Key Factors Affecting Radical Formation in Wine Studied by Spin Trapping and EPR Spectroscopy

Key Factors Affecting Radical Formation in Wine Studied by Spin Trapping and EPR Spectroscopy Ryan J. Elias, 1 Mogens L. Andersen, 2 Leif H. Skibsted, 3 and Andrew L. Waterhouse 4 * Abstract: The nonenzymatic oxidation of wine has profound effects on its sensory attributes and, thus, perceived quality. While wine oxidation has been studied for centuries, only recently has the role of free radical intermediates in wine aging been seriously addressed. In this study, the effect of various wine components on the formation or suppression of free radicals was investigated. Electron paramagnetic resonance (EPR) spin trapping was used to detect, quantify, and identify factors that alter the major radical species formed in actual wine systems. In all cases and across all treatments, the 1-hydroxyethyl radical was the sole spin adduct detected, suggesting that the Fenton reaction (i.e., formation of the hydroxyl radical and its subsequent oxidation of ethanol) is the major route for oxidation in wine. The addition of either iron, copper, or iron and copper in combination to a red wine resulted in a marked increase in observed spin adducts, demonstrating that trace levels of metals are essential catalysts in the oxidation of wine. The addition of catechin to a white wine containing excess sulfur dioxide had no effect on the initial rate of radical formation, but was prooxidative in the latter stages of the experiment. Finally, sulfur dioxide was shown to inhibit radical formation in a concentration-dependent manner. Key words: wine oxidation, metals, EPR spin-trapping, Fenton reaction, sulfite, polyphenols The presence of metals in wine are known to have deleterious effects on quality. A classic example is casse formation, or turbidity resulting from protein-metal interactions. Low concentrations of metals in wine, particularly transition metals such as iron and copper, can give protein-metal hazes at levels as low as 1.0 mg L -1 (Boulton et al. 1996). However, recent studies suggest that trace concentrations of transition metals are catalysts in a number of important oxidation reactions in wine involving oxygen, polyphenols, sulfur dioxide, and ethanol (Laurie and Waterhouse 2006b, Danilewicz 2007, Danilewicz et al. 2008, Elias et al. 2008). The essential role of transition metals in the reaction between dioxygen and biomolecules has also been well established (Buettner and Jurkiewicz 1996). According to our proposed wine oxidation scheme, reduced transition metals (e.g., iron, copper, and potentially manganese) are essential for catalyzing the thermodynamically favorable, yet kinetically unfavorable, reaction 1 Assistant professor of Food Chemistry, Department of Food Science, 336 Food Science Building, Pennsylvania State University, University Park, PA 16802; 2 Associate professor of Food Chemistry, 3 Professor of Food Chemistry, Department of Food Science, University of Copenhagen, Rolighedsvej 30, Frederiksberg, Denmark; and 4 Professor of Enology, Department of Viticulture and Enology, University of California, Davis, CA 95616. *Corresponding author (email: alwaterhouse@ucdavis.edu; fax: 530 752-0382) Acknowledgments: This research was partially funded by the American Vineyard Foundation California Competitive Grant Program for Research in Viticulture and Enology. Manuscript submitted Feb 2009, revised May 2009, accepted Jun 2009 Copyright 2009 by the American Society for Enology and Viticulture. All rights reserved. between oxygen and polyphenols to yield hydrogen peroxide. Furthermore, metals are required in order to reduce hydrogen peroxide to the reactive hydroxyl radical via the Fenton reaction. We hypothesize that this radical species is responsible for the oxidation of ethanol to acetaldehyde. The oxidation of a number of other higher alcohols, organic acids, and carbohydrates in wine may result from hydroxyl radical-mediated oxidation reactions. Examples may include glyceraldehyde from glycerol (Danilewicz 2003, Laurie and Waterhouse 2006a), glyoxylic acid from tartaric acid (Fulcrand et al. 1997, Es-Safi et al. 2000), and pyruvic acid from malic acid (Fulcrand et al. 1998, Laurie and Waterhouse 2006b). Very few studies have addressed the identification or even detection of free radical intermediates in wine. In contrast, many key free radical mechanisms have been elucidated in beer (Uchida and Ono 1996, 1999, Uchida et al. 1996, Andersen and Skibsted 1998, Heyerick et al. 2003, 2005, Huvaere et al. 2005, Frederiksen et al. 2008), which has given brewers valuable tools for the control of oxidation in their processes. The use of electron paramagnetic resonance (EPR) spectroscopy has been central to the study of beer redox chemistry by allowing for the direct observation of radical species. In particular, the technique known as spin trapping has been used with great success to detect highly reactive (i.e., short-lived) radical species in beer that would otherwise be undetectable by direct EPR observation. Spin traps are often diamagnetic nitrone or nitroso compounds that are EPR silent until they react with free radicals. The resulting paramagnetic products (known as spin adducts ) are detectable by EPR because they are exceedingly stable compared to the initial radical, often with half-lives of several hours. Free radicals can also often be 471

472 Elias et al. identified due to hyperfine splitting patterns in their EPR spectrum, thus giving powerful insight into the mechanistic details of many key oxidation reactions in beer. The theory and practice of EPR spin trapping has been extensively reviewed elsewhere (Buettner 1987, Buettner and Jurkiewicz 1996). By using a variety of spin-trapping techniques, we detected at least three different free radical species in wine under varying conditions, both in a model system and in some actual wines (Elias et al. 2009). The objective of the present study was to determine the wine composition factors related to oxidation reaction mechanisms that affect the production of free radical intermediates in a real wine system undergoing forced aerobic aging. These factors included the addition of small quantities of transition metals or various concentrations of sulfur dioxide in a red wine and the supplementation of different amounts of catechin in a white wine. The antioxidative or prooxidative effects of each treatment were individually assessed. The ultimate goal of this research is to identify the key constituents in wine that either promote or inhibit oxidation in order to suggest potential strategies for oxygen management. Materials and Methods Materials. The spin trap α-(4-pyridyl-1-oxide)-n-tertbutylnitrone (POBN) was obtained from Sigma Aldrich (St. Louis, MO) and was used as received. Iron (II) sulfate heptahydrate (Merck, Darmstadt, Germany) and copper (II) sulfate pentahydrate (Sigma) were analytical grade. Potassium metabisulfite was purchased from Bie & Berntsen (Rødovre, Denmark). (+)-Catechin hydrate (99%) was obtained from Sigma. All other chemicals and solvents were of analytical or HPLC grade. Water was purified through a Millipore Q-Plus (Millipore Corp., Bedford, MA) purification train. The red wine used in this study was a 2006 Cabernet Sauvignon from the Valle del Rapel (Chile), containing 13.5% (v/v) ethanol and a total (free and bound) sulfur dioxide (SO 2 ) concentration of 88 mg L -1. The white wine was a 2007 Viognier (Vin de Pays d Oc) from Languedoc- Roussillon (France), containing 13% (v/v) ethanol and a total (free and bound) SO 2 concentration of 96 mg L -1. Total SO 2 was measured using a kit adapted from Rebelein s method (C. Schliessmann Kellerei-Chemie, Schwäbisch Hall, Germany). Sulfur dioxide removal. Final SO 2 of the red wine was adjusted downward by the slow, stepwise addition of hydrogen peroxide (3% v/v). Briefly, three additions of H 2 O 2 (16 μl) were added to the wine (50 ml, containing 88 mg L -1 total SO 2 ) at 20-min intervals. The wine was protected from light and was mechanically agitated with a magnetic stir bar under a nitrogen gas headspace. Final total SO 2 was 20 mg L -1. Metal analysis. The concentrations of endogenous transition metals in the wines were measured by inductively coupled plasma-mass spectrometry (ICP-MS). Wine (25 ml) was concentrated to dryness using rotational vacuum concentration. The precipitate was resuspended in 250 µl ultrapure HNO 3 (70%). Acid digestion and ICP-MS analysis of Fe, Cu, and Mn were performed as previously described (Persson et al. 2006). EPR spin trapping. POBN was dissolved directly into the wine samples (1 2 ml) before the start of each experiment to achieve a 30 mm concentration of the spin trap. The samples were kept in capped glass culture tubes with a headspace volume of ~8 ml at either room temperature or at 55 C in a heated water bath. All wine samples were protected from light during the study. When required, catechin was added by directly dissolving the compound in white wine (WW). In the added metal treatments, Fe(II) and/or Cu(II) were added (10 μl) from freshly prepared stock solutions in deoxygenated water. Samples (50 μl) were loaded into 50-μL micropipettes (Brand GmbH, Wertheim, Germany) and the EPR spectra were recorded on a Miniscope MS 200 X-band spectrometer (Magnettech, Berlin, Germany) at room temperature. The EPR microwave power was set to 10 mw, the modulation frequency was 1000 mg, and a sweep time of 60 sec was used with a sweep width of 68 G. The receiver gain was set to either 90 or 900, depending on the experiment and abundance of spin adducts. EPR calibration was performed using a 2 μm solution of 2,2,6,6-tetramethylpiperidine-1-oxyl (Sigma). Simulation and fitting of the EPR spectra were performed using the PEST WinSIM program (Duling 1994). Statistics. All experiments were performed in either duplicate or triplicate. EPR data were analyzed by one-way analysis of variance (ANOVA) in which means were compared at either the 0.01 or 0.05 level of significance. The integrated statistical software component of SigmaPlot 11.0 was used to perform the analysis. Results and Discussion Identification of trapped radicals. POBN was used in all experiments to detect free radical species generated in wine samples. This spin trap yields particularly stable adducts with long half-lives, making it amenable for storage studies at high temperatures. While the formation of many different types of radicals is predicted in wine, only one radical species was trapped by POBN in the experiments reported herein. A six-line spectrum was observed (Figure 1) in all cases and across all treatments. The intensity of the signal, as was assessed by comparing peak amplitudes, varied by treatment and increased as a function of time. The hyperfine coupling constants of the observed spectrum (a N = 15.4 G, a H = 2.6 G) were nearly identical to the values for the POBN spin adducts formed from the 1-hydroxyethyl radical (MeCH OH) (Buettner 1987), which is thought to be the major species resulting from the hydroxyl radicalmediated oxidation of ethanol. This observation suggests that (1) the 1-hydroxyethyl radical is the most quantitatively abundant radical species in oxidized wine, which is consistent with previous studies involving EPR spin trapping in oxidized wine (Elias et al. 2009) and beer (Andersen and Skibsted 1998, Andersen et al. 2000, Frederiksen et al. 2008), and that (2) the ethanol in wine is oxidized by

Radical Formation in Wine 473 potent hydroxyl radicals ( OH), which are most likely generated by the metal-catalyzed reduction of hydrogen peroxide (Scheme 1). The fate of the 1-hydroxyethyl radical is acetaldehyde in both the presence and absence of oxygen (Danilewicz 2003). Effect of transition metals on radical formation. A red wine containing relatively low levels of iron (0.627 mg L -1 ) and copper (0.046 mg L -1 ) was selected, to which varying concentrations of either iron, copper, or iron and copper in combination were added (Table 1). Before addition of the metals, the total SO 2 concentration of the wine was reduced to 20 mg L -1. No POBN/MeCH OH spin adducts were observed in the control wine (no added metals) within 150 min of incubation at room temperature in the absence of light (Figure 2). However, when 5 mg L -1 Fe 2+ was added ([Fe] total = 5.627 mg L -1 ) to the same wine, POBN/MeCH OH spin adducts could clearly be observed after 60 min. Since the hydroxyl radical-mediated oxidation of ethanol is the only viable route for 1-hydroxyethyl radical production in wine, these data provide direct evidence that low concentrations of transition metals are key catalysts in wine oxidation. The final concentration of iron in this treatment (5.627 mg L -1 ) is indeed realistic, as the global average concentration of iron in wines is ~5.5 mg L -1. When 15 mg L -1 Fe 2+ was added to the same wine, a relatively high number of POBN/ MeCH OH spin adducts were observed after only 10 min of dark incubation at room temperature. Clearly, a key factor controlling the overall rate of ethanol oxidation is the endogenous metal concentration of a wine. The effect of added copper on the generation of 1-hydroxyethyl radicals in the same low-sulfite red wine was also investigated. A marked synergism between iron and copper in promoting hydrogen peroxide formation (as Scheme 1 Proposed metal-catalyzed wine oxidation mechanism: reaction between the 1-hydroxyethyl radical and the POBN spin trap. Table 1 Concentrations of trace chemicals in red (RW) and white (WW) wines. Analyte RW (mg L -1 ) WW (mg L -1 ) Boron 4.609 9.037 Aluminum 0.573 0.640 Cadmium 0.003 0.004 Calcium 63.971 75.200 Copper a 0.046 0.167 Iron a 0.627 3.926 Magnesium 64.057 127.086 Manganese a 1.492 1.429 Molybdenum 0.021 0.026 Nickel 0.046 0.038 Phosphorus 144.114 289.714 Potassium 864.857 1239.714 Selenium a nd b nd Sodium 0.072 0.042 Sulfur 168.600 337.429 Zinc 1.120 0.712 a Oxidation reactive metals. b Not detected. Figure 1 EPR spectrum of the POBN/1-hydroxyethyl radical (MeCH OH) spin adduct. The upper and lower trace represent experimental and simulated spin patterns, respectively. Figure 2 Effect of added iron, copper, or iron and copper in combination on the concentration of spin adducts measured by EPR in a reduced sulfite (20 mg L -1 total SO 2 ) red wine. Metal concentrations listed represent the amount of added metals (i.e., not final metal concentrations). The wine samples contained POBN (30 mm) and were held at room temperature under aerial conditions in the absence of light.

474 Elias et al. measured indirectly by following SO 2 loss) in a model wine system has been observed (Danilewicz 2007): the rate of SO 2 loss in a model solution containing iron (5 mg L -1 ) and copper (0.05 mg L -1 ) far exceeded the sum of the rates when the metals were tested individually (5 mg L -1 iron or 0.05 mg L -1 copper). In fact, when a relatively high concentration of copper (0.15 mg L -1 ) was added, only a modest increase in rate was observed. Danilewicz (2007) proposed that copper plays an important role in regenerating the catalytically active ferrous ion (Fe 2+ ) from the oxidized ferric (Fe 3+ ) state, possibly through the direct interaction between oxygen and copper (e.g., CuO 2 intermediates). The matrix used in our study was a commercial red wine, and therefore contained a much higher baseline copper concentration (0.046 mg L -1 ) than typically would be found in a carefully controlled model wine solution. It was therefore impossible to assess the effect that either metal had individually on the rate of spin adduct formation. However, a marked increase in POBN/MeCH OH spin adducts was observed after 150 min when 0.2 mg L -1 copper was added ([Fe] total = 0.627 mg L -1 ; [Cu] total = 0.246 mg L -1 ). The addition of 5 mg L -1 iron to the same system ([Fe] total = 5.627 mg L -1 ; [Cu] total = 0.246 mg L -1 ) resulted in roughly the same number of POBN/MeCH OH spin adducts as the 15 mg L -1 added iron treatment after 10 min, thus demonstrating the importance of copper as a catalyst in this system. A treatment containing 15 mg L -1 added iron and 0.8 mg L -1 added copper ([Fe] total = 15.627 mg L -1 ; [Cu] total = 0.846 mg L -1 ) resulted in the fastest accumulation of POBN/MeCH OH spin adducts as compared with all other treatments. Effect of sulfur dioxide on radical formation. Sulfur dioxide is well known for its antioxidant properties in wine, yet its mechanism of action has been subject for debate. Some have argued that SO 2 prevents oxidation by reacting directly with oxygen; however, this reaction is not predicted under wine conditions (Danilewicz 2007). It seems more likely that the antioxidant activity of SO 2 stems from its ability to react quickly and irreversibly with hydrogen peroxide (Scheme 2), thereby diverting this key prooxidant from subsequent reactions with reduced metals that yield radical species (e.g., hydroxyl radicals). Sulfur dioxide additions were made to a red wine in order to test the dependence of 1-hydroxyethyl radical production on SO 2 concentration. The following total SO 2 levels were established: 20, 40, 60, 80, and 100 mg L -1. The wine samples were placed in a 50 C water bath in the absence of light. POBN/MeCH OH spin adducts were observed in all treatments after 90 min (Figure 3). The concentration of spin adducts continued to increase after 13.3 hr (800 min) in all treatments, at which point an inverse correlation between SO 2 and 1-hydroxyethyl radical concentrations was clearly observable. This trend continued after 30 hr (1800 min) for the 20, 40, and 60 mg L -1 SO 2 treatments; however, no significant difference (p 0.05) was observed between the 80 and 100 mg L -1 SO 2 treatments. By 65 hr (3900 min), approximately the same number of POBN/MeCH OH spin adducts were observed in the 20, 40, and, surprisingly, 100 mg L -1 SO 2 treatments. The rate of spin adduct degradation may have exceeded spin adduct formation in the 20 and 40 mg L -1 SO 2 treatments, however, which may explain this result. A similar effect of SO 2 concentration of radical formation has been reported in beer (Andersen et al. 2000), which typically contains small amounts of naturally occurring yeast-derived sulfite. The generation of 1-hydroxyethyl radicals in beer has been shown to follow a lag phase, the length of which correlates with sulfite concentration (Uchida et al. 1996, Andersen et al. 2000). The duration of this lag phase when beer is exposed to high temperature and aerobic conditions is also correlated with the future flavor stability of that beer and is the basis of a commonly used accelerated beer aging assay (Uchida et al. 1996). No such Scheme 2 Fate of hydrogen peroxide produced by polyphenol oxidation in wine. The Fenton reaction route is represented by path A; the nonradical reaction with bisulfite is represented by path B. Figure 3 Effect of added SO 2 on the concentration of spin adducts measured by EPR in a reduced sulfite (20 mg L -1 total SO 2 ) red wine. Concentrations listed represent the final concentrations of total SO 2. The wine samples contained POBN (30 mm) and were held at 50 C in the absence of light.

Radical Formation in Wine 475 pronounced lag phase was observed in the above-described wine system, however. Effect of polyphenols on radical formation. Contrary to the popular view that polyphenols are antioxidants, many studies have demonstrated that the presence of polyphenols particularly those with catechol groups leads to the production of hydrogen peroxide under wine conditions (Wildenradt and Singleton 1974, Singleton 1987, Danilewicz 2003). In fact, the rate of SO 2 loss appears to be dependent on the concentration of polyphenols in wine, yet these compounds also seem to be important scavengers of some potentially destructive radical species (Danilewicz 2007). Clearly, the role of polyphenols in the general wine oxidation scheme is complex and nuanced. The effect of polyphenol concentration on 1-hydroxyethyl radical formation was evaluated in a white wine containing 96 mg L -1 SO 2. Increasing concentrations of polyphenols were established in the wine by the addition of (+)-catechin (100 mg L -1 to 3.5 g -1 added). The wine samples were kept in a 50 C water bath in the absence of light. Measurements were taken at 0, 30, 120, 360, 1035, and 3375 min. A lag phase of at least 120 min was observed in all treatments (including the control), in which no POBN/MeCH OH spin adducts could be detected (Figure 4). The generation of spin adducts was observed only after 360 min of incubation in all treatments. Depending on the level of significance applied, the difference between treatments was either significant (p = 0.05) or insignificant (p = 0.01) at this time point. In other words, a very similar number of spin adducts at 360 min was observed across all treatments, which suggests that catechin was neither an antioxidant nor a prooxidant during this early stage. A similar observation was made in beer, wherein the presence of ferulic acid, catechin, or prodelphinidin B-3 (all 0.2 mm) had no effect on the rate of formation of N-tert-butyl-α-phenylnitrone (PBN)/MeCH OH spin adducts within ~280 min at 55 C (Andersen et al. 2000). Therefore, it is proposed that the sole antioxidant capable of preventing ethanol oxidation in wine is SO 2, as is also the case under beer conditions. Perhaps paradoxically, a greater concentration of spin adducts were observed in the highest catechin concentration treatment (3.5 g L -1 ) after 17.25 hr (1035 min). The same trend was observed after 56.25 hr (3375 min), in which increasing concentrations of catechin resulted in increasing concentrations of spin adducts. This may demonstrate the prooxidant effect of polyphenols in wine. Wildenradt and Singleton showed that the production of volatile aldehydes in a model wine was only possible in the presence of phenols (e.g., pyrogallol, 2,3,4-trihydroxybenzoic acid, gallic acid) (Wildenradt and Singleton 1974). It was reported that the generation of aldehydes resulted from oxidation of ethanol by hydrogen peroxide, a by-product of phenol oxidation. Recently, Danilewicz demonstrated that the rate of SO 2 loss due to its reaction with hydrogen peroxide in a model wine system was dependent on the concentration of 4-methylcatechol (Danilewicz 2007). The same effect was apparently seen in the present study, in which the concentration of POBN/MeCH OH spin adducts was dependent on the concentration of catechin. A discrete lag phase was observed, presumably because of the presence of excess SO 2, followed by a steady increase in spin adducts as SO 2 was consumed by its reaction with hydrogen peroxide. By the later stages of the study, it is possible that there was very little (or any) remaining SO 2 to consume the hydrogen peroxide generated by catechin oxidation. It is hypothesized that in this system oxygen is brought into the wine by its direct reaction with a reduced metal (e.g., iron, copper) to yield a hydroperoxyl radical (Scheme 3). These endogenous transition metals are maintained in their catalytically active reduced state by the pool of wine polyphenols, which reduce Fe 3+ under wine conditions. The hydroperoxyl radicals formed in this process are rapidly quenched by phenols to give hydrogen peroxide that feeds the Fenton reaction, which may explain the concentration dependence of catechin on ethanol oxidation in this study. Figure 4 Effect of added catechin on the concentration of spin adducts measured by EPR in a white wine. Concentrations listed represent added concentrations of catechin (i.e., not final catechin concentrations). The wine samples contained POBN (30 mm) and were held at 50 C in the absence of light. Scheme 3 Proposed effect of polyphenols in promoting ethanol oxidation in wine.

476 Elias et al. Conclusions This study has demonstrated the effect of various components on the formation or suppression of free radicals in wine. In keeping with our hypothesized wine oxidation scheme, trace levels of added transition metals were shown to markedly increase the initial rate of free radical formation in a red wine. Metals appear to be essential catalysts in the oxidation of wine polyphenols (via the hydroperoxyl radical) and ethanol (via the hydroxyl radical). The sole radical observed in this study was the 1-hydroxyethyl radical, which demonstrates that it is the most quantitatively important radical species in oxidized wine and that the Fenton reaction is a major route for ethanol oxidation. The 1-hydroxyethyl radical was again trapped in the same wine to which varying concentrations of SO 2 were added. In this case, SO 2 slowed the rate of spin adduct generation and was shown to inhibit the formation of 1-hydroxyethyl radicals in a concentration-dependent manner. Perhaps paradoxically, the addition of catechin to a white wine had no effect on the initial rate of radical formation, despite the fact that polyphenols have been widely established as antioxidants. This observation is once again consistent with our proposed wine oxidation scheme, as it suggests that ethanol is oxidized by a strong oxidant (i.e., the hydroxyl radical) that reacts with wine constituents in proportion to concentration, thus in large part bypassing the pool of polyphenols (a relatively minor species in wine compared to ethanol). In the latter stages of the experiment, treatments with higher concentrations of added catechin were seen to yield more spin adducts. The apparent prooxidative effect of catechin here may result from hydrogen peroxide production coupled to the polyphenol oxidation, as has been previously reported. Literature Cited Andersen, M.L., and L.H. Skibsted. 1998. Electron spin resonance spin trapping identification of radicals formed during aerobic forced aging of beer. J. Agric. Food Chem. 46:1272-1275. Andersen, M.L., H. Outtrup, and L.H. Skibsted. 2000. Potential antioxidants in beer assessed by ESR spin trapping. J. Agric. Food Chem. 48:3106-3111. Boulton, R.B., V.L. Singleton, L.E. Bisson, and R.E. Kunkee. 1996. Principles and Practices of Winemaking. Chapman & Hall, New York. Buettner, G.R. 1987. Spin trapping: ESR parameters of spin adducts. Free Radic. Biol. Med. 3:259-303. Buettner, G.R., and B.A. Jurkiewicz. 1996. Catalytic metals, ascorbate and free radicals: Combinations to avoid. Radiat. Res. 145:532-541. Danilewicz, J.C. 2003. Review of reaction mechanisms of oxygen and proposed intermediate reduction products in wine: Central role of iron and copper. Am. J. Enol. Vitic. 54:73-85. Danilewicz, J.C. 2007. Interaction of sulfur dioxide, polyphenols, and oxygen in a wine-model system: Central role of iron and copper. Am. J. Enol. Vitic. 58:53-60. Danilewicz, J.C., J.T. Seccombe, and J. Whelan. 2008. Mechanism of interaction of polyphenols, oxygen, and sulfur dioxide in model wine and wine. Am. J. Enol. Vitic. 59:128-136. Duling, D.R. 1994. Simulation of multiple isotropic spin-trap EPR spectra. J. Magn. Reson., Ser. B 104:105-110. Elias, R.J., M.L. Anderson, L.H. Skibsted, and A.L. Waterhouse. 2009. Identification of free radical intermediates in oxidized wine using electron paramagnetic resonance spin trapping. J. Agric. Food Chem. 57:4359-4365. Elias, R.J., V.F. Laurie, S.E. Ebeler, J.W. Wong, and A.L. Waterhouse. 2008. Analysis of selected carbonyl oxidation products in wine by liquid chromatography with diode array detection. Anal. Chim. Acta 626:104-110. Es-Safi, N.E., C. Le Guerneve, V. Cheynier, and M. Moutounet. 2000. New phenolic compounds formed by evolution of (+)-catechin and glyoxylic acid in hydroalcoholic solution and their implication in color changes of grape-derived foods. J. Agric. Food Chem. 48:4233-4240. Frederiksen, A.M., R.M. Festersen, and M.L. Andersen. 2008. Oxidative reactions during early stages of beer brewing studied by electron spin resonance and spin trapping. J. Agric. Food Chem. 56:8514-8520. Fulcrand, H., C. Benabdeljalil, J. Rigaud, V. Cheynier, and M. Moutounet. 1998. A new class of wine pigments generated by reaction between pyruvic acid and grape anthocyanins. Phytochemistry 47:1401-1407. Fulcrand, H., V. Cheynier, J. Oszmianski, and M. Moutounet. 1997. An oxidized tartaric acid residue as a new bridge potentially competing with acetaldehyde in flavan-3-ol condensation. Phytochemistry 46:223-227. Heyerick, A., K. Huvaere, D. De Keukeleire, and M.D.E. Forbes. 2005. Fate of flavins in sensitized photodegradation of isohumulones and reduced derivatives: Studies on formation of radicals via EPR combined with detailed product analyses. Photochem. Photobiol. Sci. 4:412-419. Heyerick, A., Y.N. Zhao, P. Sandra, K. Huvaere, F. Roelens, and D. De Keukeleire. 2003. Photolysis of hop-derived trans-iso-α-acids and trans-tetrahydroiso-α-acids: Product identification in relation to the lightstruck flavour of beer. Photochem. Photobiol. Sci. 2:306-314. Huvaere, K., M.L. Andersen, L.H. Skibsted, A. Heyerick, and D. De Keukeleire. 2005. Photooxidative degradation of beer bittering principles: A key step on the route to lightstruck flavor formation in beer. J. Agric. Food Chem. 53:1489-1494. Laurie, V.F., and A.L. Waterhouse. 2006a. Glycerol oxidation in wine and reactions with f lavonoids. Abstr. Am. J. Enol. Vitic. 57:394A-395A. Laurie, V.F., and A.L. Waterhouse. 2006b. Oxidation of glycerol in the presence of hydrogen peroxide and iron in model solutions and wine. Potential effects on wine color. J. Agric. Food. Chem. 54:4668-4673. Persson, D.P., T.H. Hansen, P.E. Holm, J.K. Schjoerring, H.C.B. Hansen, J. Nielsen, I. Cakmak, and S. Husted. 2006. Multi-elemental speciation analysis of barley genotypes differing in tolerance to cadmium toxicity using SEC-ICP-MS and ESI-TOF-MS. J. Anal. At. Spectrom. 21:996-1005. Singleton, V.L. 1987. Oxygen with phenols and related reactions in musts, wines, and model systems: Observations and practical implications. Am. J. Enol. Vitic. 38:69-77. Uchida, M., and M. Ono. 1996. Improvement for oxidative flavor stability of beer Role of OH-radical in beer oxidation. J. Am. Soc. Brew. Chem. 54:198-204. Uchida, M., and M. Ono. 1999. Determination of hydrogen peroxide in beer and its role in beer oxidation. J. Am. Soc. Brew. Chem. 57:145-150. Uchida, M., S. Suga, and M. Ono. 1996. Improvement for oxidative flavor stability of beer Rapid prediction method for beer flavor stability by electron spin resonance spectroscopy. J. Am. Soc. Brew. Chem. 54:205-211. Wildenradt, H.L., and V.L. Singleton. 1974. Production of aldehydes as a result of oxidation of polyphenolic compounds and its relation to wine aging. Am. J. Enol. Vitic. 25:119-126.