Reaction Mechanisms of Oxygen and Sulfite in Red Wine

Transcription

1 Reaction Mechanisms of Oxygen and Sulfite in Red Wine John C. Danilewicz 1 * and Matthew J. Standing 2,3 Abstract: Studies on the aerial oxidation of a simple catechol and also (+)-catechin in model wine have shown that the SO 2 :O 2 molar is 2:1. One mole equivalent of sulfur dioxide (SO 2 ) reacts with the hydrogen peroxide (H 2 O 2 ) that is produced, and the second with the quinone. However, more recent investigations in real wine have found that these ratios can be much lower, suggesting that SO 2 may be a much less effective antioxidant in practice. This study was therefore undertaken to examine the aerial oxidation of six red wines, looking closely at the uptake of O 2 and SO 2 at different starting SO 2 concentrations. Low ratios were indeed found in some wines initially, but as oxidation proceeded, the ratio increased and finally, given sufficient time, total ratios approached or were ~2:1. The effect was accentuated with low starting SO 2 concentrations when its availability was limited. It has been proposed that simple quinones are reduced by first adding sulfite, and that this adduct then hydrolyzes to release catechol and sulfate. On the basis of this mechanism, it is proposed that real wine contains polyphenols that form more stable initial adducts, which decompose more slowly. Consequently, SO 2 is released from these initial intermediate adducts when total SO 2 concentration is determined and so appears not to have reacted, resulting in low ratios. Given sufficient time, all the adduct decomposes, releasing sulfate, and the ratio increases to 2:1. It is further proposed that SO 2 still reacts with H 2 O 2, as in model wines. The apparent molar therefore should not fall below 1:1, and ratios lower than this were not found here, contrary to previous reports. Key words: oxygen, oxygen-sulfite, sulfite, sulfur dioxide, wine oxidation The reaction of oxygen is extremely important in wine production and storage. On the one hand, controlled oxygen exposure of red wines can improve wine quality, such as by reducing astringency and bitterness, removing vegetative character, and improving color stability (Carrascon et al. 2018). On the other hand, overexposure to oxygen can be highly damaging. White wines are less able to withstand oxidation because of their much lower polyphenol content, and are best exposed to as little oxygen as possible to maintain their fresh fruity character. However, wines are repeatedly exposed to oxygen during production, and ultimately, to a varying extent, during storage (Godden 2001). Sulfite is a key additive, both as an inhibitor of bacterial growth and an antioxidant, because it counteracts the effects of oxidation, although not by reacting with oxygen directly. Polyphenols containing catechol systems are the main reductants, which are oxidized by the intermediate redox cycling of the Fe(III)/Fe(II) couple (Scheme 1) (Danilewicz 2016a). Oxygen first reacts with Fe(II) with the assistance of copper to produce Fe(III), which then oxidizes polyphenols with the assistance of sulfite and is thereby returned back to the ferrous state. Most importantly, sulfur dioxide (SO 2 ) reacts with the hydrogen peroxide (H 2 O 2 ) that is initially produced, preventing the oxidation of ethanol via the Fenton reaction (Elias and Waterhouse 2010). SO 2 as sulfite also reacts with the quinones that are produced, forming stable adducts or reducing them back to catechols. By reacting with quinones, it is also found to accelerate catechol oxidation in model wines and to accelerate the oxidation of real wine (Danilewicz et al. 2008, Danilewicz and Wallbridge 2010). In addition to these reactions, sulfite adds to carbonyl derivatives, most strongly to acetaldehyde, to form nonvolatile sulfonic acid derivatives, and thereby removes aromas that are characteristic of oxidized wine (Waterhouse et al. 2016). Consequently, under ideal model wine conditions, when catechols are oxidized, the SO 2 :O 2 molar is close to 2:1. One mole equivalent of SO 2 reacts with H 2 O 2, and the second with the quinone (Danilewicz et al. 2008, Danilewicz and Wallbridge 2010). However, in real wines, lower SO 2 :O 2 molar s have been observed (Danilewicz et al. 2008, Danilewicz 2016b), some closer to 1:1, and even lower have been more recently reported (Ferreira et al. 2015, Carrascon et al. 2015, 2018). These findings are of concern, as they indicate that SO 2 may not be as efficient an antioxidant as it is supposed to be in practice. The aim of this investigation was to look more closely at the relationship between the reaction of O 2 and SO 2 in red wines to better understand how SO 2 participates in wine oxidation Sandwich Road, Ash, Canterbury, Kent CT3 2AF. UK; 2 Department of Wine, Plumpton College (University of Brighton), Ditchling Road, Plumpton, Nr Lewes, East Sussex BN7 3AE. UK; and 3 Present Address: Dehlinger Winery, 4101 Vine Hill Road, Sebastopol, CA *Corresponding author (jdanilewicz@btconnect.com; tel: ) Manuscript submitted Oct 2017, revised Jan 2018, accepted Feb 2018 Copyright 2018 by the American Society for Enology and Viticulture. All rights reserved. doi: /ajev Scheme 1 Proposed mechanism of catechol oxidation in wine: Fe(III)/ Fe(II) redox cycling and involvement of SO

2 190 Danilewicz and Standing Materials and Methods Materials. Potassium metabisulfite (Kadifit) was purchased from Erbslöh Geisenheim AG. Wines. The wines in this study were Merlot (2012, Luis Felipe Edwards, Mountain View, Central Valley, Chile), Pinot noir (2010, Sylvain Miniot, La Grille, Vin de France), and Cabernet Sauvignon (2011, Welmoed, Stellenbosch, South Africa). In the second part of the study, wines were from wine boxes and nonvintage, except for the Tempranillo. Merlot (12.5% alcohol, ph 3.31; Groupo de Bodegas, Vinartis, Spain) and Shiraz (14.5% alcohol, ph 3.17; WO Western Cape, South Africa) were obtained from Sainsbury s Supermarket Ltd. London EC1N 2HT. Tempranillo (2011, 13.5% alcohol, ph 3.34; Felix Solis, Albali, Valdepeñas, Spain) was obtained from Co-operative Group Ltd. Manchester M60 0AG. Experimental design. Part 1. This study was conducted in 2013 when the wines Merlot (Chile), Pinot noir (France), and Cabernet Sauvignon (South Africa) were between oneand three-years-old. Five bottles of wine, the necks of which had been cleaned with ethanol, were decanted into 10 L polyethylene containers, which had been sterilized with acidified bisulfite solution and thoroughly drained. The containers were fitted with a bottom drain tap to take samples, and a lid was fitted with a 45 µm air filter (HEPA-CAP-36, Whatman Ltd.). The containers, which had ~6.25 L headspace, were shaken for 3 min to saturate the wine with air. Starting O 2 and free and bound SO 2 concentrations were determined in duplicate, and wines were quickly sealed in 15 to 18 ~68 ml brown bottles with screwcaps fitted with a plastic cone liner. The bottles were completely filled, ensuring exclusion of any air bubbles. The small air space between the plastic cone liner and the cap was filled with a polymeric filler to avoid O 2 passing through the plastic cone into solutions during the experiments. The bottles were confirmed to be airtight by oxidizing a white wine. After 17 days, the O 2 and Ss were exactly the same in two triplicate sets of bottles, one stored in air and the other under N 2. The bottles were stored in the dark at 20 ± 1 C, and the progress of oxidation was followed by taking three bottles to measure O 2 and free and bound SO 2 concentrations in triplicate at the times shown in the Results section. The 10 L containers were also stored in the dark at 20 ± 1 C and shaken daily to maintain air saturation. When free SO 2 reached ~8 to 9 mg/l (Part 1b, Table 1), a second set of ~68 ml brown bottles were filled with wine, and the above determinations were repeated. Similarly, a third set was tested when almost all the SO 2 was removed (Part 1c, Table 1). Part 2. Wines (1.25 L), that is, Merlot (Spain), Shiraz (South Africa), and Tempranillo (Spain), were transferred to N 2 -filled flasks. Potassium metabisulfite was added to raise free SO 2 to ~50 mg/l. The wines were then left overnight to equilibrate, and free and bound SO 2 were measured in triplicate. The wines were transferred to 2 L flasks that contained air, which were then shaken to saturate the wine with aerial O 2. When oxygen concentration had stabilized, the mean of the last three readings was taken as the starting O 2 concentration. Fourteen ~68 ml brown bottles were filled with wine and O 2, and free and bound SO 2 concentrations were followed over time in triplicate as described above. O 2 measurements. Part 1. An HQ30d meter with an LDO101 probe (Hach Lang) was used, for which the manufacturer specifies a resolution and limit of detection (LOD) of 0.01 mg/l O 2. The instrument was calibrated each day by measuring O 2 concentration in air-saturated deionized water at 19 to 20 C. O 2 concentration was measured in the ~68 ml brown bottles, as described below. Part 2. An HI-9146 dissolved-o 2 meter fitted with an HI Clarke-type electrode was used (Hanna Instruments Ltd.). The manufacturer specifies a resolution and LOD of the meter as 0.01 mg/l O 2, and the instrument is self-calibrating. However, to ensure that the instrument was functioning correctly and had fully stabilized, O 2 concentration (corrected for temperature) in air-saturated water was checked, and measurements were repeated several times until consistent, before wine measurements were taken. 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. This was to ensure that O 2 was not depleted on the surface of the electrode membrane. Readings stabilized within 30 sec and then remained stable for more than 5 min, indicating that, even though the system was not completely closed during the measurements, no measurable amount of external O 2 reached the measurement area during that time. SO 2 measurements. Free and bound SO 2 concentrations were measured by the aeration-oxidation method (Ough and Amerine 1988). Results and Discussion Studies in model wines have shown that when the model catechol, 4-methylcatechol (4-MeC) is oxidized on air saturation, a 2:1 SO 2 :O 2 molar is observed (Danilewicz et al. 2008). It is proposed that one mole of SO 2 reacts Table 1 SO 2 concentrations in the wines included in this study. a Starting SO 2 concn Study part Free Total 1a Merlot (Chile) 33.6 ± ± 1.1 Cabernet Sauvignon (South Africa) 43.2 ± ± 0.0 Pinot noir (France) 30.4 ± ± 0.0 1b Reduced S0 2 concn Merlot (Chile) 9.6 ± ± 0.0 Cabernet Sauvignon (South Africa) 9.6 ± ± 0.0 Pinot noir (France) 8.0 ± ± 0.0 1c Merlot (Chile) 0.0 ± ± 0.0 Cabernet Sauvignon (South Africa) 0.0 ± ± Merlot (Spain) 48.3 ± ± 0.8 Shiraz (South Africa) 58.1 ± ± 0.8 Tempranillo (Spain) 59.2 ± ± 0.0

3 Reaction of Oxygen and Sulfite in Red Wine 191 with the H 2 O 2, and the second with the quinone (Scheme 1). Consequently, when excess of benzenesulfinic acid (BSA), a strong nucleophile capable of competing with sulfite, was included, the molar was lowered to 1:1. This result was consistent with SO 2 now reacting efficiently only with the H 2 O 2, and this latter component of S should be exactly the same when O 2 reacts in real wine. Examination of the quinone reaction products when 4-MeC is oxidized revealed that ~40% was converted to the sulfonic acid, and ~50% was reduced back to the catechol (Scheme 2) (Danilewicz et al. 2008). When (+)-catechin was oxidized, an ~2:1 SO 2 :O 2 molar was also observed, but with ~96% conversion back to the catechol (Danilewicz and Wallbridge 2010). No significant amount of adduct was evident. Sulfite is generally considered an antioxidant, and its reaction with H 2 O 2 presumably involves addition to SO 2 instead of water and rearrangement of the persulfurous acid (Scheme 2). In its reaction with quinones, it reacts as sulfite, which is a strong nucleophile. Michael 1,4- addition gives the sulfonic acid adduct. However, an interesting question is the following: What is the mechanism by which sulfite reduces quinones? Sulfite reduces oxygen by way of a free radical chain reaction, first donating an electron to Fe(III) to produce sulfite radicals, which react very rapidly with O 2 (Danilewicz et al. 2008). If a quinone reacted with sulfite simply by electron transfer, the sulfite radicals produced would again quickly react with O 2 to produce persulfate radicals, which are strong oxidants that oxidize polyphenols (Danilewicz et al. 2008). A free radical mechanism seems very unlikely. Consequently, it was proposed that sulfite reduces quinones by first adding to one of the carbonyl functions. At that stage, the apparent SO 2 :O 2 molar would be 1:1, as the sulfite would be recovered in the bound fraction. Hydrolysis would then raise the ratio to 2:1 as sulfate is eliminated (Scheme 2). When ascorbic acid is oxidized in model wine, a 1:1 SO 2 :O 2 molar is observed. While one mole reacts with H 2 O 2, the second mole forms an adduct with dehydroascorbic acid, and this sulfite fraction is recovered when bound SO 2 concentration is determined (Scheme 3). The was found to be reduced to ~1.3:1 in white wines that contain ascorbic acid (Danilewicz 2016b). Initial studies with two red wines reported ~1.7:1 SO 2 :O 2 molar s following one air saturation (Danilewicz et al. 2008). In a further study, three red and two white wines gave similar ratios. However, a Sauvignon blanc did give a 2:1 ratio, which like the model system was reduced to 1:1 on adding BSA, further indicating that sulfite will react efficiently with H 2 O 2 in real wine (Danilewicz 2016b). When the air oxidation of (+)-catechin was attempted in model wine, no oxygen uptake was apparent until sulfite was added. However, pyrogallol, which contains the more readily oxidized triphenolic system present in epigallocatechin, did oxidize alone, but even then sulfite accelerated the reaction (Danilewicz 2011). Consequently, on the understanding that direct sulfite oxidation is not possible in wine, it was proposed that sulfite accelerates polyphenol oxidation by reacting with the quinone. As mentioned above, the aerial oxidation of sulfite is a radical chain reaction and is inhibited by the radical scavenging activity of polyphenols (Danilewicz 2007). Also, other nucleophiles that are not oxidized (BSA and azide) have the same accelerating effect. The accelerating action of sulfite was then confirmed in real wine when it was shown that H 2 O 2 -mediated removal of sulfite from a red wine markedly slows oxidation (Danilewicz et al. 2008). To examine further the effect of sulfite on the rate of wine aerial oxidation, three red wines were maintained at air saturation in 10 L containers under aseptic conditions. Wine samples were taken at the beginning and sealed in sets of small bottles to measure the rate of oxygen uptake and the SO 2 :O 2 molar (Part 1a, Table 1). This determination was repeated when free SO 2 was reduced to 8.0 and 9.6 mg/l (Part 1b, Table 1), and again when almost all the SO 2 had been removed (Part 1c, Table 1). The total depletion of SO 2 concentration corresponded to an oxygen uptake following no more than two to three aerial oxygen saturations based on a 2:1 SO 2 :O 2 molar. The rate of wine oxidation was found to slow markedly as SO 2 concentration fell (Figure 1), which agreed with its observed accelerating effect on polyphenol oxidation. Slowing cannot be due solely to the progressive depletion of more readily oxidized wine constituents, as just the removal of SO 2 with H 2 O 2 has the same effect in a single air saturation. This result differed from one reported previously, indicating that the rate of O 2 consumption does not appear to change over five successive aerial O 2 saturations (Ferreira et al. 2015). Results on SO 2 :O 2 s were unexpected. Following aerial O 2 saturation, the initial ratio for the Merlot (Table 2) was 1.45:1 after one day, but it increased progressively as the oxidation continued, completing at 2.10:1. The same pattern was observed with the Pinot noir (Table 3). After Scheme 2 Reaction of sulfite with hydrogen peroxide (H 2 O 2 ) and quinones. Proposed two-stage reduction of quinones. Scheme 3 Oxidation of ascorbic acid and addition of sulfite to dehydroascorbic acid.

4 192 Danilewicz and Standing two days, the ratio was 1.37:1, which increased to 2.0:1 after 10 days. With both of these wines, S increased relative to that of O 2 as the oxidation proceeded. However, with the Cabernet Sauvignon, the ratio remained ~2:1 throughout the time (Table 4). The above three wines were allowed to oxidize in air to reduce the free SO 2 concentration down to ~8 to 9 mg/l, and the determination of the SO 2 :O 2 molar was repeated. The Merlot (Table 5) and Cabernet Sauvignon (Table 6) gave low ratios of 1.24:1 initially, but again the ratios increased as the oxidation proceeded, reaching 1.70:1 and 2.01:1, respectively, after 19 days. No free SO 2 remained in these wines by the end of the studies, with only 6.4 ± 0.0 mg/l and 12.8 ± 0.0 mg/l bound SO 2 remaining. The Pinot noir Figure 1 Aerial oxidation of three red wines (Part 1a,1b, and 1c). First determination and second and third after a reduction in sulfite concentration (Table 1). Table 2 Progressive increase in the SO 2 :O 2 molar as the wine is oxidized following one aerial O 2 saturation. a Part 1a Merlot (Chile) Total SO 2 uptake ± ± ± 0.21: ± ± ± 0.12: ± ± ± 0.09: ± ± ± 0.09: ± ± ± 0.08:1 a Values are mean ± standard error of the mean. behaved differently, starting with a 2:1 ratio, despite having a very low SO 2 concentration (Table 7). In the second part of this study, SO 2 concentrations were raised to ensure that there was more than sufficient SO 2 to counteract one aerial O 2 saturation. Again, there was an indication that the ratios increased over time in the Merlot and Shiraz, although with high standard errors (Table 8). However, ratios between 1.84:1 and 2.0:1 were obtained for the three wines after five days. It appears that if oxidation is followed over sufficient time and adequate SO 2 is present, the SO 2 :O 2 molar will tend to reach 2:1, as it does in model wines. The above results differed from those previously reported (Carrascon et al. 2015, Ferreira et al. 2015), when 15 wines were subjected to five successive aerial saturations. Oxygen uptake was followed for six to seven days before the next saturation. The uptake after that time for each saturation was then plotted additively for the five saturations. After a fast uptake on the first saturation, the uptake rate appeared to be constant. This led the authors to conclude that the rate of wine oxidation was independent of sulfite concentration, which fell over the course of the study. However, this manner of presenting the data obscured some very intriguing variation in the rate of oxidation through the study. Sometimes the rate slowed toward the end of the study, sometimes it accelerated, and sometimes it accelerated midway. A complicating factor is that sulfite content varied considerably among the wines. In none of the wines was there enough sulfite to protect the wine for five aerial saturations, and in some, barely for two. Consequently, oxidation occurred against a highly variable sulfite content. Another difference from the previous studies was that low SO 2 :O 2 molar s were observed, with some lower than a 1:1 ratio. Two wines even did not react with SO 2 at all, which is not readily explainable (Ferreira et al. 2015). When wine is oxidized in air, Fe(II) is rapidly oxidized during redox cycling to produce H 2 O 2. The understanding is that O 2, in its triplet ground state, can only accept electrons singly. If Fe is not involved, the only other possibility is that O 2 reacts with free radicals, the production of which is again not readily explainable. To explain the low SO 2 :O 2 molar s, it is proposed that nucleophiles may be present, which could compete with sulfite for quinones. Amino acids have been suggested as possible nucleophile candidates (Carrascón et al. 2018). Table 4 SO 2 :O 2 molar as the wine is oxidized following one aerial O 2 saturation. a Table 3 Progressive increase in the SO 2 :O 2 molar as the wine is oxidized following one aerial O 2 saturation. a Part 1a Pinot noir (France) Total SO 2 uptake ± ± ± 0.02: ± ± ± 0.07: ± ± ± 0.01: ± ± ± 0.02:1 Part 1a Cabernet Sauvignon (South Africa) Total SO 2 uptake ± ± ± 0.20: ± 0.12 nd b nd ± ± ± 0.10: ± ± ± 0.10: ± ± ± 0.07:1 b Not determined.

5 Reaction of Oxygen and Sulfite in Red Wine 193 Part 1b Merlot (Chile) after oxidation Table 5 Progressive increase in the SO 2 :O 2 molar as the wine is oxidized following one aerial O 2 saturation with lowered SO 2 concentration. a a Values are mean ± standard error of the mean. Free SO 2 ( mg/l) Total S ± ± ± ± 0.13: ± ± ± ± 0.03: ± ± ± ± 0.03: ± ± ± ± 0.02:1 Table 6 Progressive increase in the SO 2 :O 2 molar as the wine is oxidized following one aerial O 2 saturation with lowered SO 2 concentration. a Part 1b Cabernet Sauvignon (South Africa) after oxidation Free SO 2 Total S ± ± ± ± 0.06: ± ± ± ± 0.03: ± ± ± ± 0.01: ± ± ± ± 0.05:1 Table 7 SO 2 :O 2 molar as the wine is oxidized following one aerial O 2 saturation with lowered SO 2 concentration. a Part 1b Pinot noir (France) after oxidation a Values are mean ± standard error of the mean. Free SO 2 Total S ± ± ± ± 0.02: ± ± ± ± 0.02: ± ± ± ± 0.03: ± ± ± ± 0.02:1 Table 8 SO 2 :O 2 molar in three wines with raised SO 2 levels after one and five days aerial oxidation. a Part 2 Merlot (Spain) Shiraz (South Africa) Tempranillo (Spain) Total SO 2 uptake ± ± ± 0.35: ± ± ± 0.10: ± ± ± 0.10: ± ± ± 0.33: ± ± ± 0.27: ± ± ± 0.22: ± ± ± 0.18: ± ± ± 0.10: ± ± ± 0.07:1 However, these amino acids are protonated in wine conditions and therefore cannot function as nucleophiles. Moreover, neither phenylalanine nor methionine react with a quinone in a model wine. There also was no evidence, in the presence of sulfite, for the Strecker reaction, in which amino acids add to the quinone carbonyl system, as in α-diketones (Nikolantonaki and Waterhouse 2012). Similarly, the reaction of a quinone with phloroglucinol was very slow, indicating that quinones should not react to an appreciable extent with electron-rich aryl systems such as with flavanol A-rings, and these certainly could not compete with sulfite. As mentioned above, sulfite and other nucleophiles promote (+)-catechin oxidation. Phloroglucinol has little effect in this system, again indicating very little quinone interaction (Danilewicz 2011). This leaves thiols, which are a distinct possibility, as they add to quinones at a rate similar to that for sulfite in model wine (Nikolantonaki and Waterhouse 2012). Thiol oxidation occurs by redox cycling of Cu/Fe (Kreitman et al. 2016b), resulting in and H 2 O 2 production. In model wine, this leads to acetaldehyde formation by the Fenton reaction. Assuming that the H 2 O 2 reacts with sulfite, this initial phase would result in a 1:1 SO 2 :O 2 molar, and the main products would be disulfides. Polysulfanes would also be produced, but at very low concentrations, as they would depend on the presence of hydrogen sulfide (Kreitman et al. 2016a, Jastrzembski et al. 2017). Sulfitolysis of disulfides, to give the cysteine sulfonate for instance, would take up a further sulfite equivalent and raise the to 2:1, which could explain the delay in SO 2 consumption, but not the low ratios. However, total thiol content is estimated to be on the order of 60 μm (Kreitman et al. 2016a), which could only result in the uptake of 30 μm O 2 (~1.0 mg/l). It seems likely that thiols play an important part in wine oxidation, but their involvement can only in small part explain the above findings. A crucial question is the following: Would thiols oxidize, and would they do so very readily, before they could react with quinones?

6 194 Danilewicz and Standing A competing nucleophile could explain low SO 2 :O 2 molar s, but in the present study, the ratio increased during a single air saturation, which would not be possible if quinones had reacted with some other nucleophile. Assuming that the reductants were polyphenols, it is likely that they contained a catechol system because of its reactivity, as these reductants reacted quickly during the first aerial O 2 saturation. Other possible reductants are present in relatively low concentrations. If the quinone reacted irreversibly to form a sulfonic acid adduct, a 2:1 molar would be immediately observed. However, to explain low ratios, which increase subsequently, it may be that the quinone forms an initial 1,2-adduct, as proposed for quinone reduction (Scheme 2). If this product were relatively stable, the bound SO 2 would be recovered when bound SO 2 was determined, thus appearing not to have reacted, as is the case for ascorbic acid, resulting in a low SO 2 :O 2 ratio. As this initial adduct progressively hydrolyzes during the oxidation to produce sulfate, the ratio would then increase. This would explain how sulfite would appear to react, as observed in this present study, when the reaction of oxygen had essentially stopped. It is tentatively proposed, therefore, that SO 2 reacts with H 2 O 2 as efficiently in real wine as it does in model systems. Consequently, the increasing SO 2 :O 2 s would appear to be due to the release of sulfate from a polyphenol oxidation product that had been temporarily stabilized by adding sulfite. For the Pinot noir (Table 3), O 2 consumption after two days was 4.67 mg/l, which would require 1 mole equivalent (9.34 mg/l) of SO 2 to react with the H 2 O 2 produced (Figure 2). The total SO 2 consumed was 12.8 mg/l so that only 3.46 mg/l (12.8 minus 9.34) must have been oxidized to sulfate by way of quinone reduction. The amount bound reversibly to the intermediate quinone (5.88 mg/l) would be recovered in the bound SO 2 fraction and would appear not to have reacted, resulting in a 1.37:1 ratio. Similar calculations could be made with the Merlot (Table 2). Polyphenol oxidation products are highly unstable, and if they were to subsequently react with sulfite, such oxidation products would require temporary protection before the slower reaction involving sulfite. An alternative explanation could be that the quinone is trapped intramolecularly by reacting with a hydroxyl function, as in the formation of dicatechin oxidation products such as dehydrodicatechin A (Weinges et al. 1971, Guyot et al. 1996). A well-positioned hydroxyl function close to the quinoid system could react very quickly Figure 2 Oxidation of the Pinot noir (Table 3), and proposed reaction of sulfite after two days, yielding a 1.37:1 SO 2 :O 2 molar (O 2 uptake 4.67 mg/l; 12.8 mg/l SO 2 reacted). and compete with sulfite. The cyclic product might then react more slowly with sulfite. The results obtained in this study are different from those previously reported (Carrascón et al. 2015, Ferreira et al. 2015) in that the rate of oxidation slowed as the wine was progressively oxidized. The SO 2 :O 2 molar s also increased as oxidation progressed in a single aerial O 2 saturation. However, the wines used appear to be different in that those studied by the aforementioned researchers oxidized much faster, taking up oxygen after one aerial saturation in as little as a day in the first or subsequent saturations. Furthermore, these authors commented that the faster the rate of oxidation, the lower the SO 2 :O 2 ratio appeared to be. Some of the wines studied here were from wine boxes. However, they contained between 25.6 and 38.4 mg/l free SO 2 when used. Assuming the boxes were filled at the maximum legal concentration of 45 mg/l free SO 2, it does not seem that these wines had undergone excessive oxidation, possibly equivalent to not much more than one O 2 saturation. We conclude that these wine samples were perfectly acceptable for this study. As described above, a complicating factor in the studies by Ferreira et al. (2015) and Carrascon et al. (2015) was that the sulfite content varied considerably among the wines. In none of the wines was there enough sulfite to protect the wine against five aerial saturations, and in some, barely against two. An increase in the SO 2 :O 2 molar s has been described more recently in the second saturation (Carrascón et al. 2018), and this delayed S could well be the same that was observed in the present work. However, in this later study (Carrascón et al. 2018), six of the eight wines investigated contained barely enough SO 2 to protect against two O 2 saturations, particularly when acetaldehyde content, which could decrease SO 2 availability, is taken into account. Only two wines, which contained sufficient SO 2 for more than two saturations, showed a significant increase in ratio in the second saturation. This highlights the importance of SO 2 concentration during oxidation. It has been proposed that low SO 2 :O 2 ratios might be due to the unavailability of SO 2 caused by its binding to wine constituents such as anthocyanins, and Ferreira et al. (2015) have proposed the concept of SO 2 efficiency. The results obtained here would indicate that most, if not all, of the SO 2 is available to the oxidative process if followed for sufficient time. The mechanism described above would also imply that initial molar s should not fall below 1:1 when adequate SO 2 is present, since it is proposed that the H 2 O 2 will always react. Such wines with very low ratios were sought but not identified in the present study. Conclusion This study with six red wines has shown that the SO 2 :O 2 molar s may appear to be substantially lower than those found in model systems, particularly when sulfite concentrations are limited. However, when adequate sulfite is present, the ratios progressively increased to 2:1, as observed in simple model systems, when oxidation was followed for sufficient time. Two mechanisms are suggested to explain

7 Reaction of Oxygen and Sulfite in Red Wine 195 this delay in S. As for the proposed mechanism for quinone reduction, the initial oxidation product may bind sulfite reversibly, protecting it from further degradation, but giving the impression that this sulfite has not reacted when total SO 2 is measured. The progressive decomposition of this intermediate adduct is then proposed to produce sulfate irreversibly, which allows the ratio to rise to the theoretical value. An alternative nucleophile to sulfite could not be involved in lowering the SO 2 :O 2 molar, as this could not then increase subsequently. In an alternative suggested mechanism, it is proposed that a quinone might be intramolecularly trapped immediately as formed, and the cyclic product might then react more slowly with sulfite. Literature Cited Carrascon V, Fernandez-Zurbano P, Bueno M and Ferreira V Oxygen consumption by red wines. Part II: Differential effects on color and chemical composition caused by oxygen taken in different sulfur dioxide-related oxidation contexts. J Agric Food Chem 63: Carrascón V, Vallverdú-Queralt A, Meudec E, Sommerer N, Fernandez-Zurbano P and Ferreira V The kinetics of oxygen and SO 2 consumption by red wines. What do they tell about oxidation mechanisms and about changes in wine composition? Food Chem 241: Danilewicz JC Interaction of sulfur dioxide, polyphenols, and oxygen in a wine-model system: Central role of iron and copper. Am J Enol Vitic 58: Danilewicz JC Mechanism of autoxidation of polyphenols and participation of sulfite in wine: Key role of iron. Am J Enol Vitic 62: Danilewicz JC. 2016a. Fe(II):Fe(III) ratio and redox status of white wines. Am J Enol Vitic 67: Danilewicz JC. 2016b. Reaction of oxygen and sulfite in wine. Am J Enol Vitic 67: Danilewicz JC and Wallbridge PJ Further studies on the mechanism of interaction of polyphenols, oxygen, and sulfite in wine. Am J Enol Vitic 61: Danilewicz JC, Seccombe JT and Whelan J Mechanism of interaction of polyphenols, oxygen, and sulfur dioxide in model wine and wine. Am J Enol Vitic 59: Elias RJ and Waterhouse AL Controlling the Fenton reaction in wine. J Agric Food Chem 58: Ferreira V, Carrascon V, Bueno M, Ugliano M and Fernandez-Zurbano P Oxygen consumption by red wines. Part I: Consumption rates, relationship with composition, and role of SO 2. J Agric Food Chem 63: Godden P, Francis L, Field J, Gishen M, Coulter A, Valente P, Hϕj P and Robinson E Wine bottle closures: Physical characteristics and effect on composition and sensory properties of a Semillon wine 1. Performance up to 20 months post-bottling. Aust J Grape Wine Res 7: Guyot S, Vercauteren J and Cheynier V Structural determination of colourless and yellow dimers resulting from (+)-catechin coupling catalysed by grape polyphenoloxidase. Phytochemistry 42: Jastrzembski JA, Allison RB, Friedberg E and Sacks GL The role of elemental sulfur in forming latent precursors of H 2 S in wine. J Agric Food Chem 65: Kreitman GY, Danilewicz JC, Jeffery DW and Elias RJ. 2016a. Reaction mechanisms of metals with hydrogen sulfide and thiols in model wine. Part 1: Copper-catalyzed oxidation. J Agric Food Chem 64: Kreitman GY, Danilewicz JC, Jeffery DW and Elias RJ. 2016b. Reaction mechanisms of metals with hydrogen sulfide and thiols in model wine. Part 2: Iron- and copper-catalyzed oxidation. J Agric Food Chem 64: Nikolantonaki M and Waterhouse AL A method to quantify quinone reaction rates with wine relevant nucleophiles: A key to the understanding of oxidative loss of varietal thiols. J Agric Food Chem 60: Ough CS and Amerine MA Methods for Analysis of Musts and Wines. 2nd ed. p Wiley & Sons, New York. Waterhouse AL, Sacks GL and Jeffery DW Understanding Wine Chemistry. p John Wiley and Sons Ltd., UK. Weinges K, Mattauch H, Wilkins C and Frost D Oxydative kupplung von phenolen, V. spektroskopische und chemische konstitutionsaufklärung des dehydro-dicatechins A. Liebigs Ann Chem 754: