Valorization of European Cranberry Bush (Viburnum opulus L.) Berry Pomace Extracts Isolated with Pressurized Ethanol and Water by Assessing Their Phytochemical Composition, Antioxidant, and Antiproliferative Activities

Defatted by supercritical CO2, Viburnum opulus berry pomace (VOP) was subjected to consecutive extraction with pressurized ethanol (E) and water (W) and yielded 23% of VOP-E and 8% of VOP-W, respectively. The major phytochemical groups covering 42 identified and quantified constituents in VOP extracts were organic and phenolic acids, iridoids, quercetin and (epi)catechin derivatives, flavalignans, procyanidins, and anthocyanins. The on-line HPLC-DPPH•-scavenging assay revealed the presence of numerous antioxidants. VOP-E had a higher total phenolic content, was a stronger antioxidant (equivalent to 0.77, 0.42, and 0.17 g trolox/g in oxygen radical absorbance capacity (ORAC), ABTS, and DPPH assays, respectively), and recovered the major part of phenolics from the pomace; however, both extracts demonstrated similar antioxidant activity in the cellular assay. VOP-E inhibited HT29 cancer cells at non-cytotoxic concentrations. The results of this study revealed that VOP contains valuable phytochemicals possessing antioxidant and antiproliferative activities. Consequently, extracts from VOP substances may be of interest in developing functional ingredients for healthy foods, nutraceuticals, and cosmeceuticals.


Introduction
Many berry varieties are well-known fruits and are consumed as fresh foods or used as a raw material for various processed products such as jams, juices, beverages, and others. In addition to systematically studied, comprehensively valorized, and widely commercialized berries, there are many underutilized species, which produce the fruits with specific, rather unusual sensory quality and are therefore less acceptable by the consumers as fresh commodities. However, more recently, some of the underutilized berry species have attracted an increasing interest as promising sources of health beneficial bioactive compounds, which, after comprehensive evaluation and systematic valorization, might find wider applications in the production of both regular and, particularly, rapidly developing functional foods, nutraceuticals, and cosmeceuticals [1][2][3][4][5].
European cranberry bush (Viburnum opulus L.) berries, also called guelder rose berries, may be assigned to the underutilized fruits and emerging source of food grade products and ingredients [1].

Proximate Analysis
The chemical composition of VOP was determined according to the procedures established by the Association of Official Analytical Chemists (AOAC, 1990) [22]: moisture by drying at 105 • C to constant weight; ash by mineralizing in a muffle furnace F-A1730 (Thermolyne Corp., Dubuque, IA, USA) at 500 • C for 3 h; proteins by the Kjeldahl method in a nitrogen analyzer (Leco Instruments Ltd., Mississauga, ON, Canada) using a conversion factor of 6.25; crude lipids by Soxhlet extraction with hexane for 6 h. The rest of the dry matter was assigned to carbohydrates. Each determination was carried out in triplicate.

Total Phenolic Content (TPC) and Antioxidant Capacity
Folin-Ciocalteau (TPC), DPPH, ABTS, and ORAC assays were used for evaluating antioxidant potential of VOP extracts [22]. TPC was performed as described by Singleton and Rossi (1965) with some modifications [23]. A total of 30 µL of extract solutions was mixed with 150 µL Folin-Ciocalteu reagent (1:10 in distilled water) with 120 µL 7% Na 2 CO 3 solution in a 96-well microplate and the absorbance was measured at 765 nm after 30 min in a FLUOstar Omega Reader (BMG Labtech, Offenburg, Germany). Gallic acid solutions (10-250 µg/mL) were used for the calibration curve. TPC was expressed in mg of gallic acid equivalents in g of extract and pomace, GAE/g DWE (dry weight of extract) and DWP (dry weight of pomace), respectively. Re et al.'s (1999) method with some modifications was used for ABTS •+ decolourisation assay [24]. A total of 6 µL of sample was added to 294 µL of ABTS •+ working solution. The ABTS •+ solution was prepared by mixing 50 mL of ABTS (2 mM) with 200 µL of potassium persulfate (70 mM); the mixture was kept in the dark for 15-16 h before use. The working solution was prepared by diluting with PBS (prepared from 8.18 g of NaCl, 0.27 g of KH 2 PO 4 , 1.78 g of Na 2 HPO 4 × 2H 2 O, and 0.15 g of KCl in 1 L of distilled water) to obtain the absorbance of 0.800 ± 0.030 at 734 nm. The absorbance was measured in a 96-well microplate using a FLUOstar Omega Reader during 30 min at 734 nm. A series of Trolox solutions (399-1198 µM/L) were used for calibration. The results were expressed as µM TE/g DWE and DWP.
Brand-Williams et al.'s (1995) method was applied for the DPPH • -scavenging assay with some modifications [25]. A total of 8 µL of sample was mixed with 292 µL of DPPH • methanolic solution (0.0059 g/250 mL). The working solution was prepared by diluting with methanol to obtain the absorbance of 0.7 at 515 nm. After mixing, the microplate was placed in a reader, shaken for 30 s, incubated for 60 min, and the absorbance read at 515 nm. The results were calculated as in the ABTS assay. ORAC assay was performed by using fluorescein as a fluorescent probe and AAPH as a peroxyl radical generator [26]. A total of 25 µL of sample was pipetted into 150 µL (14 µM) fluorescein solution, and after incubation during 15 min at 37 • C, 25 µL of AAPH (240 mM) were added. The fluorescence was recorded every cycle (in total, 120 cycles) using 485 excitation and 530 emission fluorescence filters. Antioxidant curves (fluorescence versus time) were first normalized, and from the normalized curves, the net area under the fluorescein decay curve (AUC) was calculated by the following formula: AUC = (1 + f 1 f 0 + f 2 f 0 . . . fi f 0 ) × CT (1), where f 0 is the initial fluorescence reading at 0 min, f i is the fluorescence reading at time i, and CT is cycle time in minutes. The final ORAC values were calculated by using the regression equation between the Trolox concentration and the net AUC. Trolox solutions (0-250 µM) were used for the calibration. The results were expressed as µM TE/g DWE and DWP. Each measurement was carried out in six replicates.

HPLC-DPPH • -Scavenging On-Line
HPLC-DPPH • -scavenging online analysis was carried out on a Waters HPLC system (Waters Corporation, Milford, MA, USA) as previously described [21] with minor changes. The Hypersil C18 analytical column (250 × 0.46 cm, 5 µm; Supelco Analytical, Bellefonte, PA, USA) was used for compound separation. The mobile phase consisted of 0.4% formic acid in water (A) and acetonitrile (B). The elution profile was at the start point 90% A, then changing to 60% in 45 min, after that, in 50 min A was decreased to 5%, then in 53 min increased to 90% A, then it was held at 90% for 3 min, and in 2 min it was returned to initial conditions and the column was equilibrated for 5 min. The decrease of absorbance after the reaction of radical scavengers with DPPH • was measured at 515 nm with the variable-wavelength Shimadzu SPD-20A UV detector (Shimadzu Corporation, Kyoto, Japan), while the identification of compounds was performed by using the Waters Acquity UPLC system (Milford, MA, USA).

UPLC-ESI-QTOF-MS Analysis
Quantitative and qualitative analysis was performed with an ultra-performance liquid chromatograph (Waters Acquity UPLC system, Milford, MA, USA) equipped with a quadrupole time-of-flight tandem mass spectrometer (MaXis 4G Q-TOF-MS). Data acquisition and instrument control were performed by using HyStar 3.2 (SR2 software, Bruker Daltonics, Bremen, Germany) [21]. Briefly, samples were eluted with a gradient of solvent A (1% formic acid in ultrapure water) and B (acetonitrile) on a 1.7 µm, 100 mm × 2.1 mm i.d. Acquity BEH C18 column (Waters) over 14 min at a flow rate of 0.4 mL/ min. The injection volume was 1 µL and column temperature was maintained at 40 • C. Gradient elution was performed as follows: 95% A in 0-4 min, 95-90% A in 4-6 min, 90-70% A in 6-10 min, 70-5% A in 10-12 min, and 5-95% A in 12-14 min. The Q-TOF mass spectrometer was equipped with an electrospray ionization (ESI) source in a negative-ion mode. ESI-Q-TOF-MS spectra were recorded in the mass range of 80-1200 m/z. Two scan events were applied, namely, full-scan analysis followed by data-dependent MS/MS of the most intense ions. The data-dependent MS/MS used 20-30.0 eV as collision energies (fragmentation was performed by using collision induced dissociation (CID)); the capillary voltage was +4000 V; end plate offset 500 kV; flow rate of drying (N 2 ) gas 10.0 L/min; nebulizer pressure 2.5 bar. The same column was used for sugars, while separation was performed using 75% acetonitrile, 25% water, and 0.1% ammonia hydroxide as mobile phase at isocratic conditions at 0.35 mL/min, 10 min, and 35 • C. The method for anthocyanins is described elsewhere [27]. The quantification was performed using cyanidin-3-glucoside as an external standard. The amounts of individual identified compounds were expressed as mg/100 g DWP. A quantitative analysis of other compounds was performed by using calibration curves (0.1-10 µg/mL) of available standards (malic acid, rutin, quinic acid, citric acid, catechin, proanthocyanidin B2, chlorogenic acid, quercetin-3-O-glucoside; for sugars glucose, fructose and sucrose were used). Concentrations of phytochemicals were determined from the peak areas, using the linear regression equations of calibration curves obtained from six concentrations of each standard. Quantification trace for quantified promising substances for developing natural antioxidants and functional food ingredients with various bioactivities. In this study, a proximate analysis of VOP was performed. The content of protein, fat, water, and ash in VOP was 16.82 ± 1.54%, 26.24 ± 0.59%, 7.76 ± 0.16%, and 1.21 ± 0.03%, respectively. Other macrocomponents should consist of various carbohydrates. For comparison, Polka et al. [31] reported approximately 3-fold and 2.5-fold lower amounts of protein and fat, respectively, and a 2.4-fold higher amount of ash in dried guelder rose berries.
Polyphenolic antioxidants and ascorbic acid, which are abundant in various fruits, are the most valuable constituents of berries, firstly, as health beneficial compounds and, secondly, as natural additives, which in some cases may be used to replace synthetic preservatives and colors. Therefore, the evaluation of an antioxidant potential of VOP was among the main objectives of this study. The antioxidant capacity of extracts (expressed for DWE) and the recovery of antioxidants from the dry VOP (expressed for DWP) were determined for each assay for this purpose. Both characteristics are important for pomace valorization, the former evaluates antioxidant properties of the product (extract, fraction), while the latter evaluates the efficiency of the extraction process. For instance, a low yield extract may possess stronger antioxidant activity, while in the case of high yields, better recovery of antioxidants from the plant may be achieved, although the extracts may show lower antioxidant capacity due to the dilution of active constituents with the neutral ones. An extraction with ethanol resulted in an approximately three times higher total extract yield comparing to water (Table 1); furthermore, antioxidant capacity values of VOP-E were significantly higher than those of VOP-W in the all assays. Values represented as mean ± standard deviation (n = 6); a, b : the mean values in the rows for Viburnum opulus pomace ethanol (VOP-E) and water (VOP-W) extracts (dry weight of extract (DWE)) and initial material (dry weight of pomace (DWP)) followed by different letters are significantly different (p < 0.05).
In addition, ethanol, due to remarkably higher yield, recovered a 3.4-3.8 times higher amount of polyphenolics from the pomace (DWP). This could be explained by the higher content and wider spectra of various soluble components in ethanol during the first step of PLE (Table 2). Despite the differences in antioxidant capacity values determined by the applied methods, which may be explained by different chemical structures of used radicals for the assay and the peculiarities of reaction mechanisms, the scavenging capacity and TPC of VOP-W were also quite high. Most likely, the increased temperature in PLE with water promotes the recovery of polyphenols from the used residue after extraction with ethanol due to the breakdown of the cell walls and increase of membrane permeability. In addition, heating might soften plant tissue and weaken the interactions between phenols and biopolymers present in the raw material, namely, proteins and polysaccharides. Consequently, the diffusion of phenolics into the solvent would increase.
Some data on antioxidant capacity of guelder rose berry juice and pomace extracts are available from the previously performed studies. Significantly lower TPC values were determined in different juice samples (5.47-10.61 mg GAE/g) while the antioxidant capacity varied from 127.37 to 260.38 and from 31.95 to 109.81 µmol TE/g in the ORAC and ABTS assays, respectively [1]. Zakłos-Szyda et al. [11] produced a polyphenol-rich fraction from the guelder rose berries by solid phase extraction under pressure; it contained 361.52 mg GAE/g compared to 5.98 mg GAE/g in juice. TPC values of methanol and acetone extracts of VOP reported in the aforementioned study agreed with our results. Kraujalis et al. [17] reported TPC in different VOP extracts from 88.6 to 74.9 mg GAE/g DW and it was higher in the water extract. The same tendencies were observed for the antioxidant capacity determined by the DPPH, ABTS, and ORAC methods. Comparing with similar products obtained from sea buckthorn berry pomace [21], guelder rose berry pomace extracts were remarkably stronger antioxidants in the all applied assays and exhibited larger TPC values. It may be also noted that the concentration of secondary metabolites as well as some macronutrients is often larger in the berry pomace compared to the fruit pulp [11,31].

Characterization and Quantification of Phenolic Compounds in VOP Extracts
The identification and quantification of phenolic acids and flavonoids in the extracts was of major interest for valorizing their potential to provide health benefits. It is well-known that the type and the structural peculiarities of phenolics may predestine extract's antioxidant and other activities. The composition of VOP extracts was analyzed by UPLC-QTOF-MS, while the radical scavenging capacity of the separated compounds was additionally monitored by the HPLC-DPPH • -scavenging on-line method. As a result, 45 phytochemical compounds were detected of which 42 were identified and 45 quantified using calibration curves produced with various analytical standards. It may be observed that most of them belong to flavonoids ( Figure 1A,B, Table 2). Combined UV (positive signals) and DPPH • quenching (negative signals) chromatograms of VOP extracts isolated from SFE-CO 2 residue are presented in Figure 1(C: VOP-W; D: VOP-E). In the latter assay, the HPLC-separated antioxidants react post-column with DPPH • , the solution of which is circulating in the coil. The induced bleaching of a radical is detected as a negative peak photometrically at 515 nm. Phenolic acids ( (14,17,29) were found to be active radical scavengers in the investigated extracts. The VOP-E extract was a stronger radical scavenger than the VOP-W extract. The activity of VOP-W was comparable with that of phenolic acids, especially chlorogenic and hydroxybenzoic, and catechin derivatives (13,15,16,22,24), while the activity of VOP-E was considerably higher, most likely due to the presence of other constituents such as iridoids (7,  in VOP-W and VOP-E, respectively, were detected as the main antioxidants responsible for extracts antioxidant capacity. Chlorogenic acid was the strongest radical scavenger, while the input of other constituents, as it may be observed from the negative peaks in the chromatograms, was remarkably lower. Chlorogenic acid and rutin were identified based on retention time and MS of authentic standards. When the standards were not available, the identification was performed by comparing the exact mass of the parent ion (M-H) and typical MS fragmentation pattern with available data in previously published articles and PubChem, Chemspider, and Metlin databases.  Malic (1), citric (2), and quinic (5) acids were identified by using standards. Citric acid was detected only in VOP-W. These acids were reported in V. opulus previously [1,32,33]. Perova et al. [33] in 100 g of eleven tested guelder rose berry varieties collected from different locations determined 578.0-2090, 279.0-1630, and 52.0-346.0 mg of malic, citric, and quinic acids, respectively. It is not surprising-the major part of soluble organic acids is transferred to the juice during pressing. Based on deprotonated ion m/z 371 and a fragment m/z 175, corresponding to the glucuronide moiety, compound 3 was tentatively identified as dihydroferulic acid 4-O-glucuronide. This compound was identified and quantified for the first time as a minor constituent in VOP-W extract. Compound 4 generated the fragment m/z 423 by RDA (retro-Diels-Alder), while other two fragments, m/z 467 and 305, were originated due to HRF (heterocyclic ring fusion) and QM (quinone methide cleavage) fragmentation, respectively. The fragment m/z 289 was formed due to the cleavage of the interflavan C-C bond corresponding to the (epi)catechin. Therefore, this compound was assigned to the procyanidin derivative. A small amount of this compound was detected only in the VOP-W extract.
Three benzoic acid derivatives were detected in VOP extracts. The compounds 6 and 8 gave similar , which coincided with the fragmentation pattern of hydroxybenzoic acid derivatives containing two and one hydroxyl groups, respectively. Structurally, hydroxybenzoic acids may contain as many as four hydroxyl groups surrounding a single benzene ring. Hydroxyl groups in natural compounds are usually attached to the 3, 4, and 5 positions which cannot be exactly determined by MS. Therefore, on the basis of the mass spectra and previously published data, the peaks 6, 8, and 12 were tentatively identified as di-and mono-hydroxybenzoic acid derivatives [34]. All three compounds were detected and quantified in VOP recovered with water, the amount of 12 was eight-fold lower than reported by Velioglu et al. [3]. and CO 2 , respectively. As compared with the MS fragments and literature, peak 39 was plausibly characterized as secologanate [35]. This compound was not reported in V. opulus previously. Peak 7 gave pseudomolecular ion m/z 441 and several fragment ions listed in Table 2. The fragmentation pathway was similar to iridoids via formation of acetic and valeric acid moieties; however, MS data was not sufficient to elucidate the exact structure of this iridoid.
Other compounds, based on the previously reported in V. opulus fruits data and fragmentation patterns, were assigned to viburtinoside and opulus iridoids [32,36]. For example, MS 2 fragments of viburtinoside derivatives (42,43) indicate the sequential neutral losses of two acetic and one (iso)valeric acid groups, while opulus iridoid III (44) had three acetic and one (iso)valeric acid moieties. Opulus III (45) contains additional pentose moiety. Bock et al. [37] and Bujor et al. [32] reported the detailed fragmentation pathways for these compounds. In our study, iridoids constituted one of the major parts of all quantified phytochemicals, and they were recovered only in ethanol. According to the recovered amount, their content in 100 g of dried VOP was from 1.21 ± 0.01 mg (43) to 10.51 ± 0.53 mg (45). These constituents were quantified for the first time in VOP.                          Peak 9 was detected only in VOP-W and exhibited the molecular ion m/z 161 and fragment ions m/z 143 (-18 Da) and 133 (-28 Da), which suggested a dimethyl ester derivative of malic acid. Four derivatives (10, 18, 21, and 25) of chlorogenic acid were detected in VOP (Table 3, Figure 2). An interpretation of their MS 2 spectra using previously developed hierarchical keys enabled to assign these peaks to a particular positional derivative [38][39][40]. Furthermore, it was found that the fragment ion type and ion intensity were in a good correlation with the position of the caffeoyl group. The characteristic fragments m/z 133, 135, 161, 173, 179, and 191 were formed due to the cleavage of an ester bond between caffeoyl and quinic acid groups, quinic acid dehydration, along with caffeoyl decarboxylation. The compounds 10 and 25 shared a similar fragmentation pattern (Table 2)  Peak 9 was detected only in VOP-W and exhibited the molecular ion m/z 161 and fragment ions m/z 143 (-18 Da) and 133 (-28 Da), which suggested a dimethyl ester derivative of malic acid. Four derivatives (10, 18, 21, and 25) of chlorogenic acid were detected in VOP (Table 3, Figure 2). An interpretation of their MS 2 spectra using previously developed hierarchical keys enabled to assign these peaks to a particular positional derivative [38][39][40]. Furthermore, it was found that the fragment ion type and ion intensity were in a good correlation with the position of the caffeoyl group. The characteristic fragments m/z 133, 135, 161, 173, 179, and 191 were formed due to the cleavage of an ester bond between caffeoyl and quinic acid groups, quinic acid dehydration, along with caffeoyl decarboxylation. The compounds 10 and 25 shared a similar fragmentation pattern (Table 2) The absence or low intensity of m/z 179 ([caffeoyl−H] − ) enabled to separate them from 3-O-caffeoylquinic acid and to identify as 1-and 5-caffeoylquinic acids [40]. Thus, compound 10 was identified as 1-caffeoylquinic acid, while the absence of m/z 179 for peak 25 let us to identify it as neochlorogenic acid. Compound 18 was identified as 3-O-caffeoylquinic acid (chlorogenic acid); it yielded the base peak m/z 191 (deprotonated quinic acid) and also gave an ion m/z 179 [M-caffeic acid-H]with half of the intensity of the base peak. Moreover, it was confirmed by the analytical standard. Compound 21 gave characteristic ion m/z 173 [M-quinic acid-H-H 2 O]and was identified as 4-O-caffeoylquinic acid (cryptochlorogenic acid) according to the previously described fragmentation pattern [39,40]. Chlorogenic acid, as the main phenolic compound in fresh berries of V. opulus, has been reported previously [1,3,33]. Chlorogenic acid and its derivatives are among the major constituents in VOP extracts as well. The content of these phenolic acids recovered from 100 g of VOP varied from 0.49 ± 0.00 to 214.8 ± 0.02 mg. Chlorogenic acid (18) was dominant in VOP-W, while 1-caffeoylquinic acid (10) was detected and quantified only in VOP-W. For comparison, the content of chlorogenic acid was similar to the previously reported in guelder rose berry juice by Velioglu et al. [3]; however, 1.16-2.7 times lower than reported by Perova et al. [33] in guelder rose fruits.
MS data for compounds 11 and 26 were not sufficient for their identification. Despite that, these compounds were quantified. Moreover, spectral information of these compounds are presented in the Supplementary Materials; the content of 11 was 52-fold higher in VOP-E, than in VOP-W, while  [41]. The cleavage of the interflavan C-C bond produced m/z 289 corresponding to the catechin (epicatechin) unit, while further loss of 44 amu (C 2 H 4 O) in the benzofuran skeleton yielded m/z 245. Consequently, the extract had some procyanidin dimers, which have different retention times due to different bonding of catechin and epicatechin moieties. In addition, compound 13 was confirmed by the authentic standard as proanthocyanidin B2. The content of recovered procyanidin dimers varied from 0.48 ± 0.00 to 8.89 ± 0.01 mg/100 g DWP. The recovered content of 13 and 22 was similar by both solvents, while the compounds 20 and 34 were found only in VOP-E with an approximately five-fold higher content of 20 than 34. Procyanidin dimers were reported previously without quantification [1,32,41]; the concentration of procyanidin dimer determined in guelder rose berry juice [3] was similar to its content in VOP determined in our study.
The  [42]. To best of our knowledge, this compound was not previously reported in VOP (Table 3). This compound was identified and quantified only in VOP-E, which recovered 1.96 ± 0.11 mg/100 g DWP. Finally, loss of the B ring (110 amu) from the parent ion yielded an ion m/z 341. The recovered content of 19 was similarly recovered by both solvents, while the content of compound 15 was eight-fold higher in VOP-W. These compounds were reported previously without their quantification [32,41].
The compounds 16 and 24 eluting at 2.0 and 3.4 min, respectively, displayed the same molecular ion m/z 289 (C 15 H 13 O 6 ). Furthermore, they gave the same characteristic MS 2 fragments m/z 245 (loss of CO 2 ), 205, and 203 (cleavage of the A-ring of flavan-3-ol). The identity of compound 16 was finally proved using the catechin analytical standard. The hydrophobicity of epicatechin is higher than catechin; therefore, 24 eluted later than 16 and was assigned to epicatechin [32,43]. The contents of catechin and epicatechin in V. opulus berry juice were 29.04 and 2.69 mg/100 g, respectively. Comparing to Velioglu et al.'s [3] study, ethanol recovered higher contents of catechin, while water produced lower amounts. Recovered epicatechin values were three-to six-fold higher than reported by Velioglu et al. [3].
The compounds 17 and 23 yielded molecular ion m/z 579, but exactly suggested two different molecular formulas: C 30 H 28 O 12 and C 26 H 28 O 15 , respectively. Compound 17 was tentatively identified as gambiriin based on the characteristic MS 2 fragmentation pattern consisting of base m/z 289, which fragmented into m/z 245 and 205 after RDA and QM [44]. To the best of our knowledge, this compound was identified and quantified for the first time in VOP. Compound 23 generated several fragments in negative ionization mode (Table 2); however, this information was not sufficient for its tentative identification. Furthermore, spectral information of this compound are presented in the Supplementary Materials. The latter compound was detected and quantified only in VOP-W, while the calculated concentration of gambiriin was approximately two times higher in VOP-E than in VOP-W.
The precursor ion m/z 865 was detected for compound 27. The fragment m/z 739 was caused by HRF reaction of the upper unit, while m/z 713 occurred due to a RDA reaction of (epi)catechin unit, while m/z 695 indicates the loss of H 2 O. The fragment m/z 577 corresponds to the (epi)catechin dimer and originates from a QM reaction. Furthermore, the fragment ion of the dimer was observed following the RDA-based reaction (m/z 425), which yielded m/z 407 owing to the subsequent loss of H 2 O. The HRF reaction was also observed for the dimer, generating m/z 451. Finally, a QM reaction with the dimer formed at single quinone, resulting in m/z 289 and 287 ions. Therefore, this compound was tentatively assigned to procyanidin C1 [1,32,41]. This compound was quantified only in VOP-E with the recovery value of 2.20 mg/100 g DWP.
Five flavalignans in the form of flavanols substituted with phenylpropanoids were found in VOP (Table 3 = a, b). The structures of these compounds together with the fragmentation pathways were proposed by Hokkanen et al. [45]. Purification and NMR would be required for more precise identification of compounds 28, 31, 36, 37, and 41. Cinchonains were previously found in lingonberries and bilberries [45], while their presence in V. opulus has not been reported. Flavalignans constituted another large group of phytochemicals quantified in VOP; their concentration varied from 0.10 ± 0.00 to 59.84 ± 1.41 mg/100 g DWP. Quantitatively, compound 36 was dominant in both extracts with higher value in VOP-E. The compounds 28 and 31 were quantified only in VOP-E, while the content of 37 and 41 was higher in VOP-E and VOP-W, respectively.
Quercetin-3-O-glucoside (32) and rutin (33) were unambiguously identified by the obtained MS data and by using available standards. These compounds were quantified in VOP-E. Both of them were reported previously [3,11,32,33,41]. The amount of 33 quantified in guelder rose berry juice by Velioglu et al. [3] was approximately three-fold lower compared to our study, while the content of 32 determined by Zakłos-Szyda et al. [11] varied from 0.01 to 1.06 mg/g of extract depending on the type of sample preparation.
Compound 29 present in VOP-E gave m/z 319 ion (C 15 H 12 O 8 ) and, based on the detailed fragmentation pattern reported by Fan et al. [46], was tentatively identified as dihydromyricetin. This compound has not been reported in V. opulus previously. The compounds 30 and 38 eluting at 6.5 and 8.9 min with a similar pseudomolecular ion [M-H]m/z 595 were assigned to quercetin pentoside hexoside. The identity is supported by the characteristic fragment ions, especially the diagnostic ion m/z 301.0356 of deprotonated quercetin aglycon, the loss of hexose (162 Da) and pentose (132 Da) [32,41]. Compound 38 was quantified only in VOP-E, while the amount of 30 was similar in both extracts.
Compound 35 detected in the VOP-E produced m/z 381 (C 18 H 21 O 9 ); based on MS, it was tentatively identified as ethylchlorogenate. This compound was identified and quantified for the first time in VOP. The most intense transition used for quantification of the compound 40 was the fragment m/z 739, which yielded m/z 289 due to the well-known quinone methide cleavage (QM). Fragments ions m/z 577, 451, 339, and 177 were formed due to the loss of hexose (162 Da), heterocyclic ring fusion (-287 Da) of ring C of upper unit, the loss of ring B from the upper unit after the 739 ion QM cleavage, and finally, the loss of hexose (162 Da) from the m/z 339. Based on this and reported data, compound 40 was tentatively identified as (epi)catechin dimer monoglycoside [41,47]; it was detected and quantified only in VOP-E.
The most important soluble sugars were quantified by using authentic standards: their amounts in 100 g DWP varied from 0.87 ± 0.12 to 2.98 ± 0.51 g of glucose, from 1.09 ± 0.11 to 5.48 ± 1.0 g of fructose, and from 1.17 ± 0.54 to 2.69 ± 0.50 g of sucrose, in VOP-E and VOP-W, respectively. Approximately 4.5-6-fold higher levels of quantified sugars were determined in VOP-W. Perova et al. [33] reported 2.3-4.0 g/100 g fructose and 2.8-4.6 mg/100 g glucose in 11 varieties of guelder rose berries collected from different locations; while sucrose was determined only in three samples in the range of 0.04-0.14 mg/100 g.
Our study demonstrated that ethanol effectively extracts phenolic compounds; however, PLE with water revealed that residual amounts of various phytochemicals, particularly higher polarity hydrophilic compounds, still remain in the residue after PLE with ethanol. On the other hand, it should be noted that high PLE temperature may increase the risk of decomposition of some thermolabile components, while some of them may undergo partial hydrolysis. In general, the results showed that VOP is rich in a large diversity of polyphenols, most of which were shown to possess strong antioxidant activity. It is known, that hydroxyl position in the molecule and some other structural properties determine antioxidant properties of flavonoids; in general, they depend on the ability to donate hydrogen or electron to a free radical. In addition, the variations in the content of reported compounds could also occur due to genotype, origin, harvesting time, and juice processing parameters.

Antiproliferative Activity and Cytotoxicity Effects of VOP Extracts
In order to evaluate the antiproliferative activity of VOP extracts, HT29 cells were used at the exponential grow phase, while cytotoxicity assessment used Caco-2 cell line as the best accepted intestinal model. When confluent, Caco-2 cells share some characteristics with crypt enterocytes and, therefore, have been a widely implemented cell model to assess the effect of chemicals, food compounds, and nano/microparticles on the intestinal function [49]. Thus, HT-29 and Caco-2 were selected as convenient cell models to evaluate the anticancer effects of VOP extracts.
To the best of our knowledge, the antiproliferative activity and cytotoxicity of VOP extracts in this study are reported for the first time. Both extracts were diluted in the solvents at their maximum acceptable concentration levels; VOP-E strongly inhibited cancer cell growth with an EC 50 value of 0.39 ± 0.03 mg/mL ( Figure 3A), without adverse effects on normal epithelia once no cytotoxicity was observed on Caco-2 ( Figure 3B). VOP-W did not exert an antiproliferative effect at 2 mg/mL concentrations, while higher doses were not applied. The antiproliferative effect of VOP may be attributed to the presence of active phytochemicals and/or their interactions (Table 2). Procyanidins, catechin, epicatechin, hydroxybenzoic acids, anthocyanins, quercetin derivatives, as well as chlorogenic acid and its derivatives strongly suppressed cancer cell growth [11]. VOP-E contained higher amounts of gambiriin (17), cinchonains (36,37), and anthocyanins. phenolic compounds, which were not found in VOP-W, e.g., cinchonains (28,31), iridoids (7, 39, 42-45), (epi)catechin derivatives (20,27,34,40), quercetin derivatives (32,33,38), scopoletin-7-Osophoroside (14), and dihydromyrecitin [50] (29) and ethylchlorogenate (35). The bioactivities of most of these phytochemicals have not been reported previously. The strong anticancer compound scopoletin [51] is an aglycone of scopoletin-7-O-sophoroside (14), which might exert anticancer activity after hydrolysis. Moreover, the information on V. opulus iridoids and their bioactivities is rather scarce. Some iridoids demonstrated remarkable effects against various cancer cells [52]. It is also worth mentioning that the compounds showing other bioactivities such as antimicrobial and antiviral usually are promising candidates for anticancer agents. For instance, Ming et al. [53] reported antibacterial, antifungal, antiviral, and hepatoprotective properties of cinchonains. In general, VOP-E showed an antiproliferative effect against HT-29 cells in vitro and could be a promising substance for expanding its testing using experimental animal models.

Cellular Antioxidant Activity (CAA) of VOP Extracts
In contrast to antiproliferative and cytoprotective properties, both VOP extracts showed strong and similar antioxidant activity with CAA values of 27.23 ± 9.01 and 30.36 ± 4.82 μmol QE/mg DWE for VOP-W and VOP-E, respectively (Figure 3 C). It is interesting noting that despite no antiproliferative activity of VOP-W in HT29 cells its antioxidant activity in Caco-2 cells was similar. It may be assumed that the antiproliferative effects of antioxidant phytochemicals of VOP, due to structural peculiarities, may be achieved by different mechanisms, which are not directly related to and cytotoxicity (B) effects using HT29 and Caco-2 cell lines, respectively. Results are expressed in terms of mean ± SD performed in triplicate. Antioxidant activity of VOP extracts evaluated by the cellular antioxidant activity (CAA) assay (C). Results are expressed as µmol QE per mg of dry extract of mean ± SD performed in triplicate. Unpaired t test was assessed with p < 0.05.

Cellular Antioxidant Activity (CAA) of VOP Extracts
In contrast to antiproliferative and cytoprotective properties, both VOP extracts showed strong and similar antioxidant activity with CAA values of 27.23 ± 9.01 and 30.36 ± 4.82 µmol QE/mg DWE for VOP-W and VOP-E, respectively ( Figure 3C). It is interesting noting that despite no antiproliferative activity of VOP-W in HT29 cells its antioxidant activity in Caco-2 cells was similar. It may be assumed that the antiproliferative effects of antioxidant phytochemicals of VOP, due to structural peculiarities, may be achieved by different mechanisms, which are not directly related to their antioxidant capacity measured by the CAA assay. For example, phenolic hydroxyl groups at ortho or para positions with respect to each other increase the antioxidant activity due to additional resonance stabilization and o-quinone or p-quinone formation [54]. Two ortho hydroxyl groups contribute more to the radical-scavenging capacity than do the two hydroxyl groups in meta position. For example, protocatechuic acid has two hydroxyl groups in the ortho position and possesses higher antioxidant activity than γ-resorcylic acid with two hydroxyls in meta position [25]. Furthermore, higher amounts of procyanidin dimers (13,22), (epi)catechin derivatives (15,19,41), catechin (16), epicatechin (24), chlorogenic (18), and cryptochlorogenic acids (25) had influence for high cellular antioxidant activity of the VOP-W extract. Additionally, the extracts demonstrated similar activity in CAA, while VOP-E showed stronger antioxidant capacity in non-cellular chemical-based assays. The CAA assay using a reduced DCFH-DA probe takes into account the bioavailability, distribution, and metabolism of antioxidants within the cell and reflects the activity of the tested extracts in the biological system [55]. It could be related to the structure of bioactive compounds; for instance, flavonoids with 2,3-double bond and 4-oxo group such as quercetin usually possessed high CAA activity [55]. Zakłos-Szyda and co-authors [56] evaluated chemoprotective activity of V. opulus fruit juice, acetone and methanol extracts of pomace, and polyphenol-rich fraction in Caco-2 cells using a strong pro-oxidant and in vitro oxidative stress inducer tert-butylhydroperoxide (t-BOOH). The extracts and especially the polyphenol-rich fraction strongly protected the cells from oxidative damage. Therefore, it was suggested that an antioxidant activity of extracts depends on the composition of polyphenolic compounds such as catechin, epicatechin, anthocyanins, procyanidin B1 and B2, neo and chlorogenic acid, rutin other quercetin derivatives, and hydroxycinnamic acids. Pre-incubation with guelder rose fruit extract at 0.365 mg/mL also protected βTC3 cells against oxidative damage induced by t-BOOH [57]. Consequently, the extracts of previously studied V. opulus fruits and in our study investigated pomace contain valuable bioactive compounds with cytoprotective properties.

Conclusions
Forty-two bioactive phytochemicals were characterized in defatted guelder rose berry pomace extracts recovered by consecutive extraction with pressurized ethanol and water. Dihydroferulic acid 4-O-glucuronide, scopoletin-7-O-sophoroside, gambiriin, five flavalignans, dihydromyrecitin, and ethylchlorogenate, as well as antiproliferative and antioxidant activity of V. opulus berry pomace (VOP) extracts are reported for the first time. Ethanol extract showed higher antiproliferative activity with no cytotoxicity at the applied doses. Both extracts showed strong antioxidant activity in Caco-2 cells protecting them upon stress stimuli. The differences in the antioxidant power and antiproliferative activity of extracts may be related to a phytochemical composition and structural peculiarities of the compounds, particularly to the distribution of hydroxyl groups in the molecules. Considering strong antiproliferative and antioxidant properties, guelder rose berry pomace extracts, especially isolated by ethanol, may find promising applications in functional foods and nutraceuticals. However, technological upscaling and economic issues should be further evaluated in order to assess the possibilities of industrial production of ingredients from V. opulus berry pomace.