondria and lysosomes may exist in the proximal tubule at birth, prior to complete development of the kidney.     

Fractionation of rat fibroblasts in a zonal rotor by means of a viscosity barrier.

Cell fractionation procedures involving differential sedimentation followed by resuspension of pellets and isopycnic centrifugation are very difficult to apply to the small amounts of material available from tissue culture cells. We have explored the possibility of successive differential and isopycnic sedimentation in a zonal rotor using a short viscosity barrier for the differential sedimentation. The marker enzymes used were cytochrome oxidase, acid phosphatase, catalase, and 5′-nucleotidase. The results of these procedures are compared to the results of one-step isopycnic separations in gradients of sucrose and Stractan. The Stractan gradient was much more effective than the sucrose gradient in separating the marker enzymes from the proteins of a postnuclear supernatant, but neither type of gradient could significantly purify the marker enzymes one from another. A two-step procedure using a viscosity barrier was effective in separating particles carrying catalase from the other marker enzymes assayed and from most of the protein. A three-step procedure resulted in similar purification of mitochondria. Modification of barrier composition and centrifugation times would probably result in further improvement of separations according to individual requirements for yield, purification, and freedom from specific contamination by other subcellular particles.     

ANALYTICAL FRACTIONATION OF HOMOGENATES FROM CULTURED RAT EMBRYO FIBROBLASTS

Homogenates of cultured rat embryo fibroblasts have been assayed for acid phosphatase, N-acetyl-β-glucosaminidase, cathepsin D, acid deoxyribonuclease, cytochrome oxidase, NADH cytochrome c reductase, 5'-nucleotidase, inosine diphosphatase, acid pyrophosphatase, neutral pyrophosphatase, esterase, catalase, cholesterol, and RNA. The validity of the assay conditions was checked. Neutral pyrophosphatase is a readily soluble enzyme. Acid hydrolases, except acid pyrophosphatase, are particle-bound enzymes, which exhibit a high degree of structural latency. They are activated and solubilized in a parallel fashion by mechanical treatments and tensio-active agents. Catalase is also particle-bound and latent; activating conditions stronger than those for hydrolases are required to activate the enzyme. Acid pyrophosphatase, 5'-nucleotidase and inosine diphosphatase are firmly particle-bound, but not latent; they are not easily solubilized. In differential and isopycnic centrifugation, the latent hydrolases, cytochrome oxidase and catalase dissociate largely from each other; this suggests the occurrence of lysosomes and peroxisome-like structures besides mitochondria. The distribution patterns of 5'-nucleotidase and cholesterol are largely similar; digitonin influences their equilibrium density to the same extent; these two constituents are thought to be related to the plasma membrane. Inosine diphosphatase and acid pyrophosphatase are also partially associated with the plasma membrane, although some part of these enzymic activities probably belongs to other structures. NADH cytochrome c reductase is associated partly with the endoplasmic reticulum, partly with mitochondria.     

Kinetic and structural differences between cytochrome c oxidases from beef liver and heart

1. The cytochrome content of beef liver mitochondria differs from that of beef heart mitochondria by an eightfold lower cytochrome aa3 and a twofold lower cytochrome b and c + c1 content. 2. The kinetic properties of cytochrome c oxidases from beef liver and heart were measured with intact cytochrome c-depleted membranes, deoxycholate-dissolved membranes, and with the isolated enzymes at various cytochrome c concentrations with an oxygen electrode. Under all conditions a higher V was found for the liver enzyme, both for the low-affinity and for the high-affinity binding site for cytochrome c. Differences were also found for the Km of the two enzymes. 3. Isolated beef heart mitochondria contained about twice as much cardiolipin than beef liver mitochondria. The isolated enzymes contained one mole cardiolipin per mole of the heart enzyme, but 2 moles cardiolipin per mole of the liver enzyme. 4. By application of a high performance sodium dodecylsulfate gel electrophoretic system the two isolated enzymes could be separated into 13 different protein components, three of which (polypeptides VIa, VIIa and VIII) were found to differ in their apparent molecular weights. The functional meaning of cytochrome c oxidase isoenzymes in liver and heart is discussed.     

Subcellular Distribution of Enzymes in Ochromonas malhamensis*

SYNOPSIS. The activity and distribution of 7 enzymes in Ochromonas malhamensis were studied. Subcellular organelles were separated by centrifugation at 648,000 g min to precipitate the larger particles; the resulting supernatant was centrifuged at 5,560,000 g min to separate the microsomal fraction from the supernatant. Sixty-four percent of the cytochrome oxidase (1.9.3.1 ferrocytochrome c:oxygen oxidoreductase, 81% of the catalase (1.11.1.6 hydrogen-peroxide: hydrogen-peroxide oxidoreductase) and 70% of the urate oxidase (1.7.3.3 urate:oxygen oxidoreductase) activity was associated with the larger particles, altho only 20% of the total protein was found in this fraction. Three acid hydrolases, cathepsin (3.4.4.9 cathepsin C, acid phosphatase (3.1.3.2 orthophosphoric monoesterphosphohydrolase) and acid ribonuclease (2.7.7.17 ribonucleate nucleotido-2′-transferase) were found mostly in the supernate (50-60%, yet their latency and their similar subcellular distribution indicated the presence of lysosomes. After 2.5 hr centrifugation in a sucrose density gradient (ρ= 1.08–1.25, the acid hydrolases showed a broad distribution which differed greatly from cytochrome oxidase associated with mitochondria. Catalase, which could not be separated from cytochrome oxidase by centrifuging on this gradient, had a different distribution after centrifugation on a kinetic gradient. Urate oxidase had a similar distribution to catalase and both these enzymes were latent, indicating the presence of peroxisomes.     

A comparison of two strains of the anaerobic ciliate Trimyema compressum

Two strains of the anaerobic ciliate Trimyema compressum, isolated from different habitats, were compared. The cytoplasm of the ciliates contained hydrogenosome-like microbodies and methanogenic bacteria; the latter were lost during continued cultivation. In addition both strains harbored a non-methanogenic endosymbiont, which was lost in strain K. The ciliates lacked cytochromes, cytochrome oxidase and catalase but contained superoxide dismutase. Hydrogenase activity could be demonstrated only in strain N. In monoxenic culture strain K needed sterols as growth factors. The cells of both strains reacted similarly with respect to oxygen tolerance (up to 0.5 mg O₂/l), inhibition of growth by cyanide and azide, and resistance to antimycin A. Only cells of strain N showed growth inhibition by chloramphenicol. It is concluded that Trimyema compressum is an anaerobic, microaerotolerant organism, its microbodies show more resemblance to hydrogenosomes than to mitochondria.     

Inhibition of respiration and destruction of cytochrome a3 by light in mitochondria and cytochrome oxidase from beef heart

Irradiation of beef-heart mitochondria and of cytochrome oxidase purified from beef-heart mitochondria with blue light inhibited electron transport from substrate (succinate for the mitochondria and reduced cytochrome c for the cytochrome oxidase) to O₂. The irradiation treatment also destroyed cytochrome a₃ as assayed by the absorption band for the reduced cyanide-cytochrome a₃ complex at 587 nm in the low-temperature absorption spectrum. Irradiation under anaerobic conditions was not inhibitory. Cytochrome a₃ was protected against photodestruction if cyanide was present during the irradiation.     

Cytochemical analysis of organelle degradation in phagosomes and apoptotic cells of the mucoid epithelium of mice.

The activity of mitochondrial cytochrome oxidase and peroxisomal catalase in the phagolysosomes and apoptotic bodies of mucoid epithelial cells was analysed. Tissue from 2-6 day old mice was used. The activity of acid phosphatase in lysosomes was also estimated. Cytochrome oxidase was demonstrated in well-preserved mitochondria inside phagosomes. Mitochondria in cells exhibiting apoptotic death also show activity of cytochrome oxidase. The enzyme activity in swollen mitochondria ceases before the membranes of the cristae disappear completely. Apoptotic bodies are phagocytosed by sister mucoid cells and, later on, they are digested inside the cell. Phagosomes which contain already degraded mitochondria show still active catalase in sequestered peroxisomes. The acid phosphatase involved in degradation of phagocytosed material originates from endocytosed lysosomes and primary and secondary lysosomes which fuse with the membranes of phagosomes.     

OXIDATIVE AND HYDROLYTIC ENZYMES IN THE NEPHRON OF NECTURUS MACULOSUS : Histochemical, Biochemical, and Electron Microscopical Studies

The distribution of oxidative and hydrolytic enzyme activities along the nephron of Necturus maculosus Rafinesque was studied histochemically. The proximal tubule possessed all the demonstrable enzyme activities associated with the hexose-monophosphate shunt and glycolysis, but lacked detectable succinic dehydrogenase and cytochrome oxidase activities. Krebs cycle enzymes other than succinic dehydrogenase were easily detectable. The distal tubule, on the other hand, possessed no detectable hexose-monophosphate shunt enzyme activities, but all demonstrable glycolytic and Krebs cycle enzymes and cytochrome oxidase were present in high activity. These data indicate that the proximal tubule of Necturus probably cannot depend, as can the distal tubule, on the Krebs cycle and cytochrome system to provide energy for its transport processes, an inference supported, in general, by available physiological evidence. The question of the importance of the hexose shunt to proximal tubular function arises. Evidence is presented that the proximal tubular blood supply is primarily venous in nature, a hypothesis which would correlate well with its anaerobic metabolic pattern. In addition, the absence of cytochrome oxidase and succinic dehydrogenase from the proximal tubular cells implies either that they possess very few mitochondria, or that their mitochondria have a very unusual enzymatic pattern. Electron microscopical observations and data obtained from the measurement of the enzyme activities of homogenates of Necturus kidney are presented which indicate that the second hypothesis is more probably correct.     

Effects of Ethylenediaminetetraacetate and Chloramphenicol on Mitochondrial Activity and Morphogenesis in Mucor rouxii

The present study demonstrates the importance of mitochondrial activities in controlling Mucor rouxii morphogenesis. The respiratory capacity of the spores of this facultatively anaerobic, dimorphic fungus becomes repressed if germination and growth take place in the absence of oxygen. The level of activity of mitochondrial enzymes such as cytochrome oxidase and malate dehydrogenase is lower in the anaerobic yeastlike cells than it is in ungerminated spores and in aerobic hyphae, but the reverse is true for glycolytic enzymes such as pyruvate kinase and alcohol dehydrogenase. Following exposure to air, yeastlike cells convert into hyphae after a lag period corresponding to aerobic adaptation. Anaerobic cultures grown in the presence of ethylenediaminetetraacetate (EDTA) at a concentration of 10(-4) M exhibit hyphal morphology. These cells, which are fully adapted to anaerobic fermentation, nevertheless have potentially active mitochondria with the same levels of respiratory enzymes as ungerminated spores. These cells are able to grow immediately after aeration, without an adaptation lag. Evidence is presented which indicates that the morphogenetic effect of EDTA is not the result of elimination of free metals. Additional evidence proving mitochondrial control of morphogenesis in M. rouxii is that chloramphenicol (4 mg/ml) induced the formation of respiratory-deficient, yeastlike cells in aerobic cultures.     

Effect of R-(-)-Deprenyl and Harmaline on Dopamine- and Peroxynitrite-Induced Membrane Permeability Transition in Brain Mitochondria

The present study examined the effect of MAO inhibitors, deprenyl and harmaline, on the membrane permeability transition in brain mitochondria. Deprenyl, harmaline, and antioxidant enzymes (SOD and catalase) attenuated alteration of the swelling, membrane potential, cytochrome c release, and Ca²⁺ transport in mitochondria treated with dopamine. In contrast, deprenyl and harmaline did not reduce the peroxynitrite-induced change in membrane permeability. Deprenyl and harmaline inhibited the decrease in thioredoxin reductase activity and the thiol oxidation in mitochondria treated with dopamine but did not decrease the effect of peroxynitrite. Deprenyl and harmaline significantly decreased the formation of melanin from dopamine. The results suggest that deprenyl and harmaline may protect brain mitochondria against the toxic action of dopamine oxidation by the maintenance of thioredoxin reductase activity, inhibition of thiol oxidation, and inhibition of dopamine oxidation product formation. In contrast, MAO inhibitors may not defend brain mitochondria against damaging action of peroxynitrite.