2019 Volume 25 Issue 6 Pages 863-869
Anthocyanins and tannins were extracted from the skins of Cabernet Sauvignon (CS) and Muscat Bailey A (MBA) grapes. Insoluble cell wall materials (IC) were prepared from the skin, pulp, and seeds of CS and MBA. The adsorption properties of anthocyanins and tannins by IC were analyzed in a model solution.
In CS, skin IC had the potential to adsorb 12% of skin total anthocyanins, and pulp and seed IC each adsorbed 4%. In MBA, skin IC had the potential to adsorb 20% of skin total anthocyanins, pulp IC, 5%, and seed IC, 3%. In both CS and MBA, skin IC also had the potential to adsorb 30% of skin tannins. In addition, pH (2.5 to5.0) and ethanol concentration (0 to 14%) affected skin tannin adsorption. Anthocyanin adsorption was maximum at pH 3.2 in CS and pH 3.0 in MBA, and minimum at pH 3.6 in CS and pH 3.4 in MBA. Tannin adsorption was maximum at pH 3.3 and minimum at pH 4.2 in both cultivars. On the other hand, both skin anthocyanin and tannin adsorption decreased with increasing ethanol concentration in both MBA and CS. These results suggest that must pH may affect the extraction and adsorption of phenolics.
Anthocyanins and tannins are major groups of phenolics in red wine and are responsible for the color and taste of red wine. The importance of these phenolics is recognized by not only winemakers but also consumers, and many people pay great attention to the concentrations of these compounds. The concentrations of anthocyanins and tannins have a very strong positive relationship with wine grade (price) in Shiraz wine, and their correlation coefficients (r2) are 0.918 and 0.818, respectively (Kassara and Kennedy, 2011). High-grade wines have a high skin tannin concentration or ratio relative to seed tannin concentration (Kassara and Kennedy, 2011). Based on this background, it is very important to elucidate the mechanisms underlying phenolics extraction to make high-quality red wines. Many attempts have been made to increase anthocyanin and tannin concentrations during maceration, including extended maceration, thermo-vinification, cold maceration, and so on. However, the extraction mechanism of anthocyanins and tannins from skin and/or seed into the liquid portion is very complicated. Phenolics are usually located in the vacuoles of skin and seed cells, and thus should be moved from the vacuoles to the liquid portion while avoiding materials that have chemical or physical affinity toward phenolics, like cell wall materials (CWM) during extraction. Bindon et al. (2010) advocated an extraction model for tannin adsorption to and desorption from CWM.
Because of the high temperature and heavy rainfall in the harvest season, Japanese red wine tends to show thinning of the concentrations of phenolics, and therefore, increasing phenolics concentrations is a crucial issue to be resolved. Indeed, tannin concentrations were much lower in Japanese Cabernet Sauvignon (CS) and Merlot wines than in those made in the USA and other countries (Harbertson et al., 2008, Ichikawa et al., 2011). The most widely produced red wine in Japan, which is red wine made of Muscat Bailey A (MBA), a hybrid of Bailey x Muscat of Hamburg (Fig. 1), has very low tannin concentration (65 mg/L on average) (Ichikawa et al., 2011). To determine the reason for the very low tannin concentration in MBA wine, comparisons of MBA and CS during maceration were made. The results showed that (1) MBA had very low tannin concentration seed, (2) skin tannin was extracted into the liquid portion but its concentration decreased drastically on day 5 of maceration for both MBA and CS, and (3) the same phenomenon was observed for anthocyanin concentration (Okuda et al., 2014). These results indicated that if the decrease of skin tannin and anthocyanin concentrations during maceration could be avoided, high-quality wine would be produced because these phenolics have been shown to be important for the taste and price of wines (Kassara and Kennedy, 2011)
Genealogical tree of Muscat Bailey A
Recently, some research has been conducted on the extractability of phenolics from skin using both chemical and physical approaches. If phenolics are bound tightly because of their high affinity with CWM, extractability would be lowered. Apolinar-Valiente et al. (2017) focused on the variation of sugar components and protein contents in skin cell walls of Monastrell (low extractability of phenolics) and its hybrid in comparison with CS, and found diversity in the cell wall composition. Very recently, Garrido-Banuelos et al. (2019) showed differences in skin cell wall components between vintage and ripeness and suggested the importance of pectin-rich components. Because proteins and polysaccharides can bind with tannin, cell wall composition may affect affinity toward tannin. Bindon et al. (2012, 2014) reported a series of studies on the interaction of tannin and CWM. They compared the specificity of origin tissue of CWM through precise analysis of cell wall composition, and suggested the importance of tannin molecular mass and porosity of the cell wall. CWM and anthocyanin binding affinity have also been reported (Padayachee et al., 2012). It was shown that anthocyanins bound with cellulose or cellulose-pectin composites, and suggested that both ionic and hydrophobic interactions might participate in these interactions.
Although the extractability of phenolics into wine and the re-adsorption of these phenolics to CWM are crucial issues, most of the studies described above were conducted in a medium or wine-like solution with controlled pH, ionic strength, etc., and little is known about the adsorption conditions of phenolics and CWM. In this study, we focused on the effects of pH and alcohol concentration on the binding reaction of phenolics and CWM, because these factors can differ between vinifications. As regards to CWM, some studies had dealt with both soluble and insoluble CWM (Bindon et al., 2016). In addition, it was reported that “insoluble-CWM (IC)” is important for tannin extraction in vinification (Hazak et al., 2005). Therefore, we focused on IC in this study.
Chemicals All reagents were analytical grade (or HPLC grade) unless otherwise specified. Folin-Ciocalteu (F-C) reagent and bovine serum albumin (fraction V) were purchased from FUJIFILM Wako Pure Chemical Corp. (Osaka, Japan).
Grape materials CS and MBA, both seven years old and grafted on Teleki 5BB, were grown in the experimental vineyard of the University of Yamanashi (35°66′N; 138°58′E) and harvested at commercial maturity on 24th September 2015. All grapes were stored at −20 °C until use.
Extraction and isolation of grape skin anthocyanins and tannins The extraction and isolation of CS and MBA skin anthocyanins and tannins were accomplished according to the method of Kennedy and Jones (2001). Thirty kilograms of grapes were harvested and destemmed, and the berries were mixed and picked randomly to make 1 kg. The picked berries were separated into skin, pulp, and seed, the weights of which are shown in Table 1. Skin crude phenolics were extracted in Erlenmeyer flasks with 2:1 v/v acetone/water at room temperature for 24 hours. To avoid phenolics oxidation, the solution was blanketed with nitrogen gas carefully and the extraction was carried out in the dark. Following extraction, the extract was concentrated under reduced pressure with nitrogen gas at 35 °C to remove acetone, and the residue was lyophilized and powdered with a mortar and pestle. To separate the crude phenolics into anthocyanin (monomer) and tannin fractions, a Toyopearl® TSK HW-40F (Tosoh Corp., Tokyo, Japan) column was used. The column (30 mm i.d. × 150 mm) was equilibrated with MeOH/water/trifluoroacetic acid (50:50:0.1 v/v/v). The crude phenolics powder was dissolved in a minimum amount of this mobile phase and then applied onto the column. The column was eluted with three column volumes of the mobile phase to obtain the anthocyanin fraction. After that, the column was eluted with three column volumes of acetone/water/trifluoroacetic acid (66.6:33.3:0.1 v/v/v) to obtain the tannin fraction. Each eluate was concentrated under reduced pressure with nitrogen gas at 35 °C to remove organic solvents, and the residue was lyophilized and powdered with a mortar and pestle. The powdered phenolics were stored at −20 °C until use.
| CS | MBA | |||
|---|---|---|---|---|
| Fresh weight (g/kg berry) |
IC (dry g/kg berry) |
Fresh weight (g/kg berry) |
IC (dry g/kg berry) |
|
| Skin | 129.4±8.7 | 8.7±1.1 | 101.7±9.0 | 6.1±0.9 |
| Pulp | 829.2±24.1 | 2.7±1.4 | 871.8±19.2 | 4.6±1.5 |
| Seed | 41.4±3.6 | 10.4±0.8 | 26.5±5.5 | 6.1±1.8 |
Preparation of IC from grape skin and pulp The preparation of IC from grape skin was accomplished according to the method of Bindon et al. (2010). Frozen pulp was homogenized at 8000 rpm for 20 s (POLYTRON® PT 10–35 GT, Kinematica, Switzerland) to form a slurry, and the slurry was immediately added to an equivalent amount of 40 mM HEPES pH 7.0. The mixture was stirred for 15 min to solubilize water-soluble material. The mixture was centrifuged twice, and the residue was retained. The HEPES-extracted pulp material and untreated frozen grape skin were washed with 70% v/v acetone/water for 24 hours. To minimize phenolics oxidation, the solution was blanketed with nitrogen and the extraction was carried out in the dark. The acetone-treated pulp and skin were washed with an additional amount of 70% v/v acetone/water. After centrifugation (6500 × g for 10 min), the residual slurry was homogenized with a mechanical homogenizer (POLYTRON® PT 10–35 GT, Kinematica, Switzerland). Thereafter, grape pulp and skin IC were prepared according to the method of Vidal et al. (2001). Briefly, the acetone-insoluble residue was extracted with 150 mL of 0.2 M Tris-HCl (pH 6.7)-equilibrated phenol to remove proteins, and then washed twice with 80% v/v ethanol/water and three times with acetone to remove grape phenolics. Next, the residue was extracted by slow shaking for 30 min with 1:1 v/v methanol/chloroform. After air-drying in a fume hood, grape pulp and skin IC were lyophilized. Finally, each lyophilisate was ground into fine powder with a mortar and pestle and stored at −20 °C until use. All preparations were carried out in triplicate.
Preparation of IC from grape seed Dried grape seeds were washed with 70% v/v acetone/water for 3 days. Thereafter, the grape seeds were ground into a rough powder with a food grinder (National MX-X103-D, Osaka, Japan). The rough powder was washed several times with hexane, 70% v/v acetone/water, and 1:1 v/v methanol/chloroform. Next, the insoluble fraction was ground again into a fine powder with a mortar and pestle and stored at −20 °C until use. All preparations were carried out in triplicate.
Preparation of model wines with various pH and ethanol concentrations The basic components of the model wine solution were 5.0 g/L potassium bitartrate and 12% (v/v) ethanol, and pH was adjusted to 3.3 with HCl. In addition, different model wines with various pHs (2.5 to 5.0) and ethanol concentrations (0 to 14%) were prepared. pH was adjusted with HCl or KOH depending on the solution.
Analysis of total phenolics and tannins The analysis of total phenolics was based on the F-C method with the reaction scale of 1/10 (Singleton and Rossi, 1965). The analysis of tannins was carried out by the bovine serum albumin (BSA) precipitation method, as described by Harbertson et al. (2002).
Analysis of anthocyanin composition by HPLC The analysis of anthocyanins was carried out according to the method of Zanatta et al. (2005) with slight modifications. The anthocyanin fraction was analyzed by HPLC using the following conditions: eluent A: 0.4% H3PO4; eluent B: acetonitrile; flow rate: 1.0 mL/min; column: Atlantis C18-T3 (Waters Corp., MA, USA); column temperature: 40 °C; detection wavelength: 520 nm; gradient program: 0 to 1 min, 0% B; 1 to 40 min, 0 to 28% B; 40 to 47 min, 50 to 100% B; 47 to 57 min, 100% B; 57 to 67 min, 0% B. Peaks obtained by HPLC analysis were quantified on the basis of the malvidin-3-O-glucoside standard curve.
Adsorption of grape skin anthocyanins and tannins to grape skin, pulp and seed IC Grape IC (50 mg) and grape skin anthocyanins or tannins (10 mg) were mixed in model wines (10 mL). For skin IC, different pH values (2.9 to 4.2) and different ethanol concentrations (0 to 14%) (10 mL) were used in addition to the experiment using control model wine. The reaction was carried out at 32 °C for 1 hour (Bindon et al., 2010). After centrifugation (15,000 × g for 5 min), total phenolics, tannins, and anthocyanins in the supernatant were analyzed. All experiments were carried out in triplicate.
Statistical analysis Statistical analysis (ANOVA) was carried out using an add-in software for Microsoft Excel, Ekuseru-Toukei 2015 ver. 2.20 (SSRI Co., Ltd., Tokyo, Japan).
Recovery of insoluble polysaccharides from grape berry Fresh weights and IC (dry weight) of CS and MBA skin, pulp, and seed are shown in Table 1. Because CS berries were smaller than MBA berries, CS had higher skin and seed fresh weights than MBA (p < 0.05). The dry weights of skin and seed IC were also higher for CS than MBA. Pulp fresh weight accounted for more than 80% of berry weight; however, the amounts of IC obtained from pulp were less than those obtained from skin or seed in both cultivars.
Adsorption potential of IC toward anthocyanins Adsorption experiments of skin total anthocyanins by IC obtained from various grape tissues were carried out in model wine solutions for both cultivars. Based on the results in Table 1, the ratios of IC in skin, pulp, and seed to total anthocyanins in berry skin were calculated. Using these ratios, the adsorption potential of IC toward skin total anthocyanins was compared (Fig. 2).
Skin total anthocyanin adsorption rates of grape IC in model wine solution (pH 3.3, 12% ethanol). Bars show SD (n=3).
In CS model wine solution (pH 3.3, 12% ethanol), skin IC had the potential to adsorb 12% of skin total anthocyanins, and pulp IC and seed IC each adsorbed 4% of skin total anthocyanins. In MBA model wine solution (pH 3.3, 12% ethanol), skin IC had the potential to adsorb 20% of skin total anthocyanins, pulp IC, 5%, and seed IC, 3%. In total, CS IC had the potential to adsorb 20% of skin total anthocyanins and MBA IC could adsorb 28% of skin total anthocyanins. In this experiment, the total amounts of anthocyanins obtained were 911 and 782 mg/kg grape for CS and MBA, respectively. Skin tannins obtained were 2071 and 1255 mg/kg grape for CS and MBA, respectively. Mathematically, 34 and 43% of anthocyanin and 51 and 84% of skin tannin can be adsorbed by IC in grapes. However, other phenolics exist in grapes, such as seed tannin. Therefore, the realistic adsorption capacity of IC toward anthocyanin and skin tannin was expected to be lower than that obtained by calculation.
Comparing CS and MBA model wine solutions, MBA skin IC had about 8% points higher adsorption potential, whereas pulp and seed IC adsorbed less than 5% anthocyanins in both CS and MBA model wine solutions. The strong adsorption potential of MBA skin IC seemed to be the nature of MBA. HPLC analysis of anthocyanins using an ODS column showed that later eluted anthocyanins showed a higher adsorption rate (data not shown). This result suggested that hydrophobicity affects the adsorption of anthocyanin by IC. There are differences in anthocyanin composition between CS and MBA grape skin (Koyama et al., 2017), and these differences may affect the binding affinities with IC. In this study, however, structural identification of anthocyanins was not confirmed. Thus, conclusions regarding the effect of anthocyanin structure on adsorption could not be made.
In addition, Padayachee et al. (2012) mentioned that the exposure of cell walls, especially cellulose and pectin, affected anthocyanin adsorption in an experiment using purple carrots. Differences in IC compositions may also affect the adsorption potential of IC toward anthocyanins, although little is known about IC compositions in these two cultivars. In the experiment described here, IC was ground into a fine powder to measure the maximum adsorption potential toward anthocyanins. Therefore, the amount of anthocyanin adsorbed by IC obtained in our study may be more than that under actual winemaking conditions.
Adsorption potential of IC toward skin tannins In the same manner as the anthocyanin adsorption experiments, the ratios of IC in skin, pulp, and seed to total tannins in berry skin were calculated and the adsorption potentials of IC toward skin tannins were investigated. Skin tannin adsorption rates (%) of grape IC (skin, pulp, and seed) in CS and MBA model wine solutions are shown in Fig. 3.
Skin tannin adsorption rates of grape IC in model wine solution (pH 3.3, 12% ethanol). Bars show SD (n=3).
In both CS and MBA model wine solutions, skin IC had the potential to adsorb 30% of skin tannins. In contrast, seed IC adsorbed 18% of skin tannins in CS model wine solution and 45% of skin tannins in MBA model wine solution. Skin tannins have diverse molecular weights, and the composition and tendency of the mean degree of polymerization (mDP) differ depending on grape variety (Koyama et al., 2017). It was reported that high molecular weight tannins are trapped easily by IC (Bindon et al., 2014, Le Bourvellec et al., 2004), whereas we found that the adsorption potentials of IC differed depending on their origin. Not only the amount of tannin adsorbed but also the quality of absorbed tannin would differ depending on the origin of IC. The high adsorption potential of MBA seed IC may partially explain the uniquely low concentrations of tannins in MBA red wine (Ichikawa et al., 2011, Ichikawa et al., 2012, Okuda et al., 2014). In total, CS IC had the potential to adsorb 64% of skin tannins and MBA IC had the potential to adsorb 91% of skin tannins.
Effect of pH and alcohol concentration on anthocyanin adsorption by skin IC
Considering the important phenomenon of anthocyanin concentration reduction during red winemaking (Okuda et al., 2014) and the high adsorption potential of skin IC, as shown in Fig. 2, the effects of pH (Fig. 4A) and ethanol concentration (%) (Fig. 4B) on anthocyanin adsorption by skin IC were analyzed. The experiments were executed in the pH range of 2.5 to 5.0 and ethanol concentrations of 0 to 14%.
Effect of pH (A) and ethanol concentration (B) on anthocyanin adsorption by skin IC (▲: CS, ○: MBA). Ethanol concentration and pH were 12% (A) and 3.3 (B), respectively. Bars show SD (n=3).
Regardless of pH, the anthocyanin adsorption rates (%) of MBA skin IC were always higher than those of CS skin IC, and this finding was supported by the results in Fig. 2. Surprisingly, the anthocyanin adsorption rates varied with pH, peaking at pH 3.2 and around 4.0 for CS and at pH 3.0 and 4.4 for MBA (Fig. 4A), but the fluctuation in CS was not distinct. On the other hand, the anthocyanin adsorption rates were low at pH 2.6, 3.5, and 4.9 for both cultivars. Anthocyanins might be easily extracted under these pH conditions. These findings indicated that red wine color may vary with must pH. The same results were obtained from experiments carried out in two consecutive years (data not shown). Anthocyanins exist in several forms, and only flavylium has a positive charge that may affect adsorption by IC. Other forms of anthocyanins have different charge states that may affect their interaction with IC. Although the reason why maximum and minimum adsorption conditions exist in the reaction between anthocyanins and IC is not known, isoelectric points (pI) of proteins and/or acid dissociation constants (pKa) of carboxy groups in IC may affect the interactions of these compounds. Little is known about the chemical and physical properties of IC, and further study should be conducted. On the other hand, in the presence of various concentrations of ethanol, anthocyanin adsorption rates decreased with increasing ethanol concentration (Fig. 4B), suggesting that anthocyanin adsorption by skin IC may occur via hydrophobic interactions. Indeed, more hydrophobic anthocyanins (acylated anthocyanins) tended to be well adsorbed as described above, whereas more hydrophilic anthocyanins tended to be poorly adsorbed with increasing ethanol concentration (data not shown).
Effect of pH and alcohol concentration on tannin adsorption by skin IC The conditions for skin tannin adsorption by skin IC were also analyzed, focusing on pH (Fig. 5A) and ethanol concentration (Fig. 5B), in the same manner as the anthocyanin adsorption experiments. Skin tannin adsorption rates were strongly affected by pH, and a peak was noted at around pH 3.3. In the case of MBA, skin tannin adsorption rates increased again at pH higher than 4.2. On the other hand, skin tannin adsorption rates decreased with increasing ethanol concentration in both MBA and CS. This result is the same as that observed in the anthocyanin adsorption experiments, and skin tannin adsorption may occur via hydrophobic interactions. Indeed, Le Bourvellec et al. (2004) reported that the interaction of procyanidin and apple cell wall is caused by weak hydrogen bonding and hydrophobic interactions. They also mentioned the importance of the ionic strength of the medium. It should be noted that in our study, the ionic strength of the model wine solution was determined on the basis of tartrate concentration, and the effect of pH adjustment was negligible.
Effect of pH (A) and ethanol concentration (B) on tannin adsorption by skin IC (○: MBA, ▲: CS). Ethanol concentration and pH were 12% (A) and 3.3 (B), respectively. Bars show SD (n=3).
The high adsorption potential of IC toward anthocyanins and skin tannins was shown in this study. These phenolics are located in the vacuoles of skin cells. Some phenolics may be retained in cells during winemaking through the adsorption by IC around vacuoles, whereas others may be extracted to the liquid during winemaking. Some of the extracted phenolics may be re-adsorbed by IC that is exposed to the liquid during winemaking. Indeed, our previous study showed that the concentration of skin phenolics decreased during red winemaking (Okuda et al., 2014). In the current study, the importance of skin IC was clarified. Knowledge of the optimum pH for anthocyanin and skin tannin adsorption/extraction is also vital for red winemaking. The practical must pH is around 3.2 to 4.2 in Japanese red winemaking. The optimum pH for anthocyanin extraction is around 3.5 and that for skin tannin extraction is around 4.2. To make good quality red wines, great caution should be exercised in the pH adjustment of must. This concerns not only the issue of extraction but also issues in the chemical or physical stabilization of anthocyanins, including copigmentation (Boulton, 2001) and polymeric pigment complex formation (Peng et al., 2002). However, anthocyanin adsorption rates (%) were diminished at pHs below 2.6, around 3.5, and above 4.5. It was clarified that grape IC adsorbed anthocyanins and tannins, and this phenomenon was clearly observed in MBA. This may be one of the reasons for the low concentrations of grape skin phenolics in MBA wine. Moreover, pH and ethanol concentration (%) affected grape skin tannin adsorption. This suggests that pH management is required during red winemaking as it may affect the extraction and adsorption of phenolics.
Acknowledgments This work was supported by JSPS KAKENHI Grant Number JP17K07812.
