ENZYME-AIDED EXTRACTION OF POLYPHENOLS FROM GRAPE POMACE

Transcription

1 ISSN: ENZYME-AIDED EXTRACTION OF POLYPHENOLS FROM GRAPE POMACE Noelia Costoya 1, Jorge Sineiro 1, Manuel Pinelo 2, Mónica Rubilar 3 and María José Nuñez 1* 1 Department of Chemical Engineering, ETS Engineering, Universidad de Santiago de Compostela, Lope Gómez de Marzoa, Santiago de Compostela, Spain 2 BioProcess Engineering, Department of Chemical Engineering, Building 229, Technical University of Denmark, 2800 Kgs. Lyngby, Denmark 3 Center of Agri-aquatic Nutritional Genomics, Universidad de La Frontera, Avda.Francisco Salazar 01145, Temuco, Chile mariajose.nunez.garcia@usc.es ABSTRACT Nowadays, grapepomace is frequently used as a source of polyphenols. With the aim of enhancing the yield of the extraction, the capacity of three commercial enzymatic preparations, Cellubrix (cellulase and β-glucosidase activities), Neutrase (protease and α-amylase activities) and Viscozyme (cellulase, hemicellulase and pectinase activities) was tested in this study. Individually applied, none of the enzymatic formulations was able to improve the efficiency of the aqueous extraction (control). However, acting in combination, a significant increase in the yields of soluble solids ( 20-45%), and phenols ( 25-65%), so as a notable improve in antiradical activity DPPH, with the subsequent decrease when measured as EC 50 ( 53-70%) was detected either in a white variety (Garnatxa). as in a red one (Cabernet Sauvignon). Synergistic effect at different dosages (0-30 g Enzyme/Kg substrate) and incubation times (1-2h) was further studied. Significance of dosage was found to be dependent on the grape variety and on the nature of the pomace (pressing, distillation). The highest phenolic yields ( g/kg of dry raw material) were obtained when distilled Cabernet Sauvignon pomace was treated with the mixture in a dosage level of 30 g/kg. KEYWORDS Grape pomace, Enzymes, Cabernet Sauvignon, Garnatxa, Phenols, DPPH antiradical activity, Synergism.

2 INTRODUCTION Most of the phenols contained in grape skin are found in the inner layer closest to the pulp, called hypodermis, composed of several layers of cells. The cell wall of grape plant is composed of 30% neutral polysaccharides (cellulose, xyloglucan, arabinan, galactan, xylan and mannan), 20% of acidic pectin substances, 15% of insoluble proanthocyanidins, lignin, structural proteins and phenols, these two latter cross-linked to the lignin-carbohydrate framework [1]. Two complementary theories have been proposed to explain the linkage between cell wall polysaccharides and phenols: Freitas et al. [2] have reported the existence of hydrogen bonds between the hydroxyl groups of the phenols and the oxygen atoms of the cross-linking ether bonds of sugars, whereas the studies of Le Bourvellec et al [3] concluded that some polysaccharides can form hydrophobic pockets able to encapsulate phenols, resulting in a stable complex. Many works dealing with the extraction of phenols from agricultural matrixes by using conventional and supercritical solvents have been conducted in the past recent years [4, 5]. However, only the most accessible and feebly attached phenols can be efficiently released by just using a solvent and applying conditions making phenolics more easily extractable, for which enzymes can certainly be a valuable help. Pectinases are the cell wall degrading enzymes more widely used in the fruit industry nowadays, followed by cellulases and proteases. Wineries produce around 14.5 million tons of grape byproducts annually; the recovery of the phenol compounds contained in the exhausted residue becomes an interesting alternative for being used as antioxidants, antimicrobials or colorants in other food systems. The objective of this work was to test the effect of different enzyme commercial preparations (individually or in combination) on the release of soluble solids and phenol compounds from two different varieties of grape pomace, and compare the results with those of a conventional aqueous extraction (control). Furthermore, the effect of the enzyme/substrate ratio, maceration time and nature of grape pomace (variety and pretreatment) has been tested in several cases. MATERIAL AND METHODS Pomace. Grape pomaces from Vitis vinifera L., of the varieties Garnatxa (white) and Cabernet-Sauvignon (red) were used in this work. Garnatxa variety was obtained from Bodegas Torres S.A. (Barcelona, Spain), whereas Cabernet-Sauvignon was supplied by Aguardientes de Galicia, S.A. (Vedra, A Coruña, Spain). First experiments were developed with pomace from pressing; once selected the enzyme with best results, distilled grape pomaces were also employed. The notation was WD for pressing pomace, and D for the distilled one. Grinding. Grape pomace was stored at -80ºC until used. Then, it was milled in a conventional coffee grinder, sieving it to select a particle size among 1 and 2 mm. Chemicals and enzymes. Gallic acid as standard and 2,2-Diphenyl-1-picrylhydrazyl radical (DPPH) were purchased from Sigma Chemical Co. HPLC-grade solvents (methanol, acetic acid), and Folin-Ciocalteau reagent were purchased from Panreac (Barcelona, Spain). 697

3 The enzymatic preparations Cellubrix (mixture of cellulose and β-glucosidase), Viscozyme (hemicellulases, cellulases and pectinases) and Neutrase (protease and variable quantities of α-amylase) were supplied by Novozymes A/S, Denmark, in bottles of 100 ml. Extraction. For the pre-selection of enzymes, the pressing pomace samples were subjected to extraction in a rotary shaker G24 New Brunswick Scientific Co. Inc. (N.J., USA) at a constant stirring rate of 140 rpm. The extraction conditions were those found as optimal for DPPH radical scavenging capacity in previous experimental designs by Rubilar et al [6]: 30 minutes, 50ºC and solvent/solid (L/S) ratio of 1:1. The enzyme/substrate ratio (total mass of the enzyme commercial preparation relative to the dry mass of grape pomace) was 15 g/kg and the enzymes were added together with solvent (water), so the incubation time of samples was the same as the extraction time. Enzymatic complexes were added individually and also as a mixture of them (1:1:1). The most effective enzyme activities were selected for subsequent experiments, in which distilled pomaces were also tested. For this case, the experiments were performed according to an experimental design which was defined for the different substrates (Garnatxa WD, Garnatxa D, Cabernet WD and Cabernet D), so structured in four groups. The enzyme/seed mass ratio was evaluated at 15 and 30 g/kg and the time of treatment evaluated at 1 and 2 hours. Control samples without enzymes (E/S= 0) were considered for all conditions. These conditions will appear joint to corresponding results in Table 2 Determination of total extractable compounds The released extractable compounds delivered during the pre-selection of enzymes were determined by gravimetry after extraction. Each sample was filtered through 0.45 µm membrane filters, and 10 ml were taken in tared 20-mL vials, introduced in the vials accessory and solvent-evaporated in a Buchi R-114 rotary evaporator. The samples were dried at 100 ºC in oven for 2 hours. Determination of total polyphenolics. Total phenolics were determined by the Folin- Ciocalteu method, as modified by Singleton & Rossi [7]. Each sample ( L) was mixed with 10-3 L of ten-fold diluted Folin-Ciocalteu reagent, shaken for 1 minute, and then L of a sodium carbonate solution containing 7.5 Kg/L was added. Samples were left at room temperature in the darkness for 2 hours and then their absorbances measured at 765 nm in spectrophotometer Shimadzu UV-160A. Results were expressed as % or as mass equivalent of gallic acid (g gallic acid/kg dry substrate). Determination of antioxidant activity as radical scavenging capacity. A DPPH radical scavenging assay was performed using the method described by Von Gadow et a1. [8] to determine the proton-donating ability of the raw extract ml of 6.1*10-5 M DPPH methanol solution was used. The reaction was started by the addition of 50 µl of sample. The bleaching of DPPH was followed at 515 nm (spectrophotometer Shimadzu UV-160A) at 25 ºC for 16 minutes. The inhibition percentage of the DPPH radical was calculated as follow: 698

4 A i = initial absorbance, time= 0, A f = final absorbance When the extracts had a high antiradical power they saturated DPPH, and the inhibition percentage values were very similar and all close to 100%. Consequently, each extract required a different series of dilutions, which were applied to obtain graphically the concentration of extract (mol gallic acid/mol DPPH) that induced a decrease of 50% of DPPH. That value, named EC 50 (efficient concentration) will be lower when higher the antiradical activity HPLC Analysis. HPLC technique was used to analyze the modification of polyphenolic profiles on Garnatxa grape pomace under different treatments. Samples were dissolved in HPLC-grade pure methanol (Panreac, Barcelona, Spain), filtered through a 0.45-µm nylonfilter and injected (20 µl) into HPLC system. The reverse-phase HPLC apparatus with a pump PU-980 connected to a quaternary gradient unit LG , a JASCO UV-1575 UV- Vis detector and a Rheodyne model 7725 loading sample injector with a 20 µl sample loop was used to obtain the HPLC profiles. The column (250 mm 4.6 mm) was a C 18 Hypersil ODS (5 µm particle size) (Supelco Inc., Pennsylvania, USA). Linear elution gradient was performed with solvents (A): 0.5 g/l acetic acid in Milli- Q water solution and (B): HPLC-grade methanol; the volumetric ratios of solvents A and B were as follows: 0-10 min, 95A/5B; min, 50A/50B; min, 30A/70B; min, 95A/5B. The flow rate was set to L/min, and detection took place at 280 nm.. Statistical analysis. The results reported in this work are the averages of, at least, three measurements, being expressed as mean ± standard deviation (SD). Tests for means were performed to analyze the data from the pre-selection of enzymes. Further assays with the selected enzymes were performed following an experimental design, evaluating three factors at two levels as previously described. Analysis of variance was performed by General Linear Models, taking into account in the model the direct effects, the first-order interactions and also second-order interactions among factors. SPSS version 13 from SPSS Inc. (Chicago, Illinois) was used to perform this analysis. RESULTS AND DISCUSSION Effect of the enzyme preparations on phenol release. The soluble solids (g/kg) obtained after extraction were much higher (about 3-4 times higher) in Garnatxa variety than in Cavernet-Sauvignon, as can be seen in Table 1. This result is explainable because of the previous contact existing between must and red grapes, which involves the transfer of many compounds to wines. For both varieties, only the mixture of the three enzymatic preparations (C+V+N) had a significant effect on the amount of compounds extracted, increasing by 21% for Garnatxa and 46% for Cabernet. When treated with each one of the single enzymatic preparations, no significant differences, compared to the control, were detected. 699

5 Table 1. Total extractable compounds, polyphenolics yield and DPPH scavenging capacity of extracts submitted to enzyme-aided extraction. % Extractable Compounds Samples Garnatxa Cabernet-Sauvignon control 25.0 ± ± 0.25 cellubrix 27.0 ± ± 0.50 viscozyme 25.5 ± ± 0.22 neutrase 27.0 ± ± 0.42 C+V+N 30.2 ± ± 0.12 % Polyphenols Garnatxa Cabernet-Sauvignon control ± ± cellubrix ± ± viscozyme ± ± neutrase ± ± C+V+N ± ± EC 50 (mol gallic acid/mol DPPH) Garnatxa Cabernet-Sauvignon control 14.0± ± 0.10 cellubrix 10.0 ± ± 0.09 viscozyme 11.6± ± 0.08 neutrase 7.8 ± ± 0.09 C+V+N 4.2 ± ± 0.06 C+V+N: mixture of Cellubrix+Viscozyme+Neutrase. This fact confirms the available data indicating that only 5-10% by weight of the skin dry matter can be normally degraded by using this kind of enzymes, suggesting that the degree of the cell-wall polysaccharide break-down that can be achieved is indeed low [9]. The results of polyphenolics compounds in the extracts are also shown in Table 1 and bear out the synergistic effect of using the combination of the three enzyme preparations, which is clear in the case of Cabernet Sauvignon (increase of 64%) and less significant in the case of Garnatxa (increase of 24%). Despite a lower content of extractable compounds ( >3 times lower) being found, red pomace had twice more phenols, in accordance with other authors [4], besides being more sensitive to enzymatic action, perhaps because cells became more damaged in the previous fermentation. When acted separately, enzymes showed only a minor influence on the release of phenols, specially Viscozyme. Synergistic effects among cell-wall degrading enzymes were already reported for other plant substrates like maize bran and apple peel, in the first case increasing the release of higher amounts of ferulic acid, and for apples enhancing the sugar matrix decomposition in apple peel [10]. Antiradical activity of the extracts measured as EC 50 shows the same tendency that polyphenolic yields, likely due to the fact that increasing amounts of phenols result in a higher free radical scavenging; Neutrase offers good numbers, but the best results corresponded to the mixture C+V+N, with strong decreases in EC 50, about 70% with Garnatxa variety and about 53% for Cabernet variety. For explaining the relatively good behaviour of Neutrase, much better than that of Viscozyme, it is necessary mention that the cell wall of grape berry is built of a cellulosexyloglucan framework that is embedded in a matrix of pectin polysaccharides, and phenols are linked to sugars by hydrogen bonds and hydrophobic interactions. To stiffen and bind 700

6 together the polysaccharide domains, structural proteins called extensins are also present in a considerable amount (around the 10% of the primary cell wall) [11]. Neutrase may catalyze the breakdown of extensins, presumably allowing cellulases, hemicellulases and pectinases present in Viscozyme and Cellubrix to act subsequently, this probably being the key point for success. On top of this, some flavonoids of the grape pomaces are known to be contained in the cell vacuoles [12], which are surrounded by a proteic membrane (tonoplast), which could be also partially degraded by the protease action. From the data obtained, it can be concluded that only the combined action of enzymes with complementary activities are able to be active. Although several references on enzyme action are available, minor enhancements in the extraction yield of phenols by using cell-wall degrading enzymes have been accomplished over the last recent years [13]. A reason could be the inhibition of the ezyme action by tannins, diferulate cross-linking of the cell wall polysaccharides [14], or the presence of different phenols [15]. In this sense, the design of a continuous process involving the constant removal of the extracted phenols could definitely favour the enzymatic action on the sugar substrate The variables polyphenols (%) and EC 50 showed a good correlation for white grape pomace (EC 50 = Phenols (%), R 2 =0.99), but no correlation existed for Cabernet. Although enzymatic effect over antiradical activity was more acute on Garnatxa, Cabernet variety had a greater antiradical power, likely aided by a major presence of anthocyanins, reported as the main antioxidants in red grapes [16]. At last, with the aim of proving the influence of enzymatic treatment on the polyphenol profile, samples of Garnatxa pressing pomace before/after 30 minutes of incubation with C+V+N were HPLC-chromatographed. Results showed in Figures 1 and 2 indicate that the profiles were similar, but signals after treatment were more intense particularly for gallic acid, catechin and epicatechin, appearing at 11, 27, and 33 min, respectively. Figure 1. HPLC chromatogram for extract of Garnatxa grape pomace 701

7 Figure 2. HPLC chromatogram for extract of Garnatxa grape pomace after enzymatic treatment (C+V+N) Because of the obtained results, the mixture C+V+N was chosen for additional experiments, in which incubation times will be higher (1-2 hours), and E/S ratio will be g/kg; assays will be done on both pressing and distilled grape pomace. Effect of enzyme dosage and incubation time on the amount of released phenols. The results obtained according to the experimental design is shown in Table 2. From the data for controls, it is outstanding the greater antiradical power obtained for distilled pomaces (2-3 times in Cabernet and 1.3 times in Garnatxa). The increase in the amount of liberated phenols from the distilled substrates corresponded with an increase in the antiradical activity measured as EC 50. Perhaps the high temperatures employed in distillation favours the disruption of the cell walls, the structure become more open, favouring the penetration of enzymes, and making the phenol release more efficient. Regarding time and E/S ratio factors, the former didn t exert a positive effect when passed from 1 to 2 hours, although when compared values of E/S=0 with those of Table 1, parameters are notably improved if 1 h is employed instead 30 minutes. By this reason 1 hour seems to be an adequate time for incubation with enzymes. E/S ratio induced a positive effect when the concentration of (C+V+N) passed from 15 to 30 g/kg, especially for distilled pomace of Garnatxa, for which EC 50 decreased by 30%. For Cabernet results are more uniform, improving both polyphenols and antiradical power in 25-30% so for pressing as for distilled pomaces. 702

8 Table 2. Experimental design and values for Polyphenols and Antiradical Capacity corresponding to different grape pomaces Enzyme E/S (g/kg) time(hours) Polyphenols(g/Kg) EC 50 (mol gallic acid/mol DPPH) Cabernet WD ± ± ± ± 0.05 C+V+N ± ± ± ± ± ± ± ± 0.02 Cabernet D ± ± ± ± 0.02 C+V+N ± ± ± ± ± ± ± ± 0.01 Garnatxa WD ± ± ± ± 0.01 C+V+N ± ± ± ± ± ± ± ± 0.00 Garnatxa D ± ± ± ± 0.00 C+V+N ± ± ± ± ± ± ± ± 0.01 D: distilled, WD: without distillation, C+V+N: mixture of Cellubrix+Viscozyme+Neutrase. EC 50 : Concentration for 50% of DPPH inhibition Table 2 indicates that by the only effect of distillation (comparing pomaces WD and D, at E/S=0) EC 50 suffered a strong decrease even for a similar values of polyphenols, perhaps because procyanidins are better antioxidants that the corresponding monomers. These conclusions could be confirmed by applying the statistical model, which results appear in Table 3, where the high significance of the model (R 2 >0.99) can be observed. The factor Time was not statistically significant for polyphenols nor for antioxidant activity; therefore 1 hour with enzyme-aided extraction was enough to improve both dependent variables. For dependent variable Polyphenols, the nature of the substrate was the main source or variability, being the distilled red variety which rendered more polyphenols. In contrast, Enzyme was the variable with the highest influence in the model for EC 50 values, having the variable Substrate a minor influence. This significance was due to the high effect of the use of a mixture of enzymes ( C+V+N ) had on both polyphenolics and EC 50 values. Therefore, the multi-activity was a key factor to increase 703

9 polyphenolics and antioxidant power of the extracts, as can be expected by the heterogeneous nature of solid matrix. Crossed interaction of these two factors, so as their second-order interaction with E/S ratio, were also significant. Table 3. Variables and interactions with significant effects at p-value< POLYPHENOLS (g/kg) Variable F-value p-value MODEL (R 2 = 0.991) <0.000 SUBSTRATE 35.9 <0.000 ENZYME <0.000 ENZYME*SUBSTRATE ENZYME*SUBSTRATE*E/S EC50 (mol gallic acid/mol DPPH) Variable F-value p-value MODEL (R 2 = 0.995) <0.000 SUBSTRATE ENZYME 364 <0.000 ENZYME*SUBSTRATE ENZYME*SUBSTRATE*E/S <0.000 For explaining the better behaviour of the distilled pomace, the possible degradation and polymerization of some phenols can be also promoted at high temperature and compounds of considerable antioxidant power can be generated. To analyze this possible influence on antioxidant activity, an HPLC chromatogram was recorded for Garnatxa D (Figure 3); if compared with Figures 1 and 2, it can be seen that the main effect was the presence of a broad hump after 60 minutes, probably due to polymeric substances generated when distilling, because this peak appeared often for procyanidins. At time, peaks of catechin/epicatechin are much smaller, which is in accordance to the previous conclusion Figure 3. HPLC chromatogram for extract of Garnatxa grape pomace after distillation. 704

10 ACKNOWLEDGMENTS This work was funded by Spanish Ministry of Education and Science (ProjectAGL C03-03) and by Xunta de Galicia (07TAL PR). We also thanks Bodegas Torres S.A and Aguardientes de Galicia, S.A for supplying us the pomaces REFERENCES [1] N.C. Carpita, & D.M. Gibeaut. Plant Journal 3: 1-30 (1993) [2] V.de Freitas, E. Carvalho. & N. Mateus. Food Chem. 81: (2003) [3] C. Le Bourvellec, B.Bouchet & C.M.G.C Renard. Biochimica et Biophysica Acta-General Subjects, 1725: (2005). [4] M.Pinelo, M.Rubilar, M.Jerez, J.Sineiro. & M.J.Nuñez. J. Agric. Food Chem., 53: (2005). [5] M Pinelo, A. Ruiz-Rodríguez, J. Sineiro, F.J. Señoráns, G Reglero & M.J. Núñez. Eur. Food Res. Technol. 226: (2007). [6] M.Rubilar, M.Pinelo, D.Franco, J.Sineiro & M. J. Nuñez. Afinidad, 60: (2002). [7] V.L Singleton & J. A. Jr. Rossi. American J. Enol. Vitic. 16: (1965). [8] A. Von Gadow, E.Joubert. & C.F. Hansmann. J. Agric. Food Chem. 45: (1997). [9] M.Pinelo, A.Arnous. & A.S Meyer. Trends Food Sci. Technol. 17: (2006) [10] Y.J. Kim, D.O. Kim, O.K. Chun, D.H.Shin, C.Y.Lee & D.B.Wilson. J. Agric. Food Chem. 53: (2005). [11] M. Lecas & J.M. Brilloue. Phytochem. 35: (1994). [12] S. Conn, W. Zhang, & B. Franco. Biotechnol. Letters 25: (2003). [13] D.Kammerer, A. Claus, A. Schieber. & R. Carle. J. Food Sci. 70: C157-C163 (2005). [14] P.Schofield, D.M. Mbugua & A.N. Pell, Animal Feed Sci.Technol. 91: (2001). [15] J.Sineiro, H.Domínguez, M.J. Núñez, & J.M Lema. Biotechnol. Letters 19: (1997). [16] N.Seeram & M.G. Nair. J. Agric. Food Chem. 50: (2002). 705