Signal processing and decision making in single cells

Abstract

Cells are not simple passive observers oblivious to their environment, but sense and adapt to environmental changes in order to thrive. In addition to sensing the presence of signals in the environment, cells can extract information relating to the dynamics and spatial location of these signals and implement a response to these extracellular perturbations. This work examines a variety of signal-processing and decision-making processes across several different organisms. To explore the connection between biological network topology and temporal signal processing, we study how periodic signals are propagated in the Hog1 osmotic response pathway of the budding yeast Saccharomyces cerevisiae. Utilizing systems identification tools from control engineering, we study how the cells rapidly and robustly maintain osmotic homeostasis. By measuring the expression level of key proteins we begin to understand how fluctuating environments regulate gene expression. The lac operon in Escherichia coli has the ability to display a bistable, "all-ornothing" response to sugar. To understand how noise drives transitions between these two stable states, we measure switching dynamics in a population of cells. A simple model is constructed that can make predictions about system behavior unavailable from a deterministic model. Further, by measuring individual switching events in a similar bistable system implemented in the Galactose utilization pathway of Saccharomyces cerevisiae, we find that correlations in switching times of related individuals can be explained in terms of correlations in levels of key regulatory proteins. Many single celled organisms, such as the slime mold Dictyostelium discoideum, can sense and respond to concentration gradients of extracellular signaling molecules. We find that the cells' ability to detect an extracellular signal is influenced by an asymmetric intracellular signal, which varies in direction and magnitude from cell-to-cell. Further, a model that accounts for both signals predicts the observed population response to directed stimuli.

(cont.) Finally, we explore a "bet-hedging" strategy for fluctuating environments with an engineered population of Saccharomyces cerevisiae cells that randomly switch between two phenotypes. Each phenotype is fit to one of two alternating environments. We find that to optimize fitness, cells must tune the phenotypic transition rates in accordance with the rate of environmental transitions.