OprD Repression upon Metal Treatment Requires the RNA Chaperone Hfq in Pseudomonas aeruginosa

The metal-specific CzcRS two-component system in Pseudomonas aeruginosa is involved in the repression of the OprD porin, causing in turn carbapenem antibiotic resistance in the presence of high zinc concentration. It has also been shown that CzcR is able to directly regulate the expression of multiple genes including virulence factors. CzcR is therefore an important regulator connecting (i) metal response, (ii) pathogenicity and (iii) antibiotic resistance in P. aeruginosa. Recent data have suggested that other regulators could negatively control oprD expression in the presence of zinc. Here we show that the RNA chaperone Hfq is a key factor acting independently of CzcR for the repression of oprD upon Zn treatment. Additionally, we found that an Hfq-dependent mechanism is necessary for the localization of CzcR to the oprD promoter, mediating oprD transcriptional repression. Furthermore, in the presence of Cu, CopR, the transcriptional regulator of the CopRS two-component system also requires Hfq for oprD repression. Altogether, these results suggest important roles for this RNA chaperone in the context of environment-sensing and antibiotic resistance in P. aeruginosa.


Introduction
Pseudomonas aeruginosa is an opportunistic pathogen that causes serious and diverse infections in host organisms by producing a broad range of virulence factors [1]. This bacterium carries intrinsic resistances to multiple classes of antimicrobial compounds, representing a major challenge for the treatment of Pseudomonas' infections [2]. For instance, resistance to carbapenem, an important class of anti-Pseudomonas compounds, is mostly caused by the decrease in production of OprD porin. In normal conditions, OprD forms a trimeric outer-membrane channel [3] which is generally involved in the import of basic amino acids and small peptides from the outer medium [4]. However, carbepenem molecules are also imported through this porin and, consequently, a reduced production of OprD causes the insurgence of bacterial resistance [5][6][7].
We have previously found that the mechanism that triggers the negative regulation of OprD is linked to Zn and Cd metal resistance. According to [8], this mechanism is a process called co-regulation between metal and antibiotic resistance. The presence of an excess of these elements activates the metal-inducible CzcRS two-component system (TCS) that induces the expression of a metal efflux pump. Furthermore, it down-regulates the production of the OprD porin, thus rendering cells resistant to both trace metals and carbapenems. Cu has also been shown to induce expression of the copRS TCS which can directly repress oprD transcription [5]. Therefore, toxic metal concentrations of Zn, Cd, or Cu may all lead to the induction of carbapenem resistance. In addition to OprD, the CzcR regulator has been shown to modulate gene expression of multiple virulence factors in response to Zn treatment,

Bacterial Strains, Growth Conditions and Minimum Inhibitory Concentration (MIC) Determination
The bacterial strains used in this study are listed in Table 1. E. coli and P. aeruginosa strains were grown at 37 • C in solid or liquid Luria-Bertani (LB) medium (US biological, Salem, MA, USA). Metal-containing LB media were supplemented with CuCl 2 or ZnCl 2 at the final concentration indicated in the figure legends. Antibiotics were used at the following concentrations (in µg per mL): gentamicin 50, tetracycline 50, carbenicillin 200, streptomycin 200 for P. aeruginosa and gentamicin 15, tetracycline 15, ampicillin 100 for E. coli.

DNA Manipulations
Restriction enzymes, PCR amplifications and cloning were performed using standard methods [31] following the manufacturers instructions. Plasmids were introduced into E. coli by heat-shock [31] and into P. aeruginosa by electroporation [32]. Plasmid inserts and genomic regions flanking the deletions were verified by sequencing.
hfq complementation: The full P. aeruginosa hfq gene and its promoter region were amplified by PCR using primers 636 and 610 (Appendix Table A1). The PCR product was digested with EcoRI and BamHI enzymes and ligated into the corresponding site of the pME4510 vector. The resulting plasmid containing the 6× His tag in the C-terminal part of the hfq gene was transformed into the P. aeruginosa WT strain and the hfq − mutant by electroporation [32].
Double mutant construction: The ∆czcRS;hfq − double mutant was constructed by homologous recombination. Briefly, the ∆czcRS strain was conjugated with the p202Sp plasmid containing the inactivated hfq gene, according to the initial protocol [26]. Clones were selected for streptomycin resistance and tetracycline sensitivity. Deletion was confirmed by PCR on genomic DNA using primers 314 and 315 (Appendix Table A1).

Western Blot Analysis
Overnight cultures were diluted to an OD 600 of 0.05 for Wild Type (WT) and ∆czcRS or OD 600 of 0.1 (absence of Zn) or 0.2 (presence of Zn) for hfq − and ∆czcRS;hfq − strains and grown for 6 h. Metal concentrations, when added, are 0.5 mM ZnCl 2 or 2.5 mM CuCl 2 . When necessary, IPTG (isopropyl-1-thio-D-galactopyranoside) was added at a final concentration of 0.1 mM. Cultures were grown for 6 h and 1 mL of culture was pelleted in a microfuge. Total proteins were solubilized to a concentration of 2 mg·mL −1 by sonication in the appropriate volume of 1× β-mercaptoethanol gel-loading buffer (an OD 600 of 1 gives 0.175 mg·mL −1 of protein). Samples were boiled for 5 min prior to loading. 15 µL (30 µg) of total protein was separated on a 4%-15% precast gel (Mini-PROTEAN TGX Gels, Biorad, Hercules, CA, USA) and transferred to a nitrocellulose membrane. Blots were incubated with anti-OprD, anti-CzcR and anti-Hsp70 antibodies as previously described [9] and revealed by chemiluminescence.

RNA Extraction and Reverse Transcription
Overnight cultures were diluted as described in Section 2.3. and grown for 6 h. RNA Protect bacteria solution (1 mL, Qiagen, Hilden, Germany) was added to 0.5 mL of culture. Cells were harvested by centrifugation and total RNA was extracted using RNeasy columns (Qiagen) according to the manufacturer's instructions. RNA (3 µg) was treated for 2 h with RQ1 DNase (Promega, Fitchburg, MA, USA), in order to remove any residual DNA, followed by phenol-chloroform extraction and ethanol precipitation. The RNA was then resuspended in 30 µL RNAse-free water. For cDNA synthesis, 500 ng of total RNA was reverse-transcribed using random hexamer primers (Promega) and Improm-II reverse transcriptase (Promega) according to the supplier's instructions. The reverse transcriptase was then heat-inactivated and the cDNAs obtained were diluted tenfold in water.

Quantitative RT-PCR
qPCR procedures were performed in triplicate starting from three independent experiments, using SYBR Green mix (Power SYBR Green PCR Master Mix, Thermo Fisher Scientific, Waltham, MA, USA), according to the manufacturer's instructions. Primers used for qRT-PCR are described in Appendix Table A1. Results were analyzed according to [33]. The well expressed porin oprF was used for normalization since its expression was affected neither by Zn nor by hfq deletion.

Semi-Quantitative RT-PCR
PCR amplifications were performed using standard procedures with 27 cycles, except for the RT-negative control (without reverse-transcriptase) for which the amplification was carried out with 30 cycles, using hsp70 primer pairs. The primers used are listed in Appendix Table A1. Each analysis was performed at least three times from three independent cultures. A representative analysis is presented.

Chromosome Immunoprecipitation
ChIP (chromosome immunoprecipitation) experiments were performed as previously described [9]. Briefly, the WT and hfq − strains were grown for 6 h in 50 mL LB medium supplemented or not with 0.5 mM Zn. To fix the protein to DNA, the cultures were treated with 1.2% (final concentration) formaldehyde for 10 min. Glycine (330 mM) was then added to quench the reaction. Bacteria were lyzed by sonication and resuspended in 500 µL ice-cold FA-lysis buffer (50 mM HEPES-KOH pH 7.5, 140 mM NaCl, 1 mM EDTA pH 8, 1% Triton X-100, 0.1% Sodium deoxycholate) supplemented with lysozyme (5 mg·mL −1 ), AEBSF (4-(2-aminoethyl) benzenesulfonyl fluoride hydrochloride, 1 mM final) and SDS (0.5% final). The resuspended pellet was sonicated as previously described [34]. After sample centrifugation 200 µL of supernatant was used for immunoprecipitation. The immunoprecipitations were performed by addition of 800 µL FA-lysis buffer with 50 µL protein A sepharose (100 mg·mL −1 ) and 5 µL of CzcR anti-serum [10]. Immunoprecipitated DNA was quantified by real-time PCR using a SYBR Green mix (Power SYBR Green PCR Master Mix, Thermo Fisher Scientific, Waltham, MA, USA) according to the supplier's specifications. Results represent the amount of oprD promoter immunoprecipitated with CzcR in comparison to the input DNA. Each value represents the average of three independent experiments (standard deviations are indicated). The primers used are listed in Appendix Table A1.

A CzcR-Independent Mechanism Represses the Production of the OprD Porin in Presence of Zn
CzcR, the transcriptional regulator of the Zn and Cd activated two-component system CzcRS, is able to directly repress the transcription of the oprD gene in the presence of Zn [7,9]. In a previous study we have shown that another mechanism, independently of CzcRS, could also be involved in oprD repression upon Zn treatment [10]. We decided to further investigate this alternative mechanism and used the ∆czcRS mutant, deleted for the metal-specific TCS, to analyze the abundance of OprD porin at different Zn concentrations in LB medium when cells are in early stationary phase, i.e., after 6 h of growth ( Figure 1A). In the wild type P. aeruginosa PAO1 strain, at 0.05 mM Zn, we observed at 6 h a drop of OprD of approximately 60%, whereas no effect (only 3% reduction) was observed in the corresponding ∆czcRS mutant ( Figure 1A). At 0.2 mM Zn, OprD could no longer be detected by western blot in the WT. Surprisingly, at this concentration, we observed a 50% decrease in OprD protein in the ∆czcRS strain. By increasing Zn to 0.5 mM, a concentration that does not affect the growth of the mutant ( Figure 1B), we found that OprD is also strongly repressed (85%) in the ∆czcRS mutant ( Figure 1A). Although this effect is not as strong as the one observed in the WT, it does however suggest that another mechanism, independent of CzcRS, might account for OprD down-regulation in the presence high Zn concentrations. concentration that does not affect the growth of the mutant ( Figure 1B), we found that OprD is also strongly repressed (85%) in the czcRS mutant ( Figure 1A). Although this effect is not as strong as the one observed in the WT, it does however suggest that another mechanism, independent of CzcRS, might account for OprD down-regulation in the presence high Zn concentrations.

Hfq is Involved in the Repression of OprD in the Presence of Zn
The RNA chaperone Hfq is known to control the expression of several porins in Gram-negative bacteria, working mostly at the post-transcriptional level, on mRNA stability or translation [17,35]. We therefore decided to test the possible involvement of this protein in the repression of OprD upon Zn treatment. To this aim, an hfq mutant was created in the czcRS genetic background to avoid the CzcR-mediated transcriptional repression. A slight growth delay was observed in the hfq mutants (hfq − and czcRS;hfq − ) in the presence of metal. This reflects the growth rate defect already detected in the P. aeruginosa hfq mutant [26]. The inoculum was therefore adapted (see experimental Section 2.3.) in order to reach the same optical density after 6h of growth ( Figure 2A). As expected, the czcRS as indicated. Blots were exposed to anti-OprD and anti-Hsp70 (loading control) antibodies. OprD quantification was performed using ImageJ software (NIH, Bethesda, MD, USA). OprD band intensity was normalized to Hsp70 intensity and expressed as a percentage of the condition of the absence of Zn; (B) Growth curves of P. aeruginosa WT strain and the ∆czcRS mutant cultivated in LB medium in the absence or presence of 0.5 mM ZnCl 2 (+Zn).

Hfq is Involved in the Repression of OprD in the Presence of Zn
The RNA chaperone Hfq is known to control the expression of several porins in Gram-negative bacteria, working mostly at the post-transcriptional level, on mRNA stability or translation [17,35]. We therefore decided to test the possible involvement of this protein in the repression of OprD upon Zn treatment. To this aim, an hfq mutant was created in the ∆czcRS genetic background to avoid the CzcR-mediated transcriptional repression. A slight growth delay was observed in the hfq mutants (hfq − and ∆czcRS;hfq − ) in the presence of metal. This reflects the growth rate defect already detected in the P. aeruginosa hfq mutant [26]. The inoculum was therefore adapted (see experimental Section 2.3.) in order to reach the same optical density after 6h of growth ( Figure 2A). As expected, the ∆czcRS mutant displayed a repression of OprD in the presence of 0.5 mM Zn, whereas the effect on OprD level was only very weak (5%) in the ∆czcRS;hfq − double mutant ( Figure 2B). These observations suggested the possible involvement of the RNA chaperone Hfq in the control of OprD production under metal stress conditions. To investigate the contribution of Hfq in this process, we first determined the amount of oprD mRNA by quantitative RT-PCR. In the ∆czcRS mutant, addition of Zn induced a drop in oprD mRNA, while disruption of the hfq gene in this strain (∆czcRS;hfq − mutant) totally abolished the decrease in the amount of oprD mRNA ( Figure 2C). Zn treatment however did not affect the expression of hfq, as measured by semi-qRT-PCR (Appendix Figure A1). Altogether these results showed that, in addition to the transcriptional repression mediated by CzcR, Zn treatment induced an Hfq-dependent mechanism leading to oprD mRNA down-regulation. Zn induced a drop in oprD mRNA, while disruption of the hfq gene in this strain (czcRS;hfqmutant) totally abolished the decrease in the amount of oprD mRNA ( Figure 2C). Zn treatment however did not affect the expression of hfq, as measured by semi-qRT-PCR (Appendix Figure A1). Altogether these results showed that, in addition to the transcriptional repression mediated by CzcR, Zn treatment induced an Hfq-dependent mechanism leading to oprD mRNA down-regulation.

Hfq is also Involved in the Negative Regulation of OprD Mediated by CzcR
In the absence of CzcRS, we observed that Hfq might contribute to the negative regulation of OprD upon Zn challenge (Figure 2). To test whether Hfq contributes to CzcRS TCS function, we examined the amount of OprD porin in the WT strain and in the single hfq mutant. Surprisingly, while the CzcR protein was induced at the same level in both strains in presence of Zn, OprD Immunoblot analysis; total protein extract of the ∆czcRS and the ∆czcRS;hfq − mutants cultivated in the absence or presence of 0.5 mM ZnCl 2 (+Zn). Blots were exposed to anti-OprD or anti-Hsp70 (loading control) antibody. OprD quantification was performed using ImageJ software. OprD band intensity was normalized to Hsp70 intensity and expressed as a percentage of the condition of the absence of Zn; (C) Amount of oprD mRNA analyzed by quantitative RT-PCR in the ∆czcRS and the ∆czcRS;hfq − mutants cultivated in the absence or presence of 0.5 mM ZnCl 2 as indicated. The amount of mRNA is represented relative to the WT strain cultivated in the absence of metal. Statistics are indicated, using p values of <0.05. (*). Statistically, there is no difference between ∆czcRS (-Zn) and ∆czcRS;hfq − (-Zn) as well as between ∆czcRS;hfq − (+/-Zn). Experiments were performed in triplicate on three independent occasions. Error bars represent the standard deviations of three independent determinations.

Hfq is also Involved in the Negative Regulation of OprD Mediated by CzcR
In the absence of CzcRS, we observed that Hfq might contribute to the negative regulation of OprD upon Zn challenge (Figure 2). To test whether Hfq contributes to CzcRS TCS function, we examined the amount of OprD porin in the WT strain and in the single hfq mutant. Surprisingly, while the CzcR protein was induced at the same level in both strains in presence of Zn, OprD repression is clearly impaired in the absence of Hfq protein ( Figure 3A). This result suggested that either Hfq is the main OprD repressor in Zn conditions or CzcR-mediated regulation requires the RNA chaperone. Complementation of the hfq − strain with a plasmid containing a functional 6x His-tagged hfq gene led to a full restoration of OprD repression in the presence of Zn ( Figure 3B), indicating that the observed result was not due to a polar effect caused by hfq deletion. The regulation of OprD in the control WT strain containing either the pME (empty plasmid) or the hfq gene (phfq) was not affected (data not shown). Moreover, the fact that the CzcR protein is fully produced suggested that Zn recognition by the CzcS sensor protein and the phosphorelay leading to CzcR induction are not affected by the absence of Hfq.  We have previously noticed that artificial overexpression of the transcriptional regulator CzcR is able to repress oprD, even in the absence of the sensor protein CzcS and in the absence of Zn [7]. We wondered whether in this situation negative regulation also required the Hfq protein. To this aim, we used the pCzcR plasmid containing the czcR gene under the control of an IPTG-inducible promoter. In the absence of Zn overexpression of czcR yielded complete repression of OprD in a czcRS background, while only a minor effect was visible in the absence of Hfq (czcRS;hfq − background) ( Figure 3C). This indicates that full repression, mediated directly by CzcR, is dependent upon the Hfq protein even in the absence of Zn challenge.

Hfq is Necessary for Localization of CzcR to the oprD Promoter
In order to investigate the role of Hfq in CzcR-mediated transcriptional repression, we evaluated the DNA-binding efficiency of CzcR to the oprD promoter. To do so, we performed a ChIP (chromosome immunoprecipitation) experiment using anti-CzcR antibody. Interestingly, addition of Zn caused a clear localization of CzcR to the oprD promoter in the WT strain compared to the hfq mutant ( Figure 4A). The weak occupancy of oprD promoter by CzcR observed in the hfq − strain is not (B) the hfq − mutant complemented with the empty plasmid (pME) or with the hfq gene (phfq) cultured in the absence or presence of 0.5 mM ZnCl 2 as indicated; (C) the ∆czcRS and the ∆czcRS;hfq − mutants transformed with either an empty (pMMB66EH) plasmid (pMM) or a plasmid containing the czcR gene (pCzcR) under an IPTG inducible taq promoter. Cultures were grown for 6 hours in presence of 0.1 mM IPTG. Blots were exposed to anti-OprD, anti-CzcR or anti-Hsp70 for the loading control.
We have previously noticed that artificial overexpression of the transcriptional regulator CzcR is able to repress oprD, even in the absence of the sensor protein CzcS and in the absence of Zn [7]. We wondered whether in this situation negative regulation also required the Hfq protein. To this aim, we used the pCzcR plasmid containing the czcR gene under the control of an IPTG-inducible promoter. In the absence of Zn overexpression of czcR yielded complete repression of OprD in a ∆czcRS background, while only a minor effect was visible in the absence of Hfq (∆czcRS;hfq − background) ( Figure 3C). This indicates that full repression, mediated directly by CzcR, is dependent upon the Hfq protein even in the absence of Zn challenge.

Hfq is Necessary for Localization of CzcR to the oprD Promoter
In order to investigate the role of Hfq in CzcR-mediated transcriptional repression, we evaluated the DNA-binding efficiency of CzcR to the oprD promoter. To do so, we performed a ChIP (chromosome immunoprecipitation) experiment using anti-CzcR antibody. Interestingly, addition of Zn caused a clear localization of CzcR to the oprD promoter in the WT strain compared to the hfq mutant ( Figure 4A). The weak occupancy of oprD promoter by CzcR observed in the hfq − strain is not due to a defect in czcR expression since the protein is produced to a similar level than in WT even in the absence of Hfq ( Figure 3B). Furthermore, the fact that the level of transcriptional repression ( Figure 4B) perfectly matches with the localization of CzcR on the oprD promoter ( Figure 4A) suggests the requirement of RNA chaperone Hfq for the complete binding of CzcR to oprD, leading to transcriptional repression of the oprD porin.

Is Hfq also Involved in the Repression Mediated by other Transcriptional Regulators?
We have previously observed that CopR, the transcriptional regulator of the CopRS TCS involved in the copper response, also repressed the transcription of the oprD gene [5]. To determine whether this negative regulation mechanism also depends on the RNA chaperone Hfq, we first evaluated the effect of CopR induction after 6 h of growth in the presence of Cu. Western blot analysis showed no OprD repression in the absence of Hfq ( Figure 5A). In addition, similarly to the overproduction of CzcR ( Figure 3A), the overproduction of CopR, was unable to decrease OprD level in the absence of Hfq ( Figure 5B). These results highlight the importance of Hfq for successful gene repression mediated by, at least, two TCS transcriptional regulators.

Is Hfq also Involved in the Repression Mediated by other Transcriptional Regulators?
We have previously observed that CopR, the transcriptional regulator of the CopRS TCS involved in the copper response, also repressed the transcription of the oprD gene [5]. To determine whether this negative regulation mechanism also depends on the RNA chaperone Hfq, we first evaluated the effect of CopR induction after 6 h of growth in the presence of Cu. Western blot analysis showed no OprD repression in the absence of Hfq ( Figure 5A). In addition, similarly to the overproduction of CzcR ( Figure 3A), the overproduction of CopR, was unable to decrease OprD level in the absence of Hfq ( Figure 5B). These results highlight the importance of Hfq for successful gene repression mediated by, at least, two TCS transcriptional regulators.

Is Hfq also Involved in the Repression Mediated by other Transcriptional Regulators?
We have previously observed that CopR, the transcriptional regulator of the CopRS TCS involved in the copper response, also repressed the transcription of the oprD gene [5]. To determine whether this negative regulation mechanism also depends on the RNA chaperone Hfq, we first evaluated the effect of CopR induction after 6 h of growth in the presence of Cu. Western blot analysis showed no OprD repression in the absence of Hfq ( Figure 5A). In addition, similarly to the overproduction of CzcR (Figure 3A), the overproduction of CopR, was unable to decrease OprD level in the absence of Hfq ( Figure 5B). These results highlight the importance of Hfq for successful gene repression mediated by, at least, two TCS transcriptional regulators.

Involvement of the Hfq Protein in Carbapenem Resistance
According to these results, Hfq appears to be an important player in P. aeruginosa carbapenem resistance. In order to confirm this, we determined the minimum inhibitory concentration (MIC) for imipenem of various mutants (Table 2). Zn or Cu increased the MIC higher than the imipenem C value (8 µg·mL −1 ), conferring resistance to P. aeruginosa only in presence of Hfq protein. As previously observed [7], the CzcRS TCS is also essential for the imipenem resistance mechanism induced by Zn as the ∆czcRS mutant remained susceptible (Table 2). Interestingly, Hfq is also important for basal imipenem tolerance. In the hfq − or ∆czcRS;hfq − strains, the MIC dropped significantly even in absence of metals. Hfq is therefore a crucial player in OprD porin regulation, modulating imipenem resistance as a consequence of metal response.

Discussion
Carbapenems are an important class of antibiotics active against both Gram-negative and Gram-positive bacteria. They are often used as last resource for the treatment of P. aeruginosa infections. Carbapenem resistance in P. aeruginosa, however, is an important and worldwide problem with a resistance rate of about 20% in Europe [36]. These antibiotics penetrate the bacterium via the OprD porin, therefore the main and most frequent mechanism of resistance to this family of antibiotics is a decrease in the amount of OprD protein [6,37]. Several regulators and small molecules are known to modulate the expression of oprD (reviewed in [38]) and the post-transcriptional regulation of this porin has been speculated on for a long time [39], suggesting OprD as a highly regulated porin. In the present study we found that, in the presence of Zn or Cu, oprD is negatively regulated by two mechanisms dependent on the RNA chaperone Hfq (Figure 6). Hfq is a well-known RNA binding protein involved in post-transcriptional regulation via modulation of sRNA-mRNA interactions [17]. Since many sRNAs control gene expression in proteobacteria, mutations affecting the Hfq expression yield pleiotropic effects. In P. aeruginosa, mutations in Hfq strongly affect quorum sensing and virulence factors production [26].
Genes 2016, 7, 82 10 of 14 Figure 6. Co-regulation mechanism linking metal resistance, virulence and carbapenem resistance. Zn activates the CzcS sensor protein that will in turn activates the CzcR transcriptional regulator. CzcR induces the expression of CzcCBA efflux pump involved in Zn, Cd and Co resistance [7] and act positively on the virulence of P. aeruginosa [9]. CzcR is also implicated in the repression oprD transcription yielding to carbapenem resistance [7]. The RNA chaperone Hfq is necessary for the binding of CzcR to the oprD promoter leading to transcriptional repression. Additionally a Zn-inducible, Hfq-dependent mechanism decreases the amount of oprD mRNA yielding to an alternative pathway inducing carbapenem resistance in presence of Zn, and even in the absence of CzcR. Overexpression of the CopR protein, either by Cu or artificially, represses OprD only in the presence of the RNA chaperone Hfq. Additionally CopR is able to induce the expression of CzcR [5].
Since Zn treatment did not affect the expression of hfq (Appendix Figure A1), the Hfq-mediated regulation in the absence of CzcR might be due to a Zn-regulated sRNA. To date, the only known sRNA able to modulate oprD levels is phrS [40]. However, overexpression of this sRNA was shown to increase OprD protein levels [40] and Zn treatment failed to modulate phrS expression (Appendix Figure A2). To our knowledge, there are no sRNAs shown to be involved in oprD negative regulation. On the other hand, several studies have demonstrated that Hfq can bind directly to the target mRNA without the help of sRNA [41][42][43].
The transcriptional effect of Hfq on the regulation of oprD reveals an important function of this RNA chaperone targeting the functionality of two-component systems transcriptional regulators. In P. aeruginosa, some 130 TCS have been listed [11]; a huge number accounting for its great versatility and ability to sense and respond to environmental stimuli. These two-component systems usually consist of a sensor histidine kinase membrane-spanning protein involved in the detection of environmental stimuli. This sensor protein activates a response regulator by phosphorylation that, in turn, modulates gene expression by binding to the target promoters [44].
The data presented here demonstrate that Hfq is necessary for the binding activity of CzcR to the oprD promoter in the presence of Zn (Figures 4 and 6), without affecting its expression or stability (Figure 3). The mechanism by which Hfq allows the binding of CzcR to the oprD promoter is not yet understood, but several possibilities could be investigated according to the known properties of this RNA chaperone [45]. Hfq is able to bind to DNA and organize the E. coli nucleoid [46], it could therefore be directly involved in the modeling of the oprD promoter, favoring CzcR binding. On the other hand, the possibility of an indirect effect on the regulation of oprD through the modulation of CzcR co-factor expression by a small RNA cannot be excluded.
We showed that Hfq is also necessary to the copper repression mediated by the CopRS TCS. In the presence of Cu, the CopR regulator represses oprD independently of CzcR [5]. We showed here that this activity is also Hfq-dependent ( Figure 5). Interestingly in the absence of Hfq, the overexpression of CopR displayed no OprD repression ( Figure 5B) while CzcR overexpression weakly decreased the amount of OprD protein ( Figure 3C). This difference could be explained by the Figure 6. Co-regulation mechanism linking metal resistance, virulence and carbapenem resistance. Zn activates the CzcS sensor protein that will in turn activates the CzcR transcriptional regulator. CzcR induces the expression of CzcCBA efflux pump involved in Zn, Cd and Co resistance [7] and act positively on the virulence of P. aeruginosa [9]. CzcR is also implicated in the repression oprD transcription yielding to carbapenem resistance [7]. The RNA chaperone Hfq is necessary for the binding of CzcR to the oprD promoter leading to transcriptional repression. Additionally a Zn-inducible, Hfq-dependent mechanism decreases the amount of oprD mRNA yielding to an alternative pathway inducing carbapenem resistance in presence of Zn, and even in the absence of CzcR. Overexpression of the CopR protein, either by Cu or artificially, represses OprD only in the presence of the RNA chaperone Hfq. Additionally CopR is able to induce the expression of CzcR [5].
Since Zn treatment did not affect the expression of hfq (Appendix Figure A1), the Hfq-mediated regulation in the absence of CzcR might be due to a Zn-regulated sRNA. To date, the only known sRNA able to modulate oprD levels is phrS [40]. However, overexpression of this sRNA was shown to increase OprD protein levels [40] and Zn treatment failed to modulate phrS expression (Appendix Figure A2). To our knowledge, there are no sRNAs shown to be involved in oprD negative regulation. On the other hand, several studies have demonstrated that Hfq can bind directly to the target mRNA without the help of sRNA [41][42][43].
The transcriptional effect of Hfq on the regulation of oprD reveals an important function of this RNA chaperone targeting the functionality of two-component systems transcriptional regulators. In P. aeruginosa, some 130 TCS have been listed [11]; a huge number accounting for its great versatility and ability to sense and respond to environmental stimuli. These two-component systems usually consist of a sensor histidine kinase membrane-spanning protein involved in the detection of environmental stimuli. This sensor protein activates a response regulator by phosphorylation that, in turn, modulates gene expression by binding to the target promoters [44].
The data presented here demonstrate that Hfq is necessary for the binding activity of CzcR to the oprD promoter in the presence of Zn (Figures 4 and 6), without affecting its expression or stability ( Figure 3). The mechanism by which Hfq allows the binding of CzcR to the oprD promoter is not yet understood, but several possibilities could be investigated according to the known properties of this RNA chaperone [45]. Hfq is able to bind to DNA and organize the E. coli nucleoid [46], it could therefore be directly involved in the modeling of the oprD promoter, favoring CzcR binding. On the other hand, the possibility of an indirect effect on the regulation of oprD through the modulation of CzcR co-factor expression by a small RNA cannot be excluded.
We showed that Hfq is also necessary to the copper repression mediated by the CopRS TCS. In the presence of Cu, the CopR regulator represses oprD independently of CzcR [5]. We showed here that this activity is also Hfq-dependent ( Figure 5). Interestingly in the absence of Hfq, the overexpression of CopR displayed no OprD repression ( Figure 5B) while CzcR overexpression weakly decreased the amount of OprD protein ( Figure 3C). This difference could be explained by the modest presence of CzcR on the oprD promoter even in the absence of Hfq ( Figure 4A). This suggests that the Hfq requirement for repression mediated by TCS regulators might differ from one regulator to the next. The fact that Hfq is necessary for the repressor activity of two distinct TCSs suggests an important role of this RNA chaperone in the sensing and adaptation of P. aeruginosa to its environment. Consistently, this might explain the pleiotropic effects observed in the hfq mutant [25,26]. Taken together our data support the fact that the Hfq protein may represent a very interesting target for drug discovery [47]. Blocking Hfq functions will affect not only P. aeruginosa quorum sensing and virulence but also several TCSs involved in carbapenem resistance.

Conclusions
We previously observed that the environmental opportunistic pathogen Pseudomonas aeruginosa becomes more virulent and resistant to carbapenem antibiotic in the presence of metal pollutants such as zinc, or copper. The data presented here, bring new light on the mechanisms of metal inducible antibiotic resistance. We found that the transcriptional repression of OprD, involved in carbapenem resistance, requires the RNA chaperone Hfq. By revealing the underlying mechanism, we discovered that the metal specific transcriptional regulator CzcR requires Hfq to bind to the oprD promoter. Furthermore, we demonstrated that another transcriptional regulator from another two-component system also requires the Hfq protein to repress OprD in presence of metal. These results highlight how Hfq chaperone acts as an essential component of the environment-sensing in P. aeruginosa, liking metal detection to antibiotic resistance.

Conflicts of Interest:
The authors declare no conflict of interest. The founding sponsors had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, and in the decision to publish the results.