Population heterogeneity in Saccharomyces cerevisiae and Escherichia coli lab scale cultivations simulating industrial scale bioprocesses

Today it is well known that a population of cells in a bioreactor is heterogeneous, opposite to traditional belief, and thus exhibiting distributions of single cell properties e.g. cell size, viability and metabolic activity rather than having a set of characteristics that can be described by averaged values. Population distributions always exist, but are significantly pronounced due to a combination of metabolic and stress responses of single cells travelling throughout the reactor experiencing gradients of substrate, pH and oxygen caused by non-ideal mixing in industrial scale bioprocesses. This thesis aimed at reaching a deeper understanding of how microbial physiology and cell dynamics are affected by the spatial heterogeneity in a bioreactor. Therefore large scale fermentation was simulated in laboratory scale using two of the most industrially relevant organisms E. coli and S. cerevisiae. Single cell distributions of cell size and fluorescence - originating from growth, cell membrane robustness and ethanol reporter strains or different fluorescence stains (for e.g. viability and metabolic activity) - were thereby followed by applying flow cytometry. Cell responses were studied in different cultivations modes, in steady state at different growth rates and in response to glucose perturbation in continuous culture, simulating the feeding zone of a large scale fed-batch fermentation and in batch culture to characterise the single cell behaviour in a dynamic environment. Furthermore, a two compartment chemostat setup, simulating different zones seen in large scale cultivations, was developed and studied under different compartmentalization conditions. The observed population heterogeneity distributions were, opposite to the common approach using mean values, described and validated in a quantitative manner through newly developed parameters, using percentile analysis followed by multivariate statistics as well as using a modeling approach.
In general the applied reporter strains as well as fluorescence stains in combination with flow cytometry showed to be valuable tools to study population heterogeneity in the different setups simulating large scale fermentation that can potentially be used in development and optimisation of industrial scale processes. Differences in growth and membrane robustness due to varying growth conditions and between slow and fast growing cells, different metabolic activities on different substrates and phenomena during compartmentalization which are hidden in a normal chemostat could be successfully visualised and quantified.