The effect of natural selection on the propagation of protein expression noise to bacterial growth.

Laurens H J Krah,Rutger Hermsen,Christopher Rao

doi:10.1371/journal.pcbi.1009208

Abstract

In bacterial cells, protein expression is a highly stochastic process. Gene expression noise moreover propagates through the cell and adds to fluctuations in the cellular growth rate. A common intuition is that, due to their relatively high noise amplitudes, proteins with a low mean expression level are the most important drivers of fluctuations in physiological variables. In this work, we challenge this intuition by considering the effect of natural selection on noise propagation. Mathematically, the contribution of each protein species to the noise in the growth rate depends on two factors: the noise amplitude of the protein's expression level, and the sensitivity of the growth rate to fluctuations in that protein's concentration. We argue that natural selection, while shaping mean abundances to increase the mean growth rate, also affects cellular sensitivities. In the limit in which cells grow optimally fast, the growth rate becomes most sensitive to fluctuations in highly abundant proteins. This causes abundant proteins to overall contribute strongly to the noise in the growth rate, despite their low noise levels. We further explore this result in an experimental data set of protein abundances, and test key assumptions in an evolving, stochastic toy model of cellular growth.

Highlights

Stochasticity is inherent to gene expression [1,2,3]
We argue that natural selection, while shaping mean abundances to increase the mean growth rate, affects cellular sensitivities
Gene expression in bacterial cells is intrinsically stochastic: copy numbers of all proteins vary between genetically identical cells, even in a homogeneous environment

Summary

Introduction

Stochastic variation in the copy numbers of proteins is observed even under constant external conditions, and among individual cells in a population of isogenic bacteria. Noise in protein expression interferes with the functioning, survival and fitness of bacteria has been of great interest for many years [3,4,5,6,7,8]. Noisy gene expression is commonly accepted as the dominant mechanism behind the strong phenotypic variation that has been observed in populations of genetically identical cells [9]. Since cellular growth rate (and its population average) is often considered an important proxy for bacterial fitness, the growth rate—and how its variation is shaped by noisy gene expression—has received much attention [11,12,13,14]. Noise in the concentration of metabolic proteins is shown to propagate from the protein-level, via the metabolic network, to the instantaneous single-cell growth rate [15]

Methods

Results

Discussion

Conclusion