Polysaccharides from Trifolium repens L. extracted by different methods and extraction condition optimization

Four different extraction methods, including hot water extraction (HWE), ultrasonic-assisted extraction (UAE), enzyme-assisted extraction (EAE) and ultrasonic-enzyme-assisted extraction (UEAE), were applied to extract polysaccharides from Trifolium repens L. (TRPs). In addition, response surface methodology (RSM) was performed to optimize the extraction conditions of TRPs. The results showed that different extraction methods had significant effects on the extraction yields and antioxidant activities of TRPs. TRPs extracted by the EAE method (10.57%) and UEAE method (10.62%) had significantly higher extraction yields than TRPs extracted by the HWE method (8.35%) and UAE method (9.43%) (P < 0.05). However, there was no significant difference between the extraction yields of the EAE method and UEAE method (P > 0.05). TRPs extracted by the EAE method had a higher content of uronic acid and exhibited better antioxidant capacities. Therefore, EAE was selected as the optimal extraction method to extract TRPs. The optimal extraction conditions of EAE to extract TPRs were liquid–solid ratio 30 mL/g, enzymolysis time 87 min, enzyme-complex dosage 1.6% and pH 6, leading to a TRPs yield of 13.15%.

study, TRPs were extracted by four methods, that is, HWE, UAE, EAE, and UEAE. The physicochemical properties and antioxidant activities of TRPs were determined to select the suitable extraction method. In addition, the processing parameters for the suitable extraction method selected were optimized by RSM. The final purpose of this study is to provide a basis for the development and utilization of Trifolium repens L.

Results and discussion
Extraction yield, pH, solubility and chemical composition of TRPs extracted by different methods. As shown in Table 1, the yields of the four TRPs were measured as HWE-TRPs (8.35%) < UAE-TRPs (9.43%) < EAE-TRPs (10.57%) < UEAE-TRPs (10.62%). Uronic acid is one of the active parts of polysaccharides, and the higher content of uronic acid might suggest higher bioactivities of polysaccharides 19 . The uronic acid contents were significantly different in the following order: HWE-TRPs (4.04%) < UAE-TRPs (4.27%) < UEAE-TRPs (5.17%) < EAE-TRPs (5.42%) (P < 0.05). There were no significant differences on the extraction yields and uronic acid contents between the EAE method and UEAE method (P > 0.05). However, TRPs extracted by the EAE method and UEAE method had significantly higher extraction yields and uronic acid contents than TRPs extracted by the HWE method and UAE method (P < 0.05). In the process of polysaccharide extraction, improving the penetration of solvent into cells is the key factor to facilitate the extraction process. HWE could accelerate improve the extraction efficiency with the increased temperature of water. The extraction process can also be further facilitated by physical methods, such as UAE, EAE and UEAE, they can largely promote the dissolution of polysaccharides through biodegradation or mechanical destruction of plant cell walls 20 . The results in the present study indicated that the EAE and UEAE methods could significantly improve the TRPs yields and uronic acid contents compared to the HWE method and UAE method (P < 0.05), possibly because the enzyme complex (cellulase, papain and pectinase) can, separately or in conjunction with ultrasound, facilitate TRPs into the extraction solvents through enzymatic hydrolysis and cavitation effects. A similar result was also reported by Li and Wang 21 .
There were no significant differences in the contents of total polysaccharides and moisture or the pH of the four TRPs (P > 0.05). Protein was not detected in the four TRPs. No significant differences were found in the solubility time of the four TRPs at each determination temperature (P > 0.05) ( The higher molecular weight of HWE-TRPs indicated that high temperature easily caused the aggregation of polysaccharides. Long time ultrasonic extraction would destroy the molecular chains of polysaccharides and degrade polysaccharide molecules 22 . In addition, enzyme can degrade the polysaccharides to some degree 23 . The lower molecular weight of UEAE-TRPs among the four TRPs might be due to the synergistic effect of ultrasonic and enzyme complexes (cellulase, papain and pectinase), which could decompose the TRPs to form smaller ones, and finally, UEAE-TRPs showed one distinct group. This result was consistent with the report of Li and Wang 21 , who studied the impact of four extraction methods (hot water, enzyme assistance, ultrasonic assistance and ultrasonic-enzyme assistance) on the molecular weight of the Hohenbuehelia serotina polysaccharides (HSP) www.nature.com/scientificreports www.nature.com/scientificreports/ and found that UEA-HSP exhibited the largest distribution of molecular weight, which also might be observed because cellulase or ultrasound could decompose the polysaccharides to form small ones.

Monosaccharide composition of TRPs extracted by different methods. HPLC analysis showed
that the four TRPs were composed of galacturonic acid (GalA), glucose (Glc), galactose (Gal) and arabinose (Ara) (Fig. 2). The ratios of GalA, Glc, Gal and Ara in the four TRPs are shown in Table 3. The GalA content of EAE-TRPs (4.82%) was higher than that of HWE-TRPs (2.97%), UAE-TRPs (3.81%) and UEAE-TRPs (4.46%) (P < 0.05). The variations in the monosaccharide composition and proportion of polysaccharides were related  www.nature.com/scientificreports www.nature.com/scientificreports/ with the differences in the extraction techniques and temperature 24,25 . In addition, monosaccharide composition is a primary factor in understanding the bioactivities of polysaccharides. The higher uronic acid content of polysaccharides might indicate its stronger biological activities 19 .
Antioxidant activities of TRPs extracted by different methods. The 1,1-diphenyl-2-picrylhydrazyl (DPPH) free radical is a stable free radical and can accept an electron or hydrogen radical to become a stable diamagnetic molecule, which has been widely accepted as a tool for estimating the free-radical scavenging activities of antioxidants 26 . The scavenging activities of the four TRPs on the DPPH radical were presented in Fig. 3A. The EAE-TRPs showed stronger DPPH scavenging ability compared with HWE-TRPs, UAE-TRPs and UEAE-TRPs (P < 0.05) at each polysaccharides concentration. The higher content of uronic acid in EAE-TRPs might contribute to its higher DPPH radical scavenging activity. Zhang et al. indicated that polysaccharides with higher uronic acid content had higher bioactivities 27 . The highest DPPH radical scavenging activity of EAE-TRPs (61.99%) was detected at 1 mg/mL, which was not significantly different from that of vitamin C (62.83%) (P > 0.05), indicating that EAE-TRPs had strong scavenging ability on DPPH radicals.
The 2,2-azino-bis-(3-ethyl-benzothiazoline-6-sulfonic acid) (ABTS) radical is a free and stable radical cation that can react with antioxidants through acceptance of a hydrogen atom or an electron 28 . The scavenging activities of the four TRPs on the ABTS radical were presented in Fig. 3B. The results showed that the ABTS radical scavenging activities of the four TRPs were in the following order: HWE-TRPs < UAE-TRPs < UEAE-TRP s < EAE-TRPs. Uronic acid could trigger the hydrogen atom of the anomeric carbon 29 . Therefore, the higher content of uronic acid in EAE-TRPs might contribute to its higher scavenging activity among the four TRPs. At the concentration of 1 mg/mL, the scavenging activities of the four TRPs were all more than 99.00%, reaching the scavenging ability of vitamin C. These results indicated that TRPs had strong ABTS scavenging ability.
The reducing power could directly reflect the donation of an electron or hydrogen and has been widely employed to investigate the antioxidant activities of natural compounds 28 . The ferric reducing powers of the four TRPs and vitamin C were presented in Fig. 3C. The ferric reducing power of TRPs was lower than that of vitamin C (P < 0.05), which was consistent with the findings of polysaccharides extracted from Inonotus obliquus 19 and Hohenbuehelia serotina 28 on reducing powers. Among the four TRPs, EAE-TRPs and UEAE-TRPs exhibited higher ferric reducing power than HWE-TRPs and UAE-TRPs, which might be due to their lower molecular weights. There are more exposed reducing ends in the polysaccharides with lower molecular weights than those with higher molecular weights. Therefore, the polysaccharides with lower molecular weights have higher ferric reducing powers 30 .
The antioxidant ability of polysaccharides might be related to their uronic acid content, monosaccharide composition and molecular weight 31 . Therefore, the antioxidant activities of polysaccharides are not a function of a single factor but a combination of several factors 32 . In the present study, TRPs obtained by EAE exhibited higher radical scavenging activities and reducing powers and were stronger than those of the other three types of TRPs, indicating that more bioactive polysaccharides could be extracted using the EAE method. The higher antioxidant activities of EAE-TRPs might due to its highest GalA content and lower molecular weight. Another possible mechanism may involve the degradation of polysaccharides and further changes in the chemical structures induced by enzymolysis treatment, which warrant further study.
Extraction condition optimization of the EAE method to extract TRPs. According to the evaluation of the physicochemical characteristics and activities of TRPs extracted by different methods, EAE-TRPs presented higher extraction yields and antioxidant activities among the four TRPs. Therefore, EAE was selected as the optimum extraction method to extract TRPs. RSM was performed to obtain the optimal extraction parameters of EAE to extract TRPs.
The results of single-factor-tests were shown in Fig. 4. Liquid-solid ratio, enzymolysis time, enzyme-complex dosage and pH all exhibited significant effects on the TRPs yield with the same trend. These results suggested that long enzymolysis time and high enzyme-complex dosage may lead to the decomposition of TRPs 33 . In addition, pH condition is a key factor ensuring optimal enzyme activity. Different pH conditions might lead to decreased www.nature.com/scientificreports www.nature.com/scientificreports/ or lost enzyme activities 34 . Moreover, excessive extraction solvent volume could increase the diffusion distance of TRPs from plant tissue, thereby inhibiting the dissolution of TRPs 35 . The maximum TRPs yields were obtained in certain ranges for the four single factors, including 20-30 mL/g for liquid-solid ratio, 60-120 min for enzymolysis time, 1-2% for enzyme-complex dosage and 5-7 for pH.

Optimization of TRPs extraction by RSM. Based on the analysis of single-factor-tests, an RSM test was
performed to obtain the optimal extraction parameters of TRPs. The TRPs yields were presented in Table 4. The predicted values were not significantly different from the corresponding actual values (P > 0.05). A multinomial regression equation was applied to demonstrate the relationship between the variables and the response. The multinomial regression equation was as follows: where Y is the TRPs yield, X 1 is the liquid-solid ratio, X 2 is the enzymolysis time, X 3 is the enzyme-complex dosage, and X 4 is the pH. The significance of the regression model was presented in Table 5. A low probability P value (<0.0001) suggested that the model was significant. The P value of the lack of fit was 0.0711, which was higher than 0.05, indicating that the model was valid 36 . The determination coefficient (R 2 ) was 0.9413, suggesting a good correlation between the TRPs yield and the four independent variables. In addition, the P-value of linear coefficients (X 2 and X 3 ), the cross-product coefficients (X 1 X 2 , X 1 X 3 and X 3 X 4 ) and quadratic coefficients (X 1 2 , X 2 2 , X 3 2 and X 4 2 ) were less than 0.05, indicating the significant effects of the coefficients on the TRPs yield.
The 3D response surface plots (Fig. 5) and 2D contour plots (Fig. 6) were the graphical presentations of the regression model. The visual interactions between the response data and the independent variables can be presented by the 3D response surface plots and 2D contour plots. The shapes of the 2D contour plots indicated the www.nature.com/scientificreports www.nature.com/scientificreports/ significance of the interactions between two variables. The circular contour plots suggest that the interactions between the two variables are not significant while the elliptical or saddle contour plots indicate that the interaction between the two variables are significant 18 . As shown in Fig. 6, the interactions of the variables (enzymolysis time and liquid-solid ratio, enzyme-complex dosage and liquid-solid ratio, and enzyme-complex dosage and pH) were significant (P < 0.05).
According to the regression model, the optimal extraction conditions of the EAE method to extract TRPs were liquid-solid ratio 29.96 mL/g, enzymolysis time 86.78 min, enzyme-complex dosage 1.6%, and pH 5.99. A predicted-maximum TRPs yield of 13.42% was obtained. To perform the extraction procedure more expediently, the actual extraction parameters were amended slightly as follows: liquid-solid ratio 30 mL/g, enzymolysis time 87 min, enzyme-complex dosage 1.6%, and pH 6. The verification experiments were repeated three times under the optimal extraction conditions, and the actual TRPs yield was 13.15 ± 0.12%, which was not significantly different from the predicted value of 13.42%. Therefore, the regression model obtained in this trial was effective in predicting TRPs yield. In recent years, the enzyme-assisted procedure is undoubtedly an emerging technology in polysaccharide extraction due to many advantages, including lower consumption of time, solvent and cost, higher properties of yields and purity, and good intervention on molecular structures 37 . A study was designed by Wang et al. to optimize the complex enzyme-assisted extraction parameters of the alfalfa polysaccharides using RSM design, and the optimal conditions were as follows: enzyme concentration of 2.5%, 2.0%, 3.0% (weight of alfalfa) of cellulase, papain and pectase, extraction temperature 52.7 °C, extraction pH 3.87, ratio of water to raw material 78.92 mL/g and extraction time 2.73 h. Under the optimal conditions, the experimental extraction yield of alfalfa polysaccharides was 5.05% 33 .
In conclusion, TRPs extracted by the EAE and UEAE methods had higher extraction yields among the four extraction methods (HWE, UAE, EAE and UEAE) used in this study, which might be observed because the enzyme complex (cellulase, papain and pectinase) could, separately or in conjunction with ultrasound, facilitate TRPs into the extraction solvents through enzymatic hydrolysis and cavitation effects. In addition, TRPs extracted by the EAE method had better antioxidant capacities, which might due to its higher content of uronic acid and lower molecular weight. Therefore, the enzyme-assisted extraction method was chosen to extract TRPs. According to the RSM analysis, the optimal extraction conditions of EAE to extract TPRs were liquid-solid ratio  Table 4. Response and experimental design of enzyme-assisted extraction method to extract TRPs. The predicted values were not significantly different from the corresponding actual values (P > 0.05).

Materials and Methods
Four different extraction methods (HWE, UAE, EAE and UEAE) were applied to extract TRPs. The physicochemical properties and antioxidant activities of TRPs were determined to select the suitable extraction method. In addition, RSM was performed to optimize the processing parameters of the suitable extraction method selected. All measurements were repeated in triplicate.

TRPs extraction with different methods. Four extraction techniques (HWE, UAE, EAE and UEAE)
were performed following the procedures below, respectively, according to methods reported previously with some modifications 21,30 . Four corresponding polysaccharides, including HWE-TRPs, UAE-TRPs, EAE-TRPs and UEAE-TRPs, were obtained. Approximately 25 g of Trifolium repens L. powder was extracted with distilled water in a liquid material ratio of 20. HWE-TRPs was extracted in a water-bath at 90 °C for 90 min. UAE-TRPs was extracted in an ultrasonic device (KQ-100KDE, Kunshan Ultrasonic, China) working at a power of 100 W at 55 °C for 90 min. EAE-TRPs was extracted with enzyme complex (cellulase, papain and pectinase, each enzyme 1% (w/w raw material powder)) at 55 °C for 90 min. UEAE-TRPs was first extracted in an ultrasonic device working at a power of 100 W at 55 °C for 45 min, then extracted with the enzyme complex mentioned above at 55 °C for 45 min. After extraction, the treatment processes, including the separation and concentration of the supernatant, the precipitation of polysaccharides, the removal of free proteins, the dialysis and drying of the polysaccharides solution, and the calculation of TRPs yield, were performed following the methods described previously 30 . The TRPs were first purified with a DEAE-52 cellulose column before further analysis (physicochemical characteristics and bioactivities) 11 .  www.nature.com/scientificreports www.nature.com/scientificreports/ Physicochemical characteristics of TRPs extracted by different methods. The chemical compositions (the content of total polysaccharides, uronic acid, protein and moisture), pH and solubility of TRPs extracted by different methods were measured based on literature reports 30 . The molecular weight of TRPs was determined using gel filtration chromatography 11 . The monosaccharide composition of TRPs was analyzed using HPLC 38 .
Antioxidant activities of TRPs extracted by different methods. The scavenging activities of DPPH radical and ABTS radical and the ferric reducing power were determined to evaluate the antioxidant activities of HWE-TRPs, UAE-TRPs, EAE-TRPs and UEAE-TRPs 35 . Vitamin C was the positive control.
Extraction condition optimization of TRPs. After evaluating the experimental results of the physicochemical characteristics and activities of TRPs, EAE was selected as the better extraction method to extract TRPs. RSM was used to obtain the optimal extraction conditions of EAE to extract TRPs.
Approximately 25 g of Trifolium repens L. powder was extracted at 55 °C following the extraction procedure of 'TRPs extraction with different methods' . A single-factor-test was performed to determine the preliminary range of variables, including X 1 (liquid-solid ratio), X 2 (enzymolysis time), X 3 (enzyme-complex dosage) and X 4 (pH). Then, the extraction conditions for the EAE method to extract TRPs were optimized by a BBD with three levels and four independent variables (X 1 , X 2 , X 3 and X 4 ) based on the results of single-factor experiments. The TRPs yield was treated as the response. The values of the experimental variables were shown in Table 4. Design Expert Software (8.0.6) was used to design the RSM experiment, perform the statistical analysis and fit the quadratic polynomial model. A quadratic polynomial model 39  where Y is the predicted response (TRPs yield); X i and X j are the variables; β 0 , β i , β ii and β ij are the regression coefficients for intercept, linear, quadratic, and interaction terms, respectively.