The Csr system regulates genome-wide mRNA stability and transcription and thus gene expression in Escherichia coli

Abstract

Bacterial adaptation requires large-scale regulation of gene expression. We have performed a genome-wide analysis of the Csr system, which regulates many important cellular functions. The Csr system is involved in post-transcriptional regulation, but a role in transcriptional regulation has also been suggested. Two proteins, an RNA-binding protein CsrA and an atypical signaling protein CsrD, participate in the Csr system. Genome-wide transcript stabilities and levels were compared in wildtype E. coli (MG1655) and isogenic mutant strains deficient in CsrA or CsrD activity demonstrating for the first time that CsrA and CsrD are global negative and positive regulators of transcription, respectively. The role of CsrA in transcription regulation may be indirect due to the 4.6-fold increase in csrD mRNA concentration in the CsrA deficient strain. Transcriptional action of CsrA and CsrD on a few genes was validated by transcriptional fusions. In addition to an effect on transcription, CsrA stabilizes thousands of mRNAs. This is the first demonstration that CsrA is a global positive regulator of mRNA stability. For one hundred genes, we predict that direct control of mRNA stability by CsrA might contribute to metabolic adaptation by regulating expression of genes involved in carbon metabolism and transport independently of transcriptional regulation.

Figure 1. Maximal growth rates (μ_max) of the MG1655, MG1655(csrA51), MG1655(csrA51) complemented with a plasmid copy of the wild type csrA gene (csrA+), and MG1655(ΔcsrD) strain in M9 minimal medium supplemented with glucose.

Figure 2. Effect of the csrA51 and ΔcsrD mutations on mRNA half-life.

(a) Box plots of transcript half-life for the MG1655, MG1655(csrA51) and MG1655(ΔcsrD) strains (n = 3351 mRNAs). Values are separated into four quartiles by horizontal bars. The central bar (in the middle of the rectangle) represents the median value, which is given above the bar. VolcanoPlot of the log₂ fold change (Log₂ FC) of mRNA half-lives (b) between the MG1655(csrA51) and MG1655 strains (n = 3028 mRNAs) and (c) between the MG1655(ΔcsrD) and MG1655 strains (n = 3333 mRNAs). A P-value ≤ 0.1 was required for fold change significance (above the horizontal dashed line). The significantly stabilized mRNAs in the mutant strain compared to the MG1655 strain are colored in red whereas the significantly destabilized mRNAs are colored in green.

Figure 3. Level of CsrB and CsrC in the MG1655, MG1655(csrA51) and MG1655(ΔcsrD) strains in continuous culture.

(a) Northern blots of total RNA probed for CsrB and CsrC. (b) Hybridization signals quantified on a PhosphorImager. The signals for MG1655 were set to 100%.

Figure 4. Effect of the csrA51 and ΔcsrD mutations on mRNA levels.

VolcanoPlot of the log₂ fold change (Log₂ FC) of mRNA amounts (n = 4254 mRNAs) (a) between the MG1655(csrA51) and MG1655 strains and (b) between the MG1655(ΔcsrD) and MG1655 strains. A P-value ≤ 0.01 (above the horizontal dashed line) and a log₂ FC higher than 0.5 or lower than −0.5 (outside the vertical dashed lines) were required for fold change significance. Between the mutant strain and the MG1655 strain, the significantly up-regulated amounts were colored in red whereas the significantly down-regulated levels were colored in green.

Figure 5

(a) Variations in mRNA stability and quantity between the strains MG1655(csrA51) and MG1655. mRNAs with variation in level but not in stability in MG1655(csrA51) compared to MG1655 are in group I, those with variations in stability and quantity in opposite directions are in group II, and those with variations in stability and quantity not in opposite directions are in group III. (b) Plots of transcript half-life (in min) as a function of transcript quantity (in arbitrary units per mg of dry cell weight) in the MG1655 strain (n = 3074 mRNAs) and MG1655(csrA51) (n = 4098 mRNAs). All the values were log-transformed and centered. The Pearson correlation coefficient is −0.81 with a P-value < 2.2E⁻¹⁶.

Figure 6. Scheme of the connections described in this study between the transcriptional and post-transcriptional regulatory networks, CsrA and CsrD, which are involved in the genome-wide regulation of gene expression.

Filled arrows represent a connection type “act on” while the empty arrow represents a correlation.