Identification and Accumulation of Phenolic Compounds in the Leaves and Bark of Salix alba (L.) and Their Biological Potential

The study examines the phenolic compounds in hydromethanolic extracts of Salix alba (L.) leaves and bark as well as their antioxidant activity and cytotoxic potential. UPLC-PDA-Q/TOF-MS analysis showed a total of 29 phenolic compounds in leaves and 34 in bark. Total phenolic compound content was 5575.96 mg/100 g of dry weight (DW) in leaves and 2330.31 mg/100 g DW in bark. The compounds were identified as derivatives of phenolic acids (seven in leaves and five in bark), flavanols and procyanidins (eight in leaves and 26 in bark) and flavonols (14 in leaves and three in bark). Both extracts exhibited strong antioxidant potential, assessed by radical scavenging activity against 1,1-diphenyl-2-picrylhydrazyl (DPPH) and 2,2′-azinobis (3-ethylbenzothiazoline-6-sulfonic acid (ABTS), but the bark extract was even stronger than the ascorbic acid used as a standard. The cytotoxicity of both extracts was evaluated against human skin fibroblasts and human epidermal keratinocytes cell lines using the Presto Blue cell viability assay. The keratinocytes were more resistant to tested extracts than fibroblasts. The leaf and bark extracts at concentrations which exhibited antioxidant activity were also not toxic against the keratinocyte cell line. Thus, S. alba extracts, especially the leaf extract, offer promise as a nontoxic natural antioxidant, in cosmetic products or herbal medicines, and as a source of bioactive secondary metabolites.


Introduction
The last few years have seen a growth in the need for natural drugs obtained from medicinal plants as an alternative to synthetic products [1]. A number of these bioactive compounds serve as starting points for the development of new drugs [2]. Clinical studies have shown that some herbal medicines are better tolerated by patients and show fewer side effects than synthetic derivatives [3].
Willow bark extract has been used for thousands of years as an analgesic, antipyretic and anti-inflammatory agent. Records suggest that, as long as 6000 years ago, white willow was used in Mesopotamia. Hippocrates recommended chewing willow bark for patients suffering from fever, inflammation and pain [4]. Currently, willow bark is used to treat many different types of pain, such as

Plant Material
Bark and leaves of Salix alba (L.) were collected in Lębork, Poland (54 • 32 N 17 • 45" E) (22 April 2019). The plant was identified by E. Piątczak according to Rutkowski [22]. The voucher specimen was deposited at the Department of Biology and Pharmaceutical Botany, Medical University of Łódź, Poland (No. EP/S.a./2019). The plant material was air dried at room temperature for two weeks protected from light.

Preparation of Extracts
Dry plant material (1 g was used for antioxidant experiments and 100 mg was weighted for UPLC analysis) was pre-extracted with chloroform overnight to remove chlorophyll. After filtration, the remaining plant material was extracted three times with a 30 mL methanol/water mixture 7:3 (v/v) for 30 min in an ultrasonic bath at 40 • C. Then, the mixture was filtered through a paper filter, the extracts were combined and evaporated to dryness under reduced pressure. The yields of the extracts were as follows: 32.8% (leaf extract) and 34.6% (bark extract).

UPLC-PDA-Q/TOF-MS Analysis of Polyphenols
The polyphenols in the S. alba extracts were identified using an ACQUITY Ultra Performance LC system equipped with a photodiode array detector with a binary solvent manager (Waters Corporation, Milford, MA, USA) with a mass detector G2 Q-TOF micro mass spectrometer (Waters, Manchester, UK) equipped with an electrospray ionization (ESI) source operating in negative mode. Individual polyphenols were separated using a UPLC BEH C18 column (1.7 mm, 2.1 × 100 mm, Waters) at 30 • C. A 10 µL volume of samples were injected, and the elution was completed in 15 min with a sequence of linear gradients and constant flow rates of 0.42 mL/min. The mobile phase consisted of solvent a (0.1% formic acid, v/v) and solvent B (100% acetonitrile). The linear gradient was as follows: 0.0-1.0 min, 99% A, 0.42 mL/min (isocratic), 1.0-12.0 min, 65.0% A, 0.42 mL/min (linear), 12.0-12.5 min, 99% A, 0.42 mL/min (linear), 12.5-13.5 min, 99% A, 0.42 mL/min (isocratic). The analysis was carried out using full scan, data-dependent MS scanning from m/z 100-1500. Leucine-enkephalin was used as the reference compound at a concentration of 500 pg/mL, and the [M-H] − ion at 554.2615 Da was detected. The [M-H] − ions were detected during a 15 min analysis performed within ESI-MS accurate mass experiments, which were permanently introduced via the LockSpray channel using a Hamilton pump. The lock mass correction was 1000 for the mass window. The mass spectrometer was operated in negative-ion mode, set to the base peak intensity (BPI) chromatograms, and scaled to 12,400 counts per second (cps) (100%). The optimized MS conditions were as follows: capillary voltage of 2500 V, cone voltage of 30 V, source temperature of 100 • C, desolvation temperature of 400 • C and desolvation gas (nitrogen) flow rate of 600 L/h. Collision-induced fragmentation experiments were performed using argon as the collision gas, with voltage ramping cycles from 0.3 to 2 V. Characterization of the single components was carried out based on the retention time and the accurate molecular masses. Each compound was optimized to its estimated molecular mass in the negative mode, before and after fragmentation. The data obtained from UPLC-MS were subsequently entered into the MassLynx 4.0 ChromaLynx Application Manager software (Waters).
The runs were monitored at the following wavelengths: flavan-3-ols and procyanidins at 280 nm, phenolic acids at 320 nm and flavonols at 360 nm. The PDA spectra were measured over the wavelength range of 200-600 nm in steps of 2 nm. The retention times and spectra were compared to those of the authentic standards.
The quantification of phenolic compounds was performed by external calibration curves, using reference compounds selected based on the principle of structure-related target analyte/standard (chemical structure or functional group). The calibration curve for 3-caffeoylquinic acid was used to quantify caffeoylquinic acid derivatives. The calibration curve for caffeic acid was used to quantify caffeoylhexose derivatives. Then, 1-caffeoylquinic and 5-caffeoylquinic acids were quantified with their own standards. The calibration curve for procyanidin B2 was used to quantify all B-type procyanidins. (+)-catechin and (−)-epicatechin were quantified with (+)-catechin standard. The calibration curves of quercetin rutinoside, 3-O-glucoside and 3-O-galactoside were used to quantify quercetin derivatives. For isorhamnetin quantification, isorhamnetin 3-O-rutinoside and 3-O-glucoside were used.
All determinations were done in triplicate (n = 3).

ABTS Assay
The antioxidant activity was determined using the ABTS radical cation decolorization test according to Re et al. [23] and Guss et al. [24] using a Shimadzu UV-1800 UV/Vis spectrophotometer. An ABTS stock solution was prepared by 19.5 mg ABTS (Sigma-Aldrich Sp. z o.o., Poznań, Poland) mixed with 50 mM phosphate buffer at pH 7.4 and 3.3 mg potassium persulfate (Sigma-Aldrich Sp. z o.o., Poznań, Poland). This mixture was kept in the dark at room temperature for 14 h before use. The ABTS stock solution was diluted with phosphate buffer at pH 7.4 to give an absorbance of the negative control (0.9 ± 0.05) at 734 nm. The assay was performed by adding 0.5 mL of each willow extract (from leaves and bark, separately) at concentrations of 7.5, 10.0, 15.0, 25.0 and 50.0 µg/mL to 0.5 mL of ABTS solution. The absorbance was measured after 10 min at 734 nm. The results were expressed as EC 50 , the concentration of the sample (in µg of dry extract/mL) at which 50% of maximum scavenging activity was recorded. Ascorbic acid was used as an antioxidant standard.

DPPH Assay
Radical scavenging activity of samples was evaluated by spectrophotometry by monitoring the decrease of absorbance at 517 nm of methanolic solution of the radical DPPH (2,2-diphenyl-1-picrylhydrazyl) incubated for 30 min in the darkness in the absence or in the presence of plant extracts [25,26]. The DPPH solution was diluted with methanol/water mixture 7:3 (v/v) to give an absorbance of the negative control (0.9 ± 0.05) at 517 nm. The assays were performed by adding 0.5 mL of each plant extract at concentrations of 1.0, 5.0, 10.0, 15.0, 20.0, 25.0, 50.0, 100.0 (µg/mL) to 0.5 mL of DPPH solution. The concentration of the extract (µg of dry extract/mL), necessary to reduce the initial concentration of DPPH by 50% (EC 50 ) under the experimental conditions was calculated. Ascorbic acid was used as a positive control.
The cells were cultured at 37 • C in a humidified 5% CO 2 incubator.

Assay for Cytotoxic Activity
The extract of S. alba leaf and bark were tested for in vitro cytotoxicity against 1BR.3.N and human epidermal keratinocytes, Neonatal cells by Presto Blue cell viability assay [27] according to the manufacturer's instructions.
The stock solution (1 mg/mL) of plant extracts were dissolved in a mixture of methanol/water (7:3 v/v). The final dilution of the stock solutions was made in culture medium. Cells were also treated with vehicle control (i.e., equivalent amounts of a 7:3 methanol/water solution mixture without extracts).
Keratinocytes and fibroblasts were seeded into 96-well plates in the full medium at a density of 10 4 cells/well. Twenty-four hours after seeding, the cells were treated with different concentrations of plant extracts (from 5 to 200 µg/mL) or with vehicle control in either full medium (for keratinocytes) or medium without FBS (for fibroblasts). After 24 h of incubation, cytotoxicity was evaluated using the Presto Blue cell viability assay on a VICTOR™ ×4 Multilabel Plate Reader (Perkin Elmer, London, UK). IC 50 values, which means the extract concentration exhibited 50% cell growth inhibition, were calculated using the GraphPad Prism 5.01 software tool.

Statistical Analysis
Each experiment was performed in triplicate and repeated at least three times. All data are presented as means ± standard deviation (SD). EC 50 values (antioxidant activity) were calculated using Microsoft Excel software. IC 50 values (cytotoxic activity) were calculated using the GraphPad Prism 5.01 software tool.
The statistical significance between both extracts was determined using a nonparametric, distribution-free Mann-Whitney U-test or the Kruskal-Wallis test. Differences between means were considered significant at p ≤ 0.05 for both tests. All the statistical analyses were calculated using Statistica software (version 13.1, 2016, STATSoft, Cracow, Poland).

Identification of Phenolics in Salix alba Extracts
In the present study, the phytochemical profile of hydromethanolic extracts of S. alba leaves and bark was analyzed using UPLC-PDA-Q/TOF-MS. The results of the qualitative and quantitative analyses are summarized in Table 1 and Figures 1-3. Twenty-nine phenolic compounds were identified in the leaf extracts and 35 in the bark extracts. They included derivatives of phenolic acids, flavanols, procyanidins and flavonols.
Caffeic acid hexoses were detected in both the bark and leaves.

Flavanols and Procyanidins
Twenty-six compounds from the flavanols and procyanidins group were found in the S. alba bark extract. These compounds consisted mainly (epi)catechin derivatives such as A-type and B-type procyanidins (dimers, trimers and tetramers) and their digallates; (epi)catechin-(epi)gallocatechins or (epi)catechin methyl-hexosides. Much fewer compounds from this group of derivatives were found in leaves that included two epigallocatechins, A-type procyanidin dimer and five B-type procyanidins (one dimer, three trimers and one pentamer).  [33]. The compounds have been also widely distributed in leaf extracts of several Salix species [15], including S. alba, S. subserrata, S. fragilis, S. daphnoides, S. caprea, S. cinerea [19,34]. Catechin was also a predominant compound in water extracts of S. aegyptiaca leaves [35] and buds of S. pyrolifolia [28].
In the bark and leaves a multitude of isomeric A-type procyanidins dimers (peaks 10, 12, 14 Proanthocyanidins are oligomers or polymers of flavan-3-ols, which are commonly distributed in plants, for example in fruits (apple, grape, cranberry), red wine, cinnamon and the hawthorn inflorescences [36]. The metabolites exhibit various biological activities, including antihypertensive, antioxidative, and anti-inflammatory properties [15]. The possible mechanism of anti-inflammatory action of proanthocyanidin dimers, B1 and B2, may be connected with the inhibition of transcription of nuclear factor-kappa B (NF-κB) [37]. The B-type procyanidin identified in white willow leaves in the present study were proanthocyanidins consisting of catechin and/or epicatechin units. Earlier, the presence of procyanidins (oligomers and polymers) (B and C-type) has been confirmed in the leaves of three willow species, including S. alba bark extracts [19,21,38]. Among other species, several procyanidin a and B-type derivatives have been identified in Gaultheria procumbens leaf extracts [39] and six different dimers and three different trimers of B-type procyanidin have been identified in Sorbus demestica leaf extract [32]. A-type procyanidin dimer digallate (peak 8, t r = 3.87 min) and procyanidin trimer digallate (peak 23, t r = 5.12 min) demonstrated pseudomolecular ions at m/z 881 and 1169, respectively. Each compound demonstrated fragmentation ions at m/z 289 assigned to epichatechin. The fragmentation patterns of trimeric compound (peak 23) at m/z 289, 577 and 865 confirmed the presence of monomeric, dimeric and trimeric epicatechin, respectively, and the loss of digallate residue (304 u) from a main pseudomolecular ion at m/z 1169.
The last compound present in the willow bark-(epi)catechin-ethyl trimer (peak 56, t r = 11.09 min) with the main ion at m/z 893 and fragmentation ions at m/z 603 and 289 revealed that the compound fragmentized into a (epi)catechin-ethyl (317 u) moiety and two (epi)catechins (289 u).

Flavonols
The S. alba leaf extract was particularly rich in flavonol derivatives, while the bark extract had a negligible amount of the compounds, including only three derivatives.
As for the content of flavonols in the bark extract, only three derivatives were detected, i.e.,

Quantitative Analysis of the S. alba Leaf and Bark Extracts
Quantitative analysis of dry extracts of S. alba leaves and bark was conducted for the presence of polyphenolic compounds, such as phenolic acids, flavanols and procyanidins, and flavonols (Table 1; Figure 3). The analyses were performed based on external calibration curves using selected reference compounds (Materials and Methods: Section 2.4). The concentration of the individual substances was expressed in mg/100 g DW.
The two extracts had different phenolic compound contents. S. alba leaf extract was a richer source of total active phenolic ingredients than bark extract: 5575.96 mg/100 g DW in the leaves compared to 2330.31 mg/100 g DW in the bark, i.e., 58.2% higher content.
Flavonol content was up to 98.5% higher in the leaves than the bark, with only trace amounts in the bark. Similarly, the leaves also had 66% more phenolic acids than the bark. Total flavanols and procyanidins content was also 11% higher in the leaves than in the bark.
In the leaves, the number of flavonols was 2074.21 mg/100 g DW, similarly as the content of the flavanols and procyanidins, which was equal to 2011.12 mg/100 g DW. Among the flavonols, quercetin derivatives accounted for 56.3% (1168.33 mg/100 g DW); the remaining 43.7% comprised isorhamnetin derivatives (905.88 mg/100 g DW). Phenolic acids constituted a significant amount of this extract and were equal to 1490.63 mg/100 g DW.
In the bark, the flavanols and procyanidins constituted as much as 77.1% of all considered phenolic compounds (1795.85 mg/100 g DW). Phenolic acid content was about 3.5 fold lower (504.87 mg/100 g DW) than in leaves. The content of flavonols in the bark was negligible (29.59 mg/100 g DW) ( Table 1).

The Antioxidant Activity
The antioxidant activity of the hydromethanolic leaf and bark extracts was determined by DPPH and ABTS − radical scavenging assays ( Table 2). The bark extract of S. alba exhibited stronger antioxidant activity (48% in DPPH and 33% in ABTS), than the leaf extract. However, leaf extract activity was also very strong, and similar (at p ≤ 0.05) to the natural reference antioxidant: ascorbic acid (EC 50 = 28 µg/mL in DPPH assay and 65 µg/mL in ABTS assay) ( Table 2). The bark extract demonstrated lower EC 50 values (EC 50 = 13.5 µg/mL in DPPH assay and 21.5 µg/mL in ABTS assay) than the reference standard, indicating even stronger antioxidant activity than ascorbic acid. A previous DPPH measurement of antioxidant activity in S. alba bark extract showed slightly lower activity (EC 50 = 19.1 µg/mL) [41]. Another study reported even lower antioxidant activity in DPPH (58-65 µg/mL) and ABTS (47-53 µg/mL) assays, depending on the granulometry classes of the powdered bark extract [21]. Among other Salix species, S. atrocinerea and S. viminalis bark extracts exhibited similar activities in the DPPH test (EC 50 = 10.98 and 14.06 µg/mL, respectively) as in the present study. On the other hand, the bark extract of S. fragilis demonstrated lower activity (EC 50 = 23.62 µg/mL) [42]. The results of the ABTS assay are comparable with those achieved earlier for other willow species, including S. purpurea barks (EC 50 = 20 µg/mL) [43].
Our paper is the first to describe the antioxidant potential of leaf extract of white willow. The extracts from leaves of other Salix species, for example S. subserrata and S. aegyptiaca [34,35], exhibited lower antioxidant activity in DPPH assay than leaf extract of S. alba in our study. The differences in the antioxidant capacity observed between even the same species, in the same assay, could be caused by differences in phytochemical profile, time of harvesting or even extraction method of the plant material.
The greatest influence on the antioxidant properties of the willow extracts was probably exerted by phenolic compounds, present in each extract. These are mainly phenolic acid derivatives and flavonols (quercetin and isorhamnetin derivatives) in leaves. In the bark extract, catechin, epicatechin and a and B-type procyanidin derivatives are probably responsible mainly for the properties.

Cytotoxic Activity
The hydromethanolic extracts of S. alba leaves and bark were tested for in vitro cytotoxicity, using two human skin cell lines: human skin fibroblasts (1BR.3.N) and human epidermal keratinocytes, neonatal cells. These two cell lines were chosen to check the cytotoxic effect of tested extracts (in the range of concentrations from 5 to 200 µg/mL) on skin cells and check the possibility of external use of the extracts, for example in cosmetology.
The S. alba bark and leaf extracts demonstrated different cytotoxic activity against fibroblasts and keratinocytes (Figure 4a,b). The obtained IC 50 values were higher for keratinocytes than for fibroblasts, which means that both extracts were less cytotoxic to keratinocytes than fibroblasts. The cytotoxic effect of both S. alba bark and leaves was similar as the vehicle control, i.e., equivalent amounts of methanol/water solution (7:3), for keratinocytes (IC 50 values of 46.08 and 69.33 µg/mL, respectively, versus 51.51 µg/mL for vehicle control). In the case of fibroblasts, the cytotoxic activity of S. alba bark was almost twice that of leaves (IC 50 values of 14.06 and 25.92 µg/mL, respectively, versus 48.22 µg/mL for vehicle control). It can be concluded that keratinocytes were more resistant to the cytotoxic effect of both S. alba extracts (Figure 4a). However, in the case of fibroblasts, the leaf extract demonstrated lower cytotoxic potential than bark extract (Figure 4b).
However, it is important to note that both extracts exhibited significant antioxidant activity at concentrations that were not toxic to the keratinocytes (EC 50 = 13.5 µg/mL in DPPH assay and 21.5 µg/mL in ABTS assay). Therefore, the extracts may be active ingredients of products used externally, which are safe for the epidermis. However, lower concentrations should be used in cosmetics acting deeper in the skin, because of the higher toxicity of the tested extracts, especially bark extract, to the fibroblast cells of the dermis.
In earlier studies, S. alba bark extract has demonstrated cytotoxic and genotoxic potential against different cell lines, for example the human leukemia (HL-60) cell line [6], human peripheral leukocyte cells and the human hepatoma cell line, with doses from 50 and 200 µg/mL [44]. However, other extracts from bark of S. atrocinerea and S. fragilis did not exert a cytotoxic effect (at 625 and 1250 µg/mL) against immortalized human keratinocytes (HaCaT) and mouse fibroblast (L929) cells [42]. The differences between cytotoxicity dose could be connected with differences in the solvent or cell line model. Our paper is the first investigation of the cytotoxic potential of S. alba extracts on skin cell lines.

Conclusions
Although extracts from Salix sp. are commonly used for treatment, mainly due to salicin content, the phytochemical profiles of polyphenol compounds in white willow leaves and bark extracts appear to have a potentially beneficial influence on human health due to the presence of quercetin glycosides, monomeric, dimeric and trimeric flavan-3-ol derivatives, including B-type procyanidins, as well as caffeoylquinic pseudodepsides. These specialized secondary plant metabolites are known to exhibit a wide range of biological activities. Moreover, both extracts of white willow described in the present study exhibited strong antioxidant activity at concentrations that were not toxic to the keratinocytes. Therefore, the extracts, especially leaf extract, may be used as a potential new source of bioactive polyphenols with possible applications in some antiaging cosmetics, which would be safe for the epidermis. Bark extracts, also rich in polyphenolic compounds, should be used at lower concentrations than leaf extracts, especially in cosmetics that penetrate deeper into the skin, because of their higher toxicity to fibroblasts. However, further studies are needed, particularly those connected with the potential for utilization and extract standardization.