Hypochlorous acid and hydrogen peroxide-induced negative regulation of Salmonella enterica serovar Typhimurium ompW by the response regulator ArcA

Background Hydrogen peroxide (H2O2) and hypochlorous acid (HOCl) are reactive oxygen species that are part of the oxidative burst encountered by Salmonella enterica serovar Typhimurium (S. Typhimurium) upon internalization by phagocytic cells. In order to survive, bacteria must sense these signals and modulate gene expression. Growing evidence indicates that the ArcAB two component system plays a role in the resistance to reactive oxygen species. We investigated the influx of H2O2 and HOCl through OmpW and the role of ArcAB in modulating its expression after exposure to both toxic compounds in S. Typhimurium. Results H2O2 and HOCl influx was determined both in vitro and in vivo. A S. Typhimurium ompW mutant strain (∆ompW) exposed to sub-lethal levels of H2O2 and HOCl showed a decreased influx of both compounds as compared to a wild type strain. Further evidence of H2O2 and HOCl diffusion through OmpW was obtained by using reconstituted proteoliposomes. We hypothesized that ompW expression should be negatively regulated upon exposure to H2O2 and HOCl to better exclude these compounds from the cell. As expected, qRT-PCR showed a negative regulation in a wild type strain treated with sub-lethal concentrations of these compounds. A bioinformatic analysis in search for potential negative regulators predicted the presence of three ArcA binding sites at the ompW promoter region. By electrophoretic mobility shift assay (EMSA) and using transcriptional fusions we demonstrated an interaction between ArcA and one site at the ompW promoter region. Moreover, qRT-PCR showed that the negative regulation observed in the wild type strain was lost in an arcA and in arcB mutant strains. Conclusions OmpW allows the influx of H2O2 and HOCl and is negatively regulated by ArcA by direct interaction with the ompW promoter region upon exposure to both toxic compounds.


Background
Hydrogen peroxide (H 2 O 2 ) and hypochlorous acid (HOCl) are reactive oxygen species that are part of the oxidative burst encountered by S. Typhimurium upon internalization by phagocytic cells. Under acidic conditions, such as those found inside the phagosome, H 2 O 2 is generated spontaneously by the reaction of two superoxide anion (O 2 − ) molecules [1]. Moreover, S. Typhimurium encodes both periplasmic and cytoplasmic superoxide dismutases that catalyze O 2 − dismutation to generate H 2 O 2 and molecular oxygen [2][3][4]. HOCl is produced by the action of myeloperoxidase (MPO) in a reaction that depends on H 2 O 2 , Cl − and acidic conditions [5,6]. Taken together, H 2 O 2 and HOCl react with thiol and heme groups, copper and iron salts generating the reactive hydroxyl radical (OH . ). As a consequence, they produce lipid peroxidation, chlorination of tyrosine residues, oxidation of iron centers, protein cross linking and DNA damage [5][6][7][8].
In order to enter Gram negative bacteria, H 2 O 2 and HOCl must be able to cross the outer membrane (OM) and even though several biological membranes are permeable to H 2 O 2 , studies in E. coli and Saccharomyces cerevisiae showed that this compound cannot diffuse freely [9,10]. For HOCl, diffusion through the OM is reported to be limited [11]. One possibility for H 2 O 2 and HOCl influx through the OM is diffusion through porins. In this context, we recently reported that OmpD, S. Typhimurium most abundant OM porin, allows H 2 O 2 diffusion [12]. OM porins are organized as homo-trimers (classic porins) or monomers (small porins) forming aqueous channels that allow the influx of hydrophilic solutes with a molecular weight ≤ 600 Da [13]. Classic porins, including OmpC and OmpF, form β-barrels with 12-22 transmembrane segments while small porins (OmpW) are composed of 8-10 [14,15]. The crystal structure of OmpW from E. coli revealed that it forms an 8-stranded β-barrel and functions as an ion channel in lipid bilayers [16,17]. In Vibrio cholerae, OmpW was described as an immunogenic 22 KDa protein [18] and its expression is altered by factors such as temperature, salinity, nutrient availability and oxygen levels [19]. Additionally, several studies show that porins are regulated by ROS. Due its oxidant nature and diffusion through the OM, regulation of porin expression must be tightly regulated as a mechanism of controlling OM permeability. Accordingly, S. Typhimurium ompD and ompW expression is regulated in response to H 2 O 2 and paraquat [12,20], respectively, and S. Enteritidis and Typhimurium exposure to HOCl results in lower levels of ompD, ompC and ompF transcripts [21].
The cellular response to oxidative stress is regulated at the transcriptional level by activating the SoxRS and OxyR regulons in response to O 2 − and H 2 O 2 , respectively [22,23], however, several studies have provided evidence for a role of the ArcAB two component system in the resistance to ROS induced damage [12,[24][25][26]. ArcA is essential for S. Enteritidis, Typhimurium and E. coli resistance to ROS [24,26,27]. ArcB is a sensor member of the histidine kinase family that is anchored to the inner membrane [28]. In response to oxygen availability, ArcB autophosphorylates in an ATP dependant intramolecular reaction at position His-292 [29,30] and transfers the phosphate group to the cytoplasmic response regulator ArcA [31][32][33], which binds to promoter regions regulating gene expression [34,35]. ArcB activity is regulated in response to oxygen conditions by the redox state of both the ubiquinone and menaquinone pools [29,[36][37][38]. However, recent studies in E. coli show that the system is regulated by the degree of aerobiosis but not by the redox state of the ubiquinone pool, challenging the idea that the system is inhibited by oxidized quinones [39].
In this work we provide further evidence of the role of the ArcAB two component system in the response to ROS under aerobic conditions and show that this system mediates regulation of ompW expression in response to a novel signal, HOCl. First we demonstrate, both in vivo and in vitro, that OmpW mediates diffusion of H 2 O 2 and HOCl and that exposure of S. Typhimurium to these compounds results in a negative regulation of ompW. By EMSA and using transcriptional fusions, we demonstrate that the global regulator ArcA binds to the ompW promoter region. Furthermore, we show that ompW negative regulation observed in wild type cells treated with H 2 O 2 and HOCl was not retained in an arcA or arcB mutant strain, indicating that the ArcAB two component system mediates ompW negative regulation in response to H 2 O 2 and HOCl. These results further expand our knowledge in both the mechanisms of ROS resistance and the role of ArcAB in this process.

Results and discussion
The OmpW porin facilitates H 2 O 2 and HOCl diffusion through the OM and reconstituted proteoliposomes Hydrogen peroxide and hypochlorous acid are ROS generated by phagocytic cells and in order to enter Gramnegative bacteria they must be able to cross the OM. Even though several biological membranes are permeable to H 2 O 2 , studies in E. coli and S. cerevisiae demonstrate that this compound cannot diffuse freely [9,10]. Additionally, the dielectric properties of H 2 O 2 are comparable to those of water and this compound has a slighter larger dipolar moment, further limiting its diffusion through the OM lipid bilayer. For HOCl, diffusion through the OM is also reported to be limited [11]. Therefore, H 2 O 2 and HOCl must be channeled through the lipid bilayer and one possibility is the influx through porins. We recently demonstrated that the most abundant OM protein in S. Typhimurium, OmpD, allows H 2 O 2 diffusion and is regulated by ArcAB [12]. Little is known about the diffusion of HOCl, but genetic evidence has suggested that in E. coli porins might be used as entry channels for hypothiocyanate ions (OSCN − ), a molecule with a similar chemical structure generated by lactoperoxidase using thiocyanate and H 2 O 2 as an oxidant [40]. In one study, ompC and ompF knockout mutants showed an increased resistance to OSCN − , however, a direct role of porins in mediating HOCl diffusion was not evaluated.
To assess whether OmpW allows the diffusion of H 2 O 2 and HOCl, scopoletin and dihydrorhodamine (DHR)-123 probes, respectively, were used to measure uptake of both toxic compounds separately in a wild type, ΔompW and a genetically complemented ΔompW (pBAD-ompW) strain as described in methods. The ΔompW strain showed an increase in extracellular fluorescence levels after exposure to H 2 O 2 and HOCl resulting in higher extra/intracellular ratios (24 and 4-fold, respectively) as compared to the wild type strain, indicating that in the absence of OmpW the influx of both toxic compounds is decreased. Genetic complementation of ΔompW resulted in nearly identical levels of both extra and intracellular fluorescence as those observed in the wild type strain, suggesting that OmpW is necessary for H 2 O 2 and HOCl uptake ( Figure 1A and C). Even though OmpW appears as a direct responsible for the influx of the compounds, a pleiotropic effect cannot be ruled out at this point because the absence of OmpW in the mutant strain could be producing a remodeling of the membrane organization.
To establish a direct contribution of OmpW in H 2 O 2 and HOCl transport, we used reconstituted proteoliposomes. OmpW-proteoliposomes showed a decrease in H 2 O 2 and HOCl extra/intraliposomal ratios (3.5 and 5-fold respectively) when compared to free liposomes ( Figure 1B and D). Proteoliposomes with S. Typhimurium OmpA porin were used as a negative control as previously described [12]. As expected, OmpA-proteoliposomes showed similar levels to those of free liposomes, indicating that OmpW facilitates H 2 O 2 and HOCl uptake.
Since OmpW channels both toxic compounds across the lipid bilayer, we hypothesized that a ΔompW strain should be more resistant to both toxic compounds when compared to the wild type strain. As shown in Figure 2, exposure of ΔompW to H 2 O 2 4 mM or HOCl 5 mM resulted in an increase in the number of colony forming units (CFU) after 60 min of treatment. However, at longer periods the CFU count between strains 14028s and ΔompW was similar. At 30 min post-treatment with either of the toxic compounds, strain ΔompW showed an increase from 1×10 6 CFU/ml to approximately 6×10 7 CFU/ml. In contrast, the CFU/ml count for strain 14028s remained almost unaltered at 1×10 6 , resulting in a 1.5-log 10 -fold increase in growth for ΔompW. A similar result was observed after 60 min of treatment where the ompW mutant strain showed an increase from 6×10 7 to 1.5×10 9 CFU/ml while the wild type strain changed from 1×10 6 to 8×10 7 CFU/ml. Our results suggest that the absence of OmpW in the mutant strain represents an advantage at short time points due to a decreased permeability towards both H 2 O 2 and HOCl. At longer periods, OM permeability should be reduced because exposure to both toxic compounds results in a negative regulation of S. Typhimurium porins including OmpD, OmpC and OmpF [12,21]. One important possibility that cannot be ruled out at this time is that in the ΔompW strain, the expression of other porins or the OM lipid composition might be altered, therefore changing OM permeability. For example, a study conducted in E. coli showed that an ompC knockout mutant had increased levels of OmpA [40], however, changes in permeability were not evaluated. Furthermore, this has not been evaluated in a S. Typhimurium or E. coli ΔompW strain.
Our data supports the proposed model where OmpW allows the influx of small polar molecules, like H 2 O 2 and HOCl. The crystal structure of OmpW from E. coli revealed that the cross-section of the barrel has approximate dimensions of 17 × 12 Å along the length of the barrel and although the interior of the channel has a hydrophobic character, the observed single channel activities shows that polar molecules traverse the barrel [17]. Taken together, these results provide biochemical and genetic evidence indicating that both toxic compounds are channeled through OmpW. From our knowledge, this is the first direct evidence of HOCl diffusion through porins. Furthermore, preliminary analyses indicate that H 2 O 2 and HOCl channeling is common for S. Typhimurium OmpD, OmpC and OmpF porins (unpublished data).

Hydrogen peroxide and hypochlorous acid exposure results in ompW negative regulation
Since the OmpW porin channels H 2 O 2 and HOCl through the OM and exposure to these molecules is detrimental to bacteria, we hypothesized that ompW should be negatively regulated when S. Typhimurium is exposed to H 2 O 2 and HOCl. To study this effect, wild type S. Typhimurium cells were grown to mid-log phase, exposed to H 2 O 2 or HOCl and ompW mRNA levels were measured by qRT-PCR. As seen in Figure 3, exposure to H 2 O 2 and HOCl resulted in lower levels of ompW transcripts (0.27 ± 0.04 and 0.156 ± 0.079, respectively) relative to control untreated cells. In agreement with our results of ompW negative regulation, similar results were observed by Wang et al. (2010) who showed that S. Enteritidis and Typhimurium cells exposed to HOCl results in modulation of ompD, ompC, ompF (negatively) and ompA (positively) expression. Furthermore, Calderón et al. (2011) demonstrated that the S. Typhimurium ompD gene is negatively regulated in response to H 2 O 2 . Therefore, our and all the published data suggest that in the presence of OClor H 2 O 2 there might be a general lowering in the concentration of porins in the outer membrane, in order to diminish the permeability. To assess the specificity of our assay, we evaluated ompD, ompC and arcB transcript levels as positive (ompD and ompC) and negative controls (arcB). The arcB gene was used as a negative control based on our microarray analysis which shows that it remains unaltered under these conditions and between strains 1408s and ΔarcA (unpublished data). Our results indicate that after exposure to both toxic compounds, arcB transcript levels remain unchanged while those of ompD and ompC are lowered as compared to untreated cells ( Figure 3). Therefore, all the evidence indicates that OM permeability is tightly regulated in response to ROS and could represent a novel mechanism of resistance when bacteria are exposed to these toxic compounds.

ArcA binds the ompW promoter region
In addition to the soxRS and oxyR systems, several studies have provided evidence that the ArcAB two component system plays an important role in the resistance to  ROS induced damage. For example, ArcA is essential for S. Enteritidis and Typhimurium resistance to ROS [24,27] and E. coli mutant strains of the sensor ArcB and the regulator ArcA, show an increased susceptibility to H 2 O 2 [26]. However, neither of these studies identified genes directly regulated by the system under oxidative stress. We recently demonstrated that ArcA negatively regulates the expression of S. Typhimurium ompD after H 2 O 2 exposure by direct interaction with its promoter region [12]. To determine if ArcA mediates ompW down-regulation in response to H 2 O 2 and HOCl, a search for putative ArcA binding sites at the ompW promoter region was performed using Virtual Footprint 3.0 [41]. The analysis predicted the presence of three ArcA binding sites (ABS) located at positions −61 to −70 (ABS-1, forward orientation), -230 to −239 (ABS-2) and −286 to −295 (ABS-3, both in reverse orientation) relative to the experimentally determined transcription start site [42]. Comparison with the extended core region 5′-GTTAATTAAATGTTA-3′ described by Evans et al. (2011) further revealed that only ABS-1 presented a high degree of identity (14 out of 15 nucleotides) with the consensus sequence. To confirm or rule out a direct interaction between ArcA and the predicted binding sites, deletions of the promoter region were generated by PCR (schematized in Figure 4B) and used to perform non-radioactive EMSAs with ArcA and phosphorylated ArcA (ArcA-P). The purity of the protein was assessed by PAGE and ArcA was the dominant product. Electrophoretic mobility shift with ArcA-P was only observed when incubated with fragments that included ABS-1 ( Figure 4C and D, W1 and W4). No shifts were observed in fragments that include both ABS-2 and ABS-3 (W3, even at threefold excess) or control fragments that did not include any ABS (W2 and W5). Non-phosphorylated ArcA only generated electrophoretic mobility shifts at higher concentrations (over 1200 nM) where the negative controls were also retarded as a result of non-specific binding ( Figure 3E). Taken together our bioinformatic and EMSA analyses indicate that ArcA-P binds to the ompW promoter region at a site located between positions −80 and -41 and suggests that this site is ABS-1 which is located between positions −70 to −55.
Evaluating ArcA binding site 1 (ABS-1) functionality To further confirm that ABS-1 ( Figure 4A) was the functional ArcA binding site mediating ompW negative regulation in response to ROS, we constructed transcriptional fusions of the ompW promoter region. We generated two different fusions which included the whole promoter from positions +1 to −600, with respect to the translation start site. One construction contained the native promoter (pompW-lacZ) while substitutions that mutated ABS-1 (shown in red and underlined,  Figure 5A) were included in the second construction (pompW/ABS1-lacZ). The constructions were transformed into the wild type strain and β-galactosidase activity was measured in response to treatment with H 2 O 2 and HOCl.
The activity of the constructions was compared to the untreated 14028s strain with the wild type fusion. Treatment of this strain with H 2 O 2 and HOCl resulted in lower activity levels (0.58 ± 0.008 and 0.53 ± 0.095, respectively), in agreement with qRT-PCR experiments. However, a 5 nucleotide substitution of the most conserved residues at ABS-1 site (pompW/ABS1-lacZ) resulted in no regulation after exposure to either of the toxic compounds (1,09 ± 0.104 and 0,93 ± 0.061), indicating that they are relevant for the transcriptional activity of ompW in response to H 2 O 2 and HOCl ( Figure 5B). Furthermore, these results are in agreement with EMSAs which indicate that ArcA only binds to fragments containing ABS-1.

The ArcAB two component system mediates ompW negative regulation
To establish a direct relationship between ompW negative regulation and ArcA-P binding to its promoter region, ompW expression was evaluated by qRT-PCR in a ΔarcA strain exposed to H 2 O 2 and HOCl. The negative regulation observed in the wild type strain was not retained in an arcA mutant treated with either of the toxic compounds and ompW transcript levels were similar as those observed in untreated cells. Genetic complementation of ΔarcA restored the negative regulation observed in wild type cells exhibiting lower ompW mRNA levels (0.161 ± 0.068 and 0.488 ± 0.027, respectively) as compared to untreated cells ( Figure 6A and C). Growth of the genetically complemented strain in the presence of glucose (non-induction) resulted in similar ompW mRNA levels between treated and untreated cells (data not shown). As controls, we measured ompD, ompC and arcB transcript levels after exposure to H 2 O 2 and HOCl in a ΔarcA strain. Transcript levels of ompD were measured since its expression is regulated by ArcA under ROS conditions [12]. Our results indicate that neither ompD or arcB transcript levels were decreased after exposure to H 2 O 2 or HOCl while those of ompC remained regulated in a ΔarcA strain treated with either of the toxic compounds ( Figure 6A), confirming that ArcA mediates ompD regulation under ROS conditions and showing that the expression of ompC is ArcA independent and regulated by different mechanisms which remain unsolved to the date, and are under study in our laboratory. Furthermore, our bioinformatic analyses in search for ArcA motifs predicted binding sites in the promoter regions of ompW and ompD, but not for ompC ( [12], data not shown).
To determine whether the negative regulation by ArcA was dependant on its cognate sensor ArcB, ompW mRNA levels were evaluated in a ΔarcB strain. In contrast to the negative regulation observed in wild type cells, ompW mRNA levels were further increased in a ΔarcB strain after exposure to HOCl (3.37 ± 1.09). Transcript levels after treatment with H 2 O 2 were similar as those observed in untreated cells ( Figure 6B). One possibility for this result is that in the absence of ArcA, ArcB might phosphorylate (i.e ArcB-OmpR, [43]) one or more response regulators, either unspecifically or due to crosstalk, which could bind to the promoter region and therefore prevent binding of positive regulators like SoxS, which has been demonstrated to regulate ompW and is up-regulated in response to HOCl [20,44]. This could result in constant ompW transcript levels as shown in Figure 6A. On the other hand, in the absence of ArcB no phosphorylation occurs and SoxS or other positive regulator(s) might have free accessibility to the ompW promoter and therefore increase its expression ( Figure 6B), although this possibility has not been evaluated in this study. Genetic complementation of ΔarcB restored the negative regulation observed in wild type cells exposed to H 2 O 2 and HOCl (0.19 ± 0.04 and 0.24 ± 0.11, respectively, Figure 6C). The ompD and ompC transcripts levels remained down-regulated after exposure to H 2 O 2 and HOCl in the ΔarcB strain, while the negative control arcA remained unaltered ( Figure 6B).
The ArcA regulon in anaerobically grown S. Typhimurium was recently determined [27]. Interestingly, neither ompD nor ompW expression was down-regulated in an ArcA dependant manner, suggesting that the ArcA regulon under anaerobic and aerobic ROS conditions could be different. Even in E. coli, ompW expression is suggested to be regulated by FNR in response to oxygen availability [39]. The difference between the ArcA regulons under aerobic and ROS conditions might be explained by studies suggesting that the mechanism of ArcA activation under aerobic conditions is different from those classically described. E. coli mutant strains in residue H-717 of ArcB are able to phosphorylate and activate ArcA through the transfer of the phosphate group from residue His-292 under aerobic conditions [45] and Loui et al. (2009) suggested that H 2 O 2 resistance is independent of ArcA phosphorylation at residue Asp-54. To the date, the detailed molecular mechanism of ArcAB activation in response to ROS remains unsolved. Therefore, further experiments to unveil the molecular mechanism by which the S. Typhimurium ArcAB two component system is activated are needed and under way in our laboratory.

Conclusion
We provide both genetic and biochemical evidence indicating that the OM porin OmpW mediates the influx of H 2 O 2 and HOCl. The results revealed that the S. Typhimurium ompW gene is negatively regulated upon exposure to both toxic compounds. Furthermore, we demonstrate that the response regulator ArcA mediates ompW negative regulation in response to H 2 O 2 and HOCl via a direct interaction with the upstream region of ompW. Taken together, with our previous observation that OmpD mediates influx of H 2 O 2 and is negatively regulated by ArcA in response to H 2 O 2 , these results further expand our knowledge regarding the coordinated regulatory mechanisms of ROS resistance and the role of ArcAB in this process.

Bacterial strains and growth conditions
Bacterial strains used in this work are listed in Table 1. Cells were grown aerobically with agitation in LB medium at 37°C. Solid media consisted of agar (20 g l −1 ) and plates were incubated at 37°C. Dilutions (1:100) of overnight cultures were used to initiate growth. When necessary, growth media was supplemented with the appropriate antibiotics (see below).

Bacterial survival after exposure to oxidative stress
Bacteria were cultured in 5 ml of LB medium at 37°C overnight with shaking. Antibiotics were added as appropriate. 1:1000 dilutions of the overnight cultures were grown in 25 ml to OD~0.4 and H 2 O 2 4 mM or NaOCl 5 mM (final concentration) were added. In all the assays the cultures were grown aerobically at 250 rpm. Aliquots  [12] of cultures were withdrawn at the different time points, diluted and plated in triplicate. Bacterial cultures were enumerated by counting the number of CFU after overnight incubation to determine the bacterial concentrations.

Construction of transcriptional fusions with reporter gene lacZ
The native ompW promoter region from positions +1 to −600 (with respect to the translation start) site was amplified by PCR with primers ompW_pLacZ_-600F_ATG 5′ CGGGGTACCCCCGATATCGAAAATTCGCG 3′ and ompW_pLacZ_-1R_ATG 5′ CCCAAGCTTACCCGCTC-CATCGTTATGGT 3′ using genomic DNA from S. Typhimurium (strain 14028s). The restriction sites (KpnI and HindIII, respectively) at the ends of the DNA fragment were introduced by the PCR primers (underlined sequences) and digested with the corresponding enzymes. The digested PCR product was cloned into the multiple cloning site (MCS) of the β-galactosidase reporter vector pLacZ-Basic (GenBank accession no. U13184), Clontech, generating plasmid pompW-lacZ. To generate plasmid pompW/ABS1-lacZ, primers ompW_pLacZ_-600F_ATG with Mut_sit_arcAR 5′ TGTTCTTATAATGCGGAATT-TATTGATCCAG 3′ and ompW_pLacZ_-1R_ATG with Mut_sit_arcAF 5′ CTGGATCAATAAATTCCGGAAT-TATAAGAACA 3′ were used to generate overlapping PCR products spanning the whole length of the ompW promoter. Mutation of ABS-1 was generated by incorporating substitutions in primers Mut_sit_arcAF and Mut_sit_arcAR (underlined sequences). The resulting PCR products were used as templates in a second reaction with primers ompW_pLacZ_-600F_ATG and ompW_pLacZ_-1R_ATG to generate the mutated ompW promoter, which was digested and cloned into the MCS of plasmid pLacZ-Basic. Constructions were confirmed by DNA sequencing. The generated constructs were transformed into wild type strain 14028s.
To evaluate activity, cells at OD 600~0 .4 were grown for 20 min in the presence of H 2 O 2 (1.5 mM) or NaOCl (530 μM). Control cells received no treatment. β-galactosidase activity was determined as previously described [20].

Protein purification
His-tagged ArcA used in EMSAs was purified as previously described [12]. Briefly E. coli BL21 cells harboring plasmid pET-TOPO-arcA were grown in 500 ml of LB medium supplemented with amplicillin (100 μg ml −1 ) to OD 600~0 .4 and protein overexpression was carried out by adding 1 mM IPTG and further growth for 6 h. Protein was purified by affinity chromatography as described by Georgellis et al., (1997).
Outer membrane proteins used in proteoliposomes were purified as described by Calderón et al. (2011). E. coli Top10 cells carrying pBAD-ompA or pBAD-ompW were grown in 500 ml to OD 600~0 .6 at 37°C and overexpression was performed for 5 h in the presence of 1 mM arabinose. His-tagged porins were purified by affinity chromatography using HisTrap HP columns (Amersham) according to the manufacturer's instructions.

RNA isolation and ompW mRNA detection
Overnight cultures were diluted (1:100) and cells were grown to OD 600~0 .4. Genetically complemented cells (ΔarcA/pBAD-arcA and ΔarcB/pBAD-arcB) were grown in the presence of arabinose (1 mM) and ampicillin (100 μg ml -1 ). At this point, H 2 O 2 (1.5 mM) or NaOCl (530 μM) was added and cells were grown for 20 min. Control cells received no treatment. After exposure to the toxic compounds, 4 ml were withdrawn from the culture and used to extract total RNA using GenElute Total RNA purification Kit® (Sigma). Total RNA treatment with DNase I and cDNA synthesis was performed as previously described [19].
Relative quantification of ompW mRNA was performed using Brilliant II SYBR Green QPCR Master Reagent Kit and the Mx3000P detection system (Stratagene). 16S rRNA was used for normalization. Specific primers were 5′ ATGAAAAAATTTACAGTGG 3′ (RTompWF) and 5′ GAAACGATAGCCTGCCGA 3′ (RTompWR) for the ompW gene; 5′ GTAGAATTCCAGGTGTAGCG 3′ (16SF) and 5′ TTATCACTGGCAGTCTCCTT 3′ (16SR) for 16S rRNA gene (16S). The reaction mixture was carried out in a final volume of 20 μl containing 1 μl of diluted cDNA (1:1000), 0.24 μl of each primer (120 nM), 10 μl of 10 x Master Mix, 0.14 μl of diluted ROX (1:200) and 8.38 μl of H 2 O. The reaction was performed under the following conditions: 10 min at 95°C followed by 40 cycles of 30 s at 95°C, 30 s at 53°C and 45 s at 72°C. Finally a melting cycle from 53 to 95°C was performed to check for amplification specificity. Amplification efficiency was calculated from a standard curve constructed by amplifying serial dilutions of RT-PCR products for each gene. These values were used to obtain the fold change in expression for the gene of interest normalized with 16S levels according to [47]. Experiments were performed in three biological and technical replicates.

In vivo and in vitro determination of hydrogen peroxide and hypochlorous acid diffusion
In vivo diffusion of H 2 O 2 was assessed as previously described [12]. For HOCl detection, overnight cultures were diluted and cells were grown to OD 600~0 .5. Two ml of bacterial culture were centrifuged for 5 min at 4500 x g and resuspended in 1 ml of 100 mM phosphate buffer (pH 7.2). A 200 μl aliquot was incubated for 5 min with 530 μM NaOCl and constant agitation. Following, cells were vacuum filtered using polycarbonate filters of 0.025 μm (Millipore) and pass through was collected (extracellular fraction). Bacteria retained in the filter were recovered with 1 ml of 50 mM phosphate buffer (pH 7.2) and disrupted by sonication (intracellular fraction). Both fractions (190 μl) were incubated separately with dihydrorhodamine-123 to a final concentration of 5 μM as previously described [49]. The fluorescent product, rhodamine-123, was measured by fluorescence detection with excitation and emission wavelengths of 500 and 536 nm, respectively. HOCl and H 2 O 2 uptake was determined as the extracellular/intracellular fluorescence ratio. The background fluorescence from a bacterial suspension not exposed to either of the toxic compounds was subtracted and results were normalized by protein concentration.
Proteoliposomes were prepared as described [50] with modifications [51]. For in vitro diffusion, proteoliposomes were incubated with 1.5 mM H 2 O 2 or 530 μM NaOCl for 5 min, vacuum filtered and pass through was recovered (extraliposomal fraction). Proteoliposomes were recovered from the filters with 2 ml of 50 mM phosphate buffer (pH 7.2) and disrupted by sonication (intraliposomal fraction).
Fluorescence was measured in both fractions as described above and H 2 O 2 or HOCl uptake was determined as the extraliposomal/intraliposomal fluorescence ratio.