Pentachlorophenol Induction of the Pseudomonas aeruginosa mexAB-oprM Efflux Operon: Involvement of Repressors NalC and MexR and the Antirepressor ArmR

Pentachlorophenol (PCP) induced expression of the NalC repressor-regulated PA3720-armR operon and the MexR repressor-controlled mexAB-oprM multidrug efflux operon of Pseudomonas aeruginosa. PCP's induction of PA3720-armR resulted from its direct modulation of NalC, the repressor's binding to PA3720-armR promoter-containing DNA as seen in electromobility shift assays (EMSAs) being obviated in the presence of this agent. The NalC binding site was localized to an inverted repeat (IR) sequence upstream of PA3720-armR and overlapping a promoter region whose transcription start site was mapped. While modulation of MexR by the ArmR anti-repressor explains the upregulation of mexAB-oprM in nalC mutants hyperexpressing PA3720-armR, the induction of mexAB-oprM expression by PCP is not wholly explainable by PCP induction of PA3720-armR and subsequent ArmR modulation of MexR, inasmuch as armR deletion mutants still showed PCP-inducible mexAB-oprM expression. PCP failed, however, to induce mexAB-oprM in a mexR deletion strain, indicating that MexR was required for this, although PCP did not modulate MexR binding to mexAB-oprM promoter-containing DNA in vitro. One possibility is that MexR responds to PCP-generated in vivo effector molecules in controlling mexAB-oprM expression in response to PCP. PCP is an unlikely effector and substrate for NalC and MexAB-OprM - its impact on NalC binding to the PA3720-armR promoter DNA occurred only at high µM levels - suggesting that it mimics an intended phenolic effector/substrate(s). In this regard, plants are an abundant source of phenolic antimicrobial compounds and, so, MexAB-OprM may function to protect P. aeruginosa from plant antimicrobials that it encounters in nature.

Expression of the mexAB-oprM operon is regulated by the product of a local repressor gene, mexR [24], occurring upstream of the efflux genes and the target of mutation in MexAB-OprM-overexpressing multidrug-resistant nalB lab [24,25] and clinical [26][27][28] isolates. MexR acts to control mexAB-oprM expression from one of two promoters (PI; most distal from the efflux genes) for this efflux operon [29,30], binding as a dimer [31] to a site that also overlaps the mexR promoter [29,32] and, so, effects negative autoregulation. Recent studies show that its ability to bind and so repress mexAB-oprM is governed by the redox status of the protein, MexR serving to regulate mexAB-oprM expression in responsive to oxidative stress [33,34]. MexR repressor activity is also modulated by the product of the armR anti-MexR repressor gene, which binds to MexR and negatively impacts its ability to bind to the mexAB-oprM PI promoter region [35,36]. armR occurs as part of the 2-gene operon PA3720-armR that is regulated by the product of the divergently-transcribed nalC repressor gene [37], with nalC mutants showing elevated PA3720-armR expression and, so, elevated mexAB-oprM expression and multidrug resistance [37] as a result of ArmR modulation of MexR's repressor activity [35]. nalC lab and clinical isolates expressing mexAB-oprM and showing a multidrug-resistant phenotype have been reported [24,37,38]. A third repressor involved in regulating mexAB-oprM expression, NalD, is encoded by a gene unlinked to mexAB-oprM and nalC/ PA3720-armR and regulates mexAB-oprM from a second, efflux operon-proximal promoter (PII) [30]. Mutations in nalD have been described in lab and clinical multidrug-resistant mutants [39].
Despite their contribution to antimicrobial resistance, RND family multidrug efflux systems in P. aeruginosa are increasingly appreciated as having other than drug efflux as an intended function, most of these systems being regulated independently of antimicrobials and instead induced in response to environmental stress (e.g, MexCD-OprJ, envelope stress [40]; MexEF-OprN, nitrosative stress [41]; MexXY-OprM, oxidative [42] and ribosome [43,44] stress). Recently, the mexAB-oprM efflux system and its regulatory locus nalC/PA3720-armR have been shown to be inducible by the uncoupler of oxidative phosphorylation and environmental contaminant, pentachlorophenol (PCP) [45], the implication being that this compound is being 'sensed' by NalC and mexAB-oprM recruited as a result of ArmR-mediated MexR modulation. While MexAB-OprM appears able to accommodate this toxin (mutants lacking the pump are more sensitive to PCP) [45], it is unclear whether PCP is the actual signal/substrate or whether downstream effects of the uncoupling of oxidative phosphorylation are responsible for nalC/ PA3720-armR and mexR/mexAB-oprM upregulation. The current study was undertaken to address this and to ascertain the mechanism by which PCP ultimately induces mexAB-oprM. We report here that NalC is a PCP-responsive repressor that mediates PCP induction of PA3720-armR and, possibly, mexAB-oprM but that PCP is also able to induce mexAB-oprM expression independently of PA3720-armR, possibly as a result of PCP-generated oxidative stress and, so, redox regulation of MexR. During the course of this work another study demonstrating PCP control of NalC repressor activity was published [46] although the mechanism(s) by which this toxin induced mexAB-oprM were not assessed, and the assumption that it was via ArmR are not supported by our study.

Bacterial strains and plasmids
The bacterial strains and plasmids used in this study are listed in Table 1. Bacterial cells were cultured in Luria broth (L broth) and on Luria agar with antibiotics, as necessary, at 37uC. Plasmid pET23a and its derivatives were maintained in E. coli with ampicillin (100 mg/ml). Plasmid pEX18Tc and its derivatives were maintained in E. coli with tetracycline (10 mg/ml). Plasmid pLysS was maintained in E. coli with chloramphenicol (30 mg/ml). DPA3720 derivatives of P. aeruginosa strains K767 and K3145 (DarmR) were generated by constructing a PA3720 deletion in plasmid pEX18Tc and mobilizing it into these strains as before [47]. Briefly, 1 kb fragments corresponding to sequences 59 and 39 to the PA3720 sequences being deleted were amplified using the polymerase chain reaction (PCR) and primers 3720 UF (59-GCTAGAATTCGTAGGTGGTGAAGCCGAGC-39; EcoRI site underlined) and 3720 UR (59-GCTAGGTACCTGACGCCGC-CATGTCCCT-39; KpnI site underlined), and 3720 DF (59-GCTAGGTACCCATGGGGGGACTCCTGCG-39; KpnI site underlined) and 3720 DR (59-GCTAAAGCTTCCTG-GTCCTGGCGACACG-39; HindIII site underlined), respectively. The 50 ml PCR reaction mixture contained 1 mg of chromosomal P. aeruginosa K767 DNA as template, 1 mM of each primer, 0.3 mM each deoxynucleoside triphosphate (dNTP), 16 Exact buffer, 16 5P buffer and 2.5 U Exact DNA polymerase (5 PRIME, Inter Medico, Markham, Ontario). Following an initial denaturation step at 95uC for 5 min, the mixture was subjected to 30 cycles of heating at 94uC for 45 sec, 63uC (annealing temperature) for 30 sec, and 72uC for 1 min, before finishing with a10-min incubation at 72uC. PCR products were gel purified, digested appropriately and cloned sequentially into pEX18Tc, to yield the DPA3720 vector, pLMS2. Plasmid pLMS2 was mobilized into P. aeruginosa K767 from into E. coli S17-1 and P. aeruginosa transconjugants harbouring chromosomal inserts of the deletion vector were selected on L-agar plates containing tetracycline (50 mg/ml) and chloramphenicol (5 mg/ml; to counterselect E. coli S17-1). These were subsequently streaked onto L agar containing sucrose (10% [wt/vol]) as before [48], with sucrose-resistant colonies screened for the appropriate deletion using colony PCR [49] with primers 3720 UF and 3720 DR and parameters detailed above (except for an annealing temperature of 55uC). A DarmR derivative of P. aeruginosa K767 was constructed using pEX18Tc:: DarmR plasmid pLC8 as described previously [37] except that transconjugants carrying chromosomal inserts of the deletion vector were selected on L-agar plates containing tetracycline and chloramphenicol as above. DarmR mutants were verified by colony PCR [49] using primers 3719UF (59-CGGTGAGCGTGG-CGCCG-39) and 3720DR . Reaction mixtures were heated to 95uC for 5 min followed by 30 cycles of 30 sec at 94uC, 30 sec at 60uC, 1 min at 72uC and a final 5-min elongation at 72uC.

DNA methods
Standard protocols were used for restriction endonuclease digestion, ligation, transformation, plasmid isolation, agarose gel electrophoresis, and preparation of chemically competent (CaCl 2 ) E. coli cells as described by Sambrook and Russell [50]. Electrocompetent P. aeruginosa cells were prepared as described by Choi and Shwiezer [51]. Chromosomal DNA was extracted from P. aeruginosa using the DNeasy Tissue Kit (Qiagen, Inc., Mississauga, Ontario). Plasmid DNAs were also prepared from E. coli or P. aeruginosa using a GeneJET Plasmid Miniprep kit (Fermentas Canada Inc., Burlington, Ontario, Canada) according to a protocol provided by the manufacturer. The WizardH SV Gel and PCR Clean-Up System kit (Promega Corp., Nepean, Ontario) was used to purify PCR products and to gel-purify DNA fragments generated by restriction endonuclease digestion. Oligonucleotides were synthesized by Integrated DNA Technologies (IDT, Coralville, Iowa, USA) and nucleotide sequencing was performed by AGCT Corp. (Toronto, Ontario).

Susceptibility testing
The antimicrobial susceptibilities of the various P. aeruginosa strains were assessed in 96-well microtiter plates using twofold serial dilutions as described [52]. In assessing the impact of PCP exposure on antimicrobial susceptibility, 0.75 mM PCP (one quarter of the MIC) was included in the growth medium used to prepare the bacterial inoculum and to generate the serial dilutions.

Quantitative RT-PCR
Overnight cultures of P. aeruginosa strains were subcultured (1:49) in fresh L-broth and incubated at 37uC with shaking for 2.5 hours, at which time cells were harvested by centrifugation. In some experiments, PCP (0.75 mM final concentration; J MIC) was added 1 to 1.5 hr before harvesting. Total bacterial RNA was isolated from 1 ml of late-log phase culture using the RiboPure TM Bacteria kit (Ambion, Inc., Streetsville, Ontario) or the High Pure RNA Isolation Kit (Roche Diagnostics) using the manufacturers' protocols, and resuspended in 50 ml elution buffer. Samples were treated with Turbo DNA-Free (Ambion, Inc; 2 U enzyme per 10 mg of RNA for 60 min at 37u). DNA-free RNA (confirmed usng PCR; 1 mg) was used to synthesize cDNA using an iScript cDNA Synthesis kit (Bio-Rad, Mississauga, Ontario) in a reaction mixture formulated as described by the kit manufacturer and incubated for 5 min at 25uC, 30 min at 42uC and 5 min at 85uC. cDNA was then stored until needed at 220uC. Quantification of the cDNA was carried out using a Bio-Rad, CFX96 TM Real-Time PCR Detection System (Bio-Rad, Mississauga, Ontario). PCR amplification reactions were performed in 20 ml reaction volumes each containing 10 ml of SsoFast EvaGreen Supermix (Bio-Rad, Mississauga, Ontario), 0.6 mM each of 2 primers per gene being amplified (mexA: mexAQ-F, 59-CAGCAGCTCTACCAGATCG-39 and mexAQ-R, 59-CGTACTGCTGCTTGCTCA-39; PA 3720: 3720Q-F, 59-GATGCCTTTCCCTTGGTC-CA-39 and 3720Q-R, 59-TCCTTGAGCCACAACACCAG-39; rpsL: rpsLQ-F, 59-GGT-TTCCTCGTACATCGGTG-39 and rpslQ-R, 59-TGCTTACGGTCTTTGACACC-39) and 5 ml of 1:49 diluted cDNA. Following an initial 3-min denaturation at 95uC, the mixture was subjected to 40 cycles of 10 sec at 95uC and 30 sec at 60uC. A melt curve, obtained following an initial 10-sec treatment at 95uC and involving 5-sec incubations of 0.5uC increments beginning at 65uC, was run at the end of 40 cycles to test for the presence of a unique PCR product. The expression levels of mexA and PA3720 were normalized to that of the reference gene rpsL. Biological duplicates and technical triplicates were performed for all samples. To confirm the absence of genomic DNA contamination, 'no-template' controls were performed in technical triplicates for all primer sets employed.

Expression and purification of polyhistidine (His)-tagged NalC protein
The nalC gene was amplified by PCR using primers: nalC-Fwd-NdeI (59-GATCCATATGAACGATGCTTCTCCCCGTCTG-3; NdeI site underlined) and nalC-Rev-SalI (59-GATCGTC-GACGCCCTGCGCGGGGCTCTG-39; SalI site underlined) in a 50 ml PCR reaction that contained 1 mg P. aeruginosa strain K767 chromosomal DNA, 0.2 mM each dNTP, 3% (vol/vol) DMSO, 16 PhusionH High Fidelity (HF) buffer and 1 U PhusionH polymerase (Finnzymes, New England Biolabs, Pickering, Ontario). The reaction mixture was heated for 30 sec at 94uC followed by 30 cycles of 30 sec at 94uC, 30 sec at 54uC and 30 sec at 72uC, before finishing with 10 min at 72uC. The nalCcontaining PCR product was purified, digested with SalI and NdeI and cloned into NdeI-XhoI-restricted pET23a (Novagen, Madison, Wisconsin, USA) to yield pLMS3 encoding His-tagged NalC (i.e, NalC-His). Following nucleotide sequencing of the cloned gene to confirm the absence of PCR-generated mutations, plasmid pLMS3 was introduced into E. coli BL21 (DE3) carrying the pLysS plasmid and NalC-His production induced with IPTG using a modified version of a previously-described protocol [29]. Briefly, an overnight culture of E. coli BL21 (DE3) carrying pLysS and pLMS3 was diluted 1:49 in LB (50 ml) supplemented with the appropriate antibiotics and incubated at 37uC until the optical density at 600 nm reached 0.5-0.6, at which time IPTG (1 mM final concentration) was added and incubation continued a further 2 hr. Cells were harvested by centrifugation (10 min at 100006g) at 4uC and pellets were resuspended in 6 ml buffer A (0.  Following centrifugation for 60 min at 160006g, the NalC-Hiscontaining supernatant was mixed with 500 ml of Ni-NTA Agarose resin (Qiagen, Inc.) equilibrated with 10 ml buffer A containing 5 mM imidazole and incubated, with shaking, for 10 min. The resin was subsequently pelleted by centrifugation (3 min at 30006g) and washed twice with 10 ml buffer A containing 5 mM imidazole and once with 2 ml buffer A containing 5 mM imidazole, the resin again being centrifuged after each wash. Bound protein was eluted stepwise with 500 ml buffer A containing increasing amounts of imidazole (50, 100, 150, 250 mM); at each step the resin was incubated with shaking at room temperature and centrifuged as above. The NalC-His protein was recovered in the supernatant following elution with buffer A containing 250 mM imidazole. Protein concentration was determined using the BCA Protein Assay Kit (Pierce, Illinois, USA), and purified protein (verified using SDS-polyacrylamide gel electrophoresis [10]) was stored at 220uC in 20% (vol/vol) glycerol.

Electromobility shift assay
The binding of purified NalC (see above) and MexR (prepared as above but using pET23a::mexR vector pKLE1 in place of pLMS3) to PCR-amplified target DNAs was assessed using the electromobility shift assay (EMSA) as described previously [30]. Briefly, 50 ng target DNA was incubated with purified NalC or MexR for 20 min at room temperature in a 15 ml reaction mixture containing 16 binding buffer (750 mM KCl, 0.5 mM dithiothreitol, 0.5 mM EDTA, 50 mM Tris-HCl, pH 7.4). Following the addition of EMSA gel-loading solution, mixtures were separated by electrophoresis on a non-denaturing 8% (wt/vol) polyacrylamide gel in 0.5 TBE buffer (22 mM Tris-HCl, 22 mM boric acid, 0.5 mM EDTA, pH 8.0) and gels were stained with 16SYBR Green EMSA nucleic acid stain. DNA was then visualized using digital photography with a S6656 SYPRO photographic filter (S6656).
The MexR target DNA, corresponding to the 351-bp mexR-mexA intergenic region, was amplified using the primers K9 (59-CTGAAGATCTGTTGCATAGCGTTG-TCCTCA-39) and K10 (59-ACGGGGTACCCGGGGTAGTTCATTGGTTTG-39) in a reaction mixture formulated as above for the 205-bp nalC-PA3720 intergenic region, except that DMSO was excluded and MgCl 2 was included at 1.5 mM final concentration. The reaction was heated at 94uC for 5 min, followed by 30 cycles of 1 min at 94uC, 2 min at 56uC, and 1.5 min at 72uC, before finishing with 30 min at 72uC. To assess the specificity of any binding observed, excess salmon sperm DNA (100 mM) was added to the reaction mixtures prior to the addition of protein. In experiments where various compounds or plant extracts were being tested for an ability to interfere with protein-DNA binding, these were also added (concentrations indicated in the appropriate figure legends) prior to adding the protein. Since preparation of PCP required the use of a solvent (10 mM NaOH), control experiments involving 10 mM NaOH only were carried out to ensure that the solvent itself was not impacting the EMSA.

Mapping the PA3720-armR transcription start site
To identify the transcription start site for the PA3720-armR operon, the 59 rapid amplification of cDNA ends (RACE) protocol [53] and a 59/39 RACE Kit, 2nd Generation (Roche Diagnostics) were used as described previously with modifications [29]. Total RNA was prepared from PA3720-armR-expressing P. aeruginosa nalC mutant K1454 using the Roche High Pure RNA Isolation Kit according to the manufacturer's instructions, and contaminating DNA was removed using the Turbo DNase Kit (Ambion). cDNA was synthesized from total RNA (1 mg) with a RACE kit-provided reverse transcriptase and a PA3720-specific primer, 3720 RACE Sp1 (59-GGCGGATAGACCAGGG CGAAAT-39; anneals 131 bp downstream of the PA3720 ATG start site), using a protocol provided by the manufacturer. A homopolymeric A-tail was added to the 39 end of the total cDNA using terminal transferase and dATP, and the dA-tailed cDNA subsequently PCR amplified using a kit-provided oligo (dT)-anchor primer and a second PA3720-specific primer, 3720 RACE Sp2 (59-CGGCCAGCTCGCGGGGCATCT-39; anneals 57 bp downstream of the PA3720 ATG start site). The resulting PCR product was purified and cloned into the pETBlue-1 AccepTor TM Vector (EMD Chemicals, Gibbstown, New Jersey, USA) as per instructions provided by the manufacturer. Following transformation of NovaBlue Giga Singles Competent Cells (Novagen), plasmid-carrying E. coli cells were selected on L-agar kanamycin (50 mg/ml) and tetracycline (15 mg/ml). Plasmids carrying inserts of the amplified 59 end of PA3720 mRNA were identified and the inserts sequenced and aligned with the PAO1 genome sequence to determine the PA3720-armR transcription start site.
PCP induction of PA3720-armR was lost in a nalC mutant (K1454) -although the mutant showed the expected increase in PA3720-armR expression compared with its wild type parent K767, PCP treatment did not enhance expression of this operon in K1454 (Fig. 1). This indicated that NalC was mediating the PCP induction of PA3720-armR, either via a direct response to PCP or via recognition of PCP-generated cellular products (PCP is an energy inhibitor). To assess this NalC was purified and the impact of PCP on its binding to the ca. 200 bp nalC-PA3720 intergenic region was examined using EMSA. As seen in Fig. 2, NalC bound the nalC-PA3720 intergenic region ( Fig. 2A) with binding compromised by the addition of PCP (some loss of binding was detected at 10 mM PCP with almost complete loss of binding seen at 750 mM) (Fig. 2B). Thus, NalC responds directly to PCP in mediating its induction of PA3720-armR.

Mapping the NalC binding site
In an attempt to localize the NalC binding site NalC binding to ca. 100 bp PA3720-proximal and ca. 100 bp PA3720 distal fragments was assessed, with binding confirmed only for the PA3720-proximal fragment (Fig. 3A, cf Fig. 3B). This fragment (with endpoints II) included putative promoters for nalC and PA3720-armR as well as an inverted repeat sequence that might serve as the NalC-binding site (Fig. 4). To assess the latter, a 45 bp fragment encompassing the inverted repeat (IR; endpoints III in Fig. 4) was used in an EMSA and NalC binding was observed (Fig. 3C). An identically-sized fragment lacking the IR and encompassing a region more PA3720-proximal (endpoints IV in Fig. 4) failed to bind NalC (Fig. 3D) and little if any binding was observed to a fragment encompassing the repeat region (endpoints III) in which the repeat sequence was mutated, and then only at the highest concentration of NalC (Fig. 3E). This confirmed the repeat region as the NalC-binding site. Interestingly, a 21-bp fragment carrying only the IR failed to bind NalC (data not shown) indicating that the proper disposition of the repeat region for NalC binding required flanking sequence. The IR-containing region to which NalC bound contains the putative promoter regions for nalC and PA3720-armR [predicted using neural network promoter prediction (http://www.fruitfly.org/seq_tools/promoter.html)] (Fig. 4) as would be expected. Using 59-RACE the PA3720-armR transcription start site was mapped to an adjacent CT pair of bases (exact base could not be determined unequivocally owing to limitations of the RACE), which occurred downstream of the putative PA3720-armR -10 region and immediately downstream of the predicted (using neural network) start site (Fig. 4). This is consistent with NalC regulating expression of the PA3720-armR operon. Figure 1. PCP induction of PA3720-armR. qRT-PCR results showing impact of PCP and/or a nalC mutation on PA3720 expression (as a measure of PA3720-armR expression). Expression of PA3720 is reported relative to the rpsL internal control for wild type P. aeruginosa K767 and its nalC derivative, K1454 exposed or not to PCP (0.75 mM; 1.5 hr). Results shown are the mean +/2 standard error of one cDNA sample for each, processed in triplicate, and are representative of 2 independent experiments. doi:10.1371/journal.pone.0032684.g001 PCP induction of mexAB-oprM is NalC and MexRdependent but PA3720-ArmR-independent In addition to its induction of PA320-armR, PCP was previously shown to induce the MexR-regulated mexAB-oprM multidrug efflux operon [45], results that have been confirmed here (modest ca. 3fold induction; Fig. 5). Mutations in nalC and mexR yielded elevated expression of mexAB-oprM (nalC, 3-fold; mexR, 5-fold; Fig. 5), as seen previously [45], and this was no longer responsive to PCP (Fig. 5). Thus, PCP operates though these regulators in promoting mexAB-oprM expression. Given that NalC controls expression of the ArmR anti-repressor that modulates MexR repression of mexAB-oprM, the obvious interpretation of these results is that PCP as a NalC effector affords PA3720-armR expression, with ArmR modulating MexR activity to effect mexAB-oprM expression. To assess this, PCP induction of the efflux operon was examined in armR mutant strain K3145. As seen in Fig. 6, mexAB-oprM was still induced in response to PCP treatment, albeit to a lesser extent, indicating that PCP induction of this efflux operon can occur independently of its promotion of ArmR expression. PA3720 encodes a protein of unknown function that, like armR, is NalC regulated [37] and PCP inducible [45] (Fig. 1) and which might, therefore, play a role in PCP induction of mexAB-oprM. Thus, PCP induction of mexAB-oprM was assessed in a mutant, K3146, lacking PA3720. Again, however, PCP induction of mexAB-oprM was retained in the mutant (Fig. 6). Similarly, PCP induction of mexAB-oprM was seen in a mutant strain, K3130, lacking both PA3720 and armR (Fig. 6), ruling out both genes independently mediating PCP induction of the efflux operon. Similarly, while loss of armR compromised the mexAB-oprM overexpression seen in a nalC mutant (Fig. 5) in agreement with previous results [37], PCP induction of mexAB-oprM was retained in the nalC DarmR mutant K2276 (Fig. 5) again indicating that PCP induction of the efflux operon can occur independently of ArmR. Interestingly, however, and in contrast to what was observed when armR was eliminated in wild type strain K767, where there was little or no impact on basal levels of mexAB-oprM expression (in the absence of PCP) (Fig. 6), loss of armR in a nalC background reduced expression of mexAB-oprM to below the basal levels seen in K767 (Fig. 5).

PCP is not a MexR effector
The observation that PCP induction of mexAB-oprM is dependent on MexR but independent of PA3720-armR suggested that PCP may be capable of directly modulating MexR activity to effect efflux gene expression. To assess this, the impact of PCP on MexR binding to a DNA fragment containing the mexAB-oprM promoter region was examined using EMSA. As seen in Fig. 7 MexR binding was observed, revealing a complex pattern of DNA 2000 (lanes 6) ng of purified NalC-his was incubated with 40 ng of C) the annealed oligonucleotide NalC-1 and its reverse complement (corresponds to sequence delineated by endpoints III in Fig. 4A), D) the annealed oligonucleotide NalC-2 and its reverse complement (corresponds to sequence delineated by endpoints IV in Fig. 4A), and E) annealed oligonucleotide NalC-3 and its reverse complement, in which the AGAACTGT sequence of NalC-1 corresponding to the first half of the inverted repeat highlighted in Fig. 4(A) (shaded sequence overlapping the putative 210 regions for nalC and PA3720-armR) is changed to TCTTGACA. doi:10.1371/journal.pone.0032684.g003 binding suggestive of variable occupancy of multiple MexR binding sites and/or cooperative binding of MexR to its target DNA. Inn any case, PCP had no impact on MexR binding at concentrations up to 100 mM, and while there were qualitative differences in the pattern of shifted DNA at the highest concentrations (500 and 750 mM) virtually all of the mexAB-oprM promoter-containing DNA remained bound to MexR at these concentrations of PCP. Thus, in contrast to NalC, PCP is not an effector for MexR.

Discussion
ArmR is an anti-repressor that modulates MexR repressor activity in nalC mutants that hyperexpress the PA3720-armR operon, accounting for the elevated expression of the mexAB-oprM multidrug efflux operon in such mutants [37]. In light of the impact of PCP on NalC repression of the PA3720-armR operon and the NalC dependence of PCP induction of mexAB-oprM it seemed reasonable to suppose that PCP induction of the efflux operon followed from elevated armR expression and subsequent ArmR-dependent modulation of MexR repression of mexAB-oprM. Indeed, this was the conclusion of a recent study, published while this work was in progress, demonstrating that PCP induction of PA3720-armR expression results from PCP directly modulating NalC's repression of this operon [46]. Still, the requirement for ArmR in PCP-inducible mexAB-oprM expression was not assessed in that study and results presented here clearly demonstrate that PCP induction of the efflux operon occurs in the absence of the anti-repressor. While there seemed to be some reduction in the Figure 4. Mapping the PA3720-armR transcriptional start site. The nalC/PA3720-armR intergenic region highlighting the RACE-determined transcription start site for PA3720-armR (bolded and italicized), the predicted transcriptional initiation site (marked with an asterisk; assessed using neural network promoter prediction software provided by M. G. Reese at http://www.fruitfly.org/seq_tools/promoter.html), the nalC and PA3720 start codons (arrows), and putative 210/235 sites for nalC (overlined) and PA3720-armR (underlined) promoters. The end-points of PCR-generated PA3720-distal (I) and -proximal (II) fragments and oligonucleotides (III and IV) used in EMSAs with purified NalC (see Fig. 2B) are identified by arrowheads above the sequence. The shaded sequence corresponds to an inverted repeat and possible NalC-binding site. doi:10.1371/journal.pone.0032684.g004 Figure 5. NalC and MexR dependence of PCP induction of mexAB-oprM. Expression of mexA (as a measure of mexAB-oprM) was assessed in P. aeruginosa strains K767 (wild type; WT), K1454 (nalC), K2276 (nalC DarmR) and K2568 (DmexR) exposed or not to PCP (0.75 mM; 1.5 hr) using qRT-PCR. Expression was normalized to rpsL controls and is reported relative (fold change) to untreated P. aeruginosa K767. Results shown are the mean +/2 standard error of one cDNA sample for each, processed in triplicate, and are representative of 2 independent experiments. doi:10.1371/journal.pone.0032684.g005 Figure 6. ArmR-/PA3720-independence of PCP induction of mexAB-oprM. Expression of mexA (as a measure of mexAB-oprM) was assessed in P. aeruginosa strains K767 (wild type; WT), K3145 (DarmR), K3146 (DPA3720) and K3130 (DPA3720-armR) exposed or not to PCP (0.75 mM; 1.5 hr) using qRT-PCR. Expression was normalized to rpsL controls and is reported relative (fold change) to untreated P. aeruginosa K767. Results shown are the mean +/2 standard error of one cDNA sample for each, processed in triplicate, and are representative of 2 independent experiments. doi:10.1371/journal.pone.0032684.g006 absolute levels of PCP-inducible mexAB-oprM expression in the absence of armR, all mutants lacking armR (i.e., K767DarmR, nalCDarmR, K767DPA3720DarmR) showed fold increases in expression of the efflux genes in the presence of PCP (2-to 3fold increase) that were similar to wild type strain K767 (ca. 3fold).
While it is unlikely that elevated armR expression in PCP-treated P. aeruginosa would not contribute to mexAB-oprM expression as a result of ArmR modulation of MexR repressor activity, these results do indicate that there are other mechanisms by which PCP can promote mexAB-oprM expression, probably in parallel with the ArmR-dependent mechanism. Such a mechanism appears not to involve PA3720, whose function still remains to be elucidated, but does require MexR, which may respond to PCP. Given the observation that PCP does not impact MexR binding to the mexAB-oprM promoter containing DNA in vitro, one possibility is that P. aeruginosa exposure to PCP promotes in vivo production of MexR effector molecule(s) that facilitate mexAB-oprM derepression. In this vein, it has been shown that MexR binding to the mexAB-oprM promoter region and its control of efflux gene expression is impacted by oxidative stress, MexR being a redox responsive regulator whose ability to bind and repress mexAB-oprM is compromised when the protein is oxidized [33,34]. Thus, exposure to hydroperoxide oxidizes MexR and alleviates promoter binding in vitro and induces mexAB-oprM expression in vivo [34]. One possibility, then, is that PCP exposure generates some form of oxidative stress in P. aeruginosa. Intriguingly, PCP has been shown to dramatically increase O 2 flux in P. aeruginosa, generating an oxidative stress that could impact MexR activity [54]. Puzzlingly, PCP induction of mexAB-oprM is absent in nalC mutant K1454, despite the fact that efflux gene expression is less than what is seen in a mexR null strain (K2568) -while much of MexR is inactive as a repressor in K1454, being bound by ArmR, clearly some MexR repression of mexAB-oprM is still occurring and would, presumably, be susceptible to modulation by PCP (or PCP-generated effectors). Still, there is some unexplained complexity to mexAB-oprM expression in a nalC strain given the greater impact of armR loss on efflux gene expression in a nalC background as compared to an otherwise wild type background where in absolute terms mexAB-oprM expression is less (2-fold; Fig. 5) in a nalC DarmR mutant (K2276) than a DarmR mutant (K3145). Whether this explains the anomaly, there is likely to be differences in the expression of at least some genes in K2276 vs. K3145, with PA3720 presumably up-regulated in the former but not the latter and any other genes impacted by loss of nalC and/or upregulation of PA3720 differentially expressed in the two strains. How these might influence mexAB-oprM expression is uncertain.
The modulation of NalC repressor activity by PCP (and other chlorinated phenols [46]) occurs at high levels (high mM to low mM) for an effector molecule suggesting that it only mimics an intended NalC effector and ultimate inducer of mexAB-oprM, likely a phenolic of some kind. Plants are common sources of phenolic compounds, many of which have antimicrobial activity [55][56][57][58], and these may be the intended inducers/substrates for MexAB-OprM. While attempts were made to assess the ability of plant extracts to 'interact' with NalC and to induce mexAB-oprM expression technical issues (extract insolubility in aqueous solutions) compromised these efforts.