Optimization of High-Pressure Ultrasonic-Assisted Simultaneous Extraction of Six Major Constituents from Ligusticum chuanxiong Rhizome using Response Surface Methodology

High-pressure ultrasound-assisted extraction technology was applied to extract ferulic acid, senkyunolide I, senkyunolide H, senkyunolide A, ligustilide and levistolide A from Ligusticum chuanxiong rhizomes. Seven independent variables, including solvent type, pressure, particle size, liquid-to-solid ratio, extraction temperature, ultrasound power, and extraction time were examined. Response Surface Methodology (RSM) using a Central Composite Design (CCD) was employed to optimize the experimental conditions (extraction temperature, ultrasonic power, and extraction time) on the basis of the results of single factor tests for the extraction of these six major components in L. chuanxiong rhizomes. The experimental data were fitted to a second-order polynomial equation using multiple regression analysis and were also examined using appropriate statistical methods. The best extraction conditions were as follows: extraction solvent: 40% ethanol; pressure: 10 MPa; particle size: 80 mesh; liquid-to-solid ratio: 100:1; extraction temperature: 70 °C; ultrasonic power, 180 W; and extraction time, 74 min.

RSM is a collection of statistical and mathematical techniques that has been extensively used for developing, improving, and optimizing extraction conditions [39][40][41][42][43]. The this article we report the first application of high-pressure ultrasound-assisted extraction combined with RSM to the optimization of the extraction of the main active ingredients of Chuanxiong Rhizoma.  [7] Note: 0.5 g sample to calculate the volume of extraction solvent.
In this study, seven factors, namely solvent type, pressure, particle size, liquid-to-solid ratio, extraction temperature, ultrasound power, and extraction time, were investigated first to optimize the extraction solvent, pressure, particle size, and liquid-to-solid ratio. Additionally, the levels of the response surface experimental design factors (temperature, power, and time) were determined according to the extraction temperature, ultrasonic power, and extraction time optimized in the single factor tests.

Effect of Solvent Type on Extraction
Solvent selection is important in the extraction of compounds from botanical materials [30]. Figure 1 shows that under similar extraction conditions 40% EtOH is superior to other solvents in extracting ferulic acid and senkyunolide H from L. chuanxiong rhizome. Both 40% EtOH and 60% EtOH are superior to the other solvents in extracting senkyunolide I, but without significant differences in their extraction yields (analysis of variance, p ≤ 0.05, the same below). Additionally, 80% MeOH and 60% EtOH are superior to other solvents in extracting levistolide A, but not significantly different from 40% EtOH. Moreover, 80% MeOH and 40% EtOH are superior to other solvents in extracting senkyunolide A, but without significant differences. The 60% EtOH solvent is superior to other solvents in extracting ligustilide and does not significantly differ from 40% EtOH.  Ethanol is a Good Manufacturing Practice-compliant solvent. Based on the results, 40% ethanol was selected as the extraction solvent to extract the six major constituents using high-pressure ultrasonic-assisted extraction in subsequent experiments. Previous studies have reported the use of ethanol (mainly for extracting ferulic acid) [1,7,[9][10][11][12][13][14][15][17][18][19][20]22], water [5,6,16], methanol [8,44], and protic ionic liquids [21] as extraction solvents for extracting ingredients from L. chuanxiong rhizomes. However, methanol is toxic, water is unable to dissolve the hydrophobic components, protic ionic liquids only extracted three lactone constituents, so although the ethanol concentrations using the above method were somewhat variable, ethanol was chosen as the best extraction solvent among the extraction solvents listed in Section 3.3.1.

Effect of Pressure on Extraction
The high pressure that promotes ultrasonic cavitation also creates shear forces that break cell walls mechanically. The cavitation bubbles from the increased pressure causes the solvent to penetrate deeper into the raw plant material/intracellular space [34,45]. Figure 2 presents how different pressure levels affect the extraction yield of the six major constituents in L. chuanxiong. Pressures of 8 and 10 MPa are superior to other pressure levels in extracting senkyunolide H and levistolide A, but the extraction yields between these two pressures do not differ significantly. The extraction yields of ferulic acid, senkyunolide I, senkyunolide A, and ligustilide continued to increase with increasing pressure and do not significantly differ between 4 MPa to 10 MPa.  The appropriate particle size is fundamental to obtain optimal extraction, and varied particle sizes could significantly affect the extraction yield. In this study, 80 mesh was better than other particle sizes in extracting ferulic acid, levistolide A, senkyunolide A, and ligustilide ( Figure 3), whereas 60 mesh was better than other particle sizes in extracting senkyunolide I and senkyunolide H. These two mesh sizes do not differ significantly.

Effect of Liquid-to-Solid Ratio on Extraction
For efficient extraction, the solvent volume must be sufficient to ensure complete immersion of materials, and an extraction solvent deficiency can lead to lower extraction yields (incomplete extraction) of ingredients [9][10][11][12]15,17,18,20,22], but redundant solvent may also lead to lower extraction yields and solvent waste [31]. Therefore, the liquid-to-solid ratio must be appropriate. The effect of liquid-to-solid ratio on the extraction yield of the six constituents was investigated, and the results are shown in Figure 4. The 100:1 (v/m) proportion is better than other liquid-to-solid ratios in extracting ferulic acid, senkyunolide H, ligustilide, and levistolide A. The 100:1 and 200:1 proportions are better than other liquid-to-solid ratios in extracting senkyunolide I, but without a statistically significant difference. The 40:1 ratio is better than other ratios in extracting senkyunolide A. To minimize solvent requirements and without compromising responses, the liquid-to-solid ratio of 100:1 (v/m) was selected.

Effect of Temperature on Extraction
The effect of temperature on the extraction yield of the six main constituents from L. chuanxiong rhizome was investigated, and the results are shown in Figure 5.

Effect of Ultrasonic Power on Extraction
Low ultrasonic power reduces extraction yield. However, excessively high power results in energy wastage. Therefore, the optimal ultrasonic power should be determined. The effect of ultrasonic power on the extraction yield of the six main constituents from the L. chuanxiong rhizomes was investigated, and the results are shown in Figure 6. The extraction yields of the six constituents from the rhizomes evidently increased with increasing ultrasonic power, but no increase was observed above 175 W. Few studies on effects of ultrasonic power for extraction constituents of L. chuanxiong are available but the extraction yields were very low [9]. Extraction time, an important parameter in solvent extraction, has two phases, i.e., the dissolution of soluble components on the surfaces of the sample particle and the mass transfer of the solute from the plant matrix into the solvent by diffusion and osmotic processes [24]. The longer the time, solvent and sample contact more fully, and this might accelerate the absorption of solvent, soften the plant tissues and weaken the cell wall integrity, as well as enhance ingredient solubility, thus larger amounts of substances would distribute to the solvent. However, too long a time may lead to lower process efficiency and wasted time. Therefore, the extraction time must be appropriate. The effects of extraction time on the extraction yields of six constituents from L. chuanxiong rhizome are shown in Figure 7. The extraction yields of ferulic acid, senkyunolide I and senkyunolide A increased as extraction time was prolonged from 30 min to 60 min and peaked at 60 min. However, the extraction yields of senkyunolide H, levistolide A, and ligustilide increased as the extraction time increased from 30 min to 70 min and peaked at 70 min.

Model Fitting of Parameters based on the Extraction Yields of the Six Constituents
The responses of the extracts (ferulic acid, senkyunolide I, senkyunolide H, senkyunolide A, ligustilide and levistolide A) in each run are presented in Table 2. The regression coefficients and results from analysis of variance (ANOVA) of the second order polynomial models (Y = A 0 + A 1 X 1 + A 2 X 2 + A 3 X 3 + A 11 X 1 2 + A 22 X 2 2 + A 33 X 3 2 + A 12 X 1 X 2 + A 13 X 1 X 3 + A 23 X 2 X 3 ) for ferulic acid, senkyunolide I, senkyunolide H, senkyunolide A, ligustilide and levistolide A are summarized in Table 3. The regression parameters of the surface response analysis of the models, namely, the linear, and quadratic, and their corresponding interaction terms have significant differences (p ≤ 0.0001, p ≤ 0.01 or p ≤ 0.05). The fitness of the model was evaluated through the lack of fit test (p < 0.05), which indicates the adequacy of model to predict accurately the variation [46]. The models were used to construct three-dimensional response surface plots to predict the relationship between the independent and dependent variables.

Effect of Process Variables on the Extraction Yield of Ferulic Acid
The experimental data were examined through regression analysis, and the coefficients of the model were evaluated for significance. Temperature (X 1 ), ultrasonic power (X 2 ), and time (X 3 ) significantly affected the extraction yield of ferulic acid (Y 1 , Table 3), with corresponding contribution rates of 1.80, 2.32, and 2.31. These results indicate that ultrasonic power and extraction time have the greatest impact on the extraction yield of ferulic acid. Table 2. Extraction yield of response surface CCD (n = 3) and expressed as mean ± SD (Units: %).   The three-dimensional response surface plots in Figure 8 illustrate the relationship between the extraction yield of ferulic acid and the experimental variables. Figure 8a illustrates the interaction effect of temperature and power, when the time was set at its 0 level (70 min), on the extraction yield of ferulic acid. Ferulic acid yields gradually increased with temperature and ultrasonic power and peaked at approximately 70 °C to 80 °C and 175 W to 200 W. However, extraction yield of ferulic acid began to decrease after further increase in these parameters.
The interaction effects between temperature and time on the ferulic acid extraction yield when the power was set at its 0 level (175 W) are presented in Figure 8b. The yield of ferulic acid increased and peaked at 60 °C to 70 °C from 65 min to 75 min.
The interaction effects of power and time at 60 °C (0 levels) on the extraction yield of ferulic acid are presented in Figure 8c. Strong interaction was observed when the power was within 175 W to 200 W and the time ranged 65 min to 75 min, which contributed to the increased extraction yield.
The ferulic acid regression model for the statistical frequency method of analysis with 95% confidence interval was obtained (X 1 : 1.33 to 1.59, X 2 : -0.82 to 0.24 and X 3 : −0.65 to 0.38) when extraction yield >0.25% (n = 21). Therefore, the optimal conditions were 73.32 °C to 75.98 °C, 154.50 W to 181.03W, and 63.50 min to 73.77 min.

Effect of process variables on the extraction yield of senkyunolide I
The senkyunolide I extraction yield results are presented in Table 2. The regression analysis results indicated that the main extraction parameters of senkyunolide I were temperature (X 1 ), ultrasonic power (X 2 ), and time (X 3 ). The relationships between the extraction yield of senkyunolide I (Y 2 , Table 3) and the variables are shown in Figure 9. The contributions of extraction temperature, ultrasonic power, and extraction time were 2.40, 2.16 and 1.22, respectively. Extraction temperature had the greatest impact on the senkyunolide I extraction yield. The senkyunolide I regression model for the statistical frequency method of analysis with 95% confidence interval was obtained (X 1 : 0.31 to 1.06, X 2 : −0.53 to 0.65 and X 3 : 0.60 to 1.30) when extraction yield > 0.09% (n = 23). Hence, the optimal conditions were 63.14 °C to 70.57 °C, 161.80 W to 191.18 W and 76.02 to 82.96 min.

Effect of Process Variables on the Extraction Yield of Senkyunolide H
The extraction yield of senkyunolide H is presented in Table 2. Regression analysis showed that the extraction yield (Y 3 , Table 3) was significantly affected by the temperature (X 1 ), ultrasonic power (X 2 ), and time (X 3 ), with corresponding contribution rates of 2.67, 1.75, and 1.67, respectively. Extraction temperature exhibited the greatest impact on the extraction yield of senkyunolide H.
The relationship of the extraction yield of senkyunolide H and process variables are depicted in Figure 10. The effects of temperature and ultrasonic power on extraction yield at 0 level fixed time are shown in Figure 10a. The extraction yield of senkyunolide H gradually increased with temperature and power and peaked at approximately 70 °C to 75 °C and 175 W to 200 W. Further increases in these parameters resulted in decreased extraction yield of senkyunolide H. The senkyunolide H regression model for the statistical frequency method of analysis with 95% confidence interval was obtained (X 1 : 0.01 to 0.85, X 2 : −0.19 to 1.02, and X 3 : −0.30 to 0.94) when extraction yield > 0.26% (n = 23). Hence, the optimal conditions were 60.11 °C to 68.51 °C, 170.28 W to 200.50 W, and 77.05 min to 79.36 min.

Effect of Process Variables on the Extraction Yield of Senkyunolide A
The extraction yield of senkyunolide A is presented in the Table 2. Regression analysis showed that the extraction yield (Y 4 , Table 3) was significantly affected by temperature (X 1 ), ultrasonic power (X 2 ), and time (X 3 ), with corresponding contribution rates of 2.65, 2.05, and 1.81. Extraction temperature had the greatest impact on the extraction yield of senkyunolide A.
The relationship of the extraction yield of senkyunolide A and process variables are depicted in Figure 11. The effects of temperature and ultrasonic power on extraction yield at 0 level fixed time are shown in Figure 11a. The senkyunolide A regression model for the statistical frequency method of analysis with 95% confidence interval was obtained (X 1 : −0.32 to 0.48, X 2 : −0.12 to 1.14, and X 3 : 0.14 to 1.24) when extraction yield >0.72% (n = 21). Hence, the optimal conditions were 56.80 °C to 64.80 °C, 172.08 W to 203.58 W, and 71.39 min to 82.38 min.

Effect of Process Variables on the Extraction Yield of Ligustilide
Regression analysis was performed using the experimental data, and the model coefficients were evaluated for significance. Temperature (X 1 ), ultrasonic power (X 2 ), and time (X 3 ) significantly affected the extraction yield of ligustilide (Y 5 , Table 3), with corresponding contribution rates of 2.72, 2.61, and 1.77. Extraction temperature showed the greatest impact on the extraction yield of ligustilide.
The three-dimensional response surface plots (Figure 12) illustrate the relationship between the extraction yield of ligustilide and experimental variables. These plots present the response as a function of two factors with another variable constant at its 0 level. Figure 12a shows the interaction effect between temperature and power, when the time was set at its 0 level (70 min) in the extraction of ligustilide. Extraction yield of ligustilide gradually increased with temperature and ultrasonic power and peaked at approximately 60 °C to 70 °C and 175 W.  The interaction effect between temperature and time at 0 level (175 W) power on the extraction of ligustilide is presented in Figure 12b. The response surface plot shows that the extraction yield of ligustilide increased and reached the maximum level at 60 °C to 70 °C for the time interval of 60 min to 80 min.
The interaction effect between power and time at 0 level (60 °C) temperature on the extraction yield of ligustilide is presented in Figure 12c. Strong interaction was observed when the power reached 175 W to 200 W and time reached 60 min to 80 min, which contributed to the increase in extraction yield. The ligustilide regression model for the statistical frequency method of analysis with 95% confidence interval was obtained (X 1 : 0.41 to 0.48, X 2 : −0.12 to 0.71, and X 3 : −0.69 to 0.44) when extraction yield > 2.72% (n = 24). Hence, the optimal conditions were 64.14 °C to 70.37 °C, 171.98 W to 192.70 W, and 63.08 min to 74.42 min.

Effect of Process Variables on the Extraction Yield of Levistolide A
The extraction yield of levistolide A is presented in the Table 2. Regression analysis showed that the extraction yield (Y 6 , Table 3) was significantly affected by temperature (X 1 ), ultrasonic power (X 2 ), and time (X 3 ), with corresponding contribution rates of 1.78, 2.44, and 1.72. Ultrasonic power had the greatest impact on the extraction yield of levistolide A.
The relationship of the extraction yield of levistolide A and process variables are depicted in Figure 13. The effects of temperature and ultrasonic power on extraction yield at 0 level fixed time are shown in Figure 13a. Extraction yield of levistolide A gradually increased with temperature and power and peaked at about 70 °C to 80 °C and 175 W to 200 W. The response surface plots for the effect of time on extraction yield (Figure 13b,c) at constant ultrasonic power and constant temperature showed that, appropriate extraction time (60 min to 80 min) had positive effects on the extraction yield. The levistolide A regression model for the statistical frequency method of analysis with 95% confidence interval was obtained (X 1 : 0.18 to 0.97, X 2 : 0.18 to 0.97, and X 3 : −0.99 to 0.21) when extraction yield > 0.04% (n = 22). Hence, the optimal conditions were 61.84 °C to 69.73 °C, 179.60 W to 199.33 W, and 60.01 min to 72.10 min. Table 4 indicates the optimum high-pressure ultrasound-assisted conditions for the extraction of the six major constituents from L. chuanxiong using RSM. Thirty six accurately weighed samples (0.5 g each filtered through 80 mesh sieve) were added to 50 mL 40% EtOH, pressurized 10 MPa, and divided into six groups. A set of samples was extracted under the optimum single ingredient conditions (obtained using statistical software), and the predicted results fitted well with the experimental results (Table 4). Based on the statistics frequency method, the following optimum extraction conditions were obtained: extraction temperature, 70 °C; ultrasonic power, 180 W; and extraction time, 74 min. Six accurately weighed samples (0.5 g each and filtered through 80 mesh sieve) were added to 50 mL 40% EtOH, pressurized 10 MPa, and extracted. The optimum extraction conditions were obtained through the statistics frequency method ( Table 4). The calculated extraction yields of statistical frequency condition were compared with the actual extraction yields under optimum conditions.

Optimization of the Extraction Process
The results showed that the significant results were obtained in the extraction yields of senkyunolide A and ligustilide, contrary to the other four constituents, are not significant. Therefore, the optimum extraction conditions for simultaneous extraction of the six ingredients can be obtained through the statistical frequency method, and optimum extraction condition for one ingredient can be obtained through the single component method.

General Information
Methanol (HPLC grade) and acetonitrile (HPLC grade) were purchased from Fisher Scientific, Inc. (Pittsburgh, PA, USA). Ethanol (HPLC grade) and acetic acid (AC grade) were purchased from Chengdu Kelong Chemical Factory (Chengdu, China). Water (HPLC grade) was purified through Milli-Q Plus system (Millipore, Bedford, MA, USA). Standard samples of ferulic acid, ligustilide and levistolide A were obtained from National Institutes for Food and Drug Control (Beijing, China). Senkyunolide I, senkyunolide H, and senkyunolide A were purchased from Beijing Beiyanxinglv Biotechnology Co., Ltd (Beijing, China). L. chuanxiong Hort. rhizomes were collected from the experimental field in Dujiangyan County, Chengdu City (Sichuan Province, China) during harvest time. The materials were confirmed based on morphologic, microscopic, and physiochemical analyses according to the Chinese Pharmacopoeia [1]. Voucher specimens were deposited in the College of Agronomy, Sichuan Agricultural University. Samples were sun-dried and ground into powder.

HPLC Quantitative Analysis of the Six Major Constituents
HPLC analysis was performed on Agilent 1200 series HPLC-DAD system comprising a vacuum degasser, quaternary pump, autosampler, thermostated column compartment and diode array detector (Agilent, Palo Alto, CA, USA) to analyze the samples extracts. The extracts (10 μL) were injected and separated on a Symmetry C 18 column (250 mm × 4.6 mm, 5 μm, Waters, Milford, MA, USA). The mobile phase consisted of methanol (A), acetonitrile (B), and 1% aqueous acetic acid (C) using a gradient program of 40: 20: 40 (A: B: C, v/v/v) at 0 min to 5 min, 60% to 100% (B), and 40% to 0% (C) from 5 min to 30 min. The flow rate was 1 mL/min, and the column temperature was 35 °C. The detection wavelength was set to 0 min to 4.3 min 321 nm and to 275 nm from 4.31 min to 30 min.

Single Factor Tests
To investigate the extraction of the six major constituents under different conditions, single factor tests were employed first to determine the optimal extraction solvent, pressure, particle size, liquid-tosolid ratio, extraction temperature, ultrasonic power and extraction time. The six major ingredients of L. chuanxiong rhizome were extracted using a SB-5200 DTD ultrasonic extractor (Ningbo Science Biotechnology Co., Ltd, Zhejiang, China). The author-customized high-pressure extraction tubes containing the sample extract were then partially immersed in the ultrasonic bath, and the liquid in the tubes was kept at the same level as that of the water in the bath. Finally, each extracted solution was filtered through a 0.45 μm syringe filter and collected into a 1.5 mL volume vial.

Pressure Selection
Each extract (0.5 g of dried powder was filtered through 60 mesh) was dissolved in 40% EtOH (25 mL, based on the results from Section 3.3.1.) and prepared for seven levels of pressure treatment (−0.05, 0.1, 2, 4

Liquid-to-Solid Ratio Selection
Each extract [0.5 g of dried powder (filtered through 80 mesh based on the results of Section 3.3.3.)] was dissolved in different volumes of 40% EtOH (100, 50, 40, 30, 20 and 10 mL) and pressurized to 10 MPa. The experiments were conducted in triplicate (18 experimental treatments). The working frequency was 40 kHz, ultrasonic power was 200 W, temperature was 40 °C, and extraction time was 60 min.

Extraction Temperature Selection
Each extract (0.5 g of dried powder) was mixed 40% EtOH (50 mL, based on the results from Section 3.3.4.) and pressurized to 10 MPa. The experiments were conducted at different extraction temperatures of 20, 30, 40, 50, 60, and 70 °C. The working ultrasonic power and time were 200 W and 60 min, respectively. The experiments were also performed in triplicate (18 experimental treatments).

Ultrasonic Power Selection
Each extract (0.5 g of dried powder) was mixed 40% EtOH (50 mL) and pressurized to 10 MPa. Different ultrasonic power levels of 75, 100, 125, 150, 175, 200, 225, and 250 W were applied. The working temperature and time were 40 °C and 60 min, respectively. The experiments were performed in triplicate (24 experimental treatments).

Extraction Time Selection
Each extract (0.5 g of dried powder) was mixed 40% EtOH (50 mL) and pressurized to 10 MPa. The working ultrasonic power and temperature were 200 W and 40 °C, respectively. The extractions were performed for 30, 40, 50, 60, 70, 80, and 90 min. The experiments were conducted in triplicate (21 experimental treatments).

Central Composite Design
The single factor tests demonstrated the marked interaction among extraction temperature, ultrasonic power, and extraction time. Hence, RSM was used to investigate the influence of these three independent variables on the six extracts. The experiments were performed according to a rotatable central composite design (CCD) [46]. The coded values of the experimental factors and factor levels ( Table 5) were used in the response surface analysis that was run 20× (Table 2) and performed in triplicate. The results of the response surface design were analyzed using Design Expert Software, version 8.0 (Stat-Ease, Inc., Minneapolis, MN, USA), and the contribution rate was calculated using the following equation: , and F jj = t jj 2 ), respectively; δ j , δ ij , and δ jj were calculated F value of the independent variables, the interaction effect terms, and the quadratic effect terms, respectively.

Conclusions
In this study, RSM was successfully applied to optimize the high-pressure ultrasonic-assisted extraction of the six major constituents from L. chuanxiong rhizome. Extraction solvent type, pressure, sample particle size, liquid-to-solid ratio, extraction temperature, ultrasonic power, and extraction time played significant roles in the extraction of constituents. We were able to extract higher yields of the constituents from L. chuanxiong, while using a green extraction solvent, less solvent, lower extraction temperature, while simultaneously reducing considerably the extraction time. According to the statistical frequency method the optimum extraction conditions for the simultaneous extraction of the six components were 40% EtOH, 10 MPa, 80 mesh, 100: 1, 70 °C, 180 W, and 74 min.