Article pubs.acs.org/JAFC
N,S,O‑Heterocycles in Aged Champagne Reserve Wines and Correlation with Free Amino Acid Concentrations Nicolas Le Menn,†,‡,§,⊥ Stephanie Marchand,*,†,‡ Gilles de Revel,†,‡ Dominique Demarville,⊥ Delphine Laborde,⊥ and Richard Marchal§ †
University of Bordeaux, ISVV, EA 4577, Unité de recherche OENOLOGIE, F-33882 Villenave d’Ornon, France INRA, ISVV, USC 1366 OENOLOGIE, F-33882 Villenave d’Ornon, France § University of Reims Champagne-Ardenne, URVVC EA 4707, B.P. 1039, 51687 Reims, Cedex 2, France ⊥ Champagne Veuve Clicquot, 13 rue Albert Thomas, 51100 Reims, France ‡
S Supporting Information *
ABSTRACT: Champagne regulations allow winegrowers to stock still wines to compensate for quality shifts in vintages, mainly due to climate variations. According to their technical requirements and house style, Champagne producers use these stored wines in their blends to enhance complexity. The presence of lees and aging at low pH (2.95−3.15), as in Champagne wines, lead to several modifications in wine composition. These conditions, combined with extended aging, result in the required environment for the Maillard chemical reaction, involving aromatic molecules, including sulfur, oxygen, and nitrogen heterocycles (such as thiazole, furan, and pyrazine derivatives), which may have a sensory impact on wine. Some aromatic heterocycles in 50 monovarietal wines aged from 1 to 27 years provided by Veuve Clicquot Ponsardin Champagne house were determined by the SPME-GC-MS method. The most interesting result highlighted a strong correlation between certain heterocycle concentrations and wine age. The second revealed a correlation between heterocyclic compound and free amino acid concentrations measured in the wines, suggesting that these compounds are potential aromatic precursors when wine is aged on lees and, thus, potential key compounds in the bouquet of aged Champagnes. The principal outcome of these assays was to reveal, for the first time, that aromatic heterocycle concentrations in Champagne base wines are correlated with wine age. KEYWORDS: heterocycles, wine aging, amino acids, Champagne, reserve wine
■
INTRODUCTION Champagne regulations allow winegrowers to stock still wines to compensate for quality shifts in vintages, mainly due to climate variations.1 Thus, by selecting young wines each year for their ability to positively influence aging, the construction of old collection vintages allows an assembly to guarantee a quality and a constant “Champagne house organoleptic style”. These same practices are met for the brandies of Cognac and Armagnac to always offer consumers a constant product each year that is not a function of a single vintage. The aim is to ensure a sustainable progress in the quality of Champagne wines. According to their technical specifications, commonly known as “house style”, Champagne producers use these stored wines (reserve base wines) in the blend to enhance complexity. During this aging period these reserve wines are stored in largevolume stainless steel tanks, on the lees but not in contact with oak. The complexity of the final assembly results from a subtle blend of wines from the current vintage (base wines) with the contribution of aged wines (reserve base wines). The aged wines contribute aging bouquet notes to the blend. Wine aging bouquet is one of the most fascinating phenomena in enology. It denotes a large set of aromas, which together form a perceptive equilibrium of all olfactory sensations, analogous to a complex perfume.2 Variations in Champagne wine aromas with age were described by the Conseil interprofessionnel du vin de Champagne (CIVC).3 Many aromatic notes that are not © 2017 American Chemical Society
present in young wines appear with aging. These aging notes are described as “toasty”, “roasted coffee”, and “pastries” as examples. Recently, Picard et al. proposed a broad definition of the bouquet of red Bordeaux wines and highlighted the importance of empyreumatic notes.4 White wines are also subject to this complexity of aroma during aging but, unfortunately, no precise definition of the aging bouquet of Champagne wines has been published, let alone that of reserve wines. However, empyreumatic notes have also been detected in aged white wines. Some of these aromas were elegantly described by Tominaga et al. and attributed to the accumulation of some volatile thiols.5 These authors also highlighted the increase in furfural levels during bottle-aging after the “prise de mousse”. However, the development of aging notes in wines before bottling had not previously been studied and still less the effects of the vintage and grape variety or winemaking and viticultural practices. Some of these parameters have a major influence on the expression of the aging bouquet, in the case of Bordeaux red wine.6 It is really difficult to predict variations in reserve wine aromas during aging, due to the small number of chemical Received: Revised: Accepted: Published: 2345
October 13, 2016 January 13, 2017 January 21, 2017 January 21, 2017 DOI: 10.1021/acs.jafc.6b04576 J. Agric. Food Chem. 2017, 65, 2345−2356
Article
Journal of Agricultural and Food Chemistry
(Lys), o-phthaldialdehyde, iodoacetic acid, 2-sulfanylethanol, sodium tetraphenylborate (NaB(C6H5)4), dimethyl sulfoxide, propionic acid, glutathione, naphthalene dialdehyde (NDA), and β-cyclodextrin were purchased from Sigma-Aldrich (Saint-Quentin-Fallavier, France). 2-Methylpyrazine was purchased from Alfa Aesar (Johnson Matthey Co., Bischheim, France), and 2,3,5-trimethylpyrazine, 2,5-dimethylthiophene, and 2-ethylpyrazine were purchased from Acros Organics (Geel, Belgium). The internal standard (2-methylpyrazine-d6) was supplied by CDN Isotopes (Quebec, Canada). All solvents were of HPLC grade. Absolute ethanol and methanol (purity > 99%) were obtained from Merck (Darmstadt, Germany). Milli-Q water was obtained from a Milli-Q Plus water system (Millipore, Saint-Quentin-en-Yvelines, France). Sodium chloride (99%), boric acid, and hydrochloric acid were supplied by VWR-Prolabo (Fontenay-sous-bois, France). Wine Samples. The 55 reserve wines analyzed were provided by the Veuve Clicquot Ponsardin Champagne house. Forty-seven wines from the “cuvée” were monovarietal Chardonnay, Pinot noir, or Meunier, from nine subregions in the Champagne vineyards. Three wines were “tailles”, made from the 5 hL second pressing (after the 20.5 hL first pressing from 4000 kg of grapes), and consisted of a blend of all three grape varieties. The wines were from several vintages, ranging from 2014 to 1988. In the Champagne region, pressing standards defined by Protected Designation of Origin (PDO) regulations impose a maximum yield of 2652 L per 4000 kg of beerys (including 2−4% of lees) leading to 2550 L after racking. Since 1993, the grape juice released during a pressing cycle is separated into three different qualities called “cuvée” (2050 L), “première taille” (300 L), and “seconde taille” (200 L), respectively, of higher, intermediate, and lower quality.20 This fractioning, which is an important point in the production of Champagne wines and a major challenge for the industry, is related to the composition of the grape berry. Most of the grape berry compounds are unevenly distributed in the different tissues. For example, at the end of the pressing, the juice (“tailles”) released by the cells close to the skin contains higher levels of flavan-3-ols than the juice obtained at the beginning of the pressing released by the cells of the intermediate zone. These phenolic compounds lead to wine with increased sensitivity to oxidative browning.21 Moreover, extraction of anthocyanins found in the skins of Pinot noir and Pinot Meunier cultivars can result in tainting of the “white” wine that is not appreciated by the consumers, even for short skin contact when exclusively whole bunches are pressed. Besides the problem of increasing oxidation as the pressing progresses, the total acidity and the protein contents strongly decreased because of less cell maturity.20,22 These observations, largely documented in the literature for three centuries, explain why winemakers carefully separate grape juices all along the pressing in the Champagne region into “cuvée” and “taille” qualities. Most of the studied wines were “cuvée”; three of them were “taille”. The grapes were hand-picked and pressed at pressing centers without destemming, supervised by one of the Veuve Clicquot Ponsardin enologists. Some winemaking parameters had been modified over the years, but the overall method was the same. The must was cold-settled, using pectolytic enzymes. The alcoholic and malolactic fermentations were performed by sequential inoculation with commercial Champagne yeast and lactic acid bacteria. The wines were racked shortly after the end of malolactic fermentation to eliminate the coarse lees. The reserve wines were then stored in thermoregulated stainless steel tanks at 18 °C (the same temperature for all wines) once their quality potential had been assessed by the technical team. The distribution of wines by vintage and variety is presented in Table 1. Samples were chosen for analysis as representative examples of the collection of aged wines. All wines, stored in stainless steel tanks, were sampled over a twoday period and kept in Champagne bottles with airtight caps. All analyses were performed within 3 months after sampling to minimize the impact of bottle aging, which would be different from tank aging on lees.
markers available and the lack of knowledge concerning links with winemaking practices. Traditionally, at Veuve Clicquot, all of the reserve wines are tasted every 6 months from the vintage until the decision is made to use them in a blend, to avoid keeping oxidized wines and select those that are ready for use. One of our objectives was to identify some odorous compounds likely to contribute to the bouquet of reserve wines and, consequently, to open the hypothesis of the contribution of this compound to the aroma of Champagne wines. In the early 2000s, it was established that aromatic heterocyclic compounds, containing sulfur, nitrogen, and oxygen, were formed under conditions similar to those of wine aging.7 The aromatic descriptors attributed to aromatic heterocycles are comparable to those perceived in aged reserve wines. In a reducing environment with low pH and temperature, heterocycles are synthesized via a Maillard-like reaction,8,9 involving the condensation of an amino acid, such as cysteine, and α-dicarbonyl compounds, both resulting from fermentation.10−13 The composition of the tripeptide known as glutathione can also be a precursor of interest for the Maillard reaction in wine, in particular its oxygen consumption rate14,15 and cysteine constitution.16 Some common compounds formed during Maillard reactions, such as oxygenated 5-membered rings (furan compounds), have been extensively studied in the context of wine aging.5 Long aging in contact with lees is known to affect the development of volatile compounds17 and redox potential,18 as well as to release proteins and free amino acids.19 The long aging of reserve wines on lees suggests that S,N-aromatic heterocycles and furan compounds may contribute to the aging bouquet of reserve wines. Analysis of the composition of the extensive collection of aged reserve wines at Veuve Clicquot Ponsardin enabled us to explore this hypothesis. This collection of Champagne reserve wines is unique in the world to reference wines stored on fine lees from 2 to 27 years, without contact with oak wood and preserved in large volumes (>90 hL). The conditions for raising the wines were not reproduced and controlled in the laboratory; the wines presented the sincere and faithful reflection of the production of Champagne wine for the study of levels of heterocycles and their precursors. Quantitation was performed by SPME-GC/MS of 23 heterocycles on 55 wines (3 grape varieties, vintages 1988−2014, and 9 Champagne-growing areas) to identify any possible links between aging and concentrations of these molecules. Correlations between heterocycle levels and their potential precursors (amino acids) were also investigated. The purpose of this work is to establish the dynamic of some odorous N,S,O-heterocycles during wine aging using the example of real Champagne reserve wines.
■
MATERIALS AND METHODS
Chemicals. Thiazole, 4-methylthiazole, 2-ethylthiazole, 2-acetylthiazole, 2-methylthiazole, 2,4,5-trimethyloxazole, 3-acetyl-2,5-dimethylfuran, 2-acetylfuran, 5-methylfurfural, 3-acetylthiophene, 2-acetylthiophene, 2,3-dimethylthiophene, 2,5-dimethylthiophene, acetylpyrazine, 2,3-diethylpyrazine, 2,6-dimethylpyrazine, 2-ethyl-3-methylpyrazine, 2-acetyl3-methylpyrazine, tetramethylpyrazine, norvaline, L-aspartic acid (Asp), L-glutamic acid (Glu), L-cysteine monochlohydric monohydrate (Cys), L-asparagine monohydrate (Asp-NH2), L-serine (Ser), L-glutamine (Gln), L-glycine (Gly), L-threonine (Thr), L-arginine monochlorhydric (Arg), L-alanine (Ala), ;-thyrosine (Tyr), L-γ-aminobutyric acid (GABA), L-ethanolamine (Etn), L-valine (Val), L-methionine (Met), L-tryptophan (Trp), L-phenylalanine (Phe), L-isoleucine (Ile), L-leucine (Leu), L-ornithine monochlorhydric (ORN), L-lysine monochlorhydric 2346
DOI: 10.1021/acs.jafc.6b04576 J. Agric. Food Chem. 2017, 65, 2345−2356
Article
Journal of Agricultural and Food Chemistry
for the last 15 min. High-pressure liquid chromatography analyses were carried out on a Dionex Ultimate 3000 coupled to an FLD Dionex 3000 fluorescence detector. The column was an Xbridge BEH C18 (150 mm × 2.1 mm × 3.5 μm), maintained at a constant temperature of 40 °C throughout. Glutathione Quantitation. Reduced glutathione was quantified using the HPLC method validated by Marchand et al.;25 it uses a precolumn derivatization with 2,3-naphthalenedialdehyde (NDA), an isocratic separation in the presence of β-cyclodextrin, and a fluorometric detection. Data Analysis. Statistical data were analyzed using analysis of variance (ANOVA): the homogeneity of variance was tested using Levene’s test, and the normality of residuals was tested using the Shapiro−Wilk test. If these conditions were not met, statistical data were analyzed using the Friedman statistical nonparametric test. The statistically significant level was 5% (p < 0.05). All statistical analyses were conducted using XLSTAT software (Addinsoft, Paris, France). Spearman’s correlation test was used to evaluate the relationships among several variables. The statistically significant level was 5% (XLSTAT software, p value 0.6). Heterocycles in cluster 2 (red color in Figure 1), including 2,3-diethylpyrazine, 2,3-dimethylthiophene, and 2,5-dimethylthiophene, projected on the opposite side, indicating a significant negative correlation (Spearman’s correlation coefficient < −0.6). Finally, cluster 3 (black color in Figure 1), including 2-acetylthiazole and acetylpyrazine, did not show any significant Spearman’s correlation. From the PCA presentation, no link was visible between the heterocycles’ chemical structure (nature of the ring or side functions) and their belonging to one or the other cluster. Variations in Heterocycle Concentrations during Wine Aging. Variations in heterocycle concentrations during wine aging were investigated in clusters 1 and 2. Figure 2 highlights the concentrations of selected heterocycles from clusters 1 (5-methylfurfural, 2-acetylfuran, and 2-methylpyrazine) and 2 (2,3-dimethylthiophene) in wines produced from 2347
DOI: 10.1021/acs.jafc.6b04576 J. Agric. Food Chem. 2017, 65, 2345−2356
2348
1.0 ± 0.5