Abstract

This study aimed to optimize the extraction solvents for Bordo grape (Vitis labrusca) seeds by response surface methodology regarding to the antioxidant activity (AA) and trans-resveratrol content. Fatty acids (FA) and AA of the oil obtained by pressing cold method were also determined. The extraction optimization was determined by the statistical simplex-centroid mixing scheme, enabling the analysis of solvents effects (water, ethanol and acetone) and their mixtures on the responses. AA was performed by DPPH, ABTS and FRAP methods, and by total phenolic compounds and flavonoids. FA were determined by GC and trans-resveratrol by HPLC. The extracts containing ternary fraction of solvents showed greater AA, increasing about 20 times compared to pure solvent. The composition that showed the best response ranged between 45-48% of water, 14-20% of ethanol and 35-38% of acetone for both grape seeds, with polarity ranged from 0.705 to 0.706. The oil from seeds showed high concentrations of PUFA, particularly linoleic acid. The optimized extraction method improved the use of this residue as a potential antioxidant source for food industry.

Keywords:
industrial by-products; solvent extraction; mixture modeling; phenolic compounds; Trans-resveratrol; fatty acids

1 Introduction

Grapes are highlighted as a worldwide consumed fruit and also present a high quantity of phenolic compounds in natura as in its derivatives (Spigno et al., 2007Spigno, G., Tramelli, L., & De Faveri, D. M. (2007). Effects of extraction time, temperature and solvent on concentration and antioxidant activity of grape marc phenolics. Journal of Food Engineering, 81(1), 200-208. http://dx.doi.org/10.1016/j.jfoodeng.2006.10.021.
http://dx.doi.org/10.1016/j.jfoodeng.200... ). Such phenolic compounds have antioxidant potential and anti-inflammatory, antimicrobial and anticarcinogenic action (Teixeira et al., 2014Teixeira, A., Baenas, N., Dominguez-Perles, R., Barros, A., Rosa, E., Moreno, D., & Garcia-Viguera, C. (2014). Natural bioactive compounds from winery by-products as health promoters: a review. International Journal of Molecular Sciences, 15(9), 15638-15678. http://dx.doi.org/10.3390/ijms150915638. PMid:25192288.
http://dx.doi.org/10.3390/ijms150915638... ).

The grape industry generates a large number of residues as peel, seeds and pulp, which also contain antioxidants (Santos et al., 2011Santos, L. P., Morais, D. R., Souza, N. E., Cottica, S. M., Boroski, M., & Visentainer, J. V. (2011). Phenolic compounds and fatty acids in different parts of Vitis labrusca and V. vinifera grapes. Food Research International, 44(5), 1414-1418. http://dx.doi.org/10.1016/j.foodres.2011.02.022.
http://dx.doi.org/10.1016/j.foodres.2011... ). This has aroused the interest of the industry itself in recycling, in order to reduce costs, environmental impacts, and add value to the product (Spigno & De Faveri, 2007Spigno, G., & De Faveri, D. M. (2007). Antioxidants from grape stalks and marc: Influence of extraction procedure on yield, purity and antioxidant power of the extracts. Journal of Food Engineering, 78(3), 793-801. http://dx.doi.org/10.1016/j.jfoodeng.2005.11.020.
http://dx.doi.org/10.1016/j.jfoodeng.200... ). Seeds of vinification were the residues that started the exploitation of the phenolic compounds of this class, followed by bagasse (Melo, 2010Melo, P. S. (2010). Composição química e atividade biológica de resíduos agroindustriais (Dissertação de mestrado). Universidade de São Paulo, Piracicaba. https://doi.org/10.11606/D.11.2010.tde-21102010-161908.
https://doi.org/10.11606/D.11.2010.tde-2... ). By this way, considering the diversity of polyphenols sources from plant matrices and their physicochemical structures, as well as its properties (such as antioxidant activity), it is hard to find a universal extraction method that can be applied in all samples (Alara et al., 2018Alara, O. R., Abdurahman, N. H., & Olalere, O. A. (2018). Optimization of microwave-assisted extraction of flavonoids and antioxidants from Vernonia amygdalina leaf using response surface methodology. Food and Bioproducts Processing, 107, 36-48. http://dx.doi.org/10.1016/j.fbp.2017.10.007.
http://dx.doi.org/10.1016/j.fbp.2017.10.... ). Aspects such as time, temperature, particle size and type of extraction are crucial because they alter the yield of the extracted antioxidant compounds (Alcântara et al., 2019Alcântara, M. A., Polari, I. L. B., Meireles, B. R. L. A., Lima, A. E. A., Silva, J. C. Jr., Vieira, É. A., Santos, N. A., & Cordeiro, A. M. T. M. (2019). Effect of the solvent composition on the profile of phenolic compounds extracted from chia seeds. Food Chemistry, 275, 489-496. http://dx.doi.org/10.1016/j.foodchem.2018.09.133. PMid:30724224.
http://dx.doi.org/10.1016/j.foodchem.201... ). In addition, the extractor solvent polarity also affects the antioxidants extracted (Bosso et al., 2016Bosso, A., Guaita, M., & Petrozziello, M. (2016). Influence of solvents on the composition of condensed tannins in grape pomace seed extracts. Food Chemistry, 207, 162-169. http://dx.doi.org/10.1016/j.foodchem.2016.03.084. PMid:27080893.
http://dx.doi.org/10.1016/j.foodchem.201... ). In this regard, the extraction optimization for each sample plays a key role because it also enhances the global economic process (Mukherjee et al., 2014Mukherjee, S., Mandal, N., Dey, A., & Mondal, B. (2014). An approach towards optimization of the extraction of polyphenolic antioxidants from ginger (Zingiber officinale). Journal of Food Science and Technology, 51(11), 3301-3308. http://dx.doi.org/10.1007/s13197-012-0848-z. PMid:26396324.
http://dx.doi.org/10.1007/s13197-012-084... ). In this manner, the development of grape seeds extraction method can add commercial and industrial values for grape culture.

By this way, this study seeks to optimize the extraction process by statistical simplex-centroid mixing scheme, assessing the obtained extracts, as well as, the oil extracted from seeds by cold pressing in terms of antioxidant activity, resveratrol content and fatty acids.

2 Materials and methods

2.1 Reagents

Folin-Ciocalteau reagent, gallic acid, quercetin, 2,2-diphenyl-1-picrylhydrazyl (DPPH), 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox), 2,4,6-tripyridyl-1,3,5-triazine (TPTZ), 2,2-Azinobis(3-ethylbenzothiazoline-6-sulphonic acid) diammonium salt (ABTS), standard nonadecanoic acid methyl ester (19:0) and trans-resveratrol internal standard were from Sigma-Aldrich. All other reagents were of analytical grade.

2.2 Samples

Unfermented Bordo seeds were obtained from an industry in the city of Toledo, Paraná State, Brazil, before fermentation process. Fermented seeds were obtained at a winery located in Quatro Pontes municipality, Paraná State, Brazil, after fermentation process. The seeds were separated from the peels and peduncles, washed, sun-dried, sieved, packed in vacuum, and stored at room temperature.

2.3 Experimental design and grape seed extracts preparation

In order to study the influence of different solvents on the extraction of antioxidants, an experimental simplex centroid design was planned, using the response surface methodology for modeling mixtures in Statistica software, version 10.0 (StatSoft, 2011StatSoft. (2011). STATISTICA (data analysis software system) (10.0). Retrieved from www.statsoft.com), with a 0.05 significance level. Table 1 presents the coded levels, where the independent variable is the proportion of solvents (water, ethanol, and acetone), and dependent variables is the antioxidant activity obtained by several methods.

Seeds were ground into powder using a knife mill (Solab - PS 30), in mesh 20 sieves. About 10 g of the samples were weighted, added to 100 mL of the extraction solvent, and kept 4 hours in an orbital shaker at 250 rpm, 45 °C, and protected from light. The extracts were filtered and placed in a rotary evaporator at 45 °C under reduced pressure. Then, they were freeze-dried for 5 days and stored in a dark room at -18 °C.

2.4 Oil extraction using a hydraulic press

To extract the oil a stainless steel template, composed of a chamber for seed storage, a piston, and a hydraulic press (EMIC, 2000 kN), was used. The load applied by the downward piston press was received by the other piston that crushed the material, extracting the oil collected through a door. A 50 stainless steel mesh was used at the bottom of the template to obtain an oil free of impurities. It was used 100 grams of dried grape seed (fermented and unfermented), at 85 tonnes of pressure and 300 seconds as extraction time.

2.5 Physicochemical analyzes

Moisture (012/IV), ash (018/IV), lipids (032/IV), and proteins contents (037/IV) were performed according to the methodology described by Instituto Adolfo Lutz (2008)Instituto Adolfo Lutz – IAL. (2008). Métodos físicos-quimicos para análise de alimentos. São Paulo: IAL.. Total nitrogen was converted into protein by the specific factor for conversion of 6.25.

2.6 Antioxidant activity

Capture of free radicals by DPPH method

The methodology employed was described by Bondet et al. (1997)Bondet, V., Brand-Williams, W., & Berset, C. (1997). Kinetics and mechanisms of antioxidant activity using the DPPH• free radical method. Lebensmittel-Wissenschaft + Technologie, 30(6), 609-615. http://dx.doi.org/10.1006/fstl.1997.0240.
http://dx.doi.org/10.1006/fstl.1997.0240... , with modifications. Methanolic solutions (40 µL) of grape seeds (oil and extracts 2.5 mg mL^-1) were pipetted and 3.0 mL of methanolic DPPH (2,2-diphenyl-1-picrylhydrazyl) solution (0.0842 mmol L^-1) were added. After 30 minutes in the dark, the absorbance was read in a UV-VIS spectrophotometer (PG Instruments Ltda, Model T 80+) at 517 nm. The blank used was methanol and Trolox (0-2500 µmol L^-1) was used for calibration curve. Results were expressed in μmol Trolox equivalent (TE) g^-1 of extract or oil (R² = 0.9997).

Capture of free radicals by ABTS•⁺ method

The antioxidant activity by ABTS (2,2-Azinobis(3-ethylbenzothiazoline-6-sulphonic acid) diammonium salt) method followed the methodology described by Re et al. (1999)Re, R., Pellegrini, N., Proteggente, A., Pannala, A., Yang, M., & Rice-Evans, C. (1999). Antioxidant activity applying an improved ABTS radical cation decolorization assay. Free Radical Biology & Medicine, 26(9-10), 1231-1237. http://dx.doi.org/10.1016/S0891-5849(98)00315-3. PMid:10381194.
http://dx.doi.org/10.1016/S0891-5849(98)... . First, the ABTS^•+ radical from a mixture of ABTS at 7 mmol L^-1 with 140 mmol L^-1 of potassium persulfate, was maintained in the dark at room temperature for 16 hours. After the incubation period, the solution was diluted with ethanol (HPLC grade) until obtaining an absorbance of 0.70 (± 0.01). In the absence of light, a sample of 30 µL was pipetted (extracts and oil 2.5 mg mL^-1) and 3.0 mL of ABTS^•+ radical solution was added. After 6 minutes, the absorbance was read at 734 nm in a UV-VIS spectrophotometer. Trolox (0-2000 μmol L^-1) was used for calibration curve and the results were expressed in μmol Trolox equivalent (TE) g^-1 of extract or sample (R² = 0.9995).

Power reduction of Fe (III) by FRAP method

The reduction power were determined by the FRAP method and evaluated according to the methodology (adapted) described by Benzie & Strain (1996)Benzie, I. F. F., & Strain, J. J. (1996). The Ferric Reducing Ability of Plasma (FRAP) as a measure of “antioxidant Power”: the FRAP assay. Analytical Biochemistry, 239(1), 70-76. http://dx.doi.org/10.1006/abio.1996.0292. PMid:8660627.
http://dx.doi.org/10.1006/abio.1996.0292... . Solutions were prepared with 3 mL of FRAP reagent, preheated at 37 °C, 300 μL of distilled water, and 100 μL of sample. The resulting solution was homogenized and incubated in a water bath at 37 °C for 60 minutes. Absorbance was read at 593 nm in a spectrophotometer and results were expressed in μmol of ferrous sulfate equivalent per gram of extract or oil (μmol EQFeSO₄^.7H₂O g^-1), by the calibration curve (0-2000 μmol L^-1) of standard FeSO₄^.7H₂O (R² = 0.9990).

Evaluation of total phenolic compounds by Folin-Ciocalteau (FC) method

The method described by Singleton & Rossi (1965)Singleton, V. L., & Rossi, J. A. (1965). Colorimetry of total phenolics with phosphomolybdic-phosphotungstic acid reagents. American Journal of Enology and Viticulture, 16(3), 144-158. to determine the concentration of total polyphenols were employed. The methanolic solutions from the seeds (extracts and oil 2.5 mg mL^-1) were prepared, 250 µL were pipetted, 250 µL of the Folin-Ciocalteu reagent (diluted at 1:1 in distilled water) was added, 500 µL of the saturated solution of Na₂CO₃ and 4.0 mL of distilled water). After 25 minutes in the dark and at room temperature, it was centrifuged for 10 minutes at 3000 rpm and the absorbance read at 725 nm. The results were determined with gallic acid (0-200 mg L^-1) as standard and expressed in mg of gallic acid equivalent (GAE) g^-1 of extract or oil (R² = 0.9997).

Flavonoid content (FLA)

The procedure adopted to determine the flavonoids content was adapted from Woisky & Salatino (1998)Woisky, R. G., & Salatino, A. (1998). Analysis of propolis: some parameters and procedures for chemical quality control. Journal of Apicultural Research, 37(2), 99-105. http://dx.doi.org/10.1080/00218839.1998.11100961.
http://dx.doi.org/10.1080/00218839.1998.... . The methanolic solutions from grape seeds were prepared (extract and oil 2.5 mg L^-1), then 500 μL were transferred to test tubes, 250 μL of aluminum chloride 5% (m/v in methanol) and 4.25 mL of methanol were added. After 30 minutes in the dark and at room temperature, the absorbance was read at 425 nm. Quercetin was used as standard (0-100 mg L^-1) and results was expressed as mg quercetin equivalents (EQ) g^-1 of extract or oil (R² = 0.9976).

2.7 Trans-resveratrol

The methodology for separation and quantification of trans-resveratrol was adapted from Souto et al. (2001)Souto, A. A., Carneiro, M. C., Seferin, M., Senna, M. J. H., Conz, A., & Gobbi, K. (2001). Determination of trans -resveratrol concentrations in brazilian red wines by HPLC. Journal of Food Composition and Analysis, 14(4), 441-445. http://dx.doi.org/10.1006/jfca.2000.0970.
http://dx.doi.org/10.1006/jfca.2000.0970... . About 18 mg of sample were diluted in 1 mL of methanol and injected into an HPLC (Dionex UltiMate 3000) system with a UV-VIS detector, at 50 °C. With isocratic elution, C18 column (250 mm × 4,6 mm) with a particle diameter of 5 μm (Agilent Eclipse XDB), the mobile phase consisted of acetonitrile and water (25%:75%), 3.0 of pH (adjusted with orthophosphoric acid), and an injection volume of 20.0 µL. The injection flow rate was 1.5 mL min^-1, with the analyte detected at 306 nm after a 6-minute run. Trans-resveratrol in methanol (0.1-10.0 mg L^-1) was used as standard and results were expressed in mg L^-1 of trans-resveratrol (R=0.9944).

2.8 Fatty acids

The procedure of Hartman & Lago (1973)Hartman, L., & Lago, R. (1973). Rapid preparation of fatty acid methyl esters from lipids. Laboratory Practice, 22(6), 475-476. PMid:4727126., with adaptations from Maia & Rodriguez-Amaya (1993)Maia, E. L., & Rodriguez-Amaya, D. B. (1993). Avaliação de um método simples e econômico para a metílação de ácidos graxos com lipídios de diversas espécies de peixes. Revista do Instituto Adolfo Lutz, 53(2), 27-35., was used to determine the fatty acids content in the seeds oils. The methylation reaction was carried out with sodium hydroxide solution in methanol, followed by esterification with acid catalysis (ammonium chloride, methanol, and sulfuric acid). The separation of the fatty acids methyl esters (FAME) was carried out by a Perkin Elmer automatic gas chromatograph, Clarus 680 model, coupled with a flame ionization detector (GC-FID) and a capillary column of fused silica Select Fame CP 7420 (100 m long, 0.25 mm inner diameter, and 0.25 μm of coating film). The gases flows were 1.1 mL min^-1 for the carrier gas (H₂), 40 mL min^-1 to H₂, and 400 mL min^-1 to the synthetic air flame. The injector and the detector remained at 240 °C and 275 °C, respectively. The FAME separation column used had a programmed temperature of 80 °C for 1 minute, followed by heating on a temperature ramp of 15 °C min^-1 until 180 °C. Then one heated to 220 °C with a 3 °C min^-1 ramp, remaining for 2 minutes, and finally a 5 °C min^-1 ramp until 250 °C for 5 minutes. The samples injections were performed with a volume of 2 μL using a 5 μL syringe (Split 1:100). Given the determined peak areas, FA could be identified by comparing with retention times of standard FAME, comparing with individual patterns and added patterns. The FAME quantification was performed in relation to the internal standard nonadecanoic acid methyl ester (19:0) (1.026 mg mL^-1 in n-heptane). The results were expressed in mg of FA per gram of total lipids (Equation 1) (Visentainer & Franco, 2012Visentainer, J. V., & Franco, M. R. B. (2012). Ácidos graxos em óleos e gorduras - identificação e quantificação (2. ed., Vol. 1). Maringá: Eduem. Retrieved from https://www.saraiva.com.br/acidos-graxos-em-oleos-e-gorduras-identificacao-e-quantificacao-1563259/p
https://www.saraiva.com.br/acidos-graxos... ).

where: C_X = concentration of the fatty acid x in mg g^-1 of total lipids; A_X = area of methyl esters corresponding to the fatty acid x; A_19:0 = internal standard area; M_19:0 = internal standard mass added to the sample (mg); M_A = mass of total lipids (g); F_TC = theoretical correction factor; F_CEA = conversion factor from FAME to FA.

2.9 Statistical analysis

All analyses were performed in triplicate and the results expressed as mean ± standard deviation. The results containing two averages were submitted to test-T (5% of probability), and the results with three or more averages were submitted to the analysis of variance (ANOVA), followed by a comparison of the averages by Tukey test (5% of probability) performed by StatSoft (2011)StatSoft. (2011). STATISTICA (data analysis software system) (10.0). Retrieved from www.statsoft.com.

3 Results and discussion

3.1 Physicochemical analyzes

Table 2 shows the physicochemical composition values of fermented and unfermented seeds of Bordô grape.

By comparing the experimental data obtained for the seeds, it is possible to verify that the averages were alike, featuring a physicochemical composition very similar for both. The differences reported between the values obtained in the physicochemical compositions analyzes can be attached to different forms of treatment, of cultivars, agroclimatic factors, and the winery management (Cheng et al., 2012Cheng, V. J., Bekhit, A. E.-D. A., McConnell, M., Mros, S., & Zhao, J. (2012). Effect of extraction solvent, waste fraction and grape variety on the antimicrobial and antioxidant activities of extracts from wine residue from cool climate. Food Chemistry, 134(1), 474-482. http://dx.doi.org/10.1016/j.foodchem.2012.02.103.
http://dx.doi.org/10.1016/j.foodchem.201... ).

Santos et al. (2011)Santos, L. P., Morais, D. R., Souza, N. E., Cottica, S. M., Boroski, M., & Visentainer, J. V. (2011). Phenolic compounds and fatty acids in different parts of Vitis labrusca and V. vinifera grapes. Food Research International, 44(5), 1414-1418. http://dx.doi.org/10.1016/j.foodres.2011.02.022.
http://dx.doi.org/10.1016/j.foodres.2011... reported in their study similar moisture values of four grape seeds varieties (from 8.87 up to 10.50%). Regarding proteins, these same authors obtained smaller values (from 6.5 up to 7.7%) analyzing those samples. Santos et al. (2011)Santos, L. P., Morais, D. R., Souza, N. E., Cottica, S. M., Boroski, M., & Visentainer, J. V. (2011). Phenolic compounds and fatty acids in different parts of Vitis labrusca and V. vinifera grapes. Food Research International, 44(5), 1414-1418. http://dx.doi.org/10.1016/j.foodres.2011.02.022.
http://dx.doi.org/10.1016/j.foodres.2011... also observed lower protein content in Izabel, Niagara, Benitaka and Brazil grape seeds than that observed in this study. The protein content indicates a potential use of these residues, once they are differentiated in a diet, able to be inserted into restricted feed systems, and also aid in maintaining weight loss (Mohd Adzim Khalili et al., 2009Mohd Adzim Khalili, R., Norhayati, A. H., Rokiah, M. Y., Asmah, R., Siti Muskinah, M., & Abdul Manaf, A. (2009). Hypocholesterolemic effect of red pitaya (Hylocereus sp.) on hypercholesterolemia induced rats. International Food Research Journal, 16(3), 431-440.).

According to Luque-Rodríguez et al. (2005)Luque-Rodríguez, J. M., Castro, M. D. L., & Pérez-Juan, P. (2005). Extraction of fatty acids from grape seed by superheated hexane. Talanta, 68(1), 126-130. http://dx.doi.org/10.1016/j.talanta.2005.04.054. PMid:18970294.
http://dx.doi.org/10.1016/j.talanta.2005... , the highest concentration of lipid on the grapes is found in seeds and may vary from 10 to 16% depending on the cultivar analyzed. It was confirmed for both samples of present study.

3.2 Antioxidant activity

According to the results (Table 3), the extracts containing ternary fraction of solvents (water, ethanol and acetone) showed best results for DPPH essays (1296.00 ± 256.62 and 1704.00 ± 80.13 μmol TE g^-1 sample) and ABTS (1684.80 ± 127.27 and 2835.77 ± 116.71 μmol TE g^-1 sample), followed by the binary compositions, highlighting the fraction of water/acetone (DPPH: 1184.88 ± 42.86 and ABTS: 1452.13 ± 14.04 μmol TE g^-1 sample), and water/ethanol (DPPH: 1453.33 ± 58.84 and ABTS: 1980.13 ± 61.49 μmol TE g^-1 sample) for fermented and unfermented seeds, respectively. The lowest antioxidant activity results were obtained for pure solvents. This same behavior was described by DiCiaula et al. (2014)DiCiaula, M. C., Lopes, G. C., Scarminio, I. S., & De Mello, J. C. P. (2014). Optimization of solvent mixtures for extraction from bark of schinus terebinthifolius by a statistical mixture-design technique and development of a uv-vis spectrophotometric method for analysis of total polyphenols in the extract. Quimica Nova, 37(1), 158-163. http://dx.doi.org/10.1590/S0100-40422014000100026.
http://dx.doi.org/10.1590/S0100-40422014... over the effects of solvents on the total polyphenols content, and antioxidant capacity of the crude extracts of Schinus terebinthifolius Raddi (Anacardiaceae) peel. Regarding the performance of solvents binary mixtures, Cheng et al. (2012)Cheng, V. J., Bekhit, A. E.-D. A., McConnell, M., Mros, S., & Zhao, J. (2012). Effect of extraction solvent, waste fraction and grape variety on the antimicrobial and antioxidant activities of extracts from wine residue from cool climate. Food Chemistry, 134(1), 474-482. http://dx.doi.org/10.1016/j.foodchem.2012.02.103.
http://dx.doi.org/10.1016/j.foodchem.201... , studying wine extracted with different solvents (water, methanol, ethanol, and acetone), also reported that the water/acetone extracts reflected in higher antioxidant activity when compared to the ethanol/water extracts.

The analysis of the best extracting solvents to FRAP methodology presented the largest average in the binary fraction of acetone and water for fermented and unfermented seed, respectively (4557.11 ± 78.12 and 5372.66 ± 215.72 μmol EQFeSO₄ g^-1). Rockenbach et al. (2008)Rockenbach, I. I., da Silva, G. L., Rodrigues, E., Kuskoski, E. M., & Fett, R. (2008). Solvent influence on total polyphenol content, anthocyanins, and antioxidant activity of grape (Vitis vinifera) bagasse extracts from Tannat and Ancelota - different varieties of Vitis vinifera varieties. Food Science and Technology, 28(Suppl.), 238-244. http://dx.doi.org/10.1590/S0101-20612008000500036.
http://dx.doi.org/10.1590/S0101-20612008... also report this behavior in a study to determine the influence of solvent (acetone and ethanol - 50% and 70%) in the total polyphenol content of grape bagasse extracts (Vitis vinifera) in Tannat and Ancelota cultivars.

The content of phenolic compounds and flavonoids showed better extraction in binary and ternary mixtures than in pure solvents, showing similar behavior to the DPPH and ABTS, with lower antioxidants extraction capacity in a pure fraction. The extract of fermented grape seed, for example, obtained with pure water and acetone had its content of phenolic compounds increased, respectively, from 1.8 and 1.4% to about 26% with combination of both solvents.

According to the antioxidant activities obtained for fermented grape seeds, the surface model analysis delineated in DPPH, ABTS, FRAP, total phenolic compounds and flavonoids were a special Cubic (P <0.001), the most significant and best fitted to predict the mixture behavior, with correlation coefficient (R) values of 0.9711; 0.9964; 0.9999; 0.9968, and 0.9930, respectively.

Equations 2, 3, 4, 5, 6 represent the models with optimal values evaluation for the antioxidant essays.

where: *W = water, E = ethanol,*A = acetone regarding volume fraction, FC = total phenolic compounds, and FLA = flavonoids content.

Figure 11b represents the results of the response surfaces from DPPH and ABTS methods, respectively, also showing optimal working conditions for the fermented seeds due to the mixture of solvents employed in the extraction. The reddish region in the graphics shows that the composition of the solvents with higher antioxidant activity for both methods is the ternary proportion. Expressing the maximum value for DPPH (1350.45 μmol TE g^-1 extract) in the composition 0.43:0.23:0.34 water/ethanol/acetone (v/v/v) with desirability of 0.843. As for the ABTS test, the maximum (1748.43 µmol TE g^-1 extract) in the composition was 0.42:0.24:0.34 water/ethanol/acetone (v/v/v), and 0.984 of desirability.

Figure 1c shows the response surface of the fermented seed for the FRAP method, showing best results with the ternary composition, with maximum value (4580.65 μmol EQFeSO₄ g^-1 extract) in the composition 0.50:0.10:0.40 water/ethanol/acetone (v/v/v), and desirability of 0.985. The water acts enhancing the antioxidant compounds in this method, once the responses with this solvent are greater than in the fractions without it. Water combined with other organic solvents contributes to forming a moderately polar medium favoring the extraction of polyphenols since mediums with extreme polarities are not characterized as good extractor (Liu et al., 2000Liu, F. F., Ang, C. Y. W., & Springer, D. (2000). Optimization of extraction conditions for active components in Hypericum perforatum using response surface methodology. Journal of Agricultural and Food Chemistry, 48(8), 3364-3371. http://dx.doi.org/10.1021/jf991086m. PMid:10956117.
http://dx.doi.org/10.1021/jf991086m... ).

Figure 11e show the data obtained for the response surface for flavonoids and phenolic compounds, respectively, according to the proportion of the solvents. The ternary mixture of solvents in the composition 0.47:0.12:0.41 water/ethanol/acetone (v/v/v) showed better response (269.15 mg EAG g^-1 extract) for the phenolic compounds test with 0.999 of desirability. This same fraction of solvents in the composition 0.48:0.11:0.41 water/ethanol/acetone (v/v/v), favors the maximum extraction of flavonoids (8.82 mg EQ g^-1 extract), with 0.966 of desirability. In both tests, it is perceived that its extraction is favored by the increase of water and acetone.

For unfermented grape seeds, the same procedure was performed to provide the response surface for the tests. The correlation presented to the special cubic model was the best fit (p<0.001 and R: 0.9939; 0.9924; 0.9977; 0.9950, and 0.9959) to determine DPPH, ABTS, FRAP, phenolic compounds and flavonoids, respectively.

Equations 7, 8, 9, 10, 11 represent the model obtained for each method for unfermented grape seeds.

where: *W = water, E = ethanol,*A = acetone regarding volume fraction, FC = total phenolic compounds, and FLA = flavonoids content.

Figure 22b shows the response surfaces generated for the DPPH and ABTS assay, respectively for the unfermented seeds, as a function of the solvents, where the best response was the ternary composition of the solvents for both. The maximum value obtained for DPPH (1741.46 µmol TE g^-1 extract) was in the composition 0.40:0.25:0.35, water/ethanol/acetone (v/v/v) with 0.973 of desirability and for ABTS (2869.78 µmol TE g^-1 extract) in the composition 0.38:0.28:0.34 water/ethanol/acetone (v/v/v), with 0.969 of desirability.

The response for the FRAP antioxidant activity is favored with the reduction of the ethanol (solvent) in the fraction. As shown in Figure 2c, the best response was in the binary mixture of the solvents water and acetone, with the maximum value obtained (5375.88 μmol EQFeSO₄ g^-1 extract) in the composition 0.52:0.48 (v/v), and 0.971 of desirability.

The response surfaces for the total phenolic compounds and flavonoids (Figure 22e), as a function of solvent proportion, show the best results in ternary fractions: (328.75 mg EAG g^-1 extract) 0.45:0.21:0.34 water/ethanol/acetone (v/v/v) of composition and 0,993 of desirability for the phenolic compounds. As for flavonoids, (13.02 mg EQ g^-1 extract) the best composition of 0.46:0.27:0.27 water/ethanol/acetone (v/v/v), with desirability of 1.000.

Linear functions of desirability were employed to define the composition of extraction solvents for fermented and unfermented seeds, maximizing the antioxidant activity. For fermented seeds, the values 57.33, 824.66, and 1592.00 μmol EQTrolox g^-1 were used for DPPH test; 62.26, 918.53, and 1774.8 μmol E.Trolox g^-1 for ABTS test; 49.26, 2347.63, and 4646.00 μmol EFeSO₄ g^-1 for FRAP test; 13.69, 141.42, and 269.15 mg EAG g^-1 for total phenolics compounds; 0.34, 4.73 and 9.12 mg EQ g^-1 for flavonoids, with desirability of 0, 0.5, and 1 for all tests corresponding to these values.

For unfermented seeds, the values 76.00, 931.33 and 1786.67 μmol E.Trolox g^-1 were used for DPPH test; 255.60, 1603.47, and 2951.33 μmol E.Trolox g^-1 for ABTS test; 173.26, 2849.63, and 5526.00 μmol EFeSO4 g^-1 for FRAP test; 19.72, 175.24, and 330.77 for total phenolic compounds, and 1.00, 7.00, and 13.00 mg EQ g^-1 for flavonoids, corresponding to desirability values of 0, 0.5, and 1 for all tests.

The predicted values found by desirability (Table 4), were similar to the results obtained experimentally for both fermented and unfermented seeds. According to DiCiaula et al. (2014)DiCiaula, M. C., Lopes, G. C., Scarminio, I. S., & De Mello, J. C. P. (2014). Optimization of solvent mixtures for extraction from bark of schinus terebinthifolius by a statistical mixture-design technique and development of a uv-vis spectrophotometric method for analysis of total polyphenols in the extract. Quimica Nova, 37(1), 158-163. http://dx.doi.org/10.1590/S0100-40422014000100026.
http://dx.doi.org/10.1590/S0100-40422014... , the choice of the extraction method has a considerable effect on the quality of the extract, but the nature of the solvent used for extraction provides the most obvious influence in the qualitative composition of the extract.

Considering the polarity of the solvents involved in the extractions with water (1.000), ethanol (0.654), and acetone (0.355) (Reichardt & Welton, 2010Reichardt, C., & Welton, T. (2010). Solvents and solvent effects in organic chemistry (4th ed.). Weinheim: Wiley. http://dx.doi.org/10.1002/9783527632220.
http://dx.doi.org/10.1002/9783527632220... ), it is clear their influence on the extraction methods. Knowing the fractions with the best responses to the extraction of antioxidant compounds, one can calculate the volumetric proportion employed in each seed, according to Snyder (1968)Snyder, V. L. R. (1968). Principles of adsorption chromatography: the separation of nonionic organic compounds. Chromatographic Science Series, 3, 1-424. http://dx.doi.org/10.1002/ardp.19693020607.
http://dx.doi.org/10.1002/ardp.196930206... . The composition obtained as the best response for the unfermented seeds (45% water, 20% ethanol, and 35% acetone) showed 0.705 of polarity, while for the fermented seeds (48% water, 14% ethanol, and 38% acetone) showed 0.706, very close numbers.

In a big picture, the use of solvents alone for extraction with high polarities, as water, or low polarities, such as acetone, did not show good results in the extraction of antioxidant compounds for grape seeds. Although the binary combination of solvents has increased the extractable compounds, it has not been fully effective in most of these extractions. However, one can see that the solvents in the ternary fraction, which showed moderate polarity, presented a better universal extraction capacity (Liu et al., 2000Liu, F. F., Ang, C. Y. W., & Springer, D. (2000). Optimization of extraction conditions for active components in Hypericum perforatum using response surface methodology. Journal of Agricultural and Food Chemistry, 48(8), 3364-3371. http://dx.doi.org/10.1021/jf991086m. PMid:10956117.
http://dx.doi.org/10.1021/jf991086m... ).

The two types of seeds, although representing residues with different treatments, presented composition and polarity of the solvent extraction mixture very close. Moreover, the results of antioxidant activity were similar. However, one can see that after the desirability analysis for the best extraction fraction of solvents, the unfermented grape seed excelled at all submitted tests.

3.3 Trans-resveratrol

The trans-resveratrol content found in the two seeds showed statistical differences (P <0.05) (Table 3). The highest indexes of the polyphenol were in the unfermented grape seed extracts, highlighting the ternary fraction of the extraction solvents water/ethanol/acetone (9.84 ± 0.37 mg L^-1 of trans-resveratrol), followed by the binary fractions water/ethanol (8.94 ± 1.12 mg L^-1 of trans-resveratrol), and water/acetone (8.28 ± 0.24 mg L^-1 of trans-resveratrol). The composition of the solvents, in particular, the ternary mixture water, ethanol and acetone favored the extraction of trans-resveratrol polyphenol. According to Casas et al. (2010)Casas, L., Mantell, C., Rodríguez, M., de la Ossa, E. J. M., Roldán, A., De Ory, I., Caro, I., & Blandino, A. (2010). Extraction of resveratrol from the pomace of Palomino fino grapes by supercritical carbon dioxide. Journal of Food Engineering, 96(2), 304-308. http://dx.doi.org/10.1016/j.jfoodeng.2009.08.002.
http://dx.doi.org/10.1016/j.jfoodeng.200... , trans-resveratrol content found in these samples ranged similarly to those found in white wine (0.01 up to 8 mg L^-1).

3.4 Seeds oil

The volume of oil obtained by cold pressing had a yield of 13.0% for both seeds. This value was very close to the obtained by the Soxhlet method which ranged from 13.46 to 13.73% for fermented and unfermented seeds, respectively. This indicates that the oil extraction method may not be a decisive factor in relation to the yield of lipid content in the seeds. However, according to (Ribeiro et al., 2019Ribeiro, P. P. C., Silva, D. M., Dantas, M. M., Ribeiro, K. D. S., Dimenstein, R., & Damasceno, K. S. F. S. C. (2019). Determination of tocopherols and physicochemical properties of faveleira (Cnidoscolus quercifolius) seed oil extracted using different methods. Food Science and Technology, 39(2), 280-285. http://dx.doi.org/10.1590/fst.24017.
http://dx.doi.org/10.1590/fst.24017... ), the yield and physicochemical quality can be affected by the oil extraction method. Table 5 shows that most of the results had no significant difference (P <0.05) between the concentrations of fatty acids in the analyzed seeds.

It were identified 12 fatty acids (FA) in both seeds studied, where linoleic acid (18:2n-6) was the major, followed by oleic acid (18:1n-9), and palmitic acid (16:0). Similar results were found in other studies of grape seeds (Cao & Ito, 2003Cao, X., & Ito, Y. (2003). Supercritical fluid extraction of grape seed oil and subsequent separation of free fatty acids by high-speed counter-current chromatography. Journal of Chromatography A, 1021(1-2), 117-124. http://dx.doi.org/10.1016/j.chroma.2003.09.001. PMid:14735980.
http://dx.doi.org/10.1016/j.chroma.2003.... ; Rockenbach et al., 2010Rockenbach, I. I., Rodrigues, E., Gonzaga, L. V., & Fett, R. (2010). Composição de ácidos graxos de óleo de semente de uva (Vitis vinifera L. e Vitis labrusca L.). Brazilian Journal of Food Technology, 13(EE01), 23-26. http://dx.doi.org/10.4260/BJFT20101304104.
http://dx.doi.org/10.4260/BJFT2010130410... ; Santos et al., 2011Santos, L. P., Morais, D. R., Souza, N. E., Cottica, S. M., Boroski, M., & Visentainer, J. V. (2011). Phenolic compounds and fatty acids in different parts of Vitis labrusca and V. vinifera grapes. Food Research International, 44(5), 1414-1418. http://dx.doi.org/10.1016/j.foodres.2011.02.022.
http://dx.doi.org/10.1016/j.foodres.2011... ). The other FA showed low concentrations for both seeds. An organism can biosynthesize saturated and unsaturated fatty acids of the omega 9 family. However, it does not produce linoleic acid (omega 6), a necessary component to be inserted in a diet (Aguiar et al., 2011Aguiar, A. C., Cottica, S. M., Boroski, M., Oliveira, C. C., Bonafé, E. G., França, P. B., Souza, N. E., & Visentainer, J. V. (2011). Quantification of essential fatty acids in the heads of nile tilapia (Oreochromis niloticus) fed with linseed oil. Journal of the Brazilian Chemical Society, 22(4), 643-647. http://dx.doi.org/10.1590/S0103-50532011000400005.
http://dx.doi.org/10.1590/S0103-50532011... ). The nutritional quality of the lipid fraction of the Bordô seeds samples (Table 5) was similar since the index of atherogenicity (IA) and thrombogenicity (IT) were identical for both. These indexes indicate the potential for stimulating platelet aggregation, i.e., the lower the IA and IT values, more antiatherogenic FA are present in the lipid fraction, and greater the prevention of the onset of coronary diseases (Tonial et al., 2010Tonial, I. B., de Oliveira, D. F., Bravo, C. E. C., de Souza, N. E., Matsushita, M., & Visentainer, J. V. (2010). Caracterização físico-química e perfil lipídico do salmão (Salmo salar L.). Alimentos e Nutrição, 21(1), 93-98.; Turan et al., 2007Turan, H., Sönmez, G., & Kaya, Y. (2007). Fatty acid profile and proximate composition of the thornback ray (Raja clavata, L. 1758) from the Sinop coast in the Black Sea. Journal of FisheriesSciences.Com, 1(2), 97-103. http://dx.doi.org/10.3153/jfscom.2007012.
http://dx.doi.org/10.3153/jfscom.2007012... ). In this manner, the obtained results for Bordô seed oils suggest they are good sources of lipids for the human organism, according to Shinagawa et al. (2015)Shinagawa, F. B., Santana, F. C., Torres, L. R. O., & Mancini-Filho, J. (2015). Grape seed oil: a potential functional food? Food Science and Technology, 35(3), 399-406. http://dx.doi.org/10.1590/1678-457X.6826.
http://dx.doi.org/10.1590/1678-457X.6826... . The ratio of hypocholesterolemic and hypercholesterolemic (HH) FA was also close in both samples (unfermented 13.50; fermented 12.88). This relationship is linked to the incidence of cardiovascular diseases since it considers the functional activity of FA in the metabolism of plasma cholesterol transport lipoproteins (quantity and type). Thereby, higher values of HH correspond to products with a desirable composition of FA (Tonial et al., 2010Tonial, I. B., de Oliveira, D. F., Bravo, C. E. C., de Souza, N. E., Matsushita, M., & Visentainer, J. V. (2010). Caracterização físico-química e perfil lipídico do salmão (Salmo salar L.). Alimentos e Nutrição, 21(1), 93-98.).

Regarding the antioxidant activity, the oils from both samples showed lower values than the extracts, while phenolic compounds and trans-resveratrol were not found in oils. According to (Mahanna et al., 2019Mahanna, M., Millan-Linares, M. C., Grao-Cruces, E., Claro, C., Toscano, R., Rodriguez-Martin, N. M., Naranjo, M. C., & Montserrat-de la Paz, S. (2019). Resveratrol-enriched grape seed oil (Vitis vinifera L.) protects from white fat dysfunction in obese mice. Journal of Functional Foods, 62, 103546. http://dx.doi.org/10.1016/j.jff.2019.103546.
http://dx.doi.org/10.1016/j.jff.2019.103... ), after quantify the levels of trans-resveratrol in grape seed-oil, only a fraction of this polyphenol remained after the oil extraction, probably due its polarity.

4 Conclusion

The experimental design obtained by the special cubic model predicted that the best results of the antioxidant activity were obtained for the ternary mixture of the solvents (water, ethanol, and acetone), increasing about 20 times this bioactivity for both seeds when compared to pure solvents. The quantification of trans-resveratrol demonstrated this polyphenol is also present in most of the ternary composition of the extractor solvent. It was possible to observe the close relationship between the polarities of the extractor solvents involved in response of the antioxidant activity. The composition that showed the best response ranged between 45-48% of water, 14-20% of ethanol and 35-38% of acetone for both grape seeds, with polarity ranged from 0.705 to 0.706. By the same way, oils of both seeds presented a similar fatty acid profile, being this vinification by-product characterized by high content of polyunsaturated fatty acids, predominating linoleic acid.

Therefore, the use of this optimized mixture of solvents can be applied in grape seeds antioxidant extraction and results at an extract with stronger bioactivity. In addition, the developed extraction method preserve bioactive compounds from grape seeds, adding value on this by-product that can be availed on food industry as a preservative.

Acknowledgements

The authors acknowledge to Unipar, Campus Toledo, for borrowing the universal press to obtain the oil and Gilmar Camargo for donating the oil extraction template.

References

Publication Dates

History