Abstract
The anthocyanin concentration of Vitis vinifera L. cvs. Cabernet Sauvignon, Merlot, Syrah, and Monastrell grape skins was determined together with anthocyanin extractability at the exact time of harvest (measured by an extractability assay based on the comparison of the anthocyanin concentration of two different solutions obtained after macerating the grapes for four hours at two different pH values). These data were compared with the anthocyanin concentration and chromatic characteristics of the resulting wines. Monastrell grapes from the Jumilla area had the highest anthocyanin concentration (whether expressed as μg/g or mg/kg of berries), but the extractability assay indicated the difficulty of their extraction. The extractability assay for Cabernet Sauvignon, Merlot, and Syrah indicated that their anthocyanins could be extracted easily, which was confirmed by the chromatic analysis of the resulting wines. Cabernet Sauvignon and Syrah wines presented the highest wine color intensity, although their anthocyanin concentration in the grapes was lower than that of Monastrell grapes. Results indicated that the above extractability assay could be useful for predicting some important chromatic wine parameters and for planning the fermentation process according to the characteristics of the grapes at the exact time of harvest and the desired wine.
Anthocyanins are the compounds responsible for wine color. They are located in the cells of the skin of red grape varieties (within vacuoles) and they diffuse into the must and wine during maceration. Evidently, skin cell walls are barriers for these compounds. These cell walls are comprised of complex polysaccharides called pectins as well as cellulose and hemicellulose and they provide structure and firmness to the cells and the fruit. The pectins are concentrated in the middle lamella walls between cells. Extraction requires that the middle lamella walls be degraded for the cells to be released and that the cell walls be broken to allow their vacuole contents to be extracted or to diffuse into the wine (Amrani Joutei et al. 1995).
Highly colored grapes do not necessarily produce highly colored wines, which may be related to the ease with which anthocyanins are extracted from grape skins into musts. A phenolic extractability assay was proposed by Saint-Criq et al. (1998) to evaluate this extractability. This assay consists of macerating the grapes for four hours at two different pH values (3.2 and 1.0), and it is based on the assumption that at pH 1.0 complete disorganization of the vacuolar membrane facilitates the release of phenolic compounds. When the pH of the macerating solution is 3.2, the natural degradation of the cells is respected, a situation similar to that occurring during maceration in winemaking (Glories and Saucier 2000). Extractability is considered optimal when the difference between these two results is small and, therefore, the extractability index is low.
According to Gonzalez-Neves et al. (2004), knowledge of the extractability of grape pigments allows for the management of red wine fermentation and prediction of the color of wines. We have measured the anthocyanin content of the skins of four varieties of grapes (one variety cultivated in two different locations) and the extractability of the anthocyanins and tannins. The extent to which the anthocyanins were transferred from skins into wine, the chromatic characteristics of the wines, and the anthocyanin profile of crushed skins were evaluated to ascertain whether the extractability assay may help clarify why highly colored grapes do not always produce highly colored wines and whether the assay could be useful for determining red wine fermentation conditions (duration, temperature, intensity and frequency of pump-over) according to the type of wine desired.
Materials and Methods
Grapes from Vitis vinifera L. cvs. Cabernet Sauvignon, Syrah, Merlot, and Monastrell were harvested in 2003 from a commercial vineyard in Jumilla (southeastern Spain). A second sample of Monastrell grapes was harvested in a commercial vineyard in Bullas (southeastern Spain). Grapes were carefully harvested in 20 kg boxes and transported to the winery. Three 1.0 kg samples from each variety were collected at random from several clusters for the analytical determinations.
Vinifications.
The grapes were crushed and destemmed and then sodium metabisulfite was added (8 g of SO2 100 kg−1 of grapes). Total acidity was corrected to 5.5 g of tartaric acid L−1 and selected yeasts were added (Fermirouge, DSM, Servian, France) (10 g of dry yeast 100 kg−1 of grapes).
Maceration was conducted in 100-L stainless steel tanks (in triplicate) and lasted 15 days. The cap was punched down twice a day throughout the skin contact period. At the end of this period, the wines were pressed at 1.5 bar in a 75-L tank membrane press. Free-run and press wines were combined and stored at room temperature. One month later, the wines were racked. After malolactic fermentation had occurred, the wines were racked again and sulfur dioxide was added.
Anthocyanin monoglucosides in berry skins, crushed skins, and wines.
Grapes were peeled with a scalpel and the skins were stored at −20°C until analysis. Macerated skin samples were taken at the end of the maceration period from the press. The macerated skins were pressed in a manual press to eliminate the remaining pulp and stored at −20°C until analysis. Samples (5 g) were immersed in methanol (50 mL) in hermetically closed tubes and placed on a stirring plate at 150 rpm and 25°C. After 14 hours, the methanolic extracts were filtered through 0.45 μm, and analyzed by high-performance liquid chromatography (HPLC). Samples of wine were filtered through 0.45 μm and directly analyzed by HPLC.
Identification and quantification of anthocyanins.
HPLC analyses were performed on a Waters 2690 liquid chromatograph (Waters, Milford, MA), equipped with a Waters 996 diode array detector and a LichroCart RP-18 column (Merck, Darmstadt, Germany), 25 x 0.4 cm, 5-μm particle size, using as solvents water plus 4.5% formic acid (solvent A) and HPLC grade acetonitrile (solvent B) at a flow rate of 1.5 mL min−1. Elution was performed with a gradient starting with 10% B to reach 15% B at 25 min, 21% B at 65 min, and then isocratic for 3 min. Chromatograms were recorded at 520 nm.
Identification of the compounds was carried out by comparison of their UV spectra recorded with the diode array detector and those reported in the literature. An HPLC-MS analysis was conducted to confirm the identity of each peak using a LC-MSD-Trap VL-01036 liquid chromatograph-ion trap mass detector (Agilent Technologies, Waldbronn, Germany), equipped with electrospray ionization (ESI). Elution was performed in the HPLC analysis conditions described above, with a flow rate of 1.0 mL min−1. The heated capillary and voltage were maintained at 350°C and 4 kV, respectively. Mass scans were measured from m/z 100 up to m/z 800. Mass spectrometry data were acquired in the negative ionization mode. Anthocyanins were quantified at 520 nm as malvidin-3-glucoside, using malvidin-3-glucoside chloride as external standard (Extra-synthèse, Genay, France).
Physicochemical determinations in grapes.
Total soluble solids (Brix) were measured using an Abbé-type refractometer. Titratable acidity and pH were measured using an automatic titrator (Metrohm, Herisau, Switzerland) with 0.1 N NaOH. Tartaric and malic acids were measured using enzymatic kits from Boehringer Mannheim GmbH (Mannheim, Germany).
The phenolic potential of grapes was calculated according to the method described by Saint-Criq et al. (1998), macerating the grapes for 4 hr at two pH values (3.6 and 1.0). The original pH 3.2 solution was exchanged for one of pH 3.6, which is better suited to the musts from the Jumilla region. The anthocyanin contents of the two solutions (ApH3.6 and ApH1) were then chemically assayed by measuring the absorbance of the samples at pH 1.0, while the total phenol content was calculated by measuring the optical density of the solution at pH 3.6 at 280 nm. The phenolic potential was calculated as follows:
Seed and skins tannin concentration was evaluated using a protein precipitation assay. Sample preparation and the protein precipitation assay were conducted according to the methods described by Harbertson et al. (2002).
Color determinations in wines.
Wines were analyzed at the end of the maceration period, at the end of alcoholic fermentation, and when malolactic fermentation was finished. Absorbance measurements were made in a Helios Alpha spectrophotometer (Thermo Electron, Waltham, MA) with 0.2-cm path-length glass cells. Color intensity (CI) was calculated as the sum of absorbance at 620 nm, 520 nm, and 420 nm (Glories 1984), and tint was calculated as the ratio between absorbance at 420 nm and at 520 nm (Sudraud 1958). Other variables calculated were red (%R), yellow (%Y), and blue (%B) percentages, according to Glories (1984). Total phenol content in wine (TPwine) was estimated by measuring the optical density of the wine at 280 nm (Ribéreau-Gayon et al. 1998).
To estimate the copigmented anthocyanin concentration in wines, samples were adjusted to pH 3.6 and membrane-filtered (0.45-μm pore size). The following measurements were made, according to Levengood and Boulton (2004):
Aacet: 20 mL of 10% acetaldehyde solution was added to 2.0 mL of wine sample in a 10-mm plastic cuvette. After 45 min, the sample was placed in a 2.0-mm cuvette and absorbance was measured at 520 nm. The reading was corrected to 10-mm path length by multiplying by 5.
A20: 100 μL of wine sample was placed into 1900 μL of a buffer solution in a 10-mm cuvette (24 mL pure ethanol was added to 176 mL distilled water and 0.5 g of potassium bitartrate was dissolved into the solution and the solution pH was adjusted to 3.6 with HCl or NaOH as needed). After several minutes the absorbance was measured at 520 nm. The reading was corrected for the dilution by multiplying by 20.
ASO2 : 160 μL of 5% SO2 solution was added to 2.0 mL of wine sample in a 10-mm cuvette. The absorbance was measured at 520 nm.
The following calculations were made: fraction of color because of copigmentation (%CA): (Aacet - A20)/Aacet; fraction of color due to free anthocyanins (%FA): (A20 - ASO2)/Aacet; and fraction of color due to polymeric pigment (%PA): ASO2/Aacet.
Statistical data treatment.
Significant differences among samples and for each variable, together with the correlation analysis, were analyzed using StatGraphics 3.1 Plus (Statistical Graphics Corp., Rockville, MD).
Results and Discussion
The physicochemical data of the grapes at the moment of harvest are shown in Table 1⇓. Berry weight was significantly higher in the Monastrell grapes and particularly in the Monastrell from the Bullas area (Monastrell-B), reaching 167 g/100 berries. The smallest berries were found in Merlot and Syrah grapes. The size of grapes is of fundamental importance for the quality of wines. Anthocyanins are synthesized in the skin, and a larger berry weight results in a lower skin-to-flesh ratio since the compounds are diluted. It is therefore important to consider the size of the berry when expressing anthocyanin concentration.
The anthocyanin concentration in grape skins, expressed as μg/g of skin or mg/kg of berry (thus taking into account the effect of the different berry sizes), and the percentage of each anthocyanin are shown in Table 2⇓. When the results were expressed as μg/g of skin, the highest values were observed in Monastrell grapes from Jumilla (Monastrell-J) and Syrah; Cabernet Sauvignon and Merlot grapes had the lowest concentration. Even when results were expressed as mg/kg of berries, the highest values were still reached in Monastrell-J and Syrah, although in this case the differences with the other varieties were lower. Monastrell grapes (from both locations) had the lowest proportion of acylated anthocyanins, as was also observed by García-Beneytez et al. (2002). It is commonly accepted that the anthocyanin composition of a given cultivar is closely linked to its genetic inheritance and, from a qualitative point of view, is quite independent of seasonal conditions or production area (Fernandez-Lopez et al. 1998). However, some small differences were detected between Monastrell-J and Monastrell-B, but it has also been shown that even two clones from the same variety may present small differences in anthocyanin qualitative composition and that seasonal conditions may have an influence (Arozarena et al. 2002).
Cabernet Sauvignon and Syrah grapes had the highest percentages of acylated anthocyanins, the latter with a high percentage of coumarate derivatives, coincident with the results of Darne (1988) and Fernandez-Lopez et al. (1995). The presence of acylated anthocyanins can be very important for wine color since they can participate in intramolecular copigmentation processes, increasing the wine color.
Results of the extractability assay are shown in Table 3⇓. The anthocyanins extracted at pH 1.0 were found in the highest concentration in Monastrell-J and Syrah grapes, while the lowest values were found in Merlot grapes. Although the results of these spectrophotometrically measured anthocyanins do not coincide with those determined by HPLC (because of the different extraction procedures), the trend for anthocyanin concentration in the different grape samples was similar. When the macerating solution was at pH 3.6, the highest anthocyanin concentration was found in Syrah and Cabernet Sauvignon grapes, which now had a higher anthocyanin concentration than Monastrell-J, while Monastrell-B showed the lowest concentration. The anthocyanin values obtained at pH 1.0 and 3.6 led to a relatively high value of the extractability index for Monastrell grapes in both locations and lower but similar values for Syrah, Cabernet Sauvignon, and Merlot grapes. These values would indicate, according to Saint-Criq et al. (1998), that the anthocyanins from Monastrell skins are more difficult to extract than those from the other varieties. Two factors could be responsible for the higher extractability index of Monastrell grapes: first, their cell walls are genetically characterized by a more rigid structure, which would hinder anthocyanin extraction, and second, Monastrell grapes presented lower Brix than Merlot, Cabernet Sauvignon, and Syrah when harvested, as shown in Table 1⇑. Some authors have maintained that the values of the extractability index decrease as maturation progresses (Saint-Criq et al. 1998, Glories 1999). However, we found in previous studies that Monastrell grapes always had a high extractability index and that, when studying the evolution of this parameter during ripening, the sugar content changed from 24.4 to 26.2 Brix, with no significant changes in the extractability index (Bautista-Ortin et al. 2004). Similarly, Gonzalez-Neves et al. (2002) found that there was no predetermined relationship between phenolic maturity and technological maturity.
The seed maturity index (SMI) represents the percentage of seed tannins that will contribute to the tannin content of a wine. It has been reported that this value decreases during ripening and that excessively high values indicate that seed tannins could easily be transferred to wine, especially if the pomace contact time was prolonged during vinification (Ribéreau-Gayon et al. 1998). The results showed that the highest values were reached in Merlot grapes. A protein precipitation assay was used to check the actual content of skin and seed tannins for all the studied grapes, with results expressed as berry content or concentration (mg/kg) (Table 4⇓). For skin tannins, Monastrell grapes had the highest content per berry and Merlot and Syrah the lowest. When expressed as concentration, Monastrell-J and Cabernet Sauvignon showed the highest values. The differences in seed tannins (mg/berry) between varieties were not great but, when expressed as mg/kg, Cabernet Sauvignon and Merlot grapes reached the highest tannin concentration, these varieties also showing the highest seed maturity index values.
The studied grapes were vinified and their chromatic parameters measured. Wine-making conditions were exactly the same, in contrast to the work of Gonzalez-Neves et al. (2004), in which the duration of maceration was decided as a function of the extractability assay results. We wanted to check whether grape characteristics, including the extractability indices, could be correlated with the chromatic characteristics of their wines when the same winemaking conditions were followed. Anthocyanin monoglucoside concentration in wines after the maceration period is shown in Table 5⇓. Monastrell-J, Syrah, and Cabernet Sauvignon presented similar values, even when different concentrations were found in the grape skins of the different varieties. The proportion of non-acylated anthocyanins was higher in Monastrell wines. It has been stated that the anthocyanin fingerprint only partially reflects the anthocyanin fingerprint of fresh grapes (García-Beneytez et al. 2002), wines usually containing a higher proportion of malvidin-3-glucoside than grapes. However, monoglucoside composition alone may not allow any clear conclusions on wine chromatic characteristics, as some anthocyanins could have been extracted from the skins and been polymerized, thus contributing to wine color.
The spectrophotometric characteristics of the wines are shown in Table 6⇓. As in the grapes, total phenols in wine (TPwine) were higher in Cabernet Sauvignon, Syrah, and Merlot wines, indicating that more phenolic compounds had been extracted into the wines. Color intensity reached its highest values in Syrah and Cabernet Sauvignon, followed by Monastrell-J. The Monastrell-B grapes, with a higher anthocyanin concentration than Merlot grapes, produced a wine of similar color intensity to Merlot but lower total phenol content. Since copigmentation percentages were very similar in Monastrell-J, Cabernet Sauvignon, and Syrah wines, as was the monoglucoside concentration (as shown in Table 5⇑), the differences in color intensity must have been due to polymeric anthocyanins, and the extent of polymerization is confirmed with the values of ASO2.
Merlot and Cabernet Sauvignon wines had the highest tannin concentration, but the chromatic characteristics of Merlot wines showed a high tint and yellow percentage. Merlot wines, with a relatively low anthocyanin concentration and high tannin concentration, may present an unbalanced anthocyanin/tannin ratio that led to the high yellow percentage.
As the wines evolved, color decreased slightly, probably because of the adsorption of pigmented compounds on yeast lees and coprecipitation. At the end of malolactic fermentation, no statistically significant differences were observed in Monastrell-J, Cabernet Sauvignon, and Syrah color intensity—the wines that showed the highest color intensity.
A study of the anthocyanins remaining in the crushed skins after maceration might also help to confirm the extractability index as a tool for measuring the ease of anthocyanin extraction in grapes (results shown in Table 7⇓). The crushed skins from Monastrell grapes showed the highest concentration of anthocyanins, and the fact that the percentage of remaining anthocyanins with respect to the original grape levels and the anthocyanin profile were very similar for both Monastrell-J and Monastrell-B may be related to the high and similar values of the extractability index. In the other varieties, the percentage of anthocyanins remaining in the crushed skins was lower than 12% of their initial concentration in the intact grapes, and malvidin-3-glucoside was the only nonacylated monoglucoside detected. Malvidin acetates were high in Cabernet Sauvignon, Syrah, and Merlot crushed skins, although somewhat lower than in grapes, and malvidin coumarates were also in high concentration.
Correlations between grape and wine characteristics are shown in Table 8⇓. The correlations between the HPLC-measured grape anthocyanins and the wine chromatic characteristics were in the same range as found for the anthocyanin content in the pH 1.0 solution (ApH1), high correlation coefficients not being observed with any of the wine parameters. On the contrary, high correlation coefficients were observed between ApH3.6 and some of the wine chromatic characteristics, being higher than 0.85 with Aacet, A20, and ASO2 (measurements directly related with wine anthocyanin content). The extractability index and the seed maturity index showed the highest correlation coefficients with TPwine (−0.92) and wine tannins (0.69), respectively. The total phenol content of the pH 3.6 solution (TP pH 3.6) was highly correlated with TPwine and wine tannins. The skin tannin content did not correlate with any of the wine parameters, whereas seed tannin was highly correlated with wine tannins. A study by Harbertson et al. (2002) did not reveal a clear relationship between the tannin concentration in grape berries at harvest with the tannin concentration in wines, but their wines were made commercially, without any attempt to standardize the winemaking practices. In our study, the wine-making practices were the same for all the grapes and, in these conditions, grape seed tannin concentration was highly correlated with wine tannin concentration.
Conclusions
Although Monastrell grapes from Jumilla presented the highest anthocyanin concentration (expressed both as μg/g and mg/kg of berries), the extractability index was high, indicating the difficulty of their extraction. Effectively, the concentration of anthocyanins and the color intensity in the respective wines did not correlate with the total anthocyanins observed in grapes (measured by HPLC or in the solution at pH 1.0) but with those extracted at pH 3.6. Therefore, our results indicate that the extractabililty index (%) points to the ease and extent of the anthocyanin transfer from grape skins to wine and that the measurement of ApH3.6, TP (pH 3.6), and seed tannin content could be used to predict key wine chromatic parameters.
Thus, consideration of the phenolic potential of grapes at the moment of harvest could allow fermentation (duration, frequency and intensity of pump-over, temperature control, enzymes) to be planned as a function of the characteristics of the grapes at that time and the wine type desired. The high extractability index value indicates that Monastrell grapes require long maceration periods for their anthocyanins and tannins to be extracted. Frequent pump-over would be necessary to extract tannins that could favor anthocyanin-tannin polymerization. Merlot grapes did not have high anthocyanin concentrations, but had a low extractability index and high seed tannin concentration. Based on that profile, Merlot grapes require short maceration periods since anthocyanins would be easily extracted, although an excess of tannins would increase the percentage of yellow color and the astringency of the wines. Syrah grapes could be favored with frequent pump-over for increasing the tannin concentration in finished wines, especially if the wine is destined to be aged. Cabernet grapes had a low extractability index and high tannin concentration, and good results could be expected when used for both short macerated wine (young wines) or long macerated wines (wines destined to be aged).
Footnotes
Acknowledgments: This work was made possible by financial assistance of the Ministerio de Ciencia y Tecnología, project AGL2003-01957.
- Received November 2004.
- Revision received February 2005.
- Revision received March 2005.
- Copyright © 2005 by the American Society for Enology and Viticulture