Natural extracts from fresh and oven‐dried winemaking by‐products as valuable source of antioxidant compounds

Abstract Winemaking by‐products are a natural source of antioxidant components; however, due to their highly perishable and seasonal nature, they may require a prior conservation step before being processed. Natural extracts from fresh and oven‐dried red grape agro‐industrial by‐products were obtained by ultrasound assisted extraction (UAE), using a hydroalcoholic solution as extracting solvent. Extracts were analyzed by HPLC‐DAD‐ESI‐MS, to know the feasibility of winemaking by‐products as natural sources of phenolic compounds, as well as the effect of the oven‐drying treatment on the phenolic composition. Oven‐drying at 45°C caused a significant decrease on the total phenolic content, which implied a reduction of the antioxidant capacity of the extracts. Also, it produced a decrease in total and individual flavan‐3‐ols, stilbenes, and flavonols, being greater in those extracts from stems. Respect to anthocyanins, which were only identified in grape pomace extracts, oven‐drying caused an important decrease, being the peonidin‐3‐O‐glucoside the more thermosensitive compound. Natural extracts from fresh or oven‐dried winemaking by‐products could be used in other food industries as a valuable source of phenolic compounds with antioxidant properties. However, further studies on other drying methods are required for addressing the preservation of phenolic compounds from winery by‐products successfully.

Other important winery by-products are the grape stems, which are removed in the early stages of the winemaking process, and are not subjected to maceration phenomena. For this reason, the stems can retain their phenolic composition almost intact, being an important natural source of stilbenes, which are compounds with significant bioactive properties (Barros et al., 2014;Makris, Boskou, & Andrikopoulos, 2007;Ruiz-Moreno et al., 2015;Spatafora, Barbagallo, Amico, & Tringali, 2013).
To obtain phenolic-rich extracts, conventional extraction methods with organic solvents are commonly used. These extraction techniques are highly polluting, labor-intensive, and time-consuming, so the development of green alternative methods is required. The ultrasound assisted extraction (UAE) could be one of the most suitable techniques, as it is very simple to use, requires relatively inexpensive apparatus, reduces the volume of solvent, and allows to use green extraction solvents with a high efficiency due to the mechanical effects that it produces in the cellular structure of the matrix (Barba, Zhu, Koubaa, Sant′Ana, & Orlien, 2016).
The UAE has been used for the recovery of different bioactive components from fresh winemaking by-products (Drosou, Kyriakopoulou, Bimpilas, Tsimogiannis, & Krokida, 2015;González-Centeno, Comas-Serra, Femenia, Rosselló, & Simal, 2015;Piñeiro, Guerrero, Fernández-Marin, Cantos-Villar, & Palma, 2013), although in most cases a comprehensive study of the remaining polyphenolic composition of extracts has not been carried out. On the other hand, it is important to highlight the highly perishable and seasonal nature of winemaking by-products. The drying treatment could be an effective method to increase their shelf life by slowing the development of microorganisms and preventing biochemical reactions that may modify the phenolic composition of those by-products.
Therefore, the objectives of this study were (1) the assessment of the feasibility of red winemaking by-products (grape pomace, seed, and stems) from Vitis vinifera L. Cv. Tempranillo, as valuable and natural sources of polyphenols and (2) study the effect of the oven-drying on the phenolic composition of the extracts. For these purposes, a green technology extraction such as ultrasonic-assisted extraction with hydroalcoholic mixtures as extracting solvent was used. Phenolic-rich extracts were characterized by HPLC-DAD-ESI-MS, and chemical data were co-related with the antioxidant capacity of the extracts.

| Fresh and oven-dried samples
Among the raw materials studied were included grape pomace, seeds, and stems from Vitis vinifera L. Cv. Tempranillo. These winery by-products were obtained from the Institute of Vine and Wine of Castilla-La Mancha (IVICAM, Tomelloso, Ciudad Real, Spain). Samples were ground with addition of dry ice and cooling jacket at 0 ° C, using a crusher Stephan UMC5 (Stephan Food Service Equipment GMBH).
This step is essential to get homogeneous samples with the same particle size (diameter less than 2 mm). Then, they were divided into two batches. One of them was conserved in a freezer at −20°C until their processing in fresh, and the remaining batch was immediately oven-dried at 45°C during 17 hr for grape pomace, and during 19 hr for seeds and stems. The drying conditions were selected after trials had been conducted to achieve a percentage moisture content of <10% using the lowest temperature and shortest possible time.
Oven-dried samples were stored at room temperature until further processing and analysis.
The initial moisture content of the fresh samples was 64.3%, 53.6%, and 66.6% for grape pomace, seeds, and stems, respectively.
While the moisture content of the dried samples was 6.0%, 5.0%, and 9.7%. The moisture content was determined (in duplicate) using a laboratory oven at 105°C.

| Ultrasound assisted extraction
Ultrasound assisted extraction (UAE) was carried out in a QSONICA sonicator Q500 (53 CHURCG HILL RD. Newtown, CT, USA). To this end, 5 g of crushed sample was mixed with 20 ml of an hydroalcoholic solvent (44% of ethanol) in a beaker, operating at 20 KHz frequency, and 81% of output amplitude, with a duty cycle of 15-s turn on and 5-s off for an extraction period of 3 min. The beaker was introduced into a cooling bath at 4°C, and the extraction temperature was maintained at 20°C. Two extraction cycles were performed, and the mixtures were centrifuged at 7000 rpm for 5 min after extraction. The supernatants were collected and filtered under vacuum.
These extracts were adjusted to the same volume (50 ml) and stored at −20°C in the dark until their analysis. Extractions from fresh and oven-dried by-products were performed in duplicate. UAE conditions were previously optimized by Response Surface Methodology (RSM). The optimization was realized with crushed oven-dried grape pomace Vitis vinifera L. Cv. Tempranillo.

| RSM experimental design
RSM was used to optimize the extraction conditions of total phenolic compounds (TPC) from oven-dried grape pomace. A Central Composite Design was used to study the effect of two independent variables, ethanol content in the hydroalcoholic solvent (25-50%, v / v), and output amplitude (60-90%) on the response variable selected (TPC). The design consisted of 11 points, four (2 2 ) factorial points, four axial points (2 × 2), and three repetitions at the central point for experimental error. The range and central point values of the two independent variables used are summarized in Table 1. To minimize the effects of unexpected variability in the responses, the experiments were run at random.
Optimal extraction conditions were estimated by multiple linear regression (MLR) using Statgraphics Centurion XVI Version 16.1.17 (Statpoint Technologies, Warrenton, VA, USA). The residual standard deviation (RSD) and the coefficient of determination (R 2 ) were used to estimate the accuracy of the model.

| Total phenolic content
To determinate the TPC of the extracts, the Folin-Ciocalteu method was followed (Singleton & Rossi, 1965). One hundred microliters (100 μl) of extract was mixed with 7.9 ml of Milli-Q water and 500 μl of Folin-Ciocalteu reagent. Then, 1.5 ml of sodium carbonate solution (5 g/L) in Milli-Q water was added to the tube and the solution was shaken. The tubes were kept at room temperature, during 2 hr, and the absorbance was measured at 765 nm in a spectrophotometer (Helios, Thermo Spectronic, Cambridge, UK), versus a blank prepared without extract. All of the analyses were carried out in duplicate. The results were expressed as milligram of gallic acid equivalents per gram of dry weight sample (mg GAE/g DW), using gallic acid as standard.

| DPPH assay
The DPPH assay was carried out using 1,1-diphenyl-2-picrylhydrazyl as a stable radical (Brand-Williams, Cuvelier, & Berset, 1995). One hundred microliters of different dilutions of extracts was mixed with 2.9 ml of a methanol DPPH radical solution (6.10 −5 mM). Then, the decrease in absorbance was measured every minute for 25 min, at 515 nm, in a spectrophotometer (Helios, Thermo Spectronic). The zero was adjusted with methanol. A calibration curve (0.1-0.8 mM Trolox) was used to calculate the antioxidant activity of the extracts, and the results were expressed in mM Trolox equivalents per gram of dry weight sample (mM/mg DW).

| ABTS assay
The ABTS assay was carried out according to the method of Re et al. (1999). First, a 7 mM ABTS solution was mixed with a 2.45 mM potassium persulfate solution, to prepare the ABTS radical cation (ABTS˙+). The mixture was kept in the dark for 12-16 hr before use, and then, it was diluted with ethanol (1:90 v/v) to achieve an absorbance value of 0.7 (±0.02). Once ABTS˙+ working solution was obtained, 2 ml was mixed with 20 μl of diluted extracts. In the case of the blank, 20 μl of ethanol was used. Absorbance was measured in a spectrophotometer (Helios, Thermo Spectronic) at 734 nm, at 0 and 5 min of reaction. Readings at both times of reaction were used to calculate the percentage inhibition value for each extract. A calibration curve (0.1-0.8 mM Trolox) was used to calculate the antioxidant activity of the extracts, and the results were expressed in mM Trolox equivalents per gram of dry weight sample (mM/mg DW).

| HPLC-DAD-ESI-MSn analysis
2.6.1 | Flavan-3-ols, mean degree of polymerization (mDP), % galloylation, % prodelphinidines, and stilbenes The first step on the analysis of flavan-3-ols (monomers, dimers, and polymeric proanthocyanidins) and stilbenes was a purification of the extracts using SPE on C18 cartridges (Sep-pak Plus C18, Waters Corp., Milford, MA, USA; cartridges filled with 820 mg of adsorbent). After conditioning the SPE cartridge with 5 ml methanol and 5 ml water, 2 ml of each extract diluted with 12 ml of water was passed through them. Then, the cartridge was dried under reduced pressure and 15 ml methanol and 5 ml ethyl acetate were added sequentially to elute the flavan-3-ols. After the solvent was evaporated in a rotary evaporator (35°C), the residue was dissolved in 2-ml methanol and stored at −18°C until its analysis.
Analysis of flavan-3-ol monomers, dimers, proanthocyanidins oligomers (condensed tannins), and stilbenes was carried out according to the previously described methodology ( The mean degree of polymerization (mDP), % galloylation, and % prodelphinidines was calculated according to the procedures previously described by Kennedy and Jones (2001).

| Flavonols and hydroxycinnamic acids
To avoid the interferences that anthocyanins may cause in the chromatographic separation and identification of other phenolic compounds, those were separated from the extracts. Polymeric cation exchange SPE cartridges (PCX 500 mg, 6 ml; Bond Elut Plexa, Agilent) were used according to the method previously described by Castillo-Muñoz, Gómez-Alonso, García-Romero, and Hermosín- Extracts from stems and seeds have not anthocyanins, and thus, the separation of these compounds was not required. In this case, 2 ml of these extracts was taken to dryness in a miVac DUO concentrator operating at 35°C (Genevac Ltd., Ipswich, UK) and resolved in 1.5 ml of a methanol/water solution (20:80). Then, these were filtered (polyester membrane, 0.20 μm, Chromafil PET 20/25, Machery-Nagel, Düren, Germany).
HPLC analysis of flavonols and hydroxycinnamic acid derivatives was performed on an Agilent 1100 Series system (Agilent, Germany). The system was equipped with a diode array detector (DAD) and a LC/MSD Trap VL electrospray ionization mass spectrometry (ESI-MS/MS) detector connected in series. Extracts were injected into a reversed-phase column Zorbax Eclipse XDB-C18 (2.1 × 150 mm; 3.5 μm particle; Agilent), with a precolumn Zorbax Eclipse XDB-C8 (2.1 × 12.50 mm; 5 μm particle; Agilent), themostatized at 40°C. Injection volume was 20 μl, flow rate was 0.16 ml/ min, and the separation was achieved using a ternary mobile phase ESI-MS/MS was used for identification, and DAD chromatograms at 360 nm and 320 nm were used for quantification of flavonols and hydroxycinnamic acids derivatives, respectively. Flavonol concentrations were expressed as equivalents of quercetin-3-glucoside (Extrasynthese, Genay, France) (μg/g of dry weight sample), while hydroxycinnamic acids were not found in the samples.

| Anthocyanins and derived compounds
HPLC analysis of anthocyanins was carried out on an Agilent 1100 Series system (Agilent, Germany), according to the methodology previously described by Rebello et al. (2013).

| Optimization of the extraction conditions
Tempranillo, and the TPC was used as a measure of the efficiency of the UAE. First, the amount of sample and the volume of extracting solvent were selected according to the probe tip diameter (12 mm) and to the operation manual of the QSONICA Q500. In order to be able to work with the smallest volume of extracting solvent (20 ml), a sample quantity of 5 g was selected. This sample quantity allowed to obtain a low mass/volume ratio and consequently an increase in the recovery of total polyphenols, without losing sensitivity in the detection of the individual extracted phenolic compounds.
Once the amount of sample was fixed, two extraction times ( Table 4 shows the individual and total concentrations (μg/g dry weight sample) of flavan-3-ols, mean degree of polymerization, % of galloylation, and % of prodelphinidines in the fresh and oven-dried red winemaking by-product extracts. The total amount of flavan-3-ols ranged from 696.9 to 3429.0 μg/g for extracts from oven-dried stems and fresh seeds, respectively.

| Phenolic profile of the extracts. Effect of the oven-drying
Fresh seed extracts present the major quantities of total flavan-3-ols, following by fresh grape pomace and stem extracts. The total quantity of flavan-3-ols decreased considerably during oven-drying. The highest losses of flavan-3-ols occurred in the stems, while seeds were the winemaking by-products less affected by oven-drying.
Respect to oligomers, procyanidin B1 was the most abundant flavan-3-ol in the extracts of fresh stems (1222.5 μg/g), which was in good agreement with those observations previously realized by González-Centeno et al. (2012). However, its content decreased considerably during oven-drying. On the other hand, procyanindin B2 was the main dimer in seed and grape pomace extracts. This compound was described as the main procyanidin dimer in seeds and grape pomace from different grape varieties (González-Paramás et al., 2004).
Seed and grape pomace extracts were richer in galloylated flavanols and prodelphinidines than extracts of stems, although the quantity of flavanols of grape pomace is totally conditioned by its proportion in seeds. It has been reported in bibliography that galloylated flavanols have an antioxidant activity higher than their nongalloylated homologues (Plumb, de Pascual-Teresa, Santos-Buelga, Cheynier, & Williamson, 1998).
Another group of important phenolic compounds are stilbenes.
Their presence in plants is related with resistance to certain fungi such as Botrytis cinerea, or to other causes of stress, such as ultraviolet irradiation (Langcake & Pryce, 1977). These compounds have also gained significant attention because of their high antioxidant activity (Aziz, Kumar, & Ahmad, 2003). Table 5 displays the individual and total concentrations (μg/g dry weight sample) of stilbenes in the fresh and oven-dried red winemaking by-product extracts.
In all extracts, only the 3-glycosylated form of resveratrol was found, with the trans isomer as majority. Extracts from fresh stems had the greatest amount of total stilbenes, highlighting their high quantity in trans-resveratrol-glucoside. Previous studies had revealed the higher content of resveratrol in stems than in seeds or skins (Cho et al., 2003). Stilbenes store at high levels in grape leaves and stems, occurring at significantly lower quantities in grapes (Langcake, 1981;Langcake & Pryce, 1977). Oven-drying treatment produced a decrease in stilbenes in all the samples, being more important in the stem extracts.
Flavonols, another important group of phenolic compounds, were only detected and quantified in extracts of fresh and ovendried grape pomace and stems. Individual and total concentrations (μg/g dry weight sample) of flavonols are shown in Table 6. Extracts TA B L E 5 Individual and total concentrations (μg/g dry weight sample) of stilbenes in natural extracts from fresh and oven-dried red winemaking by-products (n = 2) As expected, the anthocyanins were only identified in the extracts of fresh and oven-dried grape pomace. Table 7

| Correlation between antioxidant capacity and phenolic compounds of the extracts
A correlation analysis, between antioxidant values and phenolic composition, was performed in order to determine the possible contribution of the individual phenolic compounds in the antioxidant capacity of the extracts. Table 8 shows those phenolic compounds with stronger correlations.
Positive correlations were observed between antioxidant activity and some flavan-3-ols, as (+)-catechin (C), (+)-gallocatechin (GC), In the extracts tested, the correlation of monomers with the antioxidant activity was as follows: CG > ECg > GC > C > EGC This correlation order was in good agreement with the evidences previously reported by Guo et al. (1999), andNanjo et al. (1996). These authors pointed out that the presence of a gallate group in position 3′ and the occurrence of hydroxyl groupin position 5′of ring B play an important role to quench free radicals.
Dimers seemed to also play an important role in the antioxidant activity of extracts. The correlation between the procyanidin B1 content and antioxidant capacity had already been highlighted in grape seed extracts (Guendez, Kallithraka, Makris, & Kefalasa, 2005). The high values of antioxidant activity of procyanidin B1 have been attributed to the number of available hydroxyl groups (Alonso et al., 2002).
Another compound correlated to the antioxidant capacity of the extracts was trans-resveratrol-glucoside, although this compound was found at lower concentrations than flavan-3-ols.
This compound has demonstrated DPPH scavenging capacity in other studies (Wei, Zhao, Li, & Xue, 2016), while its isomer, cisresveratrol-glucoside showed less efficiency in the mechanism of transfer of an atom of hydrogen to the free radicals (Mikulski, Górniak, & Molski, 2010).

| CON CLUS IONS
Ultrasound assisted extraction allows to obtain natural phenolic-rich extracts from winemaking by-products which could be used in other food industries, as a valuable source of antioxidant components.
The optimal conditions for maximum extraction of total polyphenols were 44% of ethanol in the hydroalcoholic solvent and an output amplitude of 81%.
However, due to their highly perishable and seasonal nature, winemaking by-products require a prior conservation step before being processed. In this regard, oven-drying at 45°C caused a significant decrease on the total and individual phenolic components, which implied a reduction in the antioxidant capacity of the extracts. The changes observed during the heat treatment depended on the intrinsic characteristics of the raw materials and their phenolic composition. Further studies on other alternative drying methods would be of great value for the recovery of polyphenols from the winemaking by-products.

CO N FLI C T O F I NTE R E S T
The authors declare that they do not have any conflict of interest.

E TH I C A L S TATEM ENTS
This study does not involve any human or animal testing, and an informed consent was obtained from all study contributors.