A Descriptive Chemical Composition of Concentrated Bud Macerates through an Optimized SPE-HPLC-UV-MS2 Method—Application to Alnus glutinosa, Ribes nigrum, Rosa canina, Rosmarinus officinalis and Tilia tomentosa

Concentrated bud macerates (CBMs) are obtained from meristematic tissues such as buds and young shoots by maceration in a solvent composed of glycerin, water and ethanol (1/1/1/, v/v). Their traditional utilization in gemmotherapy has gained interest in the past years, and the knowledge of their chemical characterization can provide commercial arguments, particularly to secure their quality control. Therefore, an optimized method for phytochemical analysis including glycerol removal by a preliminary solid phase extraction (SPE) followed by compound identification using high performance liquid chromatography coupled with ultra-violet and tandem mass detectors (HPLC-UV-MS2) was developed. This method was applied on 5 CBMs obtained from Alnus glutinosa, Ribes nigrum, Rosmarinus officinalis, Rosa canina and Tilia tomentosa in order to determinate their chemical composition. Their antioxidant effects were also investigated by radical scavenging activity assays (DPPH and ORAC). Glycerol removal improved the resolution of HPLC chemical profiles and allowed us to perform TLC antioxidant screening. Our approach permitted the identification of 57 compounds distributed in eight major classes, three of them being common to all macerates including nucleosides, phenolic acids and glycosylated flavonoids. Quantification of the later class as a rutin equivalent (RE) showed a great disparity between Rosa canina macerate (809 mg RE/L), and the other ones (from 175 to 470 mg RE/L). DPPH and ORAC assays confirmed the great activity of Rosa canina (4857 and 6479 μmol TE/g of dry matter, respectively). Finally, phytochemical and antioxidant analysis of CBMs strengthened their phytomedicinal interest in the gemmotherapy field.


Introduction
Bud macerates are obtained from meristematic tissues, such as buds and young shoots, as a solution of glycerin, water and alcohol. These macerates are used in gemmotherapy, which is a popular phytomedicine in European countries [1]. Their utilization in gemmotherapy is supported by a specific and diversified chemical composition, mainly due to the unusual part of plant and extraction solvent. Bud macerates are also described to possess a wide range of biological activities [2][3][4]. However, only a few analytical studies describe their composition, and they were associated with a limited number of different raw materials (Table 1) including Alnus glutinosa (L.) Gaertn, Carpinus betulus L., Castanea spp.: C. sativa Mill., C. crenata Siebold and Zucc. and C. sativa x crenata, Cornus mas L., Table 1. Bud macerates related in the literature.
Turrini et al. [6] recently concluded that the polyphenol content of bud derivatives is strongly influenced by manufacturing processes whose parameters are often not strictly defined (for example, extraction solvent ratios, raw material/solvent mixture percentage and extraction time) and thus affect the final composition of bud macerates.
The present study focused on the phytochemical analysis of five concentrated bud macerates (CBMs) obtained by a cold maceration of fresh plant (1/20 plant, dry weight/solvent) in ethanol/glycerol/water (1/1/1) for 21 days. They were obtained from Alnus glutinosa hydroxycinnamic acids, benzoic acids, catechins, tannins and ellagic acid derivatives was highlighted [5,12,16,19]. Bud macerate obtained from Ro is here studied for the first time.
As usually performed, phytochemical profiling was achieved using HPLC-DAD-ESI-MS in comparison with standards and literature data. Moreover, a quick and preliminary solid phase extraction (SPE) was developed and applied to samples to eliminate a great part of glycerol, which is responsible of low-resolution chemical profiles. Flavonoids were part of the main compounds identified in all extracts; therefore, they were quantified as rutin equivalents by HPLC-DAD. As flavonoids are generally associated with strong antioxidant activities, an HPTLC analysis using DPPH as a revelation reagent was performed to confirm the presence of antioxidant compounds. Finally, the antioxidant capacity of the five CBMs was evaluated using DPPH and ORAC assays.

Application of a Solid Phase Extraction Procedure
The presence of glycerin in CBMs affects the resolution of the chemical profiles and avoid the calculation of extraction yields. Therefore, a procedure was developed and applied to all CBMs and the effective elimination of glycerol by SPE was then monitored by HPLC-ELSD. Figure 1 shows, as an example, the Rc extract profile before and after glycerol elimination. Based on AUC values, a 9-fold decrease was observed. In the TR 2.5-5.0 min range and when compared with total peak area values, signals associated with glycerol (as well as sugars, organic acids and vitamins) are respectively reduced to 7.8% (Rc), 15.5% (Ag), 30.4% (Rn), 33.8% (Ro) and 58.2% (Tt), leading to high-quality chromatograms. The use of this method could be particularly useful for the quality control of bud macerates by HPLC fingerprinting. macerates (CBMs) obtained by a cold maceration of fresh plant (1/20 plant, dry weight/sol-vent) in ethanol/glycerol/water (1/1/1) for 21 days. They were obtained from Alnus glutinosa (L.) Gaertn (Ag), Ribes nigrum L. (Rn), Rosa canina L. (Rc), Rosmarinus officinalis L. (Ro) and Tilia tomentosa M. (Tt) and are sold in Europe as food supplements. Bud macerates of Ag, Rc, Rn and Tt were already studied for their chemical composition. The presence of flavonols, hydroxycinnamic acids, benzoic acids, catechins, tannins and ellagic acid derivatives was highlighted [5,12,16,19]. Bud macerate obtained from Ro is here studied for the first time.
As usually performed, phytochemical profiling was achieved using HPLC-DAD-ESI-MS in comparison with standards and literature data. Moreover, a quick and preliminary solid phase extraction (SPE) was developed and applied to samples to eliminate a great part of glycerol, which is responsible of low-resolution chemical profiles. Flavonoids were part of the main compounds identified in all extracts; therefore, they were quantified as rutin equivalents by HPLC-DAD. As flavonoids are generally associated with strong antioxidant activities, an HPTLC analysis using DPPH as a revelation reagent was performed to confirm the presence of antioxidant compounds. Finally, the antioxidant capacity of the five CBMs was evaluated using DPPH and ORAC assays.

Application of a Solid Phase Extraction Procedure
The presence of glycerin in CBMs affects the resolution of the chemical profiles and avoid the calculation of extraction yields. Therefore, a procedure was developed and applied to all CBMs and the effective elimination of glycerol by SPE was then monitored by HPLC-ELSD. Figure 1 shows, as an example, the Rc extract profile before and after glycerol elimination. Based on AUC values, a 9-fold decrease was observed. In the TR 2.5-5.0 min range and when compared with total peak area values, signals associated with glycerol (as well as sugars, organic acids and vitamins) are respectively reduced to 7.8% (Rc), 15.5% (Ag), 30.4% (Rn), 33.8% (Ro) and 58.2% (Tt), leading to high-quality chromatograms. The use of this method could be particularly useful for the quality control of bud macerates by HPLC fingerprinting. Extraction yields (m/v %) were then calculated after glycerol elimination. They appeared to range between 0.3 and 0.8% (0.26 ± 0.02% for Tt, 0.48 ± 0.02% for Ro, 0.49 ± 0.02% for Rn, 0.66 ± 0.12% for Ag and 0.76 ± 0.07% for Rc). Extraction yields (m/v %) were then calculated after glycerol elimination. They appeared to range between 0.3 and 0.8% (0.26 ± 0.02% for Tt, 0.48 ± 0.02% for Ro, 0.49 ± 0.02% for Rn, 0.66 ± 0.12% for Ag and 0.76 ± 0.07% for Rc).

Composition of CBM Obtained from Ribes nigrum (Rn)
Ribes nigrum CBM was characterized by the presence of (E)-p-coumaric acid 15, gly- Coumaric acid, rutin and isoquercitrin were already described in Rn glycerin macerates in the literature [11,13,18]. The same authors evidenced the presence of benzoic acids, catechins, other cinnamic acids, other flavonols, terpenic compounds, vitamins and organic acids [9,[13][14][15]18]. In 2015, Ieri et al. [16] published the phenolic composition of "bud extracts" of Rn which is in accordance with our work. No author noticed the presence of phloridzin in bud macerate, but it was described as a phenolic compound from blackcurrant fruit [29]. These results are in accordance with Ieri et al. [16] who described glycosides of quercetin and kaempferol, gallic acid derivatives and caffeoylquinic acids together with ellagic acids derivatives, these last compounds being absent of the extracts analyzed here. Kaempferol galloyl hexoside derivative 44 and quercetin 48 are described for the first time in a Rc bud macerate. These compounds were identified respectively by Riffault et al. [26] in Rosa hybrida and Ozcan et al. [40] in Rosa canina alcoholic extracts.

Composition of CBM Obtained from Rosmarinus officinalis (Ro)
Rosmarinus officinalis CBM analysis led to the identification of a great diversity of  [41] work on Rosmarinus officinalis young shoots extract. Some polyphenolic compounds described in other types of extracts did not appear in our preparations such as carnosic acid [28,30] and diosmin [41].

Highlight on Flavonoid Content of CBMs
Since flavonoids (glycosylated and aglycones) were identified as major constituents of the glycerin macerates, they were quantified at 355 nm as rutin equivalents (RE mg/L) for comparison purpose (Table 3). This study revealed Rc as the most concentrated macerate (809 RE mg/L) whereas other extracts varied from 175 to 470 RE mg/L at 355 nm.
The lack of homogeneity in flavonoid quantification protocols and results expression makes comparisons challenging. Therefore, the flavonoid contents of the five studied plants were determined as mg of rutin equivalent (RE) per L of glycerin macerate, and then calculated per g of dry weight (DW) and per g or 100 g of fresh weight (FW) and exposed in Table 3 to compare with literature data.
For Ag, the flavonoid content was evaluated as 470 ± 6 mg RE/L equivalent to 9 mg RE/g of dry weight (DW). This could be compared to Dahija et al. [42] results, exploring the flavonoid contents of methanolic extracts of Ag leaves and showing 11.8 mg of rutin equivalent/g of DW (415 nm).
Flavonoid content of CBM obtained from Rn (175 ± 2 RE mg/L) is in the range of the description of Rn bud macerates analyzed by Ieri et al. [16]: 67-304 mg/L. Expressed in term of fresh weight, a flavonoid content of 1 RE mg/g of FW is below the one described by Liu et al. [24] who evaluated 3-4 mg/g but a different extraction solvent was used (water/acetone). Expressed in mg/g of fresh weight we found a value (150 mg/100 g) close to the one described by Donno et al. [14]: 126 mg/100 g and Turrini et al. [10]: 97 mg/100 g (calculated as the sum of the quercetin, quercitrin and rutin equivalents).
Flavonoid quantification for Rc showed a great amount of 809 ± 13 mg RE/L which was higher than that found by Ieri et al. [16] who calculated the amount of flavonol glycosides and evaluated it as 238-589 mg/L, depending on the farm (month and year of harvest).
Ro showed the third higher flavonoid content after Rc and Ag with 332 ± 2 mg RE/L corresponding to 7 mg/g of dry buds (DW). These results could be compared to the flavones concentration determined to be 5.2 mg/g of DW leaves in methanolic extracts [43]. Besides glycosylated and aglycones flavones, we found glycosylated and aglycone flavonols, plus glycosylated flavanone. A higher total flavonoid content of 24.6 mg RE/g of dry leaves (DW) extract (hexane then ethyl acetate) was reported by Kontogianni et al. [30]. These organic solvents could probably permit the enhancement of the extraction yield of Ro.
Flavonoid content for Tt corresponded to 219 ± 2 mg RE/L of glycerin macerate. This value was in accordance with those determined on flavonol glycosides by Ieri et al. [16] on 5 different buds of Tt glycerin macerates: 176-480 mg/L. It could be noted that in our work, besides glycosylated flavonols, we also detected glycosylated flavones. Another work of Turrini et al. [6] evaluated the flavonol content of Tt glycerin macerates at 52-91 mg/100 g of fresh weight depending on the process used. Our results showed a similar value (92 mg/100 g of fresh weight) even if no quercetin was detected in our macerate.

Antioxidant Activity
An HPTLC study was undertaken using Neu (flavonoids) and DPPH (antioxidant activity) as revelation reagents. CBMs were developed before (Figure 3a,b) and after (Figure 3c,d) SPE preparation. It was observed that glycerol severely hampered the TLC migration and Rc exhibited, as expected, the most interesting antioxidant profile (Figure 3d). Besides glycosylated and aglycones flavones, we found glycosylated and aglycone flavonols, plus glycosylated flavanone. A higher total flavonoid content of 24.6 mg RE/g of dry leaves (DW) extract (hexane then ethyl acetate) was reported by Kontogianni et al. [30]. These organic solvents could probably permit the enhancement of the extraction yield of Ro.
Flavonoid content for Tt corresponded to 219 ± 2 mg RE/L of glycerin macerate. This value was in accordance with those determined on flavonol glycosides by Ieri et al. [16] on 5 different buds of Tt glycerin macerates: 176-480 mg/L. It could be noted that in our work, besides glycosylated flavonols, we also detected glycosylated flavones. Another work of Turrini et al. [6] evaluated the flavonol content of Tt glycerin macerates at 52-91 mg/100 g of fresh weight depending on the process used. Our results showed a similar value (92 mg/100 g of fresh weight) even if no quercetin was detected in our macerate.

Antioxidant Activity
An HPTLC study was undertaken using Neu (flavonoids) and DPPH (antioxidant activity) as revelation reagents. CBMs were developed before (Figure 3a,b) and after (Figure 3c,d) SPE preparation. It was observed that glycerol severely hampered the TLC migration and Rc exhibited, as expected, the most interesting antioxidant profile (Figure 3d). DPPH and ORAC experiments were also performed on the five glycerin macerates. These assays gave different results, varying, respectively, from <200 to 4857 μmol TE/g and from 2487 to 6479 μmol TE/g (see Table 4), but they both confirmed the highest antioxidant activity of Rc (4857 and 6479 μmol TE/g of dry matter, respectively, using DPPH or ORAC assays). DPPH and ORAC experiments were also performed on the five glycerin macerates. These assays gave different results, varying, respectively, from <200 to 4857 µmol TE/g and from 2487 to 6479 µmol TE/g (see Table 4), but they both confirmed the highest antioxidant activity of Rc (4857 and 6479 µmol TE/g of dry matter, respectively, using DPPH or ORAC assays). The DPPH antioxidant activity seemed to be positively correlated with the flavonoid content. The same trend is not observed for the ORAC assay. The major difference in the chemical composition of Rc was the presence of galloyl quinic derivatives and galloyl flavonol glycosides. Several works showed strong antioxidant activities for gallic acid and galloyl derivatives. Furthermore, the presence of one or more galloyl moieties was correlated with the antioxidant capacity of flavonol glycosides and galloyl quinic derivatives [44][45][46].
Only a few works describing antioxidant activities of Rn and Ag bud macerates, [5,12,18] were related in the literature, but the results are not comparable. On the other hand, the antioxidant activities of Ro and Rc were studied, but they concerned extracts obtained using different solvents of extraction [30,47,48].

Solid Phase Extraction
A solid phase extraction (SPE) procedure was developed and applied to CBMs to remove glycerol. Ethanol was firstly removed from the macerate by evaporation under vacuum. SPE was carried out on a C18 column (500 mg/2.8 mL) (Thermo Scientific, Cardiff Valley Road, TN, USA), activated by methanol and balanced with water. Samples were then adsorbed on the column (2 mL of CBM corresponding to 100 mg of dried plant) and the glycerol was eluted with water. Compounds adsorbed on the C18-column were recovered by elution with methanol. This solution was then evaporated under vacuum to obtain the final dry extract. These extracts were used to calculate extraction yields, expressed in percentage of the CBM (5% of dried plant: 5 g in 100 mL of solvent (water/ethanol/glycerol (1/1/1, v/v/v)) (n = 3).

HPLC-UV-ESI-MS 2
Analytical HPLC-UV-ESI-MS 2 was run on a 2695 Waters (Guyancourt, France) coupled with a diode array detector 2996 Waters. Column, mobile phases and gradient were the same as previously described for HPLC-DAD-ELSD. Chromatograms were acquired at 254 nm. The mass analyses were performed on a Brucker (Bremen, Germany) ESI/APCI Ion Trap Esquire 3000+ in both positive and negative modes, with the conditions as follows: collision gas, He; collision energy amplitude, 1.3 V; nebulizer and drying gas, N2, 7 L/min; pressure of nebulizer gas, 30 psi; dry temperature, 340 • C; flow rate, 1.0 mL/min; solvent split ratio 1:9; scan range, m/z 100-1200. Extracts obtained after SPE were prepared in methanol, at a concentration of 10 mg/mL.

Flavonoid Content
The quantification of flavonoids was performed by HPLC-DAD using 7-point regression curves in triplicate. The column, mobile phases and gradient were the same as previously described for HPLC-DAD-ELSD. Chromatograms were acquired at 355 nm.
Statistical analysis by R software validated a linear model (r 2 = 0.9999) with a confidence interval of 95% (Line equation y = 35,662,581x with a R 2 = 0.9999). CBMs were diluted in methanol (1/1 v/v), centrifugated, then filtered (0.45 µm, PTFE membrane). Total flavonoid content (TFC) was expressed in milligrams of rutin equivalent (RE) per liter (mg RE/L) of CBM. Total flavonoid amounts are expressed as the mean of three samples ± SD (n = 3). Considering the extraction yield (m/v %), the TFC could be calculated in mg RE/g of dry weight (DW). Considering the humidity level (%), the TFC could be calculated in mg RE/g or/100 g of fresh weight (FW).

Scavenging Activity of diphenyl-picrylhydrazyl (DPPH) Radicals
The diphenylpicrylhydrazyl (DPPH) radical scavenging evaluations of the CBMs were performed as previously described [50]. In its radical form, DPPH• has an absorption band at 517 nm, which disappears upon reduction by an antiradical compound. Briefly, the tested samples and standards were diluted in absolute EtOH at 0.02 mg/mL from stock solutions at 1 mg/mL in DMSO (depending on the yields). Aliquots (100 µL) of these diluted solutions were placed in 96-well plates in triplicate. A total of 25 µL of freshly prepared DPPH solution (1 mM) were added to 75 µL of absolute EtOH using the microplate reader's injector (Infinite 200, Tecan, France) to obtain a final volume of 200 µL per well. After 30 min in the dark and at ambient temperature, the absorbance was determined at 517 nm. EtOH was used as a blank, and 10, 25, 50, and 75 µM solutions of trolox (hydrophilic α-tocopherol analog) were used for the calibration curve. Samples of chlorogenic acid ethanolic solution and rosemary ethanolic extract (both at 0.02 mg/mL) were used as the positive control standard. Results were expressed as trolox equivalents (micromoles of TE per gram of dry matter).

Measurement of Oxygen Radical Absorbance Capacity (ORAC)
ORAC assays were carried out on CBMs according to the method described by Huang et al. [51] with some modifications. This assay measures the ability of antioxidant compounds to inhibit the decline in fluorescein (FL) fluorescence that is induced by a peroxyl radical generator, 2,2 -azobis (2-methylpropionamidine) dihydrochloride (AAPH). The assay was performed in a 96-well plate. The reaction mixture contained 100 µL of 75 mM phosphate buffer (pH 7.4), 100 µL of freshly prepared fluorescein (FL) solution (0.1 µM in phosphate buffer), 50 µL of freshly prepared 2,2 -azobis(2-methylpropionamidine) dihydrochloride (AAPH) solution (51.6 mg/mL in phosphate buffer), and 20 µL of sample per well. CBMs were analysed in triplicate and diluted in phosphate buffer at different concentrations (25, 12.5, 6.25 and 3.12 µg/mL) from stock solutions at 1 mg/mL in DMSO (depending on the yields). The FL, phosphate buffer, and samples were preincubated at 37 • C for 10 min. The reaction was started by the addition of AAPH using the microplate reader's injector (Infinite ® 200, Tecan, France). Fluorescence was then measured and recorded for 40 min (λexc 485 nm, λem 520 nm). The 75 mM phosphate buffer was used as a blank, and 12.5, 25, 50, and 75 µM solutions of Trolox were used as calibration solutions. A chlorogenic acid solution (8.8 µM) and a rosemary ethanolic extract (12.5 µg/mL) in phosphate buffer were used as positive control standard. The final ORAC values were calculated using a regression equation between the trolox concentration and the net area under the FL decay curve and were expressed as micromole of trolox equivalents per gram of dry matter. Areas under curves were calculated using Magellan data analysis software (Tecan, France).

Conclusions
In this research, an optimized method was used for the phytochemical analysis of concentrated bud macerates. A preliminary step of solid phase extraction was applied before recording HPLC-DAD-ELSD profiles, HPLC-UV-ESI-MS 2 analysis and HPTLC revelation, allowing high resolution fingerprints. The phytochemical analysis of Ag, Rn, Rc, Ro and Tt CBMs led to the identification of 57 compounds distributed in 8 chemical classes: flavonoids, nucleosides, phenolic acids, gallotannins and galloyl flavonol glycosides, glycosylated dihydrochalcones, lignans, quinones and abietane type diterpenes Although phenolic acids and flavonoids are common in all macerates, some compounds are defined particularly in a plant such as a dihydrochalcone 19 in Rn; gallotannins (6, 8-10, 14) and galloyl flavonol glycosides (34 and 44) in Rc; and an abietane-type diterpene 57, a lignan 13 and a quinone 55 in Ro.
Moreover, the phytochemical analysis of CBMs of Ag, Rc, Rn and Tt was in agreement with previously reported data and the chemical composition of Ro CBMs was investigated for the first time. This study highlighted a great chemical diversity, which is in accordance with the traditional description of bud macerates. Strong antioxidant activities, especially for Rc CBM, also support their use in gemmotherapy.