Abstract
When model wines that contain polyphenols are oxidized, hydrogen peroxide and quinones are produced. Sulfur dioxide reacts with the hydrogen peroxide, preventing ethanol oxidation by way of the Fenton reaction, and in the case of (+)-catechin, sulfite reduces the quinone near quantitatively back to the catechol. Consequently, the O2:SO2 molar reaction ratio is close to 1:2 in ideal experimental conditions. Here, eight wines (three red and five white wines) were studied to investigate whether this ratio might be similar in practice, so as to assess how effective SO2 might be as an antioxidant in real wine. The reaction ratio was found to be decreased down to 1:1.7 in most wines. To determine the reason for this decrease, a white wine was treated with a large amount of benzenesulfinic acid. This substance reacts very efficiently with quinones and would therefore prevent their interaction with sulfite. The molar reaction ratio was then reduced to 1:1, as has been previously observed in model wine. This result was taken to indicate that sulfite is fully effective in removing hydrogen peroxide and that the reduction in the molar reaction ratio from the theoretical 1:2 ratio was due to limited interaction with polyphenol oxidation products. Two white wines, which were found to be rapidly oxidized with much reduced O2:SO2 molar reaction ratios, were found to contain ascorbic acid. The effect of adding ascorbic acid to a white wine on the reaction of oxygen was therefore also examined.
It has been proposed that the interaction of polyphenols with oxygen is mediated by the redox cycling of the Fe(III)/Fe(II) couple (du Toit et al. 2006, Waterhouse and Laurie 2006). Oxygen oxidizes Fe(II) to the ferric state, producing hydrogen peroxide, a process that is markedly accelerated by copper (Ribéreau-Gayon et al. 2000a, Danilewicz 2013). In turn, Fe(III) coordinates with polyphenols and oxidizes them to quinones (Scheme 1; Danilewicz 2011). This second step is much slower than Fe(II) oxidation, and although pyrogallol systems will oxidize alone, weaker reductants such as catechols require the assistance of nucleophiles that react with quinones to drive the reaction forward. These nucleophilic substances, which may be termed “oxidation promoters,” include sulfite and thiols in wine and benzenesulfinic acid (BSA) and azide under experimental conditions (Danilewicz 2011, Nikolantonaki and Waterhouse 2012). Overall, sulfite therefore not only removes hydrogen peroxide, thereby preventing ethanol oxidation (Elias and Waterhouse 2010), but also accelerates polyphenol oxidation and then recycles the resulting quinone back to the catechol.
Consequently, under simplified conditions when (+)-catechin is oxidized in model wine, two sulfites are taken up for each oxygen that is consumed (Danilewicz et al. 2008). Sulfite is very effective in this system, as it is estimated that ~96% of the quinone is reduced, so that, paradoxically, the concentration of the catechol does not appear to change as it is oxidized. With (−)-epicatechin, only ~79% of the catechol is regenerated, suggesting that the quinone is less stable, and some is lost before it can react with sulfite (Danilewicz and Wallbridge 2010). When 4-methylcatechol is oxidized in model wine, not all of the quinone is reduced, some adds sulfite to give the sulfonic acid (Danilewicz et al. 2008). However, the O2:SO2 molar reaction ratio is still ~1:2. It is proposed that the mechanism of quinone reduction involves the addition of sulfite to the quinone, as occurs with other carbonyl compounds, but this addition is followed by sulfate elimination to regenerate the catechol (Danilewicz 2011). Experimentally, as mentioned above, other nucleophiles, such as BSA and azide, also accelerate catechol oxidation. In the case of BSA, the O2:SO2 molar reaction ratio is reduced to 1:1 in model wine, as SO2 reacts only with hydrogen peroxide because the quinone is no longer available. In wine, substances such as thiols are present that can intercept quinones and compete with sulfite to reduce the O2:SO2 molar reaction ratio (Danilewicz et al. 2008, Nikolantonaki et al. 2010, Nikolantonaki and Waterhouse 2012). Thiols are an important group of compounds. They include cysteine and derivatives (Sarakbi et al. 2013), varietal thiols (Nikolantonaki et al. 2012), which require to be preserved during wine production, and undesirable mercaptans along with H2S, formed during fermentation and wine storage, both of which need to be removed (Ribéreau-Gayon et al. 2000b).
In a small initial study, the O2:SO2 molar reaction ratio was shown to be ~1:1.7 in two red wines (Danilewicz et al. 2008). The reaction of sulfite with hydrogen peroxide is extremely fast, as is evident in the oxidation–aspiration method for measuring SO2 concentration (Ough and Amerine 1988a). Therefore, departure from the theoretical 1:2 molar reaction ratio should be most likely due to the competition of wine constituents with sulfite for polyphenol oxidation products, presumably mainly quinones. The aim of this study was to measure the O2:SO2 molar reaction ratio in a larger number of wines to assess how effective sulfite is in protecting polyphenols from oxidative degradation in practice.
Materials and Methods
Materials
Water (Emsure, Fe ≤ 1 μg/L, Cu ≤ 0.4 μg/L; E. Merck, Darmstadt, Germany), Cu(II) sulfate pentahydrate, sodium hydroxide, L -(+)-tartaric acid (BDH AnalaR grade), and ethanol (96% GPR grade) were obtained from VWR International (Lutherworth, UK). Ascorbic acid, Fe(II) sulfate heptahydrate (>99% ACS reagent), 2,6-dichlorophenol-indophenol, and BSA sodium salt were obtained from Sigma-Al-drich (Poole, Dorset, UK) and potassium metabisulfite (Kadifit) from Erbslöh Geisenheim AG (Geisenheim, Germany).
Wines
Sauvignon blanc (WO Western Cape, South Africa), Pinot Grigio (Hungary), Chardonnay (SE Australia), and Merlot (Spain) were obtained from Sainsbury’s Supermarket Ltd., London, UK EC1N 2HT. Soave (Italy) and Shiraz 2014 (Australia, Jacaranda Hill) were purchased from the Co-operative Group Ltd., Manchester, UK M60 0AG. A Cabernet Sauvignon 2013 (Blossom Hill, CA) was obtained from Blossom Hill Co., London, UK NW10 7HQ, and a Chardonnay 2013 (Turner Road, CA) from Accolade wines Ltd., Guilford, UK GU3 1LR. Where the year is not given, the wines were nonvintage, obtained from wine boxes.
Preparation of model wine solutions
(+)-Tartaric acid (8.0 g) was dissolved in water (~750 mL) in a 1-L volumetric flask, and ethanol was added to give a 12% (v/v) final concentration. The pH was increased to 3.20 with 2.5 N sodium hydroxide, also adding water progressively to the measuring mark as the required pH was approached. The concentration of tartaric acid was selected to simulate the action of all the wine acids and to produce a final titratable acidity of ~4 g/L. Both malic and tartaric acids bind to Fe, and both similarly affect the reduction potential of the Fe(III)/Fe(II) redox couple (Danilewicz 2014).
Measurement of the wine O2:SO2 molar reaction ratio
The free and bound SO2 concentrations were checked initially, and potassium metabisulfite was added as required to raise the free SO2 concentration to between 30 and 45 mg/L. After equilibration for two to three hr, the wine was shaken in air to saturation. When required, ascorbic acid (60 mg/L) and BSA (as the sodium salt, 1.64 g/L) were then added. The air-saturated wine was quickly sealed in sixteen ~68-mL brown glass bottles with screw caps fitted with a plastic cone liner. The bottles were completely filled, ensuring the exclusion of any air bubbles and then stored in the dark. The small air space between the plastic cone liner and the cap was filled with a polymeric filler to avoid O2 passing through the plastic cone into the solutions during the experiments. The first three bottles were taken immediately to measure starting O2 and free and bound SO2 concentrations in triplicate.
The uptakes of O2 and SO2 were followed over time in triplicate, and the O2:SO2 reaction ratio was calculated from the uptake of O2 and total SO2 at the final reading, which in the case of the red wines was at 88.3 to 89.3 hr (Figure 1 and Table 1). For white wines, this was at the last reading shown in Figure 2 and Table 1. The importance of using total SO2 uptake is discussed below.
O2 measurement
An HI-9146 dissolved-O2 meter fitted with an HI-76408 Clarke-type electrode was used (Hanna Instruments Ltd., Leighton Buzzard, UK). The manufacturer specifies the resolution and limit of detection of the instrument as 0.01 mg/L O2, and the instrument is self-calibrating. However, to ensure that the instrument was functioning correctly and had fully stabilized, O2 concentration (corrected for temperature) in air-saturated water was checked, and the measurements were repeated several times until consistent before the O2 concentration was measured in wines. For the measurements, the bottle caps were quickly removed, and a small stirrer bar was inserted, followed by the electrode, which displaced ~13 mL of liquid. The electrode tip was lowered to ~5 mm above the briskly stirring magnetic bar to ensure that O2 was not depleted on the surface of the electrode membrane. Readings stabilized within 30 sec and then remained stable for more than 5 min, showing that, although the system was not completely closed during the measurements, no measurable amount of external O2 reached the measurement region during that time.
SO2 measurement
Free and bound SO2 concentrations were measured by the aeration-oxidation method (Ough and Amerine 1988a).
Measurement of the O2:SO2 molar reaction ratio as ascorbic acid is oxidized in model wine
Fe (3.0 mg/L) as FeSO4, Cu (0.15 mg/L) as CuSO4, and SO2 (~70 mg/L) as potassium metabisulfite were added to 500 mL model wine, which was saturated with air and allowed to equilibrate overnight. Ascorbic acid (88.1 mg/L) was added, and the model wine was sealed in brown bottles as described above. Changes in O2 concentration and free and bound SO2 concentrations were followed as single measurements because of the length of time required for SO2 determinations relative to the speed of the reaction (~5 hr). Small amounts of bound SO2 (~3 mg/L) were observed initially and were due to impurities in the ethanol. A duplicate experiment gave essentially the same result.
Ascorbic acid measurement
Ascorbic acid concentration was determined by the 2,6-dichlorophenolindophenol method (Ough and Amerine 1988b).
UV-vis spectroscopy
Spectra were taken with a Jenway 7315 spectrometer (Keison Products, Chelmford, UK), using 10-mm quartz cuvettes. Wine samples (5 mL) were diluted to 20 mL with water, and spectra were obtained against water blanks. E280 and E320 were the absorbance values at 280 nm and 320 nm, respectively, multiplied by the dilution factor.
Fe and Cu measurements
Fe and Cu analyses were performed by Campden BRI (Nutfield, UK), using a Perkin Elmer 3110 atomic absorption spectrometer (Perkin Elmer Corp, Norwalk, CT).
Measurements were taken in triplicate, and mean values (± SD) were calculated and figures drawn with Excel software (Microsoft, Redmond, WA). Where error bars denoting ± SD are not shown, they were smaller than the data point symbol dimensions. Experiments were conducted at ambient temperature (20 to 25°C).
Results and Discussion
The three red wines studied oxidized at similar rates and gave O2:SO2 molar reaction ratios close to 1:1.7 from O2 uptake and total SO2 consumption at 88.3 to 89.3 hr (Figure 1 and Table 1). The observed ratios were similar to those previously reported for a Beaujolais and Shiraz wine (Danilewicz et al. 2008). It should be noted that when free SO2 is reduced, some bound SO2 dissociates, so that bound SO2 concentration is also reduced. This was particularly so for these wines. Therefore, in calculating reaction ratios, it was noted that it is essential to take the reduction in total SO2 into account. For instance, with the Shiraz wine, the ratio was decreased from 1:1.73 to 1:0.8 if only free SO2 was considered.
The five white wines that were examined showed much greater variability in oxidation rates than the red wines (Figure 2). The white wines were therefore profiled to determine if some of their basic properties might be responsible for these differences (Table 2). The Sauvignon blanc wine consumed O2 a little more slowly than the Soave (Figure 2), even though it had a higher pH and higher polyphenol concentration (Table 2). However, the Soave had double the Fe and Cu concentrations of the Sauvignon blanc (Table 2). The Californian Chardonnay, which showed a similar profile to the Sauvignon blanc wine, but with a higher Fe content, oxidized much faster than both the Soave and Sauvignon blanc wines (Figure 2). Evidently, many more wines will need to be studied with more detailed profiling to establish what controls the oxidation rate. The Sauvignon blanc wine showed a perfect 1:2 molar reaction ratio, but the Californian Chardonnay and Soave wines had reaction ratios of ~1:1.7, similar to those of the red wines (Table 1).
Studies with model wines have shown that the rate of oxidation, as measured by SO2 consumption, is dependent on polyphenol concentration and is independent of SO2 concentration, as the rate of oxidation remained essentially constant as the concentration of free SO2 decreased from ~100 to 10 mg/L (Danilewicz 2007). These observations provide clear evidence that oxygen does not react with SO2 directly but with polyphenols. The 1:2 O2:SO2 molar reaction ratio also showed that SO2 is very efficient in reacting with the hydrogen peroxide and quinones in the simple model system. It can be inferred, therefore, that SO2 should be similarly efficacious in reacting with hydrogen peroxide in real wine. Any departure from the ideal reaction ratio should be due to the reactivity of SO2 with the oxidation products of the more complex polyphenol mix in real wine.
(+)-Catechin is not oxidized on its own in model wine. It requires sulfite to react with its quinone to drive the reaction forward. However, pyrogallol, which is a stronger reductant, is oxidized alone, even though its rate of oxidation is still accelerated by sulfite (Danilewicz 2011). This observation explains why the removal of SO2 from red wine, which contains more reactive polyphenols incorporating pyrogallol systems, slows oxidation but does not entirely prevent it (Danilewicz et al. 2008). Therefore, the dependence of the rate of oxidation and of the O2:SO2 molar reaction ratio on sulfite concentration in real wine needs to be examined more closely. To avoid any complications in this regard, an effort was made to adjust the starting concentrations of free SO2 to similar levels in all the wines.
It has long been known that the rate of polyphenol oxidation is pH-dependent, accelerating as pH increases (Cilliers and Singleton 1991). The rate of oxidation also depends on Fe concentration and is greatly influenced by Cu, which has been shown to accelerate Fe(II) oxidation (Danilewicz 2013). Temperature would also be important, and although not controlled in this study, it remained in a reasonably narrow range.
Two wines stood out as being rapidly oxidized with low O2:SO2 molar reaction ratios: the Pinot Grigio with a ratio of 1:1.29 and the Australian Chardonnay with a ratio of 1:1.36. It appeared that these two wines contained a more reactive reductant, the oxidation product of which consumed less SO2. Ascorbic acid is added to wine with sulfite at bottling to provide additional antioxidant activity (Bradshaw et al. 2011), and its presence in these wines was immediately suspected. Indeed, titration against 2,6-dichlorophenolendophenol (Ough and Amerine 1988b) showed that the Pinot Grigio and Australian Chardonnay wines contained 52 and 34 mg/L of the acid, respectively.
As with catechols, when ascorbic acid is oxidized, oxygen is reduced to hydrogen peroxide, which would be removed by sulfite. However, unlike the (+)-catechin quinone, the dehydroascorbic acid produced forms an adduct (Bradshaw et al. 2011), and so bound SO2 should increase (Scheme 2). When ascorbic acid was oxidized in model wine, the free SO2 concentration was reduced by 12.8 mg/L for an O2 uptake of 3.81 mg/L, resulting in an O2:SO2 molar reaction ratio of 1:1.7, the same ratio as has been previously reported (Bradshaw et al. 2004). However, bound SO2 increased by 4.8 mg/L. Subtracting this amount from the reduction in free SO2 should give the amount that reacted with hydrogen peroxide, which was 8.0 mg/L, giving an O2:SO2 molar reaction ratio of 1:1.05 as required. However, the increase in bound SO2 corresponded to only 63% binding to ascorbic acid oxidation products. Presumably, some dehydroascorbic acid reacted further before it could be intercepted by sulfite. This finding differed from that previously reported, which indicated that SO2 did not bind significantly to ascorbic acid oxidation products (Barril et al. 2012).
To examine the effect of added ascorbic acid to wine, 60 mg/L ascorbic acid was added to the Sauvignon blanc wine. The rate of oxidation markedly increased after this addition (Figure 3), and the molar reaction ratio was reduced from 1:2.03 to 1:0.99. Ascorbic acid both accelerated oxygen removal and substantially reduced the O2:SO2 molar reaction ratio. The effect was different from that observed in model wine, as the concentration of bound SO2 did not rise but slightly decreased (Table 1). It appeared that the ascorbic acid oxidation products that could react with SO2 in the model wine had reacted further in the real wine.
To examine why the reaction ratio should be reduced in wine, BSA was added to the Sauvignon blanc wine (Figure 3). BSA reacts very efficiently with quinones because in model wines, a quantitative yield of the phenylsulfone adduct is obtained (Danilewicz et al. 2008). The quinone–catechol couple of this product has a higher reduction potential than that of the original catechol system and therefore should not be further oxidized. The O2:SO2 molar reaction ratio was reduced to ~1:1 (Table 1), the same ratio as has been previously observed in model wine (Danilewicz et al. 2008). Evidently, with the quinone no longer being available, sulfite reacted with the full equivalent of hydrogen peroxide that was formed. Consequently, when the O2:SO2 molar reaction ratio is reduced in wine, this reduction is due to the incomplete reaction of SO2 with catechol oxidation products and not due to inefficient hydrogen peroxide removal. It is apparent, therefore, that sulfite is not wholly effective in protecting polyphenols, and a substantial proportion will not be regenerated but will react further as wine is oxidized. The inability of sulfite to protect polyphenols completely would account for changes in polyphenol composition that would be observed even when a wine was thought to be well protected by SO2. However, it is intriguing that one white wine did give the theoretical 1:2 reaction ratio, but a much larger study will be required to investigate why this was the case and how prevalent this observation is. There is the interesting possibility that a high reaction ratio could be beneficial during the production and storage of white wines.
Conclusion
Sulfite is not as effective in protecting polyphenols in wine as it is in model systems. It is possible, therefore, that in wine, more powerful nucleophiles compete with sulfite or that less stable quinones are formed which decompose before they can react with sulfite. However, it also appeared that sulfite, when used at recommended concentrations, should be fully effective in removing hydrogen peroxide and thus in preventing ethanol oxidation. If all the remaining sulfite reduces quinones, then up to ~70% of polyphenols would be recycled in most wines at the SO2 concentrations used in this investigation.
- Received July 2015.
- Revision received September 2015.
- Accepted October 2015.
- Published online December 2015
- ©2016 by the American Society for Enology and Viticulture