Novel Hydrogen Peroxide Metabolism in Suspension Cells ofScutellaria baicalensis Georgi*

We identified a rapid and novel system to effectively metabolize a large amount of H2O2 in the suspension cells ofScutellaria baicalensis Georgi. In response to an elicitor, the cells immediately initiate the hydrolysis of baicalein 7-O-β-d-glucuronide by β-glucuronidase, and the released baicalein is then quickly oxidized to 6,7-dehydrobaicalein by peroxidases. Hydrogen peroxide is effectively consumed during the peroxidase reaction. The β-glucuronidase inhibitor, saccharic acid 1,4-lactone, significantly reduced the H2O2-metabolizing ability of theScutellaria cells, indicating that β-glucuronidase, which does not catalyze the H2O2 degradation, plays an important role in the H2O2 metabolism. As H2O2-metabolizing enzymes, we purified two peroxidases using ammonium sulfate precipitation followed by sequential chromatography on CM-cellulose and hydroxylapatite. Both peroxidases show high H2O2-metabolizing activity using baicalein, whereas other endogenous flavones are not substrates of the peroxidase reaction. Therefore, baicalein predominantly contributed to H2O2 metabolism. Because β-glucuronidase, cell wall peroxidases, and baicalein pre-exist inScutellaria cells, their constitutive presence enables the cells to rapidly induce the H2O2-metabolizing system.

Plants display a broad range of defense responses to protect themselves against mechanical damage or pathogen attack. One of the earliest responses is the production of large amounts of reactive oxygen species (ROS), 1 which is called the oxidative burst (1)(2)(3)(4). The physiological roles of the oxidative burst have been well examined to date. For example, several studies demonstrated that ROS directly reduce pathogen viability (5) and that micromolar concentrations of H 2 O 2 inhibit spore germination of a number of fungal pathogens (6). In soybean and tomato, cell wall proteins such as hydroxyproline-rich proteins (extensin) are rapidly immobilized by the action of peroxidases as soon as H 2 O 2 is produced, resulting in strengthening of the cell walls to pathogen attack (7,8). Moreover, it has been reported that the oxidative burst can induce lipid peroxidation, leading to loss of membrane integrity and finally to the death of host plant cells, which is known as the hypersensitive response (5,9). Thus, although the oxidative burst plays important roles in plant defense, the plant cells themselves are consequently exposed to serious oxidative stress. Most plant cells produce ROS as by-products of redox reactions under normal conditions, but they are maintained at a low level by ROS-metabolizing enzymes such as catalase and ascorbate peroxidase (10,11). However, the detoxification mechanism of ROS produced by the oxidative burst has not been clearly understood. In this work, we establish that the suspension cells of Scutellaria baicalensis Georgi (skullcap plants) have a rapid and novel system to detoxify a large amount of H 2 O 2 .
S. baicalensis contains numerous flavones, and their pharmacological properties have been extensively investigated. Among them, baicalein (BA) has attracted considerable attention, as it has a variety of interesting activities such as antibacterial (12), antiviral (13), anticancer (14), and lipoxygenaseinhibitory effects (15). In addition, this plant has long been known to possess a ␤-glucuronidase, called baicalinase (16). Previously, we purified ␤-glucuronidase from the callus culture and demonstrated that this enzyme displays high activity for baicalein 7-O-␤-D-glucuronide (BAG), the main flavonoid of this plant (17). To establish the physiological roles of ␤-glucuronidase and BAG, we attempted further studies. Consequently, we found that the suspension cells of S. baicalensis effectively metabolize H 2 O 2 by a sequential reaction, including the hydrolysis of BAG to BA by ␤-glucuronidase and then the oxidation of BA by cell wall peroxidases. In this paper, we describe the detoxification mechanism of H 2 O 2 in Scutellaria cells. We also report on the purification of the peroxidases involved in H 2 O 2 degradation and their kinetic properties.

EXPERIMENTAL PROCEDURES
Plant Materials-The calluses were induced from the shoot stem segments of S. baicalensis, as described previously (17). The suspension cells were obtained by incubating the 4-week-old calluses in liquid Murashige-Skoog medium (18) containing 2,4-dichlorophenoxyacetic acid (0.5 mg/liter) and N 6 -benzyladenine (0.5 mg/liter) at 25 Ϯ 1°C under 16-h light conditions. The suspension cells were subcultured every 4 weeks in liquid Murashige-Skoog medium under the same culture conditions.
Structural Determination of 6,7-Dehydrobaicalein-Because the amount of the unknown compound produced from the elicited cells by oxidative burst was not enough to analyze its structure, we enzymatically synthesized this compound from BA using CM-52 eluate as a crude peroxidase preparation. The CM-52 eluate (50 ml), which was prepared from 4-week-old cells (300 g) as described below, was added to 50 mM citrate buffer (pH 4.0, 200 ml) containing 2 mM H 2 O 2 . BA (30 mg) dissolved in dimethyl sulfoxide (1 ml) was slowly added to the enzyme solution and stirred at room temperature for 5 min. We confirmed by HPLC analysis that the reaction product had the same retention time * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
as the unknown compound produced by the oxidative burst. The reaction mixture was applied to the reversed-phase polystyrene gel, MCIgel CHP 20P (Mitsubishi Chemical Co., Tokyo, 2.0 ϫ 20.0 cm) previously equilibrated with distilled water. After the column was washed with 100 ml of distilled water, fractions containing the unknown compound alone were eluted with 100 ml of methanol. However, we did not concentrate this methanol solution because the unknown compound was unstable to concentration procedures, including evaporation and lyophilization. Therefore, NMR analysis could not be carried out, but the methanol solution showed an [MϩH] ϩ ion peak at m/z 269 in the fast atom bombardment mass spectrum. Finally, the unknown compound could be isolated as a phenazine derivative. After o-phenylenediamine (150 mg) was dissolved in the above methanol solution containing the unknown compound, 1 ml of acetic acid was added. The reaction mixture was stirred at room temperature for 10 min, and the methanol was then removed by evaporation. The residue was loaded on an MCIgel CHP 20P column (2.0 ϫ 15.0 cm) pre-equilibrated with 50% (v/v) aqueous methanol. After washing the column with 100 ml of the same solvent, elution with 80% (v/v) aqueous methanol afforded a phenazine derivative (19 mg). This compound was characterized as the 6,7-phenazine derivative of BA by its NMR analysis. Because o-phenylenediamine reacts with compounds that have an ortho-or a 1,2-diketone structure to form phenazine derivatives (19), the unknown compound was determined as 6,7-dehydrobaicalein (DBA). Structures of these compounds are shown as insets in Fig. 3.
Elicitor Treatment of Suspension Cells-The 4-week-old suspension cells were harvested by filtration and washed with fresh growth medium. Aliquots (each 5 g) of the cells were subdivided into 100-ml flasks, and 20 ml of the growth medium was added to each flask. After 200 l of 10% (w/v) yeast extract (Nacalei Tesque, Kyoto) dissolved in the growth medium was added (final concentration of yeast extract, 0.1%), each cell suspension was incubated at 25°C for 0 -180 min. The saccharic acid 1,4-lactone (SAL) treatment of the elicited cells was carried out as described above, except a yeast extract solution containing 100 mM SAL (final SAL concentration, 1 mM, Sigma) was used instead of the yeast extract solution alone. All experiments were repeated with three replicates in each samples.
Determination of BAG, BA, and DBA in the Elicited Cells-Elicited samples incubated for different times were homogenized with methanol (20 ml). After centrifugation of the homogenate at 20,000 ϫ g for 5 min, the supernatant was used to determine the amounts of BAG, BA, and DBA. The quantification of these flavones was carried out with an HPLC system (Tosoh, Tokyo) composed of a CCPM pump and a UV-8000 absorbance detector equipped with a Cosmosil 5C18 AR column (0.46 ϫ 15.0 cm, Nacalai Tesque). BAG, BA, and DBA were eluted at a flow rate of 1 ml/min with 25, 35, and 15% (v/v) aqueous acetonitrile containing 50 mM phosphoric acid, respectively. The effluent was monitored by absorption at 254 nm, and the peak intensity was determined with a Chromatocorder 21 (Tosoh). The amount of BAG and BA was calculated from the standard curves obtained with authentic samples. As we could not isolate DBA in a free form, the following standard curve was used for the estimation of DBA levels. Various known amounts of BA were enzymatically oxidized to DBA, and each peak intensity of DBA was measured. The molar concentrations of DBA were considered to be equal to those of BA because we had confirmed that the oxidation of BA to DBA proceeds quantitatively under the conditions used. Therefore, the standard curve for DBA was constructed with the concentrations of BA and the peak intensities of DBA.
Determination of Extracellular H 2 O 2 -After the elicited samples were centrifuged at 20,000 ϫ g for 5 min, the supernatant (1.0 ml) was passed over a Pasteur pipette column containing 0.1 g of Cosmosil 75C18 OPN (Nacalei Tesque) to remove phenolic compounds, which interfere with the precise quantification of H 2 O 2 . The column eluent was used to determine the amount of H 2 O 2 . We confirmed that H 2 O 2 is quantitatively recovered by this chromatography method. The column eluent (100 l) was incubated in a buffer (100 l) containing purified peroxidase 1 (0.2 g), 400 M BA, 0.6% (w/v) Triton X-100, and 200 mM sodium citrate buffer (pH 4.0) at 30°C for 10 min. Under these assay conditions, the amount of H 2 O 2 was shown to be linear with respect to the DBA production. Therefore, the peak intensity of DBA in the reaction mixture was measured by HPLC as described above; the amount of H 2 O 2 was calculated from the calibration curve, which was constructed with known amounts of H 2 O 2 and the peak intensities of DBA produced by the enzymatic reaction.
Assay of ␣-Mannosidase-The assay consisted of 100 mM sodium citrate buffer (pH 5.5), 2 mM 4-nitrophenol-␣-D-mannoside, 1 mM mercaptoethanol, and the enzyme solution (500 l) in a final volume of 1 ml. The samples were incubated at 30°C for 60 min, and the reaction terminated with 100 l of 0.5 N NaOH. The absorption of the sample was measured at 415 nm. The amount of 4-nitrophenol liberated was calculated from the standard curve.
Assay of ␤-Glucuronidase and the Cell Wall Peroxidase-The assay for ␤-glucuronidase was conducted, as described previously (17). The peroxidase activity was determined by incubating an assay mixture containing the enzyme solution, together with 200 M BA, 200 M H 2 O 2 , 0.3% (v/w) Triton X-100, and 100 mM sodium citrate buffer (pH 4.0) at 30°C for 10 min in a final volume of 200 l. The amount of H 2 O 2 in the reaction mixture was measured as described above. The enzyme activity (katal) was defined as the amount (mole) of H 2 O 2 consumed per second.
Purification of Cell Wall Peroxidases-All procedures were carried out at 4°C, unless otherwise indicated. Four-week-old suspension cells (300 g) were shaken in 1 M NaCl (1,000 ml) at 100 rpm for 60 min and then filtered with Nylon screen. After the filtrate was centrifuged at 20,000 ϫ g for 15 min, the supernatant was fractionated by addition of ammonium sulfate. Proteins precipitating between 10 and 75% saturation were collected by centrifugation at 20,000 ϫ g for 30 min and then dialyzed overnight against three changes of 10 mM sodium phosphate buffer (pH 6.0). The dialyzed sample was applied to a Whatman CMcellulose column (2.5 ϫ 25.0 cm) equilibrated with the same buffer. After the column was washed with the above buffer (200 ml), bound proteins were eluted with a 1,000-ml linear gradient of NaCl (0 -0.4 M) at a flow rate of 1.0 ml/min. The eluent was collected in 20-ml fractions. The most active fractions (fractions [23][24][25] were concentrated by ultrafiltration (Advantec, Tokyo) and dialyzed against 10 mM sodium phosphate buffer (pH 7.0). The dialysate was applied to a 1.0 ϫ 20.0-cm column containing hydroxylapatite (Nacalai Tesque, Kyoto) pre-equilibrated with 10 mM sodium phosphate buffer (pH 7.0). Fractions containing peroxidase 1 were eluted with 150 ml of the same buffer at a flow rate of 1.0 ml/min. The most active fractions (fractions 5-7, each 20 ml) were pooled and concentrated by ultrafiltration. Peroxidase 2 was eluted from the same column with a 400-ml linear gradient of ionic strength sodium phosphate buffer (pH 7.0, 10 -100 mM). The most active fractions (fractions [12][13][14] were pooled, concentrated by ultrafiltration, and dialyzed against 10 mM sodium phosphate buffer. Protein Assay-Protein concentrations were measured according to Bradford (20) using bovine serum albumin as the standard.
Determination of Molecular Mass and Isoelectric Point-SDS-PAGE analysis was carried out with the system of Laemmli (21) in a 12.5% acrylamide gel of 0.75-mm thickness (Bio-Rad). The subunit molecular mass of the enzyme was determined by comparison with low molecular mass protein standards (Sigma). Isoelectric focusing was conducted according to O'Farrell (22) using 7.5-mm glass tubes (Atto, Tokyo). The pI of the purified enzyme was determined by comparison with marker proteins (Sigma). The native molecular masses of peroxidases 1 and 2 were estimated by gel filtration chromatography on a 1.5 ϫ 75.0-cm column of Sephacryl S-200 HR (Amersham Pharmacia Biotech) equilibrated with 10 mM phosphate buffer (pH 7.0) at a flow rate of 0.3 ml/min. Fractions of 15 ml were collected. Molecular mass markers from 29 to 700 kDa were resolved under the same conditions prior to running the fractions containing peroxidase activity.

Elicitor Treatment of Cell Suspension of S. baicalensis in the Presence of ␤-Glucuronidase
Inhibitor-S. baicalensis is known to contain flavonoid-specific ␤-glucuronidase, although its physiological importance is not understood (16). To reveal the roles of this enzyme in S. baicalensis, we first investigated effects on the cells induced by inhibition of ␤-glucuronidase. In this study, SAL was used as the ␤-glucuronidase inhibitor.
A significant effect of ␤-glucuronidase inhibition was observed in the elicited cells but not in the nonelicited cells. When Scutellaria cells were treated with the elicitor yeast extract in the presence of SAL, numerous components that should exist within the cells were released to the medium (Fig. 1A). In addition, ␣-mannosidase (0.60 nanokatal/g fresh cells), which is reported to be present in the vacuoles and cytosol (23), was also detected in the medium. Accordingly, the elicitation in the presence of SAL was thought to result in serious damage to the cells. In contrast to SAL treatment, elicitation in the absence of SAL resulted in much less damage; the elicited cells released only several components (Fig. 1B) and a lower level of ␣-mannosidase (0.09 nanokatal/g fresh cells) to the medium. Similar experiments using H 2 O 2 instead of the elicitor also demonstrated that the elicited cells undergo more serious damage in the presence of SAL. These results indicated that Scutellaria cells have a rapid H 2 O 2 -detoxifying system that is inhibited by SAL.
We measured the amount of H 2 O 2 produced by the elicitation. As shown in Fig. 2, the SAL-treated cells immediately started to produce H 2 O 2 after addition of the elicitor. The amount of H 2 O 2 was maximum after 10 min of incubation and thereafter was slowly degraded. The SAL-untreated cells also produced the maximum level of H 2 O 2 at 10 min after elicitation, although the amount was lower, as compared with SALtreated cells. After 30 min of incubation, the H 2 O 2 amount decreased to the same level as that before elicitation, showing that the SAL-untreated cells effectively metabolize H 2 O 2 . Based on these results, we concluded that the rapid degradation of H 2 O 2 depends largely on a hydrolytic reaction by ␤-glucuronidase. Because S. baicalensis possesses a large amount of BAG along with BAG-specific ␤-glucuronidase (17,24), it was suggested that the aglycone BA contributes to the H 2 O 2 metabolism.
Production of 6,7-Dehydrobaicalein in Elicited Cells-HPLC analysis showed that the elicited cells in the absence of SAL released an unknown compound (retention time, ϳ8 min) to the medium (Fig. 1B). In contrast, this unknown peak was much smaller on the chromatogram of the SAL-treated cells (Fig. 1A). As this compound was also assumed to be involved in the H 2 O 2 metabolism, we attempted a structural elucidation. After several unsuccessful attempts, its structure was determined as DBA by the method as described under "Experimental Procedures." DBA is the first flavone with an ortho-diketone moiety. Judging from its structure, this flavone was considered to be derived from the oxidation of BA.
Changes in the Amounts of BAG, BA, and DBA during Elicitation-We assessed changes in the amounts of BAG, BA, and DBA during elicitation. In the absence of SAL, the cells immediately initiated the hydrolysis of BAG after addition of the elicitor, and BAG continued to decrease until 30 min after incubation (Fig. 3A). The decrease of BAG correlated well with the H 2 O 2 degradation such that H 2 O 2 induced by the oxidative burst also decreased until 30 min after incubation, as shown in Fig. 2. On the other hand, BA increased rapidly with BAG hydrolysis, and the BA amount at 30 min was about two times higher than that before elicitation (Fig. 3B). Although at least ϳ2.5 mol of BAG should be hydrolyzed after elicitation, only a slight increase (ϳ0.06 mol) in the BA amount was observed even at the maximum level, suggesting that most of the BA is immediately metabolized. SAL-untreated cells produced more DBA than BA, and its amount reached a maximum level after 60 min of incubation (Fig. 3C). Thus, the production of DBA also occurred rapidly in response to the elicitor. In contrast, addition of SAL extensively inhibited the hydrolysis of BAG (Fig. 3A) and the production of DBA (Fig. 3C), indicating that the hydrolysis of BAG is essential for the production of DBA. Therefore, we concluded that DBA is produced by the hydrolysis of BAG to BA, followed by the oxidation of BA. Hydrogen peroxide may be consumed during the conversion of BA to DBA because this step is an oxidative reaction. DBA was also assumed to undergo further metabolism, based on the facts that the decrease in the DBA amount was observed on incubation over 120 min and that only ϳ15% of BAG hydrolyzed was recovered as DBA (0.38 mol at the maximum level). However, the mechanism of DBA metabolism was not established in this study.
Identification of the H 2 O 2 -metabolizing Enzyme-We investigated whether the oxidation of BA to DBA required H 2 O 2 . Consequently, crude enzyme extract from Scutellaria cells was shown to catalyze the formation of DBA from BA by consuming H 2 O 2 , indicating that H 2 O 2 is metabolized by peroxidase. Furthermore, we confirmed that BA metabolism by enzymes except for peroxidase is ineffective because the crude enzyme extract did not metabolize BA in the absence of H 2 O 2 . To extract this peroxidase effectively, extraction conditions were optimized. Crude enzyme extracts were prepared using various solvents, and then the peroxidase activity in each extract was measured by quantifying the amount of H 2 O 2 consumed by the oxidation of BA. The higher activity was observed in the extract prepared with 1 M NaCl, which is often used for solubilization of proteins ionically bound to the cell walls. Moreover, the protoplasts prepared by digestion of the cell walls displayed much lower metabolizing activity for H 2 O 2 (0.5 nanokatal/g cells) as compared with the intact Scutellaria cells (11.5 nanokatal/g cells), and most of the activity was recovered in the digestion medium. These results apparently indicate that the H 2 O 2 -metabolizing enzymes exist in the cell walls.
The Activity of ␤-Glucuronidase and the Cell Wall Peroxidase after Elicitation-Because it was confirmed that ␤-glucuronidase and the cell wall peroxidase contribute to the detoxification of H 2 O 2 , we measured the activity of both enzymes after elicitation. As shown in Fig. 4, neither activity increased during the incubation period (180 min) tested, and the peroxidase activity decreased more than the ␤-glucuronidase activity. Hence, we concluded that the metabolism of H 2 O 2 produced by the oxidative burst is catalyzed by these enzymes that are present consititutively in Scutellaria cells.
Purification of Cell Wall Peroxidases-Previously, we purified ␤-glucuronidase from the Scutellaria calluses and characterized its properties (17). At present, there is no information on the cell wall peroxidases in this plant. Therefore, to characterize precisely the cell wall peroxidases, we attempted to purify them. The 4-week-old cells of S. baicalensis were shaken in 1 M NaCl and then filtered. The filtrate was fractionated by ammonium sulfate saturation. More than 70% of the peroxidase activity was precipitated between 10 and 75% saturation of ammonium sulfate, resulting in a 3-fold purification. As a first chromatographic step, the solubilized ammonium sulfate fraction was applied to a CM-cellulose (CM-52) column, where a peroxidase was eluted with a linearly increasing gradient of NaCl (0 -0.4 M). A lower level of peroxidase activity was recovered in the void volume, whereas most of the peroxidase activity eluted at ϳ0.2 M NaCl. This step increased the specific activity in the latter fractions by a factor of 23-fold with a recovery of 43%. Because the total enzyme activity in the latter fractions was about 50 times higher than that in the former fraction, further purification of the former fractions was not carried out. As a final purification step, the latter CM-52 eluate was applied to a hydroxylapatite column. A peroxidase activity was eluted with 10 mM phosphate buffer, and SDS-PAGE showed that the peroxidase (termed peroxidase 1 in this paper) in these fractions was purified to homogeneity (Fig. 5). Further elution with an increasing gradient of the ionic strength of phosphate buffer (10 -100 mM) gave another peroxidase, named peroxidase 2. The purification of peroxidase 2 to homogeneity was also confirmed by SDS-PAGE. Peroxidases 1 and 2 were purified 121-and 89-fold by the three-step procedure with a final recovery of 17 and 8% of the enzyme activity, respectively. Both peroxidases are the first flavone-metabolizing enzymes to be purified.
Molecular Mass and Isoelectric Point of Peroxidases 1 and 2-SDS-PAGE of peroxidases 1 and 2 showed subunit molecular masses of 38 and 34 kDa, respectively (Fig. 5). The native molecular masses were estimated from the elution volume of both peroxidases on Sephacryl S-200 HR chromatography, where peroxidases 1 and 2 eluted as a single molecular species with a molecular mass of about 35 kDa in each case. These results suggested that each peroxidase exists as a monomeric enzyme. The pI values for peroxidases 1 and 2 were determined to be 8.7 and 8.9, respectively by comparison with marker proteins of known pI on isoelectric focusing gels.
Standard Assay Conditions of Peroxidases 1 and 2-We determined the optimum pH of each peroxidase using BA. The activity of peroxidase 1 was maximum between pH 4.0 and 4.5, with half-maximal activities at pH values around 3.0 and 6.0. Peroxidase 2 also showed the maximum activity between pH 4.0 and 4.5, although its activity was slightly lower than that of peroxidase 1. Based on these results, standard assays were carried out with citrate buffer (pH 4.0).
Effects of Various Flavones on H 2 O 2 -metabolizing Activity of Peroxidases 1 and 2-In S. baicalensis, wogonin and oroxylin A are also biosynthesized as minor flavones (24). To reveal whether these endogenous flavones are involved in the detoxification of H 2 O 2 , the substrate specificity was examined with them as well as BA. As shown in Table I, peroxidases 1 and 2 showed the high H 2 O 2 -metabolizing activity using BA with peroxidase 1 showing somewhat higher activity than peroxidase 2. In contrast, BAG does not undergo oxidation by either peroxidase. Scutellaria cells contain a much lower level of BA (ϳ0.06 mol/g fresh cells) than BAG (ϳ9.5 mol/g fresh cells), indicating that the cells have to hydrolyze BAG to effectively degrade H 2 O 2 . In addition, peroxidases 1 and 2 could not oxidize any of the endogenous flavones except for BA. In contrast to BA, these flavones possess a methoxyl or a glucuronyl group in their molecule; therefore, it is assumed that both enzymes can only oxidize flavones with hydroxyl groups. Moreover, we evaluated the peroxidase activity using apigenin (5,7,4Ј-trihydroxyflavone) and luteolin (5,7,3Ј,4Ј-tetrahydroxyflavone), which have not been identified in S. baicalensis. Both peroxidases showed a high activity with luteolin, whereas apigenin was not oxidized by either peroxidase. Like BA, luteolin has an ortho-dihydroxyl moiety in its molecule, thus suggesting that an ortho-dihydroxyl group is required for the peroxidase reaction. DISCUSSION ROS play important roles in plant defense such as the pathogen growth inhibition, cell strengthening, and the hy-persensitive reaction. However, they are also thought to inflict serious damage on the host plant cells. In particular, huge amounts of ROS are quickly produced by the oxidative burst, although their detoxification mechanism is not fully understood. As results of our studies on flavonoid metabolism, we identified in Scutellaria cells a novel H 2 O 2 -metabolizing system that is closely linked with the metabolism of the flavone BAG. Such a H 2 O 2 -metabolizing system has not been hitherto reported. Like Scutellaria cells, rye leaves (23), Pueraria lobata cells (25,26), apple leaves (27), and garbanzo plants (28) also possess endogenous-flavonoid-specific glycosidase in addition to flavonoid glycosides. Interestingly, their aglycones are assumed to be oxidized by a peroxidase reaction or to have a potent antioxidant activity, suggesting that these plants may have ROS-detoxifying systems similar to that in Scutellaria cells.
The first step in the metabolic pathway of H 2 O 2 is the hydrolysis of BAG by ␤-glucuronidase. Previously, we reported that ␤-glucuronidase in S. baicalensis shows a high activity for the endogenous flavone BAG, but its roles have remained unclear (17). In the present study, we found that ␤-glucuronidase catalyzes the production of an antioxidant flavone BA. Keppler and Novacky reported that exogenous addition of antioxidants reduces death of hypersensitively responding cells (29), but surprisingly, Scutellaria cells can produce the antioxidant flavone BA in response to an elicitor or H 2 O 2 . To our knowledge, hydrolases involved in H 2 O 2 metabolism have not been reported so far. Under normal conditions, BAG and ␤-glucuronidase are thought to exist in the different cellular compartments, because despite the presence of high ␤-glucuronidase activity, the amount of BA is much lower than that of BAG. We assumed that the oxidative burst inflicted serious damage on the compartmentation, resulting in the hydrolysis of BAG, based on the fact that after addition of an elicitor or H 2 O 2 , the cells released components that should exist within the cells into the extracellular medium. Flavonoid hydrolysis initiated by stress has also been reported for other plants. In P. lobata the hydrolysis of isoflavone glucosides is initiated by the elicitor treatment (25,26), whereas the infection of pathogens to apple leaves causes hydrolysis of chalchone glucoside (27).
As a second step, released BA was rapidly converted to DBA by cell wall peroxidases, and H 2 O 2 was confirmed to be detoxified at this step. To characterize the properties of the enzymes in pure forms, we attempted to purify them from Scutellaria cells. After a combination of ammonium-sulfate precipitation and two chromatographic steps, two ionic forms of peroxidases (peroxidases 1 and 2) were resolved. The structural properties of peroxidases 1 and 2 are similar to each other. The molecular mass (38 and 34 kDa) of each peroxidase resembles those of  a The numbering of a flavone skeleton was described in Fig. 3A. b GluA, ␤-D-glucuronyl. c -, no activity was detected.
ascorbate peroxidase (34 kDa), guaiacol peroxidase (33 kDa), and extensin peroxidases (34 -37 kDa) (10). On the other hand, pI values (8.7 and 8.9) of peroxidases 1 and 2 indicated that they are cationic peroxidases. Among endogenous flavones tested, peroxidases 1 and 2 showed the highest H 2 O 2 -metabolizing activity with BA, and peroxidase 1 displayed somewhat higher activity than peroxidase 2. In contrast, both peroxidases could not oxidize other endogenous flavones including BAG. Scutellaria cells contain much more BAG than BA, indicating that H 2 O 2 metabolism depends significantly on the hydrolysis of BAG. These findings provided a reasonable elucidation for the result that SAL treatment extensively reduced the ability of Scutellaria cells to metabolize H 2 O 2 . It is notable that H 2 O 2 is effectively metabolized using luteolin as a proton donor. Neither luteolin nor its glycosides have been identified in S. baicalensis, but they do occur in numerous plants (30), thus suggesting that a similar detoxification of H 2 O 2 may occur commonly in the plant kingdom. Anhalt and Weissenböck (23) reported a metabolic pathway of luteolin glucuronide in rye leaves, where luteolin 7-Odi-glucuronyl-4Ј-O-glucuronide is hydrolyzed by endogenous ␤-glucuronidase, and the hydrolysate luteolin 7-O-diglucuronide is finally polymerized by a peroxidase reaction.
Flavonoid polymerization by peroxidase is also reported for daidzein, the isoflavone of P. lobata. Park et al. (26) demonstrated that in vitro peroxidase reaction with daidzein gives dimeric daidzein and unidentified polymers. We attempted similar reaction with BA and peroxidases, but such polymerization was not recognized, and only DBA was quantitatively produced. Therefore, we concluded that neither BA nor DBA is polymerized by peroxidases, contrary to the metabolism of luteolin 7-O-diglucuronide and daidzein, although the precise mechanism of DBA metabolism has remained still unclear.
Concerning the roles of cell wall peroxidases in other plants, the insolubilization of extensin (7,8) and the biosynthesis of the cell wall including lignification (31) have been hitherto reported. Our study unequivocally demonstrated a novel function in which cell wall peroxidases rapidly metabolize a huge amount of H 2 O 2 produced by the oxidative burst. These peroxidases and ␤-glucuronidase pre-exist in Scutellaria cells, and neither activity was increased by the oxidative burst. The constitutive presence of these enzymes, in addition to a large amount of BAG, possibly enables cells to immediately induce the H 2 O 2 degradation system.
Ascorbate peroxidase and catalase are well known as H 2 O 2metabolizing enzymes. SAL is not an inhibitor for both enzymes, although the SAL treatment induces a serious damage in the elicited Scutellaria cells. Therefore, H 2 O 2 detoxification by ascorbate peroxidase and catalase seems less effective as compared with the cell wall peroxidases. BA belongs to a quite different class of natural products from ascorbic acid, which is required for H 2 O 2 metabolism by ascorbate peroxidases, but it is interesting that both compounds act as a proton donor. The cell wall peroxidases of S. baicalensis and ascorbate peroxidase may metabolize H 2 O 2 by a similar mechanism because both peroxidase reactions afforded a product (DBA and dehydroascorbic acid) with a diketone moiety.
In conclusion, BAG and BA were shown to be involved in the protection of Scutellaria cells against oxidative stress, whereas other interesting roles were also suggested for BAG metabolism. It seems particularly important that BA is rapidly formed in response to the elicitor because BA has antibacterial (12) and antiviral (13) effects. We also confirmed that BA shows antibacterial activity for Clavibactor michiganensis subsup. nebraskense and C. michiganensis subsup. michiganensis, 2 suggesting that BA may contribute to a chemical defense against pathogens. Because more DBA was produced than BA during the oxidative burst, we are now examining the antimicrobial activity of DBA as well as BA using various pathogens.