<p style="text-align: justify;"><em>Oenococcus oeni</em> is the bacterium most often associated with spontaneous malolactic fermentation (MLF) of wine. During MLF, malic acid is transformed into lactic acid and several metabolites are modified, modulating wine’s total acidity and improving its sensory properties. Previous works have suggested that certain genetic groups of <em>O. oeni</em> strains are associated to different kinds of products. In the present study we have spotted two groups of strains isolated mainly from Burgundy wines, one associated to red wines and the other to white wines. Sequencing 14 genomes of red and white wine strains revealed that they share a common ancestor that probably colonised two different substrates –red and white wine-associated environments–, diverging over time and disseminating to various regions. Their capacity to perform MLF and modify the volatile profile of wine was determined by fermenting a chardonnay wine and analysing its volatile fraction with a non-targeted metabolomics approach by GC-MS. The strains had a different impact on the volatile composition depending on their group of origin. These results show for the first time a correspondence between the product of origin of the strains and the volatile profile of the wines they produce. Furthermore, the genetic features that might be implied in these different phenotypes are examined.</p>

Oenococcus oeni, the lactic acid bacterium primarily responsible for malolactic fermentation in wine, is able to grow on a large variety of carbohydrates, but the pathways by which substrates are transported and phosphorylated in this species have been poorly studied. We show that the genes encoding the general phosphotransferase proteins, enzyme I (EI) and histidine protein (HPr), as well as 21 permease genes (3 isolated ones and 18 clustered into 6 distinct loci), are highly conserved among the strains studied and may form part of the O. oeni core genome. Additional permease genes differentiate the strains and may have been acquired or lost by horizontal gene transfer events. The core pts genes are expressed, and permease gene expression is modulated by the nature of the bacterial growth substrate. Decryptified O. oeni cells are able to phosphorylate glucose, cellobiose, trehalose, and mannose at the expense of phosphoenolpyruvate. These substrates are present at low concentrations in wine at the end of alcoholic fermentation. The phosphotransferase system (PTS) may contribute to the perfect adaptation of O. oeni to its singular ecological niche.

Brettanomyces bruxellensis is the main spoilage microbial agent in red wines. The use of fungal chitosan has been authorized since 2009 as a curative treatment to eliminate this yeast in conventional wines and in 2018 in organic wines. As this species is known to exhibit great genetic and phenotypic diversity, we examined whether all the strains responded the same way to chitosan treatment. A collection of 53 strains of B. bruxellensis was used. In the conditions of the reference test, all were at least temporarily affected by the addition of chitosan to wine, with significant decrease of cultivable population. Some (41%) were very sensitive and no cultivable yeast was detected in wine or lees after 3 days of treatment, while others (13%) were tolerant and, after a slight drop in cultivability, resumed growth between 3 and 10 days and remained able to produce spoilage compounds. There were also many strains with intermediate behavior. The strain behavior was only partially linked to the strain genetic group. This behavior was little modulated by the physiological state of the strain or the dose of chitosan used (within the limits of the authorized doses). On the other hand, for a given strain, the sensitivity to chitosan treatment was modulated by the chitosan used and by the properties of the wine in which the treatment was carried out.

Although it is well known that O. oeni is well adapted to wine, this study shows that strains of some genetic lineages within this species have evolved to adapt better than others to specific types of wines.

Nowadays, the use of sulfur dioxide (SO2) during the winemaking process is a controversial societal issue. In order to reduce its use, various alternatives are emerging, in particular bioprotection by adding yeasts, with different impacts on yeast microbiota in early winemaking stages. In this study, quantitative-PCR and metabarcoding high-throughput sequencing (HTS) were combined with MALDI-TOF-MS to monitor yeast population dynamic and diversity in the early stages of red winemaking process without sulfites and with bioprotection by Torulaspora delbrueckii and Metschnikowia pulcherrima addition. By using standard procedures for yeast protein extraction and a laboratory-specific database of wine yeasts, identification at species level of 95% of the isolates was successfully achieved by MALDI-TOF-MS, thus confirming that it is a promising method for wine yeast identification. The different approaches confirmed the implantation and the niche occupation of bioprotection leading to the decrease of fungal communities (HTS) and Hanseniaspora uvarum cultivable population (MALDI-TOF MS). Yeast and fungi diversity was impacted by stage of maceration and, to a lesser extent, by bioprotection and SO2, resulting in a modification of the nature and abundance of the operational taxonomic units (OTUs) diversity.

Oenococcus oeni is the main bacterial species that drives malolactic fermentation in wine. Most O. oeni strains produce capsular exopolysaccharides (EPS) that may contribute to protect them in the wine hostile environment. In O. oeni genome sequences, several genes are predicted to encode priming glycosyltransferases (pGTs). These enzymes are essential for EPS formation as they catalyze the first biosynthetic step through the formation of a phosphoanhydride bond between a hexose-1-phosphate and a lipid carrier undecaprenyl phosphate. In many microorganisms, mutations abolishing the pGT activity also abolish the EPS formation. We first made an in silico analysis of all the genes encoding putative pGT over 50 distinct O. oeni genome sequences. Two polyisoprenyl-phosphate-hexose-1-phosphate transferases, WoaA and WobA, and a glycosyltransferase (It3) were particularly examined for their topology and amino acid sequence. Several isoforms of these enzymes were then expressed in E. coli, and their substrate specificity was examined in vitro. The substrate specificity varied depending on the protein isoform examined, and several mutations were shown to abolish WobA activity but not EPS synthesis. Further analysis of woaA and wobA gene expression levels suggests that WoaA could replace the deficient WobA and maintain EPS formation.

La spectrométrie de masse de type MALDI-TOF a été adaptée afin d’être utilisée comme outil innovant d’identification au niveau de l’espèce des levures et bactéries isolées d’échantillons variés (moûts, vins, boissons). L’analyse d’un grand nombre de clones permet d’apprécier la diversité des espèces de levures, bactéries acétiques et lactiques présentes dès les phases pré-fermentaires, au cours des fermentations, pendant l’élevage ou après conditionnement. Dans le cas d’altération de produits, cet outil innovant participera à une meilleure maitrise des risques microbiologiques.

No abstract