William Pontius

Biological pathways perform calculations with often-small numbers of constituent molecules, leading to potentially significant variability in their output. In this thesis, I use the chemotaxis pathway of the bacterium Escherichia coli as a model to investigate the molecular origins of large temporal fluctuations and their consequences for cellular behavior. The bacterial chemotaxis pathway is a simple sensory network that performs temporal comparisons of external chemical stimuli, enabling the bacterium to perform a random walk biased toward increasingly favorable conditions. In this thesis, I first analyze measurements of living cells and argue that the statistics of pathway noise and the cellular response to stimuli, which both arise from the same biochemical pathway, are intrinsically linked. I then use theoretical models to argue that noise in the bacterial chemotaxis pathway may have significant positive consequences for the behavior of the cell: by coordinating the behavior of independent, stochastically switching flagellar motors, noise may enable the cell to respond more quickly to stimuli, track weak chemoattractant gradients more effectively, and explore sparse environments more efficiently. Finally, I construct a detailed, calibrated stochastic model of the mechanism through which the chemotaxis system adapts to persistent stimuli and identify the specific architectural features—densely clustered chemoreceptors and an enzyme localization mechanism—that give rise to large pathway fluctuations. I further argue that these features giving rise to pathway noise also underlie other well-known properties of the chemotaxis system: precise adaptation and functional robustness to expression levels of the pathway constituents. The simplicity of the bacterial chemotaxis system and the ubiquity of many of its architectural features suggest that these results will be relevant to the study of pathway noise in other sensory systems and throughout biology.