Research Note Key Compounds of Provence Rosé Wine Flavor

Research Note Key Compounds of Provence Rosé Wine Flavor Gilles Masson 1 and Rémi Schneider 2 * Abstract: The aromas of French Provence rosé wines were subject to in-depth gas chromatography and sensory analyses at the Research and Experimentation Center on Rosé Wine (Centre de Recherche et d Expérimentation sur le Vin Rosé). The study has demonstrated the olfactive contribution of different volatile compounds of fermentative and varietal origins. Dearomatization and reconstitution of flavor has revealed the preponderant character of ethyl esters and higher alcohol acetates, which are simple to determine and are at the origin of the fruity and amylic notes of rosé wines. Moreover, the search for specific varietal compounds well known in other types of wines has made it possible to identify some key compounds of rosé wine flavor: two volatile thiols (3-mercaptohexanol and 3-mercaptohexyl acetate), two furanic compounds (furaneol and homofuraneol), and one C13 norisoprenoid (β-damascenone). These compounds have now been measured in most of the experiments conducted at the Center on Rosé Wine and are used as qualitative indicators in evaluation protocols for the exploitation of vines or for rosé winemaking techniques. Key words: rosé wines, ethyl esters, acetates, furaneol, 3-mercaptohexanol, β-damascenone Rosé wines are characterized by strong fruit expression. White and red wines often exhibit fruity notes as well, and many rosé wines are enriched with f loral or spicy flavors, but the fruity signature seems to be a common identifier of rosé wines throughout the world. Recent studies have demonstrated the contribution of certain volatile compounds. In addition to the well-known and easily measured fermentative compounds such as ethyl esters and the higher alcohols and their acetates (Cacho 2004), other compounds contribute to the flavor of rosé wines, including the 3-mercaptohexanol and its acetate in Bordeaux rosé wines (Murat et al. 2001b, Dubourdieu and Murat 2004, Murat 2005) and Grenache grapes (Ferreira et al. 2002, Cacho 2004). β-damascenone and furaneol, which are present in many wines and provide interesting olfactory nuances, have also been identified in rosé wines. Thus, among the compounds that contribute to the fruity aroma of wines, a certain number are so-called varietal compounds, the presence and content of which may depend on the variety used for winemaking. In addition, in the case of thiol compounds, conditions of fermentation (e.g., yeast strain, temperature, and nitrogen content of must) determine the yield of these compounds from their precursors (Murat et al. 2001a, Masneuf-Pomarède et al. 2006, Subileau et al. 2008). Key compounds of rosé wines have been determined in particular contexts, such as Spain and the Bordeaux region (Ferreira et al. 2002, 1 Centre de Recherche et d Expérimentation sur le Vin Rosé, 70 avenue Wilson, 83550 Vidauban, France, and 2 Nyséos, Bât. 28, 2 place Viala, 34060 Montpellier cedex, France. *Corresponding author (email: remi.schneider@supagro.inra.fr) Manuscript submitted Jul 2008, revised Sept, Nov 2008, accepted Nov 2008. Publication costs of this article defrayed in part by page fees. Copyright 2009 by the American Society for Enology and Viticulture. All rights reserved. 116 Murat et al. 2001b), where the varietals and winemaking processes may be different from those used in the Provence region of France. While previous research represents accurate preliminary studies, it seemed important to identify the key compounds of the fruity aroma, as Provence rosé wines are a blend of several varieties (not only Grenache) with a particular winemaking process, which leads to their particular color and taste. Such identification may help in determining winemaking practices that favor the presence of these compounds. Materials and Methods Wines. Ten Côtes de Provence AOC rosé wines were sampled in 2004, 2005, and 2006 (total of 30 wines); these wines corresponded to the 2003, 2004, and 2005 vintages, respectively. The wines were stored at 12 C for a maximum of 6 months before analysis. All wines had received medals at the Saint Tropez wine competition. The annual competition occurs in May and wines are tasted by a professional jury (producers, enologists, specialized journalists). Flavor reconstitution. Sensory analysis. Tests were performed at the Center on Rosé Wine in a room dedicated to sensory analysis. The jury was comprised of 10 panelists, trained for five years on sensory analysis of rosé wines. Black INAO glasses were used and wines (~10 ml) were served at 12 C in a randomized order. The Quantitative Descriptive Analysis (QDA) profile (Stone and Sidel 1998) was used for descriptive analysis. Panelists on the jury used a standardized sheet that contained the descriptors to be considered: rose, grapefruit, orange, lemon, raspberry, strawberry, cherry, bramble, watermelon, mushroom, caramel, licorice, and banana. Each panelist was asked to quantify the intensity of each descriptor, using an integer score from 0 to 7, in order to generate a sensory profile.

Key Compounds of Provence Rosé Wine Flavor 117 Wine dearomatization. A sample of 250 ml from one of the 10 wines selected in 2004 (considered as especially aromatic) was submitted to two successive solvent extractions (2 x 55 ml) under magnetic agitation. Several solvents were used: dichloromethane (CH 2 ), ethyl acetate, and solutions of pentane dichloromethane (2:1 and 3:5). Following the separation of the organic phase by decantation, the residues of solvents, present in the treated wine, were eliminated by vacuum evaporation at 40 C for 15 min. Dearomatization efficiency was assessed by analysis of the main residual volatile compounds of fermentative origin present in the dearomatized wine versus the initial wine, using a liquid/liquid extraction with ether/ hexane mix (50/50) followed by quantitative analysis by gas chromatography-flame ionization detection (GC-FID) with ethyl heptanoate as internal standard. Estimation of sensory impact of readdition of individual volatile compounds. Fifteen volatile compounds (Table 1) were added individually to the dearomatized wine, at concentrations close to those in the initial wine. The 15 samples thus obtained were evaluated in triangle tests by the jury versus the dearomatized wine. For the wines considered as different from the initial wine, the nature of the aromatic change was qualified by the jury, using the standardized sensory sheet. Flavor reconstitution. In a second step, the set of 15 compounds was added to the dearomatized wine, at concentrations close to those initially determined in the initial wine. The levels of compounds added were based on the difference in concentration between the initial wine and the dearomatized wine. The three wines (initial, dearomatized, reconstituted) were then submitted to sensory analysis by the jury according to the QDA profile, using the same standardized tasting sheet as previously described. Instrumental analysis techniques. Main volatile compounds of fermentative origin. Quantification was performed at the Center on Rosé using a method derived from Murat (2001). The wines (50 ml) were submitted to a liquid/liquid extraction using an ether/hexane mix (50/50) and the extract, concentrated to 5 ml, was analyzed by GC-FID. Ethyl heptanoate was used as internal standard for semiquantification. Varietal thiols, β-damascenone, and β-ionone. Analyses were carried out by the SARCO laboratory (Floirac, France). The analysis of thiols was performed using a published method (Tominaga et al. 1998), which consisted basically in extracting the volatile thiols by liquid-liquid separation with methylene chloride and a further purification by a reversible combination of thiols with p-hydoxybenzoate. 3-Mercapto-1-methoxybutane was used as internal standard. Instrumental analysis was then performed by GC-MS in SIM mode. β-damascenone and β-ionone were analyzed according to the protocol described by Murat (2001). The two norisoprenoidic compounds were extracted from 250 ml of wine, previously adjusted to ph 7, using methylene chloride/pentane (1/1 v/v). 2-Octanone was used as internal standard. After concentration of the extracts to 0.5 ml, quantification was performed by GC-MS in SIM mode. Furaneol and homofuraneol. Analysis of 2,5-dimethyl-4-hydroxy-3(2H)-furanone (furaneol) (HDMF) and 2-ethyl-4-hydroxy-5-methyl-3(2H)-furanone (homofuraneol) (HEMF) was performed at the Nyseos laboratory using a published method (Schneider et al. 2002). 250 ml of wine sample spiked with 14.5 µg [ 2 H 6 ]-HDMF and 13.2 µg [ 2 H 5 ]-HEMF, synthesized by the INRA laboratory, Montpellier, France (Schneider et al. 2002) was extracted with 3 x 60 ml of CH 2 (magnetic stirring at 500 rpm) at O C under nitrogen atmosphere. The organic Added compound Table 1 Results of triangle tests carried out by addition of each compound to the dearomatized wine. Concn (mg/l) Correct answers Results at 5% threshold a cis-3-hexen-1-ol 0.02 4/10 NS Main olfactive notes added Amplified olfactive notes 2-, 3-Methylbutanol 24.1, 124.2 7/10 S Banana, rose, cherry Strawberry, raspberry, bramble, melon Ethyl acetate 3.10 2/10 NS Ethyl butanoate 0.86 6/10 NS Ethyl decanoate 1.39 4/10 NS 2-Phenylethanol 24.5 4/10 NS Ethyl octanoate 5.13 9/10 S Citrus fruits Lemon, orange, grapefruit Isoamyl acetate 0.83 10/10 S Rose 2-Phenylethyl acetate 0.44 5/10 NS Ethyl hexanoate 0.99 8/10 S Raspberry, strawberry Melon, caramel, licorice Diethyl succinate 0.74 5/10 NS Isobutanol 20.4 6/10 NS Hexyl acetate 0.09 8/10 S Ethyl dodecanoate 0.02 6/10 NS a NS: nonsignificant; S: significant (p < 0.05).

118 Masson and Schneider layers were combined, dried over Na 2 SO 4, and concentrated at 40 C to 4 ml. Gas chromatography ion trap mass spectrometry/mass spectrometry (GC-IT-MS/MS) analysis was carried out using a Varian 3800 gas chromatograph fitted with a 30 m DBWAX column (0.25 mm i.d., 0.5 µm film thickness) coupled to an IT-MS/MS detector (Saturn 2000; Varian, Palo Alto, CA). Chromatographic conditions were described previously (Kotseridis 1999). Detection of HDMF, HEMF, and their deuterated analogs was performed by chemical ionization (isobutane, 0.35 bar) and MS/MS with the following parameters (Table 2). Calibration curves were determined for HDMF and HEMF, using dichloromethane solutions (4 ml) of 14.5 µg d 6 -HDMF and 13.2 µg d 5 -HEMF containing from 1.2 to 51.4 µg HDMF and from 4.6 to 54.3 µg HEMF. For each compound, the peak area natural/deuterated ratios were plotted against the respective concentration ratios. Compound Table 2 MS/MS parameters for the quantification of furaneol (HDMF) and homofuraneol (HEMF) in wine extracts. Parent ions Refragmentation (m/z) Mode a excitation Quantifiers (m/z) Qualifiers (m/z) HDMF 129 NR 35 V 111 83, 101 d 6 -HDMF 135 NR 35 V 63 88, 105, 116 HEMF 143 NR 35 V 97 125 d 5 -HEMF 148 NR 35 V 101+102+103 129 a NR: nonresonant. Results and Discussion Flavor reconstitution. Dearomatization efficiency. None of the solvent solutions used allowed the complete elimination of the volatile compounds that were assayed, likely because the target compounds exhibited different polarity. The most polar, such as methanol or 2,3-butanediol, were not extracted efficiently or substantially as the solvent used was apolar, but those compounds are widely recognized as noncrucial compounds in wines (Etiévant 1991). For ethyl esters and acetates, which are markers of the fruity aroma of wine (Etiévant 1991, Cacho 2004, Guth 1997a, 1997b), the most efficient extraction was obtained globally using a pentane/dichloromethane (3:5) solution (Table 3). Among the compounds most strongly eliminated by solvent extraction, 15 identified generally for their sensory impact were selected as potential markers for the rest of the study: ethyl butanoate, hexanoate, octanoate, decanoate, and dodecanoate; isoamyl; ethyl, hexyl, and phenylethyl acetates; ethyl succinate; cis-3-hexen-1-ol; isobutanol; 2- and 3-methylbutanol; and 2-phenylethanol. Sensory impact of eliminated compounds. To estimate the sensory impact of the compounds eliminated by the dearomatization process, the 15 compounds were added individually to the dearomatized wine, and the wines thus obtained were submitted to triangle tests versus the initial wines (Table 1). When the difference was significant, the jury was asked to describe the difference using the standardized sheet. Only five compounds appeared to have a sensory impact when added individually (Table 1). Those five aromatic markers are ethyl esters or acetates, known to contribute to the fruity aroma of wine. Surprisingly, the odor that allowed the differentiation in the triangle tests was not always consistent with that of the pure compounds, possibly because of the interaction effects with the matrix or with residual volatile compounds. As expected, none of the compounds added individually could reconstitute the aroma of the initial wine (data not shown). Thus a global flavor reconstitution was performed. Reconstitution. The complementation of the dearomatized wine was conducted to estimate the contribution of the 15 volatile compounds tested at concentrations close to those determined in the initial wine. Verification was made by comparing their dosage in the initial wine and in the reconstituted wine (Table 4). The values found in the reconstituted wine are closer to those found in the initial wine except for some acetates (ethyl and isoamyl) and to a lesser extent for ethyl hexanoate. Despite these imperfections, the wine was submitted for sensory analysis to the jury. Results show that the reconstituted wine exhibited a flavor significantly more intense than the dearomatized wine (Figure 1). The supplementation has made it possible to increase significantly most of the olfactory notes beyond those found in the initial wine. The added compounds clearly contributed to the olfactory quality of the wine, and therefore can be classified among the key aroma compounds of this type of wine. However, three notes remained limited despite the supplementation: lemon, orange, and grapefruit. These olfactory notes recall those of the thiol-type compounds (e.g., 3-mercaptohexanol and its acetate), which also have been identified as key compounds of some rosé wines, such as those produced from Merlot, Cabernet Sauvignon, Cabernet franc (Murat et al. 2001b), and Grenache (Ferreira et al. 2002). 3-Mercaptohexanol (3MH) and 3-mercaptohexyl acetate (3MHA) are known for their intense odors of grapefruit and passion fruit and exhibit low perception thresholds of 60 and 4 ng/l, respectively (Tominaga et al. 1998). The addition option was not selected to validate this hypothesis, as it is difficult to manipulate compounds that are easily oxidizable and we were not able to confirm the supplementation level online. We chose to perform the dosage in a certain number of wines to determine if their concentration enabled validating them as key compounds. Varietal thiols. Quantification of the three varietal thiols (4-methyl- 4-mercapto-pentan-2-one [4MMP], 3MH, and 3MHA) was carried out on

Key Compounds of Provence Rosé Wine Flavor 119 Ethyl esters (mg/l) Table 3 Concentrations (mg/l) of the main fermentative compounds in the initial wine and the wines dearomatized by liquid-liquid extraction using various solvents. After extraction with Initial wine CH 2 Ethyl acetate Pentane/CH 2 (2:1) Pentane/CH 2 (3:5) Ethyl butanoate 0.181 0.001 0.046 0.037 0.008 Ethyl hexanoate 0.646 0.016 0.004 0.021 0.003 Ethyl octanoate 0.862 0.004 0.006 0.046 0.001 Ethyl decanoate 0.314 0.001 0.008 0.018 0.006 Ethyl dodecanoate 0.005 0.002 0.002 0.001 0.003 Ethyl succinate 0.503 0.009 0.011 0.010 0.001 Ethyl lactate 4.429 0.296 3.067 6.348 1.906 Acetates (mg/l) Ethyl acetate 27.709 8.796 5.430 1.509 0.440 Isoamyl acetate 1.779 0.001 0.004 0.032 0.003 Hexyl acetate 0.052 0.001 0.010 0.010 0.001 Phenylethyl acetate 0.667 0.018 0.030 0.014 0.014 Alcohols and polyols (mg/l) Methanol 3.875 8.943 1.200 13.392 9.686 Pentanol 15.789 3.523 2.299 8.767 4.364 Isobutanol 32.882 2.714 0.365 1.748 2.046 cis-3-hexanol 0.072 0.006 0.009 0.007 0.008 Butanol 1.307 2.432 0.724 0.217 0.871 2-Methylbutanol 31.279 0.520 0.142 2.631 0.618 3-Methylbutanol 193.575 8.678 1.672 20.309 6.108 Hexanol 0.863 1.234 0.365 0.069 0.356 2,3-Butanediol 70.739 53.505 70.830 45.395 32.317 2-Phenylethanol 48.309 2.483 0.094 14.789 1.979 Carbonyl compounds (mg/l) Acetone 1.791 0.535 0.421 1.315 0.246 Acetaldehyde 2.950 4.266 3.341 3.381 0.398 the 2003 and 2004 samples. As expected, 4MMP, which produces the typical boxtree aroma of Sauvignon wine (Darriet 1995) was not found in any of the samples analyzed. 3MH was found in all samples analyzed (Table 5), and its concentration was higher than the published perception threshold (Tominaga et al. 1998) in 90% of cases, which indicates a significant contribution of this compound to the aroma of Provence rosé wines. 3MHA was found in 14 of the wines, and its content therein was systematically higher than the perception threshold. In terms of vintage, the 2003 wines had thiol concentrations much lower than those found in the 2004 wines. Such a difference could be linked to differences in wine origin. However, this difference was also found between samples 4 and 10 of 2003 and samples 18 and 20 of 2004 (data not shown), which were from the same winery. It is therefore more likely that the lower concentrations in the 2003 wines resulted from a heatwave that occurred in 2003. This observation is confirmed by other findings (Dagan 2006, R. Schneider, unpublished data, 2005), which led to a hypothesis of lower biosynthesis of the thiol precursors in grapes under these growing conditions. However, in our study, we are unable to confirm this point as we lack data on those precursors in grapes that are at the origin of these compounds. Other explanations may involve yeast strain (or feral yeast) or winemaking processes, which could be different for each winery. Results presented in Table 5 are consistent with those reported elsewhere (Murat et al. 2005) and obtained from 30 Bordeaux wines and 10 Provence wines, although the wines did not belong to the same vintage. The authors also showed that 3MHA concentration was higher in Provence rosés than in Bordeaux rosés, a difference that could be due to the specificity of the vine varieties or more likely of the yeast strain used in each region. In addition, among the Provence vines, notable differences have been identified (Masson 2006): Syrah and Grenache rosé wines have higher volatile thiol concentrations than Cinsaut rosés, and 3MHA concentration seems higher in the Syrah wines.

120 Masson and Schneider Table 4 Concentrations (mg/l) of the 15 volatile compounds of fermentative origin in the initial wine and reconstituted wine. Initial wine (mg/l) Reconstitued wine (mg/l) Ethyl butanoate 0.929 0.860 Ethyl hexanoate 1.564 0.988 Ethyl octanoate 4.177 5.129 Ethyl decanoate 1.370 1.390 Ethyl dodecanoate 0.022 0.020 Ethyl succinate 0.705 0.740 Ethyl acetate 31.170 3.100 Isoamyl acetate 2.764 0.833 Hexyl acetate 0.111 0.086 Phenylethyl acetate 0.430 0.440 Isobutanol 23.171 20.417 2-Methylbutanol 19.280 24.060 3-Methylbutanol 136.53 124.19 cis-3-hexanol 0.017 0.016 2-Phenylethanol 20.640 24.490 Figure 1 Comparison of the olfactory profile carried out by the expert jury on the three wines: A (initial), D (dearomatized), S (reconstituted). Despite the marked effects between the vintages, it appears that the 3-mercaptohexanol and, to a lesser extent, its acetate are key compounds of rosé wine aroma, which agrees with findings for Grenache rosé wines (Cacho 2004). It could also explain the poor notes of citrus fruits due to the supplementation of the wine for the flavor reconstitution test. Indeed, the wine reconstituted without the addition of the 3-mercaptohexanol and its acetate exhibits citrus fruit notes significantly lower than the initial wine, whereas the dosage indicates that they are present at concentrations much higher than their perception thresholds. β-damascenone and β-ionone. In order to complete the study on the key compounds of the aroma of Provence rosé wines, we also examined β-damascenone and β-ionone as well as furaneol and homofuraneol. The impact of these compounds on the aroma of rosé wines has been studied previously (Ferreira et al. 2002, Murat 2001). β-damascenone has been found in numerous vines and corresponding wines and is formed by acid hydrolysis of numerous precursors (reviewed by Bayonove 1998). Known for its f lavor of stewed apples, this compound exceeded its perception threshold in the rosé wines that were analyzed (Table 5). Thus, this compound could play a significant role in the bouquet of Provence rosé wines. However, detection thresholds published in the literature vary widely, depending on the matrix from which they have been determined: 45 ng/l in a model hydroalcoholic solution (Kotseridis 1999) and 4.5 µg/l in sweet wines (Etiévant 1991). It is therefore difficult to evaluate the contribution of that compound in our samples. A recent study has established the thresholds in a white wine dearomatized using coal (140 ng/l) and in different red wines dearomatized by evaporation (850 7000 ng/l) (Pineau et al. 2007). There is no data for rosé wines, but considering the nature of rosé wine, we might use the olfactory threshold of white wines. In such a case, β-damascenone may be a key compound of the wine aroma, given the concentrations found (between 1.4 and 4.6 µg/l). Moreover, Pineau et al. (2007) suggested that β-damascenone may act an enhancer of red fruit aromas in red wines, which may also be the case in rosé wines. The other norisoprenoid, β-ionone, is also a powerful aroma compound, reminiscent of violet. Its origin is not well established, but it could be generated by a direct degradation of β-carotene present in grapes during the vinification, as hypothesized elsewhere (Baumes et al. Vintage Table 5 Volatile compound concentrations in Côtes de Provence AOC rosé wines (mean ± std) (na: not analyzed). 3-Mercaptohexanol (ng/l) 3-Mercaptohexyl acetate (ng/l) β-damascenone (µg/l) β-ionone (ng/l) Furaneol (µg/l) Homofuraneol (µg/l) 2003 (10 samples) 211 ± 201.7 40 ± 29.4 na na na na 2004 (10 samples) 675 ± 419.4 89 ± 80.6 3 ± 1.6 14 ± 4.5 na na 2005 (10 samples) na na na na 94 ± 47.3 4 ± 1.9

Key Compounds of Provence Rosé Wine Flavor 121 2002). In the rosé wines analyzed, β-ionone exhibited concentrations too low to provide an olfactory contribution. Its contribution to the aroma of rosé wines may therefore be relatively specific to rosés from Bordeaux vines such as Merlot, Cabernet Sauvignon, and Cabernet franc (Murat 2005) and should not be considered as a key compound in Provence rosé wines. Furaneol and homofuraneol. These compounds are known for their pleasant flavor of caramel and strawberry (furaneol) and caramel and toasted bread (homofuraneol). Furaneol was first identified in wines from Vitis labrusca (Rapp et al. 1980) and seems to characterize hybrid wines (Rapp et al. 1980, Guedes de Pinho and Bertrand 1995). Its presence in the Vitis vinifera wines was identified only recently (Guth 1997a, 1997b, Cutzach et al. 1998a, 1998b), as it is difficult to analyze (Guedes de Pinho and Bertrand 1995). Homofuraneol is also an important volatile compound for Merlot and Cabernet Sauvignon wines (Kotseridis 1999). Apart from when wines are aged in oak barrels, the origin of furaneol and homofuraneol in wines is not known, despite reports of a glycosidic precursor of furaneol (Guedes de Pinho and Bertrand 1995, Kotseridis 1999). The Provence rosé wines studied here all had furaneol concentrations higher than published perception thresholds, whereas homofuraneol concentrations were slightly lower than published perception thresholds (Schneider et al. 2002) (Table 5). These results indicate that furaneol may play a significant role in the aroma of the rosé wines analyzed. These observations confirm the conclusions found by other researchers (Ferreira et al. 2002), who demonstrated that the elimination of these two compounds in a Grenache rosé wine reduced significantly the fruity and caramel notes. Conclusion Some key components of Provence rosé wines have been identified. These compounds are derived from fermentation (ethyl esters and higher alcohol acetates) or are varietal in origin (3-mercaptohexanol, 3-mercaptohexyl acetate, furaneol, and homofuraneol, β-damascenone). Today they are used as qualitative markers in most experimental protocols implemented by the Center on Rosé Wine and concern factors as varied as terroir, juice extraction, fermentation, and rosé wine conservation. Other compounds analyzed (β-ionone, some higher alcohols) are present in rosé wines in low concentrations and are clearly lower than their perception threshold. Research is ongoing to identify the compounds responsible for some aromatic notes often described in these wines, especially in floral, spicy, and yellow fruit ranges. Literature Cited Baumes, R., J. Wirth, S. Bureau, Y. Gunata, and A. Razungles. 2002. Biogeneration of C 13 -norisoprenoid compounds: Experiments supportive for an apo-carotenoid pathway in grapevines. Anal. Chim. Acta 458:3-14. Bayonove, C. 1998. L arôme variétal: le potentiel aromatique du raisin. In Œnologie: Fondements scientifiques et techniques. C. Flanzy (ed.), pp. 165-181. Lavoisier Tec & Doc, Paris. Cacho, J. 2004. Relation entre l analyse physico-chimique et sensorielle des vins rosés. In Proceedings for the Rencontres Internationales du Rosé. G. Masson (ed.), pp. 51-63. Centre des Rosés, Vidauban, France. Cutzach, I., P. Chatonnet, and D. Dubourdieu. 1998a. Etude sur l arôme des vins doux naturels non muscatés. 1. Analyse qualitative de l arôme des Banyuls et Rivesaltes au cours de leur vieillissement. J. Int. Sci. Vigne Vin 32:99-110. Cutzach, I, P. Chatonnet, R. Henry, M. Pons, and D. Dubourdieu. 1998b. Etude sur l arôme des vins doux naturels non muscatés. 2. Dosages de certains composés volatils intervenant dans l arôme des vins doux naturels au cours de leur vieillissement. J. Int. Sci. Vigne Vin 32:211-221. Dagan, L. 2006. Potentiel aromatique des raisins de Vitis vinifera L. cv. Petit Manseng et Gros Manseng. Contribution à l arôme des vins de pays Côtes de Gascogne. PhD thesis, Ecole Nationale Supérieure Agronomique de Montpellier, France. Darriet, P. 1995. Recherche sur l arôme et les précurseurs d arôme du Sauvignon. Ph.D. thesis, University of Bordeaux II, France. Dubourdieu, D. and M.L. Murat. 2004. Acquisitions récentes sur l arôme des vins rosés. In Proceedings for the Rencontres Internationales du Rosé. G. Masson (ed.), pp. 44-50. Centre des Rosés, Vidauban, France. Etiévant, P. 1991. Wine. In Volatile Compounds in Foods and Beverages. H. Maarse (ed.), pp. 483-546. Marcel Dekker, New York. Ferreira, V., N. Ortin, A. Escudero, R. Lopez, and J. Cacho. 2002. Chemical characterization of the aroma of Grenache rosé wines: Aroma extract dilution analysis, quantitative determination, and sensory reconstitution studies. J. Agric. Food Chem. 50:4048-4054. Guedes de Pinho, P., and A. Bertrand. 1995. Analytical determination of furaneol (2,5-dimethyl-4-hydroxy-3(2H)-furanone). Application to differentiation of white wines from hybrid and various Vitis vinifera cultivars. Am. J. Enol. Vitic. 46:181-186. Guth, H. 1997a. Identification of character impact odorants of different white wine varieties. J. Agric. Food Chem. 45:3022-3026. Guth, H. 1997b. Quantitation and sensory studies of character impact odorants of different white wine varieties. J. Agric. Food Chem. 45:3027-3032. Kotseridis, Y. 1999. Etude de l arôme des vins de Merlot et Cabernet Sauvignon de la région bordelaise. PhD thesis, University of Bordeaux II, France. Masneuf-Pomarède, I., C. Mansour, M.L. Murat, T.T. Tominaga, and D. Dubourdieu. 2006. Influence of fermentation temperature on volatile thiols concentrations in Sauvignon blanc wines. Int. J. Food Microbiol. 108:385-390. Masson, G. 2006. Quoi de neuf sur l arôme des vins rosés? Rosé. com 9:3-7. Murat, M.L. 2001. Recherches sur la vinification des vins rosés et clairets de Bordeaux. Diplôme d Etudes et de Recherches, University of Bordeaux II, France. Murat, M.L. 2005. Acquisitions récentes sur l arôme des vins rosés. 1. Caractérisation de l arôme, étude du potentiel aromatique des raisins et des moûts. Rev. Œnol. 117:27-30. Murat, M.L., I. Masneuf, P. Darriet, V. Lavigne, T. Tominaga, and D. Dubourdieu. 2001a. Effect of Saccharomyces cerevisiae yeast

122 Masson and Schneider strains on the liberation of volatile thiols in Sauvignon blanc wine. Am. J. Enol. Vitic. 52:136-139. Murat, M.L., T. Tominaga, and D. Dubourdieu. 2001b. Mise en évidence de composés clefs dans l arôme des vins rosés et clairets de Bordeaux. J. Int. Sci. Vigne Vin. 35:99-105. Pineau, B., J.C. Barbe, C. Van Leeuwen, and D. Dubourdieu. 2007. Which impact for β-damascenone on red wines aroma? J. Agric. Food Chem. 55:4103-4108. Rapp, A, W. Knipser, L. Engel, H. Ullemeyer, and W. Heimann. 1980. Off-flavor compounds in the berry and wine aroma of grape vine hybrids. I. The strawberry-like aroma. Vitis 19:13-23. Schneider, R., Y. Kotseridis, A. Belancic, C. Augier, and A. Razungles. 2002. Quantification of furaneol and homofuraneol in Vitis vinifera wines by stable isotope dilution assay using GC-MS/MS. In 10 th Weurman Flavor Research Symposium. J.M. Quéré and P.X. Etiévant (eds.), pp. 710-714. Lavoisier Tec & Doc, Paris. Stone, H., and J.L Sidel. 1998. Quantitative descriptive analysis: Developments, applications and the future. Food Technol. 52 (8):48-52. Subileau, M., R. Schneider, J.M. Salmon, and E. Degryse. 2008. Nitrogen catabolite repression modulates the production of aromatic thiols characteristic of Sauvignon blanc at the level of precursor transport. FEMS Yeast Res. 8:771-780. Tominaga, T., F. Furrer, R. Henry, and D. Dubourdieu. 1998. Identification of new volatile thiols in the aroma of Vitis vinifera L. var. Sauvignon. Flavor Fragrance J. 13:159-162.