Quantitative analysis of microbial sensing and motility

Abstract

Microbes need to extract relevant information from their environment and use this information to produce adequate behavioral responses that ensure their survival. Quantitative, mathematical analysis of microbial sensory systems (such as various signaling pathways) and their effectors (such as bacterial motors) forms the basis of the field of systems biology. Because of their relative simplicity, in comparison with analogous systems in multicellular organisms, these structures are more amenable to quantitative modelling. In this dissertation I present the quantitative analysis of three microbial sensory-effector systems, two in bacteria and one in the unicellular eukaryote Saccharomyces cerevisiae. In all three cases I look at a behavior that is an evolutionarily selected response to a given problem that the microorganism is confronted with. I then explain the mechanistic basis of this response in the sensory or effector system by a mathematical model. In the first case, mating in yeast cells, the problem the cells need to solve is to establish the likelihood of mating and invest cellular resources accordingly to prepare for the mating event. The solution that wild-type yeast MATa cells have evolved to tackle this problem is fractional sensing, the ability to sense robustly the fraction of partner cells in a mixed population. The mechanism that enables this behavior is the degradation of the partner cells’ pheromone signal by a secreted enzyme. I show mathematically that the necessary consequence of this mechanism is the rescaling of the signal strength proportionally to the fraction of partner cells, as opposed to their absolute quantity. Additionally, I also explain the experimentally observed difference between the fractionally sensing wild-type cells and the mutants performing absolute sensing, due to the latter’s lack of a signal attenuation mechanism. Moreover, by a cost-benefit model of mating, I show that the strategy of fractional sensing and resource investment is optimal, as compared to sensing the absolute amount of partners. In the second case, I look at the most prevalent bacterial signaling systems, the so-called two-component systems and their capacity to generate bistability, or, in behavioral terms, memory. In the case of two-component systems that control developmental processes, an irreversible shift is required at the level of individual cells: once the system is turned ‘on’, it should not revert to its ‘off’ state, within some range of the input. At the population level, because of the stochasticity of chemical reactions and variation in expression levels, a bistable control system can result in a bimodal distribution with some cells in ‘on’ and others in ‘off’ state. In fluctuating and unpredictable environments this strategy of ‘bet hedging’ is another advantageous feature of bistability. I first describe post-translational mechanisms that can generate bistable behavior and analyze the parametric properties of bistable systems. Second, I show that the transcriptional auto-induction of pathway components can lead to bistability in the ‘canonical’ two-component system with a bifunctional sensor kinase as well, a question not resolved in the previous literature. In the third case, I analyze the motility of the marine bacteria Shewanella putrefaciens. Higher efficiency of spreading and chemotaxis is expected to lead to higher fitness as it enables a bacterial population to better explore and exploit the resources of its environment. Wild-type Shewanella cells achieve this higher efficiency by inducing a lateral flagellar system, leading to a lower mean turning angle. By lowering the mean of the turning angle distribution, the presence of the lateral flagella leads to higher directional persistence and hence increased spreading efficiency. By both analytical calculations and stochastic simulations I reproduce the experimentally observed trends of spreading. Furthermore, I show that in shallow gradients the higher directional persistence also leads to higher chemotactic efficiency. By mathematical analysis I was able to identify the mechanisms underlying these evolutionarily selected behaviors. Moreover, in the case of yeast mating, I also showed that the observed behavior of fractional sensing is optimal in cost-benefit terms. In the case of transcriptionally induced bistability in bacterial two-component systems, the analysis identified parametric properties of bistable systems that can be potentially used to engineer monostable signaling systems into bistable ones experimentally.