The roots of O. japonicus were collected in Dongtou, Zhejiang province (China). Sephadex G-150, DEAE- cellulose, fucose, arabinose, rhamnose, xylose, glucose, galactose, mannose, trifluoroacetic acid (TFA) was purchased from Sigma. RPMI-1640 and DMEM medium was purchased from Gibco. The assay kits for superoxide dismutase (SOD), malondialdehyde (MDA) and nitric oxide (NO) were purchased from Jiancheng Biologic Project Company, Nanjing, Jiangsu Province. Cell line HaCaT and LO2 was purchased from Shanghai Institute of Cell Biology, Chinese Academy of Sciences. All the other chemicals used were of analytical grade.

Gas chromatography (GC) and Gas chromatography-mass spectrometry (GC-MS) were described previously [20] . The products were dried by lyophilization. Gas chromatography (GC) (Shimadzu GC-2010) equipped with RTX-50 column (30.0 m × 0.25 mm × 0.25 μm) and flame-ionization detector (FID). Evaporation was performed at around 45°C under reduced pressure. The operation was performed using the following conditions: column temperature was programmed from 140˚C (maintained for 2 min) to 170˚C at a rate of 6˚C/min, and increased to 173˚C at a rate of 0.2˚C/min, then increased to 233˚C at a rate of 6˚C/min, held for 40 min at 233˚C; the rate of N₂ carrier gas was 1.0 ml/min; injection temperature was 250˚C; detector temperature was 300˚C. Gas chromatography-mass spectrometry (GC-MS) was run on the instrument Shimadzu GCMS-QP2010 (Shimadzu, Japan) and equipped with RTX-50 column (30 m × 0.25 mm × 0.25 μm), and at temperatures programmed from 140˚C (maintained for 2 min) to 250˚C (kept for 20 min) at a rate of 3˚C /min. Nitrogen was the carrier gas.

HaCaT cell were maintained in DMEM medium supplemented with heat-inactivated fetal bovine serum (10%) at 37˚C in humidified air containing 5% CO₂. The cells were plated into triplicate wells (100 μL/well) in 96-well ((5 × 10⁵ cells/well) flat bottom tissue culture plates and co-cultured with samples at indicated concentrations and 250 μM H₂O₂. Serum free DMEM was used as control. Following 24 h incubation at 37˚C and 5% CO₂, the supernatants were collected for detection of SOD activity, NO and MDA levels using commercial assay kits.

LO2 Cell were maintained in DMEM medium supplemented with heat-inactivated fetal bovine serum (10%) at 37˚C in humidified air containing 5% CO₂. The cells were plated into triplicate wells (200 μL/well) in 96-well flat bottom tissue culture plates and co-cultured with samples at indicated concentrations and 30 mmol/L glucose. Serum free DMEM was used as control. Following 24 h incubation at 37˚C and 5% CO₂, the cell was measured by MTT assay. The absorbance at 570 nm test wavelength was translated into inhibition ratio for comparison.

where A₀ and A_s are absorbance of blank and sample respectively.

The polysaccharide was prepared as described previously [20] . The roots of O. japonicus were soaked with 95% ethanol to remove the pigments and small lipophilic molecules. The residue was then extracted with 10 vol. of distilled water at 90˚C for 3 h thrice. All water-extracts were combined, filtrated, concentrated, and precipitated with 95% EtOH (1:4, v/v) at 4˚C for overnight. The precipitate was collected by centrifugation and deproteinated by Sevag method [21] . Finally the supernatant was lyophilized to give crude polysaccharides.

The crude polysaccharides were dissolved in distilled water and filtered through a membrane (0.45 μm). Then the solution was applied to a DEAE-52 cellulose column. Fractions were eluted with increased concentration of NaCl (0.01 - 1.0 M). The polysaccharide fractions were collected, concentrate, dialysed and finally lyophilized. Then, the sample was further purified by Sephadex G-150, and the second fraction was collected and lyophilized to give a polysaccharide named OJP2.

The polysaccharide was monitored by the phenol-sulfuric acid method [22] .

The homogeneity and molecular weight of OJP2 was evaluated and determined by high performance gel permeation chromatography (HPGPC) as described previously [20] . The sample solution was applied to Waters High Performance Liquid Chromatography (HPLC) equipped with a TSK-GEL G5000 SWXL column (7.8 ´ 300 mm), eluted with 0.1 mol/L Na₂SO₄ solution at a flow rate of 0.4 ml/min and detected by a Waters 2414 Refractive Index Detector. The columns were calibrated with Dextran T-series standard of known molecular weight (200,000, 70,000, 40,000, 10,000, 5000 Da). The molecular weight of OJP2 was estimated by reference to the calibration curve made above.

The monosaccharide of OJP2 was analyzed by GC as described previously [20] . OJP2 was hydrolyzed with 2 M TFA (2 ml) at 120˚C for 2 h. After removing TFA with methanol, the hydrolyzed product was reduced with NaBH₄ (50 mg), followed by neutralization with dilute acetic acid and evaporated at 45˚C. The reduced products (alditols) were added with 1 ml pyridine and 1 ml acetic anhydride in a boiling water bath for 1 h. The acetylated products were analyzed by GC.

OJP2 (20 mg) was methylated three times according to the method of Needs and Selvendran [23] . The methylated products were extracted by chloroform and examined by IR spectroscopy. The absence of the absorption peak corresponding to hydroxyl indicated the complete methylation. The product was hydrolyzed using 2 M TFA, followed by reduction using NaBH₄ and finally acetylated with acetic anhydride. The partially methylated alditol acetates were analyzed by GC-MS.

Lipid peroxidation was determined by quantifying MDA concentrations, which was spectrophotometrically measured by the absorbance of a red-colored product with thiobarbituric acid [24] . SOD and NO were measured with commercial kits, and the test method was done according to the reagent protocol prepared by the manufacturing firm.

Results were presented as mean ± standard deviation (S.D.). Data were analyzed by one-way ANOVA using Student’s t-test. P-values less than 0.05 were considered significant.

The polysaccharides extracted from the root of O. japonicus were purified through with DEAE-cellulose column and Sephadex G-100 column and named OJP2. OJP2 has no absorption at 280 and 260 nm in the UV spectrum, indicating the absence of protein and nucleic acid.

The average molecular weight of OJP2 was determined as 88.1 kDa by HPGPC. Results from phenol-sulfuric acid assay showed that OJP2 contained 95.3% carbohydrate.

OJP2 was hydrolyzed by TFA into individual monosaccharides that were further reduced and acetylated for GC analysis. The results showed that OJP2 is composed of Rha, Ara, Xyl, Glc, Gal with a relative molar ratio of 0.5:5:4:1:10 (Figure 1).

The IR spectrum of OJP2 was shown in Figure 2. The IR spectrum revealed a typical major broad stretching peak around 3400 - 3500 cm⁻¹ for the hydroxyl group, and the small band at around 2950.21 cm⁻¹ was attributed to the C-H stretching and bending vibrations. The relatively strong absorption peak at around 1614.55 cm⁻¹ reflects the absorption of the C=O group that is part of glycosides [20] [25] . The absorptions at 1020.91 and 1099.76 cm⁻¹ indicated a pyranose form of sugars [26] . The region between 950 and 1200 cm⁻¹ is dominated by ring vibrations overlapped with stretching vibrations of (C-OH) side groups and the (C-O-C) glycosidic band vibration [20] [27] .

The fully methylated OJP2 was hydrolyzed with acid, converted into alditol acetates and analyzed by GC-MS. As shown in Table 1, the ratios of methylated fragments were calculated based on the areas of the methylated products and corrected using the effective-carbon response method [28] . The GC-MS results (Table 1) indicated that 2,5-Me-Xyl (1,3-linked Xyl), 2,3,4-Me-Gal (1,6-linked Gal) and part of Ara were major components of the backbone structure, part of Ara and Gal were distributed in branches, and residues of branches terminated with Rha, Glc and Gal.

As shown in Table 2, the activity of SOD was significantly (P < 0.01) decreased in H₂O₂-treated group compared to the normal control group. OJP2-treated (250, 500, 1000 μg/mL) significantly (P < 0.01) increased the activities of SOD in H₂O₂-treated HaCaT cells with a dose-dependent manner. The treatment of HaCaT cells with H₂O₂ afforded the decrease in the NO production. However, incubation of the H₂O₂-treated cells with OJP2 polysaccharide markedly reduced in a dose-dependent manner the decrements in NO production of H₂O₂-treated cells (P < 0.01).

The levels of MDA in H₂O₂-treated groups was significantly (P < 0.01) increased (Table 2). Incubation of the H₂O₂-treated cells with OJP2 polysaccharide significantly decreased (P < 0.01 or P < 0.05) the levels of MDA, although it did not act as a dose-dependent manner.

As shown in Table 3, the treatment of LO2 cells with 30 mmol/L glucose caused inhibition of the cells growth and the inhibition ratio up to 52.8%. However, incubation of the glucose-treated cells with OJP2 polysaccharide markedly decreased in a dose-dependent manner the inhibition ratio of glucose-treated cells.

The treatment of LO2 cells with 30 mmol/L glucose afforded the decrease in the activity of SOD and NO production. However, incubation of the glucose-treated cells with OJP2 polysaccharide markedly reduced in a dose-dependent manner the decrements in SOD activity and NO production of glucose-treated cells (P < 0.01 or P < 0.05). The treatment of OJP2 not only recovered the glucose-decreased SOD activity and NO production, but also increased SOD activity and NO production, especially for NO production.