Abstract
Gene expression can be noisy1,2, as can the growth of single cells3,4. Such cell-to-cell variation has been implicated in survival strategies for bacterial populations5–7. However, it remains unclear how single cells couple gene expression with growth to implement these survival strategies. Here we show how noisy expression of a key stress response regulator, rpoS8, allows E. coli to modulate its growth dynamics to survive future adverse environments. First, we demonstrate that rpoS has a long-tailed distribution of expression in an unstressed population of cells. We next reveal how a dynamic positive feedback loop between rpoS and growth rate produces multi-generation rpoS pulses, which are responsible for the rpoS heterogeneity. We do so experimentally with single-cell, time-lapse microscopy9 and microfluidics10 and theoretically with a stochastic model11,22. Finally, we demonstrate the function of the coupling of heterogeneous rpoS activity and growth. It enables E. coli to survive oxidative attack by causing prolonged periods of slow growth. This dynamic phenotype is captured by the rpoS-growth feedback model. Our synthesis of noisy gene expression, growth, and survival paves the way for further exploration of functional phenotypic variability.