Characterization of Polyphenolics in Grape Pomace Extracts Using ESI Q-TOF MS/MS

Red and white grape pomace mainly containing grape skins, seeds as well as some stems were generously provided by Woodward canyon winery (Lowden, WA). The red pomace was generated from Cabernet Sauvignon and the white was from Chardonnay. Both red and white grape pomace were freeze-dried and ground to 40-60 mesh powders. GSE (GravinolSuper^TM) was purchased from OptiPureChemco Industries Inc. (Los Angeles, CA).

Ethanol, catechin, rutin, gallic acid, formic acid, glacial acetic acid, vanillin, aluminum chloride, DPPH• (2,2-diphenyl-1-picrylhydrazyl) radical, and Folin-Ciocalteu’s reagent were purchased from Sigma-Aldrich Co. (St. Louis, MO, USA). Sodium carbonate and sodium acetate were from JT Baker (Center Valley, PA, USA).

After testing the efficiency of extraction by using different extraction durations (5 to 30 min) and temperatures (40 and 50?C), with and without acidified solvents, in combination with or without ultrasonic treatment (Ultrasonic cleaner XSPS-180-6L, SharperTek, Pontiac, MI, USA), we selected 80% EtOH-1% Formic Acid (FA) in combination with ultrasonic extraction for 15 min at 40?C for extracting polyphenols from grape pomace. Briefly, pomace powders were first treated with hexane to remove non-polar compounds such as lipids at room temperature. 80% EtOH-1% FA was added into dried hexane extracted grape pomace powder at 10:1 ratio of solvent to sample (volume/weight). The sample was vortexed and incubated at room temperature for one hour, followed by 15 min ultrasonic treatment (40?C), then centrifuged at 12,000 rpm for 15 min. The supernatants containing extracted polyphenols were collected. The residues were re-extracted twice. Supernatants from three extractions were combined and kept at -20?C till analyses.

All analyses were performed in 96-well microplates using Synergy H1 Hybrid microplate reader (BioTek Instruments Inc., Winooski, VE, USA). All reactions were conducted at room temperature, and incubation time for specific reaction was determined by its kinetics with the desired wavelength.

Total phenolic content: Total phenolic content was determined using the modified Folin-Ciocalteu procedure [10]. Gallic acid (1.0 to 50.0 μg/ml) was used to generate the standard curve. In brief, 200 µL of diluted GPEs and standard solutions were added into each well, followed by 12.5 µL of Folin-Ciocalteu’s reagent and then 37.5 µL of 20% Na₂CO₃. The absorbance at 760 nm was read after 2 hour incubation. The results were expressed as milligrams of gallic acid equivalent per gram of dried pomace weight (mg GAE/g DW).

Total flavonoid content: The modified AlCl₃-acetate method [11] was used to measure total flavonoid content. Briefly, 50 µL of diluted GPEs and standard solutions were mixed with 150 µL of 5% AlCl₃ and 50 µL of acetate buffer (pH 5.0) sequentially. The absorbance at 420 nm was measured after one hour incubation at room temperature. Rutin was used as the standard with a range of 4.0 to 148.0 μg/ml and total flavonoid content was expressed as milligrams of rutin equivalent per gram of dried pomace weight (mg RE/g DW).

Total PAC content: Total PAC content was analyzed by the modified Vanillin-HCl procedure [12]. In brief, 50 µL of diluted GPEs and standard solutions were mixed with 150 µL of 4% vanillin and 50 µL of concentrated hydrochloride. The absorbance was measured at 500 nm after incubation for 30 min. The total PAC content was determined using catechin as a standard ranging between 5.0 and 250.0 μg/ml. Data were expressed as milligrams of catechin equivalent per gram of dried pomace weight (mg CE/g DW).

DPPH radical scavenging assay: The total antioxidant activity in GPEs was determined by the ability to scavenge DPPH• [13]. 200 µL of DPPH• solution (6 x 10^-5 M) were added into microplate wells containing 50 µL of diluted GPEs or standards. The DPPH•scavenging activity was measured at 515 nm after 90 min incubation at room temperature. Data were expressed as mg CE/g DW that was calculated by the standard curve of catechin in the range of 0.3 to 6.0 μg/mL. The percent inhibition of DPPH• radical scavenging activity was calculated by the following equation: inhibition (%) = X 100, where A is the absorbance.

The mass spectra were collected via the direct infusion on a Waters ESI Q-TOF Premier (Waters, USA) with electrospray ionization source equipped with MassLynxv4.1. Both positive and negative ion ESI mode MS/MS analyses were performed under the following conditions: the capillary voltage, +3.5 KV/-3.0 KV [ESI+/ESI-]; the source temperature, 110°C; the sample cone, 30V/40V [ESI+/ESI-]; the desolvation (L/hr), 300/350 [ESI+/ESI-]; the scan range, 90-2190 amu; the scan rate, 1 sec/scan. Samples were diluted in 25% methanol with 0.5% formic acid and directly infused into the electrospray source at the flow-rate of 3 µL/min. The m/z number of precursor ion marked with * indicates that precursor ion generated from ESI+ mode.

The relative percentages of (epi)cat and PACs were calculated from their respective peak intensities divided by total peak intensities of (epi)cat and PACs including galloylated PACs.

All data were presented as mean with their corresponding standard deviations from three independent experiments. The student’s t-test was used to identify difference. Differences at p ≤ 0.05 were considered significant.

The extraction yield (dried weight of crude extract/dried weight of pomace x 100) for the dried red and white grape pomaces was 30.65 ± 0.95% and 54.76 ± 0.18%, respectively. The extraction yield of white grape pomace is much higher than previous reported [14,15] while the red grape pomace is similar to others [16].

The total phenolic, flavonoid and PAC contents are presented in table 1. The total phenolic and flavonoid contents were higher in RGPE than those in WGPE. The total flavonoid contents in RGPE and WGPE accounted for about 63% and 25% of total phenolic content, respectively. The total phenolic level in both RGPE and WGPE was higher than the value reported previously using the same extraction solvent (30.4 for Cabernet Franc and 24.5 for Chardonnay) [17]. Our data are consistent with the results obtained from Tinta Cao (red) and Chardonnay (white) [18], Cabernet sauvignon, Pinot Noir and Merlot (red) [16], and white grapes cultivated in Turkey [19]. Furthermore, the total phenolic content in WGPE was largely consistent with the content identified in four white grape cultivars (30.9-46.5 mg GAE/g DW) [14]. The difference in phenolic contents by different reports is likely due to the variation in grape cultivars, climate and culture conditions, as well as extraction methods [17,18]. The total PAC content is higher in RGPE than that in WGPE (Table 1), in agreement with a previous report [14].

	RGPE	WGPE
Total phenolics (mg GAE/g DW)	69.83 ± 4.53	58.15 ± 5.21*
Total flavonoids (mg RE/g DW)	43.89 ± 1.22	14.32 ± 1.67*
Total Proanthocyanidins (mg CE/g DW)	133.79 ± 6.74	92.10 ± 6.00*
Antioxidant activity (mg CE/g DW)	74.48 ± 1.12	58.66 ± 1.92*
DPPH• inhibition (%)	68.28 ± 0.52	62.74 ± 1.34*

RGPE: Red Grape Pomace Extract; WGPE: White Grape Pomace Extract; Total content of phenolic, flavonoids and PAC (Proanthocyanidins) is expressed respectively as mg of Gallic Acid Equivalent (GAE), mg of Rutin Equivalent (RE) and mg of Catechin Equivalent (CE) per gram of Dried pomace Weight (DW). Total antioxidant activity is determined with the DPPH radical scavenging activity and expressed as mg of Catechin Equivalent (CE) per gram of dried weight. Data are means of three independent experiments. Data were presented as Mean ± SEM; ^*: P< 0.05

In agreement with their higher content of total phenolics, flavonoids and PACs, the RGPE had higher DPPH• free radical scavenging capacity and greater antioxidant activity than WGPE (Table 1).

Table 2 lists the phenolic compounds putatively identified by direct infusion tandem MS in both negative and positive modes. Figure 1 shows the direct infusion ESI Q-TOF mass spectra of RGPE, WGPE and GSE in both positive and negative ion modes. Inserts in figure 2A, are enlarged spectra in negative mode showing overlapped isotope patterns of PAC dimers to hexamers and doubly charged tetramers and monogalloylated heptamers and nonamers containing A-and B-types. The PACs were further examined with their fragments processing through main fragmentation patterns of Retro-Diels-Alder (RDA) fission, Heterocyclic Ring Forming fission (HRF), and Quinonemethide (QM) fission as demonstrated in figure 3, as well as Benzofuran Forming (BFF) fission [5,20,21].

 

GSE: Grape Seed Extract; PAC: Proanthocyanidin; RGPE: Red Grape Pomace Extract; WGPE: White Grape Pomace Extract; Precursor ion marked with * means [M+H]⁺, otherwise stands for [M-H]^-; +: means this compound was detected with reasonable intensity

 

At ESI-, gallic acid ([M-H]^- ion at m/z 169) was presented in all grape extracts and was confirmed by its MS/MS fragment at m/z 125. The [M-H]^- ions of m/z 133,149 and 191 were detected in both RGPE and WGPE while the [M-H]^- at m/z 179 was presented only in WGPE (Table 2); they are malic acid, tartaric acid, citric acid and caffeic acid, respectively (Figure 1 A,B) [22,23].

The [M-H]^- ion at m/z 301 were detected only in RGPE, which likely is quercetin. Its fragment ions mainly are m/z 273 [M-H-28(CO)]^-, 257 [M-H-44(CO2)]^-, 229 [M-H-44(CO2)-28(CO)]^- (Table 2), which was similar to a previous report [24].

At ESI-, the [M-H]^- ions at m/z 463 and 477 were observed in RGPE, WGPE and GSE, which might be quercetin 3-glucoside and quercetin 3-glucuronide with the fragment at m/z 300 (loss of a glucosyl unit) and 301 (loss of a glucuronate group), respectively (Table 2). Both of them were reported previously in grape skin at ESI- [25].

ESI+ signals attributable to anthocyanins were observed in grape pomace extracts. The [M+H]⁺ ion at m/z 479*(stands precursor ions generated from ESI+ mode) was detected in both RGPE and WGPE (Figure 1 C,D), which was assigned to petunidin-3-glucoside confirmed by its fragments at m/z 303 and 317 (Table 2) [26]. The [M+H]⁺ ions at m/z 463*, 493*, 505*, 535*, 625*, 639*, and 655* could be assigned to peonidin 3-glucoside, malvidin 3-glucoside, peonidin 3-acetylglucoside, malvidin 3-acetylglucoside, petunidin 3-p-coumaroylglucoside, and malvidin 3-p-coumaroylglucoside,and malvidin 3-(6-O-caffeoyl) monoglucoside (Figure 1 C,D) confirmed by their fragments of 301 (loss of 162 Da, a glucosyl unit), 331 (loss of 162 Da), 301 (loss of 204 Da, an acetyl glucosyl unit), 331 (loss of 204 Da), 317 (loss of 308 Da, a coumaroylglucosyl unit), 331 (loss of 308 Da), and 331 (lose of 224 Da, a caffeoylglucosyl unit) (Table 2), respectively [26-28]. Of which, peonidin 3-glucoside and peonidin 3-acetylglucoside were detected in both RGPE and WGPE, while malvidin derivatives and petunidin 3-p-coumaroylgluside only found in RGPE (Table 2).

The [M-H]^- ion at m/z 491 could be assigned to malvidin 3-glucoside with the fragment at m/z 329 (loss of a glucosyl unit), which was only detected in RGPE and WGPE and was reported previously in grape skin as well [25].

Catechin and epicatechin were found in all grape extracts at both ESI+ and ESI- with the precursor ion at m/z 291* and 289, respectively (Figure 1) and backed by their characteristic fragmentations (Table 2) mainly via loss of one water for both of them, RDA (loss of 152 Da), HRF (loss of 126 Da) and BFF for the m/z 291* precursor ion, and loss of a -CH₂-CHOH group or CO₂, loss of C₄H₄O₂ from the A-ring and C₆H₆O₂ from B-ring for the m/z 289 to generate corresponding fragments [20,29,30]. 

 

Monogalloylated B-type (epi)cat oligomers: The [M-H]^- ions at m/z 729, 1017 and 1305 could be assigned to monogalloylated B-type dimers, trimers and tetramers with 2,3 and 4 possible structures, respectively. They were all detected in RGPE, WGPE and GSE (Figure 1 A,B and 2A) and their main fragments are listed in table 2. The fragments of [M-H]^- ions at m/z 729 and 1017 have been characterized previously via loss of 152 Da (RDA or galloyl group), loss of 126 Da (HFR at the top unit), loss of water, and QM (upper and lower unit after loss of galloyl group) [7,31-33]; the fragments of [M-H]^- at m/z 1305 have not been described, which processes the similar fragmentaion patterns of [M-H]^- ions at m/z 729 and 1017. The [M-H]^- ions assignable to monogalloylated pentamers (m/z 1593) and hexamers (m/z 1881) were also detected in RGPE and WGPE (data not shown).

Under ESI+ the monogalloylated B-type dimers at m/z 731* were detected in RGPE, WGPE and GSE; while the monogalloylated B-type trimers (m/z 1019*) and tetramers (m/z 1307*)were presented only in RGPE and GSE (Table 2). They all have similar fragmentation pattern as those at ESI-.

A-type PACs: A-type PACs were previously reported in other foods such as peanut skins, hops, and raspberry. However, they have been barelyreported in grapes and their products [6,34], which were characterized in this study.

The observed [M-H]^- ions at m/z 575, 863, 1151, 1439, and 1727 revealed a series of compounds with a mass difference of 288 Da that can be attributed to A-type PAC dimers, trimers, tetramers, pentamers and hexamers, respectively. They displayed 2 amu difference from the corresponding B-type PACs at m/z 577, 865, 1153, 1441 and 1729 [5,7,24,31,35,36]. Further, the observed [M+H]⁺ ions for A-type PAC dimers to hexamers at m/z 577*, 865*, 1153*, 1441* and 1729* also present 2 amu difference from the corresponding B-types at m/z 579*, 867*, 1155*, 1443* and 1731* [6,20,21] (Figure 1,2 and Table 2).

Figure 3, showed the fragmentation patterns for three selected precursor ions at ESI- mode. Figure 3A and 3B, showed the fragment pathways of A-type (m/z 575) and B-type (m/z 577) dimers. Their characteristic fragmentations are mainly via HRF (loss of 126 Da), RDA (loss of 152 Da) and QM cleavages at the top and bottom units (Table 2), which are consistent with previous reports at ESI- for A-type [5,24,37-39] and for B-type [5,7,29,33,35,39]. At ESI+, the main fragments for both A-type (m/z 577*) and B-type (m/z 579*) dimers (Table 2) are also similar to previous reports [20,21].

RDA: Retro-Diels-Alder fission; QM: Quinone-Methide fission; HFR: Heterocyclic Ring Fission; [M-H-152]^-in C: this fragment comes from two possible ways: 1) loss of one galloyl group, 2) via RDA if the terminal unit is (epi)cat. RDA¹ means loss of 152 Da from precursor ions (A,B) or the fragment of m/z 289 (C); RDA²: this fragment generates from the fragment of m/z 559 (loss one galloyl group) through RDA if the terminal unit is or to become (epi)afz (lose 136 Da) (C); RDA³: this is from the fragment of m/z 559 (loss one galloyl group) through RDA if the terminal unit is or to become (epi)cat (lose 152 Da) (C), HRF, [M-H-126]-; QM, lose one (epi)catechin; “x 4” means the fragment of m/z 559 is zoomed in by 4 times

One monogalloylated A-type PAC dimer with the [M-H]^- at m/z 711 was observed in WGPE and GSE, which gave the MS/MS fragments at m/z 693 ([M-H-18]^-, loss of one water), 559 ([M-H-152]^-, loss of a galloyl group or via RDA if the terminal unit is (epi)cat), 423 (QM of m/z 711 or from the fragment ion at m/z 559 via RDA if the terminal unit is or to be (epi)afz), 407 (from the ion at m/z 559 via RDA if the terminal unit is or to be (epi)cat), 289 (QM cleavage from terminal unit of (epi)cat while the top unit is (epi)afz after loss of galloyl group and 285 (QM cleavage from terminal unit of (epi)afz while the top unit is (epi)cat after loss of galloyl group), and 137 (RDA of fragment at m/z 289) (Figure 3C and Table 2). It could be (epi)catG-A-(epi)afz, (epi)afzG-A-(epi)cat, (epi)cat-A-(epi)afzG, or (epi)afz-A-(epi)catG [(epi)afz G, (epi)afzelechin 3-O-gallate].

A-type dimers and B-type PAC dimers to hexamers were detected in all grape extracts under both ESI- and ESI+. In addition, single A-type linked PAC trimers and tetramers can be detected in GSE, RGPE and WGPE under ESI- (Table 2). Otherwise, all A-type PACs described below only detected in GSE under ESI+ and/or ESI- (Table 2 and Figure 2 A,B).

Under ESI- mode, precursor ions at m/z 863 and 861 could be assigned to PAC trimers with 1 and 2 A-type linkages, respectively (insert in Figure 2A). Their main fragments are aligned with previous reports [24,35,36,40]. Trimers with one A-type linkage could be assigned as (epi)cat-A-(epi)cat-(epi)cat or (epi)cat-(epi)cat-A-(epi)cat depending on the fragments: (epi)cat-A-(epi)cat-(epi)cat has fragement ions at m/z 573 and 289 via QM cleavages between middle and terminal units while (epi)cat-(epi)cat-A-(epi)cat has m/z 575 and 287 fragments via QM cleavages between top and middle units. In addition, both of them generate fragments at m/z 737 (loss of 126 Da through HRF) and 711 (loss 152 Da by RDA). Under ESI+ mode,the precursorions at m/z 865* and 863* are PAC trimers with 1 and 2 A-type linkages. Based on their fragments in table 2, the [M+H]⁺ at m/z 865* might be (epi)cat-(epi)cat-A-(epi)cat [6].

PAC tetramers, pentamers and hexamers with 1 to 3 A-type linkages were detected under ESI- with the corresponding precusor ions at m/z 1151, 1149 and 1147; 1439, 1437 and 1435; 1727, 1725 and 1723 (insert in Figure 2A). The main fragments of [M-H]^- ions at m/z 1147, 1149, 1435, 1437 and 1725 are listed in table 2, which are generally produced from [M-H-152]^- (RDA), [M-H-(288)n]- (progressively loss (epi)cat units) or loss water molecules. A portion of these precursor ions and possible isomers of tetramers and pentamers with one and two A-type linkages are mentioned previously in other foods such as peanuts and cranberry [5,24,35,36].

Under ESI+ mode, tetramers, pentamers and hexamers with 1 and 2 A-type linkages were also detected in GSE with precursor ions at m/z 1153* and 1151*; 1441* and 1439*; 1729* and 1727*, respectively (Table 2). In addition, hexamers with three A-type linkages were detected in GSE at m/z 1725* under ESI+. Overall, with increasing the degree of polymerization, the detected amount of A-type PACs decreased.

Doubly charged A-type PACs: Doubly charged ([M-2H]^2-) A-type PACs were detected only in GSE (inserts in Figure 2). Under ESI-, [M-2H]^2- PAC pentamers with 1-3 A-type linkages (m/z 1439, 1437 and 1435, respectively) were occurred at m/z 719, 718 and 717, respectively (insert in Figure 2A). [M-2H]^2-heptamers with 1-3 A-type linkages (m/z 2015, 2013 and 2011, respectively) were detected at m/z 1007, 1006 and 1005, respectively under ESI- mode (insert in Figure 2A). Under ESI+, the double charged heptamers with 1-3 A-type linkages were at m/z 1009*, 1008* and 1007* (insert in Figure 2B). [M-2H]^2-PAC nonamers with 1-4 A-type linkages (m/z 2591, 2589, 2587, and 2585, respectively) were also detected respectively at m/z 1295, 1294, 1293 and 1292 (insert in Figure 2A). Up to now, there was no detailed report about doubly charged precursor ions in grapes, especially doubly charged A-type PACs, though doubly charged A-type PAC tetramers (m/z 1149) and pentamers (m/z 1439) were recently reported in dry-blanched peanut skins [24].

Some singly charged precursor ions overlaped with the doubly charged ones in some cases along with some unkown precusor ions with high intensities such as at m/z 313, 325, 359, and 439 in grape pomaces warrant future characterization.

Relative content of PACs analyzed by ESI Q-TOF MS: The relative content of monomeric and polymeric (epi)catechins were different under different ionization mode (Table 3). The percentage of monomeric (epi)catechin calculated from ESI+ mode was much higher than that from ESI- mode, but the relative content of oligomers from ESI- was generally higher than that from ESI+ (Table 3). Overall, oligomers were the major PACs in all grape extracts, dominant by dimers and trimers.