Green Extraction of Alkaloids and Polyphenols from Peumus boldus Leaves with Natural Deep Eutectic Solvents and Profiling by HPLC-PDA-IT-MS/MS and HPLC-QTOF-MS/MS

Peumus boldus Mol., is a Chilean medicinal tree used for gastrointestinal and liver diseases. Such medicinal properties are associated with the presence of bioactive flavonoids and aporphine alkaloids. In this study, a new green and efficient extraction method used seven natural deep eutectic solvents (NADES) as extraction media. The extraction efficiency of these NADES was assessed, determining the contents of boldine and total phenolic compounds (TPC). Chemical profiling of P. boldus was done by high-performance liquid chromatography coupled to photo diode array detector and electrospray ion-trap mass spectrometry (HPLC-PDA-ESI-IT/MS) and electrospray ionization quadrupole time-of-flight high-resolution mass spectrometry (HPLC-ESI-QTOF-MS). Among the NADES tested, NADES4 (choline chloride-lactic acid) and NADES6 (proline-oxalic acid) enable better extraction of boldine with 0.427 ± 0.018 and 2.362 ± 0.055 mg of boldine g−1 of plant, respectively. Extraction of boldine with NADES4 and NADES6 was more efficient than extractions performed with methanol and water. On the other hand, the highest TPC were obtained using NADES6, 179.442 ± 3.79 mg of gallic acid equivalents (GAE g−1). Moreover, TPC in extracts obtained with methanol does not show significant differences with NADES6. The HPLC-PAD-MS/MS analysis enable the tentative identification of 9 alkaloids and 22 phenolic compounds. The results of this study demonstrate that NADES are a promising green extraction media to extract P. boldus bioactive compounds and could be a valuable alternative to classic organic solvents.


Introduction
Peumus boldus Mol., (Monimiaceae) is a Chilean medicinal tree used for gastrointestinal and liver diseases [1,2]. In Chile, this tree also is called Boldo, Peta, Voldu or Boldu and botanically was described for the first time by Molina in 1782. The phytochemical profiling of Boldo usually is based on its aporphine alkaloids and phenolic compounds, whose concentration varies depending on the analyzed part of the tree [3][4][5][6]. For example, the concentration of boldine is higher in the bark than in the leaves. Moreover, certain classes of alkaloids are concentrated in other parts such derived from morphinane, protoberberine, bisbenzylisoquinoline, indole and quinolizidine alkaloids using 75 different DES [26]. Authors found that DES based on Choline-lactic acid were superior for the extraction of alkaloids derived from morphinane, protoberberine, indole, and quinolizidine alkaloids. It is worth noting that one of the parameters that most influenced the extraction with DES was the water content used, which was optimized at close to 46%. In another recent work, different NADES were used for the extraction of alkaloids from Amarillydaceae [27]. These authors reported that NADES with the best efficiency to extract lycorine, crinine and crinamine were those derived from Choline Chloride: fructose and H 2 O (35%). The optimization of such solvent showed that the best conditions were: a temperature of 45 • C, extraction time of 51 min and a water content of 21%. Moreover, the same group evaluated the cytotoxicity of extracts made with NADES and surfactants for the same alkaloids, finding that the solvent can significantly influence the biological activity of the extract [28]. Considering the above mentioned variables, in this work we assess for the first time the extraction of the alkaloids and polyphenols from Boldo leaves using different NADES. These new green solvents not only improve the extraction of bioactive compounds but also allow to obtain extracts with different phytochemical profiles. Additionally, in the present work, we evaluated the presence of the main alkaloids and phenolic compounds present in P. boldus leaves extracts by using HPLC-DAD-IT-MS/MS and HPLC-QTOF-MS/MS. Plants 2020, 9, x FOR PEER REVIEW 3 of 17 derived from morphinane, protoberberine, bisbenzylisoquinoline, indole and quinolizidine alkaloids using 75 different DES [26]. Authors found that DES based on Choline-lactic acid were superior for the extraction of alkaloids derived from morphinane, protoberberine, indole, and quinolizidine alkaloids. It is worth noting that one of the parameters that most influenced the extraction with DES was the water content used, which was optimized at close to 46%. In another recent work, different NADES were used for the extraction of alkaloids from Amarillydaceae [27]. These authors reported that NADES with the best efficiency to extract lycorine, crinine and crinamine were those derived from Choline Chloride: fructose and H2O (35%). The optimization of such solvent showed that the best conditions were: a temperature of 45 °C, extraction time of 51 min and a water content of 21%. Moreover, the same group evaluated the cytotoxicity of extracts made with NADES and surfactants for the same alkaloids, finding that the solvent can significantly influence the biological activity of the extract [28]. Considering the above mentioned variables, in this work we assess for the first time the extraction of the alkaloids and polyphenols from Boldo leaves using different NADES. These new green solvents not only improve the extraction of bioactive compounds but also allow to obtain extracts with different phytochemical profiles. Additionally, in the present work, we evaluated the presence of the main alkaloids and phenolic compounds present in P. boldus leaves extracts by using HPLC-DAD-IT-MS/MS and HPLC-QTOF-MS/MS.  Figure 1 shows the structure of the two main types of alkaloids present in P. boldus. In Figure 2 an illustrative chromatogram is shown for the methanol extract (control solvent) of P. boldus registered at 304 nm. Peaks were numbered according to its elution order from 1 to 31. As summarized in Tables 1 and 2 as well as Figures S1-S4, identification was based on UV spectra obtained by HPLC-PDA, comparison of the retention times with available standards, accurate masses (HPLC-QTOF-MS/MS) and MS/MS spectra (HPLC-IT-MS/MS). Therefore, these chromatographic analyses allow the identification of nine alkaloids. Among these compounds, six were identified as  Figure 1 shows the structure of the two main types of alkaloids present in P. boldus. In Figure 2 an illustrative chromatogram is shown for the methanol extract (control solvent) of P. boldus registered at 304 nm. Peaks were numbered according to its elution order from 1 to 31. As summarized in Tables 1 and 2 as well as Figures S1-S4, identification was based on UV spectra obtained by HPLC-PDA, comparison of the retention times with available standards, accurate masses (HPLC-QTOF-MS/MS) and MS/MS spectra (HPLC-IT-MS/MS). Therefore, these chromatographic analyses allow the identification of nine alkaloids. Among these compounds, six were identified as aporphines: laurolitsine, isoboldine, boldine, isocorydine, laurotetanine and N-methyllaurotetanine; and three were identified as benzylisoquinoline derivatives: coclaurine, N-methylcoclaurine and reticuline. Additionally, 22 phenolic compounds were identified by HPLC-IT-MS/MS in Boldo leaf extracts (Table S1). Most of them had been previously reported by Simirgiotis and coworkers [6], who identified 52 phenolic compounds in male and female Boldo trees. As expected, since this author used aqueous extraction, a greater presence of proanthocyanidins oligomers (19 trimers + tetramers) is plausible. Also, a greater number of tri and tetra-glycosides of quercetin, isorhamnetin and kaempferol was observed.

Phytochemical Profiling of Peumus boldus Methanol Extract
Plants 2020, 9, x FOR PEER REVIEW 4 of 17 aporphines: laurolitsine, isoboldine, boldine, isocorydine, laurotetanine and N-methyllaurotetanine; and three were identified as benzylisoquinoline derivatives: coclaurine, N-methylcoclaurine and reticuline. Additionally, 22 phenolic compounds were identified by HPLC-IT-MS/MS in Boldo leaf extracts (Table S1). Most of them had been previously reported by Simirgiotis and coworkers [6], who identified 52 phenolic compounds in male and female Boldo trees. As expected, since this author used aqueous extraction, a greater presence of proanthocyanidins oligomers (19 trimers + tetramers) is plausible. Also, a greater number of tri and tetra-glycosides of quercetin, isorhamnetin and kaempferol was observed.  Tables  1 and 2. HPLC separation was performed in reverse phase under gradient.  [29,32] * Peak numbers are the same as the ones depicted in Figure 2.  Tables 1  and 2. HPLC separation was performed in reverse phase under gradient.  [29,32] * Peak numbers are the same as the ones depicted in Figure 2. Bold values represent the base peak of the mass spectra.  Figure 2. Bold values represent the base peak of the mass spectra.

Identification of P. boldus Phenolic Compounds
In our study, compounds corresponding to the chromatographic peaks 1, 2 and 5 show molecular ions [M + H] + at m/z 579 and the characteristic MS/MS ion fragment at 291, suggesting that these compounds are procyanidin dimers of catechin or epicatechin (compounds corresponding to the chromatographic peaks 3 and 4, m/z 290.4 and 291.1). In a previous work, after phloroglucinolysis, we demonstrate that the structure of such compounds corresponds to catechin-derived procyanidins [18]. Compounds corresponding to the chromatographic peaks 11,16,17,18,23,25,26 were identified as luteolin derivatives. Compound corresponding to chromatographic peak 11 was identified as luteolin-pentosyl-glucosyl-rhamnose with molecular ion [M + H] + at m/z 727. correspond to the loss of NH 3 and CH 3 NH 2 , respectively. These data agrees with the fragmentation patterns of coclaurine and N-methylcoclaurine [29][30][31]. In addition to this fragmentation pattern, these compounds suffer neutral loss of CH 3 OH, which explain the fragment ion at m/z 237. Subsequently this last fragment gave the ion at m/z 209 corresponding to -CO loss. Fragment ion at m/z 175 could be explained by the cleavage of the double bond present in the mother fragment ion at m/z 269. Finally, β-cleavage of the fragment ion at m/z 269 explain the origin of the ion fragment at m/z 137 (see scheme in Figure S4). On the other hand, for compound corresponding to peak 8 in positive ionization mode, the molecular formula of ion at m/z 313.13653 was predicted as C 18 H 19 NO 4 . In HPLC-PDA-IT-MS/MS analysis, peak 8 show a precursor ion at m/z 314 [M + H] + and in MS/MS gave a fragment ion at m/z 297 corresponding to the loss of 17 Da [MH + NH 3 ] + and a main fragment ion at m/z 265 generated by the sequential loss of two methyl radicals. Subsequently this last fragment gave the ion at m/z 237 corresponding to -CO loss. These data suggests that peak 8 is laurolitsine [32]. For compounds corresponding to peaks 9 and 10 in positive ionization mode, the molecular formulas of ions at m/z 327.14718 and 327.14716 were predicted as C 19  , generated by the sequential loss of two methyl radicals and 28 Da from -CO loss, respectively. These data and the comparison of elution order for standards in C-18 columns suggest that the identity of peaks 9 and 10 could unambiguously be assigned to isoboldine and boldine [29,31,33,34]. For compound corresponding to peak 15 in positive ionization mode, the molecular formula of ion at m/z 329.16323 was predicted as C 19 H 23 NO 4 . In HPLC-PDA-IT-MS/MS analysis, peak 15 showed a molecular ion at m/z 330 [M + H] + and in MS/MS a prominent product ion at m/z 192 [M + H-138] + , which is consistent with the loss of C ring with methoxyl and hydroxyl groups previous to a putative loss of 31 Da from CH 3 NH 2 (very low abundance of fragment at m/z 299). In addition, fragment ion at m/z 192 is a diagnostic ion used to confirm the presence of a methoxyl and hydroxyl groups at the A ring in benzylisoquinoline alkaloids [31,35,36]. According with MS data, peak 15 is reticuline. For compound corresponding to peak 21 in positive ionization mode, the molecular formula of ion at m/z 341.16282 was predicted as C 20 H 23 NO 4 . In HPLC-PDA-IT-MS/MS analysis, peak 21 showed a molecular ion at m/z 342.1 [M + H] + , ions at m/z 311 and 296 in MS/MS product and a prominent fragment ion at m/z 279 caused by the sequential loss of CH 3 NH 2 and a methoxyl radical. Fragments ions at m/z 264 and m/z 248.1 correspond to the consecutive loss of two methyl groups. According to these data, the identity of peak 21 is assigned to isocorydine [36]. For the compound corresponding to peak 27 in positive ionization mode, the molecular formula of ion at m/z 327.14735 was predicted as C 19 H 21  The observed fragments ions at m/z 296, 280 and 265 are coherent with the consecutive loss of three methyl groups. These data and the elution time suggest that this peak is N-methyl-laurotetanine (rogersine) [29,32]. Proposed fragmentation of aporphine alkaloids of P. boldus can be observed in the scheme presented in Figure S5.

Extractability of Boldine from Peumus boldus Leaves with Diverse NADES
Once identified the characteristic alkaloids in Boldo leaves, their extraction with seven selected NADES was evaluated (Table 3) [25][26][27][28][37][38][39][40][41][42][43]. The efficiency of these NADES was compared with methanol and water. If the literature related to Boldo extraction is carefully reviewed, it can be seen that both methanol and ethanol are conventionally used solvents and therefore we selected methanol as the control solvent (Table 4). For example, Rogalisnki and coworkers [44], evaluated the performance of boldine extraction using hot pressurized water and supercritical CO 2 compared with methanol extraction. Also in the work of Fuentes-Barros [7], the extraction of P. boldus alkaloids for HPLC analysis was performed with methanol. As in shown in Figure 3b, in the chromatogram obtained with NADES6, boldine appeared at around t R = 18 min. For the quantitative analysis of boldine in the different NADES extracts, we used a UV signal (304 nm) from HPLC-DAD-IT/MS (Figure 3a,b).     Figure 6 presents the quantitative analysis of boldine, where it is clear that methanol is twotimes more efficient than NADES1, NADES2, NADES3, NADES5 and NADES7 (0.1533 mg, 0.1607 mg, 0.1291 mg, 0.1473 mg and 0.1650 mg g −1 dry plant). NADES1 and NADES2 are alcohol-based solvents with polarity quite similar to ethanol [26][27][28]. Interestingly, we found two NADES that enable better extraction of boldine from Boldo leaves. These solvents were NADES4 (choline chloridelevulinic acid, 1:1) and NADES6 (proline-oxalic acid, 1:1) with 0.4270 mg and 2.3615 mg of boldine per gram of plant, respectively. From these results, it is remarkable that boldine extraction with NADES6 is eight-times more efficient than methanol. Moreover, boldine extraction yields varied   Figure 6 presents the quantitative analysis of boldine, where it is clear that methanol is twotimes more efficient than NADES1, NADES2, NADES3, NADES5 and NADES7 (0.1533 mg, 0.1607 mg, 0.1291 mg, 0.1473 mg and 0.1650 mg g −1 dry plant). NADES1 and NADES2 are alcohol-based solvents with polarity quite similar to ethanol [26][27][28]. Interestingly, we found two NADES that enable better extraction of boldine from Boldo leaves. These solvents were NADES4 (choline chloridelevulinic acid, 1:1) and NADES6 (proline-oxalic acid, 1:1) with 0.4270 mg and 2.3615 mg of boldine per gram of plant, respectively. From these results, it is remarkable that boldine extraction with NADES6 is eight-times more efficient than methanol. Moreover, boldine extraction yields varied greatly depending on the type of HBD used for NADES preparation. For instance, the alcohol-based  Figure 6 presents the quantitative analysis of boldine, where it is clear that methanol is two-times more efficient than NADES1, NADES2, NADES3, NADES5 and NADES7 (0.1533 mg, 0.1607 mg, 0.1291 mg, 0.1473 mg and 0.1650 mg g −1 dry plant). NADES1 and NADES2 are alcohol-based solvents with polarity quite similar to ethanol [26][27][28]. Interestingly, we found two NADES that enable better extraction of boldine from Boldo leaves. These solvents were NADES4 (choline chloride-levulinic acid, 1:1) and NADES6 (proline-oxalic acid, 1:1) with 0.4270 mg and 2.3615 mg of boldine per gram of plant, respectively. From these results, it is remarkable that boldine extraction with NADES6 is eight-times more efficient than methanol. Moreover, boldine extraction yields varied greatly depending on the type of HBD used for NADES preparation. For instance, the alcohol-based NADES exhibited poor extraction capacity, reaching only a 54% alkaloids in comparison with methanol. On the other hand, the extraction efficiencies of carboxylic acid-based such as NADES4 and NADES6 were significantly higher than other NADESs, as well as methanol.
Plants 2020, 9, x FOR PEER REVIEW 10 of 17 NADES exhibited poor extraction capacity, reaching only a 54% alkaloids in comparison with methanol. On the other hand, the extraction efficiencies of carboxylic acid-based such as NADES4 and NADES6 were significantly higher than other NADESs, as well as methanol. Figure 6. Effect of different NADESs on the extraction of boldine from Peumus boldus leaves with using heat + stirring (H+S, blue bars) and ultrasonic assisted extraction (UAE, green bars). In all extraction performed with NADES 1-7, 20% water was added to reduce viscosity. White and black bars corresponds to extraction performed with water and methanol, respectively.
These results are in agreement with the results published by Duan and coworkers [37], for other types of alkaloids such as jatrorrhizine hydrochloride, palmatine hydrochloride and berberine hydrochloride from herb Berberidis Radix. The differences observed between NADESs and MeOH can be due to the lack of extractability of partially ionized compounds by MeOH, where electrostatic interactions could significantly contribute to their extraction [38]. Then, when this result is compared with other extraction methods [7,10,[45][46][47][48][49], it is observed that the extraction performance for boldine is still more efficient than the other extraction procedures based on conventional solvents (Table 4). This difference can be explained by the variations in the substrate, the extraction procedure or method of analysis. With respect to the extraction process, the difference may be due to the limited selectivity of the adopted method and the resulting contamination of the alkaloid fraction and -at least in part-to the improvements in the solubility of the alkaloids. The latter can be explained by the increase in the solvation of non-polar organic solvents for alkaloids (naturally present as salts in Boldo leaves) after pH adjustment of aqueous alcohol solutions in the pretreatment steps. For instance, in acidic conditions, boldine is protonated and its water solubility is significantly better than in the neutral solvent [26,37].

Extraction Yields of Total Polyphenols from Peumus boldus Leaves
In Figure 7, it is observed the results obtained for total content polyphenols (TPC) are expressed as gallic acid equivalents (GAE). The best yield for the extraction of total polyphenols was again obtained with the NADES6 (L-proline: oxalic acid). Interestingly, no significant difference was observed in the results of TPC between heating + stirring extraction (179.442 ± 3.79 mg g −1 GAE dw) and UAE extraction (172.659 ± 2.55 mg g −1 GAE dw). Moreover, the TPC in extracts obtained with control solvent (methanol) does not show significant differences with NADES6 (one-way ANOVA, p < 0.05), suggesting that NADES6 could be used to replace methanol. On the other hand, our results showed that H2O, NADESs 1-4 and 7 were less suitable for the extraction of polyphenols. Since high viscosity of NADES is one of the main drawbacks for its use, it is worth mentioning that in all the extractions performed in the present study the viscosity was reduced by adding a maximum of 20% of water (Table 3). It has been reported that this strategy does not alter the supramolecular NADES network; on the contrary, it has a dramatic effect by increasing the mass transfer and mobility of the molecules [41,42]. Therefore, the addition of water combined with temperature help to reduce the Figure 6. Effect of different NADESs on the extraction of boldine from Peumus boldus leaves with using heat + stirring (H+S, blue bars) and ultrasonic assisted extraction (UAE, green bars). In all extraction performed with NADES 1-7, 20% water was added to reduce viscosity. White and black bars corresponds to extraction performed with water and methanol, respectively.
These results are in agreement with the results published by Duan and coworkers [37], for other types of alkaloids such as jatrorrhizine hydrochloride, palmatine hydrochloride and berberine hydrochloride from herb Berberidis Radix. The differences observed between NADESs and MeOH can be due to the lack of extractability of partially ionized compounds by MeOH, where electrostatic interactions could significantly contribute to their extraction [38]. Then, when this result is compared with other extraction methods [7,10,[45][46][47][48][49], it is observed that the extraction performance for boldine is still more efficient than the other extraction procedures based on conventional solvents (Table 4). This difference can be explained by the variations in the substrate, the extraction procedure or method of analysis. With respect to the extraction process, the difference may be due to the limited selectivity of the adopted method and the resulting contamination of the alkaloid fraction and -at least in part-to the improvements in the solubility of the alkaloids. The latter can be explained by the increase in the solvation of non-polar organic solvents for alkaloids (naturally present as salts in Boldo leaves) after pH adjustment of aqueous alcohol solutions in the pretreatment steps. For instance, in acidic conditions, boldine is protonated and its water solubility is significantly better than in the neutral solvent [26,37].

Extraction Yields of Total Polyphenols from Peumus boldus Leaves
In Figure 7, it is observed the results obtained for total content polyphenols (TPC) are expressed as gallic acid equivalents (GAE). The best yield for the extraction of total polyphenols was again obtained with the NADES6 (L-proline: oxalic acid). Interestingly, no significant difference was observed in the results of TPC between heating + stirring extraction (179.442 ± 3.79 mg g −1 GAE dw) and UAE extraction (172.659 ± 2.55 mg g −1 GAE dw). Moreover, the TPC in extracts obtained with control solvent (methanol) does not show significant differences with NADES6 (one-way ANOVA, p < 0.05), suggesting that NADES6 could be used to replace methanol. On the other hand, our results showed that H 2 O, NADESs 1-4 and 7 were less suitable for the extraction of polyphenols. Since high viscosity of NADES is one of the main drawbacks for its use, it is worth mentioning that in all the extractions performed in the present study the viscosity was reduced by adding a maximum of 20% of water (Table 3). It has been reported that this strategy does not alter the supramolecular NADES network; on the contrary, it has a dramatic effect by increasing the mass transfer and mobility of the molecules [41,42]. Therefore, the addition of water combined with temperature help to reduce the strong intermolecular interactions ruled by the H bond network in eutectic solvents [43,50]. NADESs prepared with levulinic acid-choline chloride-or 1, 4 butanediol also give good results and could be used too in combination with more exhaustive methods such as ultrasound and microwave-assisted extraction [49][50][51]. In the case of TPC, NADES6 (H+S and UAE) does not shown significant differences when compared with control solvent (methanol) and could be used as a greener replacement for this solvent (Figure 7). It should be noted that viscosity and negligible volatility of NADES are two properties that constitute a disadvantage compared to traditional solvents (methanol or ethanol). This disadvantages has led to profusely search new eutectic solvents with low viscosity. On the other hand, the volatility of traditional solvents allows their distillation, but in turn it is an environmental problem since they can cause air pollution and cause damage to human health. While it is true, solvents such as water, ethanol or methanol are cheaper and widely used solvents, they have some additional drawbacks. In particular, although they can be as efficient as a NADES, these alcohols have the serious disadvantage of being flammable and in cases where large-scale extraction processes must be scaled, they are considered to be dangerous. Nevertheless, it is clear that there are several points of debate regarding the use of NADES which require attention in the future. For instance, it is necessary to know more about the toxicity and permanence of NADES residues in the environment (degradation) and how these residues could affect living organisms. Furthermore, the use of NADES as bio-compatible solvents require to know if they affect the bioactivity of the bioactive products. The removal or recovery of NADES are points frequently addressed in recent publications. However, several strategies have been proposed. For instance, Liu and coworkers [52] reported the application of counter-current separation (CCS) to recover secondary metabolites from NADES. They also propose use CCS to recycle NADES because it remain intact after CCS and can be extruded from the column of high speed counter current chromatography (HSCCC) apparatus. Liquid-Liquid extraction also has been recently proposed by Smink et al. [53]. Other methods such as ultrafiltration, precipitation, solid phase extraction and electro-dialysis have been reported for NADES [54][55][56][57][58].
NADESs prepared with levulinic acid-choline chloride-or 1, 4 butanediol also give good results and could be used too in combination with more exhaustive methods such as ultrasound and microwave-assisted extraction [49][50][51]. In the case of TPC, NADES6 (H+S and UAE) does not shown significant differences when compared with control solvent (methanol) and could be used as a greener replacement for this solvent (Figure 7). It should be noted that viscosity and negligible volatility of NADES are two properties that constitute a disadvantage compared to traditional solvents (methanol or ethanol). This disadvantages has led to profusely search new eutectic solvents with low viscosity. On the other hand, the volatility of traditional solvents allows their distillation, but in turn it is an environmental problem since they can cause air pollution and cause damage to human health. While it is true, solvents such as water, ethanol or methanol are cheaper and widely used solvents, they have some additional drawbacks. In particular, although they can be as efficient as a NADES, these alcohols have the serious disadvantage of being flammable and in cases where large-scale extraction processes must be scaled, they are considered to be dangerous. Nevertheless, it is clear that there are several points of debate regarding the use of NADES which require attention in the future. For instance, it is necessary to know more about the toxicity and permanence of NADES residues in the environment (degradation) and how these residues could affect living organisms. Furthermore, the use of NADES as bio-compatible solvents require to know if they affect the bioactivity of the bioactive products. The removal or recovery of NADES are points frequently addressed in recent publications. However, several strategies have been proposed. For instance, Liu and coworkers [52] reported the application of counter-current separation (CCS) to recover secondary metabolites from NADES. They also propose use CCS to recycle NADES because it remain intact after CCS and can be extruded from the column of high speed counter current chromatography (HSCCC) apparatus. Liquid-Liquid extraction also has been recently proposed by Smink et al. [53]. Other methods such as ultrafiltration, precipitation, solid phase extraction and electro-dialysis have been reported for NADES [54][55][56][57][58].

Preparation of NADES
The preparation of all NADES tested was based on previously reported procedures [24][25][26]37,38]. Briefly, choline chloride and l-proline (hydrogen bond donor-HBD) was mixed with lactic acid, 1,2-propanediol, glycerol, levulinic acid, citric acid or oxalic acid (hydrogen bond acceptors-HBA) at predetermined molar ratios. Mixtures were mildly heated under stirring, until a perfectly transparent liquid was formed. NADES were kept in sealed glass vials in the dark, at ambient temperature. The list of the NADES used in this study, along with details regarding their preparation and references, are presented in Table 3.

Plant Material and Extraction
The plant material (leaves of Peumus boldus) was collected at the University of Concepción in June 2017 and authenticated in the Herbarium of the Department of Botany at the University of Concepción, Chile. The Voucher specimen was kept under code CONC N • 187541. After collection, the plant samples were air dried for 14 days at room temperature in the dark, and then ground to a fine powder using a Blender (Waring, McConnellsburg, PA, USA). This material was used for all further procedures. All NADES were used as 80% (v/v) aqueous solutions in order to reduce viscosity. Extractions were carried out according to a previously described methodology [39,40]. In brief, plant material (0.1 g) was placed in a 50 mL conical centrifuge tubes and 10 mL of NADES solvent was added. With the aim to compare the extraction yield of alkaloids from P. boldus, methanol and water were used as control solvents under the same conditions set up for all NADES. The suspension was vortexed vigorously for 30 s until a homogeneous thick mixture was obtained. Then, homogeneous samples were extracted through Heating and stirring extraction in a Syncore Polyvap R24 (Büchi, Flawil, Switzerland), under the following conditions: 60 • C for 50 min at 340 rpm. In addition, Ultrasound-assisted extraction (UAE) was performed using an Ultrasonic homogenizer bar JY92-IIDN (XinZhi Institute, NingBo, China) at room temperature for 20 min with a sonication power of 140 W and frequency of 37 kHz. After extraction, samples were clarified by centrifugation (Eppendorf 5804 R, Long Island, New York, NY, USA) at 8000 rpm for 10 min. The supernatants were filtered through a Millipore 0.45 µm cellulose acetate membrane filter and two-fold diluted with mobile phase prior to HPLC analysis. The extraction procedure describes above was performed in triplicates. For total polyphenol content determination, samples were 10-fold diluted with distilled water.

Total Polyphenol Content
The extracts were re-dissolved in water, and total phenolic content (TPC) was determined by using the Folin-Ciocalteu method with slight modifications [59]. In brief, 20 µL of properly diluted samples were mixed with 780 µL of distilled water and 50 µL of Folin-Ciocalteu reagent in 1.5 mL conical tubes. After 1 min, 150 µL of 7.5% sodium carbonate solution were added and mixed. Samples were leave in the dark at room temperature for 1 h. Aliquots of 200 µL were charged in 96 well microplates and the absorbance was measured at λ 750 nm using an EPOC microplate reader (Biotek). The analysis were performed in triplicate and normalized against negative controls (distilled water or diluted NADES) according to the Table 1. TPC was expressed as milligrams of gallic acid equivalents per gram of extract (mg g −1 GAE of extract) based on a standard curve of gallic acid (50-400 mg L −1 ; y = 0.013x + 0.0073; R 2 = 0.9991).

Qualitative and Quantitative HPLC-PDA-IT-MS/MS Analysis
The samples of P. boldus were analyzed by HPLC-PDA-IT-MS/MS in an Agilent 1100 Series system (Agilent, Waldbronn, Germany) equipped with an automatic degasser, a quaternary pump, an auto-sampler and a photodiode array detector (G1315B) and LC/MSD Trap VL (G2445C VL) ESI-MS n system, and it was coupled to an Agilent Chem Station (version B.01.03) data-processing station. The stationary phase employed was a Zorbax Eclipse XDB-C18 Narrow-Bore (150 mm × 2.1 mm; 3.5 µm particle size) column, while the mobile phase consisted of solvent A (water/formic acid/acetonitrile, 87:10:3, v/v/v) and solvent B (acetonitrile/water/formic acid, 50:40:10, v/v/v). The elution profile was (time, % of solvent B): 0 min, 3%; 10 min, 15%; 35 min, 40% B; 39-41 min, 100% B, and 47 min, 3% B [60]. The flow rate was 0.190 mL min −1 and the column temperature was set at 40 • C while the injection volume was 20 µL. The mass spectrometer was run in the positive ion mode with the following parameters: the capillary voltage was set at 3500 V, drying gas flow N 2 , 8 mL min −1 ; drying temperature, 325 • C; nebulizer, 50 psi; and scan range, 100-1200 m/z. The collision energy (CE) increased linearly in the range of 30-45 eV depending on the m/z range (100-1200). The range of detection wavelength were 200-600 nm. However, for alkaloids detection and boldine quantification, 304 nm was selected. Boldo extracts (10 mg) were dissolved in methanol (1 mL), diluted with mobile phase and filtered with a 0.45 µm syringe filter of polytetrafluoroethylene (13 mm) (Millex). The results were expressed as milligrams per gram of extract (mg g −1 extract). The linearity of the method was assessed from the correlation coefficients (R 2 ) of three set of calibration curves obtained for seven levels of boldine concentrations ranging from 0.0469 mg L −1 to 15.00 mg L −1 (y = 352.33x − 75.904; R 2 = 0.9971). Each point was injected three times. Limit of detection (LOD) and Limit of quantification (LOQ) were estimated at signal to noise (S/N) ratios of 3:1 and 10:1, respectively [61]. With this procedure, LOD and LOQ values were 0.003 mg L −1 and LOQ = 0.023 mg L −1 , respectively.

Q-ToF High-Resolution Mass Spectrometry Measurements
The analytical system used consisted of a 1260 Infinity high performance liquid chromatography system coupled to a diode array detector (DAD) and a 6545 quadrupole-time of flight (Q-TOF) mass spectrometer detector (Agilent, Waldbronn, Germany). The control software was Mass Hunter Workstation version B.06.11 (Agilent, Santa Clara, CA, USA). The Q-TOF used a Dual Jet Stream Electrospray Ionization (Dual AJS-ESI) source operated in the positive ionization mode and the following parameters were set: capillary voltage, 3500 V; fragmentor, 200; gas temperature, 350 • C; drying gas, 8 L min −1 ; nebulizer, 40 psig; sheath gas temperature, 400 • C; sheath gas flow, 10 L min −1 ; acquisition range, 100-1000 m/z; and CID, with a linear range of 30-45. Samples were analyzed after injection (10 µL) on a Zorbax Eclipse Plus C18 Rapid Resolution HD column (2.1 mm × 50 mm, 1.8 µm) protected with a 5 mm guard column of the same material thermostated at 40 • C and his flow rate was 0.3 mL min −1 . The solvent system was 1mM of ammonium formate + 0.1% formic acid in water (solvent A) and 1 mM of ammonium formate + 0.1% formic acid in methanol (solvent B). The elution gradient was (time, % of solvent B): 0 min, 7%; 10 min, 20%; 40 min, 75%; 46.5 min, 95%; 56 min, 7%; and a post time of 8 min. Compounds were identified using the algorithm "Find by Formula" that evaluated the mass accuracy together with the isotopic relative abundance and isotopic separation. All the compounds were identified by the QTOF-MS and the MS/MS spectra acquired with the IT-MS and their absorption spectra in UV-visible region, as well as considering the data provided by literature [6,7,[29][30][31][32][33][34][35][36].

Statistical Analysis
Statistical comparison was performed using GraphPad Prism 5. (GraphPad Software, San Diego, CA, USA). Variables were expressed as mean and standard deviation (SD). The comparisons between the means in each assay were performed by one-way analysis of variance (ANOVA) at a 95% confidence level. Tukey's multiple comparison post-hoc test was applied to determine the differences amongst extraction yields. Data points plotted in Figures 5 and 6 represent the means of at least three independent experiments, each conducted in triplicate.

Conclusions
In this work, advanced analytical methods have been used to carry out a thorough characterization of a P. boldus extracts. In this report, the identity of main P. boldus alkaloids and phenolics compounds was confirmed by HPLC coupled to DAD-IT-MS/MS and Q-ToF HRMS. Finally, from our results, it can be concluded that NADESs are a potential green alternative to conventionally used organic solvents as extraction media to improve the extraction of alkaloids and phenolic compounds. Among the NADESs tested in our study, proline-oxalic acid (1:1) with 20% water was the most promising solvent, attaining higher extraction yields of boldine and TPC from P. boldus leaves. Overall, an adequate fine-tuning of HBD/HBA components in a NADES is a powerful strategy that allows us to perform selective extractions of certain molecules with pharmacological interest. This latter, along with its superior extraction efficiency and reduced environmental and lower economic impacts, make NADESs an interesting alternative to organic solvents for the extraction of Boldo bioactive metabolites.