La levure Saccharomyces cerevisiae joue un rôle essentiel dans la production de composés aromatiques, tels que les esters, les alcools supérieurs et les acides organiques, ainsi que dans la transformation de précurseurs d'arômes du raisin pendant la fermentation du vin. Afin d'identifier les bases génomiques et métaboliques de ces propriétés, un croisement a été réalisé entre deux souches de levures de vin, sélectionnées pour leurs besoins en azote différents lors de la fermentation. 130 ségrégants de génération F2 ont été génotypés par séquençage complet du génome et individuellement phénotypés pendant la fermentation en mesurant les métabolites extracellulaires par HPLC et GC-MS. Les flux métaboliques intracellulaires ont été estimés à l’aide d’un modèle à base de contraintes. Une analyse QTL (quantitative trait locus) a été utilisée pour identifier les allèles influençant les variations d'arômes et de flux métaboliques. Plus de 80 QTL expliquant la variation de 59 caractères quantitatifs ont été détectés. Ces caractères comprennent des paramètres fermentaires, de consommation de substrat, la production de principaux métabolites et d’arômes fermentaires, ainsi que le métabolisme de composés aromatiques du raisin. L’intérêt de la cartographie QTL pour identifier les déterminants génétiques de variations de flux intracellulaires (f-QTLs) a par ailleurs été démontrée. Les QTL détectés ont été disséqués et des gènes dont les allèles contribuent spécifiquement aux variations phénotypiques ont été identifiés. Ces résultats soulignent la complexité génomique et métabolique de la synthèse et de la transformation d'arômes par la levure. L'identification de ces déterminants génétiques permet de mieux comprendre les liens entre variation génétique des levures et traits technologiques et fournit une base précieuse pour le développement de souches optimisées par des stratégies génétiques de croisement assisté par marqueurs. / The yeast Saccharomyces cerevisiae plays a vital role in the production of aroma compounds, such as esters, higher alcohols and organic acids, and the conversion of grape-derived aroma precursors during wine fermentation. To identify the genomic and metabolic bases for these processes, a cross was performed between two wine yeast strains selected because of their different nitrogen requirement during fermentation. 130 F2-segregants were genotyped by whole genome sequencing and individually phenotyped during wine fermentation by measuring extracellular metabolites using HPLC and GC-MS. Intracellular metabolic fluxes were estimated by constraint-based modeling. Quantitative trait locus (QTL) mapping was used to identify allelic variants influencing variations in the aroma profile and metabolic fluxes. More than 80 QTLs explaining variation in 59 quantitative traits were detected. These traits consisted of general fermentation parameters, substrate consumption, the production of main metabolites and fermentative aromas and the metabolism of grape aroma compounds. The applicability of QTL mapping to detect regions influencing intracellular fluxes (f-QTLs) was furthermore demonstrated. Found QTLs were dissected and genes with allele specific contributions to the phenotype were identified. These results emphasize the genomic and metabolic complexity of yeast aroma formation. In addition, the identification of genetic determinants increases knowledge about the links between genetic variation and industrial traits and provides a valuable foundation for the development of optimized strains by marker-assisted selection breeding strategies.

Thesis (MScAgric)--Stellenbosch University, 2013. / ENGLISH ABSTRACT: Grape must is a complex medium, and during wine production numerous biochemical pathways and metabolic reactions are taking place simultaneously to produce a specific taste and aroma. Microorganisms, specifically yeast, play a key role in the formation of metabolites formed during alcoholic fermentation. Sauvignon blanc, a well studied grape cultivar, is known to have a versatile range of aroma profiles ranging from “green” to “tropical”. It has been broadly stated that a “green” Sauvignon blanc can be created in the vineyard and a “tropical” Sauvignon blanc can be created by selecting a specific yeast strain, and that the balance between “green” and “tropical” characters is essential for the final aroma profile. Except for grape-derived varietal aromatic compounds such as methoxypyrazines (green), volatile thiols (tropical) and monoterpenes (floral), yeast derived volatile compounds including esters, higher alcohols, fatty acids and carbonyl compounds will also contribute to the final wine aroma. The main aim of this study was to assess how viticultural treatment-derived differences in grape must, can impact on aroma production when this grape must is fermented with different commercial wine yeast strains. The viticulture treatment focused on light intensity modulated through canopy treatment. Volatile aroma differences were compared for canopy and yeast treatments, specifically focusing on the fermentation derived bouquet (esters, higher alcohols, volatile fatty acids, carbonyl compounds and monoterpenes). Results showed significant differences between initial must compositions, including titratable acidity, malic acid and yeast assimilable nitrogen. The volatile aroma compounds were also significantly impacted although no noticeable effect on the overall fermentation kinetics was observed. Depending on the yeast strain differences in volatile compounds varied. A clear vintage effect is noticeable between volatile compounds affected by the treatments. Data generated in 2012 shows clear differences between ethyl- and acetate esters and could clearly be grouped according to yeast strain through multivariate analysis. Sensory evaluation results could clearly be distinguished according to canopy treatment and to a lesser degree according to yeast strain used. This indicates that although yeast has a more prominent impact on the fermentative bouquet that develops during alcoholic fermentation the overriding aroma is primarily derived from grape-derived metabolites which can be manipulated by canopy treatments. None the less the difference in fermentation bouquet does add to the complexity of the wine especially in the case of fermentation derived “tropical” aromas including guava and passion fruit. In some cases where shaded grapes had higher ester concentrations, the resultant wine also had higher aroma quality. This study has contributed to a better understanding of the complex relationships between canopy manipulation and yeast selection on aroma formation. The analysis of volatile aroma alone however is not enough to understand the final perception of wine taste and further indepth studies of the viticultural and oenological factors is needed. In particular, this project has focused on a single vineyard over only two vintages. The general validity of the conclusions derived from this study therefore will require additional data sets. / AFRIKAANSE OPSOMMING: Druiwemos is ‘n komplekse medium en tydens wynbereiding is daar verskeie biochemiese weë en metaboliese reaksies wat gelyktydig plaasvind om ‘n spesifieke smaak en aroma te produseer. Mikro-organismes, veral gis, speel ‘n sleutelrol in die vorming van metaboliete tydens alkoholiese gisting. Sauvignon blanc, ‘n goed bestudeerde druifkultivar, besit ‘n veelsydige reeks aromaprofiele wat wissel van “groen” tot “tropies”. Oor die algemeen word dit voorgehou dat ‘n “groen” Sauvignon blanc in die wingerd geskep word, terwyl ‘n “tropiese” Sauvignon blanc geskep kan word deur ‘n spesifieke gisras te selekteer, en die balans tussen “groen” en “tropiese” karakters is noodsaaklik vir die finale aromaprofiel. Behalwe vir druifafgeleide kultivarafhanklike aromatiese verbindings soos metoksipirasiene (groen), vlugtige tiole (tropies) en monoterpene (blomagtig), sal gisafgeleide vlugtige komponente, waaronder esters, hoër alkohole, vetsure en karbonielverbindings, ook tot die finale wynaroma bydra. Die hoofdoelwit van hierdie studie was om te bepaal hoe verskille in druiwemos wat afkomstig is van wynkundige behandeling ‘n impak op aromaproduksie kan hê wanneer hierdie druiwemos met verskillende kommersiële wyngisrasse gegis word. Die wynkundige behandeling het gefokus op ligintensiteit wat deur lowerbehandeling gereguleer is. Vlugtige aromaverskille is op grond van lower- en gisbehandelings vergelyk, met ‘n spesifieke fokus op die gistingsafgeleide boeket (esters, hoër alkohole, vlugtige vetsure, karbonielverbindings en monoterpene). Die resultate het beduidende verskille getoon tussen aanvanklike mossamestellings, waaronder titreerbare suurheid, appelsuur en gis-assimileerbare stikstof. Daar was ook ‘n noemenswaardige impak op die vlugtige aromaverbindings, hoewel geen merkbare effek op die algehele gistingskinetika waargeneem kon word nie. Die verskille in vlugtige verbindings het gewissel op grond van die gisras. ‘n Duidelike oesjaareffek was merkbaar tussen vlugtige verbindings wat deur die behandelings geaffekteer is. Data wat in 2012 gegenereer is, toon duidelike verskille tussen etiel- en asetaatesters en kon duidelik m.b.v. meervariantanalise volgens gisras gegroepeer word. Die resultate van die sensoriese evaluering kon duidelik volgens lowerbehandeling onderskei word, en tot ‘n mindere mate volgens die gisras wat gebruik is. Dít dui daarop dat hoewel gis ‘n meer prominente impak het op die gistingsboeket wat tydens alkoholiese gisting ontwikkel, is die oorheersende aroma hoofsaaklik afgelei van druifafgeleide metaboliete wat deur lowerbehandelings gemanipuleer kan word. Nietemin dra die verskil in gistingsboeket by tot die kompleksiteit van die wyn, veral in die geval van gistingsafgeleide “tropiese” aromas, insluitend koejawel en grenadella. In sommige gevalle waar beskadude druiwe hoër esterkonsentrasies gehad het, het die gevolglike wyn ook ‘n hoër aromakwaliteit gehad. Hierdie studie dra by tot ‘n beter begrip van die effek van die komplekse verhoudings tussen lowermanipulasie en gisseleksie op aromavorming. ‘n Analise van vlugtige aroma alleen is egter nie voldoende om die finale persepsie van wynsmaak te begryp nie en bykomende diepgaande studies van die wingerdkundige en wynkundige faktore word benodig. Hierdie projek het in die besonder gefokus op ‘n enkele wingerd oor slegs twee oesjare. Die algemene geldigheid van die afleidings wat van hierdie studie gemaak word, sal dus bykomende datastelle vereis. / The National Research Foundation and Postgraduate Merit Bursary for financial support