Simultaneous Optimization for Ultrasound-Assisted Extraction and Antioxidant Activity of Flavonoids from Sophora flavescens Using Response Surface Methodology

The ultrasonic-assisted extraction process and antioxidant activity of flavonoids from Sophora flavescens were investigated in this study. In order to optimize the extraction of flavonoids from Sophora flavescens, the influence of extraction time, methanol concentration, ultrasonic temperature, and solvent-to-material ratio was analyzed. Results showed that the extraction yields reached a maximum with the extraction time of 30 min, methanol concentration of 80%, temperature of 80 °C, and solvent-to-material ratio of 26 mL/g. The flavonoids were determined by HPLC, and the mean yields of trifolirhizin, formononetin, isoxanthohumol, maackiain, and kurarinone under the optimal conditions were 2.570, 0.213, 0.534, 0.797, and 3.091 mg/g, respectively. The evaluation of vitro antioxidant activity exhibited Sophora flavescens flavonoids had a strong 1,1-diphenyl-2-picrylhydrazyl (DPPH) and hydroxyl radical-scavenging ability with IC50 of 0.984 and 1.084 mg/g, respectively. These results indicate that ultrasonic-assisted extraction is an efficient approach for the selective extraction of flavonoids, and response surface methodology further optimized the extraction.


Introduction
Sophora flavescens, the dry root of Sophora flavescens Ait., is a widely used traditional Chinese medicine that has the functions of clearing heat and diuresis, expelling dampness and anthelmintic [1]. The chemical constituents in Sophora flavescens are mainly alkaloids and flavonoids. Previous studies and clinical research mostly focused on the alkaloids [2][3][4][5]. However, recent attention has been attracted by isopentenyl flavonoids contained in Sophora flavescens [6,7]. Isopentenyl flavonoids from Sophora flavescens have a variety of biological and chemical activities, and are the reason for the comprehensive evaluation of its quality [1]. Furthermore, it has been well demonstrated that isopentenyl flavonoids have extensive pharmacological activity and clinical applications such as bacteriostasis, antitumor, anti-inflammation, anti-arthritic, and in the regulation of dyslipidemia [8][9][10].

Effect of Ultrasonic Time
The effect of ultrasonic time on total flavonoids extraction yield is shown in Figure 1a. The extraction was continued for 20, 30, 40, 50, 60, 70, and 80 min, respectively. The extraction temperature was fixed to 50 • C and ultrasonic power fixed to 120 W. Sixty per cent methanol was selected as solvent, and the liquid-to-material ratio was fixed to 20:1 mL/g, all other conditions being equal. With the prolongation of ultrasonic extraction time, the yield of flavonoids increased. After the extraction time lasted for 50 min, the yield of flavonoids reached its maximum, followed by a slow declining trend. It is speculated that prolonged ultrasonic time may degrade flavonoids or accompany the dissolution of other impurities. For comprehensive consideration, ultrasonic time of 30-50 min was used in the RSM experiment.

Effect of Solvent
In general, solvent is considered an important parameter for UAE: it may affect the solubility of the target ingredients [20]. Generally, ethyl acetate, ethanol, and methanol are the most hackneyed solvents for extraction, but different solvents are suitable for different flavonoids compounds. According to the theory of similar dissolve mutually and previous research, the extraction using methanol has a higher yield and better solubility of flavonoids [21]. Based on the structure of the five flavonoids compounds (Figure 2), it is speculated that higher methanol ratio is propitious to the dissolution of these substances. In this single-factor experiment, we investigated the effect of different methanol concentration (30-90% (v/v)) on flavonoid extraction. As shown in Figure 1b, the yield of flavonoids kept steady increasing when the concentration of methanol ranged from 30% to 60%. When methanol concentration exceeded 60%, the yield of flavonoids decreased weakly. It could be caused by the increasing dissolution of other alcohol-soluble substances. Therefore, the variable range of methanol concentration used in the RSM experiments was selected as 60-80%.

Effect of Temperature
As the results shows in Figure 1c, a significant increase in flavonoid yield was observed by increasing the extraction temperature from 20 • C to 70 • C because higher extraction temperatures accelerate molecular motion, penetration, dissolution, and diffusion in favor of releasing the flavonoids. The peak yield (1.85 ± 0.072 mg/g) was reached at an extraction temperature of 70 • C, and the yield decreased with a further increase in the extraction temperature. The yield of flavonoids in 80 • C was significantly higher than yield in 50 • C (p < 0.05), thus the variable temperature range used in the RSM experiments was selected as 60-80 • C.

Effect of Liquid-to-Material Ratio
UAE procedure is executed at different liquid-to-material ratios while fixing the other extraction parameters as follows, extraction temperature, 50 • C; methanol concentration, 60%; and extraction time, 30 min. As shown in Figure 1d, extraction yield was affected by the liquid-to-material ratio. With an increased ratio from 10:1 mL/g to 20:1 mL/g, the flavonoids yield increased and reached the maximum value (0.965 ± 0.062 mg/g) at 20:1 mL/g. When the liquid-to-material ratio was further increased, an induction in flavonoids was observed. It indicates that a liquid-to-material ratio that is too low or too high may not favor the movement of the plant cells and the flavonoids to the solvents under ultrasonic treatment [15]. No significant difference was observed between an extraction yield of 25:1 mL/g and 35:1 mL/g (p > 0.05). Therefore, the variable range of liquid-to-material ratio used in the RSM experiments was selected as 20:1-30:1 mL/g because solvent consumption can increase production cost. In this single-factor experiment, we investigated the effect of different methanol concentration (30-90% (v/v)) on flavonoid extraction. As shown in Figure 1b, the yield of flavonoids kept steady increasing when the concentration of methanol ranged from 30% to 60%. When methanol concentration exceeded 60%, the yield of flavonoids decreased weakly. It could be caused by the increasing dissolution of other alcohol-soluble substances. Therefore, the variable range of methanol concentration used in the RSM experiments was selected as 60-80%.

Effect of Temperature
As the results shows in Figure 1c, a significant increase in flavonoid yield was observed by increasing the extraction temperature from 20 °C to 70 °C because higher extraction temperatures accelerate molecular motion, penetration, dissolution, and diffusion in favor of releasing the flavonoids. The peak yield (1.85 ± 0.072 mg/g) was reached at an extraction temperature of 70 °C , and the yield decreased with a further increase in the extraction temperature. The yield of flavonoids in 80 °C was significantly higher than yield in 50 °C (p < 0.05), thus the variable temperature range used in the RSM experiments was selected as 60-80 °C .

Effect of Liquid-to-Material Ratio
UAE procedure is executed at different liquid-to-material ratios while fixing the other extraction parameters as follows, extraction temperature, 50 °C ; methanol concentration, 60%; and extraction time, 30 min. As shown in Figure 1d, extraction yield was affected by the liquid-to-material ratio. With an increased ratio from 10:1 mL/g to 20:1 mL/g, the flavonoids yield increased and reached the maximum value (0.965 ± 0.062 mg/g) at 20:1 mL/g. When the liquid-to-material ratio was further increased, an induction in flavonoids was observed. It indicates that a liquid-to-material ratio that is too low or too high may not favor the movement of the plant cells and the flavonoids to the solvents under ultrasonic treatment [15]. No significant difference was observed between an extraction yield of 25:1 mL/g and 35:1 mL/g (p > 0.05). Therefore, the variable range of liquid-to-material ratio used in the RSM experiments was selected as 20:1-30:1 mL/g because solvent consumption can increase production cost.

Model Fitting
The BBD in the optimization experiment consisted of four factors, three levels, and five center point runs that were carried out in triplicate. The experimental conditions and results of 29 runs are presented in Table 1. The results of the analysis of variance (ANOVA) are shown in Table 1. p-values were used to check the significance of each coefficient. Specifically, values of "probability (p) > F" of less than 0.05 and 0.01 indicate that the model terms are significant and highly significant, respectively, and values greater than 0.05 indicate that the model terms are not significant [23,24]. F values revealed that the model was statistically significant (p < 0.05) for five flavonoids of trifolirhizin, formononetin, Isoxanthohumol, maackiain, and kurarinone. The lack of fit of each model was not significant (p > 0.05), indicating that the developed model adequately explains the relationship between the independent variables and responses. The values of determination coefficients (R 2 ) and adjusted determination coefficients (Adj. R 2 ) are shown in Table 1. These indicated the response was sufficiently explained by the model. The generated response surface 3D graphs corresponding to each response showed the interactive effects of the variables, if any (Figure 3a-c).

Model Fitting
The BBD in the optimization experiment consisted of four factors, three levels, and five center point runs that were carried out in triplicate. The experimental conditions and results of 29 runs are presented in Table 1. The results of the analysis of variance (ANOVA) are shown in Table 1. p-values were used to check the significance of each coefficient. Specifically, values of "probability (p) > F" of less than 0.05 and 0.01 indicate that the model terms are significant and highly significant, respectively, and values greater than 0.05 indicate that the model terms are not significant [23,24]. F values revealed that the model was statistically significant (p < 0.05) for five flavonoids of trifolirhizin, formononetin, Isoxanthohumol, maackiain, and kurarinone. The lack of fit of each model was not significant (p > 0.05), indicating that the developed model adequately explains the relationship between the independent variables and responses. The values of determination coefficients (R 2 ) and adjusted determination coefficients (Adj. R 2 ) are shown in Table 1. These indicated the response was sufficiently explained by the model. The generated response surface 3D graphs corresponding to each response showed the interactive effects of the variables, if any (Figure 3a    Note: * p < 0.05 significant, ** p < 0.01 highly significant.

Trifolirhizin
In Table 1, the ANOVA results showed significant linear (X2 and X4) and interactive (X1X2) effects on trifolirhizin. Based on the regression coefficient (β) values, linear (X2) revealed a major effect, which was followed by X4 and X1X2. The extraction yield value of trifolirhizin can be expressed as the following second order polynomial equation.  (a) extraction time (X 1 ) and methanol concentration (X 2 ); (b) extraction temperature (X 3 ) and solvent-to-material ratio (X 4 ); and (c) methanol concentration (X 2 ) and solvent-to-material ratio (X 4 ). Note: * p < 0.05 significant, ** p < 0.01 highly significant.

Trifolirhizin
In Table 1, the ANOVA results showed significant linear (X 2 and X 4 ) and interactive (X 1 X 2 ) effects on trifolirhizin. Based on the regression coefficient (β) values, linear (X 2 ) revealed a major effect, which was followed by X 4 and X 1 X 2 . The extraction yield value of trifolirhizin can be expressed as the following second order polynomial equation.

Formononetin
The solvent-to-materials ratio exerted a linear and quadratic effect on the yields of formononetin. The extraction yield value of trifolirhizin can be expressed by the following second order polynomial As show in Table 1, the solvent-to-materials ratio (X 3 ) has remarkable significance (p < 0.01) on isoxanthohumol, and the quadratic effect of temperature (X 3 2 ) and the interaction (X 3 X 4 ) also make a difference on the extraction of isoxanthohumol. The quadratic equation obtained in terms of actual factors is described as follows

Maackiain
The linear effects of the methanol concentration (X 2 ) (p < 0.05), temperature (X 3 ) (p < 0.05), and solvent-to-material ratio (X 4 ) (p < 0.01), as well as the interaction of X 2 X 4 showed significant effects on maackiain. The quadratic equation is described as follows

Kurarinone
Kurarinone is primarily affected by methanol concentration (X 2 ) and solvent-to-liquid ratio (X 4 ) (p < 0.05). The quadratic effect of temperature(X 3 ) also works in the extraction of kurarinone. The model of kurarinone can be represented as follows

Optimal Processing Conditions and Model Verification
Based on the regression analysis and 3D surface plots of the independent variables, the optimum conditions for the maximum flavonoids (trifolirhizin, formononetin, isoxanthohumol, maackiain, and kurarinone) using comprehensive evaluation value were obtained with an extraction time 30 min, methanol concentration of 79.98% (v/v), extraction temperature of 80 • C, and solvent to material ratio of 26.25 mL/g. For the convenience of experiments, parameters were modified slightly in the verification experiment as follows, extraction time 30 min, methanol concentration of 79.98% (v/v), extraction temperature of 80 • C, and solvent-to-material ratio of 26.25 mL/g. All experiments under the optimized conditions were performed in quintuplicate, and the results are displayed in Table 2.

DPPH Radical Scavenging Activity
DPPH scavenging ability is a common used method to appraise antioxidant activity of natural compounds. The antioxidant activity of flavonoids from Sophora flavescens was evaluated with DPPH-scavenging assay, in comparison with known antioxidant, ascorbic acid (VC). As shown in Figure 4a, the DPPH radical scavenging activity of flavonoids presented notable DPPH radical-scavenging activity, and the capacity were increased with increasing concentrations. Obviously, the scavenging effects of extractions were weaker than that of VC in same concentration. However, when the concentrations of flavonoids increased to 1.00 mg/mL, the antioxidant activity (95.83 ± 0.27%) of flavonoids was approximate to that of VC (96.04 ± 2.20%) (Figure 3). These results indicated that Sophora flavescens extracts had a strong DPPH radical scavenging activity, with an IC50 value of 0.984 mg/g. activity, and the capacity were increased with increasing concentrations. Obviously, the scavenging effects of extractions were weaker than that of VC in same concentration. However, when the concentrations of flavonoids increased to 1.00 mg/mL, the antioxidant activity (95.83 ± 0.27%) of flavonoids was approximate to that of VC (96.04 ± 2.20%) (Figure 3). These results indicated that Sophora flavescens extracts had a strong DPPH radical scavenging activity, with an IC50 value of 0.984 mg/g.

Hydroxyl Radical Scavenging Activity
The hydroxyl radical, which is well known as one of the most reactive free radicals, can trigger free radical chain reactions and attack biological molecules to induce severe damage. As shown in Figure 4b, Sophora flavescens extracts exhibited a lower capacity of hydroxyl radical scavenging activity than that of VC. It displayed a concentration-dependent antioxidant activity with a similar trend to VC, and an IC50 value of 1.084 mg/g. These results proved that flavonoids obtained from Sophora flavescens possess the ability to be good antioxidants.

Hydroxyl Radical Scavenging Activity
The hydroxyl radical, which is well known as one of the most reactive free radicals, can trigger free radical chain reactions and attack biological molecules to induce severe damage. As shown in Figure 4b, Sophora flavescens extracts exhibited a lower capacity of hydroxyl radical scavenging activity than that of VC. It displayed a concentration-dependent antioxidant activity with a similar trend to VC, and an IC50 value of 1.084 mg/g. These results proved that flavonoids obtained from Sophora flavescens possess the ability to be good antioxidants.

HPLC Analysis of Flavonoids
Five flavonoids of Sophora flavescens were identified by HPLC and detected at wavelengths of 295 nm ( Figure 5). The content of trifolirhizin, formononetin, isoxanthohumol, maackiain, and kurarinone was determined by a corresponding conversion of the peak area. Pharmacological and nutritional studies have found that these five compounds exhibited various activities such as antioxidant, bacteriostasis, antitumor, and anti-inflammatory effects [8,[25][26][27]. The combined effects of these flavonoids and other compounds may affect antioxidant activities by scavenging free radicals and inhibiting lipid peroxidation [28][29][30]. Therefore, these five constituents in the extracts may be partly responsible for the antioxidant activity observed in flavonoids obtained by UAE.
nutritional studies have found that these five compounds exhibited various activities such as antioxidant, bacteriostasis, antitumor, and anti-inflammatory effects [8,[25][26][27]. The combined effects of these flavonoids and other compounds may affect antioxidant activities by scavenging free radicals and inhibiting lipid peroxidation [28][29][30]. Therefore, these five constituents in the extracts may be partly responsible for the antioxidant activity observed in flavonoids obtained by UAE.

Plant Materials
Sophora flavescens was purchased from the Traditional Chinese Herbal Medicine Co., Ltd. (Hangzhou, China) of Zhejiang Chinese Medical University and identified by Xilin Chen, the professor of Zhejiang Chinese Medical University. It was crushed using a mill (TAISITE, Tianjin, China) and sifted through a 50-mesh seive, stored in a dry place before experimental study.

Plant Materials
Sophora flavescens was purchased from the Traditional Chinese Herbal Medicine Co., Ltd. (Hangzhou, China) of Zhejiang Chinese Medical University and identified by Xilin Chen, the professor of Zhejiang Chinese Medical University. It was crushed using a mill (TAISITE, Tianjin, China) and sifted through a 50-mesh seive, stored in a dry place before experimental study.

Ultrasound-Assisted Extraction (UAE) of Flavonoids
Dry Sophora flavescens powder (0.5 g) was placed into a 50 mL stuffed conical flask and then the different solvents were added. The container was capped, and the system was started-up. Ultrasonic processing was executed in the thermostatic ultrasonic cleaner (ZL3-120, Shanghai Zuole instruments Co., Ltd., Shanghai, China), equipped with a fixed ultrasound power (120W) and digital controlled low-frequency sonotrode (40 kHz). The instrument also equipped with timing device and temperature displayed digital controller to control the medium temperature. The extraction temperature and time were set at different degrees according to different conditions. After extraction, the extraction vessels were left for several minutes to cool down to room temperature. Then each extract was filtered and methanol was added to original volume [18].

Determination of Flavonoids
The content of flavonoids was determined by spectrophotometry using the aluminum nitrate colorimetric method with some modifications [17,27]. Briefly, 1 mL extracted solution containing flavonoids was added to a 10 mL volumetric flask. Four milliliters of 60% (v/v) methanol and 0.16 mL of 5% (w/w) NaNO 2 were mixed for 5 min, and then 0.1 mL of 10% Al(NO 3 ) 3 (w/w) was added and mixed, 6 min later, 1 mL of 4% (w/w) NaOH was added. After 10 min, the absorbance of the solution at 510 nm was measured with a double beam UV/Vis spectrophotometer (TU-1900, Pgeneral, Beijing, China) against the same mixture without the sample as a blank. The calibration curve was prepared by preparing rutin solutions at concentrations from 20 to 100 µg/mL in methanol. The concentration of total flavonoids in extract was expressed as mg of rutin equivalents per gram dry weight of extract. The calibration curve was determined to y = 5.895x + 0.0733 and R 2 = 0.993, where y is absorbance value of sample and x is sample concentration (10-100 µg/mL).

BBD Experimental Design
A four-factor-three-level Box-Behnken Design (BBD) was used [20]. The effect of independent variables (extraction time, X1; extraction temperature, X2; methanol concentration, X3; and solvent-to-material ratio, X4) on the UAE of total flavonoids from Sophora flavescens was estimated. On the basis of preliminary experiments, all variables were fixed at 3 levels, denoted by −1, 0, and 1, the experimental designs of the coded (x) and actual (X) levels of variables are shown in Table 3.
The response values (Y) of trifolirhizin (Y 1 ), formononetin (Y 2 ), isoxanthohumol (Y 3 ), maackiain (Y 4 ), kurarinone (Y 5 ) in each trial were analyzed using Design Expert 8.0.6 and fitted to a second-order polynomial regression model as follows where β 0 , β i , β ii , and β ij are constant regression coefficients of the model, while Xi and Xj are the independent variables.

Assay of DPPH Radical Scavenging Activity
The scavenging effect of the flavonoids extracted by UAE on DPPH radical was based on the procedure described by previous studies with some modifications [32]. Briefly, 2.0 mL of 0.2 mM DPPH in anhydrous ethanol was added to 2.0 mL of sample at different concentrations (0.1-1.0 mg/mL). The mixture was shaken and incubated for 30 min at room temperature in dark, and the absorbance of the resulting solution was measured at 517 nm. Ascorbic acid was used as the control substance. The scavenging percentage was calculated according to the following equation.
where, A i was the absorbance of 2.0 mL sample + 2.0 mL DPPH; A 0 was the absorbance of 2.0 mL sample + 2.0 mL anhydrous ethanol; and A j was the absorbance of 2.0 mL DPPH + 2.0 mL anhydrous ethanol.

Assay of Hydroxyl Radical Scavenging Activity
Hydroxyl radical scavenging activity was determined according to the literature with some modifications [33]. Two milliliter sample solutions of different concentrations (0.1-1.0 mg/mL)-1 mL of salicylic acid-ethanol solution (9.0 mM), 1 mL of FeSO 4 solution (9.0 mM), and 3.0 mL of distilled water-were mixed in a tube. The reaction was initiated by the addition of 1.0 mL H 2 O 2 (8.8 mM) to the mixture above, and the absorbance at 510 nm was measured. The hydroxyl radical scavenging activity was calculated as follows where, A i is the absorbance of the sample; A 0 is the absorbance of the group with deionized water instead of sample; and A j is the absorbance of water instead of H 2 O 2 .

Statistical Analysis
The results are expressed as mean ±SD (standard deviation). Design-Expert 8.0.6 (Stat-Ease Inc., Minneapolis, MN, USA) was used to calculate the coefficients of the quadratic polynomial model and the optimization. p-values of less than 0.05 were considered to be statistically significant.

Conclusions
RSM is a comprehensive mathematical and statistical method that has been successfully applied for optimize multifaceted processes and evaluate the influence of multiple variables and their interactions. Experimental results indicated that both methanol concentration and solvent-to-material ratio had highly significant effects on the dependent variables. Furthermore, temperature also has an influence on the yields of maackiain, while the extraction time has little correlation with the response values. The optimum extraction conditions are as follows, extraction time, 30 min; methanol concentration, 80%; extraction temperature, 80 • C; and solvent-to-material ratio, 26:1 mL/g. Under these conditions, the verification experimental results of five flavonoids in Sophora flavescens detected by HPLC agreed with the predicted values. It can be said that UAE is an effective and practical method for simultaneously extracting five flavonoids from Sophora flavescens. Thus, this optimized extraction method could be conducive for the extraction and pharmaceutical analysis of flavonoids from Sophora flavescens.