UHPLC-HRMS Analysis of Fagus sylvatica (Fagaceae) Leaves: A Renewable Source of Antioxidant Polyphenols

European beech (Fagus sylvatica L.) is a deciduous tree, widely distributed in Europe and largely appreciated for its wood and nutritive nuts. Beech leaf also enjoys food use as salad, but an understanding of its nutraceutical value is still far from being achieved. Indeed, and also taking into account beech leaf as a consistent biomass residue available beechwood production and use, it needs to be explored as a valuable renewable specialized source of bioactive molecules. In this context, an untargeted ultra-high-performance liquid chromatography hyphenated with high resolution mass spectrometry (UHPLC-HRMS) approach was favorably applied to a beech leaf alcoholic extract, which also was evaluated for its antiradical capability (by means of assays based on 2,2-diphenyl-1-picrylhydrazyl (DPPH) and [2,2’-azinobis-(3-ethylbenzothiazolin-6-sulfonic acid)] (ABTS) radical cation) and its ferric ion reducing power. Redox mitochondrial activity towards Caco-2 cells paved the way to explore the extract’s capability to inhibit intracellular Reactive Oxygen Species (ROS) using 2’,7’dichlorofluorescin diacetate (DCFH-DA) assay. Hydroxycinnamoyl derivatives, mainly belonging to the chlorogenic acid class, and flavonoids were the main constituents. Uncommon flavanone C-glycosides were also found, together with a plentiful flavonol diversity. Cell-free and cell-based assays highlight its dose-dependent antioxidant efficacy, providing a foundation for further investigation of beech leaf constituents and its valorization and use as a reservoir of bioactive natural products with potential nutraceutical applications.


Introduction
Nowadays, the replacement of synthetic and artificial chemicals with natural products with less impact on human or animal health and environment, makes their recovery a major challenge. Indeed, great efforts have been devoted to the discovery and exploitation of renewable sources of valuable bioactive compounds, such as the so-called bioactive specialized natural products. These latter include a number of compound classes and sub-classes that, in the last years, have attracted a lot of attention for a possible application in various sectors (i.e., nutri-cosmeceutical, medical and pharmacological), due to their recognized benefits for human and animal health [1]. Thus, the research and valorization of new plant matrices, also those not directly used for these purposes, need to be explored as a virtuous source of these molecules. Innovative examples are agro-food wastes [2][3][4], or biorefining of forest biomass [5], from which compounds with traditional use in the prevention and/or treatment of different diseases could be favorably isolated [6]. Byproducts from forestry and wood processing industry are promising feedstocks for the A linear gradient was used, in which the percentage of solvent B increased as follows: 0-5 min, 5%→15% B; 5-10 min, 15% B; 10-12 min, 15%→17.5% B; 12-17 min, 17.5%→45%B; 17-18.50 min, 45% B; 18.51-20 min, column re-equilibration. The flow rate was set at 400 µL/min.
High-Resolution Mass Spectrometry (HR-MS) data were obtained by an AB SCIEX Triple TOF ® 4600 mass spectrometer (AB Sciex, Concord, ON, Canada), equipped with a DuoSpray TM ion source (AB Sciex, Concord, ON, Canada) operating in the negative ElectroSpray (ESI) mode. A full scan Time-Of-Flight (TOF) survey (accumulation time 100 ms, 100-1000 Da) and 8 information-dependent acquisition MS/MS scans (accumulation time 50 ms, 80-850 Da) were acquired, using the following parameters: curtain gas 35 psi, nebulizer and heated gases 60 psi, ion spray voltage 4500 V, ion source temperature 600 • C, declustering potential −70 V, collision energy −35 ± 5 V. The instrument was controlled by Analyst ® TF 1.7 software (AB Sciex, Concord, ON, Canada), whereas MS data were processed by PeakView ® software version 2.2 (AB Sciex, Concord, ON, Canada). The compounds were identified mainly through the study of their tandem mass spectrometry (TOF-MS/MS) fragmentation patterns, and the comparison with literature data whenever possible.
ABTS radical cation was generated as previously reported [18]. The ABTS •+ solution was diluted with Phosphate-buffered saline (PBS; pH 7.4) until an absorbance of 0.7 at 734 nm was read. The extract at different doses was directly dissolved in the ABTS •+ solution, and after 6 min the absorbance was measured by a Victor3 spectrophotometer (Perkin Elmer/Wallac, Waltham, MA, USA) in reference to a blank, in which the samples were replaced with solvent [18]. DPPH • scavenging capability was estimated as previously reported [18], and the absorption at 517 nm was measured on the Victor3 spectrophotometer (Perkin Elmer/Wallac, Waltham, MA, USA) in reference to a blank, in which the samples were replaced with the solvent.
Trolox (4, 8, 16, 32 µM) was used as positive standard, and Trolox Equivalent Antioxidant Capacity (TEAC) of beech-leaf extract was calculated, based on both ABTS and DPPH tests. For each antiradical test, three replicate measurements for three samples (n = 3) of the extract (in total, 3 × 3 measurements) were performed. All data were expressed as mean ± standard deviation (SD).

Fe(III) Reducing Power
Beech leaf alcoholic extract (at 200, 100, 50, 25, 12.5, 6.25, and 3.125 µg/mL final concentration levels) was investigated for its ability to reduce the Fe 3+ using ferricyanide FRAP assay, as previously reported [19]. The absorbance was measured at 700 nm. The increase in absorbance with reference to the blank was considered to value the reducing power. Trolox (4, 8, 16, 32 µM) was used as positive standard, and TEAC value of beech-leaf extract was calculated. The test was carried out performing three replicate measurements for three samples (n = 3) of the extract (in total, 3 × 3 measurements). All data were expressed as mean ± standard deviation (SD).

Determination of Total Phenols
Total phenol content was determined according to the Folin-Ciocalteau procedure [19]. Samples (0.25 mg and 0.125 mg) were mixed with 2.25 mL of Na 2 CO 3 (7.5% w/v) and 0.25 mL of Folin-Ciocalteu reagent. The tubes were mixed and allowed to stand for 3 h at room temperature. The absorbance was read at 765 nm using a Synergy spectrophotometer (Biotek, Winooski, VT, USA). The test was carried out performing three replicate measurements for three samples (n = 3) of the extract (in total, 3 × 3 measurements). Data were expressed as milligrams of gallic acid equivalents (GAEs) per g of extract (mean ± standard deviation). To this purpose, a gallic acid calibration curve (R 2 = 0.9716) was built up in the range 0.78-25 µg/mL (final concentration levels).

Determination of Total Flavonoids
The extract (0.5 mL) was dissolved in distilled water (5 mL), and NaNO 2 solution (5%, w/v; 0.3 mL) was added. After 5 min, AlCl 3 solution (10%, w/v; 0.6 mL) was poured into the flask, and after 6 min, NaOH solution (1.0 M; 2.0 mL) and distilled water (2.1 mL) were added. The absorbance was read at 510 nm against the blank (water), and flavonoid content is expressed as milligrams of quercetin equivalents per 100 g of fresh material [20]. To this purpose, a quercetin calibration curve (R 2 = 0.9979) was built up in the range 0.78-100 µg/mL (final concentration levels). The test was carried out performing three replicate measurements for three samples (n = 3) of the extract (in total, 3 × 3 measurements). All data were expressed as mean ± standard deviation (SD).

Cell Culture, Cytotoxicity and Intracellular ROS Assessment
Human epithelial cell line Caco-2 (ATCC ® HTB¬37™, American Type Culture, Manassas, VA, USA) was cultured in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 10% fetal bovine serum, 50.0 U/mL of penicillin and 100.0 µg/mL of streptomycin, at 37 • C in a humidified atmosphere containing 5% CO 2 . Cells were seeded in 96-multiwell plates at a density of 2.5 × 10 4 cells/well. After 24 h, cells were treated for 24 h with different doses of the beech leaf alcoholic extract (25, 50, 100 and 200 µg/mL) or pure quercetin standard (5, 10, and 50 µM). When incubation was completed, inhibition of mitochondrial redox activity was determined by the MTT cell viability test, which was based on the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) dye, as previously described [3,4]. In order to evaluate intracellular ROS inhibition [21], Caco-2 cells were seeded at a density of 2.5 × 10 4 well on a black 96-well microplate in 100 µL growth medium/well. Cells were cultured for 24 h at 37 • C in 5% CO 2 and observed under the inverted phase contrast microscope. After 24 h, the growth medium was removed, and cells were twice washed with PBS (100 µL). Then, the cells were co-exposed to the investigated extract (25, 50, 100 and 200.0 µg/mL; final concentration levels) or quercetin (10 µM) and 2 ,7 -dichlorofluorescin diacetate (DCFH-DA; 60 µM) for 60 min. The treatment medium was then removed, the cells were washed with PBS, and 2,2 -azobis(2-methyl propionamidine)dihydrochloride (AAPH, 500 µM; 100 µL) was added. The 96-well microplate was placed into a PerkinElmer Victor3 Multilabel Plate Reader at 37 • C. The fluorescence intensity was measured at 485 nm excitation and 535 nm emission wavelength every 20 min for 100 min. Two independent experiments were carried out performing in each six replicate measurements for three samples (n = 3) of the extract (in total, 6 × 3 measurements). Data were expressed as mean ± standard deviation (SD).

Chemical Composition of F. sylvatica Leaf Methanolic Extract
Nowadays, the growing consciousness of environmental sustainability promotes the recovery of waste from production chains, also as basis and foundation of a circular economy. Renewable forest materials are rich in nutrients and bioactive molecules, whose recovery opens up to the development of functional products with high added value [22], to be used in various production sectors, from food, to nutraceuticals, cosmetics, up to the creation of packaging. In this context, getting insight into the chemical composition of beech leaf, its diversity in polyphenols could represent its intrinsic value.
The analytical determination of the diversity in bioactive molecules, after suitable extraction of the matrix, represents the crucial step of the entire process. Actually, techniques in tandem mass spectrometry hyphenated with chromatographic separation methods allows the chemical composition to be finely unraveled through an untargeted approach that provides accuracy in phenolomic data and compounds identification [23].
In the present study UHPLC-ESI-HRMS and High-Resolution tandem mass spectrometry (HR-MS/MS) techniques were first applied to unravel the chemical composition of this undervalued plant source [24][25][26]. An untargeted metabolic approach was used. Sixty-nine compounds were tentatively identified (Figure 1), mainly belonging to phenol and polyphenol classes. These latter have been described also as constituents of other forest trees and their wastes. Indeed, recently, bulk samples of bark waste from Pinus contorta, Pinus sylvestris, and Quercus robur, were investigated in depth for their polyphenol content [27], and various biomass residues (shavings, edged cuts, and pruning wastes) from walnut were analyzed as sources of antioxidant compounds by means of a green extraction process [28]. In this context, Pycnogenol ® (PYC) and Flavangenol ® are good examples of commercially available pine bark-based products [29]. The bark of douglas fir (Pseudotsuga menziesii Franco), one of the premier timber trees in the world, was found a rich source in taxifolin, which is broadly applied in pharmaceutical preparations [30].
10, x FOR PEER REVIEW 5 of 20 (PYC) and Flavangenol ® are good examples of commercially available pine bark-based products [29]. The bark of douglas fir (Pseudotsuga menziesii Franco), one of the premier timber trees in the world, was found a rich source in taxifolin, which is broadly applied in pharmaceutical preparations [30].
In Table 1, ESI negative ion mode MS and MS/MS data, molecular formulas, unsaturation degree (RDB-Ring and Double Bond) values and mass accuracy are listed.   In Table 1, ESI negative ion mode MS and MS/MS data, molecular formulas, unsaturation degree (RDB-Ring and Double Bond) values and mass accuracy are listed.

Benzoic and Hydroxycinnamic Acids Derivatives
Compounds 4, 5 and 6 were tentatively identified as (di)hydroxybenzoic acid hexosides and dihydroxybenzoic acid, respectively. In fact, deprotonated glycosides underwent homolytic and heterolytic cleavages of the hexose moiety, providing fragment ions at m/z 153.0191 and m/z 152.0112.
Compounds 7 and 16, previously identified in F. sylvatica leaves, were tentatively identified, in a ratio 1:2, as 3-O-and 5-O-caffeoyl quinic acid (Figure 2, panels A and B) based on the elution order and different fragmentation patterns [31]. The fragment ion at m/z 179.03 (deprotonated caffeic acid) was also identified in TOF-MS/MS spectra of compounds 8 (at m/z 341.0879), 9 (at m/z 297.0613) and 17 (at m/z 253.0718), which were tentatively identified as caffeoyl acid hexoside (C 15 H 18 O 9 ), caffeoyl threonic acid (C 13 H 14 O 8 ) and caffeoyl propanoic acid (C 12 H 14 O 6 ), respectively. The fragmentation pattern of metabolite 9 showed also product ions at m/z 135.0304 and m/z 117.0192 ( Figure 2D), corresponding to deprotonated threonic acid and its dehydrated derivative. Instead, for metabolite 17 the propionyl moiety (C 3 H 6 O 2 ) was identified by neutral loss of 74.04 Da ( Figure 2E). Other caffeoyl derivatives were the metabolites 3, tentatively identified as hydroxycaffeoyl quinic acid, and 12, that putatively corresponded to a dimer of caffeoyl quinic acid. In particular, in the TOF-MS/MS spectrum of metabolite 3, whose isomer, with antimicrobial activity against Staphylococcus aureus and Escherichia coli, was isolated from Hymenocrater calycinus (Boiss.) Benth. [32], the deprotonated molecular ion at m/z 371.1052 underwent neutral loss of 18 Da, providing the less intense ion at m/z 353.0872 (caffeoyl quinate), and gave rise to the abundant ion at m/z 191.0554 (quinate) ( Figure 2F). Deprotonated molecular ion of metabolite 12, at m/z 707.1845, underwent neutral loss of 174 Da and 192 Da, whose presence was further confirmed by fragment ion at m/z 191.0557. Moreover, according to TOF-MS/MS spectrum, the ring arrangement could be of the β-truxillic acid type; cleavages along the two axes of central core provided the fragment ions at m/z 463.1085, 353.0870 and 243.0651 ( Figure S1).
Furthermore, coumaroyl derivatives were also identified: compounds 11 and 20, at m/z 337.0921 (18), were tentatively identified as 3-O-and 5-O-p-coumaroyl quinic acid (pCoQAs) in accordance with the molecular formula C 16 Figure 2I). In all TOF-MS/MS spectra, the fragment ion at m/z 163.04 was detected, highlighting the presence of deprotonated coumaric acid (C 9 H 8 O 3 ). Compounds 21 and 22, whose deprotonated molecular ion was in accordance with the molecular formula C 16 Metabolites 24, 25 and 34 were hypothesized to be C-glycosylated flavanones, which are much less studied than O-glycosides, but endowed with several health benefits, such as antioxidant, anticancer, antitumor and anti-diabetic activities [33]. The occurrence of naringenin-C-glycosides was previously described by Hoffman et al. [25], but though an HPLC-MS/MS via Multiple Reaction Monitoring (MRM) analysis was performed, these authors did not provide sufficient MS/MS details to prove the presence of these compounds in beech leaves. The almost superimposable MS/MS fragmentation patterns were in accordance with naringenin C-hexoside isomers, likely bearing different sugar moieties that reasonably explained the different retention times (Figure 3). In fact, the deprotonated molecular ion at m/z 433.11 provided the ions [Ag+71] − at m/z 343.08 and [Ag+41] − at m/z 313.07 (base peak) by the 0,2 X and 0,3 X cross-ring cleavage of hexoside, likely linked at C-8 position. This hypothesis was supported by very low intensity of dehydrated fragment ions at m/z 415.10 and 325.07, which are known to be much more pronounced for 6-C isomers [34]. Although the occurrence of this kind of compounds is quite unusual, some literature data consoled our hypothesis. Indeed, naringenin 8-C-β-glucopyranoside (isohemiphloin) was isolated from Eucalyptus hemiphloia F. Muell. (Myrtaceae), and its 6-C isomer (hemiphloin) was identified also in Ononis vaginalis M.Vahl. (Fabaceae), Tulipa gesneriana L. and Ulmus wallichiana Planch., beside eriodictyol 6-C-β-Dglucopyranoside [35]. Different flavan-3-ols mono-and diglycosides have been identified too. In Figure 4, neutral losses and related molecular formulas of their main fragmentations are schematized. In particular, considering flavonols group, monoglycosides were 82.2%, diglycosides account 17.2%, and (acyl)-glycoside flavonols were only 0.51%. The deprotonated compounds 26 (m/z 479.0844) and 27 (m/z 449.0746) were putatively identified as myricetin 3-O-hexoside and myricetin 3-O-pentoside, respectively. In fact, the loss of 162 Da (hexose moiety) and 132 Da (pentose moiety) provided in both cases the fragment ion at m/z 317.02, attributable to myricetin, together with its aglycone radical anion at m/z 316.02, whose abundance allowed us to hypothesize the C-3 linkage of sugars moieties.  tized. In particular, considering flavonols group, monoglycosides were 82.2%, diglycosides account 17.2%, and (acyl)-glycoside flavonols were only 0.51%. The deprotonated compounds 26 (m/z 479.0844) and 27 (m/z 449.0746) were putatively identified as myricetin 3-O-hexoside and myricetin 3-O-pentoside, respectively. In fact, the loss of 162 Da (hexose moiety) and 132 Da (pentose moiety) provided in both cases the fragment ion at m/z 317.02, attributable to myricetin, together with its aglycone radical anion at m/z 316.02, whose abundance allowed us to hypothesize the C-3 linkage of sugars moieties.  The fragment ion at m/z 425.0863 was due to retro-Diels Alder (RDA) mechanism, and the monomeric unit at m/z 289.0710 was generated through the quinone methide fission [37]. Moreover, the deprotonated molecular ion at m/z 289.0708 for compound 15 was in accordance with the molecular formula C 15 H 14 O 6 and (epi)catechin compound. The decarboxylation (−44 Da) and subsequent loss of an ethenone unit (C 2 H 2 O), generated fragment ions at m/z 245.0818 and 203.0709, respectively; from the latter, the loss of lateral chain as 2-methylene-2H-pyran (C 6 H 6 O; −94.04 Da) for nucleophilic attack of hydroxyl group to benzylic carbon, provided base peak at m/z 109.0290. Following the benzofuran-forming fission reaction from deprotonated molecular ion the ion at m/z 123.0447 was formed [37].

Lignans
Compounds 29, 32, 36, and 47 were tentatively identified as lignans. In particular, metabolite 29 was tentatively identified as isolariciresinol hexoside, whose deprotonated molecular ion at m/z 521.2046 underwent neutral loss of sugar moiety providing the fragment ion at m/z 359.1504, from which the loss of formaldehyde provided the abundant ion at m/z 329.1396. Compounds 32 and 36 could be neolignan-O-deoxyhexoside isomers; these metabolites were previously reported as unidentified compounds [11]. Instead, two neolignan-9 -O-deoxyhexoside stereoisomers were isolated and identified by NMR in Fagus hayatae Palib. ex Hayata lea [38]. This finding is in line with the occurrence of these metabolites, which were also isolated in Pinus thunbergii [39]. Beside fragment ions at m/z 491.1926(61) and 473.1824 (30), which are derived from dehydration reactions, the bond cleavage between the two phenylpropanoid units provided fragment ions at m/z 179.0712(13) (base peak) and 313.1290(99). In Figure S2 TOF-MS/MS spectra were reported, and the hypothesized chemical structures for the most abundant fragment ions were provided.
The metabolite 47, whose tentative characterization was possible with a slight modification of Q-TOF parameters ( Figure S3), was identified as 4,9,9 -trihydroxy-3,3 ,5 -8-O-4 -neolignan-7-O-deoxyhexoside [40]. The loss of 212.06 Da, likely corresponding to trimethoxygallic acid, generated the fragment ion at m/z 343.1410, which in turn lost a methyl radical to generate the ion at m/z 328.1177. The presence of the deoxyhexose moiety was confirmed by loss of 146.06 Da, providing the ion at m/z 197.0821. As for compounds 32 and 36, the fragment ion derived by sugar loss has a very low intensity, likely justified by an intramolecular hydrogen bond. Compound 48 with the [M−H] − ion at m/z 551.2167, was supposed to be a lignan ( Figure S4): 9 -hydroxy-7 -propen-3 ,5 -dimethoxyphenyl-3methoxyphenyl-7,9-propanediol-4-O-hexoside [41]. The fragment ion at m/z 209.0815 was attributable to a sinapyl alcohol moiety, from which ions at m/z 194.0579 and 179.0344 were generated. To the best of our knowledge, this lignan has never been identified in F. sylvatica.

Fatty Acids
The investigated extract also showed the occurrence of mono-or poly-hydroxylated fatty acids, whose presence, together with epoxy derivatives, was previously identified by Matzke et al. [42]. In particular, metabolite 52, was tentatively identified as trihydroxy octadecadienoic acid (m/z 327.2185). It underwent dehydration processes, providing fragment ions at m/z 309.2087 and 291.1962. The C12-C13 bond cleavage provided the ion at m/z 229.1448, which in turn lost 58 Da (C 3 H 6 O) and 46 Da, to generate fragment ions at m/z 183.1385 and 171.1026, respectively ( Figure S5).
Compound 56 was putatively identified as hydroxyhexadecanoic acid, e.g., 16-OHhexadecanoic acid [33]), whose dehydration provided a low intensity ion at m/z 269.2114, whereas TOF-MS/MS spectrum of compound 61 indicated it could be dihydroxyoctadecedienoyl quinic acid. The loss of dehydrated quinic acid, identified by fragment ion at m/z 191.0559, produced an abundant ion at m/z 311.2230 that corresponded to the fatty acid. In addition to fragment ions at m/z 293.2117 and 275.2012, obtained by losses of hydroxy groups, also a fragment ion at m/z 223.1703, to cleavage between C-4 and C-5, was detected ( Figure S6).
Compound 63 was putatively a hydroxyoctadecatrienoic acid, maybe with hydroxyl group at carbon C-10. In fact, the β-scission of the alcoholic group provided the ion [M−H-110] − at m/z 183.1385 identified as base peak. An allyl scission provided ions at m/z 221.1539 and 211.1338 from dehydrated ion (m/z 275.2014) and deprotonated molecular ion, respectively. Compound 65 showed TOF-MS/MS spectrum in accordance with 15,16-dihydroxy-9,12-octadienoic acid [43]. Product ions deriving from water losses were detected at m/z 293.2125 and 275.2017. Furthermore, allyl scission gave a low abundance ion at m/z 253.1804, whereas the β-fission of the alcoholic hydroxyl group provided an abundant ion at m/z 223.1707. Compound 69 (at m/z 675.3629) was putatively identified as a linolenic acid glyceryl-dihexoside. In fact, the product ion at m/z 277.2158 corresponded to the fatty acid moiety, whereas glyceryl dihexoside could be identified by fragment ions at m/z 415.1454 and 397.1344; from the latter the loss of 92 Da, identifying a glycerol unit, provided the ion at m/z 305.0848. In Figure S7, the TOF-MS/MS spectrum and a putative fragmentation pathway are reported.

Other Minor Compounds
None of the remaining compounds were assignable to any of the previously discussed classes. Briefly, compound 1 was likely quinic acid, whereas metabolite 2 was putatively identified as tyrosine hexoside, based on the presence of the ion at m/z 180.0664 (deprotonated tyrosine), formed after the neutral loss of the hexose moiety. Metabolite 19 with the [M−H] − ion at m/z 387.1666 was tentatively identified as 12-hydroxyjasmonate (tuberonic acid) [18].

Relative Quantitation of F. sylvatica Leaf Chemical Constituents
Hydroxycinnamic acid (HCA) derivatives constitute a considerable part of low molecular weight phenol compounds. Caffeoyl-based HCA compounds were the most abundant, as they account for 88.0% of the compounds with C 6 C 1 and C 6 C 3 carbon skeleton. In particular, chlorogenic acids, such as 3-O-and 5-O-caffeoyl quinic acid, were 27.9% and 54.9%, respectively, whereas coumaroyl derivatives were less represented (~4.0%). The interest in chlorogenic acids is due to the plethora of beneficial effect ascribed to these substances, for which certain fruits, vegetables, spices are the main dietary sources. In particular, 5-O-caffeoyl quinic acid, firstly analyzed for its antioxidant, anti-inflammatory, and antitumor activity [46] was found to play multiple and key roles in protecting humans at neuronal, cardiovascular, and gastrointestinal levels. It was also implied in glucose and lipid metabolic regulation [47]. Besides caffeoyl derivatives, flavonoids appeared to be in appreciable amount. Flavonols were the main constituents of this class, with a relative percentage of 86.0%. Figure 5 shows the relative content of benzoic/hydroxycinnamic acids class, and flavonoids.
Kaempferol derivatives were the most abundant (66.0%), followed by quercetin derivatives (24.0%), and myricetin derivatives (10%). The abundance of flavonoids in different organs of F. sylvatica, such as leaves or bark [12,48,49], was stated by different literature data, in which kaempferol, quercetin, myricetin, luteolin or naringenin derivatives were identified. Unusual naringenin-C-glycosides are as part of flavanone class, which is commonly associated to different benefits due to their ability to act as free radical-scavenger. Juice and peel of citrus fruits are the main dietary sources of these compounds [50], which were also clinically evaluated for cardiovascular disease protection. The anti-inflammatory, antioxidant and soothing effects of flavonoids are broadly exploited for food, drugs or cosmetics production [51]. Kaempferol derivatives were the most abundant (66.0%), followed by quercetin derivatives (24.0%), and myricetin derivatives (10%). The abundance of flavonoids in different organs of F. sylvatica, such as leaves or bark [12,48,49], was stated by different literature data, in which kaempferol, quercetin, myricetin, luteolin or naringenin derivatives were identified. Unusual naringenin-C-glycosides are as part of flavanone class, which is commonly associated to different benefits due to their ability to act as free radical-scavenger. Juice and peel of citrus fruits are the main dietary sources of these compounds [50], which Apart from antiradical and reducing activities, food-derived flavonoids were shown to prevent non-communicable diseases on-set, and to exert pro-oxidant effect in cancer cells, thus increasing ROS levels and apoptosis rate. Several epidemiological evidences suggest that kaempferol-rich foods reduce the risk of liver, colon and skin cancer [52], whereas the uncountable properties of quercetin and its derivatives give rise recently-filed patents for disparate therapeutic applications [53]. Towards a green and sustainable waste valorisation chain, the employment of F. sylvatica leaves should be pursued.

Beech Leaf Alcoholic Extract Showed Antioxidant Efficacy in Cell-Free Assays
The alcoholic leaf extract was evaluated for its antioxidant capability by means of assays, involving stable radicals, such as DPPH • and ABTS •+ tests, and by ferricyanide FRAP assay. Folin-Ciocalteu test, which employs phospho-tungsto-molybdate, was also performed, together with the colorimetric determination of total flavonoid content. Data acquired showed that the extract was effective in scavenging ABTS •+ with an ID 50 and TEAC value equal to 31.4 ± 1.10 µg/mL and 0.27, respectively, and it was able to markedly reduce ferric ions, also at the lowest tested dose ( Figure 6A). The total phenol content (TPC) was 69.64 ± 3.1 mg of gallic acid equivalents per g of extract ( Figure 6B). Tanase et al. [24], who investigated TPC in beech bark hydroalcoholic extract, found it was 76.49 mg GAE/g plant material. Comparably, bark extracts of F. sylvatica, obtained by means of different extractive methods were screened for their antiradical activity and phenol content, and their diversity in catechins, taxifolin glycosides, procyanidins, syringic acid or coniferyl alcohol glycosides were further estimated by chromatographic techniques hyphenated to mass spectrometry [10,24,48,54]. Among the few studies on beech leaf, a comparative qualitative study was carried out on hydroalcoholic extracts (EtOH:water, 7:3, v/v) of beech leaves collected in Romania; it was evidenced that TPC significantly varied based on leaf harvest times [12,55], reaching the maximum TPC value (33.55 mg/g GAE), when leaves were collected in September. According to UHPLC-ESI-QqTOF analysis, a high content of flavonoid compounds was spectrophotometrically determined ( Figure 6B). ated to mass spectrometry [10,24,48,54]. Among the few studies on beech leaf, a comparative qualitative study was carried out on hydroalcoholic extracts (EtOH:water, 7:3, v:v) of beech leaves collected in Romania; it was evidenced that TPC significantly varied based on leaf harvest times [12,55], reaching the maximum TPC value (33.55 mg/g GAE), when leaves were collected in September. According to UHPLC-ESI-QqTOF analysis, a high content of flavonoid compounds was spectrophotometrically determined ( Figure 6B).

Beech Leaf Alcoholic Extract Decreased Intracellular ROS in Caco-2 Cells
The extract richness in polyphenols and the positive responses of cell-free antioxidant tests paved the way to the evaluation of its antioxidant potential in cell-based assays. To this purpose, Caco-2 cell line represents a useful model, able to predict in vivo behavior. In fact, besides being widely used as an in vitro model for the intestinal epithelial barrier, they were reported to offer a distinctive advantage in screening intestinal absorption of

Beech Leaf Alcoholic Extract Decreased Intracellular ROS in Caco-2 Cells
The extract richness in polyphenols and the positive responses of cell-free antioxidant tests paved the way to the evaluation of its antioxidant potential in cell-based assays. To this purpose, Caco-2 cell line represents a useful model, able to predict in vivo behavior. In fact, besides being widely used as an in vitro model for the intestinal epithelial barrier, they were reported to offer a distinctive advantage in screening intestinal absorption of natural antioxidants after oral intake [21]. Thus, with the aim to exploit beech leaf phytochemicals in the formulation of nutraceuticals, the ability of the extract to counteract reactive oxygen species (ROS) overproduction was preliminarily evaluated in Caco-2 cells [49]. In fact, metabolic activity is strongly related to ROS production, which intracellularly originate in the mitochondrial respiratory chain, also producing toxic metabolic byproducts [56]. Thus, redox mitochondrial activity was first assessed by means of MTT test ( Figure 7A). The beech leaf alcoholic extract exhibited dose-response inhibition of the metabolic activity of Caco-2 cells, with an ID 50 equal to 148.4 µg/mL. In the cell model, after co-exposing Caco-2 cell monolayer to increasing dose levels of the extract and to the fluorescent probe DCFH-DA, oxidative stress is induced by 2,2-azobis(2-amidinopropane) dihydrochloride (AAPH). Based on fluorescence levels, which are related to the oxidation degree, it was observed that the extract markedly reduced the cellular fluorescence compared to the untreated cells, when it was tested at dose level equal to 200 µg/mL ( Figure 7B). The other doses tested, while exhibiting less pronounced efficacy, showed a strongly time-dependent activity. Overall, the dose-dependent capability of the beech leaf extract was ascertained. Considering the cell basal rate in oxidizing species, the outcome of the investigated beech on redox balance could be due to extract-induced mitochondrial alterations. The effect was compared to that exerted by pure standard quercetin, which was tested at 10 µM, based on cell viability data and according to literature [21,57]. In fact, the peroxyl radical scavenging activity of quercetin in Caco-2 cells was broadly proved [21], and it was found that doses of the flavonol higher than 20 µM markedly affected colon cells viability [57] Antioxidants 2021, 10, x FOR PEER REVIEW 17 of 20 investigated beech on redox balance could be due to extract-induced mitochondrial alterations. The effect was compared to that exerted by pure standard quercetin, which was tested at 10 µM, based on cell viability data and according to literature [21,57]. In fact, the peroxyl radical scavenging activity of quercetin in Caco-2 cells was broadly proved [21], and it was found that doses of the flavonol higher than 20 µM markedly affected colon cells viability [57]

Conclusions
Fagus sylvatica leaves, investigated through an untargeted UHPLC-HR-MS/MS analysis, revealed their richness in flavonoids, mainly flavonol monoglycosides. Thus, the recovery of this renewable plant source could favor an application to create beech leaf-derived products, in the form of food or dietary supplements or herbal remedies for human or animal health. Cell-free assays evaluating the antiradical and reducing capability, as well as preliminary antioxidant capability in Caco-2 cells, strengthened this exploitation

Conclusions
Fagus sylvatica leaves, investigated through an untargeted UHPLC-HR-MS/MS analysis, revealed their richness in flavonoids, mainly flavonol monoglycosides. Thus, the recovery of this renewable plant source could favor an application to create beech leaf-derived products, in the form of food or dietary supplements or herbal remedies for human or animal health. Cell-free assays evaluating the antiradical and reducing capability, as well as preliminary antioxidant capability in Caco-2 cells, strengthened this exploitation hypothesis. Thus, the promising antioxidant results herein reported could be the basis for the further phytochemical investigation of F. sylvatica, and the use of different extractive and chromatographic approaches aimed at recovering high amounts of pure beech bioactive compounds.