Analysis of molecular interactions between yoghurt bacteria by an integrated genomics approach

The lactic acid bacteria (LAB) are a group of Gram-positive bacteria that ferment sugars such as lactose to produce mainly lactic acid. LAB are a group of industrially important microorganisms that are applied for the production of many fermented foods. These include foods produced with substrates from plant origin (e.g. sauerkraut and wine) and animal origin (e.g. fermented meats and dairy products such as yoghurt). The current market trends regarding sustainability and health-promoting foods demand more efficient and a more diverse range of fermentations. Most fermentations are carried out by multiple strains of different species. The interactions between consortium members are at the base of the performances of the individual microorganisms within a microbial ecosystem and therewith of the whole fermentation. These microbial interactions are often poorly understood. Streptococcus thermophilus and Lactobacillus delbrueckii subsp. bulgaricus are two LAB species that upon fermentation convert (bovine) milk into yoghurt. These two bacteria stimulate each other in growth and acid production. They produce exopolysaccharides (EPS), important for the texture of yoghurt, and characteristic flavor compounds such as acetaldehyde and diacetyl. However, the molecular basis of the mutualistic interactions between these two bacteria was poorly characterized. In this thesis research, a combination was used of screening, mixed culture transcription profiling, whole-genome metabolic modeling, experimental evolution and next-generation sequencing. This was done to unravel the molecular basis of the interactions between S. thermophilus and L. bulgaricus in milk. The results showed that interactions were primarily based on the exchange of metabolites (see Figure 1). Moreover, it was shown which genes or pathways were affected. Evidence was found that S. thermophilus provided L. bulgaricus with formic acid, folic acid (both involved in purine metabolism), long-chain fatty acids (by the action of lipolytic enzymes to break down milk fat) and CO2. The proteolysis by the exoprotease of L. bulgaricus, in turn, provided both species peptides, which are taken up by the cell and broken down into amino acids (AA) by intracellular peptidases. However, this probably did not yield a sufficient supply of branched-chain and sulfur AA, leading to a higher expression of the genes for biosynthesis of these AA in both species when grown in mixed culture. Moreover, EPS biosynthesis genes were induced in the mixed culture, leading to increased EPS production and a higher viscosity of the yoghurt. Figure 1. Schematic representation of the mutualistic interactions between S. thermophilus and L. bulgaricus in yoghurt. Solid arrows indicate interactions; dotted arrows indicate pathways that are affected by the interactions. Pathways that were for the first time shown to be regulated at the transcriptome level upon co-culture are indicated in bold. Pathways that were confirmed in our study to be regulated at the transcriptome level upon co-culture are underlined. EPS is hypothesized to promote the exchange of both bacteria. There was no evidence at the transcriptome level for the exchange of putrescine and ornithine. AA, amino acids; BCAA, branched-chain AA; EPS, exopolysaccharides; LCFA, long-chain fatty acids. A mixed culture genome-scale metabolic model confirmed that cross-feeding interactions between the yoghurt bacteria were based on purine and AA metabolism. Moreover, this model was used to show that the interactions provided a significant benefit to both bacteria, i.e. their biomass yield on lactose increased by around 50% in mixed culture. Experimental evolution revealed that it is possible to co-adapt a novel combination of strains of S. thermophilus and L. bulgaricus. It was shown that their mutual stimulation increased by optimizing their interactions by fine-tuning expression of pathways involved in the interactions. Furthermore, as little as ~1000 generations of co-culture was sufficient to transform the relatively slow growing mixed culture into one that showed similar performance as commercial starters with respect to key characteristics such as acidification rate and viscosity. Improved understanding of the described interactions that are at the base of the yoghurt fermentation provides us targets for the rational optimization of existing mixed culture fermentations and the rational development of industrially relevant mixed cultures, such as those containing probiotics. Moreover, the results are in particular interesting for the field of microbial ecology as they show how mutual nutritional dependencies evolve and structure the microbial composition of this ecosystem.