Optimizing the Extraction of Polysaccharides from Bletilla ochracea Schltr. Using Response Surface Methodology (RSM) and Evaluating their Antioxidant Activity

Bletilla ochracea Schltr. polysaccharides (BOP) have a similar structure to Bletilla striata (Thunb.) Reichb.f. (Orchidaceae) polysaccharides (BSP). Therefore, BOP can be considered as a substitute for BSP in the food, pharmaceuticals and cosmetics fields. To the best of our knowledge, little information is available regarding the optimization of extraction and antioxidant activity of BOP. In this study, response surface methodology (RSM) was firstly used for optimizing the extraction parameters of BOP. The results suggested that the optimal conditions included a temperature of 82 ◦C, a duration of 85 min and a liquid/material ratio of 30 mL/g. In these conditions, we received 26.45% ± 0.18% as the experimental yield. In addition, BOP exhibited strong concentration-dependent antioxidant abilities in vitro. The half-maximal effective concentration (EC50) values of BOP against 1,1-diphenyl-2-picrylhydrazyl (DPPH·), 2,2′-azino-bis (3-ethylbenzothiazoline-6-sulphonate) (ABTS+·), hydroxyl (·OH) and superoxide anion (·O2−) radicals and ferrous ions (Fe2+) were determined as 692.16, 224.09, 542.22, 600.53 and 515.70 μg/mL, respectively. In conclusion, our results indicate that BOP can be a potential natural antioxidant, deserving further investigation.

Response surface methodology (RSM), a statistical model for predicting the optimal extraction parameters with high reliability and reproducibility, has been widely applied for gaining maximal yield in the extraction of polysaccharides [13]. For example, Hu et al. optimized the extraction process of the polysaccharide (CAPs) from Camptotheca acuminata fruits (microwave power, 600 W; liquid/solid ratio, 40:1 mL/g; time, 14 min; and temperature, 70 °C). Under those conditions, the yield (8.53%) was very close to the predicted value (8.61%) [14].
In this study, BOP was isolated by extracting with hot water. Box-Behnken design (BBD), a typical RSM, was firstly applied for optimizing the independent variables, including temperature, time and liquid/material. Moreover, the in vitro antioxidant activity was qualitatively and quantitatively evaluated by electron paramagnetic resonance (EPR) technique and radicalscavenging experiments (DPPH·, ABTS + ·, ·OH, ·O2 − and Fe 2+ ). In summary, the results are expected to provide the theoretical and technical supports for the further utilization of BOP.

Experimental Materials and Chemicals
The pseudobulbs of B. ochracea were procured from the National Engineering Laboratory for Resource Development of Endangered Chinese Crude Drugs in Northwest of China (Xi'an, Shaanxi Province).

Material Pretreatment and Polysaccharides Extraction
According to previous reports [15], the tubes of B. ochracea were washed, air-dried, and powdered. The powders were successively preprocessed with petroleum ether (60-90 °C) and 80% ethanol for depigmenting and delipidating. Then, the polysaccharides were extracted from B. ochracea Response surface methodology (RSM), a statistical model for predicting the optimal extraction parameters with high reliability and reproducibility, has been widely applied for gaining maximal yield in the extraction of polysaccharides [13]. For example, Hu et al. optimized the extraction process of the polysaccharide (CAPs) from Camptotheca acuminata fruits (microwave power, 600 W; liquid/solid ratio, 40:1 mL/g; time, 14 min; and temperature, 70 • C). Under those conditions, the yield (8.53%) was very close to the predicted value (8.61%) [14].
In this study, BOP was isolated by extracting with hot water. Box-Behnken design (BBD), a typical RSM, was firstly applied for optimizing the independent variables, including temperature, time and liquid/material. Moreover, the in vitro antioxidant activity was qualitatively and quantitatively evaluated by electron paramagnetic resonance (EPR) technique and radical-scavenging experiments (DPPH·, ABTS + ·, ·OH, ·O 2 − and Fe 2+ ). In summary, the results are expected to provide the theoretical and technical supports for the further utilization of BOP.

Experimental Materials and Chemicals
The pseudobulbs of B. ochracea were procured from the National Engineering Laboratory for Resource Development of Endangered Chinese Crude Drugs in Northwest of China (Xi'an, Shaanxi Province).

Material Pretreatment and Polysaccharides Extraction
According to previous reports [15], the tubes of B. ochracea were washed, air-dried, and powdered. The powders were successively preprocessed with petroleum ether (60-90 • C) and 80% ethanol for depigmenting and delipidating. Then, the polysaccharides were extracted from B. ochracea using hot water. The extracting solution was collected, filtered and precipitated by 4 volumes of ethanol at 4 • C for 24 h. Finally, the sample was re-dissolved, deproteinized by repeated freeze-thawing for ten times [16] and dialyzed against distilled water (exclusion limit 8-14 kDa) for 72 h for further removing the small molecules. The resulting solution was lyophilized and named BOP. The yield of BOP was calculated by the following equation: where A represents the weight of BOP and B represents the weight of B. ochracea powder.

Response Surface Methodology (RSM) Design
Based on the results of the single-factor test, a three-levels-three-factors Box-Behnken design (BBD) was applied to statistically optimize the extraction of the BOP. The independent variables with their levels and BBD experimental design are listed in Tables 1 and 2, respectively. All data, analysis and prediction were performed by Design-Expert software 11.0 (Stat-Ease, Minneapolis, MN, USA).  According to a reported method [17], 100 µL of different concentrations (0-2 mg/mL) of BOP and vitamin C (V C ) were mixed with 100 µL of DPPH-ethanol solution (0.3 mM) at 37 • C for 30 min in the dark. Subsequently, the absorbance was measured at 517 nm. The DPPH· scavenging rate was calculated with the following equation: where A 0 is the absorbance of the mixture without samples (replaced by distilled water), A 1 is the absorbance of the reaction solutions and A 2 is only the absorbance of the BOP.

ABTS radical Scavenging Activity
According to a reported method [18], ABTS solution (7.2 mM) and K 2 S 2 O 8 solution (in a 1:1 ratio) were co-incubated at room temperature in the dark for 24 h to produce ABTS + ·. The resulting solution was diluted to obtain an absorbance of 0.7 at 734 nm. Then, 50 µL of different concentrations (0-2 mg/mL) of BOP and V C were mixed with 150 µL of the ABTS + · solution and placed in the dark for 5 min. Subsequently, the absorbances of the reaction solutions were measured at 734 nm. The ABTS + · scavenging rate was calculated with the following equation: where A 0 is the absorbance of the mixture without samples (replaced by distilled water), A 1 is the absorbance of the reaction solutions and A 2 is only the absorbance of the BOP.

Ferrous Chelating Activity
According to a reported method [19], 100 µL of different concentrations (0-2 mg/mL) of BOP and ethylenediaminetetraacetic acid disodium salt (EDTA-2Na) were mixed with 50 µL of FeSO 4 (0.125 mM) and 50 µL of ferrozine (1 mM) and incubated for 10 min at 37 • C. Subsequently, the absorbance was measured at 562 nm. The Fe 2+ chelating rate was calculated with the following equation: where A 0 is the absorbance of the mixture without the samples (replaced by distilled water), A 1 is the absorbance of the reaction solutions and A 2 is only the absorbance of the BOP.

Hydroxyl Radical Scavenging Activity
According to a reported method [20], 50 µL of different concentrations (0-2 mg/mL) of BOP and V C were mixed with 50 µL of sodium salicylate (9 mM), 0.05 mL of FeSO 4 (9 mM) and 0.35 mL of H 2 O 2 (0.025%) and incubated for 10 min at 37 • C. Subsequently, the absorbance was measured at 562 nm. The ·OH scavenging rate was calculated with the following equation: where A 0 is the absorbance of the mixture without the samples (replaced by distilled water), A 1 is the absorbance of the reaction solutions and A 2 is only the absorbance of the BOP. For spin-trapping ·OH experiments by electron paramagnetic resonance (EPR, JES-FA series, Japan), 100 µL of BOP (2.0 mg/mL) and deionized water were mixed with the fresh mixture of FeSO 4 (100 µL, 9 mM) and the spin traps DMPO (100 µL, 50 mM). Then, H 2 O 2 was added to generate hydroxyl radicals. The reaction mixtures were immediately transferred to the EPR tube and fitted into the tester. The EPR signal was detected at 20 • C. The settings of the spectrometer were as follows: microwave power, 20 mW; modulation frequency, 100 kHz; modulation amplitude, 1.0 G.

Superoxide Anion Radical Scavenging Activity
According to a reported method [21], 50 µL of different concentrations (0-2 mg/mL) of BOP and V C were mixed with 0.05 mL of NADH (470 µM), 0.05 mL of NBT (156 µM) and 0.05 mL of PMS (60 µM) and incubated for 5 min. Subsequently, the absorbance was measured at 562 nm. The ·O 2 − scavenging rate was calculated with the following equation: Superoxide anion radical scavenging rate (%) where A 0 is the absorbance of the mixture without the samples (replaced by distilled water), A 1 is the absorbance of the reaction solutions and A 2 is only the absorbance of the BOP. For spin-trapping ·O 2 − experiments, 100 µL of BOP (2.0 mg/mL) and deionized water were mixed with a fresh solution of NADH (100 µL, 156 M) and the spin traps DMPO (100 µL, 50 mM), separately. Then, the PMS (50 µL, 60 µM) was added to initiate the generation of superoxide anion radicals. The reaction mixtures were immediately transferred to the EPR tube and fitted into the tester. The EPR signal was recorded at 20 • C. The settings of the spectrometer were as follows: microwave power, 20 mW; modulation frequency, 100 kHz; modulation amplitude, 1.0 G.

Statistical Analysis
All experiments were performed in triplicate; the results are presented as the means ± standard deviation (SD). One-way analysis of variance was used for the statistical comparisons (IBM SPSS Statistics 23). The differences between the means were considered statistically significant at P < 0.05.

Single-Factor Experiments for BOP Extraction
The influence of temperature on the yield of BOP is illustrated in Figure 2A. When time and the ration of liquid/material were fixed at 90 min and 30:1 mL/g, respectively, the yield increased with the temperature (50-80 • C) and reached the maximum value at 80 • C. However, the yield showed a decreasing tendency at the range of 80-90 • C. Thus, 70, 80 and 90 • C were applied in further experiments.  The influence of time on the yield of BOP is illustrated in Figure 2B. When the temperature and the ratio of liquid/material were fixed at 80 °C min and 30:1 mL/g, respectively, the yield increased with time (30-90 min) and reached the maximum value in 90 min. However, the yield showed a decreasing tendency during 90-150 min. Thus, 80, 90 and 120 min were applied in further experiments.
The influence of liquid/material on the yield of BOP is illustrated in Figure 2C. When temperature and time were fixed at 80 °C min and 90 min, respectively, the yield increased when the The influence of time on the yield of BOP is illustrated in Figure 2B. When the temperature and the ratio of liquid/material were fixed at 80 • C min and 30:1 mL/g, respectively, the yield increased with The influence of liquid/material on the yield of BOP is illustrated in Figure 2C. When temperature and time were fixed at 80 • C min and 90 min, respectively, the yield increased when the ration of liquid/material (10-30 mL/g) was increased and reached the maximum value for 30 mL/g. However, the yield showed a decreasing tendency for 30-50 mL/g. Thus, 20, 30 and 40 mL/g were applied in further experiments.

Model Fitting and Statistical Analysis
The predicted and experimental yields of BBD, including 17 experimental runs, are summarized in Table 2. The extraction yields of BOP ranged from 23.25% to 26.82% with a different combination of the independent parameters. Based on multiple regression analysis of the experimental data, a second-order polynomial equation was established to correlate the yield of BOP and the test variables.
where Y, B, C and D represented the yield of BOP, temperature, time and the ratio of liquid/material, respectively.
Additionally, the analysis of variance (ANOVA) for the model is presented in Table 3. A high F-value (57.98) and a low P-value (<0.0001) indicated the statistical significance of the model. The lack of fit of F-value (3.13) and P-value (0.1496) implied excellent agreement between the experimental and predicted values. Meanwhile, the coefficient of determination (R 2 = 0.9868) suggested that only 1.32% of total variables could not be explained by the model. Moreover, the high adjusted (adj) and low coefficient variation (CV = 0.8) also supported the above conclusion that we successfully developed a model with high precision, reliability, and reproducibility for optimizing the yield of the BOP. Accordingly, two linear terms (A and B), one interaction term (AB) and all quadratic terms (A 2 , B 2 and C 2 ) exhibited that the extraction rate could be highly significant [22].

Response Surface Analysis and Validation Experiments
The 3D response surface and 2D contour plots of the model were constructed to especially investigate the correlation between the response Y and three independent variables (Figure 3). The shapes of the contour plots, such as elliptical or circular, indicated whether the variable mutual interactions were significant or not [23]. Obviously, only the interaction of temperature and time was significant for the yield of BOP by observing the contours of the graphs and the density, which were in agreement with the ANOVA results. Through solving Equation (1), the optimized extractions were obtained as: temperature of 81.64 °C, time of 84.96 min and liquid/material ratio of 29.95 mL/g; the maximum yield of BOP was predicted as 26.6698%. Considering the practicality, we adjusted the extraction parameters to the temperature of 82 °C, duration of 85 min and ratio of liquid/material of 30 mL/g, and we obtained the experimental yield as 26.45% ± 0.18%, which was close the predicted value, demonstrating that the model could be used for optimizing the extraction of BOP.

DPPH, ABTS Radicals Scavenging and Fe 2+ Chelating Abilities of the BOP
The results of DPPH radical scavenging experiments are illustrated in Figure 4A. Clearly, BOP exhibited excellent DPPH· scavenging rate concentration-dependently from 62.5 to 2000 µg/mL. At the maximum concentration (2000 µg/mL), the scavenging ability was 78.06%, which was lower than that of the positive control (VC, 95.94%). Additionally, the half-effective concentration (EC50) of BOP was calculated as 692.16 µg/mL. Through solving Equation (1), the optimized extractions were obtained as: temperature of 81.64 • C, time of 84.96 min and liquid/material ratio of 29.95 mL/g; the maximum yield of BOP was predicted as 26.6698%. Considering the practicality, we adjusted the extraction parameters to the temperature of 82 • C, duration of 85 min and ratio of liquid/material of 30 mL/g, and we obtained the experimental yield as 26.45% ± 0.18%, which was close the predicted value, demonstrating that the model could be used for optimizing the extraction of BOP.

DPPH, ABTS Radicals Scavenging and Fe 2+ Chelating Abilities of the BOP
The results of DPPH radical scavenging experiments are illustrated in Figure 4A. Clearly, BOP exhibited excellent DPPH· scavenging rate concentration-dependently from 62.5 to 2000 µg/mL. At the maximum concentration (2000 µg/mL), the scavenging ability was 78.06%, which was lower than that of the positive control (V C , 95.94%). Additionally, the half-effective concentration (EC 50 ) of BOP was calculated as 692.16 µg/mL.  The results of ABTS radical scavenging experiments are illustrated in Figure 4B. Obviously, BOP exhibited outstanding ABTS + · scavenging rate concentration-dependently from 62.5 to 2000 µg/mL. At the maximum concentration (2000 µg/mL), the scavenging ability was 88.50%, which was similar to VC (96.19%). Additionally, the EC50 of BOP was calculated as 224.09 µg/mL.
The results of ferrous ion chelating experiments are illustrated in Figure 4C. Evidently, BOP exhibited remarkable Fe 2+ chelating rate concentration-dependently from 62.5 to 2000 µg/mL. At the maximum concentration (2000 µg/mL), the chelating ability was 89.94%, which was similar than the positive control (EDTA-2Na, 94.03%). Additionally, the EC50 of BOP was calculated as 515.70 µg/mL.

Hydroxyl and Superoxide Anion Scavenging Effects of BOP and EPR Detection
The results of the hydroxyl radical scavenging experiment are illustrated in Figure 5A. Apparently, BOP exhibited an excellent ·OH scavenging rate concentration-dependently from 62.5 to 2000 µg/mL. At the maximum concentration (2000 µg/mL), the chelating ability was 82.89%, which was lower than the positive control (VC 91.93%). Additionally, the EC50 of BOP was calculated as 542.22 µg/mL. Furthermore, we used EPR technology to qualitatively determine the OH radical scavenging effect of the BOP. In Figure 5a, the signal acquisition by the control group was very high and obvious, while the signal becomes relatively lower and weaker in the BOP group, indicating that the polysaccharide could effectively eliminate ·OH in vitro, which agreed with the aforementioned quantitative results.
The results of the superoxide anion radical scavenging experiment are illustrated in Figure 5B. Apparently, BOP exhibited an excellent ·O2 − scavenging rate concentration-dependently from 62.5 to 2000 µg/mL. At the maximum concentration (2000 µg/mL), the chelating ability was 89.94%, which was close than positive control of 94.03%. Additionally, the EC50 of BOP was calculated as 600.53 µg/mL. Furthermore, we used EPR technology to qualitatively determine the ·O2 − radical scavenging effect of the BOP. In Figure 5b, the signal acquisition by the control group was very high and significant, while the signal becomes relatively lower and weaker in the BOP group, indicating that the polysaccharide could effectively eliminate ·O2 − in vitro, which agreed with the aforementioned quantitative results. The results of ABTS radical scavenging experiments are illustrated in Figure 4B. Obviously, BOP exhibited outstanding ABTS + · scavenging rate concentration-dependently from 62.5 to 2000 µg/mL. At the maximum concentration (2000 µg/mL), the scavenging ability was 88.50%, which was similar to V C (96.19%). Additionally, the EC 50 of BOP was calculated as 224.09 µg/mL.
The results of ferrous ion chelating experiments are illustrated in Figure 4C. Evidently, BOP exhibited remarkable Fe 2+ chelating rate concentration-dependently from 62.5 to 2000 µg/mL. At the maximum concentration (2000 µg/mL), the chelating ability was 89.94%, which was similar than the positive control (EDTA-2Na, 94.03%). Additionally, the EC 50 of BOP was calculated as 515.70 µg/mL.

Hydroxyl and Superoxide Anion Scavenging Effects of BOP and EPR Detection
The results of the hydroxyl radical scavenging experiment are illustrated in Figure 5A. Apparently, BOP exhibited an excellent ·OH scavenging rate concentration-dependently from 62.5 to 2000 µg/mL. At the maximum concentration (2000 µg/mL), the chelating ability was 82.89%, which was lower than the positive control (V C 91.93%). Additionally, the EC 50 of BOP was calculated as 542.22 µg/mL. Furthermore, we used EPR technology to qualitatively determine the OH radical scavenging effect of the BOP. In Figure 5a, the signal acquisition by the control group was very high and obvious, while the signal becomes relatively lower and weaker in the BOP group, indicating that the polysaccharide could effectively eliminate ·OH in vitro, which agreed with the aforementioned quantitative results.

Discussion
Traditionally, most of the Chinese herbs were extracted by decoction, and polysaccharide is the main component of the aqueous extracts of these herbs. Modern pharmacological studies suggest The results of the superoxide anion radical scavenging experiment are illustrated in Figure 5B. Apparently, BOP exhibited an excellent ·O 2 − scavenging rate concentration-dependently from 62.5 to 2000 µg/mL. At the maximum concentration (2000 µg/mL), the chelating ability was 89.94%, which was close than positive control of 94.03%. Additionally, the EC 50 of BOP was calculated as 600.53 µg/mL. Furthermore, we used EPR technology to qualitatively determine the ·O 2 − radical scavenging effect of the BOP. In Figure 5b, the signal acquisition by the control group was very high and significant, while the signal becomes relatively lower and weaker in the BOP group, indicating that the polysaccharide could effectively eliminate ·O 2 − in vitro, which agreed with the aforementioned quantitative results.

Discussion
Traditionally, most of the Chinese herbs were extracted by decoction, and polysaccharide is the main component of the aqueous extracts of these herbs. Modern pharmacological studies suggest that polysaccharides possess multiple bioactivities, including antidiabetic [24], anticancer [25], antioxidant [26] and anti-inflammatory [27] effects. Particularly, antioxidants have various promising applications for food preservation and pharmacotherapy.
Hot water extraction (HWE), a conventional and convenient method, is widely applied to extract the polysaccharides from different herbs. During extraction, the temperature, time and ratio of liquid/material can significantly influence the yield. Su et al. stated that elevating the temperature could accelerate the dissolution and diffusion of polysaccharide, thus increasing the yield [28]. However, other reports indicated that the extremely high temperature can degrade the polysaccharide, thus decreasing the yield (Figure 2A) [29]. Analogically, the bioactive compounds were completely dissolved during a reasonable extraction period, whereas a long extraction period could hydrolyze the polysaccharides [30]. Thus, when the extraction time was lengthened, the yield of BOP increased at first, then reduced gradually after reaching the peak value ( Figure 2B). Furthermore, the ratio of liquid/material was a crucial factor for polysaccharide's extraction. Generally, a low ratio of liquid/material suggests that the volume of water was too low for extracting the polysaccharide. Sufficient liquid/material implied greater concentration difference between the plant cell and exterior water, accelerating the polysaccharide's diffusion from cell to external water. In contrast, with the saturation of water, the mass transfer rate changed negatively, and a high volume of liquid would result in the wastage of water and polysaccharide in the subsequent procedure ( Figure 2C) [31]. Compared to the traditional orthogonal testing, RSM exhibited many advantages, such as establishing a high precision regression equation and describing the interactions between several factors in detail. In this study, the extraction conditions of BOP were firstly optimized by RSM for achieving the maximal yield (26.45%).
Highly unstable free radicals are qualitatively detected by EPR technology. We used DMPO as traps for generating the signal intensity to indicate the existence of free radical; a high corresponding signal peak indicated a higher concentration of the free radical [32]. DPPH·, a stable reactive nitrogen species (RNS), can become a DPPH-H molecule by receiving a hydrogen atom [33]. Thus, the DPPH· scavenging experiment can assess the antioxidant ability of polysaccharide ( Figure 4A). ABTS was oxidized by K 2 S 2 O 8 , generating green ABTS + ·. Antioxidants reduced the ABTS + ·, and the system became blue in color [34]. The change in color to blue/green reveals that the polysaccharide possesses an antioxidant effect ( Figure 4B). However, both DPPH· and ABTS + · are nonphysiological radicals. Therefore, a more physiological-radical scavenging test should be conducted for comprehensively evaluating the antioxidant ability of the sample. Specifically, OH, one of the most harmful ROS, can be produced by Fenton reaction (Fe 2+ + H 2 O 2 = Fe 3+ + OH − +·OH). Meanwhile, ·O 2 − , an initial free radical, forms other stronger ROS, including singlet oxygen and H 2 O 2 .
Meanwhile, BOP exhibited an outstanding Fe 2+ chelating effect, inhibited the Fenton reaction and then decreased the production of ·OH ( Figure 4C). Additionally, ·O 2 − can be scavenged by BOP via redox reaction ( Figure 5B,a). Qu et al. extracted the BSP by infrared-assisted extraction (IRAE) and assessed their antioxidant activity in vitro. In the same concentration of sample, BOP showed stronger abilities for scavenging DPPH·, ·OH and ·O 2 − and chelating Fe 2+ than BSP [8]. Generally, the in vitro antioxidant activity of polysaccharides might be affected by multiple factors, such as molecular weight, configuration and even factors associated with the selected antioxidant evaluation system [36]. Thus, BOP exhibited better antioxidative capacity than BSP and should be developed as a potential natural antioxidant.

Conclusions
In the present study, an optimal extraction of BOP was verified by RSM with BBD. The optimum conditions were demonstrated as: temperature of 82 • C; time of 85 min; and liquid/material ratio of 30 mL/g. A practical yield of 26.45% was obtained, corresponding to the predicted value (26.6698%). Meanwhile, BOP exhibited outstanding antioxidant activity in vitro, such as scavenging of DPPH·, ABTS + ·, ·OH and ·O 2 − and chelating of Fe 2+ . Finally, our findings demonstrated that BOP could be considered as a potential antioxidant ingredient for applications in food, medical and cosmetics industries. Further investigation about the in vivo antioxidant ability of BOP is in progress.