Pichia membranifaciens as spoilage yeast in Patagonian wines: isolation, origin and volatile phenols production

JULIETA S. SAEZ; CHRISTIAN A. LOPES; MARCELA P. SANGORRÍN

Carlos Paz, Córdoba, Argentina.

Congreso; VI Congreso Argentino de Microbiología General; 2009

Sociedad Argentina de Microbiología General (SAMIGE)

Yeasts belonging to the species Dekkera bruxellensis (anamorf Brettanomyces bruxellensis) have long been associated with serious economic damages in winemaking due to its ability to grow and to produce phenolic compounds.  In order to evaluate the potential contamination of two wines in a Patagonian familiar cellar, different yeast isolation protocols developed for D. bruxellensis were evaluated. Wine samples were filtered using 0.2 µm pore size membranes and subsequently seeded in two selective agar media: A (GPY + cicloheximide + ampicillin) and B (GPY + CaCO3 + ampicillin). Alicuots of the same wines were inoculated into two selective liquid media: C (yeast extract 3%, glucose 1,5%, NO3K 0,09%, ampicillin 0,2%) and D (YNB, glucose 1%, cicloheximide 0,001%, ampicillin 0,2%). After incubation at 26ºC, yeast cultures were isolated in the same (A and B) solid media. Yeast colonies from each wine were selected according with their macroscopic features and frequencies and identified by ITS1-5.8S rDNA-ITS2 PCR-RFLP analysis.  Only two yeast species (Saccharomyces cerevisiae and Pichia membranifaciens) were detected in both wines using three out of four different media. In medium A only S. cerevisiae colonies were observed. Because P. membranifaciens has been related to food spoilage (including wine film formation), we evaluated the spoilage potential of our isolates (production of volatile phenols in wine). We also evaluated the possible source of contamination (vineyard or cellar surfaces). Healthy and damaged (birds and moulds) grapes located in both shadowed and sunny vineyard areas were sampled before harvest in the same familiar cellar. Samples from cellar surfaces (13 fermentation vats and filters) were obtained by streaking with sterile cotton plugs and incubation in agitation using selective liquid medium. After growth, samples were seeded in the same solid media and identified by the same methodology described before. Hanseniaspora uvarum and Zygoascus hellenicus were detected in both healthy and damaged grapes from sunny areas; the same two species as well as Pichia guilliermondii were isolated from damaged grapes in the shadow and only H. uvarum in healthy grapes in the shadow.  Thirty-eight percent of the samples from vat surfaces did not showed yeast isolates; however, 62% evidenced the presence of Candida boidiniiand the species P. guilliermondii, Rhodotorula mucilaginosa and H. uvarum were only detected one vat each.  The only species detected in filters was P. membranifaciens.  The intraspecific characterization of the P. membranifaciens isolates by mtDNA-RFLP demonstrated that a same strain (molecular pattern D) was detected in both vines and filter. Finally, the capacity of producing volatile phenols by different P. membranifaciens isolates belonging to mtDNA-RFLP pattern D was evaluated by HPLC in laboratory scale fermentations using sterile wine added with the precursor p-coumaric acid (100 mg/L).  Average values of 0,6 mg/L of 4-etilphenol and 2,7 mg/L of 4-vinilphenol were detected, evidencing for the first time the capacity of this yeast species to produce these compounds in wine conditions. According with our results, P. membranifaciens must be considered as a potential spoilage yeast able to grow and to produce ethyl phenols in wine.