Plant Sulfite Oxidase as Novel Producer of H2O2

Sulfite oxidase (EC 1.8.3.1) from the plant Arabidopsis thaliana is the smallest eukaryotic molybdenum enzyme consisting of a molybdenum cofactor-binding domain but lacking the heme domain that is known from vertebrate sulfite oxidase. While vertebrate sulfite oxidase is a mitochondrial enzyme with cytochrome c as the physiological electron acceptor, plant sulfite oxidase is localized in peroxisomes and does not react with cytochrome c. Here we describe results that identified oxygen as the terminal electron acceptor for plant sulfite oxidase and hydrogen peroxide as the product of this reaction in addition to sulfate. The latter finding might explain the peroxisomal localization of plant sulfite oxidase. 18O labeling experiments and the use of catalase provided evidence that plant sulfite oxidase combines its catalytic reaction with a subsequent non-enzymatic step where its reaction product hydrogen peroxide oxidizes another molecule of sulfite. In vitro, for each catalytic cycle plant SO will bring about the oxidation of two molecules of sulfite by one molecule of oxygen. In the plant, sulfite oxidase could be responsible for removing sulfite as a toxic metabolite, which might represent a means to protect the cell against excess of sulfite derived from SO2 gas in the atmosphere (acid rain) or during the decomposition of sulfur-containing amino acids. Finally we present a model for the metabolic interaction between sulfite and catalase in the peroxisome.

Sulfite oxidases (SO) 3 from vertebrates (published as EC 1.8.3.1) play an essential role in sulfur detoxification by catalyzing the reaction SO 3 2Ϫ ϩ H 2 O ϩ 2 (cyt c) ox 3 SO 4 2Ϫ ϩ 2H ϩ ϩ 2 (cyt c) red (1), which is the terminal step in the oxidative degradation of cysteine and methionine. Different electron acceptors were reported to interact with vertebrate SO including cytochrome c, ferricyanide, and oxygen (2)(3)(4). In mammals, SO is localized in the intermembrane space of mitochondria (5) where electrons derived from sulfite are passed to cytochrome c, the physiological electron acceptor. Vertebrate SO is a homodimeric protein with monomers subdivided into a Moco domain and a heme domain, as verified by the atomic structure of chicken SO (6).
Recently we have described plant SO (7) from Arabidopsis thaliana, which is the fourth molybdenum enzyme present in plants in addition to nitrate reductase, xanthine dehydrogenase, and aldehyde oxidase. Cloning and characterization of plant SO was possible by using sequence homologies to the mammalian counterpart. However, in contrast to the animal enzyme plant SO lacks the heme domain, which is evident from the amino acid sequence, its enzymological and spectral properties (7), and the atomic structure (8). Also its subcellular localization differs from that of animals, in plants we showed SO to be localized in peroxisomes (9). SO is wide spread and highly conserved within the plant kingdom; the SO gene is present in higher and lower plants, and the protein encoded seems to be highly conserved because antibodies directed against Arabidopsis SO detect proteins of the correct size in a wide range of herbaceous and also woody plants (7). Obviously, SO has an important function if it was conserved during evolution. The major function of plant SO seems to be different from the role it plays in animals where it oxidizes sulfite derived from the decomposition of sulfur-containing amino acids. In plants, sulfur metabolism is different because plants assimilate sulfur by reducing sulfate via sulfite to sulfide to form cysteine (for recent reviews, see Leustek (10) and Saito (11)). This sulfur assimilation pathway is localized in chloroplasts. How do we unravel the role of plant SO? Previous data showed that plants subjected to sulfur-labeled SO 2 gas readily converted the radioactive label into sulfate (12,13), yet the corresponding enzyme catalyzing this reaction was not characterized at that time. Most importantly, this reaction was interpreted as the key step ("safety valve") in protecting plants from acid rain that has been one major cause of forest damage (14). Excess of sulfite derived from SO 2 gas is detrimental for the cell because sulfite is toxic. Now that plant SO is isolated and characterized (7)(8)(9), it is important to study the special properties that make SO a unique catalyst and that were responsible for conserving SO during evolution among plants. The clue to understand the role of SO seems to lie in its peroxisomal localization, either because only there the substrate is available or because in this organelle the reaction product could be further metabolized. SO 2 gas can readily penetrate through biomembranes (15), and sulfite is generated during sulfate assimilation in the chloroplasts. Hence substrate availability is no convincing argument for explaining peroxisomal localization of SO. However, peroxisomes are known as the compartment in which reactive oxygen is produced, and therefore, identifying both the electron acceptor and the reaction products could be the clue for describing the special properties that make plant SO unique.
Here we identified molecular oxygen as the terminal electron acceptor for plant SO and showed that it converts molecular oxygen into hydrogen peroxide. Furthermore we provide evidence that this hydrogen peroxide is able to non-enzymatically oxidize sulfite. Thus, under in vitro conditions plant SO combines both an enzymatic and a non-enzymatic reaction.

EXPERIMENTAL PROCEDURES
Protein Expression and Purification-For overexpression and purification of recombinant A. thaliana SO the QIAexpress kit (Qiagen) for metal ion chelating chromatography of His-tagged proteins was used. The protein was expressed and purified from pQE80-At-sox in Escherichia coli strain TP1000 (genotype ⌬mobAB) as described earlier (7). Purification of soluble rAt-SO was performed at 4°C by using nickelnitrilotriacetic acid superflow matrix (Qiagen) followed by anion exchange chromatography on a 10 ml SourceQ15 column (Amersham Biosciences). Moco-containing SO did not bind to the column, whereas the inactive Moco-free form of SO bound to the column and eluted between 250 and 300 mM NaCl (8). The protein preparation was stable at 4°C for several weeks after dialyzing and concentrating to 2-10 mg/ml. Concentration of recombinant plant SO was determined by UV spectroscopy using the calculated extinction coefficient of ⑀ ϭ 69.820 Detection of Plant SO Protein-Protein of leaves and roots were extracted and analyzed as described earlier (7). For immuno-detection, protein was separated on a 10% SDS-PAGE, and the polyclonal anti-rAt-SO antibody (Eurogentec) was diluted 1:500 prior to use.
SO Assays-SO was routinely assayed for the activity to reduce ferricyanide as described previously (7). In parallel, oxygen measurements were performed with an inoLab OxiLevel2 electrode (Wissenschaftlich-Technische Werkstätten GmbH) in 2.2-ml volume by using 2 g of recombinant plant SO, oxygen-saturated 0.1 M Tris acetate buffer (pH 7.25) over a time of 2 min and expressed as oxygen consumption (mg/ liter) per second. In the mixed SO ϩ catalase assay, 2 g of recombinant plant SO and 94 units of bovine catalase (Sigma) were incubated in 0.1 M Tris acetate buffer. For quantification of sulfite and sulfate used and produced by recombinant plant SO, respectively, the reaction was stopped by heating at 95°C for 5 min, and sulfite and sulfate were determined as described below. All data presented are replicates from at least three independent experiments and expressed as mean Ϯ S.D.
Detection of Hydrogen Peroxide-Determination of H 2 O 2 formation was performed according to Nag et al. (16) with TiOSO 4 . SO reaction (2 g in 900 l) was stopped at different time points by adding 100 l of the TiOSO 4 stock solution (0.2 g of TiOSO 4 in 5 ml of 95% H 2 SO 4 and diluted to a final volume of 35 ml with water). Formation of the yellow/ orange peroxodisulfatotitanat(IV) complex was measured at 405 nm. H 2 O 2 formation was calibrated using a standard curve. H 2 O 2 formation was also determined using lucigenin as described by Rost et al. (17).
Determination of 18 (18). After 30 min of incubation, 2 l of 37% HCl per ml of reaction mix was added, and sulfate was precipitated with BaCl 2 (Sigma) in a final concentration of 0.85 mM. BaSO 4 was collected by centrifugation (10 min, 15,000 rpm). Air-dried pellets of approximately 0.5 mg were transferred into silver capsules (ThermoFinnigan). Samples were injected into a high temperature conversion/elemental analyzer (ThermoFinnigan) for ␦ 18 O analysis, coupled by a Conflo II interface to an isotope ratio mass spectrometer (Delta Plus; ThermoFinnigan). Isotopic values are expressed in atom % 18 O.
Sulfate and Sulfite Analysis-50-l aliquots of samples were injected into an ion exchange chromatography system (DX 120; Dionex). Sulfate and sulfite were separated on an IonPac column (AS9-Sc 250 ϫ 4 mm; Dionex) and eluted with a solution containing 1.8 mM Na 2 CO 3 plus 1.7 mM NaHCO 3 at a flow rate of 0.9 ml min Ϫ1 . Both compounds were detected by a conductivity detector module (Dionex) combined with an upstream-inserted micromembrane suppressor (ASRS-Ultra 4 mm, Dionex). The detection limit of this setup is 0.3 mol liter Ϫ1 for both ions.

Constitutive Expression of SO in Roots and
Leaves-The expression of SO on plant level was analyzed. SO protein was detected in all plant organs of Arabidopsis including roots, leaves, and flowers (Fig. 1A). The overall protein concentration of SO in leaves was found to be in a range of 0.1% of the total leaf protein. This quantification was performed by comparing a linear dilution of recombinant SO with the total extracted SO protein, as detected with the purified peptide antibody against recombinant Arabidopsis SO (data not shown). Using this antibody, immunoblot analysis was employed to measure the diurnal variation of SO expression. With 20 g of total crude proteins of Arabidopsis leaves and roots, respectively, only slight variations were detected over the day (Fig. 1B). Six to nine hours after light setting, SO expression was highest, while during the night and in the morning the protein amount was slightly decreased.
Molecular Oxygen as Terminal Electron Acceptor-Vertebrate SO is an intensively studied enzyme. Different electron acceptors were described to interact with purified animal SO including cytochrome c, oxygen, ferricyanide, methylene blue, and 2,6-dichlorophenol indophenol (2, 3). Electrons derived from sulfite are passed via the heme domain of the enzyme onto cytochrome c, the physiological electron acceptor (19). Plant SO, however, is unable to interact with cytochrome c because it lacks a heme domain. Furthermore, plant SO is not localized in mitochondria like animal SO but in peroxisomes. Therefore an electron acceptor other than cytochrome c has to be assumed. As animal SO was described to function with oxygen, although with much lower affinity, we tested the consumption of molecular oxygen by plant SO. The experiments showed that plant SO is able to use molecular oxygen as terminal electron acceptor very efficiently. The K m value of 22.6 M was in the same range as the K m value with the artificial acceptor ferricyanide determined in previous experiments (7).
H 2 O 2 Is Formed during Sulfite Oxidation-When plant SO uses molecular oxygen as terminal electron acceptor, the question arises what could be the products of this reaction? Clearly, sulfate is the main reaction product, and as plant SO is localized in peroxisomes (9) Table 1 show that in the absence of catalase only 48 -62% of the label is found in sulfate, thus demonstrating that plant SO combines an enzymatic with a subsequent non-enzymatic sulfite oxidation.
The turnover number of plant SO was calculated on the basis of V max determined in the presence of catalase to measure only the enzymatic reaction (without catalase V max is higher because of the contribution of the subsequent non-enzymatic reaction). The turnover number of plant SO was calculated to be 4,500 s Ϫ1 and that for catalase is one of the highest known, namely 40,000,000 s Ϫ1 (24). This ensures that at the catalase concentrations used in the experiments virtually all H 2 O 2 generated by SO will be immediately decomposed. Furthermore, when determining the rate of sulfite oxidation by plant SO in the absence and in the presence of catalase, values of 0.15 Ϯ 0.015 M s Ϫ1 and 0.064 Ϯ 0.009 M s Ϫ1 , respectively, were calculated. Thus the non-enzymatic reaction of sulfite with H 2 O 2 has an apparent rate of 0.086 Ϯ 0.012 M s Ϫ1 .
Most remarkably, plant SO and catalase are found in the same cell organelle, namely in peroxisomes (7,9,25). These organelles harbor catabolic pathways that characteristically produce H 2 O 2 , and catalase is one of the enzymes detoxifying excess H 2 O 2 . Catalase is inhibited by sulfite with half-maximum inhibition below 500 M sulfite (26). Therefore, we wanted to know whether catalase is inhibited by sulfite under our assay conditions. Fig. 3 shows that increasing amounts of sulfite inhibit catalase (half-maximal inhibition at 260 M sulfite). In the combined SO ϩ catalase assay (Fig. 3) the increase of sulfite concentration is accompanied by an initial increase in oxygen consumption because catalase is inhibited and does not contribute to the net oxygen consumption. This increase, however, is only transient because at higher sulfite concentrations of plant SO is also inhibited by its substrate.

DISCUSSION
From the data presented it is evident that plant SO is a housekeeping protein, expressed in every plant organ at any time with only low diurnal variation. Different from vertebrate SO that consists of a heme domain and a molybdenum cofactor domain, the heme domain is missing in the plant enzyme (7,8). While vertebrate SO is a mitochondrial enzyme with cytochrome c as the physiological electron acceptor (19), plant SO is localized in peroxisomes and does not react with cytochrome c (7,9). Therefore, plant SO should need an electron acceptor with a redox potential similar to a heme, but no evidence was found for the copurification of plant SO with a heme-like protein (7), 4 while in bacteria a separate heme domain was co-purified with the Moco domain of SO from Thiobacillus novellus (27). Our results identified oxygen as the terminal electron acceptor for plant SO that efficiently converts oxygen into H 2 O 2 . Thus, plant SO is the first eukaryotic member fully matching the EC nomenclature for sulfite:oxygen oxidoreductases (EC 1.8.3.1), while the animal enzymes with cytochrome c as the electron acceptor should fall into the class of sulfite:ferricytochrome c oxidoreductases (EC 1.8.2.1).
We found that in vitro the non-enzymatic oxidation of sulfite by H 2 O 2 contributes to the net reaction of plant SO. Earlier, Miszalski and Ziegler (28) had speculated that oxidation of sulfite in plants might be initiated by superoxide anions formed on the reduction site of the electron transport system in chloroplasts, or by free radicals, or by H 2 O 2 . However, no experimental proof was presented. Later it was shown that in clouds and rain droplets H 2 O 2 is one of the most effective oxidants for HSO 3 Ϫ . This reaction is the key step during the development of acid rain.
H 2 O 2 is highly soluble in water with ambient concentrations about 6 orders of magnitude higher than ozone. The reaction proceeds even at low temperatures (228 K) on the surface of ice clouds (20). The rate of   (29). The proposed mechanism proceeds via equilibrium with a peroxymonosulfurous acid ion (SO 2 OOH Ϫ ), where the velocity of the second step increases as pH decreases (29,30).
Our labeling experiments and the use of catalase provide evidence that under in vitro conditions plant SO combines its enzymatic sulfite oxidation with a subsequent non-enzymatic step using its reaction product H 2 O 2 as intermediate for oxidizing another molecule of sulfite. Under these in vitro conditions, the non-enzymatic step increases the production of sulfate and lowers the net consumption of oxygen. In planta, additional H 2 O 2 -generating reactions known to occur in peroxisomes may be involved in sulfite oxidation as well.
What could be the biological relevance of this process? Sulfite is a toxic metabolite that can break disulfide bridges, which is termed sulfitolysis; sulfite inhibits numerous enzymes, and it can attach to aldehydes forming hydroxysulfonates, which are metabolic inhibitors (14). Therefore, its fast removal by oxidation to non-toxic sulfate is a means to protect the cell against excess of sulfite derived from SO 2 gas in the atmosphere or during the decomposition of sulfur-containing amino acids. Prior to characterization of plant SO, this reaction was interpreted as the key step in protecting plants from acid rain that has been a major cause of forest damage (14,31). Not only sulfite but also H 2 O 2 is viewed as a toxic metabolite; however, in recent years it became clear that at lower concentrations H 2 O 2 also plays an important role as signaling molecule-mediating cellular responses to different stresses (for review, see Neill et al. (32) and Laloi et al. (33)). The steady-state concentration of H 2 O 2 for Arabidopsis has been reported to be in the range between 60 M and 7 mM (34,35). These variations may reflect technical difficulties in quantifying H 2 O 2 concentrations. Furthermore, the production of H 2 O 2 during sulfite oxidation also explains the peroxisomal localization (9) of plant SO because this organelle harbors catalase as one of the enzymes detoxifying excess H 2 O 2 .
To this end we have no direct proof for the existence of non-enzymatic sulfite oxidation in the living cell. For theoretical reasons, how-ever, we want to propose a model (Fig. 4) summarizing the possible interaction between plant SO and catalase in the peroxisome. At low sulfite concentrations, H 2 O 2 as a reaction product of plant SO is degraded by catalase. At higher sulfite concentrations, however, catalase becomes inhibited, and accumulating H 2 O 2 derived from the SO reaction oxidizes non-enzymatically a second sulfite molecule. Furthermore, it has to be taken into account that under catalase-inhibited conditions accumulating H 2 O 2 derived from other peroxisomal processes could contribute to the non-enzymatic detoxification of sulfite as well. The detailed regulation of this process is not fully understood and needs further investigation.