Abstract
Assessment of the phenolic content of red grapes is an important prerequisite for understanding how grape phenols impact wine quality. The influence of selected factors on extraction efficiency of phenols from eight different red grape cultivars was investigated to determine a rapid and robust extraction method. The effects of solvent contact time, extraction temperature, and concentration of ethanol and hydrochloric acid on extraction of total phenols and anthocyanins were investigated in a series of statistically designed factorial experiments. The results were compared to the “total” concentrations as measured by a modified “total” extraction protocol and the data were expressed as relative extraction efficiencies. Both extraction temperature and concentration of the solvents ethanol and hydrochloric acid exerted highly significant effects on the extraction of both total phenols and anthocyanins. The optimized extraction procedure was as follows: mix 50% v/v acidified aqueous ethanol (0.1 M HCl) in a 1:1 v/w ratio with crushed grapes to give a final ethanol of ∼25% v/v for 5 min at 40°C, followed by neutralization and clarification. By this rapid procedure, it was possible to extract an average 81.8% of the total phenols and 91.5% of the anthocyanins from the grapes.
Red wine color is mainly dependent on the content and composition of the phenolic substances present, including notably the total concentration of anthocyanins and polymeric pigments (Bakker et al. 1986, Mazza et al. 1999, Sacchi et al. 2005). Although anthocyanins dominate the color of young red wines, other substances such as tannins play a major role in long-term wine color development and, hence, influence the wine color stability during aging (Cheynier et al. 2006). The color forms an important part of the perceived quality of red wine (Somers and Evans 1974), but the phenolic compounds also affect aroma and mouthfeel (Preys et al. 2006). Several changes in the phenolics occur during red wine production, particularly during fermentation and maturation. Nevertheless, a main hypothesis in our ongoing work on grape and wine phenols is that the phenolics present in the grapes significantly influence the quality of the finished wine and that it may be possible to predict wine quality from analysis of the phenolics in grapes.
A first step in testing this hypothesis is to establish a relationship between the phenolics in grapes and those in the resulting red wines. However, since assessment of the content and profile of phenolics present in grapes is strongly dependent on the extraction method employed, a prerequisite for obtaining a proper evaluation of the phenolics present in grapes is to define a robust extraction method for grape phenols.
Condensed tannins and anthocyanins are the two most abundant classes of polyphenols found in grapes. After the onset of veraison, anthocyanins accumulate in the skins while condensed tannins of both seeds and skin decrease during ripening (Adams 2006, Kennedy et al. 2000). Anthocyanins are easily extracted from grape skins with different solvents (acidic methanol, acidic ethanol, or acetone/water) that are also the conventionally most widely used solvents for phenols extraction from grapes (Macheix et al. 1990). The extraction of phenols from the seeds is more challenging. The seeds are generally not crushed during winemaking and the extraction of gallic acid, condensed tannins, and catechins from intact seeds into the fermenting wine is only achieved slowly during 5 to 12 days of maceration with a gradually increasing ethanol concentration that facilitates extraction from the seeds and skins (Gonzalez-Manzano et al. 2004).
The majority of seed polyphenols are located in the outer seed coat (Thorngate and Singleton 1994), and it is possible to extract the majority of the seed polyphenols under simulated but realistic winemaking conditions (Singleton and Draper 1964). However, when grape seeds are left intact during rapid extraction for phenols analysis, little or no phenols are extracted from the seeds (Meyer et al. 1997). In contrast, when the grape seeds are crushed, flavan-3-ols and gallic acid are rapidly extracted from the seed tissue. Hence, crushing of the grape seeds is required to obtain a representative extraction of phenols from the seeds during rapid grape extraction.
A complete extraction method for grape anthocyanins and total phenols, which requires a solvent contact time of 1 hour and a high ratio (10:1 v/w) of solvent to solids, has been reported (Iland et al. 2004). Other reported methods also rely on long extraction times, use of different organic solvents, or tedious, multistep sample preparation (Kallithraka et al. 1995, Saint-Cricq de Gaulejac et al. 1998). Surprisingly few studies have systematically addressed how different parameters of the grape phenols extraction process such as temperature, time, solvent-acidification, or type of cultivar quantitatively affect extraction of phenols from grapes intended for wine production.
In the context of establishing a consistent, rapid method for obtaining high extraction efficiencies of phenols from grapes, the objective of this study was to evaluate the influence of selected parameters expected to affect extraction efficiency and robustness and to identify an optimal extraction procedure. To mimic the events taking place during winemaking and avoid use of potentially toxic, highly volatile, or flammable organic solvents, ethanol was chosen as the most relevant extraction solvent for the study. The effects of extraction temperature, addition of acid, ethanol concentration, and extraction time (that is, solvent contact time) were evaluated on eight different red grape cultivars in statistically designed experiments using total phenols and anthocyanins as responses.
Materials and Methods
Grape samples and chemicals.
Grape samples of eight different red cultivars (Alicante, Merlot, Syrah, Cinsault, Grenache, Carignan, Cabernet Sauvignon, and Mourvedre) of Vitis vinifera were collected in the south of France in August and September 2005. Mature grapes were manually picked from vines from the same field location, frozen in polyethylene bags, and stored at −30°C until use. Technical grade 96% v/v ethanol (V&S Distillers, Aalborg, Denmark) and analytical grade hydrochloric acid (Merck, Darmstadt, Germany) were used to prepare extraction solvents. HPLC-grade acetonitrile, o-phosphoric acid, gallic acid, (+)-catechin hydrate, (−)-epicatechin, rutin hydrate, caffeic acid, and malvidin-3-glucoside hydrochloride for HPLC solvents and standards were purchased from Sigma-Aldrich (St. Louis, MO). Chemicals for tannin analysis (bovine serum albumin [BSA, fraction V powder], tartaric acid, sodium dodecyl sulfate [SDS], acetic acid, sodium chloride, triethanolamine [TEA], and ferric chloride hexahydrate) were all of analytical grade and also purchased from Sigma-Aldrich.
Grape homogenization.
Grapes were manually destemmed while frozen, placed in polyethylene bags, and sample aliquots for each extraction experiment were gently thawed in a water bath at room temperature for approximately 1 hour. The grapes were then crushed with an Ultra-Turrax T25 high speed homogenizer (IKA-Werke & Co. GmbH KG, Janke & Kunkel, Staufen, Germany) at 24,000 rpm under a stream of nitrogen for a defined time (i.e., 0.5 to 4 min), according to the experimental plan. The resulting grape purée of juice, skin, seeds, and pulp is referred to as grape homogenate in the following.
Estimation of juice content.
50 g grape homogenate was weighed and centrifuged (20 min, 4,800 g) and the resulting juice was passed through Whatman grade 4 filter paper (Whatman International, Kent, UK). For each sample, the juice density was estimated by weighing a 25-mL sample. The remaining solids were rinsed with 25 mL water, centrifuged 20 min at 4,800 g, filtered, and dried at 105°C overnight to give the mass of insoluble solids. The juice content was expressed as mL juice/g grape and the values obtained were used to calculate the concentrations of anthocyanins and total phenols/g grape.
Total extraction protocol.
For benchmarking, an estimate of “total” anthocyanins and phenols present in the different grape samples was made via a modified version of an existing “total” extraction protocol (Iland et al. 2004). The modification was that the solid grape residues from the first extraction were subjected to a second extraction, so that the final extraction protocol was as follows: 2 g grape homogenate was weighed and extracted for 1 hour at room temperature (∼25°C) with 20 mL aqueous ethanol (50% v/v, adjusted to pH 2 with HCl). During the solvent contact, the mixture was continuously mixed on an oscillating inverter. The extract was then centrifuged 5 min at 16,100 g and diluted 21 times with 1 M HCl. The resulting solid residues were rinsed with distilled water, recovered by filtration through Whatman grade 4 filter paper, and re-extracted overnight with 20 mL aqueous ethanol at room temperature during continuous mixing. The second extract was centrifuged 5 min at 16,100 g and diluted 7.67 times in 1 M HCl (the dilutions were made to target the optimal range for subsequent absorbance measurements). The absorbances at 280, 520, and 700 nm of the diluted samples, referred to as abs(280 nm), abs(520 nm), and abs(700 nm), respectively, were measured in 10-mm quartz cuvettes on a spectrophotometer (Cary 300, Varian, St. Helens, Australia). The abs(700 nm) data were used to evaluate the turbidity of the samples. The anthocyanin concentration was expressed as mg malvidin-3-glucoside equivalents/g grape from the absorbance at 520 nm via use of a generic extinction coefficient ε = 50 mL/(mg · cm) (equation 1) (Iland et al. 2004, Somers and Evans 1977). Total phenols were expressed as 0.01 absorbance units at 280 nm/g grape (equation 2) (Iland et al. 2004).
(eq 1)
(eq 2)
V is the volume of the added extraction solvent in mL, m is the mass of the extracted grape homogenate in g, juice content is the volume of juice in mL/g grape, DF is the dilution factor of the extract in 1 M HCl, and ε is the generic extinction coefficient of malvidin-3-glucoside in mL/(mg · cm). In the calculations, it was assumed that the material used for the re-extraction did not contain any juice (juice content = 0 mL/g). The total concentrations of anthocyanins and total phenols, respectively, were then calculated as the sum of the first and second extraction yields. For each grape cultivar, the benchmark “total” phenols and anthocyanins were determined from at least duplicate extractions.
Experiment 1. Fast extraction of phenols and anthocyanins.
The effect of extraction temperature (20, 40, or 60°C), solvent composition (0, 25, and 50% v/v ethanol), and hydrochloric acid concentration (0 or 0.1 M) on yields of total phenols and anthocyanins was examined in a randomized, full factorial design with one determination on each factor combination and three center points. For all extractions, a 1:1 weight:volume ratio of grape homogenate:solvent was used, as this ratio resulted in maximum phenols extraction in preliminary experiments (data not shown). For each extraction experiment, 400 g grapes were homogenized for 2.5 min and immediately thereafter aliquots of 15 g homogenate were transferred into individual beakers containing a stirring magnet. Solvent was added according to the experimental plan and the samples were sealed, placed in a magnetic stirring water bath, and incubated with stirring at defined temperatures of 20, 40, or 60°C for 30 min according to the experimental plan. After the solvent contact period, 2-mL aliquots of each sample were centrifuged for 5 min at 16,100 g and diluted 101 times in 1 M HCl. The different dilution factors for total and fast extractions were necessary because of the inherently higher solvent to solid ratio used in the total extraction protocol.
After 1 hour, the absorbances of the diluted samples were read at 280, 520, and 700 nm using 10-mm quartz cuvettes and the amount of anthocyanins and total phenols for both the fast solvent extractions and total extractions were calculated from equations 1 and 2. The abs(700 nm) data were used to evaluate sample turbidity. Each individual extraction yield was calculated as a relative percentage of the total determined on the same homogenate in triplicate using the total extraction protocol (Figure 1⇓). The experiment was repeated for nine samples, covering the different grape cultivars under investigation: Alicante, Merlot (two different samples), Syrah, Cinsault, Grenache, Carignan, Cabernet Sauvignon, and Mourvedre (Table 1⇓).
Effect of homogenization and solvent contact times.
180 g grapes (Alicante only) were destemmed, thawed, and homogenized as described above. During homogenization, 15-g samples were taken after 0.5, 1, 3, and 4 min of homogenization (single run). Each sample was supplemented with 15 mL 25% aqueous ethanol containing 0.1 M HCl, sealed, and stirred with a magnet on a 40°C water bath. A 2-mL sample was subsequently withdrawn from each beaker containing ∼30 g after 0, 2, 5, 15, and 30 min, centrifuged 1 min at 16,100 g, and diluted 51 times in 1 M HCl. The absorbances of the diluted samples were then measured at 280, 520, and 700 nm after 1 hour. The concentrations of anthocyanins and total phenols were calculated from equations 1 and 2 and the results were expressed as relative percentages of the anthocyanins and total phenols obtained in triplicate with the modified extraction protocol. The abs(700 nm) data were used to evaluate the turbidity of samples.
Experiment 2. Time-optimized extraction conditions.
150 g grapes were destemmed, thawed, and homogenized for 2 min as described above. Homogenate portions (16 g) were then transferred quickly to four beakers. Each sample was mixed with 1 mL 1.6 M HCl and 15 mL aqueous ethanol (53.3% v/v) that had been preheated to 60°C to minimize the heating time of the grape sample. These mixtures were then sealed and stirred with magnetic stirring at 40°C. After 5 and 15 min, respectively, the extracts (each in duplicate) were centrifuged for 5 min at 16,100 g and diluted 51 times in 1 M HCl. After 1 hour, the absorbances of these acidified samples were read at 280, 520, and 700 nm in 10-mm quartz cuvettes. The amounts of anthocyanins and total phenols were calculated from equations 1 and 2 and the results were expressed as relative percentages of the anthocyanins and total phenols obtained in duplicate with the total extraction protocol. To test the consistency of extraction among cultivars, the extractions were carried out on Alicante, Merlot, Syrah, Cinsault, Grenache, Carignan, Cabernet Sauvignon, and Mourvedre grapes (Table 1⇑).
Experiment 3. Final extraction protocol.
150 g grapes were destemmed, thawed, and homogenized for 2 min as described above. 50 g homogenate was weighed in a glass bottle and mixed with 50 mL acidic (0.1 M HCl) aqueous ethanol (50% v/v), which had been preheated to 60°C. The mixture was sealed and then stirred vigorously with a magnetic bar at 40°C. After 5 min, 1 mL 5 M sodium hydroxide solution was slowly added while stirring for 1 min to neutralize the added hydrochloric acid. The sample was centrifuged 10 min at 15,000 g and filtered through a Whatman grade 4 cellulose filter. Finally, the extracts were centrifuged for 5 min at 23,000 g and diluted 101 times in 1 M HCl. After 1 hour, the absorbances of these acidified samples were read at 280, 520, and 700 nm in 10-mm quartz cuvettes. The concentrations of anthocyanins and total phenols were calculated from equations 1 and 2 and the results were expressed as relative percentages of the anthocyanins and total phenols obtained in duplicate with the total extraction protocol. The abs(700 nm) data were used to evaluate the turbidity of the samples. To test the consistency of extraction among cultivars, the extractions were carried out on Alicante, Merlot, Syrah, Cinsault, Grenache, Carignan, Cabernet Sauvignon, and Mourvedre grapes (Table 1⇑).
Stability of phenolics during prolonged solvent contact.
100 g Merlot grapes were destemmed, thawed, and homogenized as described above. The homogenate (50 g) was extracted with 50 mL 50% v/v EtOH at ambient temperature by stirring for 1 hour, centrifuged for 10 min at 15,000 g, filtered through a Whatman grade 1 filter paper, further centrifuged for 10 min at 23,000 g, and finally filtered through a Titan2 0.45-μm nylon syringe filter (Sun Sri, Rockwood, TN) to obtain a solution that was practically free of particles. Two 10-mL aliquots were mixed with 10 mL extraction solvent (25% v/v ethanol, 0.1 M HCl, preheated to 60°C) and stirred at 40°C for 5 and 30 min, respectively. Immediately after the solvent contact period, the samples were neutralized with 0.2 mL 5 M sodium hydroxide and the volumes were adjusted to 25 mL with water. For comparison, an untreated sample was prepared by diluting a 10-mL aliquot to 25 mL with water. All samples were flushed with nitrogen and frozen for later phenolic analysis by HPLC and protein precipitation.
Analysis of phenolic compounds by HPLC.
Samples from the stability study were analyzed on an Agilent 1100 series HPLC instrument (Agilent, Waldbronn, Germany) equipped with a vacuum degasser, a quaternary pump, an autosampler, a thermostatted column compartment, and a diode array detector as described (Lamuela-Raventos and Waterhouse 1994), with some modifications. A Gemini C18 column (150 mm x 4.6 mm, 3-μm particle size, 110 Å pore size) (Phenomenex, Torrance, CA) with a 4 x 3 mm guard column of the same material was used as stationary phase at 40°C. The solvents were: solvent A (water with 0.20 M o-phosphoric acid and 3% v/v acetonitrile, adjusted to pH 1.50 with aqueous sodium hydroxide) and solvent B (a 1:1 v/v mixture of solvent A and acetonitrile). A constant flow of 0.5 mL/min was applied with a linear gradient elution profile of 0 min (11% solvent B), 40 min (40% solvent B), 50 min (60% solvent B), 53 min (100% solvent B), 60 min (100% solvent B), 61 min (11% solvent B), and 65 min (11% solvent B). Prior to injection, each sample was centrifuged at 23,000 g for 5 min, filtered through a Phenex 0.45-μm nylon syringe filter (Phenomenex), and stored under nitrogen until analysis. The injection volume was 10 μL. The compounds were identified according to their retention times and spectral properties. Gallic acid, (+)-catechin, and (−)-epicatechin were quantified at 280 nm from external standard curves of authentic standards. Hydroxycinnamates were collectively estimated at 316 nm (for peaks having absorption maxima at 316 nm) and expressed as mg caffeic acid equivalents/L by comparison with an external standard curve of caffeic acid. Flavonols were collectively estimated at 365 nm (for peaks having characteristic absorption maxima at 365 nm) and expressed as mg rutin equivalents/L by comparison with an external standard curve of rutin hydrate. Anthocyanins were quantified at 520 nm and expressed as mg malvidin-3-glucoside equivalents/L by comparison with a standard curve of malvidin-3-glucoside hydrochloride.
Analysis of tannins by protein precipitation.
Samples from the stability study were analyzed by the precipitation method as described (Harbertson et al. 2003) with a few modifications due to equipment limitations. Prior to analysis, the samples were diluted in a model wine solution of 12% v/v ethanol with 5 g/L tartaric acid, which had been adjusted to pH 3.3 with sodium hydroxide. The tannin-protein precipitate was formed by mixing 0.5 mL diluted wine and 1 mL BSA solution (containing 1 mg BSA/mL dissolved in a buffer of aqueous 0.2 M acetic acid and 0.17 M sodium chloride adjusted to pH 4.9) for 30 min. The precipitate was centrifuged at 14,000 g for 5 min to form a pellet and the supernatant was discarded. The pellet was washed with 0.25 mL of the pH 4.9 buffer, centrifuged at 14,000 g for 3 min, and the supernatant was discarded. The washing step was repeated one time. The rinsed pellet was redissolved in 1.5 mL buffer containing 5% w/v sodium dodecyl sulfate and 5% v/v triethanolamine by gentle mixing for 20 min. The background was measured as the absorbance at 510 nm of 1 mL redissolved solution. The sample was then mixed with 125 μL 11.4 mM ferric chloride in 11.4 mM aqueous HCl, and final absorbance at 510 nm was recorded after 10 min. The tannin was calculated as 1.125 times the final absorbance minus the background absorbance. The tannin concentration was expressed as mg catechin equivalents (CE)/L from a standard curve of the color reaction between catechin and ferric chloride.
Statistical data analysis.
The full factorial designs in experiment 1 were fitted to a linear model accounting for main and interaction effects using SAS JMP software (version 5.1, SAS Institute, Cary, NC). The response levels yi for all i observations were estimated in a linear model of the three factor levels (x1, x2, and x3) accounting for main and interaction effects (equation 3):
(eq 3)
The best linear model was then found by multiple linear regression to minimize the sum of squares of the residual values εi by the method of least squares (Montgomery 2001). The effect of each factor combination was estimated using mean-centered factor levels scaled between −1 and 1 to allow direct comparison of βvalues. Significances of differences were accepted on a 95% confidence level (p < 0.05).
Results and Discussion
The published “total” extraction protocol includes one extraction step only (Iland et al. 2004). To establish the benchmark for assessing the extent of phenols extraction from the different grape cultivars, the yields of total phenols and anthocyanins obtained in individual extraction steps were evaluated when repeating the “total” extraction procedure. The data obtained showed that almost all anthocyanins were removed during the first extraction, as the average yields of the second extraction on the different grape varieties were less than 3% of the sum of the first and second extractions (Table 2⇓). The yields of the third and fourth extractions were essentially nothing (data not shown). In contrast, 3.8 to 14.4% of the total phenols (the percentages calculated from the sum of the first and second extraction) were removed in the second extraction. For this reason, we decided to modify the published method by using the sum of the first and second extraction of the anthocyanins and total phenols as the benchmark “total.”
Experiment 1. Rapid solvent extraction of phenols and anthocyanins.
The mean extraction efficiency from the factorial design experiment evaluating extraction temperature (20, 40, or 60°C), ethanol concentration (0, 25, or 50% v/v), and hydrochloric acid concentration (0 or 0.1 M) ranged from 33 to 99% for total phenols and from 42 to 100% for anthocyanins (Table 3⇓). The relative standard deviations across the nine samples ranged from 4.7 to 17.1% for total phenols and from 1.7 to 8.6% for anthocyanins.
As expected, wide variations in the responses were recorded with the different extraction treatments. The three main factors of extraction temperature, ethanol concentration, and hydrochloric acid concentration exerted a significant influence on the extraction efficiency of both anthocyanins and total phenols (p < 0.0001) (Table 4⇓). Comparison of the estimated β parameters signified that ethanol had the greatest influence on the extent of extraction achieved (β = 12.8 to 18.8), while the influence of hydrochloric acid (β = 6.7 to 8.0) and extraction temperature (β = 5.5 to 6.6) were of lower magnitude. No interaction effects were significant for total phenols, but for anthocyanins significantly negative interaction effects (β = −1.8 to −5.5) were recorded for all pairwise combinations. For total phenols, the increase in the extraction efficiency by acidifying the solvent was 10 to 15% irrespective of temperature and ethanol concentration (Figure 2A⇓), while anthocyanin extraction increased between 5 and 35% with acidification (Figure 2B⇓). Moreover, the extraction of anthocyanins appeared to be more affected by acidification at the lower ethanol concentrations and lower temperatures.
During one-way analysis of variance (ANOVA), mean relative standard deviations were significantly lower for anthocyanins than for total phenols ( p < 0.0001). There was also a significant, negative correlation between mean extraction efficiency and relative standard deviation for anthocyanins (r = −0.730, p < 0.0001) and total phenols (r = −0.801, p < 0.0001) (Figure 3⇓). However, the correlation for anthocyanins was not significant for samples in which >80% of the anthocyanins were extracted. These findings confirm that to make the method robust, a desirable protocol yields high extraction efficiency, especially with respect to total phenols. The factor combinations with 50% ethanol and 0.1 M hydrochloric acid at 40 and 60°C showed high extraction efficiency and low variation across the different samples: at 40°C, 91.7% total phenols (rel SD 6.5%) and 97.6% anthocyanins (rel SD 2.9%), and at 60°C, 98.7% total phenols (rel SD 4.7%) and 99.9% anthocyanins (rel SD 3.7%). These two treatments therefore have good potential to meet the dual needs of high extraction efficiency and a robust extraction method.
Effect of solvent contact time.
The effect of solvent contact time (15, 30, and 150 min) was tested for Alicante grapes. No significant difference in extraction after 15 and 30 min could be established for either total phenols (p = 0.546) or anthocyanins (p = 0.483), but for 15, 30, and 150 min there was a significant, negative effect of solvent contact time on both total phenols ( p < 0.0001) and anthocyanins (p < 0.0001). In a supplementary experiment on Alicante grapes, we found a significantly negative effect of further increasing hydrochloric acid in the solvent from 0.1 M to 0.2, 0.5, or 1 M on the extraction of total phenols, but no significant effect on extraction of anthocyanins (data not shown).
Effect of homogenization time and solvent contact time.
Since short extraction time was one goal of this research, the potential for using even shorter solvent contact time for extraction of phenols was further investigated. A full factorial design varying the homogenization time (0.5, 1, 3, and 4 min) and the solvent contact time (0, 2, 5, 15, and 30 min) on a single cultivar (Alicante) was carried out using constant conditions of 40°C, 25% v/v ethanol, and 0.1 M hydrochloric acid.
The data from this factorial design were difficult to fit in a linear model accounting for both main and interaction effects. The R2(total phenols) was 0.67 and the R2(anthocyanins) was 0.61, likely due to a nonlinear response at low solvent contact times. Due to the imprecision of the linear models, responses were configured as contour plots (Figure 4⇓). These plots clearly show that, under the given conditions, the major part of total phenols and anthocyanins are extracted within approximately the first 5 min of extraction and the first minute of homogenization.
Experiment 2. Time-optimized extraction conditions.
The time-optimized extractions were conducted at 40°C using preheated 50% v/v aqueous ethanol with 0.1 M hydrochloric acid and gave mean extraction efficiencies of 93.5% for total phenols and 98.9% for anthocyanins after a 5-min extraction (Table 5⇓). After 15 min of extraction, an average of 95.9% total phenols and 100.5% anthocyanins were obtained. Even though extraction yields of both total phenols ( p < 0.001) and anthocyanins (p < 0.01) were significantly increased by the longer solvent contact time, the average increase amounted to less than 2.5%. The relative standard deviations across the different cultivars at 5.9% for total phenols and 4.6% for anthocyanins after a 5-min extraction were almost unaffected by the solvent contact time. Hence, the high extraction efficiencies found after only 5 min of solvent contact were acceptable for a fast extraction protocol. We additionally analyzed the amount extracted in the grape homogenate before solvent extraction and found mean extraction efficiencies of only 35% total phenols and 41% anthocyanins, both with high relative standard deviations across cultivars (17.0% for total phenols and 11.5% for anthocyanins).
In several instances, extraction efficiencies exceeded 100%, which resulted from high turbidity in many of the fast extracts (abs(700) averaged 2.6% and 6.0% of abs(280) and abs(520), respectively), probably caused by insufficient sample clarification. Therefore, extraction efficiencies in Table 5⇑ were to some extent overestimated.
Experiment 3. Final extraction protocol.
To avoid potential risk of perturbing the phenolic profile with acidic conditions and because very acidic samples may be incompatible with some analytical methods, a neutralization step was included as a postextraction treatment. The neutralization encompassed addition of sodium hydroxide right after the 5-min extraction. In addition, to overcome the turbidity problems from experiment 2, a filtration step was introduced and a higher centrifugation speed was used prior to dilution with hydrochloric acid. This resulted in acceptable low turbidities, where the abs(700) amounted on average to 0.6% and 1.3% of the abs(280) and abs(520), respectively (data not shown). The 5-min extractions conducted at 40°C using preheated 50% v/v aqueous ethanol with 0.1 M hydrochloric acid, followed by acid neutralization and sample clarification, gave mean extraction efficiencies of 81.8% total phenols and 91.5% anthocyanins (Table 6⇓). The relative standard deviations of the extraction efficiencies were 6.0% for total phenols and 3.8% for anthocyanins, which shows that the optimized extraction protocol was robust across different grape cultivars.
Stability of phenolics during prolonged solvent contact.
To assess the robustness of grape phenolics to contact with acidified solvent, an evaluation of the effect on the phenolic profiles and tannins of extended solvent contact for 5, 10, and 30 min was conducted for a Merlot grape extract (Table 7⇓). From the phenolic profiles obtained, it was not possible to discern any significant changes in the concentrations of the different compounds (Figure 5⇓). An examination of the integrated data indicated that the total concentrations of the main compounds (gallic acid, catechins/flavan-3-ols, hydroxycinnamates, flavonols, and acylated and nonacylated anthocyanins) remained significantly constant during the extended solvent contact (Figure 5⇓, Table 7⇓). The total concentrations of tannins were also constant during the initial 5 min of solvent contact, but decreased by ∼9% during the longer solvent contact periods. These results thus clearly demonstrate that the phenols were generally stable during the extraction protocol.
Conclusions
In statistically planned experiments, extraction temperature and the concentrations of ethanol and hydrochloric acid were found to significantly affect extraction of total phenols and anthocyanins from grapes. An optimized extraction protocol, giving high average extraction efficiencies of 81.8% total phenols and 91.5% anthocyanins from grapes, was obtained with only 5 min of solvent contact time using 50% v/v aqueous ethanol with 0.1 M HCl at 40°C and a 1:1 w/v grape homogenate:solvent ratio, followed by acid neutralization and sample clarification. The relative standard deviations of the extraction efficiencies across eight grape cultivars were 6.0% for total phenols and 3.8% for anthocyanins, which corroborated the robustness of the protocol across different grape varieties.
Footnotes
Acknowledgments: Laboratories Dubernet and Cabinet d’Ingénieurs Conseils en Viticulture are gratefully acknowledged for collecting the grapes.
- Received November 2006.
- Revision received May 2007.
- Copyright © 2007 by the American Society for Enology and Viticulture