The chemostat Evolutionary pressures on microbial metabolic strategies in the chemostat In collaboration with :

Protein expression is shaped by evolutionary pressures. Due to limitations in biosynthetic capacity, the costs and benefits of enzyme production are important determinants of fitness. While these processes are well understood in batch conditions, in chemostats, which are extensively used to study microbial evolution in a laboratory setting, a feedback of the microbial physiology on the conditions in the chemostat hinders an intuitive understanding. Here, we aim to provide a solid theoretical framework of the selective pressures and optimal evolutionary strategies in the chemostat. We show that the optimal enzyme levels can be described with control theory and that optimal strategies are implemented by well-defined metabolic subsystems, known as elementary flux modes, similar to batch conditions. However, as we illustrate with a realistic coarse-grained model of the physiology and growth of the yeast Saccharomyces cerevisiae, the evolutionary dynamics and final outcome of evolution in a chemostat can be very different. Simulated evolution of respiro-fermentative yeast cells in a chemostat at an intermediate dilution rate shows an evolutionary stable coexistence of a strictly respiring and a strictly fermenting strain. Our results connect a kinetic, mechanistic view of metabolism with cellular physiology and evolutionary dynamics. We provide a theoretical framework for interpreting and reasoning about selection and evolution experiments in the chemostat. Introduction Growth rate relates intimately to fitness for unicellular organisms. The specific growth rate is the rate of biomass synthesis per unit biomass, which implies a strong selective pressure on the efficient usage of biomass. Expression of non-functional proteins reduces growth-rate in batch (Dong et al , 1995; Shachrai et al , 2010) and incurs a fitness cost in chemostat conditions (Novick and Weiner, 1957; Dean et al , 1986; Dean, 1989; Lunzer et al , 2002; Stoebel et al , 2008). Deviations from wild type enzyme expression also reduce fitness (Jensen et al , 1993; Snoep et al , 1995). The importance of the cost and benefit of enzyme expression is apparent from an evolutionary study where Escherichia coli attained predicted, environment-dependent optimal expression levels within 500 generations (Dekel and Alon, 2005). In two recent publications (Berkhout et al , 2013b; Wortel et al , 2014) we studied the questions: “Which enzymes should be expressed and to what concentration to achieve growth rate maximisation?” The first question pertains to the choice of metabolic “strategy”, i.e. which pathways should be expressed under the prevailing conditions. For example, should a yeast cell adopt a respiratory or a fermentative mode of metabolism, or a combination thereof? A growth maximizing metabolic strategy always turn out to be a so called Elementary Flux Mode (EFM) (Wortel et al , 2014), which is a minimal “route” through a metabolic network able to operate at steady state. In a sense, an EFM is a “pure” metabolic strategy, e.g. fermentation or respiration, but not respiro-fermentation (Wortel et al , 2014). The second question pertains to the biosynthetic cost and catalytic benefit of the different enzymes within a metabolic pathway. We proposed a general definition for the benefit and cost of enzyme expression and showed that maximization of benefit minus cost for each enzyme maximizes the growth-rate (Berkhout et al , 2013b). Furthermore, for each enzyme, the optimal relative enzyme concentration equals the