Using Torulaspora delbrueckii killer yeasts in the elaboration of base wine and traditional sparkling wine.

For still wines, killer strains of Torulaspora delbrueckii can be used instead of non-killer strains to improve this species' domination during must fermentation, with an ensured, reliable impact on the final wine quality. The present work analysed the usefulness of these killer yeasts for sparkling-wine making. After the first fermentation, the foaming capacity of T. delbrueckii base wines was very low compared to Saccharomyces cerevisiae base wines. Significant positive correlations of foaming parameters were found with the amounts of C4 -C16 ethyl esters and proteins, and negative with some anti-foaming alcohols produced by each yeast species. There were, however, no evident positive effects of polysaccharides on those parameters. The organoleptic quality of the T. delbrueckii base wines was judged inappropriate for sparkling-wine making, so that the following second-fermentation experiments only used a single assemblage of S. cerevisiae base-wines. While second fermentation was completed with inoculation of S. cerevisiae (both alone and mixed with T. delbrueckii) to yield dry sparkling wines with high CO2 pressure, single inoculation with T. delbrueckii did not complete this fermentation, leaving sweet wines with poor CO2 pressure. Yeast death due to CO2 pressure was much greater in T. delbrueckii than in S. cerevisiae, making any killer effect of S. cerevisiae over T. delbrueckii irrelevant because no autolysed cells were found during the first days of mixed-inoculated second fermentation. Nonetheless, the organoleptic quality of the mixed-inoculated sparkling wines was better than that of wines single-inoculated with S. cerevisiae, and showed no deterioration in foam quality. This seemed mainly to be because T. delbrueckii increased the amounts of ethyl propanoate and some acids (e.g., isobutyric and butanoic), alcohols (e.g., 3‑ethoxy‑1‑propanol), and phenols (e.g., 4‑vinylguaiacol). For these sparkling wines, no significant correlations between foaming parameters and aroma compounds were found, probably because the differences in foaming parameter values among these wines were fairly small. This is unlike the case for the base wines for which there were large differences in these parameters, which facilitated the analysis of the influence of aroma compounds on base-wine foamability.