Flavonoids from Epimedium pubescens: extraction and mechanism, antioxidant capacity and effects on CAT and GSH-Px of Drosophila melanogaster

Ultrasonic extraction

Leaf powders of 0.5 g were mixed with special volume (5, 10, 15, 30, 35 ml) of aqueous ethanol solution at certain concentration (0, 20%, 40%, 60%, 80%, 100%) in a flask, and marinated at room temperature for 30 min. Afterwards, the flask was placed in the ultrasonic cleaning bath, and leaf powders were extracted at certain temperature (15, 25, 30, 35, 45, 55 °C) for special time (10, 20, 30, 40, 60 min). In the process of extraction, the position of the flask was changed randomly at regular intervals to ensure homogenous exposure of the mixture of leaf powders and ethanol solution to ultrasound irradiation. After ultrasonication, the mixture was centrifuged at 4000 rpm for 10 min, and the supernatants were filtered through 0.45 µm microporous membranes. The filtrates were concentrated by the rotary evaporator under vacuum and then lyophilized at −48 °C until the solvent was completely removed. Extraction yield of flavonoids from E. pubescens is calculated using the following equation: (1)${\displaystyle \mathrm{Y}\left(\text{\%}\right)={\mathrm{W}}_{\mathrm{f}}÷{\mathrm{W}}_{\mathrm{l}}\times 100\text{\%}}$ where Y is extraction yield of flavonoids, W_f is the weight of flavonoids (mg), and W_l is the weight of leaf powders (mg).

ABTS^⋅+ radical-scavenging activity

Modified method of Yang et al. (2017) were used to determine ABTS^⋅+ radical-scavenging capacity. ABTS^⋅+ radical solution was prepared using the reaction between ABTS and potassium persulfate. Flavonoid solutions at different concentrations (10, 60, 110, 160, 210, 260, 310 µg/ml) of 1.0 ml were respectively mixed with ABTS^⋅+ radical solution of 4.0 ml, and incubated in the dark at 30 °C for 6 min. Immediately, the mixtures were tested for the absorbance at 734 nm. In negative and positive controls, ultrapure water and Vc solution were used instead of flavonoid solution, respectively. ABTS^⋅+ radical-scavenging capacity is calculated using the following equation: (2)${\displaystyle \mathrm{RA}\left(\text{\%}\right)=\left({\mathrm{A}}_{\mathrm{n}}-{\mathrm{A}}_{\mathrm{f}}\right)÷{\mathrm{A}}_{\mathrm{n}}\times 100\text{\%}}$ where RA (%) is radical-scavenging activity, A_f is the absorbance of the mixture of ABTS^⋅+ radical solution and flavonoid solution, and A_n is the absorbance of negative control. The value of EC₅₀ (effective concentration that reduced chemiluminescence by 50%) is calculated according to Shahidi & Zhong (2015). All experiments were carried out in triplicate, and the results of ABTS assay were plotted with means of three replicates.

DPPH^⋅ radical-scavenging activity

DPPH assay was modified from Yang et al. (2017). Flavonoid solutions of 1.0 ml at various concentrations (10, 60, 110, 160, 210, 260, 310 µg/ml) were respectively mixed with DPPH solution of 2.0 ml, and incubated in the dark at 37 °C for 30 min. The absorbance of the mixture was recorded at 517 nm. Ultrapure water and Vc solution were used in negative and positive controls, respectively. DPPH^⋅ radical-scavenging capacity is calculated using the following equation: (3)${\displaystyle \mathrm{RD}\left(\text{\%}\right)=\left({\mathrm{A}}_{\mathrm{n}}-{\mathrm{A}}_{\mathrm{f}}\right)÷{\mathrm{A}}_{\mathrm{n}}\times 100\text{\%}}$ where RD (%) is radical-scavenging activity, A_f is the absorbance of the mixture of DPPH solution and flavonoid solution, and A_n is the absorbance of negative control. EC₅₀ value is calculated as described above. The results of DPPH assay were also plotted with means of replicates.

Drosophila melanogaster strain and culture condition

Animal experiments were conducted as described with slight modification (Peng et al., 2011; Deepashree et al., 2019). The study was reviewed and approved by the Ethics Committee of Health Science Center, Xi’an Jiaotong University (Protocol no.: 2017-375). And wild-type fruit flies used for all in vivo experiments were provided by Health Science Center, Xi’an Jiaotong University. Newly eclosed fruit flies (1- to 2-day-old) were randomly divided into four groups (n = 20): male control group; female control group; male experiment group; and female experiment group. Fruit flies in the control groups were fed on the cornmeal medium (10.0% of cornmeal, 1.5% of agar powders, 13.5% of sugar, 0.5% of propanoic acid, 1.0% of dried yeast powders and 73.5% of deionized water). And fruit flies in the experiment groups were raised on the flavonoid medium, which was actually the cornmeal medium supplemented with Epimedium flavonoids (final concentration of 0.5 mg/ml). All fruit flies were reared for 20 days under a L12:D12 photoperiod (12 h light/dark cycle), at relative humidity of 60% and temperature of 25 °C. Afterwards, they were frozen to death at −20 °C for further test.

Preparation of fruit fly homogenates

According to the instructions of commercial enzyme assay kits, fruit flies were homogenized on ice in the pre-chilled kit buffer at a ratio of 1.0 g fruit flies per 33 ml, and immediately centrifuged at 3,000 rpm for 15 min at 4 °C. The supernatants were collected as fruit fly homogenates. Contents of protein in the homogenates were determined by the Bradford method (Bradford, 1976).

Determination of antioxidant enzyme activities

Influences of extraction parameters on extraction yield

It is known that ultrasonic extraction of bioactive components from plants may be affected by a great number of factors, such as ultrasound frequency, ultrasonication time, liquid to solid ratio, extractant and extraction temperature (Zhang et al., 2008; Kazemi, Khodaiyan & Hosseini, 2019). Because low frequency of ultrasound was helpful to the improvement of extraction yield of target compounds and the degradation of toxic alkaloids in many cases (Vinatoru, 2001), 30 kHz was chosen as the working frequency. Effects of four extraction parameters (i.e., ultrasonication time, liquid to solid ratio, ethanol concentration and extraction temperature) on extraction yield of flavonoids from E. pubescens were systematically investigated (Fig. 1).

Figure 1A illustrates influence of ultrasonication time on extraction yield of flavonoids from E. pubescens. Extraction yield increased with extending ultrasonication time from 10 to 30 min, and decreased thereafter. When ultrasound travels in the liquid, it may create bubbles or cavities (Luque-García & Luque de Castro, 2003). The process by which bubbles or cavities form, grow and collapse is called acoustic cavitation. Acoustic cavitation possesses physical and chemical effects (Wang & Weller, 2006). The physical effects include the production of high-speed jets of liquid, which may result in the disruptions of Epimedium samples and the efficient exudation of flavonoids out of the matrix. However, with the prolongation of ultrasonication time, more and more free radicals are generated by sonolysis of solvent (Qiao et al., 2014), which may interact with Epimedium flavonoids and induce their degradation.

Effect of liquid to solid ratio on extraction yield of Epimedium flavonoids is shown in Fig. 1B. With the increase of liquid to solid ratio from 10 to 60 ml/g, extraction yield improved gradually. The highest extraction yield was obtained when liquid to solid ratio reached 60 ml/g. This is probably due to the fact that a larger volume of ethanol solution generally dissolves a larger quantity of flavonoids. However, extraction yield declined significantly (p < 0.05) while liquid to solid ratio was more than 60 ml/g. Likewise, Xia et al. (2011) observed that extraction yield of phillyrin from Forsythia suspensa at first increased and then reduced with improving liquid to solid ratio from 5 to 30 ml/g. One of the possible reasons for the lower extraction yield of flavonoids with the higher liquid to solid ratio is that more flavonoids are degraded by exposure to ultrasound irradiation when flavonoids are dispersed in excessive amount of ethanol solution (Luque-García & Luque de Castro, 2003).

Figure 1C reveals that ethanol concentration greatly affected extraction yield of Epimedium flavonoids. Extraction yield of flavonoids notably rose as ethanol concentration increased from 0% to 40% (v/v) (p < 0.01), and then levitated in the range of 40%–60% (p > 0.05). Extraction yield achieved a maximum at ethanol concentration of 60%, and hereafter decreased with the increase of ethanol concentration from 60% to 100%. The possible reasons include (1) Epimedium samples are prone to be hydrated with 60% ethanol solution; (2) 60% ethanol solution has relatively great affinity for flavonoids; and (3) flavonoids have relatively high solubility in 60% ethanol solution (Zhang et al., 2013).

Influence of extraction temperature on extraction yield of Epimedium flavonoids is illustrated in Fig. 1D. The highest extraction yield was achieved at 30 °C. And extraction yield declined at lower or higher temperatures. In the range of 15–30 °C, relatively high temperature may increase the number of bubbles or cavities formed by ultrasound and reduce the viscosity of ethanol solution, which facilitate mass transfer between flavonoids and ethanol solution and subsequently increase extraction yield. In the range of 30–55 °C, extremely high temperature may reduce the number of bubbles or cavities and attenuate the impact of acoustic cavitation on Epimedium samples (Palma & Barroso, 2002).

BBD

To understand the interaction between extraction variables in a consecutive range and to further optimize the condition for ultrasonic extraction of flavonoids from E. pubescens, response surface methodology was carried out depending on the results of the above single factor experiments (Li et al., 2019). The four factors in BBD included ultrasonication time (X₁, 20–40 min), liquid to solid ratio (X₂, 50–70 ml/g), ethanol concentration (X₃, 40%–80%) and extraction temperature (X₄, 20–40 °C) (Table 1). The results of 29 experimental runs in BBD were presented in Table 2 and the statistical data was shown in Table A.1. R² is 0.9494, suggesting that BBD model is capable of elucidating the interaction of selected variables; and adj-R² is higher than 0.8 (adj- R² = 0.8988), indicating that the model do not include non-significant terms (Mahfoudhi et al., 2014). It means that there is a high degree of correlation between the observed and predicted values, and the fitness and reliability of the model are satisfactory. On the basis of regression analysis, BBD model can be expressed by the following equation: (6)${\displaystyle Y=6.8058-0.1546{\mathrm{X}}_1-0.0779{\mathrm{X}}_2+0.1680{\mathrm{X}}_3+0.0268{\mathrm{X}}_4+0.0012{\mathrm{X}}_1{\mathrm{X}}_2+0.0005{\mathrm{X}}_1{\mathrm{X}}_3-0.0008{\mathrm{X}}_1{\mathrm{X}}_4+0.0004{\mathrm{X}}_2{\mathrm{X}}_3+0.0002{\mathrm{X}}_2{\mathrm{X}}_4+0.0006{\mathrm{X}}_3{\mathrm{X}}_4+0.0013{\mathrm{X}}_1^2+0.0002{\mathrm{X}}_2^2-0.0018{\mathrm{X}}_3^2-0.0003{\mathrm{X}}_4^2}$ where Y is response values and X_i (i = 1, 2, 3 and 4) is the coded variables. According to Eq. (6), response values were related to the coded variables by the second-order polynomial. According to F and p values, extraction variables possessing significant effects on extraction yield were the quadratic term of ethanol concentration ($X_3^2$), followed by the linear terms of extraction temperature (X₄), ethanol concentration (X₃) and liquid to solid ratio (X₂) (Table A.1). On the contrary, the linear term of ultrasonication time (X₁), the quadratic term of ultrasonication time ($X_1^2$), the quadratic term of liquid to solid ratio ($X_2^2$) and the quadratic term of extraction temperature ($X_4^2$) and the interaction terms of two variables (X₁X₂, X₁X₃, X₁X₄, X₂X₃, X₂X₄ and X₃X₄) were found to be not significant (p > 0.05) (Table A.1).

The relationship between every two variables is illustrated by 3D response surface plots (Fig. 2) and contour plots (Fig. 1A). As seen in Fig. 2A, when ultrasonication time was less than 30 min, liquid to solid ratio hardly affected extraction yield of Epimedium flavonoids. However, when ultrasonication time exceeded 30 min, extraction yield increased from 8.71% to 9.17% with an increase of liquid to solid ratio. Figure 2B show 3D response surface plot achieved by varying ultrasonication time and ethanol concentration and fixing liquid to solid ratio (0 level) and extraction temperature (0 level), respectively. Figure 2D illustrate 3D response surface plot with respect to liquid to solid ratio and ethanol concentration, respectively. Extraction yield at first rose and then declined as ethanol concentration increased (Figs. 2B and 2D). When ethanol solution was at the lower concentration (<55%), extraction yield was almost level with the increase of liquid to solid ratio. When ethanol solution was at the higher concentration (>55%), extraction yield increased markedly (Fig. 2D). The interaction between ultrasonication time and ethanol concentration exhibited the similar trend (Fig. 2B). With the increase of extraction temperature from 20 to 40 °C (liquid to solid ratio = 60 ml/g, ethanol concentration = 60% and ultrasonication time = 20 min), extraction yield rose from 8.47% to 9.27% (Fig. 2C). As shown in Fig. 2E, an increase of extraction temperature from 20 to 40 °C (ultrasonication time = 30 min, ethanol concentration = 60% and liquid to solid ratio = 70 ml/g) promoted the improvement of extraction yield from 8.55% to 9.22%. At extraction temperature below 30 °C, extraction yield slightly increased with the increase of liquid to solid ratio. However, at 40 °C, an increase in liquid to solid ratio from 50 to 70 ml/g caused a considerable increase of extraction yield from 8.97% to 9.22% (Fig. 2E). Figures 2B, 2D and 2F indicate the quadratic effect of ethanol concentration on extraction yield. It appears that ethanol concentration is a significant factor (Fig. 2F).

The software used for RSM generated the condition for recovery of flavonoids from E. pubescens: ultrasonication time of 39.71 min, liquid to solid ratio 69.41 ml/g, ethanol concentration 63.40%, and extraction temperature 37.98 °C. And the predicted extraction yield of flavonoids was 9.40%. Taking into account the operational feasibility and simplicity, the generated condition was slightly modified: ultrasonication time of 39 min, liquid to solid ratio 70 ml/g, ethanol concentration 63%, and extraction temperature 38 °C. In order to confirm the effectiveness of the modified condition, ultrasonic extraction was performed under the modified condition. The average extraction yield was found to be 9.44%, which was very close to the predicted value (9.40%) and was higher than all of response values in 29 experimental runs in BBD (Table 2). The results provided evidence that the modified condition was the optimal one.

Comparison between ultrasonic extraction and heating extraction

Ultrasonic extraction was compared with classical heating extraction in terms of extraction yield and chemical composition (see below). The condition for ultrasonic extraction (with ultrasound) and heating extraction (without ultrasound) was as follows: extraction duration of 39 min, liquid to solid ratio 70 ml/g, ethanol concentration 63%, and extraction temperature 38 °C. Noticeably, extraction yield obtained by ultrasonic extraction (9.44 ± 0.04%) is higher than that by heating extraction (8.74 ± 0.13%) (p < 0.01).

Impact of ultrasound on particle size distribution

In order to identify the mechanism underlying ultrasonic extraction, particle sizes of Epimedium samples treated by ultrasonic extraction and heating extraction were determined by laser diffraction particle size analysis using deionized water as a dispersant. As shown in Fig. 3, frequency of small particle sizes (1.0–5.0, 5.0–10.0, 10.0–50.0 and 50.0–100.0 µm) and big particle sizes (500.0–1000.0 µm) of the samples processed by ultrasonic extraction was higher than that by heating extraction, while frequency of moderate particle sizes (100.0–250.0 and 250.0–500.0 µm) by ultrasonic extraction was lower than that by heating extraction. That is to say, with the assistance of ultrasound, the number of microparticles ranged from 1.0 to 100.0 µm (1.0–5.0, 5.0–10.0, 10.0–50.0 and 50.0–100.0 µm) in the samples most probably increases, whilst the number of macroparticles ranged from 100.0 to 500.0 µm (100.0–250.0 and 250.0–500.0 µm) most probably decreases. It is known that ultrasound may cause acoustic cavitation in the liquid. When the expanded bubbles or cavities collapsed, their potential energy would transform into kinetic energy in high-speed jets of liquid (approximately 400 km/h) (Luque-García & Luque de Castro, 2003). When Epimedium samples were stricken by high-speed jets of liquid, damage or ruptures might take place in the samples. The disruptions possibly resulted in the reduction of particle size of the samples and the improvement of their specific surface area. Thus the surface area for the sample-extractant contact increased drastically, which might aid in mass transfer between the samples and ethanol solution and hence increase extraction yield of flavonoids.