Significance of plant sulfite oxidase

Sulfite oxidizing activities are known since years in animals, microorganisms, and also plants. Among plants, the only enzyme well characterized on molecular and biochemical level is the molybdoenzyme sulfite oxidase (SO). It oxidizes sulfite using molecular oxygen as electron acceptor, leading to the production of sulfate and hydrogen peroxide. The latter reaction product seems to be the reason why plant SO is localized in peroxisomes, because peroxisomal catalase is able to decompose hydrogen peroxide. On the other hand, we have indications for an additional reaction taking place in peroxisomes: sulfite can be nonenzymatically oxidized by hydrogen peroxide. This will promote the detoxification of hydrogen peroxide especially in the case of high amounts of sulfite. Hence we assume that SO could possibly serve as "safety valve" for detoxifying excess amounts of sulfite and protecting the cell from sulfitolysis. Supportive evidence for this assumption comes from experiments where we fumigated transgenic poplar plants overexpressing ARABIDOPSIS SO with SO(2) gas. In this paper, we try to explain sulfite oxidation in its co-regulation with sulfate assimilation and summarize other sulfite oxidizing activities described in plants. Finally we discuss the importance of sulfite detoxification in plants.     

Tryptic cleavage of rat liver sulfite oxidase. Isolation and characterization of molybdenum and heme domains.

Treatment of rat liver sulfite oxidase with trypsin leads to loss of ability to oxidize sulfite in the presence of cytochrome c as electron acceptor. Ability to oxidize sulfite with ferricyanide as acceptor is undiminished, while sulfite leads to O2 activity is partially retained. Gel filtration of the proteolytic products has led to the isolation of two major fragments of dissimilar size derived from sulfite oxidase. The smaller fragment has a molecular weight of 9500 and appears to be monomeric when detached from sulfite oxidase. It contains the heme in its cytochrome b5 structure, has no sulfite oxidase activity, and is reducible with dithionite but not with sulfite. The heme fragment can mediate electron transfer between pig liver microsomal NADH cytochrome b5 reductase and cytochrome c. The larger fragment has a molecular weight of 47,400 under denaturing conditions but elutes from Sephadex G-200 as a dimer. It contains no heme but retains all of the molybdenum and the modified sulfite-oxidizing capacity present in the proteolytic mixture. All of the EPR properties of the molybdenum center of native sulfite oxidase are retained in the molybdenum fragment. The molybdenum center is a weak chromophore with an absorption sectrum suggestive of coordination with sulfur ligands. Reduction by sulfite generates a spectrum attributable to molybdenum (V). Spectra of oxidized and sulfite-reduced preparations are sensitive to anions and pH. NH2-terminal analysis of native sulfite oxidase and the two tryptic fragments has permitted the conclusion that the sequence represented by the heme fragment is the NH2 terminus of native enzyme. These studies have demonstrated that the two cofactor moieties of sulfite oxidase are contained in distinct domains which are covalently held in contiguity by means of an exposed hinge region. Isolation of functional heme and molybdenum domains of sulfite oxidase after tryptic cleavage has demonstrated conclusively that the cytochrome b5 region of the molecule is required for electron transfer to the physiological acceptor, cytochrome c.     

Oxidative Half-reaction of Arabidopsis thaliana Sulfite Oxidase: GENERATION OF SUPEROXIDE BY A PEROXISOMAL ENZYME*

Vertebrate forms of the molybdenum-containing enzyme sulfite oxidase possess a b-type cytochrome prosthetic group that accepts reducing equivalents from the molybdenum center and passes them on to cytochrome c. The plant form of the enzyme, on the other hand, lacks a prosthetic group other than its molybdenum center and utilizes molecular oxygen as the physiological oxidant. Hydrogen peroxide is the ultimate product of the reaction. Here, we present data demonstrating that superoxide is produced essentially quantitatively both in the course of the reaction of reduced enzyme with O₂ and during steady-state turnover and only subsequently decays (presumably noncatalytically) to form hydrogen peroxide. Rapid-reaction kinetic studies directly following the reoxidation of reduced enzyme demonstrate a linear dependence of the rate constant for the reaction on [O₂] with a second-order rate constant of k_ox = 8.7 × 10⁴ ± 0.5 × 10⁴ m⁻¹s⁻¹. When the reaction is carried out in the presence of cytochrome c to follow superoxide generation, biphasic time courses are observed, indicating that a first equivalent of superoxide is generated in the oxidation of the fully reduced Mo(IV) state of the enzyme to Mo(V), followed by a slower oxidation of the Mo(V) state to Mo(VI). The physiological implications of plant sulfite oxidase as a copious generator of superoxide are discussed.     

Sulfite oxidase from chicken liver. The role of imidazole and carboxyl groups for the reaction with cytochrome c

Oxidation of sulfite to sulfate by sulfite oxidase is inhibited when the enzyme is treated with reagents known to modify imidazole and carboxyl groups. Modification inhibits the oxidation of sulfite by the physiological electron acceptor cytochrome c, but not by the artificial acceptor ferricyanide. This indicates interference with reaction steps that follow the oxidation of sulfite by the enzyme's molybdenum cofactor. Reaction with diethylpyrocarbonate modifies ten histidines per enzyme monomer. Loss of activity is concomitant to the modification of only a single histidine residue. Inactivation takes place at the same rate in free sulfite oxidase and in the sulfite-oxidase--cytochrome-c complex. Blocking of carboxyl groups with water-soluble carbodiimides inactivates the enzyme. But none of the enzyme's carboxyl groups seems to be essential in the sense that its modification fully abolishes activity. The pattern of inactivation by chemical modification of sulfite oxidase is quite similar to that observed previously for cytochrome c peroxidase from yeast [Bosshard, H. R., B?nziger, J., Hasler, T. and Poulos, T. L. (1984) J. Biol. Chem. 259, 5683-5690; Bechtold, R. and Bosshard, H. R. (1985) J. Biol. Chem. 260, 5191-5200]. The two enzymes have very different structures yet share cytochrome c as a common substrate of which they recognize the same electron-transfer domain around the exposed heme edge.     

The role of tyrosine 343 in substrate binding and catalysis by human sulfite oxidase

In the crystal structure of chicken sulfite oxidase, the residue Tyr(322) (Tyr(343) in human sulfite oxidase) was found to directly interact with a bound sulfate molecule and was proposed to have an important role in mediating the substrate specificity and catalytic activity of this molybdoprotein. In order to understand the role of this residue in the catalytic mechanism of sulfite oxidase, steady-state and stopped-flow analyses were performed on wild-type and Y343F human sulfite oxidase over the pH range 6-10. In steady-state assays of Y343F sulfite oxidase using cytochrome c as the electron acceptor, k(cat) was somewhat impaired ( approximately 34% wild-type activity at pH 8.5), whereas the K(m)(sulfite) showed a 5-fold increase over wild type. In rapid kinetic assays of the reductive half-reaction of wild-type human sulfite oxidase, k(red)(heme) changed very little over the entire pH range, with a significant increase in K(d)(sulfite) at high pH. The k(red)(heme) of the Y343F variant was significantly impaired across the entire pH range, and unlike the wild-type protein, both k(red)(heme) and K(d)(sulfite) were dependent on pH, with a significant increase in both kinetic parameters at high pH. Additionally, reduction of the molybdenum center by sulfite was directly measured for the first time in rapid reaction assays using sulfite oxidase lacking the N-terminal heme-containing domain. Reduction of the molybdenum center was quite fast (k(red)(Mo) = 972 s(-1) at pH 8.65 for wild-type protein), indicating that this is not the rate-limiting step in the catalytic cycle. Reduction of the molybdenum center of the Y343F variant by sulfite was more significantly impaired at high pH than at low pH. These results demonstrate that the Tyr(343) residue is important for both substrate binding and oxidation of sulfite by sulfite oxidase.     

Identification and biochemical characterization of Arabidopsis thaliana sulfite oxidase. A new player in plant sulfur metabolism

In mammals and birds, sulfite oxidase (SO) is a homodimeric molybdenum enzyme consisting of an N-terminal heme domain and a C-terminal molybdenum domain (EC ). In plants, the existence of SO has not yet been demonstrated, while sulfite reductase as part of sulfur assimilation is well characterized. Here we report the cloning of a plant sulfite oxidase gene from Arabidopsis thaliana and the biochemical characterization of the encoded protein (At-SO). At-SO is a molybdenum enzyme with molybdopterin as an organic component of the molybdenum cofactor. In contrast to homologous animal enzymes, At-SO lacks the heme domain, which is evident both from the amino acid sequence and from its enzymological and spectral properties. Thus, among eukaryotes, At-SO is the only molybdenum enzyme yet described possessing no redox-active centers other than the molybdenum. UV-visible and EPR spectra as well as apparent K(m) values are presented and compared with the hepatic enzyme. Subcellular analysis of crude cell extracts showed that SO was mostly found in the peroxisomal fraction. In molybdenum cofactor mutants, the activity of SO was strongly reduced. Using antibodies directed against At-SO, we show that a cross-reacting protein of similar size occurs in a wide range of plant species, including both herbacious and woody plants.     

Effects of large-scale amino acid substitution in the polypeptide tether connecting the heme and molybdenum domains on catalysis in human sulfite oxidase

Sulfite oxidase (SO) is a molybdenum-cofactor-dependent enzyme that catalyzes the oxidation of sulfite to sulfate, the final step in the catabolism of the sulfur-containing amino acids, cysteine and methionine. The catalytic mechanism of vertebrate SO involves intramolecular electron transfer (IET) from molybdenum to the integral b-type heme of SO and then to exogenous cytochrome c. However, the crystal structure of chicken sulfite oxidase (CSO) has shown that there is a 32 ? distance between the Fe and Mo atoms of the respective heme and molybdenum domains, which are connected by a flexible polypeptide tether. This distance is too long to be consistent with the measured IET rates. Previous studies have shown that IET is viscosity dependent (Feng et al., Biochemistry, 2002, 41, 5816) and also dependent upon the flexibility and length of the tether (Johnson-Winters et al., Biochemistry, 2010, 49, 1290). Since IET in CSO is more rapid than in human sulfite oxidase (HSO) (Feng et al., Biochemistry, 2003, 42, 12235) the tether sequence of HSO has been mutated into that of CSO, and the resultant chimeric HSO enzyme investigated by laser flash photolysis and steady-state kinetics in order to study the specificity of the tether sequence of SO on the kinetic properties. Surprisingly, the IET kinetics of the chimeric HSO protein with the CSO tether sequence are slower than wildtype HSO. This observation raises the possibility that the composition of the non-conserved tether sequence of animal SOs may be optimized for individual species.     

Radiation inactivation of hepatic sulfite oxidase

Sulfite oxidase (EC 1.8.3.1), purified from chicken liver, is comprised of two identical subunits of 55 kDa, each of which contains a molybdenum and heme prosthetic group. The functional size of sulfite oxidase was determined by radiation inactivation analysis using both full, sulfite:cytochrome c reductase, and partial, sulfite:ferricyanide reductase, catalytic activities. Inactivation of full enzyme activity indicated a target size of 42 kDa while the partial activity indicated a target size of 25 kDa. These results confirm the earlier findings of two equivalent subunits and suggest the presence of a functional domain within the subunit structure that contains the molybdenum center and exhibits a smaller molecular mass than that of the enzyme subunit.     

The formation of hydrogen peroxide during the oxidation of reduced nicotinamide adenine dinucleotide by cytochrome o from Vitreoscilla.

The formation of hydrogen peroxide during the oxidation of NADH by purified preparations of cytochrome o has been demonstrated by employing three independent methods: polarographic, colorimetric, and fluorometric. The first two methods were used to assay for the accumulation of hydrogen peroxide and showed that hydrogen peroxide did accumulate as a product, but only about 30% of the oxygen consumed or 15 to 20% of the NADH oxidized was recoverable as hydrogen peroxide. This lack of 1:1 stoichiometry was not due to residual catalase activity in these preparations which could be eliminated by freeze-thawing. Thus, hydrogen peroxide may not be the sole or primary product of the NADH-cytochrome o oxidase reaction. The fluorometric assay could be coupled directly to the NADH-cytochrome o oxidase reaction in one medium, and this method showed that hydrogen peroxide was generated continuously from the beginning of the reaction in a 1:1 stoichiometry, hydrogen peroxide generated to NADH oxidized. This result suggests that hydrogen peroxide is an intermediate that can be trapped efficiently under the conditions of the fluorometric assay, whereas under the conditions of the first two assays most of the hydrogen peroxide generated undergoes further reaction. Exogenously added FAD or FMN increased the percentage of hydrogen peroxide that accumulated in the NADHcytochrome o oxidase reaction. Flavin is believed to act on the reductase side of cytochrome o so the increased percentage of hydrogen peroxide is not likely to result from the direct reaction of reduced flavin with oxygen.     

Cytochrome c oxidase: catalytic cycle and mechanisms of proton pumping--a discussion

Cytochrome c oxidase catalyzes the reduction of molecular oxygen to water, a process in which four electrons, four protons, and one molecule of oxygen are consumed. The reaction is coupled to the pumping of four additional protons across the membrane. According to the currently accepted concept, the pumping of all four protons occurs after the binding of oxygen to the reduced enzyme and is exclusively coupled to the last two electron transfer steps. A careful analysis of the existing data shows that there is no experimental evidence for this paradigm. It is more likely that only three protons are pumped during the second half of the catalytic cycle of cytochrome c oxidase after the reaction with oxygen. In this article a variant of a recent mechanistic model of proton pumping by electrostatic repulsion is discussed. It is based on the electroneutrality principle in a way that in the catalytic cycle each electron transfer to the membrane-embedded electron acceptors is charge-compensated by uptake of one proton. The mechanism takes into account the findings with mutant cytochrome c oxidases and explains the results of many recent experiments, including the effects of hydrogen peroxide.     

Oxidation of molybdopterin in sulfite oxidase by ferricyanide. Effect on electron transfer activities

The attenuation of the sulfite:cytochrome c activity of sulfite oxidase upon treatment with ferricyanide was demonstrated to be the result of oxidation of the pterin ring of the molybdenum cofactor in the enzyme. Oxidation of molybdopterin (MPT) was detected in several ways. Ferricyanide treatment not only abolished the ability of sulfite oxidase to serve as a source of MPT to reconstitute the aponitrate reductase in extracts of the Neurospora crassa mutant nit-1 but also eliminated the ability of sulfite oxidase to reduce dichlorobenzenoneindophenol after anaerobic denaturation. Additionally, the absorption spectrum of anaerobically denatured ferricyanide-treated molybdenum fragment of rat liver sulfite oxidase was typical of fully oxidized pterins. Ferricyanide treatment had no effect on the protein of sulfite oxidase or on the sulfhydryl-containing side chain of MPT. Quantitation of the ferricyanide reaction showed that 2 mol of ferricyanide were reduced per mol of MPT oxidized, yielding a fully oxidized pterin. These results corroborate the previously reported conclusion that the native state of reduction of MPT in sulfite oxidase is at the dihydro level (Gardlik, S., and Rajagopalan, K.V. (1990) J. Biol. Chem. 265, 13047-13054). As a result of oxidation of the pterin ring, the affinity of MPT for molybdenum is decreased, leading to eventual loss of molybdenum. Because the loss of molybdenum is slow, a population of sulfite oxidase molecules can exist in which molybdenum is complexed to oxidized MPT. These molecules retain sulfite:O2 activity, a function apparently dependent solely on the molybdenum-thiolate complex, yet have greatly decreased sulfite:cytochrome c activity, a function requiring heme as well as the molybdenum center of holoenzyme. These observations suggest that the pterin ring of MPT participates in enzyme function, possibly in electron transfer, directly in catalysis, or by controlling the oxidation/reduction potential of molybdenum.     

Copper Proteins and Oxygen : Correlations between structure and function of the copper oxidases

A comprehensive survey of the interaction of the copper proteins and oxygen is presented including a correlation of structure, function, and other properties of the known copper oxidases and of hemocyanin. The origin of their blue color and the structure of copper complexes and copper proteins are related to the oxidation state of copper ion and relevant electronic transitions probably arising from the formation of charge transfer complexes. The oxygen reactions of hemocyanin, ceruloplasmin, and cytochrome oxidase show half-saturation values far below the other Cu enzymes. The formation of hydrogen peroxide as a reaction product is associated with the presence of one Cu atom per oxidase molecule or catalytic system. Water is the corresponding product of the other Cu oxidases with four or more Cu atoms per molecule, except for monoamine oxidase. Mechanisms for the oxidase action of the two and four electron transfer Cu oxidases and tyrosinase are proposed. These reactions account for the number, the oxidation-reduction potential, and the oxidation state of Cu in the resting enzyme, the cyclical change from Cu(II) to Cu(I), the diatomic nature of O₂, the sequence of the oxidation and reduction reactions, and other salient features. The catalytic reactions involved in the oxidation of ascorbic acid by plant ascorbate oxidase, ceruloplasmin, and Cu(II) are compared. Finally the substrate specificity, inhibitory control, and the detailed mechanism of the oxidase activity of ceruloplasmin are summarized.     

Kinetic investigations of the reactions of cytochrome c oxidase with hydrogen peroxide

The reaction of H2O2 with reduced cytochrome c oxidase was investigated with rapid-scan/stopped-flow techniques. The results show that the oxidation rate of cytochrome a3 was dependent upon the peroxide concentration (k = 2 X 10(4) M-1 X s-1). Cytochrome a and CuA were oxidised with a maximal rate of approx. 20 s-1, indicating that the rate of internal electron transfer was much slower with H2O2 as the electron acceptor than with O2 (k greater than or equal to 700 s-1). Although other explanations are possible, this result strongly suggests that in the catalytic cycle with oxygen as a substrate the internal electron-transfer rate is enhanced by the formation of a peroxo-intermediate at the cytochrome a3-CuB site. It is shown that H2O2 took up two electrons per molecule. The reaction of H2O2 with oxidised cytochrome c oxidase was also studied. It is shown that pulsed oxidase readily reacted with H2O2 (k approximately 700 M-1 X s-1). Peroxide binding is followed by an H2O2-independent conformational change (k = 0.9 s-1). Resting oxidase partially bound H2O2 with a rate similar to that of pulsed oxidase; after H2O2 binding the resting enzyme was converted into the pulsed conformation in a peroxide-independent step (k = 0.2 s-1). Within 5 min, 55% of the resting enzyme reacted in a slower process. We conclude from the results that oxygenated cytochrome c oxidase probably is an enzyme-peroxide complex.     

Mechanism of cytochrome c peroxidase. O-benzoylhydroxylamine as an analog of hydrogen peroxide.

A number of reagents, some of which are electronic analogs of hydrogen peroxide, will replace it in the reactions of cytochrome c peroxidase. These compounds include N-bromosuccinimide, sodium hypochlorite, and the novel oxidizing agent O-benzoylhydroxylamine. If fragments of the oxidant played a functional role in the structure of the oxidized form of the enzyme, it would be expected that the product formed from O-benzoylhydroxylamine would differ from that formed from hydrogen peroxide. The products formed on reaction of the two oxidizing agents with cytochrome c peroxidase are indistinguishable. This results carries implications for the structure of the so-called ES compound. The extension in the range of specific substrates for cytochrome c peroxidase allows identification of the structure which compounds must possess to be oxidizing substrates for the enzyme. A mechanism for the first step of the reaction is suggested. O-Benzoylhydroxylamine is also a reducing agent, and its reaction with the enzyme is analogous to that of hydrogen peroxide with catalase. The final product of the reaction is the inert nitric oxide complex of ferrous cytochrome c peroxidase.     

Existence of a new type of sulfite oxidase which utilizes ferric ions as an electron acceptor in Thiobacillus ferrooxidans

A new type of sulfite oxidase which utilizes ferric ion (Fe3+) as an electron acceptor was found in iron-grown Thiobacillus ferrooxidans. It was localized in the plasma membrane of the bacterium and had a pH optimum at 6.0. Under aerobic conditions, 1 mol of sulfite was oxidized by the enzyme to produce 1 mol of sulfate. Under anaerobic conditions in the presence of Fe3+, sulfite was oxidized by the enzyme as rapidly as it was under aerobic conditions. In the presence of o-phenanthroline or a chelator for Fe2+, the production of Fe2+ was observed during sulfite oxidation by this enzyme under not only anaerobic conditions but also aerobic conditions. No Fe2+ production was observed in the absence of o-phenanthroline, suggesting that the Fe2+ produced was rapidly reoxidized by molecular oxygen. Neither cytochrome c nor ferricyanide, both of which are electron acceptors for other sulfite oxidases, served as an electron acceptor for the sulfite oxidase of T. ferrooxidans. The enzyme was strongly inhibited by chelating agents for Fe3+. The physiological role of sulfite oxidase in sulfur oxidation of T. ferrooxidans is discussed.     

Existence of a new type of sulfite oxidase which utilizes ferric ions as an electron acceptor in Thiobacillus ferrooxidans.

A new type of sulfite oxidase which utilizes ferric ion (Fe3+) as an electron acceptor was found in iron-grown Thiobacillus ferrooxidans. It was localized in the plasma membrane of the bacterium and had a pH optimum at 6.0. Under aerobic conditions, 1 mol of sulfite was oxidized by the enzyme to produce 1 mol of sulfate. Under anaerobic conditions in the presence of Fe3+, sulfite was oxidized by the enzyme as rapidly as it was under aerobic conditions. In the presence of o-phenanthroline or a chelator for Fe2+, the production of Fe2+ was observed during sulfite oxidation by this enzyme under not only anaerobic conditions but also aerobic conditions. No Fe2+ production was observed in the absence of o-phenanthroline, suggesting that the Fe2+ produced was rapidly reoxidized by molecular oxygen. Neither cytochrome c nor ferricyanide, both of which are electron acceptors for other sulfite oxidases, served as an electron acceptor for the sulfite oxidase of T. ferrooxidans. The enzyme was strongly inhibited by chelating agents for Fe3+. The physiological role of sulfite oxidase in sulfur oxidation of T. ferrooxidans is discussed.     

Evaluation of relative contributions of two enzymes supposed to metabolise hydrogen peroxide in Paracoccus denitrificans

A biosensor exploiting an electrochemically mediated enzyme-catalysed reaction was used to quantify relative contributions of cytoplasmic catalase and periplasmic cytochrome c peroxidase to the overall rate of hydrogen peroxide breakdown in cells of Paracoccus denitrificans. The effects of antimycin (an inhibitor of electron flow to cytochrome c peroxidase), the reaction rate versus substrate concentration profiles for the whole cells and subcellular fractions, and the time courses of oxygen concentration demonstrated a profound decrease in the capacity of cytochrome c peroxidase to reduce H2O2 under in vivo conditions. The reason is suggested to be a competition for available electrons between the enzyme and terminal oxidases metabolising oxygen produced by catalase.     

Flavocytochrome b2: reactivity of its flavin with molecular oxygen

Flavocytochrome b2, a flavohemoprotein, catalyzes the oxidation of lactate at the expense of the physiological acceptor cytochrome c in the yeast mitochondrial intermembrane space. The mechanism of electron transfer from the substrate to monoelectronic acceptors via FMN and heme b2 has been intensively studied over the years. Each prosthetic group is bound to a separate domain, N-terminal for the heme, C-terminal for the flavin. Each domain belongs to a distinct evolutionary family. In particular, the flavodehydrogenase domain is homologous to a number of well-characterized l-2-hydroxy acid-oxidizing enzymes. Among these, some are oxidases for which the oxidative half-reaction produces hydrogen peroxide at the expense of oxygen. For bacterial mandelate dehydrogenase and flavocytochrome b2, in contrast, the oxidative half-reaction requires monoelectronic acceptors. Several crystal structures indicate an identical fold and a highly conserved active site among family members. All these enzymes form anionic semiquinones and bind sulfite, properties generally associated with oxidases, whereas electron transferases are expected to form neutral semiquinones and to yield superoxide anion. Thus, flavocytochrome b2 is a highly unusual dehydrogenase-electron transferase, and one may wonder how its flavin reacts with oxygen. In this work, we show that the separately engineered flavodehydrogenase domain produces superoxide anion in its slow reaction with oxygen. This reaction apparently also takes place in the holoenzyme when oxygen is the sole electron acceptor, because the heme domain autoxidation is also slow; this is not unexpected, in view of the heme domain mobility relative to the tetrameric flavodehydrogenase core (Xia, Z. X., and Mathews, F. S. (1990) J. Mol. Biol. 212, 837-863). Nevertheless, this reaction is so slow that it cannot compete with the normal electron flow in the presence of monoelectronic acceptors, such as ferricyanide and cytochrome c. An inspection of the available structures of family members does not provide a rationale for the difference between the oxidases and the electron transferases.     

Xanthine oxidase/hydrogen peroxide generates sulfur trioxide anion radical (SO3.-) from sulfite (SO3(2-)).

In the presence of hydrogen peroxide (H2O2), xanthine oxidase has been found to catalyze sulfur trioxide anion radical (SO3.-) formation from sulfite anion (SO3(2-)). The SO3.- radical was identified by ESR (electron spin resonance) spin trapping, utilizing 5,5-dimethyl-l-pyrroline-l-oxide (DMPO) as the spin trap. Inactivated xanthine oxidase does not catalyze SO3.- radical formation, implying a specific role for this enzyme. The initial rate of SO3.- radical formation increases linearly with xanthine oxidase concentration. Together, these observations indicate that the SO3.- generation occurs enzymatically. These results suggest a new property of xanthine oxidase and perhaps also a significant step in the mechanism of sulfite toxicity in cellular systems.