Stochastic Properties of Transcription Factor Expression Revealed by Single-Molecule Noise Analysis

Abstract

Transcription factors (TFs) play important roles in gene regulation. Monitoring TF production in real time will help to elucidate how the stochasticity in the expression TFs influences the expression of their target genes. However, many TFs are expressed at low levels and are thus difficult to detect. Further, labeling TFs with fluorescent proteins can potentially disrupt their regulatory functions. Here, we developed a novel strategy, Co-Translational Activation by Cleavage (CoTrAC), to monitor the stochastic expression of a TF, λ repressor CI, in live E. coli cells at the single-molecule level. CI becomes fully functional upon co-translational cleavage from a membrane-targeted reporter, which can be counted individually. Using this strategy, we monitored the production of CI in real time in the regulatory context of the λ phage genetic switch. We developed a robust noise analysis to decompose the total noise in CI production into intrinsic, extrinsic and correlation sources. We show that intrinsic noise likely arises from both transcriptional and translational bursting. Extrinsic noise accounts for most of the total noise on the time scale of one cell cycle. CI enhances transcription primarily by increasing bursting frequency. Furthermore, negative autoregulation increases intrinsic noise, but counteracts extrinsic noise and ensures that molecular memory diminishes on a time scale shorter than one cell cycle. The experimental and theoretical approaches presented here offer a unique opportunity to investigate noise in gene regulatory networks.