6533b856fe1ef96bd12b2344

RESEARCH PRODUCT

Characterization of the biofilm mode of life in the spoilage yeast Brettanomyces bruxellensis

subject

description

The management of contamination by the spoilage yeast Brettanomyces bruxellensis is a real challenge for the wine industry. The biofilm mode of life, known to increase the resistance of micro-organisms and to allow their persistence in the environment, is a strategy that can be adopted by B. bruxellensis.In this thesis project, microscopic observations revealed the presence of matrix around the cells, an essential element in the definition of a biofilm. The study also revealed that different morphotypes are involved in the structure of the biofilm, in particular filaments forming a network. Chlamydospore-like elements, never described before in B. bruxellensis, were observed within the biofilm but also in planktonic cultures. The production of such elements could be a strategy of the yeast to better persist in stressful environments. Significant differences in the amount of adhered cells were observed depending on the nature of the supports and media used, demonstrating the impact of the environment on biofilm formation in B. bruxellensis. In particular, the absence of glucose seems to decrease the adhesion capacity of several B. bruxellensis strains.In addition, agar invasion in B. bruxellensis newly described in this work is characterised by the development of diverse multicellular structures within the agar medium, composed notably of filaments. Optimised analysis through an image acquisition and processing pipeline revealed that the presence of glucose and oxygen favours agar invasion. B. bruxellensis also appears to be able to form biofilm-like structures such as air-liquid biofilms and complex colonies.Finally, the adhesion, biofilm formation and invasion abilities appear to be strain dependent, supporting the knowledge about the high intraspecific diversity in B. bruxellensis. Two rapid and reliable methodologies were adapted to discriminate strains within previously defined genetic groups: a RAPD-PCR protocol and a deep learning tool. The latter is based on the diversity of cell morphology to predict the genetic group of an isolate with an accuracy of 96.6%. This new approach opens the way for the implementation of simple routine methods accessible to the wine industry for the prevention of B. bruxellensis contamination risks.