REALTIME PCR SYSTEMS TO MONITOR YEASTS IN GRAPE MUST AND WINE
Porret, Naomi Azur
The quality of beverages, particularly wine, depends largely on their microflora. Yeast and bacterial fermentations shape the product flavor profile, and can cause spoilage. Current detection methods for microorganisms are usually either inaccurate, slow, or both. In this project, realtime PCR systems were designed to test wines for the presence of certain microorganisms. The sampling method during wine production was optimized to ascertain that the microflora in the samples is representative of the whole tank. Four different sampling locations within the vertical axis of stainless steel tanks were tested during both the alcoholic and the malolactic fermentation. Saccharomyces cerevisiae and Oenococcus oeni were enumerated by plating. The results showed that a representative number of viable cells is obtained by taking a sample from the sampling valve. Part of the actin gene (Act1) was sequenced as the basis for the realtime PCR. Detection systems for the three undesirable wine yeasts Brettanomyces bruxellensis, Hanseniaspora uvarum, and Pichia anomala were designed and validated. The extraction methods were optimized to ensure that they were quantitative within a suitable range. The specificity of the detection systems as well at their threshold and range were very satisfactory. The realtime PCR system for B. bruxellensis was used to determine whether there is a quantitative connection between B. bruxellensis contamination and the marker phenols 4-ethylphenol, 4-ethylguaiacol, and 4-ethylcatechol or the subjective sensory impression. The experiments show that although B. bruxellensis does produce the known marker phenols, there is no mathematical correlation between the volatile phenol concentration and cell counts. The influence of sanitation procedures on the occurrence of B. bruxellensis in wineries was investigated using plating and realtime PCR. B. bruxellensis was found in most locations in wineries, including the wines. The inheritance of the teinturier phenotype was studied in a cross segregating for this character. Segregation results suggest that the primary gene controlling berry skin color also controls berry flesh pigmentation. This new detection method is faster, more accurate, and more sensitive than previous microbial and chemical detection methods. Routine testing of beverages would minimize the effects of spoilage microorganisms, resulting in decreased product losses for the beverage producers.
realtime PCR; spoilage yeasts; Brettanomyces
dissertation or thesis