Fig 1.
The L. pneumophila csrA- mutant shows a defect in intracellular growth.
The L. pneumophila csrA- mutant has a defect in replication in Acanthamoeba castellanii. Grey, wt L. pneumophila strain Paris; green, ΔletA-mutant; Blue, ΔrsmYZ double mutant; Red, csrA—mutant. Infections were performed at 37°C and the number of intracellular, viable bacteria was determined by the standard plate count assay. Each time point represents the mean +/- SD of three biological replicates.
Fig 2.
Half of the L. pneumophila proteins are differentially expressed upon CsrA deletion.
Protein intensities in the wt and csrA- strains (three biological replicates, csrA- = Mut1-3; wt = WT1-3) were measured by differential shotgun proteomics and visualized in a heat map (left) and a profile plot (right) after non-supervised hierarchical clustering. Every row represents a quantified protein (n = 1448) for which the normalized (LFQ) intensity in each biological replicate is color indicated in the columns.
Table 1.
Summary of RNA targets of CsrA identified by using Co-immunoprecipitation with Flag-tag antibodies and subsequent deep sequencing (RIPseq).
Table 2.
Summary of the genes influenced by CsrA and discussed in detail.
Fig 3.
CsrA interacts directly with fleQ-mRNA in vitro.
A) Electromobility shift assays (EMSA) with 200nM of biotinylated fleQ-mRNA combined with varying concentrations of purified CsrA-His were undertaken in a 6% Native Tris-PAGE gel. Lane 1: no CsrA, lane 2: 0.2 μM CsrA, lane 3: 0.5 μM CsrA, lane 4: 1.0 μM CsrA, lane 5: 2.0 μM CsrA, lane 6: 5.0 μM CsrA, lane 7: 5.0 μM CsrA + 2.0 μM unlabeled RsmZ. B) Mfold secondary structure prediction of the fleQ-mRNA fragment used for the EMSA. Red, two potential CsrA-binding sites, which are mutated in C. Blue, transcriptional start codon. C) EMSA with recombinant CsrA and 200 nM of non-mutated RNA (pFleQ) or mutated in the indicated regions (mFleQ). AGGA motifs were replaced by an AAAA sequence using PCR mutagenesis. Lane 1: no CsrA + non-mutated fleQ-mRNA, lane 2: 5.0 μM CsrA + non-mutated fleQ-mRNA, lane 3: 5.0 μM CsrA + mfleQ-mRNA mutated in region 1, lane 4: 5.0 μM CsrA + mfleQ-mRNA mutated in region 2, lane 5: 5.0 μM CsrA + mfleQ-mRNA mutated in both region 1 and 2.
Fig 4.
CsrA interacts directly with lqsR mRNA in vitro.
A) Electromobility shift assay (EMSA) with 200nM of biotinylated lqsR mRNA combined with varying concentrations of purified CsrA-His lqsR mRNA and recombinant CsrA in 6% Native Tris-PAGE. Lane 1: no CsrA, lane 2: 0.2 μM CsrA, lane 3: 0.5 μM CsrA, lane 4: 1.0 μM CsrA, lane 5: 2.0 μM CsrA, lane 6: 5.0 μM CsrA, lane 7: 5.0 μM CsrA + 2.0 μM unlabled RsmZ. B) Mfold secondary structure prediction of the lqsR mRNA fragment used for the EMSA. Red, two potential CsrA-binding sites, which are mutated in C. Blue, transcriptional start codon. C) EMSA with recombinant CsrA and 200 nM of non-mutated RNA (pLqsR) or mutated in the indicated regions (mLqsR). AGGA motifs were replaced by an AAAA sequence using PCR mutagenesis. Lane 1: no CsrA + non-mutated lqsR mRNA, lane 2: 5.0 μM CsrA + non-mutated lqsR mRNA, lane 2: 5.0 μM CsrA + mlqsR mRNA mutated in region 1, lane 3: 5.0 μM CsrA + mlqsR mRNA mutated in region 2, lane 4: 5.0 μM CsrA + mlqsR mRNA mutated in both region 1 and 2.
Fig 5.
CsrA feedback regulation and autoregulation of the stringent response.
A) Electromobility shift assay (EMSA) with 200nM of biotinylated RNA demonstrating the interaction of purified CsrA and relA mRNA in vitro. 200nM of biotinylated relA mRNA and increasing concentrations of recombinant CsrA in 6% Native Tris-PAGE were used. Lane 1: no CsrA, lane 2: 0.5 μM CsrA, lane 3: 1.0 μM CsrA lane 4: 2.0 μM CsrA, lane 5: 5.0 μM CsrA, lane 6: 5.0 μM CsrA + 2.0 μM unlabeled RsmZ. B) EMSA with 200nM of biotinylated RNA demonstrating the interaction of purified CsrA and rpoS mRNA in vitro. 200nM of biotinylated rpoS mRNA and increasing concentrations of recombinant CsrA in 6% Native Tris-PAGE were used. Lane 1: no CsrA, lane 2: 1.0 μM CsrA, lane 3: 2.0 μM CsrA, lane 4: 5.0 μM CsrA, lane 5: 5.0 μM CsrA + 2.0 μM unlabeled RsmZ. C) Model of the stringent response and quorum sensing network in L. pneumophila and the role of CsrA on their regulation. During the transmissive phase, amino acid and fatty acid starvation triggers the GTP pyrophosphokinase RelA and the ppGpp synthetase/hydrolase SpoT to produce the alarmone (p)ppGpp. Amongst others, the (p)ppGpp production results in a higher transcription rate of the small ncRNAs RsmX, RsmY and RsmZ which dissociate the RNA-binding protein CsrA from its target-RNAs. This leads to an activation of RpoS, LqsR and PmrA expression (positive feedback) that were formerly repressed by CsrA and an inhibition of RelA (negative feedback). Predicted negative effects in the regulatory cascade are represented by red lines, positive effects by black arrows.
Fig 6.
Glyceraldehyde 3-phosphate (Gap) and transketolase (Tkt) transcription is regulated differently by CsrA.
A) Schematic organization of the PPP/Glycolysis-operon in L. pneumophila Paris. TSS indicates the transcriptional start site under the control of an RpoD-dependent promoter. Green, bold arrows show the CsrA-binding region and black arrows highlight the region where qRT-PCR was conducted. B) EMSA with 200nM of biotinylated RNA demonstrating the interaction of purified CsrA and gap mRNA: Lane 1: no CsrA, lane 2: 0.5 μM CsrA, lane 3: 1.0 μM CsrA, lane 4: 2.0 μM CsrA, lane 5: 5.0 μM CsrA, lane 6: 5.0 μM CsrA + 2.0 μM unlabled RsmZ. Right side, run-off transcript produced under optimal in vitro transcription conditions performed with the MEGAshortscript Kit (ambion) to show transcript length compared to the low range ssRNA ladder (NEB). C) qRT-PCR results of the gap and the tkt transcripts at different growth stages (OD) between wt and csrA- show lower expression levels of the gap gene in E-phase (OD1-3) in absence of CsrA whereas tkt is not affected. No differences are noticed during transition (OD3) and PE-phase. Complementation of the csrA- strain restored the wt transcript levels
Fig 7.
CsrA acts as a positive regulator for Glyceraldehyde 3-phosphate (Gap) by preventing premature transcriptional termination of the PPP/Glycolysis-operon.
A) RNA secondary structure Mfold-prediction of the CsrA-binding region inside the gap gene reveals two major conformations: the left one contains a potential hairpin-terminator while the A(N)GGA-motif is covered in a double-strand region with low affinity to CsrA. The right one, shows the A(N)GGA-motif located in an open loop with high CsrA-interaction affinity and the hairpin structure is disrupted. Below, the nucleic acid sequence is shown that was used for Mfold modeling and the transcription termination assays. Red, CsrA-binding site A(N)GGA; green, the potential transcription terminator hairpin; blue, the putative auxiliary element of Rho-dependent termination. B) Left panel: In vitro transcription termination assay in presence of 1 μM of purified NusG-protein and varying concentration of Rho- and CsrA-protein (+ 0.5 μM, ++ 1μM; Lane 1: no Rho, no CsrA, lane 2: 1 μM Rho, no CsrA, lane 3: 1 μM Rho, 0.5 μM CsrA, lane 4: 1 μM Rho, 1 μM CsrA, lane 5: 1 μM Rho, no CsrA, lane 6: 1 μM Rho, 1 μM CsrA). A representative 10% urea-PAGE gel shows the formation of the truncated transcript from the Rho-dependent termination without CsrA and the full-length transcript with CsrA. Right panel: In vitro transcribed of the run-off fragment and the marker showing the size of the fragment. C) Regulatory model of the transcription of the PPP/Glycolysis operon. In absence of CsrA, Rho-dependent termination within the operon is responsible for polarity effects downstream of the transcriptional block. This leads to reduced transcript-levels of the gap gene whereas the tkt gene is not affected. When CsrA binds to the RNA, an anti-terminator structure is favored preventing that the elongation complex stalls at the hairpin structure. As a consequence, only the presence of CsrA ensures the efficient transcription of the glycolysis/gluconeogenesis genes of the operon.
Fig 8.
CsrA modulates the expression of a thiamine pyrophosphate (TPP) riboswitch element.
A) Schematic representation of the thi-operon in L. pneumophila including the transcriptional start site (TSS), the CsrA-binding region, the thi element of the predicted TPP riboswitch and a predicted transcription termination site upstream of the start codon (AUG). The CsrA-binding site is overlapping the thi element of the TPP riboswitch. This organization suggests that CsrA is implicated in the fine-tuning of the expression of the downstream thi-operon most probably due to conformational changes in the secondary RNA structure. B) EMSA with 200nM of biotinylated thi-element (TPP) RNA and purified CsrA: Lane 1: no CsrA, lane 2: 1.0 μM CsrA, lane 3: 2.0 μM CsrA, lane 4: 5.0 μM CsrA, lane 5: 5.0 μM CsrA + 2.0 μM unlabeled RsmZ. C) Beta-lactamase (BlaM) assay in minimal medium grown Legionella without, with 1 mM and with 2 mM of TPP. BlaM activity in 10μg total protein of wt and csrA- strain containing the 5'UTR of the thi-operon in a pXDC61 plasmid was measured. Each value represents the mean +/- SD of three independent experiments. BlaM activity is significantly decreased in the mutant at the different conditions indicating a positive effect of CsrA on the thi-operon expression in L. pneumophila. D) Model of the TPP riboswitch modulated by CsrA. Mfold prediction of the secondary structure of the 5'UTR thi-region: When TPP is bound, the OFF state of the riboswitch is favored in which the expression of the operon is inhibited (most likely due to premature termination at the predicted termination site). The presence of CsrA in contrast might stabilize the ON state where the structure of the thi-element is dispersed, hence, higher amounts of TPP would be necessary to shift the element back to the OFF state leading to the down-regulation of the thi-genes expression.
Fig 9.
Model how Legionella CsrA influences the pyruvate metabolism and which different regulatory functions it exerts.
A) Carbon flux into the energy production is favored by CsrA, whereas production of the storage molecule PHB is repressed. Additionally, amino acids and glucose, but not glycerol are the preferred carbon source in presence of CsrA. Proteins in red represent a negative effect of CsrA on the pathway, whereas in green, proteins are shown to be under positive control of CsrA. In black, CsrA interacts with the RNA, but no quantitive difference was observed under our condition. PrsA (Lpp0607), Ribose-phosphate pyrophosphokinase, RpiA (Lpp0108), Ribose-5-phosphate isomerase A, GlpD (Lpp1368), Glycerol-3-phosphate dehydrogenase, GlpK (Lpp1369), glycerol kinase, Tpi (Lpp2838), Triosephosphate isomerase, Zwf (Lpp0483), Glucose-6-phosphate 1-dehydrogenase, Pgl (Lpp0484), 6-Phosphogluconolactonase, Edd (Lpp0485), 6-Phosphogluconate dehydratase, Gap (Lpp0153), Glyceraldehyde 3-phosphate dehydrogenase, Pgk (Lpp0152), 3-Phosphoglycerate kinase, Eno (Lpp2020), Enolase, Pyk (Lpp0151), Pyruvate kinase, PpsA (Lpp0567), Phosphoenolpyruvate synthase, Pdh (Lpp1461), Pyruvate dehydrogenase complex, Pyc (Lpp0531), Pyruvate carboxyltransferase, Ppc (Lpp1572) Phosphoenolpyruvate carboxylase, SfcA (Lpp3043), NAD-specific malic enzyme, AcnA (LPP1659), Aconitate hydratase, Icd (Lpp0878), Isocitrate dehydrogenase, Sdh (Lpp0595), Succinate dehydrogenase, Suc (Lpp0597), 2-Oxoglutarate dehydrogenase, Ald (Lpp0986), Alanine dehydrogenase, SdaA (Lpp0854), Serine dehydratase, PhbB (Lpp0621), acetoacetyl-CoA reductase, PhbC (Lpp2038), Polyhydroxyalkanoate synthase, Adc (Lpp0728) Acetoacetate decarboxylase, Bdh (Lpp2264), 3-Hydroxybutyrate dehydrogenase. B) CsrA can act as a negative regulator of the translation initiation process by blocking the ribosome binding site of the RNA and hence, interfering with its ribosome interaction. Examples in L. pneumophila are the transcriptional regulator FleQ and the quorum sensing response regulator LqsR. Binding of CsrA leads to a conformational re-organization of the target-RNA. As a consequence, the RBS is better accessible for the ribosome yielding in a translational activation due to CsrA interaction. This mode of action might be relevant for the relA mRNA in L. pneumophila. CsrA interaction with the RNA can stabilize the target-RNA by blocking RNase-specific binding sites. Contrary, also a destabilization can be triggered by CsrA when its binding leads to conformational changes of the RNA that facilitate the attack of an RNase. In Legionella, we suggest that the fur mRNA is protected by CsrA against degradation by binding to an A(N)GGA motif overlapping a putative RNase E recognition site. Finally, CsrA can affect transcriptional elongation in a negative (promoting termination) or in a positive way (stabilizing an anti-terminator structure). The transcription of the gap gene in L. pneumophila is only guaranteed in presence of CsrA as binding of the protein prevents the Rho-dependent termination downstream of the tkt gene part of the PPP/Glycolysis operon.