Strains of S. enterica used in this study (Table 1) belong to serovar Typhimurium and derive from a His⁺ derivative of SL1344 (Hoiseth and Stocker, 1981; Cano et al., 2002). Escherichia coli RK4353 [Δ(argF-lac)169 Δ(fimB-fimE)632(::IS1) Δ(fruK-yeiR) F⁻ λ⁻ [araD139] deoC1 flhD5301 gyrA219 rbsR22 relA1 rpsL150] was the host of pBAD18 (Guzman et al., 1995). Escherichia coli DH5α [endA1 hsdR17 supE44 thi1 recA1 gyrA96 relA1 ΔlacU189 (Ф80 lacZΔM15)] was also used for cloning. E. coli CC118 λ pir [phoA20 thi-1 rspE rpoB argE(Am) recA1 (λ pir)] and E. coli S17-1 λ pir [recA pro hsdR RP4-2-Tc::Mu-Km::Tn7 (λ pir)] were used for directed construction of point mutations. Plasmid pDMS197 is a suicide vector containing the sacB gene of Bacillus spp. (Edwards et al., 1998). Transduction was performed with S. enterica phage P22 HT 105/1 int201 (Schmieger, 1972). The P22 HT transduction protocol was described elsewhere (Torreblanca et al., 1999). To obtain phage-free isolates, transductants were purified by streaking on green plates (Garzon et al., 1995). Phage sensitivity was tested by cross-streaking with P22 H5.

Targeted disruption of yiiU (zapB) and ygfE (zapA) was achieved using the pKD13 plasmid (Datsenko and Wanner, 2000) and oligonucleotides listed in Supplementary Table S1. The antibiotic resistance cassette introduced during strain construction was excised by recombination with pCP20 (Datsenko and Wanner, 2000). For construction of the lacZ fusions in the Salmonella chromosome, FRT sites generated by excision of the Km^R cassette were used to integrate plasmid pCE40 (Ellermeier et al., 2002). Addition of 3xFLAG tag to the protein-coding zapB DNA sequence was carried out using the pSUB11 plasmid as template (Uzzau et al., 2001) and the oligonucleotides zapB-FLAG-FOR and zapB-FLAG-REV. Introduction of the lon::Tn10 allele into SL1344 was achieved by P22 HT-mediated transduction from the ATCC 14028 derivative SV5663 (Garcia-Calderon et al., 2009). The P_LtetO-zapB construct was engineered by inserting the P_LtetO promoter from the pXG1 plasmid (Lutz and Bujard, 1997) upstream of the transcriptional start site of zapB::lacZ on the Salmonella chromosome using primers zapB-PLtetO-FOR and zapB-PLtetO-REV. Details on the procedure have been described elsewhere (Lopez-Garrido and Casadesus, 2012).

To clone the zapB promoter with its 5' untranslated region (UTR) upstream of the lacZ gene of pIC552 (Macian et al., 1994), the primers pIC-zapB-FOR and pIC-zapB-REV were used. Plasmids carrying the S. enterica and E. coli zapA and zapB genes were constructed by cloning the respective genes under the control of the arabinose-dependent p_BAD promoter of pBAD18 (Guzman et al., 1995). Plasmids pBR328-CyaR, pBR328-InvR, pBR328-MicA, pBR328-RyhB-1, and pBR328-RyhB-2 are pBR328 derivatives.

Lysogeny broth (LB) and M9 minimal medium (Sambrook and Russell, 2001) were used as liquid media. Solid LB contained agar at 1.5% final concentration. To prepare LB containing sodium DOC (Sigma-Aldrich), an appropriate volume from a 25% stock was added. Liquid cultures were grown with aeration by shaking in an orbital incubator. Antibiotics were used at concentrations described previously (Garcia-Calderon et al., 2005). Arabinose was used at a final concentration of 0.2%.

Subcellular fractionation was performed as described elsewhere (Pucciarelli et al., 2002), with some modifications. Briefly, bacteria were grown in LB medium at 37°C, spun down by centrifugation at 15,000 × g for 5 min at 4°C and resuspended twice in cold phosphate-buffered saline (PBS, pH 7.4). The bacterial suspension was disrupted by sonication, and unbroken cells were further removed by low-speed centrifugation (5,000 × g, 5 min, 4°C). The supernatant was centrifuged at high speed (100,000 × g, 30 min, 4°C) and the new supernatant was recovered as the cytosol fraction. The pellet containing envelope material was suspended in PBS with 0.4% Triton X-100 and incubated for 2 h at 4°C. The sample was centrifuged again (100,000 × g, 30 min, 4°C) and divided into the supernatant containing mostly inner membrane proteins and the insoluble fraction corresponding to the outer membrane fraction. An appropriate volume of Laemmli buffer was added to each fraction. After heating (100°C, 5 min) and clearing by centrifugation (15,000 × g, 5 min, room temperature), the protein content was analyzed by SDS-PAGE.

Aliquots from exponential cultures in LB, each containing around 3 × 10² colony-forming-units (CFUs) were transferred to polypropylene microtiter plates (Soria Genlab, Valdemoro, Spain) containing known amounts of DOC. After 12 h incubation at 37°C, growth was visually monitored.

Levels of β-galactosidase activity were determined using the CHCl₃-sodium dodecyl sulfate permeabilization procedure (Miller, 1972).

For quantitative reverse transcriptase-PCR (qRT-PCR), Salmonella RNA was extracted from late exponential cultures (OD₆₀₀ ≅ 1). Gene-specific primers (Supplementary Table S1) were designed with ProbeFinder software¹ from Roche Applied Science. The efficiency of each primer pair was determined to be higher than 90%, following the procedure described in the real-time quantitative PCR (qPCR) guide of Applied Biosystems. Relative RNA levels were determined using the ∆∆Ct method. Each ∆∆Ct determination was performed at least in three different RNA samples.

Total protein extracts were prepared from cultures grown until stationary phase (OD₆₀₀ > 1) at 37°C in LB or LB + 5% DOC. Around 3 × 10⁹ bacterial cells were collected by centrifugation (16,000 × g, 2 min), resuspended in 50 μl of Laemmli sample buffer [1.3% SDS, 10% (v/v) glycerol, 50 mM Tris-HCl, 1.8% β-mercaptoethanol, 0.02% bromophenol blue, pH 6.8] and boiled (100°C, 5 min). Before loading, the samples were cleared by centrifugation (15,000 × g, 5 min). For analysis of protein stability, cultures were grown in LB at 37°C with shaking to a OD₆₀₀ of 0.6–0.8. Chloramphenicol (40 μg ml⁻¹) was added, and the culture was divided into two types: the DOC sample, to which 25% of DOC was added to reach a final concentration of 5%, and the LB sample, to which LB was added to equalize the volume with the DOC sample. A 1 ml aliquot was removed as zero-time control, and the samples were further incubated at 37°C without shaking. One-milliliter samples were extracted after 10, 20, 30, 45, 60, 75, and 90 min and treated as indicated above. Proteins were resolved by Tris-Tricine-PAGE (12%) electrophoresis and transferred onto polyvinylidene-difluoride membranes using a semidry electrophoresis transfer apparatus (Bio-Rad). Specific proteins were detected with anti-Flag M2 monoclonal antibody (1:5,000, Sigma-Aldrich) and anti-GroEL polyclonal antibody (1:20,000; Sigma-Aldrich). Goat anti-mouse horseradish peroxidase-conjugated antibody (1:5,000; Bio-Rad) or goat anti-rabbit horseradish peroxidase-conjugated antibody (1:20,000; Santa Cruz Biotechnology) were used as secondary antibodies. Proteins recognized by the antibodies were visualized by chemoluminiscence using the luciferin-luminol reagents (Thermo Scientific) in a LAS 3000 Mini Imaging System (Fujifilm). For quantification, the intensity of the hybridization bands was determined using MultiGauge software (Fujifilm). GroEL was used as loading control.

Aliquots from exponential cultures in LB and LB + 5% DOC were centrifuged at 16,000 × g, 37°C, during 2 min, and washed twice with 500 μl of NaCl 0.9%. The pellets were resuspended in 100 μl of a solution of lysozyme (Sigma-Aldrich), 3 mg ml⁻¹. Cell lysis was facilitated by freezing at −20°C for >2 h. After lysis, RNA was extracted using 1 ml of TRIsure reagent (Bioline). Lastly, total RNA was resuspended in 25 μl of RNase-free water. The quality of the preparation and the RNA concentration were determined using a ND-1000 spectrophotometer (NanoDrop Technologies).

For Northern blot analysis, 10 μg of total RNA was loaded per well and electrophoresis was performed in denaturing 1% agarose-formaldehyde gels. Vacuum transfer and fixation to Hybond-N⁺ membranes (GE Healthcare) were performed using 0.05 M NaOH. UV crosslinking was used to immobilize RNAs on the membrane. For mRNA stability experiments, rifampicin (500 mg ml⁻¹) was added to exponential cultures grown in LB at 37°C (zero-time control). At this point, the culture was divided into two types: the DOC sample, to which 25% of DOC was added to reach a final concentration of 5%, and the LB sample, to which LB was added to equalize the volume with the DOC sample. Incubation was continued, and culture aliquots were taken at appropriate times. DNA oligonucleotides (Racer2-zapB for zapB mRNA and 5S-probe for the loading control) were labeled with [³²P]-γ-ATP using T4 polynucleotide kinase (Biolabs). Labeled probes were purified over microspin G-25 columns (GE Healthcare) to remove unincorporated nucleotides. As a control of RNA loading and transfer efficiency, the filters were hybridized with an oligoprobe for the 5S rRNA. Membranes were hybridized with oligoprobes at 42°C. Signals were visualized with a FLA-5100 image system (Fujifilm), and quantification was performed using MultiGauge software (Fujifilm).

RNA was isolated from S. enterica stationary cells (DO₆₀₀ ~2) grown in LB. Fifteen microgram of RNA was used to determine cDNA ends using a standard protocol. RNAs were prepared either with or without tobacco acid pyrophosphatase (TAP) to distinguish primary transcript 5' ends from internal 5' processing sites. The sequences of the primers are listed in Supplementary Table S1. The DNA primer for cDNA synthesis, Racer2-zapB, is complementary to zapB RNA. A second DNA primer for subsequent PCR amplification of cDNAs, Racer-out, is homologous to the adaptor RNA primer Racer-RNA-adaptor used for 5' rapid amplification of cDNA ends (5'-RACE). PCR products detected both with and without TAP treatment were further amplified with the internal primers Racer-nested and Racer1-zapB, and purified. PCR products were cloned on pGEM-T Easy (Promega), and 29 clones from two independent assays were sequenced.

Introduction of point mutations into MicA sRNA was performed by PCR using as template the wild type micA gene cloned on pBR328. Two pairs of primers were employed: MicA5101319-FOR and micAREVSalI; and MicA5101319-REV and micAFORBamHI. MicA5101319-FOR and MicA5101319-REV are complementary primers that introduce the point mutations in the gene, while micAREVSalI and micAFORBamHI introduce ad hoc restriction sites. The resulting DNA fragments were ligated into one using Gibson assembly (New England Biolabs, Ipswich, MA, United States). The fragment was digested with SalI and BamHI and cloned onto pBR328. The plasmid containing substitutions in MicA sRNA (5, G→C; 8, G→C; 13, U→A; and 19, A→U) was transformed into the zapB + 18::lacZ background to yield strain SV9486 (zapB + 18::lacZ pBR328-MicA*).

To insert point mutations into the chromosomal zapB gene, PCR was performed as described for MicA, using the pairs of primers For-zapB(MicA)-50,545,661 and zapB-XbaI-REV; and Rev-zapB(MicA)-50,545,661 and zapB-SacI-FOR. As above, ligation by Gibson assembly was followed by digestion and cloning onto pDMS197. The strains carrying pMDS197 derivatives were used as donors in matings with the S. enterica strain SV9265 (which harbors a Km^R cassette replacing the zapB coding sequence) as recipient. Tc^R transconjugants were selected on M9 plates supplemented with tetracycline. Several Tc^R transconjugants were grown in nutrient broth containing 5% sucrose. Individual tetracycline-sensitive segregants were then screened for kanamycin sensitivity and examined for the incorporation of the mutant zapB allele by DNA sequencing using external oligonucleotides (zapB-E1 and zapB-E2). The resulting mutant (SV9647, zapB*50A54T58A61G) contained four substitutions (50: U→A, 54: A→U, 58: U→A, and 61, C→G) that restored complementation with mutant MicA.

DNA templates for in vitro transcription of zapB mRNA and MicA sRNA were generated by PCR using chromosomal DNA from both the wild type and a zapB mutant (SV9647), and plasmid DNA from strains containing wild type and mutant variants of sRNA MicA cloned onto pBR328. The primers used were Transc-zapB-FOR and Transc-zapB-REV, and Transc-micA-FOR and Transc-micA-REV (Supplementary Table S1). The PCR products were purified with the Wizard® SV Clean-Up System (Promega). For the in vitro transcription reaction, the HiScribe T7 High Yield RNA Synthesis kit was employed. Two sets of reactions were prepared: (i) for synthesis of non-labeled zapB transcripts, the zapB PCR product was mixed with equimolar concentrations of all four ribonucleotides (10 mM) and (ii) for synthesis of internally labeled MicA sRNA and zapB mRNA, each PCR product was mixed with ATP, GTP, and CTP 10 mM and 0.4 mM of UTP. Radioactive labelling was performed by adding 3 μl of EasyTide® uridine 5'-triphosphate (α-³²P) 3,000 Ci mmol⁻¹ (PerkinElmer) to a final concentration of 0.25 μM. One microliter of T7 RNA Polymerase Mix and 4 μl of Master Mix were added to each reaction tube. Nuclease-free water was added to a final volume of 40 μl. The reactions were then incubated at 37°C for 2 h and treated with DNase (TURBO DNase Ambion 2 U μl⁻¹) for 15–20 min. To discard unincorporated nucleotides, the reactions were passed through GE Healthcare ilustra™ MicroSpin™ G-25 columns. Subsequently, the RNA transcripts (both labeled and unlabelled) were purified by electrophoresis on a 7.9 M urea, 8% polyacrylamide denaturing gel. The hybridization fragments were visualized using flour-coated TLC plates (Ambion). Gel slices were crushed, and RNA was eluted overnight at 4°C with crush-soak solution (200 mM NaCl, 10 mM Tris-HCl, and 10 mM EDTA). The RNA was ethanol-precipitated and resuspended in nuclease-free water.

Gel mobility shift assays were performed with 0.015 pmol of [α-³²P]-labeled wild type or mutant MicA (MicA*) RNA, and 0.075 pmol of unlabelled zapB mRNA transcript in 1× binding buffer (200 mM Tris-HCl pH 8, 10 mM DTT, 10 mM MgCl₂, 200 mM KCl, and 100 mM Na₂HPO₄-NaH₂PO₄) at 37°C. The reactions were incubated, and samples were taken at 10, 30, 60, and 100 min. The binding reactions were mixed with 2 μl of loading dye (48% glycerol and 0.01% bromophenol blue) and subjected to electrophoresis at room temperature in a 6% non-denaturing polyacrylamide Tris-borate EDTA (TBE) gel for 2 h at 55 V, 15 mA in 1x TBE buffer (90 mM Tris-borate/2 mM EDTA). In parallel to the binding reactions, a positive control was prepared by mixing zapB mRNA and [α-³²P]-labeled MicA (mutant and wild type) followed by boiling denaturation and renaturation at 37°C. Single [α-³²P]-labeled zapB RNA and MicA were also loaded on the gel. After electrophoresis, gels were dried and analyzed in a FLA-5100 Scanner (Fujifilm).

Student’s t-test (unpaired) was performed using GraphPad Prism version 8.0 for Mac to determine the statistical differences between two groups. Ordinary one-way ANOVA was also performed using GraphPad Prism.

The zapB gene of E. coli encodes an 81-amino-acid protein that acts as a non-essential cell division factor involved in Z-ring assembly (Adams and Errington, 2009). The name zapB is for “Z ring-associated protein B” (Ebersbach et al., 2008). Alignment of the amino acid sequences of the ZapB protein of E. coli and the predicted protein encoded by the yiiU locus of Salmonella enterica reveals a 91% identity value. For this reason, the S. enterica yiiU locus will be henceforth named zapB.

Minimal inhibitory concentration (MIC) analysis was performed to determine whether the S. enterica ZapB protein is necessary for bile resistance. A null ZapB⁻ mutant (SV6269) was found to be sensitive to DOC (Table 2). The conclusion that ZapB is necessary for bile resistance was confirmed by complementation analysis with plasmid-borne E. coli and S. enterica zapB genes expressed from the arabinose-inducible promoter of pBAD18 (Table 2).

The observation that a ZapB⁻ mutant is sensitive to bile raised the possibility that lack of another Z-ring assembly modulator, ZapA, might also cause bile sensitivity. A bioinformatic search followed by Clustal W alignment revealed that the ygfE gene of S. enterica encodes a protein with 88% amino acid identity with E. coli ZapA. YgfE was thus renamed ZapA. Disruption of the zapA gene caused sensitivity to DOC, and complementation tests using plasmid-borne E. coli and S. enterica zapA genes restored bile resistance (Table 2).

The MIC of DOC for a zapA zapB mutant was similar or identical to that of a zapB mutant, an epistatic effect which suggests that bile sensitivity in the absence of either ZapA or ZapB results from disruption of the same cellular process.

The levels of zapA and zapB mRNAs in LB and LB + DOC were monitored by qPCR. Relative RNA levels, averaged from six independent experiments, were as follows: zapA, 1.00 (arbitrary) in LB and 1.21 ± 0.36 in LB + DOC; and zapB, 2.09 ± 0.49 in LB and 8.80 ± 1.78 in LB + DOC. These experiments confirmed previous observations indicating that expression of zapB (but not of zapA) increases in the presence of bile (Hernandez et al., 2012). Further work was centered on zapB only.

The stability of the ZapB protein in the presence and in the absence of DOC was analyzed by Western blot analysis using a ZapB version tagged with the 3 × FLAG epitope (ZapB::3 × FLAG). Note that the strain carrying this ZapB variant is resistant to DOC and the protein may be, therefore, considered wild type (Table 2). The amount of ZapB protein detected in LB + DOC was lower than in LB (Figure 1A), suggesting that ZapB might be degraded in the presence of DOC. To investigate this possibility, we carried out protein stability assays to monitor the half-life of the ZapB::3 × FLAG protein upon addition of DOC. As shown in Figures 1B,C, ZapB is degraded faster in LB + DOC (half-life around 30 min) than in LB (half-life >90 min).

Because the Lon protease is involved in the degradation of unfolded and misfolded proteins (Tsilibaris et al., 2006; Gur et al., 2011) and bile salts cause protein misfolding and denaturation (Bernstein et al., 1999; Leverrier et al., 2003), the stability of ZapB protein was monitored in a Lon⁻ background (strain SV7017). In the absence of Lon protease, the ZapB stability in LB + DOC increased (Figures 1B,C). Hence, we tentatively conclude that the Lon protease may degrade ZapB in the presence of DOC.

To investigate the mechanism underlying zapB upregulation in the presence of DOC, the transcriptional start site of the zapB gene was identified by 5'-RACE. Twenty-four out of twenty-nine clones analyzed from two independent 5'-RACE assays identified a 5'-end A residue located 56 nucleotides upstream of the zapB translational start codon. In silico analysis of the region revealed the occurrence of nucleotide sequences compatible with the consensus sequences of −35 and −10 modules in σ⁷⁰-dependent promoters, and with appropriate spacing (Collado-Vides, 1996). The proposed structure of the zapB promoter is presented in Figure 2A.

To confirm the existence of the predicted zapB promoter, the putative promoter region was cloned on the promoter-probe vector pIC552 (Macian et al., 1994) to generate a transcriptional fusion with the lacZ gene. The pIC552 derivative bearing the zapB promoter was introduced into the wild type strain, and β-galactosidase activity measurements confirmed that the cloned region was able to drive lacZ expression. Interestingly, transcription of the pIC552-borne p_zapB-lacZ gene was not regulated by DOC (Figure 2B). This observation raised the possibility that upregulation of zapB expression in the presence of DOC might be postranscriptional.

Upregulation of zapB mRNA in the presence of bile was confirmed by β-galactosidase assays using a translational zapB::lacZ fusion and by Northern blot analysis (Figures 3A,B). When the translational zapB::lacZ fusion was placed under the control of a heterologous promoter, pL_tetO (Lutz and Bujard, 1997), increased zapB expression was still observed in the presence of DOC (Figure 3C). This observation was in agreement with the above observation that the zapB promoter cloned on pIC552 was not responsive to DOC (Figure 2B), leading us to conclude that DOC-dependent regulation of zapB is not transcriptional.

Evidence for postranscriptional upregulation of zapB in the presence of bile prompted a comparison of the stability of the zapB transcript in LB and in LB + DOC. For this purpose, Northern blot analysis was performed. Decay of zapB mRNA was faster in LB than in LB + DOC (Figure 4), suggesting that the higher amount of zapB mRNA found in the presence of DOC might result from increased mRNA stability.

Among the mechanisms that control bacterial mRNA stability, interaction with small regulatory RNAs (sRNAs) is common (Storz et al., 2004, 2011; Caron et al., 2010; Podkaminski and Vogel, 2010; Prevost et al., 2011). Hence, we considered the possibility that the stability of zapB mRNA might be controlled by a sRNA, either destabilizing the transcript in LB or stabilizing the transcript in the presence of DOC. Because most bacterial small regulatory RNAs require the Hfq chaperone for interaction with the target and for stability of the sRNA itself (Valentin-Hansen et al., 2004; McCullen et al., 2010; Soper et al., 2011), we analyzed the stability of zapB mRNA in an Hfq⁻ strain using Northern blotting (Figure 4). In the Hfq⁻ background, the stability of the zapB transcript with and without DOC was similar to the stability of zapB mRNA in the wild type strain grown in LB, suggesting that stabilization of zapB mRNA in the presence of DOC requires the Hfq RNA chaperone (Figures 4A,B).

To identify the hypothetical zapB mRNA region that might be subjected to postranscriptional control, we constructed a set of translational fusions inserting the lacZ gene at different locations within the zapB coding sequence: 9, 18, 27, 30, and 66 nucleotides after the start codon (Supplementary Figure S1). Because 3'UTRs are often targets for sRNA binding (Opdyke et al., 2004; Papenfort and Vanderpool, 2015), we also constructed a strain in which the 3'UTR of zapB was deleted downstream of the zapB::lacZ fusion (Supplementary Figure S1). Measurement of β-galactosidase activities upon growth in LB and in LB + DOC provided the following observations:

Altogether, the above observations provide evidence that the 5'UTR and the upstream (5') region of the zapB coding sequence (CDS) may be the main (perhaps the only) region of zapB mRNA involved in postranscriptional control by DOC.

We did not rule out the possibility that Hfq binding might be sufficient to stabilize zapB mRNA. However, this possibility seemed unlikely as Hfq often acts in concert with sRNAs (Basineni et al., 2009; Chao and Vogel, 2010; Soper et al., 2010; Storz et al., 2011). A computational search for potential interactions between Salmonella sRNA candidates and zapB mRNA using the IntaRNA program (Smith et al., 2010) identified sRNAs with potential targets in zapB mRNA (Supplementary Table S2). Such sRNA candidates were shortlisted using two criteria: (i) the potentially interacting mRNA sequence should be located in the 5' region of the zapB transcript and (ii) the potentially interacting sRNA should be regulated by bile. These criteria were met by five candidates: MicA, InvR, CyaR, RyhB-1, and RyhB-2, previously reported to be upregulated in the presence of bile (Kroger et al., 2013).

MicA, InvR, CyaR, RyhB-1, and RyhB-2 sRNAs were cloned onto pBR328, and the resulting plasmids were transformed into a Salmonella strain harboring a zapB::lacZ fusion. β-Galactosidase activities in LB and in LB + DOC were then determined (Figure 5B). The presence of DOC resulted in higher β-galactosidase activities when the sRNAs MicA and InvR were overexpressed, and the MicA effect was especially strong. The other sRNAs tested in the assay did not show any effect. These results suggest that MicA may be involved in stabilization of zapB mRNA in the presence of DOC.

In silico analysis with the IntaRNA tool (Smith et al., 2010) predicts that MicA might interact with zapB mRNA by base pairing at a zapB mRNA region that contains a putative ribosome binding site (5'GAGG3') and the AUG start codon of the coding sequence (Figure 6A). To probe the predicted interaction, point mutations were introduced in MicA sRNA by site-directed mutagenesis: 5, G→C; 8, G→C; 13, U→A; and 19, A→U. The mutant version of micA (micA*) was then cloned on pBR328 and transformed into a strain carrying a zapB::lacZ fusion. Expression of zapB::lacZ in LB and in LB + DOC was compared with that of a strain that overexpressed wild type MicA. β-Galactosidase activity analysis indicated that the upregulation of the zapB transcript detected upon MicA overexpression disappeared when MicA was mutated (Figure 6B). Actually, the β-galactosidase activity in the presence of MicA* was similar to that of the control harboring the empty vector (Figure 6B). These observations suggest that the mutations introduced into MicA had disrupted the interaction with zapB mRNA, thus supporting the existence of a zapB mRNA:MicA sRNA interaction at the region predicted by bioinformatic analysis.

Further evidence for zapB mRNA:MicA sRNA interaction was obtained in vitro. An electrophoretic mobility shift assay (EMSA) was carried out using wild type and mutant versions of both zapB mRNA and MicA sRNA, the latter labeled with [α-³²P] (Figure 6C). The mutations introduced in zapB*, the mutant version of zapB mRNA (50: U→A, 54: A→U, 58: U→A, and 61, C→G) were complementary to those present in MicA*, so that the interaction predicted by IntaRNA could be restored. Binding was detected for both zapB mRNA:MicA and zapB* mRNA:MicA*, and the fact that binding was largely independent from the incubation time may suggest a strong interaction (Figure 6C). Altogether, the experiments shown in Figures 5, 6 support the following tentative conclusions: (i) zapB mRNA and MicA sRNA do interact; (ii) the main (perhaps only) interaction involves the upstream region of zapB mRNA; and (iii) the zapB:MicA interaction increases zapB expression in the presence of DOC, probably by MicA-mediated stabilization of zapB mRNA.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.