Purification, structure and anti-oxidation of polysaccharides from the fruit of Nitraria tangutorum Bobr.

In this paper, polysaccharides were extracted from the fruits of Nitraria tangutorum Bobr. (NTWP) using a hot water extraction method and extraction conditions were optimized by RSM. The optimal conditions were determined as follows: extraction time 7 h, extraction temperature 60 °C, ratio of water to raw material 15 : 1, and with these conditions, the yield was 14.01 ± 0.11%. After purification using DEAE-cellulose column and Sephadex G-200 column, NTWP-II was successfully obtained. The results of GC-MS and SEC-LLS analysis suggested that monosaccharide composition of NTWP-II was composed of Rha, Ara, Man, Glc and Gal with the molar ratio of 1.14 : 2.5 : 3.00 : 2.69 : 5.28 and Mw, Mw/Mn and Rz 2.29 × 105, 1.32, 15.22. The detailed structure of NTWP-II was characterized by FT-IR, NMR. Based on these analyses, the structure of the repeating unit of NTWP-II was established.


Introduction
Oxidation is an important energy production process in organisms. However, it is well-known that free radicals, especially reactive oxygen species such as hydrogen peroxide, hydroxyl radicals and superoxide which generate during metabolism, could damage fatty acids, proteins, DNA and other macromolecules in our bodies and result in various diseases including cardiovascular diseases, neurodegenerative diseases, cancers and aging-related disorders. Recent papers prove that synthetic antioxidants are potential hazards in liver damage, carcinogenesis et al. and are restricted. 1,2 Thus, it is very important to exploit natural antioxidants, especially those of plant origins. Constituents and crude numerous plant extracts have been recognized as natural antioxidants, with benecial effects against free radicals in biological systems.
Nitraria (Zygophyllaceae) includes 15 species and only Nitraria tangutorum Bobr. grows in China. 3 It is used to sand stabilization because of natural ability in withstanding wind and sand. 4 In addition, the fruits and seeds of Nitraria tangutorum Bobr. were oen used by local residents to treat ailments of the spleen and stomach, indigestion, neurasthenia and colds, and the leaves were used as an antispasmodic, antineuropathic, and antiarrhythmic agent in folk medicine.
Thus, there was a growing interest in the area of research on the positive effect of Nitraria tangutorum Bobr. on human health and other benets. Recent studies focused on ingredients such as seed oil, anthocyanins, tangutorine, polysaccharides et al. from their leaves, seeds, fruits and juice by-products. [2][3][4][5][6][7][8] Some researches indicated that seed oil has been shown to improve immune response, reduce oxidants, and mitigate fatigue. [4][5][6] Nine anthocyanins were separated from the seed oil of Nitraria tangutorum by subcritical uid extraction. 5 Cyanidin derivatives, the main components, has a good antioxidant. 4 François' research showed that hydroalcoholic extract from the fruits of Nitraria sibirica Pall. induced vasorelaxation. 9 From the leaves of Nitraria tangutorum Bobr., Duan and Liu isolated tangutorine (1) and identied its structure. 6,7 The fruits of Nitraria tangutorum Bobr. were traditional medicinal food of Tibetans and used to alleviate fatigue caused by oxygen deciency for thousands of years. Wei hua Ni' study indicated that water-soluble polysaccharides from the fruits of Nitraria tangutorum Bobr. in Tibetan plateau signicantly exhibited anti-fatigue activities for the rst time, through triglyceride mobilization during exercise and protecting corpuscular membrane by prevention of lipid oxidation via modifying several enzyme activities. 8 Zhang et al. reported that the polysaccharides from Nitraria fruit was composed of mannose, rhamnose, galacturonic acid, glucose, galactose and arabinose with approximate molar ratios of 9.2 : 3.3 : 1.1 : 1 : 1.9 : 2.3. 10 Ni et al. isolated the polysaccharides (NTWP) from Nitraria tangutorum Bobr. by hot water extraction, puried by DEAE-cellulose ion exchange chromatography and found that NTWP was composed of mannose, rhamnose, glucuronicacid, galacturonic acid, glucose, galactose and arabinose with approximate molar ratios of 3.9 : 1.8 : 0.2 : 3.3 : 70.6 : 7.6 : 13.1. 8 Although previous studies have reported the structural properties of the polysaccharides from the fruits of Nitraria, the results of these studies were not consistent. To scan the structural characterization of polysaccharides from Nitraria tangutorum Bobr., we puried NTWP by DEAE-cellulose anion-exchange  chromatography and Sephadex G-200 column, identied their  structural features by GC-MS, SEC-LLS, FT-IR and NMR. Furthermore, we evaluated antioxidant activity of NTWP-II in vitro.

Plant materials
The fruits of Nitraria tangutorum Bobr. were collected from Min qin in Gansu, China at October 2012. The fresh fruits were air-dried, crushed and kept in plastic bags at room temperature for use.

Preparation of NTWP
The extraction process was carried out by the former report with slight modication. 8 The dried fruit powder was defatted with petroleum ether, and reuxed with petroleum benzine to remove lipids, some colored ingredients and small molecular impurities. Aer removing the solvent, the resulting pretreated powder was dried and used for the extraction of NTWP. The pretreated samples were soaked in distilled water at room temperature. For hot water treatment, a sample (1 g) was put into a triangular ask (1 L), diluted with distilled water at a liquid-solid ratio varying from 8 : 1 to 25 : 1, v/w, and extracted at 30-80 C for 1-12 h. The mixture was centrifuged at 5000 rpm for 15 min, and the insoluble residue was retreated as mentioned above for 2 times. The supernatants were collected, concentrated by a rotary evaporator under reduced pressure to an appropriate volume, then the supernatant was deproteinized by using Sevage's method. Then the deproteinized supernatant was dialyzed, concentrated, mixed with four times of absolute ethanol and kept overnight at 4 C. The resulting precipitates were collected by centrifugation at 5000 rpm for 10 min, washed sequentially with anhydrous ethanol and acetone, and dried to afford crude NTWP. The extraction yield (%) was calculated with the formula below: where Y was the yield of polysaccharide (%), W 1 was the polysaccharides of extraction (g), and W 0 represented dried sample weight (g).

Experimental design of NTWP
Based on the results of single factor experiments, a three-level BBD with three factors was applied to determine the optimal levels for extraction temperature (X 1 ), extraction time (X 2 ) and ratio of water to raw material (X 3 ), which signicantly affected the extraction efficiency, and the optimal range of each variable was determined. The experimental designs of the code and the actual levels of each factor are presented in Table 1. As shown in Table 1, the three factors chosen for this study were designated as X 1 , X 2 and X 3 and were prescribed into three levels, coded +1, 0, and À1 for high, intermediate and low value, respectively. Test variables were coded according to the following equation: where x i was the coded value of an independent variable; X i was the actual value of an independent variable; X 0 was the actual value of an independent variable at centre point; DX was the step change value of an independent variable. The whole design consisted of 17 experimental points carried out in random order ( Table 2). All experiments were performed in triplicate and the averages of polysaccharide yields were taken as response. For predicting the optimal point, a second order polynomial model was tted to correlate relationship between independent variables and response (polysaccharide yield). For the three factors, the equation was: where Y was the response variables (yields of polysaccharides in real values). A 0 , A i , A ii , A ij were the regression coefficients of variables for intercept, linear, quadratic and interaction terms, respectively. X i and X j were independent variables (i s j).
The signicance in the model was evaluated by analysis of variance (ANOVA). The accuracy and general ability of the polynomial model could be evaluated by a determination coefficient R 2 and adjusted coefficient of determination R adj 2 .
Subsequently, several conrmation experiments were conducted to verify the validity of the statistical experimental strategies. The regression coefficients were then used to make statistical calculation to generate dimensional and contour maps from the regression models. Statistica (Version 8.0, USA) soware package was used to analyze the experimental data. P-values of less than 0.05 were considered to be statistically signicant.

Isolation and purication of NTWP
The crude polysaccharide (NTWP), obtained under the optimal conditions, was redissolved in distilled water, and the supernatant was loaded onto a DEAE-cellulose column (2.6 cm Â 50 cm), eluting stepwise with a gradient of aqueous solution of sodium chloride (NaCl, 0-1.0 M) at a ow rate of 0.2 mL min À1 . The eluate was collected by using an automated fraction collector. In order to detect polysaccharides, a 0.2 mL sample collected from each eluted fraction (2 mL per tube) was mixed with sulfuric acid and phenol to produce color reaction. The carbohydrate content of each tube was monitored at 490 nm. The fractions with rose color were combined, concentrated, dialyzed, and lyophilized, aer which the main polysaccharide was redissolved in deionized water and loaded onto a Sephadex G-200 column (20 Â 500 mm). Aer loading the sample onto the column, the column was eluted with deionized water at a ow rate of 0.5 mL min À1 . The main polysaccharide fractions were collected, combined, and lyophilized to obtain a puried polysaccharide (NTWP-II).
2.6. Structural characterization of NTWP-II 2.6.1. Components analysis. The carbohydrate and proteins contents of NTWP-II were determined by the phenol-sulphuric acid method and Bradford's method according to the method by our report. The composition was analyzed according to the earlier report from our laboratory. 11 Briey, 4 mg of sample were dissolved in 4 mL of 4 M triuoroacetic acid acetate (TFA) in a test tube and then hydrolysed at 120 C for 10 h under air-tight conditions. TFA was then evaporated through decompression and distillation. When the tube was dry, 10 mg of ammonium hydrochloride and 0.5 mg pyridine were added and allowed to react in a 90 C water bath for 30 min. Then 0.5 mL of cold (kept at 4 C in a refrigerator) acetic anhydride was added to the test tube and the mixture was incubated in the 90 C water bath for a further 30 min to allow the acetylation reaction to occur. The end-product was decompressed and distilled to dryness. The acetate derivatives were analysed by GC-MS with an HP-5 capillary column (HP 6820, Hewlett-Packard). The temperature programme was set to increase from 120 C to 250 C with an increment of 5 C min À1 and He was the carrier gas. The standard monosaccharides were measured following the same procedure. D-Amr, L-Rha, D-Lyx, D-Ara, D-Xyl, D-Man, D-Glu, D-Gal were used as references.
2.6.2. FT-IR analysis. The sample was ground with KBr powder and then pressed into pellets for Nicolet NEXUS 670 FT-IR in the frequency range of 4000-400 cm À1 to detect functional groups. Sixteen scans at a resolution of 4 cm À1 were averaged and referenced against air.
2.6.3. Molecular weight determination. Size exclusion chromatography, combined with laser light scattering (SEC-LLS), measurements were carried out with a multi-angle laser photometer (MALLS, k ¼ 690 nm; DAWN EOS, WyattTechnology Co., USA). An UltrahydrogelTM column (7.8 Â 300 mm, Waters, USA) was used as the SEC instrument. An Optilab refractometer (Dawn, Wyatt Technology Co., USA) was simultaneously connected. The polysaccharide solutions with desired concentrations were prepared, and optical clarication of the solutions was achieved by ltration into a scattering cell. The injection volume was 200 mL, and the ow rate was 0.5 mL min À1 . The refractive index increment (dn/dc) value of the sample was determined, by using an Optilab refractometer at 690 nm and 25 C, to be 0.147 mL g À1 . The basic light scattering equation is as follows: where, K is an optical constant equal to [4p 2 n 2 (dn/dc)]/(l 4 N A ), c is the polysaccharide concentration in mg mL À1 , R q is the Rayleigh ratio, l is the wavelength, n is the refractive index of the solvent, dn/dc is the refractive index increment, N A , the Avogadro number, N A is the second virial coefficient. As the column separates the polymer according to molecular weight, each fraction was led to the light scattering detector for instantaneous measurement of the scattering intensities. The refractive index detector, connected in series, gave the estimate of the polymer concentration. In chromatography mode, we have a single and sufficiently low concentration at a particular slice because of the further dilution by the SEC column of the already dilute injected solutions. 2.6.4. NMR spectroscopy. NMR spectra were recorded with a BRUKER400 MHz spectrometer at a probe temperature of 80 C. All exchangeable H of the samples were replaced by D 2 O before obtaining the spectra in deuterium oxide. Chemical shis were calibrated using coaxial NMR tubes with (3trimethylsilyl)-propane sulfonic acid sodium salt (DSS, dH ¼ 0.00, dC ¼ 0.00) in the inner tube. The 2D 1 H-1 H correlated spectroscopy (COSY), 1 H-1 H total correlation spectroscopy (TOCSY), 1 H-13 C heteronuclear single quantum coherence (HMQC) and 1 H-13 C heteronuclear multiple quantum coherence (HMBC) measurements were used to assign signals and to determine the sequence of sugar residues. 12

Assay for antioxidant activities
Polysaccharides were dissolved in deionized water at the concentration of 0.02-5 mg mL À1 . Antioxidant assay in vitro was carried out on scavenging DPPH, hydroxyl, superoxide radical and chelating metal, using V E , V C , BHT and EDTA as the positive controls according to the earlier report. 11 Each concentration was measured in triplicate and averaged.

Results and discussion
3.1. Inuence of three single factors on the extraction efficiency of NTWP 3.1.1. Effect of ratio of water to raw material on extraction yield of NTWP. Ratio of water to raw material was a routine parameter that signicantly affects the extraction efficiency. In the present study, extraction of NTWP was carried out at different ratios of water to raw material (5, 10, 15, 20, 25, 30 mL g À1 ) when other parameters were as follows: extraction temperature 55 C and extraction time 5 h. As shown in Fig. 1a, the extraction yield increased with the increase of ratio of water to raw material from 5 to 15 mL g À1 and while above 15 mL g À1 , the yield increased slowly up to its maximum amount of 13.42 AE 0.4% at 20 mL g À1 when ratio of water to raw material continued to rise. This phenomenon could be explained that the higher the ratio of water to raw material was, the lower the concentration and viscosity of the extraction solvent would be. Hence more polysaccharides molecules could dissolve in water and the extraction yield increased. 13 But the amount of polysaccharides in raw material was denite and further increase of ratio of water to material would not increase the extraction yield. Thus, ratio of water to raw material range of 15 : 1 to 25 : 1 was favorable for extracting NTWP.
3.1.2. Effect of extraction temperature on extraction yield of NTWP. Extraction time is another factor that inuenced the extraction efficiency. To investigate the effect of temperature on extraction yield of NTWP, extraction was conducted at different temperatures (30, 40, 50, 60, 70 and 80 C, respectively), when the other extraction variables were set as follows: ratio of water to raw material 20 mL g À1 and extraction time 5 h. The effect of extraction time on the yield is presented in Fig. 1b. As shown in Fig. 1b, the extraction yield increased when temperature increased from 30 to 60 C, reaching a maximum (13.30 AE 0.6%) at 60 C, and then declined when extraction temperature continued to rise. Therefore, it could be explained that as the temperature increased, the polysaccharides diffusion coefficient would increase, resulting in an enhanced solubility of the polysaccharides in the solvent. However, when temperature became higher, the extraction yield of NTWP decreased drastically, probably due to that high temperature could destroy the structure of polysaccharides and lead to degradation. Therefore, extraction temperature range of 50-70 C was favorable for extraction of NTWP, being selected for further optimization in BBD design.
3.1.3. Effect of extraction time on extraction yield of NTWP. To investigate the inuence of time on extraction efficiency, extraction process was carried out at 1, 3, 5, 7, 9, 11 h, while other parameters were as follows: ratio of water to raw material 20 mL g À1 and extraction temperature 60 C. The effect of extraction time on the yield was presented in Fig. 1c. When extraction time varied from 1 to 9 h, the variance of extraction yield was relatively rapid and reached the maximum extraction yield (13.72 AE 1.3%) at 7 h, and then decreased as the extraction proceeded. These results can be explained that extended extraction time (>7 h) would result in the degradation of the polysaccharides, which was induced by their thermal instability. Therefore, aer the maximum extraction yield was achieved, longer time of how water extraction was not necessary. Thus, temperature range of 5-7 h was selected as optimal in the BBD experiment. Table 1, each experiment in the design matrix was performed and the experimental data were obtained. By applying multiple regression analysis on the experimental data, the Design-Expert soware generated a second order polynomial equation that could express the relationship between process variables and the response. The nal equation obtained in terms of coded factors was given below:

Predicted model and statistical analysis. As shown in
where Y was the predicted yield of polysaccharides, X 1 was ratio of water to raw material (W), X 2 was extraction temperature ( C) and X 3 was extraction time (h).
Analysis of variance (ANOVA) was performed to evaluate the predictive model and the variables. The P-values was used as a tool to check the signicance of each coefficient and indicated the pattern of interactions between variables. The smaller the value of P was, the more signicant the corresponding coefficient would be. A signicant lack of t demonstrated that the tted model failed to failed to represent the data in the experimental domain at which points were not included in the regression. The ANOVA for the tted quadratic polynomial model of extraction efficiency was presented in Table 3. As shown in Table 3, the quadratic regression model has a very low P-value (P < 0.0001), indicating that the tness of the model was highly signicant. The "Model F-Value" of 69.92 implied the model was signicant. There was only a 0.01% (P < 0.0001) chance that a "Model F-Value" could occur due to noise. At the same time, the "Lack of Fit F-Value" of 0.48 implied the lack of t was not signicant relative to the pure error. A "Lack of Fit F-Value" this large has 71.48% (P ¼ 0.7148) chance to occur because of noise. Coefficient of determination (R 2 ) indicated that 98.90% of the variations can be explained by the tted model. The value of adjusted determination coefficient (R adj 2 ) was 0.9749, which also conrmed that the model was highly signicant. Moreover, a relatively low value of coefficient variation (1.68%) indicated high degree of precision and good deal of reliability for the experimental values. "Adeq Precision" measured the signal to noise ratio. A ratio greater than 4 was desirable. The ratio of 26.961 indicated an adequate signal. Thus, the results indicated that this model could be used to navigate the design space. The signicance of each coefficient was also determined by using F-value and P-value. It could be seen that the linear coefficients (X 2 and X 3 ), quadratic terms coefficients (X 1 2 , X 2 2 , and X 3 2 ) and cross product coefficients (X 1 X 3 , X 2 X 3 ) were signicant, with small P-values (P < 0.05). And the signicant interaction between factors means that effect of them on the extraction yield of polysaccharides was not simply an additive. The empirical model was converted to three-dimensional (3D) and contour plots to predict the relationships between the independent variables and the response. 14, 15 3.2.2. Optimization of extraction procedure. The graphical representation of regression equation was obtained using Design-Expert to evaluate the effects of independent variables and their interactions on extraction efficiency of NTWP. As graphical representations of the regression equation, three dimensional response surface and two dimensional contour plots were very useful to visualize the relationship between independent and dependent variables and the interactions between two variables. Different shapes of the contour plots indicated different interactions between the variables. Circular contour plot means the interactions between the corresponding variables are negligible, while elliptical contour suggested the interactions between the corresponding variables are signicant. Among these three variables (ratio of water to raw material, extraction temperature and extraction time) as shown in Fig. 2, one variable was kept constant at zero level, when the other two variables within the experimental range were depicted in the plots. The extraction yield (Y) of NTWP affected by extraction time (X 1 ) and extraction temperature (X 2 ) was shown in Fig. 2A and a with ratio of water to raw material (X 3 ) xed at a zero level. The extraction yield (Y) increases rapidly when ratio of extraction time (X 1 ) and extraction temperature (X 2 ) increase in the range of 5-6.8 h and 50-60 C, respectively; but beyond 6.8 h and 60 C, extraction yield (Y) decreases slightly. The elliptical contour plot shown in Fig. 2A, a indicated the mutual interactions between extraction time (X 1 ) and extraction temperature (X 2 ) were signicant. Fig. 2B, b showed the 3D response surface plot and the 2D contour plot at varying extraction time and ratio of water to raw material at xed extraction temperature (0 level). The same trends with Fig. 2A, a were depicted in Fig. 2B, b, of which Fig. 2B, b showed a similar elliptical contour plot. The extraction yield of polysaccharides affected by extraction temperature and ratio of water to raw material is shown in Fig. 2C, c with extraction time xed at a zero level (7 h). The yield of polysaccharides increased evidently with the extraction temperature from 50 C to 65 C. However, beyond 65 C, the extraction yield would not increase as the temperature ascended.
3.2.3. Verication of predictive model. By employing the soware Design-Expert, the solved optimum values of the tested variables were extraction time 7.08 h, extraction temperature 60.45 C, liquid-solid ratio 15 : 49. Under the optimal conditions, the maximum predicted yield of NTWP was 14.13%. Taking account of the operating convenience, the optimal parameters were determined as following: extraction time 7 h, extraction temperature 60 C, liquid-solid ratio 15 : 1.   To ensure the predicted result was not biased toward the practical value, experimental rechecking was performed using this deduced optimal condition. A mean value of 14.01 AE 0.11% (n ¼ 3), obtained from real experiments, demonstrated the validation of the RSM model. The good correlation between experimental and predicted values conrmed that the response model was accurate and adequate for the extraction of NTWP. The validation result revealed that there was no signicant difference between experimental and predicted values, suggesting that the response model was adequate for reecting the expected optimization.

Isolation and purication of polysaccharide fraction NTWP
NTWP solution (5 mg mL À1 ) was loaded into a DEAE cellulose-52 column equilibrated with a linear gradient elution of NaCl from 0 M to 1.0 M. Two independent elution peaks (NTWP-I 24.11% and NTWP-II 64.57%, Fig. 3A) detected at 490 nm by the phenol-sulfuric acid assay were obtained.
NTWP-I consisted of glucose and NTWP-II consisted of rhamnose, mannose, glucose, galactose and arabinose by GC-MS. The next research focused on NTWP-II. Thus, only the NTWP-II fraction was collected for the subsequent purication and antioxidant activity assays. The NTWP-II fraction was collected, dialyzed, concentrated, and loaded into a Sephadex G-200 column. The fraction produced a single elution peak (Fig. 3B).

Chemical composition of NTWP-II
NTWP-II isolated from the fruit of Nitraria tangutorum Bobr. by a series of purication procedures, including water extraction, ethanol sedimentation, deproteinization, and dialysis. The total sugar content of NTWP-II separated by Sephadex G-200 column was 94.3 AE 3.42% and was poor in protein (2.22 AE 0.34%). Through acid hydrolysis, NTWP-II was subjected to GC-MS analysis and the results were shown in Fig. 4. As can be seen in Fig. 4a, NTWP-II was composed of rhamnose, arabinose, mannose, glucose and galactose with the molar ratio of 1.14 : 2.5 : 3.00 : 2.69 : 5.28. These results didn't show a good correlation with those reported by Ni et al. They found that NTWP was composed of mannose, rhamnose, glucuronicacid, galacturonic acid, glucose, galactose and arabinose with approximate molar ratios of 3.9 : 1.8 : 0.2 : 3.3 : 70.6 : 7.6 : 13.1. This result might be due to regional difference.

Molecular weight determination
The SEC-LLS chromatogram patterns of NTWP-II were shown in Fig. 5a and b. The chromatograms of NTWP-II exhibited a single peak indicating that there was no aggregation and the homogeneity of the puried samples. The molecular weight of NTWP-II was determined by SEC-LLS. The weight average molar mass (M w ), polydispersity (PD, M w /M n ) and z-average radius of gyration (R z ) were 2.29 Â 10 5 , 1.32, 15.22, respectively.

FT-IR analysis of NTWP-II
NTWP-II was characterized by FT-IR spectroscopy as shown in Fig. 6. The infrared spectra showed strong and wide stretching peak around 3413 cm À1 for O-H stretching vibrations as well as a weak absorption peak at 2902 cm À1 for C-H stretching vibrations. A strong absorption peak at 1618 cm À1 was attributed to asymmetric and symmetric stretching of the carboxylate anion group (COO), indicating NTWP-II be acidic polysaccharides. Each particular polysaccharide has a specic band in the 1000-1200 cm À1 region. This region was dominated by ring vibrations overlapped with stretching vibrations of (C-OH) side groups and the (C-O-C) glycosidic band vibration. The absorptions at 1072.99 cm À1 indicated a pyranose form of sugar. 16,17 A specic band in the 1080 cm À1 region indicated a pyranose form of sugar. 17,18 In the anomeric region (950-700 cm À1 ), the spectrum exhibited the characteristic absorption at 883 cm À1 due to the presence of mannose. 12 This showed a good correlation with monosaccharide composition. These results indicated that NTWP-II possessed typical absorption peak of polysaccharides.

Nuclear magnetic resonance (NMR) spectroscopy analysis
NMR spectroscopy could provide detailed structural information including the monosaccharide composition, aor banomeric congurations, linkage patterns, and sequences of the sugar units. Signals of NTWP-II in 1D 1 H and 13 C NMR and 2D NMR (HMQC, HMBC, COSY, NOESY and TOCSY) spectra were assigned as completely as possible, based on the monosaccharide analysis and chemical shis reported in the literature. 20,21 The 1 H and 13 C NMR spectra of NTWP-II were shown in Fig. 7a and b. The 1 H NMR spectrum was crowded in a narrow region ranging from 3 to 5 ppm which was typical of polysaccharides, and this conrmed the presence of many similar sugar residues. 22, 23 1 H NMR spectrum (Fig. 7b) NTWP-II contained ve signals at 2.23, 0.86 and 3.84 Hz, respectively), suggesting that residues A, B, D and E were a-linked. Meanwhile the anomeric proton of residue C has the chemical shi smaller than 5.0 ppm and suggested residue C being b-linked. 20,21,24 The 1D 13 C NMR spectra (Fig. 7a)  In TOCSY and HMQC spectrum (Fig. 7d and c), d 61.44 ppm and d 68.35 ppm were attributed to no-substitution of C-6 and weak substitution of C-6. Based on the chemical shis of H, C and 3 J H-1,H-2 of residues C, residues C was identied as b-galactose. According to the molar ratio of monosaccharide, signal intensity of 13   This journal is © The Royal Society of Chemistry 2018 linkage of A was C-1. Based on standard monosaccharide and former research, residues A was identied as a-arabinose, because the chemical shis of C-1 (d 97.6 ppm) in a-arabinose was larger than in b-arabinose (d 93.4 ppm). The chemical shis of C-4 (d 71.19 ppm) down low eld conrmed the substitution at position C-3. Therefore, we inferred that residue A was (1/4)-a-arabinose. Compared with standard monosaccharide, residues E was identied as (1/)-a-mannose, residues D and B were identied as (1/)-aglucose and (/6)-a-rhamnose, respectively from the chemical shis of residues D (C-6 d 63.53 ppm) and B (C-1 d 107.87 ppm).
In NOESY spectrum (Fig. 7f), the crosspeaks between C H-3 and E H-1, A H-1 and C H-3 indicated that the presence of the bond (1/3) between C and E, A and C. In HMBC spectrum (Fig. 7e), some crosspeaks were observed between H-1 and a 13 C peak in different residues: E C-1 and B H-2, and E C-1 and D H-6. These HMBC data revealed the existence of two sequences in the structure: E (1/2) B, E (1/6) D.
According to previous studies of NTP-II, (1/3)-b-galactose was the main sugar units of backbone and linked with a-mannose, aarabinose and b-galactose by the glycosidic linkage of (1/3) and (/1)-a-glucose of the side chain was obtained. It was proposed that the residue sequence in the repeating unit was as: 3.8. Antioxidant activities in vitro of NTP-II DPPH radical was a stable free radical, of which alcohol solution has characteristic absorption maximum at 517 nm and used to evaluate the free radical scavenging activities of natural compounds. When DPPH ethanol solution was reduced, absorbance was decreased and the solution changed from purple to light yellow. 17 The reduction extent of absorbance reects the hydrogen or electron donating abilities of antioxidants. 25 As shown in Fig. 8a, at the range from 0.02 mg mL À1 to 1 mg mL À1 , the color signicantly became light with the increase of NTWP-II concentrations. Beyond the concentration of 1 mg mL À1 , the trend was not signicant. However, the scavenging effect of NTWP-II on DPPH radical was lower than that of BHT. Previous studies demonstrated that the antioxidant activity of polysaccharides has been attributed to their composition and structural features. 26 The results implied that NTWP-II could have stronger ability to donate electron or hydrogen. 25 Hydroxyl radical could cause severe damage to adjacent biomolecules or cell death. 19 For hydroxyl radical, there were two types of antioxidation mechanism: one suppresses the generation of the hydroxyl radical, and the other scavenges the hydroxyl radicals generated. Thus, it forms a stable radical to terminate the radical chain. It has been found that the antioxidant activities of polysaccharides are affected by various factors such as chemical contents, molecular mass and structure. 27 The scavenging effect on hydroxyl radical of NTWP-II was shown in Fig. 8b. The hydroxyl radical, known to be generated through the Fenton reaction in this system, was scavenged by samples. For NTWP-II, the effects of scavenging hydroxyl radicals were in a concentration-dependent manner from 0.04-3mg mL À1 . The IC 50 value of NTWP-II was 0.82 mg mL À1 . According to the previous study, polysaccharides with moderate molecular weights have been found to have strong antioxidant activity. In this study, NTWP-II showed strong hydroxyl radical scavenging activity. This might be attributed to moderate molecular weight and high contents of hydrogen. However, the relationships between antioxidant activity and physicochemical property or structural features of the polysaccharide have not been comprehensively understood. 28 Superoxide anion radicals are weak oxidants and are thus not harmful to the body. However, its combination with hydroxyl molecules may damage DNA and other biomolecules. Scavenging effects of NTWP-II and V C on superoxide radical were shown in Fig. 8c. For these two samples, scavenging activities of NTWP-II and V C followed a dose-dependent manner at all tested concentrations. The scavenging effects signicantly increased with increasing concentration from 0.04 mg mL À1 to 3.0 mg mL À1 and IC 50 values for scavenging superoxide radical was 0.35 mg mL À1 . The results indicated that NTWP-II has a noticeable superoxide radical scavenging activity.
Chelation is an important biological process, as iron is an essential metallic element for respiration, oxygen transport, and the activity of many enzymes for metabolism. Among the transition metals, Fe 2+ is known as the most powerful prooxidant due to its high reactivity, which accelerates the reactions of oxidation through the Fenton reaction. 29 The ferrous ion chelating activity of NTWP-II at different concentrations was shown in Fig. 8d and compared with EDTA in the equivalent concentration as a positive control. The metal chelating ability was recognized as a correlative activity to antioxidant. From Fig. 8d, we knew the chelating activity of NTWP-II exhibited a much weaker metal chelating ability. The chelating rate for NTWP-II was only 57.3% even at 5.0 mg mL À1 , while that of EDTA was 60.1% at 0.6 mg mL À1 . Metal chelation activity reduces the concentrations of the transition metals, thereby catalysing lipid peroxidation and thus reducing oxidative reactions. The reducing power results revealed that NTWP-II can act as electron donor compound, react with free radicals and convert them to more stable products terminating, therefore, reactions of the radical chain.

Conclusion
In order to increase the extraction yield of polysaccharides form the fruit of Nitraria tangutorum Bobr. The optimal extraction conditions with hot water extraction were optimized by BBD. The optimum variables given by BBD were as follows: extraction time 7 h, extraction temperature 60 C, liquid-solid ratio 15 : 1. Under these conditions, the experimental yield was obtained as 14.01 AE 0.11% (n ¼ 3), which corresponded with the predicted value well. The puried homogeneous polysaccharides NTWP-II was successfully obtained by DEAE-cellulose column and Sephadex G-200 column chromatography. In addition, chemical analysis indicated that the weight average molar mass (M w ), polydispersity (PD, M w /M n ) and z-average radius of gyration (R z ) were 2.29 Â 10 5 , 1.32, 15.22 by SEC-LLS. Monosaccharide analysis revealed that NTWP-II was composed of rhamnose, arabinose, mannose, glucose and galactose with the molar ratio of 1.14 : 2.5 : 3.00 : 2.69 : 5.28. Based on the analysis of monosaccharide composition, 1D and 2D NMR, the backbone structure of NTWP-II consisted mainly of (1/3)-b-galactose. Besides, and linked with a-mannose, a-arabinose and b-galactose by the (1/3) glycosidic linkage and (/1)-a-glucose of the side chain was obtained. NTWP-II exhibited positive radical scavenging activities against DPPH radical, superoxide anion and hydroxyl radicals, and metal chelating ability in vitro. Thus, NTWP-II could be explored as a natural antioxidant food ingredient. These might further provide theoretical basis for the widely application of NTWP-II in medicine and health care products.

Conflicts of interest
There are no conicts to declare.