Article pubs.acs.org/JAFC
Impact of Bottle Aging on Smoke-Tainted Wines from Different Grape Cultivars Renata Ristic,†,‡ Lieke van der Hulst,†,‡ Dimitra L. Capone,§ and Kerry L. Wilkinson*,†,‡ †
The Australian Research Council Training Centre for Innovative Wine Production and ‡School of Agriculture, Food and Wine, The University of Adelaide, PMB 1, Glen Osmond, South Australia 5064, Australia § The Australian Wine Research Institute (AWRI), P.O. Box 197, Glen Osmond, South Australia 5064, Australia S Supporting Information *
ABSTRACT: Smoke taint is the term given to the objectionable smoky, medicinal, and ashy characters that can be exhibited in wines following vineyard exposure to bushfire smoke. This study sought to investigate the stability of smoke taint by determining changes in the composition and sensory properties of wines following 5 to 6 years of bottle aging. Small increases in guaiacol and 4-methylguaiacol (of up to 6 μg/L) were observed after bottle aging of smoke-affected red and white wines, while syringol increased by as much as 29 μg/L. However, increased volatile phenol levels were also observed in control red wines, which indicated that changes in the composition of smoke-affected wines were due to acid hydrolysis of conjugate forms of both naturally occurring and smoke-derived volatile phenols. Acid hydrolysis of smoke-affected wines (post-bottle aging) released additional quantities of volatile phenols, which demonstrated the relative stability of glycoconjugate precursors to the mildly acidic conditions of wine. Bottle aging affected the sensory profiles of smoke-affected wines in different ways. Diminished fruit aroma and flavor led to the intensification of smoke taint in some wines, but smoke-related sensory attributes became less apparent in smoke-affected Shiraz wines, post-bottle aging. KEYWORDS: acid hydrolysis, bottle aging, cultivars, guaiacol glycoconjugates, smoke taint, volatile phenols, wine
■
INTRODUCTION
ences in the guaiacol glycoconjugate content of fruit harvested from the different cultivars (at maturity). Variation among grape glycoconjugate concentrations could not be explained by differences in canopy architecture or crop yield, which suggested that differences might instead reflect varietal responses to smoke exposure. Low levels of guaiacol glycoconjugates were found in fruit from control (unsmoked) vines, with the exception of control Shiraz vines, for which fruit contained between 7- and 16-fold higher glycoconjugate concentrations than for other cultivars.18 This observation was in agreement with previous research that suggests guaiacol is a natural component of Shiraz grapes.19 Importantly, there is no evidence to suggest that compositional effects of grapevine exposure to smoke are carried over between seasons. Neither guaiacol or 4-methylguaiacol nor obvious smoke-related sensory attributes were observed in a 2007 wine made from fruit harvested from grapevines repeatedly exposed to smoke during the 2005/06 growing season,8 while guaiacol glycoconjugates could not be detected in fruit harvested from grapevines grown in a vineyard partially burned during a bushfire in the growing season following the fire.20 During fermentation, glycoconjugate precursors can be hydrolyzed to release volatile phenols,9,21 but a significant portion of the glycoconjugate pool remains in the final wine.5,9,11,18 The volatile phenol content of a smoke tainted Pinot Noir wine was monitored following amelioration by
Over the past decade, considerable research has been undertaken in response to the increasing incidence of vineyard exposure to smoke reported in wine producing countries including Australia, Canada, North America and South Africa. A number of volatile phenols, including guaiacols, cresols, and syringols, have been identified in smoke-affected grapes and wine.1−6 These compounds were subsequently used as markers of smoke taint in studies that (i) demonstrated that the intensity of smoke-related sensory attributes in wine is influenced by the timing and duration of grapevine exposure to smoke7,8 and techniques employed during winemaking9,10 or (ii) evaluated methods for amelioration of smoke taint using defoliation of grapevines11 or treatment of smoke tainted wine by reverse-osmosis and solid phase adsorption12 or with commercial fining agents.13 A seminal study reporting the progressive release of volatile phenols during alcoholic and malolactic fermentation of smokeaffected Merlot juice2 provided the first evidence of smokederived volatile compounds being accumulated in grapes in precursor forms. This was confirmed following identification of the β-D-glucopyranoside of guaiacol in juice from smokeaffected grapes.14 Several studies have since shown that volatile phenols are taken up by grapevine leaves and fruit but are quickly glycosylated to give glucoside, glucose-glucoside, pentose-glucoside, or rutinoside precursors.15−17 Indeed, glycosylated forms of guaiacol were detected in Merlot and Viognier grapes within a few days of grapevines being exposed to smoke.15 A recent study involving the application of smoke to seven different grapevine cultivars 18 (for 1 h, at approximately 7 days postveraison) found significant differ© 2017 American Chemical Society
Received: Revised: Accepted: Published: 4146
March 20, 2017 April 29, 2017 May 3, 2017 May 3, 2017 DOI: 10.1021/acs.jafc.7b01233 J. Agric. Food Chem. 2017, 65, 4146−4152
Article
Journal of Agricultural and Food Chemistry
coupled to a 5973 mass selective detector using SIDA methods described previously.4,30 Benzyl mercaptan was measured in a subset of wines considered to reflect the more heavily smoke-affected wines (based on chemical and sensory data), that is, the 2010 Shiraz, 2010 Pinot Gris, 2011 Shiraz, and 2011 Merlot wines, together with their corresponding control wines. Quantitation was performed (in 2016, i.e., post-bottle aging) according to published SIDA methodology25 using an Agilent 1200 HPLC coupled to an Applied Biosystems 4000 QTrap hybrid tandem mass spectrometer operating in ESI mode. The preparation of isotopically labeled internal standards (d4guaiacol β-D-glucopyranoside, d3-guaiacol, d3-4-methylguaiacol, d7-ocresol, d3-syringol, and d5-benzyl mercaptan), validation of all SIDA methods, and instrumental operating conditions are described extensively in the aforementioned publications.4,15,24,25,30 Sensory Analysis. Descriptive analysis (DA) of 2010 and 2011 control and smoke-affected wines was performed in previous studies18,26 but repeated in the current study following bottle aging with a trained panel comprising 11 research staff and students (6 female and 5 male) from the University of Adelaide and AWRI (6 of whom had prior experience tasting smoke tainted wines, including as participants in the DA of wines in 2010 and 2011). Prior to formal evaluation, the panel underwent a series of training sessions (4 × 2 h sessions for experienced panelists and an additional 2 × 2 h for new panelists) according to published methodology.31 The panel first evaluated control and smoke-affected wines and reached consensus that the sensory attributes generated during prior DA of wines (Table 1) were appropriate for use in the current study. During subsequent
reverse osmosis and solid phase adsorption, and a gradual increase in guaiacol and 4-methylguaiacol concentrations was observed over a 30 month period.12 This was attributed to the slow acid hydrolysis of glycoconjugate precursors, similar to that reported for glycoconjugates of 3-methyl-4-hydroxyoctanoic acid during oak maturation of wine.22 The susceptibility of glycoconjugate forms of smoke-derived volatile phenols to hydrolysis in the acidic wine medium would therefore be expected to influence the intensity of smoke taint with time. As such, this study aimed to investigate the stability of smoke taint in wine by determining changes in wine composition and sensory properties during bottle aging. The study also included quantitation of benzyl mercaptan, a potent volatile thiol associated with the aroma of smoke,23 to determine to what extent, if any, this compound might represent an additional marker of smoke taint.
■
MATERIALS AND METHODS
Chemicals. Chemicals (analytical grade) were purchased from Sigma-Aldrich (Steinheim, Germany and Castle Hill, NSW, Australia). Solvents (HPLC grade) were sourced from Merck (Darmstadt, Germany). Deuterium-labeled internal standards (d3-guaiacol, d3-4methylguaiacol, d7-o-cresol, d3-syringol, d4-guaiacol β-D-glucoside, and d5-benzyl mercaptan) were synthesized in house, as previously reported.4,15,24,25 Wine Samples. Control and smoke-affected wines were available from two previous studies18,26 involving the application of smoke to grapevines (for 1 h, at approximately 7 days postveraison) under experimental conditions described previously.2,9 Shiraz, Cabernet Sauvignon, Pinot Noir, Chardonnay, Sauvignon Blanc, and Pinot Gris wines were obtained from field trials conducted in 2010,18 while Shiraz, Merlot, Chardonnay, and Sauvignon Blanc wines were available from a field trial conducted in 2011.26 Wines were made according to small-lot winemaking procedures (alcoholic fermentation only, no wines underwent malolactic fermentation) before being bottled in 375 mL glass bottles, under screw cap closures, and cellared (in darkness) at 15 °C. Chemical and sensory analyses were performed 6 months after bottling (i.e., in 2010 or 2011), then again following 5−6 years bottle aging (i.e., in 2016). Preparation of Acid Hydrolysates. The 2010 wines were subjected to strong acid hydrolysis following 6 years of bottle aging using methodology described in previous studies.2,27 Briefly, hydrolysates were prepared by heating control or smoke-affected wine (10 mL, adjusted to pH 1.0 with the dropwise addition of concentrated sulfuric acid) at 100 °C for 1 h. Hydrolysates were cooled to ambient temperature, then frozen prior to chemical analysis. Chemical Analysis. Red wine color and/or total phenolics content of control and smoke-affected wines was determined by UV−vis spectrophotometry28 in 2010 and 2011 (i.e., post-bottling) and by rapid 96-well plate format versions of the methyl cellulose precipitable tannin assay and modified Somers color assay29 in 2016 (i.e., postbottle aging). The guaiacol glycoconjugate concentrations of 2010 wines were determined by liquid chromatography−tandem mass spectrometry (LC−MS/MS) according to the stable isotope dilution analysis (SIDA) method developed by Dungey and colleagues.15 Analyses performed in 2010 used a 4000 Q TRAP hybrid tandem mass spectrometer equipped with a Turbo ion source (Applied Biosystems/ MDS Sciex, Foster City, CA, USA) combined with an Agilent 1200 HPLC system equipped with binary pump, degasser, autosampler, and column oven (Agilent Technologies, Forest Hill, Vic, Australia), while 2016 analyses were performed on an Agilent 1290 Infinity UHPLC combined with a 6490 QQQ LC−MS with iFunnel technology. The concentrations of several volatile phenols (guaiacol, 4-methylguaiacol, p-, m-, and o-cresols, and syringol) were measured in 2010 and 2011 wines (before and after bottle aging) and in acid hydrolysates of 2010 wines (after bottle aging) by the Australian Wine Research Institute’s Commercial Services Laboratory (Adelaide, Australia). Volatile phenols were measured using an Agilent 6890 gas chromatograph
Table 1. Aroma and Palate Attributes Used for Sensory Analysis of Control and Smoke-Affected Wines attribute Aroma fruit smoke cold ash earthy medicinal Palate fruit flavor smoky flavor ashy aftertaste woody aftertaste medicinal flavor metallic bitter drying acidity
description overall intensity of fruit aroma perception of smoke aroma, including smoked meat/bacon, toasty, charry, cigar box burnt aroma associated with ash, including ashtray, tarry, campfire aromas associated with musty, dusty, wet wood, barnyard, mushroom, dank, moldy aromas characteristic of band-aids, disinfectant, cleaning products, solvents overall intensity of fruit flavor perception of smoke flavor, including bacon and smoked meat length of taste associated with residue of ashtray perceived in the mouth after expectorating, including coal ash, ashtray, tarry, acrid, campfire length of taste associated with woody residue, including wood, oak, pencil shavings flavors characteristic of band-aids, disinfectant, cleaning products, solvents the “tinny” flavor associated with metals intensity of bitter taste/aftertaste drying, puckering mouthfeel after expectoration of the wine intensity of sour/acid taste
training sessions, panelists gained familiarity in recognizing and scoring the intensity of sensory attributes. Formal evaluation commenced once panel performance was deemed satisfactory (i.e., based on panel by sample interactions). Wines were formally evaluated by panelists in isolated booths under controlled conditions (i.e., 22−23 °C and sodium lighting) during four consecutive sensory sessions (two sessions for whites wines and two sessions for red wines). Wines (30 mL) were presented (in triplicate) in three-digit coded, covered, International Standards Organization standard wine glasses, in random order. Panelists rated the intensity of aroma, flavor, and mouthfeel attributes using an unstructured line scale, with anchor points of “low” 4147
DOI: 10.1021/acs.jafc.7b01233 J. Agric. Food Chem. 2017, 65, 4146−4152
Article
4148
3 ± 0.3 19 ± 2.0 2 ± 0.3 13 ± 2.0 1 ± 0.1 6 ± 0.6 nd 4 ± 0.7 nd 5 ± 1.2 nd 9 ± 1.5 nd 17 ± 2.5 nd 7 ± 1.8 nd 3 ± 0.3 nd 4 ± 0.6 nd 4 ± 0.9 nd 6 ± 1.0 19 ± 1.0 106 ± 11.9 3 ± 0.3 36 ± 7.2 3 ± 0.0 14 ± 2.2 nd 13 ± 1.0 nd 15 ± 3.2 nd 25 ± 4.0 ± 0.6 ± 2.6 ± 1.5 ± 1.0 ± 0.7 ± 2.1 nd 4 ± 0.3 nd 4 ± 0.9 nd 9 ± 0.7 31 39 25 32 10 14
± 0.0 ± 0.3 ± 0.0 ± 1.2 ± 0.3 ± 1.0 nd nd nd nd nd 4 ± 0.7 a
Pinot Gris
Sauvignon Blanc
Chardonnay
Pinot Noir
Cabernet Sauvignon
Shiraz
Values are means (±standard error) of three replicates (n = 3); nd = not detected; tr = trace (i.e., < 1 μg/L).
2 6 2 9 1 3
nd 3 ± 0.3 nd 4 ± 1.0 nd 2 ± 0.0 nd tr nd 1 ± 0.5 nd 3 ± 0.6 9 ± 0.3 30 ± 1.3 3 ± 0.3 19 ± 3.8 3 ± 0.3 8 ± 1.0 nd 2 ± 0.0 nd 4 ± 0.7 nd 12 ± 1.2 8 ± 0.3 10 ± 0.7 7 ± 0.3 10 ± 0.3 2 ± 0.0 3 ± 0.3 nd tr nd 1 ± 0.7 nd 2 ± 0.3 nd 2 ± 0.0 nd tr nd tr nd nd nd nd nd tr C S C S C S C S C S C S
9 ± 0.6 26 ± 2.0 2 ± 0.3 20 ± 4.4 nd 6 ± 1.0 nd 1 ± 0.6 nd 2 ± 0.3 nd 10 ± 0.9
3 ± 0.3 10 ± 0.7 5 ± 0.3 17 ± 2.7 1 ± 0.0 8 ± 1.0 nd tr nd tr nd 8 ± 1.2
total cresols (μg/L) 4-methylguaiacol (μg/L) 4-methylguaiacol (μg/L) guaiacol (μg/L)
total cresols (μg/L)
syringol (μg/L)
guaiacol (μg/L)
4-methylguaiacol (μg/L)
total cresols (μg/L)
syringol (μg/L)
guaiacol (μg/L)
post-acid hydrolysis post-bottle aging
RESULTS AND DISCUSSION Influence of Bottle Aging on Color and Total Phenolics of Wines. The color and/or phenolic profile(s) of wines changes during storage and aging as a consequence of chemical transformations of anthocyanins (for red wines), tannins, and other phenolic compounds.28 It was therefore not surprising that changes were observed in the color density, hue, and total phenolics of control and smoke-affected wines following bottle aging (Supplementary Table S1). The color density of Shiraz and Pinot Noir wines declined with aging, while hue increased, but significant differences were not observed between control and smoke-affected wines. The total phenolic content of red wines also decreased substantially. A significant difference in total phenolics was observed between 2010 control and smoke-affected Pinot Noir wines post-bottle aging, but at 0.4 and 2.0 au, respectively, differences may not have been meaningful from a sensory perspective. In the case of white wines, differences were not observed between the total phenolic content of 2010 control and smoke-affected wines post-bottling. The phenolic content of these wines increased slightly with bottle aging, with the control Sauvignon Blanc containing slightly higher phenolic levels than the smokeaffected Sauvignon Blanc post-bottle aging. In contrast, 2011 smoke-affected wines had moderately higher phenolic concentrations than their corresponding control wines after both bottling and bottle aging. Influence of Bottle Aging on Concentrations of Volatile Phenols and Guaiacol Glycoconjugates in 2010 Wines. Small increases, that is, up to 4 μg/L, in guaiacol or 4-methylguaiacol concentrations were observed following bottle aging of smoke-affected wines as well as control Cabernet Sauvignon and Pinot Noir wines (Table 2). Considerably larger increases in syringol levels were observed, that is, 11−29 and 3−7 μg/L increases for smoke-affected red and white wines, respectively, and 8−23 μg/L increases for control red wines. The increases that were observed in the guaiacol or syringol concentrations of control red wines not only indicate the presence of precursor forms of these volatiles as natural grape constituents, but also that hydrolysis of these precursors during bottle aging would account for at least some of the volatile phenol formation observed in smoke-affected wines, especially in the case of syringol. Interestingly, bottle aging resulted in decreased concentrations of cresols in both smoke-affected wines and control Shiraz and Cabernet Sauvignon wines, while levels remained the same, at 1 μg/L, in control Pinot Noir wine. Of the volatile phenols measured, none was detected in control white wines either before or after bottle aging. Acid hydrolysis treatment of control white wines did not yield detectable levels of volatile phenols either (Table 2). This is likely attributable to exceptionally low glycoconjugate levels in control wines; certainly guaiacol glycoconjugate levels were barely detectable, at ≤10 μg/L, post-bottling, and
post-bottling
■
treatment
Table 2. Concentrations of Volatile Phenols in 2010 Control (C) and Smoke-Affected (S) Wines Following Bottling, Bottle Aging, and Acid Hydrolysisa
syringol (μg/L)
(at 10% of the line) and “high” (at 90% of the line). Panelists rinsed thoroughly with a solution of pectin (1 g/L) and with water and rested for at least 60 s between samples. Data were acquired with RedJadeFizz software (Version 2.4, Biosystemes, Couternon, France). Data Analysis. Chemical data were analyzed by one- or two-way analysis of variance (ANOVA) using GenStat (15th Edition, VSN International Limited, Herts, UK). Sensory data were analyzed using SenPAQ (version 5.01, Qi Statistics, Reading, UK) and XLSTAT (version 2015.1, Addinsoft, New York, USA). Mean comparisons were performed by least significant difference (LSD) multiple comparison test at P < 0.05.
38 ± 0.7 153 ± 17.5 26 ± 0.6 120 ± 15.6 2 ± 0.3 35 ± 4.5 nd 22 ± 1.3 nd 30 ± 6.1 nd 44 ± 5.3
Journal of Agricultural and Food Chemistry
DOI: 10.1021/acs.jafc.7b01233 J. Agric. Food Chem. 2017, 65, 4146−4152
Article
Journal of Agricultural and Food Chemistry
guaiacol glycoconjugate levels were higher in Shiraz, followed by Cabernet Sauvignon and then Pinot Gris. Similar levels of guaiacol glycoconjugates were detected in smoke-affected Chardonnay and Sauvignon Blanc wines, which yielded similar levels of guaiacol following both bottle aging and acid hydrolysis. The fact that the acid hydrolysate of smoke-affected Pinot Noir wine contained comparable levels of guaiacol to Chardonnay and Sauvignon Blanc hydrolysates, despite lower levels of guaiacol glycoconjugates, can be explained by the higher guaiacol concentrations of wine after both bottling and bottle aging. The release of volatile phenols during acid hydrolysis treatment of wines further demonstrates the presence of both naturally occurring volatile phenol precursors in wines made from red cultivars, and smoke-derived volatile phenol precursors in wines made from smoke-affected fruit. However, the release of substantial quantities of volatile phenols following acid hydrolysis of wines after 6 years of bottle aging demonstrates the relative stability of these precursors at wine pH. Indeed, there were no significant differences between the guaiacol glycoconjugate concentrations of 2010 wines post-bottling versus post-bottle aging (Table 3). The glycoconjugate profiles of wines may, however, have changed during bottle aging; that is, in a manner similar to that observed during the maturation of model wine extracts of oak wood, in which the partial hydrolysis of galloylglucoside and rutinoside precursors of 3-methyl-4-hydroxyoctanoic acid, to the corresponding glucoside precursor, was observed.22 The stability of volatile phenol glycoconjugates would undoubtedly be influenced by the number and position of ring substituents relative to the hydroxyl group. Since previous research has shown that enzymes present in human saliva can hydrolyze volatile phenol glycoconjugates,32 it should also be noted the glycoconjugates themselves can contribute to the smoke-related flavor and aftertaste associated with smoke taint, that is, as a consequence of in-mouth release of volatile phenols.32 Influence of Bottle Aging on Concentrations of Volatile Phenols in 2011 Wines. Similar changes to volatile phenol composition occurred during bottle aging of 2011 wines (Table 4). Small increases (up to 6 μg/L) in guaiacol or 4methylguaiacol concentrations were again observed following bottle aging of smoke-affected wines and control Shiraz and Merlot wines, while more substantial increases in syringol levels were observed, that is, 14−23 and 7−9 μg/L increases for smoke-affected red and white wines, respectively, and 8−17 μg/ L increases for control red wines. However, during bottle aging of 2011 wines, small increases (up to 4 μg/L) in total cresols
they were not detectable post-bottle aging (Table 3). In contrast, significant increases in volatile phenol levels were Table 3. Concentrations of Guaiacol Glycoconjugates in 2010 Control (C) and Smoke-Affected (S) Wines Following Bottling and Bottle Aginga guaiacol glycoconjugates (μg/L)
treatment
Shiraz
Cabernet Sauvignon
Pinot Noir
Chardonnay
Sauvignon Blanc
Pinot Gris
a
post-bottling
post-bottle aging
C
334 ± 5
379 ± 108
S
1480 ± 151
1405 ± 421
C
39 ± 6
54 ± 27
S
396 ± 78
555 ± 56
C
19 ± 1
18 ± 18
S
111 ± 17
147 ± 35
C
9±1
nd
S
213 ± 34
341 ± 53
C
10 ± 2
nd
S
180 ± 37
184 ± 57
C
8±1
nd
S
306 ± 66
497 ± 56
P value
0.001 (control vs smoked-affected) 0.952 (post-bottling post-bottle aging) 0.001 (control vs smoked-affected) 0.120 (post-bottling post-bottle aging)
