EFFECTS OF AGING IN OAK BARRELS ON THE TRANS -RESVERATROL AND ANTHOCYANIN CONCENTRATION OF RED WINES FROM HUNGARY

Polyphenol compounds in grapes and wines are of paramount importance: they have a key role in determining wine quality, and also the bene ﬁ cial health effects of moderate red wine consumption are well-known. The polyphenol concentration of wines is determined mostly by: a) their concentration in the grapes and b) the production technology, particularly the time and type of aging. Our goal was to determine the trans-resveratrol and anthocyanin contents of Hungarian red wines under different manufacturing conditions, by monitoring the 24-month aging process – using barrique and oak barrels – with semi-annual sampling, without considering vintage. We have chosen to determine polyphenol components that could originate from either the grapes, or be produced during the wooden barrel aging. Both the aging time and the wine variety had non-negligible effects on the changes in the concentrations of the studied components, whereas the wooden barrel type had not.

Acta Alimentaria 48, 2019 polymerization and condensation (WATERHOUSE, 2002).Flavonoids are particularly strong antioxidants, and are responsible for the sensory changes and browning of wines as well (SINGLETON & ESAU, 1969).The bitter and astringent taste is determined by catechins and leucoanthocyanines (KENNEDY et al., 2006).Resveratrol is a stilbene-type non-fl avonoid phenol.It is mostly present in the grape skin; consequently, the resveratrol content of wines is determined primarily by the applied grape processing technology -pectin-degrading enzyme treatment, CO 2 maceration, application of hyperoxidation, and the size of the press.trans-Resveratrol occurs naturally in the grape skin as its glycoside, called piceid.During alcoholic fermentation, the β-glycosidase activity of yeast releases trans-resveratrol (BAVARESCO et al., 2016).trans-cis rearrangement can occur as a result of the isomerase activity of yeast, while piceid-resveratrol conversion was observed for malolactic fermentation as well (KEITA et al., 2004).On one hand, resveratrol is a fl oral immune substance, participating in the defence mechanism of the grape against fungal infections (YANG et al., 2009).On the other hand, in humans it has a protective role against cardiovascular diseases, as well as antiinfl ammatory, antiviral, antibacterial, and antioxidant properties (KING et al., 2006;KISBENEDEK et al., 2014).As demonstrated by SEIGNEUR and co-workers (1990) and GAUTAM and co-workers (2000), trans-resveratrol has an important inhibitory role against leukemic cells, and was found to be effective against certain cancerous conditions ( GUERRERO et al., 2009).Resveratrol was observed to inhibit β-amyloid aggregation in Alzheimer's disease (RIVIERE et al., 2010), and to protect neurons from oxidative stress (MIZUTANI et al., 2000).
Our goal was to chemically monitor the aging phases of wine, to observe and quantify the changes in polyphenol concentrations.We have investigated the total polyphenol contents of three different wine varieties from the same region.The wines were produced under manufacturing conditions with controlled fermentation, and aged in different oak barrels.During the 24-month wooden barrel aging, the wines were sampled semi-annually to determine their polyphenol content, without accounting for vintage.The components to be determined were selected to be representative of polyphenols that are introduced either naturally, from the grapes, or during the applied technology, or from the wooden barrel aging.Present study deals with the fi ne composition of the Kadarka, Kékfrankos (Blaufrankisch), and Cabernet Franc varieties, focusing on the concentrations of trans-piceid and transresveratrol during the two-year wooden barrel aging.Hungarian wines were also analysed from two wine regions, from Villány and Eger, with the highest trans-resveratrol contents in Merlot and in Cabernet Sauvignon, and trans-piceid content in Blaufrankisch from Villány wine region (MONTSKO et al., 2010).
The aim of the study was to fi nd answers to the following questions: how the total polyphenolic content -and, in particular, the concentration of trans-piceid, trans-resveratrol, and anthocyanins -changes over the two-year wooden barrel aging; and does the applied wooden barrel type have any effect on the amount of polyphenols over the two-year aging process.

Materials and methods
Three characteristic wine varieties of the Szekszárd wine region in Southwest Hungary (Kadarka, Kékfrankos, and Cabernet Franc) were studied.The wines have not received any treatment other than racking and basic sulphuring.The wines were sulphured up to 20-25 mg l -1 of free sulphur concentration.Harvest dates were different for the three varieties: early October for Cabernet Franc, mid-October for Kadarka, and late October for Kékfrankos.For Unauthenticated | Downloaded 09/15/23 05:22 AM UTC each variety, healthy base material was processed.Alcoholic fermentation was carried out under controlled conditions for 8-14 days, at 18-24 °C temperature, using UVAFERM BDX variety-specifi c yeast.Following fermentation, the different varieties (Kadarka, Kékfrankos, and Cabernet Franc) were stored in separate oak and barrique barrels for 24-month aging, while the control samples were stored in steel barrels.Sampling was carried out twice a year -in April 2014, September 2014, April 2015, and September 2015 -, both for the wooden barrel and for the control samples.
Spectrophotometric measurements were carried out on a MOM Spektromom 195 device.The total polyphenol content was determined with the Folin-Ciocalteu phenol reagent and published as gallic acid equivalents (KÁLLAY & TÖRÖK, 1997).Anthocyanin content was measured at 550 nm according to the modifi ed method of FLANZY and co-workers (1969), following dilution with 96 % ethanol, containing 2 v/v% HCl.Leucoanthocyanines were determined spectrophotometrically after heating with a 40/60 mixture of hydrochloric acid and butanol according to the modifi ed method of FLANZY and co-workers (1969).The concentration of both components were expressed in malvidin-3,5-diglycoside equivalents.Catechin content was measured spectrophotometrically at 500 nm, after reaction of the alcohol-diluted wine with vanillin (REBELEIN, 1965).Colour index and colour tone (wine hue) were measured according to the Hungarian standard (MSZ 14848:1979) based on the work of SUDRAUD (1958).Absorbances for colour index and colour tone were measured at 420 nm and 520 nm, respectively.Colour index is calculated as the sum of the absorbances for the two wavelengths, while colour tone is expressed by the ratio of the absorbances.The qualitative and quantitative analyses of resveratrols and anthocyanin monomers were carried out with high-performance liquid chromatography (HP Series 1050) in the research laboratory of the Department of Oenology at Szent István University.
Determination of resveratrols was carried out according to the method described by KÁLLAY and TÖRÖK (1997).After fi ltration, the wines were injected directly into the HPLC device working in isocratic mode.For fi ltration, a Sartorius membrane fi lter with 0.45 μm pore diameter was used.Determination of anthocyanin monomers was carried out according to the method described by KÁLLAY and TUSNÁDY (2001).Similarly to resveratrols, no sample preparation was needed other than micro-fi ltration.Prior to injection, the wine samples were fi ltered with a Sartorius membrane fi lter with 0.45 μm pore diameter.The used solvents (HPLC grade) and chemicals (analytical grade) for the spectroscopic and HPLC measurements were purchased from VWR International (Radnor, PA, USA).The applied standards were obtained from Sigma-Aldrich group (Saint Louis, MO, USA).
Quantitative results were evaluated with several standard statistical methods, such as correlation statistics and factorial analysis of variance (ANOVA).The latter is based on the pairwise comparison of the average values of the different groups of samples.STATISTICA 13 software (Dell Inc., Tulsa, OK, USA) was used for the analysis.
For all three varieties, 500-litre oak barrels and 225 litre barrique barrels, produced from sessile oak (Quercus petraea), were applied for the 24-month aging process.We have used fi rst-load oak barrel for Kadarka, and fi ve-year, fourth-load barrel for Kékfrankos and Cabernet Franc.The barrique barrels were fi rst-load for Kadarka and third-load for Kékfrankos and Cabernet Franc.The oak and barrique barrels used for the aging of Kadarka were produced from two varieties of Hungarian sessile oak (Zemplén and Mecsek) in 50-50% ratio, their stave seasoning has lasted for 36 months and they received toasting afterwards.Those used for Kékfrankos and Cabernet Franc were produced purely from the Zemplén variety, with minimum 24 months of stave seasoning.

Results and discussion
From the studied varieties, the starting trans-resveratrol concentration of Kékfrankos is the highest (this is in agreement with earlier fi ndings).Table 1 summarizes the changes in the concentrations of trans-piceid and trans-resveratrol over time for the three wine varieties.trans-Piceid and trans-resveratrol concentrations display an increasing tendency, because of the small extent of sulphuring during the two-year observation.Kékfrankos clearly exhibits higher concentrations in the case of trans-resveratrol.Phenolic compounds are introduced into the wine primarily from the grapes and must (while keeping their bioactivities); this can be further promoted by grape processing and alcoholic fermentation.Wooden barrel aging can also increase the starting concentration of resveratrol, as demonstrated by the data for Kadarka and Kékfrankos.In addition, resveratrol can be released from piceid during fermentation and aging.We have calculated the basic statistical properties of polyphenol content and the nonparametric Spearman rank-correlation coeffi cients between the variables.(The abbreviated variable names are explained in the Abbreviations section.)The basic properties and the correlation coeffi cients are summarized in Tables 2 and 3, respectively.The higher standard deviations for some variables can be explained with differences between the wine varieties.To proceed, it was necessary to rescale the variables, for which standardization was employed.
There is a signifi cant positive correlation between catechin (CAT) and leucoanthocyanin (LEU) concentrations; these have decreased due to aging, owing to polymerization (Table 3).The decrease of CAT concentration is presumably due to the polymerization of catechin monomers to tannins of various degrees of polymerization (and molecular mass).The decrease of LEU concentration is due to polymerization or oxidative reactions.
Colour tone (TON) and polymerization (P) exhibit a negative correlation: with longer storage times polymerization increases, while colour tone decreases or remains unchanged.This can be explained by the polymerization reactions of anthocyanins.The loss of anthocyanins during aging cannot be considered a loss of colour, as most of the anthocyanin monomers polymerize (for red wines receiving months of aging, 30-40% of the colorants are polymers).This quantity can be considered stable in the absence of physical, chemical, or microbiological perturbations.During a change in this equilibrium, the polymerization degree might increase.Notes: The range of the coeffi cients is between -1 and +1.The most correlated variables are marked with a bold font.
Based on the correlation coeffi cients, there is a strong correlation between trans-piceid (TPI) and cis-resveratrol (CRES) and between cis-resveratrol and trans-resveratrol (TRES).Concentrations of the less stable cis-resveratrol were much lower than trans-resveratrol, but for both an increase in concentration could be observed at the beginning, followed by a gradual decrease at the end of the aging process.
We have applied factorial analysis of variance (ANOVA) to the standardized variables.The dataset contained a total of 11 dependent variables (measured values) and two factors (wine variety, barrel type).Effects of the two factors were analysed together and separately.Based on the results in Suppl.Material Table S1, barrel type (oak, barrique, and control) did not result in statistically signifi cant difference (α=0.05 in each case), unlike wine variety.In addition, the two factors together did not convey a signifi cant difference, either.Supplementary material is available online at https://akademiai.com/doi/suppl/10.1556/066.2019.0004.
Figure 1 presents an illustrative example of the similarity of the concentrations (in this case that of trans-resveratrol), when plotted against barrel types.Barrel type 1 was the initial sample.The trans-resveratrol concentration of Kadarka has barely changed compared to the starting sample, and the two other varieties did not display signifi cant differences between barrel types either (including the steel barrel control samples).The markers on Figure 1 denote the average values in each case.On the other hand, there is a signifi cant difference at the α=0.05 level based on measurement time (sampling time) and wine variety, as summarized by the factorial ANOVA results in Suppl.Material Table S2.This result prompted us to investigate the trans-resveratrol and anthocyanin concentrations of the Kékfrankos variety separately, compared to the other two varieties, as shown in Figure 3AB.It is clear that Kékfrankos displays statistically signifi cantly higher trans-resveratrol and anthocyanin concentrations compared to the other two varieties.Despite the fact that we have not accounted for vintage in this study, our results are in agreement with those of NIKFARDJAM and co-workers, 2006, in particular that the variety is a prime determi nant of the concentrations of trans-resveratrol and anthocyanin.On the other hand, the trans-resveratrol content can be infl uenced by the harvest date as well, based on the work of GEANA and co-workers (2015).

Conclusions
The effects of manufacturing conditions, particularly the type and length of wooden barrel aging were examined for three wine varieties (Kadarka, Kékfrankos, and Cabernet Franc).The varieties have displayed different starting polyphenol concentrations.After aging, anthocyanin concentrations have decreased due to the long wooden barrel aging process, and the extensive polymerization of polyphenols.From the three varieties, Kékfrankos displayed the highest initial concentration of anthocyanin.Trans-resveratrol was also, on average, a more prominent constituent in this variety.Our fi ndings are in good agreement with earlier results.Based on statistical analysis, the type and size of oak and barrique barrels used for the aging process did not have a signifi cant effect on the changes in polyphenol concentrations.

Fig. 2 .
Fig. 2. Factorial ANOVA with the factors of sampling time and wine variety for anthocyanin concentrations.The semiannual measurements are denoted starting with April 2014 and ending with Sept. 2015.Kadarka is marked with (blue) continuous line, Kékfrankos is marked with (red) dotted line and Cabernet Franc is marked with (green) dashed line.Vertical bars denote 0.95 confi dence intervals : Kadarka; : Kékfrankos; : Cabernet Franc

Fig. 3 .
Fig. 3. Factorial ANOVA for the different wine varieties based on A: trans-resveratrol and B: anthocyanin concentrations.Vertical bars denote 0.95 confi dence intervals

Table 1 .
Changes in trans-piceid and trans-resveratrol concentrations at the four sampling dates for the three varieties

Table 2 .
Basic statistical properties of the variables N: number of samples.The full names of the variables can be found in the Abbreviations.