The stochastic nature of intracellular control circuits

Sammanfattning: Chemical reactions are probabilistic by nature, causing random fluctuations inconcentrations to emerge spontaneously. In conventional (macroscopic) kinetics this internal noise is ignored under the assumption that average copy numbers are so high that fluctuations are negligible in comparison. By contrast, intracellular regulatory circuits often operate on components present in a few to a few hundred copies. In close analogy with the critical phenomena studied in statistical physics, many metabolic and genetic pathways are also kinetically designed in ways that inevitably generate random ultravariation. Macroscopic kinetics is then not only unsuitable because averages take asubordinate position to fluctuations, but also because it does not accurately account for averages when reaction rates depend nonlinearly on noisy concentrations. In such cases it is necessary to take a stochastic molecular-level (mesoscopic) approach to kinetics where rate equations are replaced by birth and death master equations.This thesis is mainly a master equation analysis of internal noise in intracellular processes. By connecting Biochemical Systems Theory to stochastic kinetics, it is shown how basic principles of metabolic and genetic control circuits, such as sensitivity, robustness and homeostasis, all directly relate to the significance of internal noise. The commonly held notion that signal noise invokes an increased response noise is shown to be based on a linear perspective on fluctuations. Since regulation is generally nonlinear, signal noise may also affect the sensitivity with which reactionrates respond to changes in the average signal concentration. Insensitive regulatory mechanisms can in fact exploit the signal noise that arises from biochemical reactions for increased sensitivity: Stochastic Focusing. In regulatory circuits this may result in `kinetic uncertainty principles' where the uncertainty in one component can be decreased only by increasing the uncertainty in another component.The origins and effects of internal noise are mainly exemplified here by copynumber control of plasmids Co1E1 and R1. Plasmids are simple, well-characterized systems and subject to a strong selective pressure to reduce both copy number averages and fluctuations, making them ideal for mesoscopic modeling.