Extraction of Bioactive Components from Chamaenerion angustifolium (L.) Scop. with Choline Chloride and Organic Acids Natural Deep Eutectic Solvents

Chamaenerion angustifolium (L.) Scop. (fireweed) is a perennial herbaceous plant of the Onagraceae family widely used in folk and scientific medicine. It is a promising source of bioactive components. One of the modern trends in extraction is the use of natural deep eutectic solvents (NADESs) combined with ultrasound-assisted extraction (UAE). However, works devoted to the extraction of biologically active substances from C. angustifolium using NADESs are scarce. The aim of this work is a comprehensive study of UAE of bioactive components from C. angustifolium using NADESs based on choline chloride and malonic, malic, tartaric, and citric acids. The antioxidative properties, total phenols, and flavonoids content were estimated for NADES-based extracts. The reference solvents were water and 90% v/v ethanol. Volatile extracted components were identified using GC-MS. The kinetics of the UAE were studied at 45 °C for 20–180 min with water added to 30 wt% NADES. The power of the ultrasound was 120 W, and the frequency was 40 kHz. It was found that NADES choline chloride + citric acid is more effective for the extraction of bioactive components. For this, NADES UAE conditions were optimized following a Box–Behnken design of the experiment and a response surface methodology. The temperature ranged from 30 to 60 °C, the time of extraction ranged from 20 to 60, and the addition of water ranged from 30 to 70 wt%. We established the optimal extraction conditions: temperature 58 °C, time of extraction 35 min, and 70 wt% water. The obtained results expand the knowledge about the use of NADES for the extraction of biologically active compounds from cheap and available plant raw materials.


Introduction
Chamaenerion angustifolium (L.) Scop. (fireweed) is a perennial herbaceous plant of the Onagraceae family. This plant is widely distributed throughout the northern hemisphere, including large parts of the boreal forests. Fireweed is a widely used folk medicine plant; it also is used for drinks and honey production. Due to folk medicine, this plant improves genitourinary system action, stimulates immunity, prevents viral diseases, and increases vitality [1]. It contains several bioactive compounds such as phenolic acids, cinnamic acid derivatives, ascorbic acid, flavonoids (quercetin, kaempferol, and myricetin derivatives), tannins, coumarins, carotenes, sterols, and triterpenes; they provide a high pharmacological activity such as antioxidant, antibacterial, anti-inflammatory, analgesic, and anticancer properties [2][3][4][5][6]. The content of flavonoids in this plant is one of the highest among the herbaceous plants of the Kola Peninsula [7]. C. angustifolium quickly accumulates a large

Preparation of NADESs
Choline chloride served as an HBA. It was mixed with the appropriate amount of malonic (MA), malic (Mal), or citric (CA) acids as the HBD at a molar ratio of 1:1 or with tartaric (Tar) acid at a molar ratio of 2:1, as stated in [13]. The reagents were weighed by an Acculab Atilon laboratory balance (Sartorius group, Goettingen, Germany). The mixtures were heated at 80 • C for several hours to obtain homogeneous liquids. NADESs are highly viscous liquids, they may be dissolved with water to decrease the viscosity [25]. In our work, 30-70 wt% water was added to NADESs. The prepared NADESs were characterized using the 1 H NMR method with a 500 MHz Bruker AVANCE III NMR spectrometer (Bruker, Billerica, MA, USA) equipped with a BBI probe head with an inner coil for 1H nuclei and using the FTIR method with a Bruker Alpha FT-IR spectrometer (Bruker, Billerica, MA, USA) with a Platinum ATR attachment, as described in our previous work [26].

Ultrasound-Assisted Extraction and Optimization Using Box-Behnken Design
In order to choose the most suitable NADES for further work, it was necessary to compare the efficiency of extraction of biologically active components with various solvents. However, in order to properly compare the efficiency of the solvents, it was necessary to get close enough to achieving phase equilibrium, which meant we had to evaluate the kinetics of extraction. Theoretically, in dependence on the solvent nature, two cases are possible: the kinetic curves do not cross (Figure 1a) or they do cross (Figure 1b). In case (a), the extraction yield for solvent 1 is higher than that for solvent 2 all the time (both at times t 1 and t 2 ), and the curves do not cross. In case (b), the curves do cross. At the beginning of the extraction (t 1 ) the extraction yield for the solvent 1 is higher than that of solvent 2, but as the phase equilibrium is achieved (t 2 ), the extraction yield for solvent 2 is higher than that of solvent 1. It is important to know when the phase equilibrium is achieved so that we can adequately compare the extraction efficiency of the different solvents. Thus, it is important to evaluate the extraction yield at the correct extraction time.
Extraction was described in detail in [26]. Briefly, the ratio of plant material and the solvent was 1:10 (w/v). UAE was performed in the VBS-3DP thermostated ultrasound bath (Vilitech, Moscow, Russia) with an ultrasound power of 120 W and an ultrasound frequency of 40 kHz. After extraction in 1.5 mL Eppendorf tubes, the samples were centrifuged at 4000 rpm for 5 min. The extraction kinetics were researched at 45 • C using NADESs with 30% water. The reference solvents were water and 90% ethanol. The extraction times were 20, 40, 60, 120, and 180 min. We determined that the best time is 60 min.
The optimization of extraction was carried out by the Box-Behnken design of experiment with three levels of three parameters. Due to the method, the central point was replicated five times. The parameters and their levels are presented in Table 1. The temperature limits were chosen to be 30-60 • C following our previous work [26]. The addition of 30% by weight of water to the studied NADES was chosen as the minimum additive since a smaller additive leads to a significant increase in viscosity, which makes it difficult to work with the extragent. The maximum water addition was 70 wt%. Extraction was described in detail in [26]. Briefly, the ratio of plant material and the solvent was 1:10 (w/v). UAE was performed in the VBS-3DP thermostated ultrasound bath (Vilitech, Moscow, Russia) with an ultrasound power of 120 W and an ultrasound frequency of 40 kHz. After extraction in 1.5 mL Eppendorf tubes, the samples were centrifuged at 4000 rpm for 5 min. The extraction kinetics were researched at 45 °C using NADESs with 30% water. The reference solvents were water and 90% ethanol. The extraction times were 20, 40, 60, 120, and 180 min. We determined that the best time is 60 min.
The optimization of extraction was carried out by the Box-Behnken design of experiment with three levels of three parameters. Due to the method, the central point was replicated five times. The parameters and their levels are presented in Table 1. The temperature limits were chosen to be 30-60 °C following our previous work [26]. The addition of 30% by weight of water to the studied NADES was chosen as the minimum additive since a smaller additive leads to a significant increase in viscosity, which makes it difficult to work with the extragent. The maximum water addition was 70 wt%.

Chemical Analysis
All chemical analyses are described in detail in [26]. The total phenolic content (TPC) was determined by a reaction with a Folin-Chocalteu reagent [27]. The total flavonoid content (TFC) was determined using the complexation reaction with aluminum chloride [7]. For these analyses, row extracts were diluted 100 times. The total antioxidant capacity (TAC) was estimated using the phosphomolybdate method [28]. For the TAC, 5 μL of raw extract was mixed with 2 mL of reagent solution without additional dissolving. Free radical scavenging (FRS) was estimated with the DPPH method [23] for extracts diluted 400 times.
The calibration curves for TPC were prepared using solutions of gallic acid (10-200 μg·mL −1 ) and were expressed as mg/g of gallic acid equivalent (GAE) per one gram of plant weight; curves for TFC used rutin (100-1000 μg·mL −1 ) and were expressed as mg/g of rutin equivalent (RE) per one gram of plant weight; and curves for TAC used ascorbic

Chemical Analysis
All chemical analyses are described in detail in [26]. The total phenolic content (TPC) was determined by a reaction with a Folin-Chocalteu reagent [27]. The total flavonoid content (TFC) was determined using the complexation reaction with aluminum chloride [7]. For these analyses, row extracts were diluted 100 times. The total antioxidant capacity (TAC) was estimated using the phosphomolybdate method [28]. For the TAC, 5 µL of raw extract was mixed with 2 mL of reagent solution without additional dissolving. Free radical scavenging (FRS) was estimated with the DPPH method [23] for extracts diluted 400 times.
The calibration curves for TPC were prepared using solutions of gallic acid (10-200 µg·mL −1 ) and were expressed as mg/g of gallic acid equivalent (GAE) per one gram of plant weight; curves for TFC used rutin (100-1000 µg·mL −1 ) and were expressed as mg/g of rutin equivalent (RE) per one gram of plant weight; and curves for TAC used ascorbic acid (1.25-10 mg mL −1 ) and were expressed as mg/g of ascorbic acid equivalent (AAE) per one gram of plant weight. It should be particularly noted that calibration curves were obtained for each type of solvent, and it was established that the nature of the solvents influences the parameters of the linear regression.

GC/MS Analysis
The components of extracts were analyzed with a GCMS-QP2010 (SHIMADZU, Kyoto, Japan) equipped with an HP-5MS column (Agilent J&W, Santa Clara, CA, USA) with a (5%-phenyl)-methylpolysiloxane phase (film thickness 0.25 µm), length 30 m, internal diameter 250 µm. Helium was used as the carrier gas at a flow rate of 1 mL/min. The injector temperature was set at 280 • C. The temperature program was as follows: isothermal step at 40 • C for 3 min, ramp 40 to 280 • C with a heating rate of 10 • C/min, and isothermal step at 280 • C for 5 min. Thus, the total running time was 32 min.
For the gas-chromatography analysis, 0.1 mL of water-based or NADES-based extracts were mixed with 1 mL of hexane and were intensively shaken for a few minutes. After this, the hexane phase was analyzed. Ethanolic extracts were injected into the chromatographic system directly. The aliquot volume was 1 µL. It was injected in split mode 1/10 for water-based and NADES-based extracts and 1/80 for ethanolic extracts.
Each of the peaks of the mass spectra were compared with the NIST 27.147 database. Since no internal standard was included, only compounds with similarity indices of 80% and above were taken into account.

Kinetical Analysis
The extraction kinetics data were approximated by a second-order model according to [29,30]. Such an approach was successfully applied to describe extraction kinetics in NADES [31]. According to a second-order model, the dependence of the concentration (Y t ) is expressed in terms of the equilibrium concentration (Y(eq)) and the rate constant (k): The parameters k and Y(eq) can be found form the linearized Equation (1) in the coordinates t /Y t vs. t: Thus, the equilibrium concentration is found from the slope of the line, and the rate constant is found from the shift of the line.

Statistical Analysis
The measurements for the comparison of extraction effectiveness were made three times for each analysis. The statistical comparison was performed using a factorial analysis of variance (ANOVA) and a post-hoc Tukey's HSD test. These two methods were used to estimate statistically significant differences at p ≤ 0.05. The calculations were performed using MS Excel 2010 (Microsoft, Redmond, DC, USA) with the Real Statistics Resource Pack add-on [32]. For the Box-Behnken design of the UAE condition optimization, an ANOVA test and a response surface methodology were used via DesignExpert 11 (Stat-Ease, Minneapolis, MN, USA) software.

Results and Discussion
The results of the kinetic parameters estimation are presented in the Table 2. The higher the value of a rate constant, the lower the equilibrium achieving time. Among the TPC, the highest rate constant was observed for ethanol (2.5×10 −3 g·mg −1 ·min −1 ), and the lowest was observed for NADESs (0.6-1.2 × 10 −3 g·mg −1 min −1 ). This may be partly related with the fact the NADESs are more viscous than ethanol. However, it may be noted that the rate constant for water was lower than for NADES choline chloride + malonic acid. For the TFC, the highest rate constants were obtained for ethanol (2.7 × 10 −3 g·mg −1 ·min −1 ) and NADES choline chloride + citric acid (3.6 × 10 −3 g·mg −1 ·min −1 ). For the TAC, the rate constant for water (144.6 × 10 −3 g·mg −1 ·min −1 ) was much higher than for other solvents. The rate constants for DPPH were very different for different solvents. The highest values were obtained for water (155.6 × 10 −3 g·mg −1 ·min −1 ) and NADES choline chloride + tartaric acid (186.4 × 10 −3 g·mg −1 ·min −1 ), and the lowest one value was obtained for NADES choline chloride + malic acid (2.3 × 10 −3 g·mg −1 ·min −1 ). Generally, it may be noted that the rate constants for TAC and DPPH were much higher than for TPC and TFC, which may indicate that the antioxidant and antiradical activities do not only depend on the polyphenol or flavonoid contents in extracts. Some substances responsible for antioxidant properties were extracted faster than the polyphenolic compounds.
The kinetic curves for TPC, TFC, TAC, and DPPH ( Figure 2) show a good quality of approximation. Moreover, it should be noted that in some cases the curves for several solvents intersect, and this was mentioned above as the theoretically assumed case (Figure 1b). For example, the curve for NADES choline chloride + malic acid intersects the curve for NADES choline chloride + malonic acid in the case of flavonoid extraction kinetics. In general, we can assume that 60 min is enough time for the main part of the extraction to be completed and the system to approach the state of equilibrium. Extracts obtained during this time can be compared according to the analyzed parameters to assess the effectiveness of certain solvents. Table 2. Parameters of second-order kinetic model applied for TPC, TFC, TAC, and DPPH; "(Eq)" means "equilibrium values"; k: rate constant. flavonoid contents in extracts. Some substances responsible for antioxidant properties were extracted faster than the polyphenolic compounds. The kinetic curves for TPC, TFC, TAC, and DPPH ( Figure 2) show a good quality of approximation. Moreover, it should be noted that in some cases the curves for several solvents intersect, and this was mentioned above as the theoretically assumed case ( Figure  1b). For example, the curve for NADES choline chloride + malic acid intersects the curve for NADES choline chloride + malonic acid in the case of flavonoid extraction kinetics. In general, we can assume that 60 min is enough time for the main part of the extraction to be completed and the system to approach the state of equilibrium. Extracts obtained during this time can be compared according to the analyzed parameters to assess the effectiveness of certain solvents. Thus, the effectiveness of different solvents was compared after 60 min of extraction. Moreover, using the calculated equilibrium values of TPC, TFC, TAC, and DPPH, it was possible to compare the experimental and calculated effectiveness of the different solvents. Figure 3 presents the experimental values of TPC, TFC, TAC, and DPPH for 60 min of extraction (± SD) are along with the calculated values. Despite the calculated equilibrium values being a little higher than the experimental ones (because 60 min is, strictly speaking, not enough time for the full equilibrium setting), the general trend is the same for both types of data. The results for TFC can be considered as the exception: the calculated equilibrium values for ethanol and water were higher than for NADESs.
The values of TPC in the case of NADESs choline chloride + citric and tartaric acids Thus, the effectiveness of different solvents was compared after 60 min of extraction. Moreover, using the calculated equilibrium values of TPC, TFC, TAC, and DPPH, it was possible to compare the experimental and calculated effectiveness of the different solvents. Figure 3 presents the experimental values of TPC, TFC, TAC, and DPPH for 60 min of extraction (± SD) are along with the calculated values. Despite the calculated equilibrium values being a little higher than the experimental ones (because 60 min is, strictly speaking, not enough time for the full equilibrium setting), the general trend is the same for both types of data. The results for TFC can be considered as the exception: the calculated equilibrium values for ethanol and water were higher than for NADESs.   The values of TPC in the case of NADESs choline chloride + citric and tartaric acids reached 250-300 mg GAE mL −1 , which was almost twice as much as for ethanol and water (near 150 mg GAE mL −1 ). The TFC values for all solvents were near 60-70 mg RE mL −1 , and it may be concluded that flavonoids are extracted from this plant equally well with water or ethanol and NADESs. The highest TAC value was obtained for NADES choline chloride + citric acid (above 50 mg AAE mL −1 ). The DPPH values for NADESs were higher than for ethanol and water, and again, the highest values were obtained for NADESs choline chloride + citric and tartaric acids.
The ANOVA and the Tukey's HSD test (p < 0.05) were carried out to evaluate the differences between solvents (Table 3). It was observed that the TPC values were significantly different for all pairs of solvents, with the exception of the ethanol-water pair. The differences for TFC values were insignificant for most solvent pairs, with the exception of the pairs of NADESs choline chloride + citric acid and malic acid, water, and ethanol-water. For the TAC, there were no differences between ethanol, water, and NADESs with malonic or malic acids, but NADESs with citric and tartaric acids showed higher TAC values. For the DPPH values, ethanol and water showed equal effectiveness, as did NADESs with citric, malonic, and tartaric acids.
Thus, the most promising NADES for further investigations is choline chloride + citric acid. Since it was found from kinetic experiments that the main part of the extraction takes place within an hour, this time was chosen as the limit in the optimization experiments.
Following the response surface methodology, the responses were described with polynomial equations. The second-order response surface model was chosen as the most suitable for all responses types; the parameters were found by experimental data approximation: where A is the temperature, B is the time, and C is the water content in NADES.
The results of the ANOVA test (Table 4) demonstrate a good approximation for TPC and TFC but not for TAC and DPPH. This may be due to the fact that the TAC parameters depend on different groups of substances and the overlap of the different extraction processes leads to more complex dependencies. For TPC, A, C, and C 2 are significant model terms; for TFC, C, B 2 , and C 2 are significant model terms; for TAC, C and B 2 are significant model terms; and for DPPH there are no significant model terms. The value of R 2 was 0.8961 for TPC, 0.8270 for TFC, 0.7679 for TAC, and 0.6166 for DPPH. The relatively high p-values for the "Models" term and their insignificance for TAC and DPPH may be a common tendency for these extract properties. Similar results can be found in [33].
The dependence of TPC, TFC, TAC, and DPPH on the temperature, time of extraction, and water content in NADES was illustrated in the response surface contour plots generated by the model for the extraction. They are presented in Figures 4-7. The response surfaces of TPC, TFC, and TAC displayed similar behaviors, while DPPH showed distinct response profiles.
Increases in the temperature and water content increased polyphenol yields. A similar trend was observed for flavonoid yields. Moreover, the higher the temperature, the greater the TAC and DPPH values, which may seem somewhat strange since some substances that exhibit antioxidant properties are thermolabile.
From Equations (3)-(6), the optimal conditions of extraction were found: temperature 58 • C, time of extraction 35 min, and 70 wt% water. At these conditions, TPC reached 301 mg GAE mL −1 , TFC reached 74 mg RE mL −1 , TAC reached 54 mg AAE mL −1 , and DPPH reached 70%. Table 5 shows that the composition of the ethanol extract was very different from other extracts, while the NADES extracts were similar in composition to the aqueous extracts. The GC-MS chromatograms and peaks identification for each solvent are also presented in the Supplementary Materials (S1-S6).
The main components of the ethanol extract were 5-(hydroxymethyl)-2-furancarboxaldehyde, 1,2,3-benzenetriol, octadecanal, and tricosanol. The main components of water-and NADES-based extracts were bis(2-ethylhexyl) ester of adipic acid, 3-(octadecyloxy)propyl ester of oleic acid, α-farnesene, and squalene. For the water extracts and NADES choline chloride + malic acid, the identical components were 2-hexanone, 3,3-dimethyl-2-hexanone, 4-butoxy-2-butanone, and 3-hexen-2-one. Moreover, these components were not extracted by the other studied solvents. Specific components of the extract based on NADES choline chloride + citric acid were aliphatic hydrocarbons such as heptadecane, hexadecane, and eicosane.  Table 4. Model summary and analysis of variance (ANOVA) of TPC, TFC, TAC, and DPPH of the Chamaenerion angustifolium (L.) Scop. leaf extracts. TPC  TFC  TAC  DPPH  TPC  TFC  TAC  DPPH  TPC  TFC  TAC  DPPH  TPC  TFC  TAC  Some of the volatile substances found in this work were previously found in extracts and essential oils of C. angustifolium. Thus, in [1], the presence of myristic, lauric, palmitic, and oleic acids was mentioned. Palmitic, oleic, and stearic acids were found in C. angustifolium samples from the Ternopil region (Ukraine) [3]. Benzeneacetaldehyde was found in ethanolic extracts of C. angustifolium collected in Lithuania, but no other components common to this work have been described [4]. Aliphatic hydrocarbons (heptadecane, hexadecane, eicosane, nonacosane, and heneicosane), and aliphatic alcohols (phytol, tricosanol, and hexacosanol) were identified in [6] in C. angustifolium collected in the Central Siberian Botanical Garden, Siberian Branch, Russian Academy of Sciences, as in this work.

F-Value p-Value
Thus, it should be noted that plant samples of C. angustifolium from different regions differ in composition, and a detailed phytochemical analysis of the volatile components as well as the influence of growing conditions on the chemical composition of this plant needs more detailed study.

Conclusions
Bioactive components were, for the first time, extracted with the NADES from Chamaenerion angustifolium (L.) Scop. (fireweed) growing in the Kola Peninsula. Solvents were formed with choline chloride and malonic, malic, tartaric, and citric acids, and with common solvents, such as water and 90% v/v ethanol.
The kinetics of the extraction processes were studied, and the parameters of the secondorder reaction kinetics model for each solvent were obtained. It was shown that in 60 min the extraction of polyphenols and flavonoids almost reaches equilibrium. The values of the antioxidant and antiradical activity reach a plateau within a shorter time, which indicates a high rate of extraction of substances responsible for these parameters.
It was found that NADES choline chloride + citric acid is the most effective solvent for the extraction of biologically active compounds. The extraction conditions were optimized using BBD. The optimal conditions are: temperature 58 • C, extraction time 35 min, and 70 wt% water.
A GC-MS analysis of the extracts obtained using various solvents was performed. It was found that the compositions of the NADES-based extracts were close to the aqueous extracts. The main components of the NADES-based extracts were bis(2-ethylhexyl) ester of adipic acid, 3-(octadecyloxy)propyl ester of oleic acid, α -farnesene, and squalene.
The obtained results expand our knowledge about the use of NADES for the extraction of biologically active compounds from various types of plant raw materials and can potentially be useful for the development of environmentally friendly methods for the production of natural biologically active additives and pharmaceuticals.