Assessing genetic and phenotypic diversity of Brettanomyces yeast

Brettanomyces yeast species, with Brettanomyces (Dekkera) bruxellensis being the most important, are generally reported to be spoilage yeasts in the beer and wine industry due to the production of phenolic off-flavors. The aromas imparted, which can be described as ‘medicinal and ‘barnyard’, are colloquially known as “Brett” character and are generally considered negative for beer and wine quality. However, the same compounds are regarded positively when produced during certain fermentation processes, such as the production of some styles of beer (e.g. lambic and gueuze). Despite its economic importance, surprisingly little is known about the biology, physiology and ecology of Brettanomyces yeasts. Herein, several aspects of Brettanomyces yeast biology and ecology were studied,  with particular emphasis on B. Bruxellensis, thus contributing to a better understanding of the biology and ecology of these important influencers of flavor profile.

In the first chapter (Chapter I), we give a comprehensive literature overview of the state-of-the-art of Brettanomyces research, emphasizing areas that were particularly well explored at the start of this PhD study, including aroma-associated aspects and methods for detection and identification. We also focused on recent genetic and genomic studies providing novel insights into the biology and evolution of B. bruxellensis. In Chapter II, we assessed the genetic relationships between 50 Brettanomyces strains belonging to all species presently identified within the genus and isolated from different food products and beverages using established DNA fingerprinting methods. These methods included ribosomal RNA (rRNA) gene sequencing, random amplified polymorphic DNA (RAPD) PCR, arbitrarily primed (AP) PCR and repetitive element PCR fingerprinting (rep-PCR). Our results support earlier findings that Brettanomyces yeasts form a genetically diverse clade, even within a species, and are represented by several subgroupings. Further our results revealed an intriguing correlation between B. bruxellensis genotype groups and the respective source of isolation, suggesting niche adaptation. To further explore this relationship, we first sequenced a (beneficial) beer isolate of B. bruxellensis (VIB X9085; ST05.12/22) and compared its genome sequence with the genome sequences of two wine spoilage strains (AWRI 1499 and CBS 2499) (Chapter III). In addition to strain-specific single nucleotide polymorphisms (SNPs) and insertions/deletions (InDels), structural genome variation was found between our strain and both wine strains, with some genomic regions specifically deleted in the beer strain. These included, but were not limited to, a region harboring the B. bruxellensis nitrate assimilation cluster and a region representing a cluster of genes mainly involved in carbon metabolism. Next, in Chapter IV, metabolic differences in carbon and nitrogen assimilation between different B. bruxellensis strains from different beverages (beer, wine and soft drink) were thoroughly assessed using Biolog Phenotype Microarrays. While some similarities of physiology were noted, many traits were variable among strains. Interestingly, some phenotypes were found that could be linked to strain origin, especially for the assimilation of particular α- and β-glycosides as well as α- and β-substituted monosaccharides. Based upon gene presence or absence, an α-glucosidase and β-glucosidase were found explaining the observed phenotypes. Further, using a PCR screen on a large number of B. bruxellensis isolates we have been able to specifically link a genomic deletion (e.g. harboring a β-glucosidase gene) to the beer strains, suggesting that this region may have a fitness cost for B. bruxellensis in certain fermentation systems such as brewing. Additionally, our work indicates that most beer strains are diploid, whereas the vast majority of wine strains are known to be triploid, suggesting that the additional set of chromosomes may confer a selective advantage in environments such as wineries. Finally, in Chapter V a series of fermentation tests were performed, in which different B. bruxellensis strains were inoculated in different media representative of different ecological niches of the yeast (i.e. beer, wine and soft drink). Utilisation of different sugars was quantified and production of (off-) flavors was monitored after one month of fermentation. Our results revealed that not only the medium (which may contain different (off-) flavor precursors), but also the yeast strain mediates the formation of the typical Brettanomyces (off-) flavors. Moreover, a (moderate) correlation was found between the origin of the strains and their impact on the volatile composition of the media. Only strains originally isolated from wine produced typical Brettanomyces off-flavors when inoculated in wine, whereas strains originally isolated from beer or soft drink did not, for example. Vice versa, in strong golden pale ale (Duvel), beer strains behaved differently from the other strains. Altogether, these results suggest that the interaction between medium and strain affects the outcome of potential (off-) flavor production.

Together, the components of this PhD study provide tools to discriminate Brettanomyces strains and reveal a first glimpse into the genetic diversity and genomic and phenotypic plasticity of B. bruxellensis. Our findings are relevant for the wine industry as well as for the beer industry. After all, deeper understanding of the ecology of B. bruxellensis not only provides novel insights into the evolution of this intriguing yeast species, it will also facilitate avoidance of wine spoilage and improvement of B. bruxellensis strains for brewing.