Merging fast sensory profiling with non-targeted GC-MS analysis for multifactorial experimental wine making

The aromatic complexity of a wine results from the perception of the association of volatile molecules and each aroma can be categorized into different families. The “green” aromas family in red wines has retained our attention by its close link with the fruity perception. In that study, the “green” olfactory concept of red wines was considered through a strategy combining both sensory analysis and hyphenated chromatographic techniques including HPLC and MDGC (Multidimensional Gas Chromatography). The aromatic space of this concept was specified by lexical generation through a free association task on 22 selected wines by a panel of wine experts. Then, 70 French red wines were scored on the basis of the intensity of their “green” and “fruity” attributes.

INTRODUCTION: Foam stability of sparkling wines is significantly favored by the presence of surface active agents such as proteins and polysaccharides [1]. For that reason, the renowned sparkling wines are aged after the second fermentation in contact with the lees for several months (even years). Thereby wines are enriched in these macromolecules due to yeast autolysis. Since this practice is slow and costly, winemakers are seeking for alternative procedures to increase their concentration in base wines. In that sense, the supplementation with inactive yeast during alcoholic fermentation has been proposed [2]. The aim of this study was to determine whether this new strategy is really useful for enriching base wines in macromolecules and for improving foam properties of the base wines.

Proteins play a significant role in the colloidal stability and clarity of white wines [1]. However, under conditions of high temperatures during storage or transportation, the proteins themselves can self-aggregate into light-dispersing particles causing the so-called protein haze [2]. Formation of these unattractive precipitates in bottled wine is a common defect of commercial wines, making them unacceptable for sale [3]. Previous studies identified the presence of phenolic compounds in the natural precipitate of white wine [4], contributing to the hypothesis that these compounds could be involved in the mechanism of protein haze formation.

The composition of wine is largely determined by the composition of pre-fermentation juice, which is influenced by extraction of grape components. Different grape harvesting and processing conditions could affect the extraction of grape components into juice. Among these grape components, pathogenesis-related (PR) proteins are of great concern for white wine maker as they are the main cause of haze formation in finished white wine. If not removed before bottling, these PR proteins may progress into haze through the formation of complex with phenolics under certain conditions. Thaumatin-like proteins (TLPs) and chitinases are the main constituents of PR proteins found in protein haze.

Recent studies have demonstrated the role of certain lactones, particularly 2-nonen-4-olide, and volatile thiols (3-sulfanylhexan-1-ol) in the over ripped aromas of noble rot sweet wines (Stamatopoulos et al. 2014ab). These compounds are partly formed during the maturation and under the activity of B. cinerea on grapes. This research was carried out in the vineyard of Sauternes with aim to better understand their genesis depending on the grape over-ripening on two different soil types during 3 vintages. Thus, the study was conducted, with the Sémillon grape, during vintages 2012, 2014 & 2015, at 4 stages of over-maturation of the grapes (healthy, pourri plein, pourri roti, pourri roti + 15 days) considering two vineyard plots with different soil characteristics (calcosol & peyrosol) planted with the 315 Sémillon clone and grafted on 101-14 rootstock respectively in 1981 and 1980 and cultivated with the same vineyard management. Volatile lactones were assayed by liquid-liquid extraction followed by GC/MS analysis and the precursors of 3-sulfanylhexanol by an adaptation of the method by Capone et al. 2010 (SPE-
UPLC/FTMS).