In order to study the CRP-cAMP mediated regulation of CsrB and CsrC in Salmonella, the expression of both sRNAs was monitored using transcriptional reporter fusions. The regulatory elements controlling CsrB and CsrC expression in Salmonella have been identified in silico and characterized experimentally (Martínez et al., 2014). Accordingly, the described regulatory regions of CsrB (−403, +18) and CsrC (−347, +60) were cloned as lacZ transcriptional fusions in the pQF50 vector (Farinha and Kropinski, 1989), allowing to monitor the levels of CsrB and CsrC during growth. The contribution of CRP-cAMP was assessed by determining the expression level of CsrB and CsrC in wild-type (WT) and in a Δcrp mutant strain, lacking the transcriptional factor CRP. In rich media, no differences in growth rate could be observed in WT and Δcrp strains carrying either csrB-lacZ or csrC-lacZ (Figure 1A). Transcriptional expression was monitored during logarithmic growth (OD_600 nm 0.4) and upon entry into stationary phase (OD_600 nm 2.0). In the WT, CsrB expression was apparently identical in the two growth phases (Figure 1B). By contrast, CsrC displayed a growth-dependent expression pattern, being less expressed during logarithmic growth and being induced after entering stationary phase (Figure 1B). Remarkably, CRP-cAMP contributes to the CsrC growth-dependent expression pattern. In Δcrp, the expression of CsrC is induced when compared to WT at the logarithmic phase and to a lesser extent at the stationary phase, indicating that CRP-cAMP represses CsrC expression (Figure 1B). By contrast, no effect on CsrB regulation by CRP-cAMP was observed. These results were unexpected since it was previously described that CRP-cAMP acts as an activator of both CsrB and CsrC expression when Salmonella is grown on LB agar media (Teplitski et al., 2006). To discern if the discrepancy could be consequence of differences in the genetic constructs used to monitor gene expression, a similar experiment as in Teplitski et al. (2006) was performed using our strains. The transcriptional expression was monitored after growth on LB agar media. Consistently with the previous report, the transcriptional expression of CsrB and CsrC was strongly diminished in a Δcrp derivative strain compared to WT when Salmonella cells were grown on LB agar media as noted by the white colony phenotype of Δcrp when compared to the blue colony phenotype of WT (Supplementary Image S1; Teplitski et al., 2006). Altogether, CRP-cAMP seem to be required for activation of the expression of CsrB and CsrC in solid media, while it acts as a repressor of CsrC particularly at logarithmic growth phase. This indicates that CRP-cAMP-mediated regulation of CsrB and CsrC is dependent on the growth conditions.

The differential regulation of CsrC and CsrB by CRP-cAMP at the logarithmic growth phase was further corroborated by the direct RNA detection of CsrB and CsrC. In the WT, CsrC transcript was barely detected, indicating that CsrC expression is tightly silenced. In the Δcrp strain, high levels of CsrC were detected indicating that CRP-cAMP is involved in the CsrC silencing during logarithmic growth (Figure 1C). Contrasting the CRP-cAMP-dependent changes in CsrC levels, CsrB was not influenced by knockout of Δcrp (Figure 1C), thereby agreeing with the results of our transcriptional fusions (Figure 1B). In addition, deletion of ΔcsrB cause a mild upregulation of CsrC, presumably by affecting the positive feed forward loop that free CsrA protein exerts on its repressors CsrB and CsrC (Romeo and Babitzke, 2018).

CRP becomes active upon binding to cAMP, which is produced by the adenylate cyclase Cya. Therefore, absence of crp or cya should display similar expression profiles. Accordingly, the expression of CsrC in logarithmic growth phase is induced in the Δcya derivative strain when compared to WT while no change was observed in the expression of CsrB (Figure 1D). To further confirm the involvement of CRP-cAMP in CsrC regulation, chemical complementation of Δcya was carried out by ectopic addition of cAMP. The addition of cAMP repressed the expression of CsrC in the Δcya derivative strain while no effect was observed in CsrB expression (Figure 1E). Our data indicate that CRP-cAMP is involved in the growth-dependent regulation of CsrC by repressing its expression.

To further characterize the CRP-cAMP mediated regulation of CsrC, a chromosomal csrC-lacZ fusion was generated. In agreement with earlier results, chromosomal csrC-lacZ has a growth dependent expression pattern, where it is lowly expressed at the logarithmic growth phase and induced upon entry into early stationary phase (Figure 2A). Interestingly, we further observed an eightfold induction of the chromosomal csrC-lacZ fusion in the Δcrp background when compared to the WT, which was only true during logarithmic growth (Figure 2A).

The BarA-SirA two-component system was described to positively regulate the expression of CsrC (Teplitski et al., 2006). The possible involvement of BarA-SirA in the regulation of CsrC by CRP-cAMP was assessed. SirA deletion leads to a decrease of the overall transcriptional expression of the chromosomal csrC-lacZ fusion, 28.2 ± 0.8 Miller units in ΔsirA compared to 308.8 ± 19.3 Miller units in WT. Interestingly, despite the overall lower expression levels of CsrC in the ΔsirA background, the deletion of Δcrp led to a sixfold increase in the transcriptional expression of csrC-lacZ compared to WT (Figure 2B), indicating that CRP-cAMP modulates csrC expression via a SirA-independent mechanism.

In E. coli, it was proposed that CRP-cAMP represses CsrC expression through binding to the upstream region of csrC and competing with UvrY, the SirA homolog (Pannuri et al., 2016). In Salmonella, SirA induces CsrC expression by binding ∼160–168 bp upstream of the csrC promoter (Figure 3A; Martínez et al., 2014). Alignment of the upstream promoter regions of CsrC from E. coli and Salmonella showed that SirA binding site is conserved, whereas the CRP-cAMP binding site displays a lesser extent of conservation (Supplementary Image S2). To characterize if CRP-cAMP binding to the csrC promoter is required for the described regulation, an additional csrC-lacZ fusion was generated where the putative binding sites of these transcriptional factors are not present. As shown in Figure 3A, the pQF50 cloned fragment csrC₃₄₇-lacZ (−347, +60) maintains SirA/CRP binding sites while csrC₉₁-lacZ (−91, +60) does not. Remarkably, in both csrC₃₄₇-lacZ and csrC₉₁-lacZ an induction of expression is observed in the Δcrp strain when compared to WT (Figure 3B). These results suggest that CRP-cAMP represses the expression of CsrC during logarithmic growth phase via a mechanism that is independent of the competition between SirA and CRP for the csrC promoter.

In Salmonella, CRP-cAMP represses the expression of the regulatory sRNA Spot 42 in a growth-dependent manner (El Mouali et al., 2018). The similarity of the expression profiles of Spot 42 and CsrC, let us to hypothesize that CRP-cAMP might regulate CsrC via modulation of Spot 42 levels. To this end, we determined CsrC expression during logarithmic growth upon ectopic expression of Spot 42. Remarkably, CsrC transcript levels were strongly induced upon Spot 42 overexpression (pBRSpot 42) when compared to the vector control strain (pBRVC) (Figure 4A). In contrast, no accumulation of CsrB was detected upon overexpression of Spot 42 (Figure 4A). Further supporting our hypothesis, Spot 42 seemed to differentially regulate CsrB and CsrC expression (Figure 4A), similar to the described CRP-cAMP-mediated regulation (Figure 1C). To verify this observation, we overexpressed Spot 42 in the chromosomal csrC-lacZ fusion background, resulting in a threefold induction of CsrC when compared to the vector control strain (Figure 4B). The Spot 42-mediated regulation of CsrC may account for the described repression of CsrC by CRP-cAMP. The contribution of Spot 42 on the induction of CsrC in Δcrp was further assessed. The CsrC derepression in a Δcrp mutant strain drops in absence of Spot 42 (Δspf) as compared to a Spot 42 proficient strain indicating that Spot 42 contributes to the CsrC derepression in a Δcrp background (Figure 4C).

Interestingly, it has been described that SirA can bind to the promoter region of spf (Zere et al., 2015), suggesting that SirA might be regulating Spot 42. However, transcriptomic data of ΔsirA compared to WT in Salmonella indicates that Spot 42 it is not regulated by SirA (Colgan et al., 2016). The effect of SirA on Spot 42 expression was assessed, using a chromosomal spf-lacZ in the presence and absence of SirA in both WT and Δcrp genetic backgrounds at logarithmic growth phase. As previously shown, Spot 42 expression is induced in the Δcrp mutant compared to WT (El Mouali et al., 2018; Figure 5A). Remarkably, no regulation was observed in the absence of SirA and induction of Spot 42 expression in the Δcrp mutant strain is still observed in absence of SirA (Figure 5A). These results indicate that SirA it is not regulating Spot 42 expression under the studied conditions. Furthermore, SirA was not required for the positive effect of Spot 42 on CsrC expression as concluded from csrC-lacZ expression studies under Spot 42 overexpression in both WT and ΔsirA. In both genetic backgrounds, the overexpression of Spot 42 induces the expression of csrC-lacZ to a similar fold (Figure 5B), consistent with no involvement of SirA in the CRP-cAMP mediated regulation of CsrC.

Altogether, our data indicate that CRP-cAMP differentially regulates CsrB and CsrC at the logarithmic growth phase where it represses specifically CsrC but not CsrB. The CRP-cAMP regulated sRNA Spot 42 arise as new regulator of the Csr regulon in Salmonella.

The bacterial strains, derivatives of Salmonella enterica serovar Typhimurium SL1344, were cultivated in Lysogeny broth (LB; tryptone 10 g/l, yeast extract 5 g/l and sodium chloride 10 g/l). When required, media was supplemented with the antibiotics indicated, ampicillin (Amp) 100 μg/ml, chloramphenicol (Cm) 15 μg/ml or kanamycin (Km) 50 μg/ml. Bacterial cultures were inoculated at an OD_600 nm of 0.001, cultures reaching OD_600 nm 0.4 were considered logarithmic growth and reaching OD_600 nm 2.0 was considered early stationary phase. The bacterial strains and plasmids used in this study are listed in Supplementary Tables S1, S2, respectively.

Deletion strains were generated by standard gene replacement as previously described (Datsenko and Wanner, 2000). For chromosomal csrC-lacZ transcriptional fusion. The csrC:Cm strain was cured from the antibiotic resistance as previously described (Cherepanov and Wackernagel, 1995). The resulting csrC:frt strain carrying the FRT scar was further used to obtain chromosomal csrC-lacZ by integration of pKG136 as previously described (Ellermeier et al., 2002). Chromosomal spf-lacZ was obtained as for CsrC (El Mouali et al., 2018). Oligonucleotides used for strains construction are listed in Supplementary Table S3.

Plasmidic transcriptional fusions were generated for csrB and csrC. The regulatory regions of interest were PCR amplified, BamHI/HindIII digested and ligated within pQF50 (Farinha and Kropinski, 1989). Spot 42 was cloned in pBRplac (Guillier and Gottesman, 2006; Beisel and Storz, 2011). Spot 42 was PCR amplified, AatII/EcoRI digested and ligated in pBRplac. Spot 42 was expressed constitutively in Salmonella (El Mouali et al., 2018). Oligonucleotides used for cloning are listed in Supplementary Table S3.

Strains of interest were grown to logarithmic growth phase (OD_600 nm 0.4) or early stationary phase (OD_600 nm 2.0). β-galactosidase activity was measured as described previously (Miller, 1992). Shortly, 100 μl of culture was added to 900 μl of buffer Z (60 mM Na₂HPO₄, 40 mM NaH₂PO₄, 10 mM KCl, 1 mM MgSO₄, and 50 mM 2-mercaptoethanol, pH 7) and cells were lysed by the addition of 10 μl of toluene. Reactions were incubated at 28°C and β-galactosidase activity was measured upon addition of 200 μl of ONPG (4 mg/ml). Reactions were stopped by the addition of 500 μl of Na₂CO₃ (1 M). The OD_420 nm and OD_550 nm was measured and Miller Units were calculated as previously described (Miller, 1992). β-galactosidase activity determination was performed in technical duplicates for each of three biological replicates.

Strains of interest were grown to logarithmic growth phase (OD_600 nm 0.4). The biomass of 4 units of OD_600 nm was collected, and total RNA extracted by classic hot phenol method. Shortly, the cells were resuspended in 600 μl of TE (10 mM Tris-HCl, 1 mM EDTA, pH 8.0) containing 0.5 mg/ml lysozyme. Then, 60 μl of 10% SDS (w/v) was added; mixed by inversion and incubated at 64°C for 1–2 min. After the incubation, 66 μl of sodium acetate pH 5.2 (1 M) was added and mixed by inversion. For RNA extraction, 750 μl of Roti-Aqua phenol was added to the samples, mixed by inversion and incubated at 64°C for 6 min. Upon centrifugation (15 min, 13,000 rpm, 4°C), the top aqueous layer was transferred to a fresh 2 ml microcentrifuge tube, and 750 μl of chloroform was added. The samples were mixed and upon centrifugation (12 min, 13,000 rpm, 15°C), the upper aqueous layer was transferred into a new tube and precipitated with 30:1 mix of ethanol and sodium acetate (1 M, pH 6.5). The samples were incubated for 2–3 h or overnight at −20°C. The samples were then centrifuged and precipitated RNA was resuspended in water and concentration measured by NanoDrop. Samples of 10 μg of total RNA were subjected to electrophoretic separation in Tris-Borate-EDTA (TBE) 8% acrylamide gels containing 8.3 M urea. RNAs were transferred to Hybond N + (GE Healthcare) filters by semi-dry TBE based transference. Transcripts of interest were detected by hybridization with 5′ radiolabeled oligos as probes. Images were obtained with the FLA-5100 imaging system (Fujifilm). Oligonucleotides used as probes are listed in Supplementary Table S3.

Graph Pad 8.0 software was used for data analysis. Unpaired two-tailed Student’s t-test were carried out for two groups comparison and p < 0.05 were considered significant.