|Home | About | Journals | Submit | Contact Us | Français|
Oxidative stress is one of the main challenges bacteria must cope with during infection. Here, we identify a new oxidative stress sensing and response ospR (oxidative stress response and pigment production Regulator) gene in Pseudomonas aeruginosa. Deletion of ospR leads to a significant induction in H2O2 resistance. This effect is mediated by de-repression of PA2826, which lies immediately upstream of ospR and encodes a glutathione peroxidase. Constitutive expression of ospR alters pigment production and β-lactam resistance in P. aeruginosa via a PA2826-independent manner. We further discovered that OspR regulates additional genes involved in quorum sensing and tyrosine metabolism. These regulatory effects are redox-mediated as addition of H2O2 or cumene hydroperoxide leads to the dissociation of OspR from promoter DNA. A conserved Cys residue, Cys-24, plays the major role of oxidative stress sensing in OspR. The serine substitution mutant of Cys-24 is less susceptible to oxidation in vitro and exhibits altered pigmentation and β-lactam resistance. Lastly, we show that an ospR null mutant strain displays a greater capacity for dissemination than wild-type MPAO1 strain in a murine model of acute pneumonia. Thus, OspR is a global regulator that senses oxidative stress and regulates multiple pathways to enhance the survival of P. aeruginosa inside host.
Pseudomonas aeruginosa is a Gram-negative bacillus that is ubiquitous in diverse environments. It is also an opportunistic pathogen of humans, most notably those afflicted with cystic fibrosis or whose immune systems have been compromised (Reynolds et al., 1975; Cross et al., 1983; Govan and Harris, 1986). Similar to many other human pathogens, P. aeruginosa must overcome the oxidative stress response generated by phagocytic cells for successful infection. Phagocytes utilize the cytotoxic effects of reactive oxygen species (ROS), such as superoxide, H2O2, and hydroxyl radical, to contain bacterial infections. In order to counter this innate immune response, P. aeruginosa possesses a multifaceted defence system against oxidative stress, with proteins such as catalase and superoxide dismutase, thioredoxin, glutaredoxin, and small molecules such as glutathione and melanin (Hassett and Cohen, 1989; Scandalios, 1997; Rodríguez-Rojas et al., 2009).
As key regulators, OxyR and SoxR have been well known to modulate oxidative stress response in bacteria (Storz and Imlay, 1999). Homologues of these two proteins have been identified in many bacterial species, including P. aeruginosa. In P. aeruginosa, OxyR is important for oxidative stress defence (Ochsner et al., 2000) and is required for full virulence in rodent and insect models of infection (Lau et al., 2005). Recently, OxyR has also been shown to regulate secretion of potent cytotoxic factors in P. aeruginosa (Melstrom et al., 2007). On the contrary, P. aeruginosa does not rely on SoxR for an oxidative stress response (Kobayashi and Tagawa, 2004; Palma et al., 2005). Instead, P. aeruginosa SoxR responds to phenazines (Dietrich et al., 2006), which act as signalling molecules in the bacterium (Dietrich et al., 2006). Phenazines produced by P. aeruginosa are colourful, redox-active antibiotics. These compounds have profound effects on the structural organization of colony biofilms (Dietrich et al., 2009) and have been identified as virulence factors in a number of in vivo model systems (Wilson et al., 1988; Mahajan-Miklos et al., 1999; Ran et al., 2003; Gibson et al., 2009). A connection between phenazine biosynthesis and oxidative stress response in P. aeruginosa is as yet unclear.
In our previous work, we showed that the MarR family transcriptional regulators MgrA and SarZ play key roles in virulence regulation in Staphylococcus aureus using an oxidation-sensing mechanism (Chen et al., 2006; 2008a). Both MgrA and SarZ are members of a subfamily of MarR proteins that utilize cysteine oxidation to sense oxidative stress and regulate bacterial responses. The prototype, OhrR in Bacillus subtilis, regulates bacterial resistance to organic hydroperoxides (Fuangthong et al., 2001; Sukchawalit et al., 2001; Mongkolsuk and Helmann, 2002; Newberry et al., 2007). However, in a pathogenic bacterium such as S. aureus, these regulators seem to play regulatory roles that have much broader and profound effects on global properties of the pathogen. These discoveries in S. aureus raise the possibility that the OhrR/MgrA homologues in pathogenic P. aeruginosa may also assume global roles through sensing oxidative stress. Here, we present a new redox-active regulator, OspR, in P. aeruginosa. This protein, using an oxidation-sensing mechanism, is involved in oxidative stress response, pigment production, β-lactam resistance and dissemination of P. aeruginosa during infection. OspR also affects expression of genes involved in tyrosine metabolism (hmgA, PA2010) and quorum sensing (phzM, phzS, PA1897). These results should help shed light on the multifaceted oxidative stress response in P. aeruginosa and contribute to understanding its role in P. aeruginosa physiology and pathogenesis.
As demonstrated in our previous work, the global regulator MgrA plays a key role in virulence regulation in S. aureus using an oxidation-sensing mechanism (Chen et al., 2006). To identify the MgrA homologues in P. aeruginosa, we performed BLASTP analyses with S. aureus mgrA against the genome of P. aeruginosa PAO1. Two hits were obtained with PA2825 showing 37.59% identity to MgrA while PA2849 sharing 31.25% identity with MgrA (Fig. S1A). Pfam analysis indicated that both PA2825 and PA2849 proteins possess the MarR-type helix–turn–helix motif and as such belong to the family of MarR proteins.
Downstream of PA2849 is PA2850, and there exists a 142 bp intergenic region between PA2849 and PA2850. The encoding protein of PA2850 shows a 78% similarity to Ohr (organic hydroperoxide resistance) protein in Xanthomonas campestris pv. phaseoli. The gene encoding OhrR (transcriptional regulator, MarR family) from X. campestris pv. phaseoli is co-transcribed with its downstream adjacent gene, ohr (Panmanee et al., 2002; Klomsiri et al., 2005). Considering that numerous bacteria maintain this genetic organization of ohrR and ohr (Fuangthong et al., 2001; Sukchawalit et al., 2001; Chuchue et al., 2006), it is very likely that PA2849 is an OhrR homologue in P. aeruginosa.
According to the annotation of P. aeruginosa PAO1 genome, PA2825 is a likely transcriptional regulator gene. The coding regions of PA2825 and PA2826 (a putative glutathione peroxidase gene) overlap by four base pairs; the coding region of PA2825 starts at the fourth nucleotide before the end of PA2826 (Fig. S1B). Sequence analysis of this genetic organization in Pseudomonas species indicates that PA2825 orthologues are present in various P. aeruginosa and P. fluorescens strains but absent in P. putida, P. syringae and P. entomophila L48 strains (Fig. S2). We focused on PA2825 and named it as ospR based on observed phenotypes presented in this study.
To probe the biological functions of ospR, we generated an ospR null mutant strain (ΔospR, Fig. S1B) through the method of allelic replacement as described in Experimental procedures. Further, the complementing plasmid of p-ospR (Table 1) was constructed and then transformed into ΔospR, yielding the ΔospR/p-ospR strain. Introduction of p-ospR into ΔospR provided for constitutive expression of ospR without IPTG induction (data not shown). We performed the Northern blot analysis of mRNA of the ospR neighbouring genes PA2827, PA2826 and PA2824 respectively. The Northern blot results showed that there was no significant difference in mRNA levels of PA2827 and PA2824 between the ΔospR strain (harbouring pAK1900) and the wild-type MPAO1 strain (harbouring pAK1900) (Fig. 1A). However, a significant overexpression of the PA2826 transcript was detected in the ΔospR strain as compared with the wild-type strain (Fig. 1A). Complementation with p-ospR in the ΔospR strain restored the low mRNA level of PA2826 similar to that observed in the wild-type strain (Fig. 1A). These results indicate that OspR represses expression of PA2826. Interestingly, the deletion of ospR in P. aeruginosa MPAO1 also resulted in a slight increase of the mRNA level of PA2850 (ohr) (Fig. S3A).
Deletion of ospR in P. aeruginosa leads to de-repression of PA2826, a gene encoding a glutathione peroxidase (GPx) protein based on the annotation of the P. aeruginosa PAO1 genome and sequence analysis. The main biological role of glutathione peroxidase is to protect the organism from oxidative damage. In order to assess whether ospR plays a role in bacterial response to oxidative stress, we tested the sensitivity of the ΔospR strain to H2O2 and paraquat by using stress plate assay as described in Experimental procedures. From the assay ΔospR strain exhibited an increased resistance to H2O2 (Fig. 1B and Table S1), but showed a hypersensitivity to paraquat when compared with the wild-type MPAO1 strain (Fig. 1C). These phenotypes could be complemented by the introduction of p-ospR (Fig. 1B and C, and Table S1).
To examine whether the altered resistance/sensitivity to H2O2/paraquat is due to the overexpression of PA2826 in the ΔospR strain, a plasmid for constitutive expression of PA2826 (p-PA2826) was constructed and then transformed to the wild-type strain. Additional stress plate assays were performed and results are shown in Fig. 1D. The MPAO1/p-PA2826 strain displayed an increased resistance to H2O2, and an increased sensitivity to paraquat when compared with the control strain MPAO1/pAK1900. The increase of H2O2 resistance is more evident in a strain overexpressing PA2826 (p-PA2826) than the ospR mutant strain (Fig. 1 and Table S1). This phenotype correlates with the transcription levels of PA2826 in these stains as shown by Northern blot analysis (Fig. S3).
To further demonstrate that PA2826 mediates the altered resistance/sensitivity to H2O2/paraquat, we generated a PA2826-ospR null mutant strain (ΔPA2826-ospR), as described in Experimental procedures. Stress plate assay indicated no significant difference in resistance/sensitivity to H2O2/paraquat between the ΔPA2826-ospR strain and the wild-type MPAO1 strain. Moreover, the introduction of p-PA2826 to the ΔPA2826-ospR strain led to increased resistance to H2O2 and increased sensitivity to paraquat when compared with the ΔPA2826-ospR strain harbouring pAK1900 (Fig. S4). These results showed that OspR contributes to the oxidative stress response of P. aeruginosa and this process is mediated through PA2826.
PA2826 encodes a glutathione peroxidase (GPx) that catalyses the reduction of hydroperoxides, including H2O2, by using glutathione, and functions to protect bacterial cells from oxidative damage. To examine this role, we measured the redox state of the GSH/glutathione disulphide (GSSG) couple in MPAO1/pAK1900, ΔospR/pAK1900, ΔospR/p-ospR and MPAO1/p-PA2826 strains, respectively, when bacteria were grown to an early stationary phase (OD600 = 2.2–2.3). As a result, ΔospR/p-ospR displayed a similar GSH/GSSG ratio as that of wild-type MPAO1/pAK1900. However, both ΔospR/pAK1900 and MPAO1/p-PA2826 exhibited a 1.7-fold and 5.5-fold higher intracellular GSH/GSSG ratio compared with that of the wild-type strain.
Genetic evidence indicates that OspR regulates the expression of PA2826 (Fig. 1A). In order to probe the putative OspR binding site, we examined the PA2827–PA2826 intergenic region. An AT-rich inverted repeat sequence, 5′-ttcaatcaagttgtgtgcAATTgAATTgcgagctactt tattt-3′ (B2, uppercase letters indicate the conserved inverted repeat), was found (Fig. 2A). It is well known that the MarR family proteins specifically bind palindromic or pseudopalindromic sites using conserved winged helix fold (Wilkinson and Grove, 2006; Newberry et al., 2007). The inverted repeat overlaps the core promoter elements for PA2826 (Fig. 2A). The direct interaction of OspR with PA2826 promoter was tested by gel mobility shift assay using purified 6His-OspR and a 43 bp DNA fragment (B2) containing the putative OspR binding site. As shown in Fig. 2B, OspR bound to the 43 bp PA2826 promoter sequence in the presence of non-specific salmon sperm DNA, while failed to shift a 43 bp DNA fragment (B1: 5′-agagcgacgacccgtctttggttcgggtcgtttttcattcaat-3′, Fig. 2A) of the Rho-independent terminator sequence as a control (data not shown).
We have shown that OspR contributes to bacterial oxidative stress response through regulating the expression of PA2826. To further assess whether ospR responds to oxidative stress, we treated P. aeruginosa MPAO1 (log phase, OD600 = 1.2–1.5) with different amounts of H2O2 for 10 min. Subsequently, the bacterial cells were harvested and the total RNAs were isolated. Northern blot analysis showed elevated transcripts of PA2826 and ospR when the wide-type bacterial cells were treated with H2O2 as compared with the untreated bacteria (Fig. 2C). The transcription enhancements of PA2826 and ospR are positively correlated with the concentrations of H2O2 applied to the bacterium. Increased transcriptions of PA2826 mRNA were also observed when P. aeruginosa MPAO1 was treated with paraquat. As shown in Fig. 2D, the expression of PA2826 is significantly induced by 0.05 µM paraquat in Luria–Bertani (LB) medium; however, unlike the situation with H2O2, higher concentrations of paraquat failed to further induce the expression of PA2826. We currently do not have an explanation for this observation with the paraquat treatment. Since increased transcription of PA2826 is also detected in the ΔospR strain (Fig. 1A), it is reasonable to suggest that OspR represses PA2826-ospR, and oxidation with H2O2 leads to dissociation of OspR from the promoter DNA and de-repression of PA2826 in P. aeruginosa. To confirm this hypothesis, we performed gel mobility shift assay with 6His-OspR and the same 43 bp promoter DNA fragment (B2) used in Fig. 2B. Addition of H2O2 or cumene hydroperoxide (CHP) led to dissociation of OspR from the DNA (Fig. 2E). In addition, the binding of OspR to the promoter DNA could be restored by the addition of excess reducing agent (DTT) (Fig. 2E).
Unexpectedly, the introduction of p-ospR into ΔospR in the complementary experiments described above also led to the production of dark red, water-soluble pigments when bacterial cells were grown on the LB agar plate (Fig. 3A). We were intrigued by this observation which suggests that OspR may have a broader role of regulation. We introduced p-ospR into the wild-type MPAO1 strain (MPAO1/p-ospR). MPAO1/p-ospR displayed the same dark red pigmentation on the LB agar plate (Fig. 3B) as observed with ΔospR/p-ospR. This phenotype is independent of PA2826 as MPAO1/p-PA2826 exhibits the wild-type pigmentation (Fig. 3B), and ΔPA2826-ospR/pAK1900 and ΔPA2826-ospR/p-PA2826 show the same normal pigmentation as MPAO1/pAK1900 while ΔPA2826-ospR/p-ospR displays the dark red pigment like the strain with constitutive expression of ospR when grown on the LB plate (Fig. S5A).
A previous study with mutation of two P. aeruginosa phenazine-modifying genes, phzM and phzS, led to accumulation of yellow and dark red phenazines (Mavrodi et al., 2001). To examine whether ospR affects the expression of these two P. aeruginosa phenazine-modifying genes, we performed the Northern blot analysis of mRNAs of phzM and phzS. As shown in Fig. 3C, the constitutive expression of ospR severely reduces the expression of both phzM and phzS in the MPAO1 strain. Consistent with this, the mRNA levels of phzM and phzS (Fig. 3D) were increased in the ΔospR strain while further reduced in the ΔospR strain complemented with p-ospR as compared with the wild-type strain (Fig. 3D).
The OhrR/MgrA family proteins are known to control oxidative stress response, bacterial virulence and antibiotic resistance (Fuangthong et al., 2001; Sukchawalit et al., 2001; Chen et al., 2006; 2008b). To assess whether OspR contributes to drug resistance in P. aeruginosa, we examined bacterial growth on TSA plates supplied with chloramphenicol, ciprofloxacin and cefotaxime respectively. Results showed that ΔospR did not exhibit noticeable difference from the wild-type strain (data not shown) while MPAO1/p-ospR displayed higher resistance to cefotaxime (Fig. 3E) and ceftazidime (data not shown) when compared with the wild-type strain. This cefotaxime resistance phenotype is not mediated through PA2826 as ΔPA2826-ospR/p-ospR showed the same resistance to cefotaxime as MPAO1/p-ospR (data not shown), and ΔPA2826-ospR/p-ospR displays enhanced resistance to cefotaxime when compared with ΔPA2826-ospR/pAK1900 (Fig. S5B).
OspR harbours two cysteine residues, Cys-24 and Cys-134 (Fig. S1). To investigate the contributions of these two residues to the function of OspR, each residue was mutated to serine. The ospR C24S mutant (in pAK1900, p-ospRC24S), ospR C134S mutant (in pAK1900, p-ospRC134S) and ospR C24S/C134S double mutant (in pAK1900, p-ospRC24SC134S) were prepared and introduced into the wild-type MPAO1 strain respectively. The introduction of p-ospRC24S, p-ospRC134S or p-ospRC24SC134S into the wild-type MPAO1 strain led to the production of the dark red, water-soluble pigments, a phenotype similar to that observed in MPAO1/p-ospR when bacteria were grown on LB agar plate (Fig. S6). In liquid culture, the MPAO1/p-ospR strain displayed the usual blue-green pigmentation (Fig. 4A). However, both MPAO1/p-ospRC24S and MPAO1/p-ospRC24SC134S strains exhibited dark red pigmentation, while a yellow-green pigmentation was observed for the MPAO1/p-ospRC134S strain grown in LB liquid (Fig. 4A). This observation indicates that Cys-24 plays the key role in the regulatory function of OspR. We further examined β-lactam resistance of the two cysteine mutant strains. As shown in Fig. 4B, both OspR C24S mutant and OspR C24SC134S mutant strains showed enhanced β-lactam resistance but OspR C134S mutant showed normal sensitivity to β-lactam as the control strain (MPAO1/p-ospR).
To further investigate the role of Cys-24 in the regulatory function of OspR, we introduced p-ospRC24S into the ospR mutant strain and performed Northern blot analysis of transcripts of PA2826 and PA2850 (ohr). As shown in Fig. 4C, elevated transcriptions of PA2826 and ohr were observed when the wide-type bacterial cells were treated with H2O2 as compared with the untreated bacteria, and this observation is consistent with previous transcriptional profiling result (Palma et al., 2004). There was no significant increase of the mRNA level of PA2826 in the ΔospR/p-ospR strain treated with H2O2 as compared with the untreated bacteria; however, a decrease of the mRNA level of PA2826 was observed in the ΔospR/p-ospRC24S strain treated with H2O2 compared with that of the untreated bacteria (Fig. 4D). Interestingly, an increase of ohr transcription was observed in the ΔospR/p-ospRC24S strain treated with H2O2, but to a much less extent as compared with those of the ΔospR/p-ospR and the wild-type strains treated with H2O2 (Fig. 4D). Thus, OspR also affects ohr and Cys-24 plays a major role in the regulatory function of OspR.
Next, we purified 6His-OspR C24S mutant protein and tested its interaction with the PA2826 promoter (43 bpDNA fragment, B2) using gel mobility shift assay. This mutant protein binds the promoter DNA tighter than the wild-type OspR (Fig. 5A). Importantly, treatment with either H2O2 or CHP failed to dissociate OspRC24S from the 43 bp DNA fragment (B2) (Fig. 5B), indicating that Cys-24 is the key residue involved in the oxidant sensing. Not surprisingly, Cys-24 is the conserved cysteine used for oxidant sensing in MgrA, SarZ and OhrR (Fuangthong et al., 2001; Sukchawalit et al., 2001; Chen et al., 2006; 2008a). Residues that form the hydrogen-bonding network (Tyr) with the redox-active cysteine and the nearby hydrophobic pocket (Tyr/Phe) are also conserved in OspR (Fig. S1).
Our next step was to confirm the redox activity of Cys-24 in OspR. We performed experiments to detect the potential formation of Cys-24-SOH in the wild-type OspR following CHP treatment by using NBD chloride (Chen et al., 2006; 2008a; Panmanee et al., 2006). However, this assay failed to yield an R-S(O)-NBD derivative in the CHP-oxidized OspR (data not shown). We suspected that this could be due to a rapid reaction of the generated sulphenic acid intermediate (R-SOH) with the second Cys residue, Cys-134, in OspR. We used non-reducing sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) gel to investigate whether intermonomer disulphides are formed in the oxidized OspR and OspRC24S in vitro (Panmanee et al., 2006; Chen et al., 2008a). Indeed, upon CHP treatment, OspR was oxidized to give a covalent dimer (Fig. 5C). Treatment with DTT prior to electrophoresis converted the oxidized protein back to its monomeric form, indicating a disulphide linkage between the two monomers in the oxidized OspR (Fig. 5C). We note that two distinct bands appeared in the size range corresponding to dimeric OspR (Fig. 5C). Formation of the lower dimeric protein band was dependent on the addition of the oxidant and the presence of Cys-24, while formation of the upper band did not require the presence of Cys-24 (Fig. 5C). The upper dimeric protein band could be a single-disulphide cross-link between two Cys-134 residues in the two monomers, whereas the lower dimeric band is doubly cross-linked OspR between Cys-24 from one monomer and Cys-134 from the other monomer (Soonsanga et al., 2008). To test if Cys-24-SOH is produced upon oxidation and subsequently trapped by Cys-134, the NBD chloride trapping experiment was repeated using CHP-oxidized OspRC134S. The UV-visible absorbance spectrum of NBD-labelled, oxidized OspRC134S exhibited a maximal absorbance at 347 nm, indicating formation of a sulphenic acid species in the oxidized OspRC134S protein (Fig. 5D).
Six additional potential OspR binding targets (Table 1), corresponding to the promoter regions of 10 genes, were identified by a search in the P. aeruginosa genome sequence for the putative conserved OspR binding motif (AATTnAATT) located upstream (−1 bp to −400 bp) of the coding region by using RAST (http://rsat.ulb.ac.be/rsat/index.html). Interestingly, the promoter region of PA1897 and the hmgA(PA2009)–PA2010 intergenic region were found as a potential binding site for OspR (Table 1 and Fig. S7). While PA1897 is a quorum-sensing-regulated gene, hmgA encodes an enzyme homogentisate-1,2-dioxygenase involved in tyrosine metabolism. Mutation of hmgA yielded strains with the hyperproduction of a dark-brown pyomelanin pigment in various Pseudomonas species including P. aeruginosa, P. chlororaphis O6 and P. putida (Arias-Barrau et al., 2004; Kang et al., 2008; Rodríguez-Rojas et al., 2009). An adjacent but oppositely transcribed gene PA2010 (hmgR) was proposed to encode a regulator that induces hmgA expression when homogentisate is present (Arias-Barrau et al., 2004). We tested whether OspR binds to these two promoter regions using gel mobility shift assays. As shown in Fig. 6A, OspR could shift the 70 bp hmgA(PA2009)–PA2010 intergenic region sequence (hmgA-p: 5′-tggcacgctggctgcgttatt tttatcgtAATTcAATTacgcataacgtaatttgagtggaaggcagcgt-3′) (uppercase letters indicate the conserved inverted repeat), and the 70 bp PA1897 promoter sequence (PA1897-p: 5′-ggcaggttgtccctgccgggctgtgacAATTtAATTc gaccaggcatttcattgtccgtgccgattttca-3′) (uppercase letters indicate the conserved inverted repeat), respectively, while OspR failed to shift a 43 bp control DNA fragment (B2). To test whether the AATTnAATT direct repeat is required for the binding of OspR to the B2 DNA fragment, we performed the gel shift assay using a 34 bp B2d DNA fragment (B2 lacking AATTnAATT, 5′-ttcaatcaagttgtgtg cgcgagctactttattt-3′) and B2 DNA fragment as a control. As shown in Fig. 6B, OspR failed to bind to B2d DNA lacking the AT-rich repeat. Interestingly, the AT-rich inverted repeat sequence, 5′-AATTnAATT-3′, is similar to the putative OhrR box sequence thought to be involved in the binding of OhrR to the target promoters in B. subtilis (Fuangthong et al., 2001), X. campestris (Sukchawalit et al., 2001) and Agrobacterium tumefaciens (Chuchue et al., 2006). The role of this AATTnAATT motif in regulation awaits further study.
To further examine if ospR affects the expression of PA1897, hmgA and PA2010, we performed the Northern blot analysis of mRNAs of these three genes in MAPO1 (harbouring pAK1900), ΔospR (harbouring pAK1900) and ΔospR/p-ospR respectively. The result showed that deletion of ospR led to increased expression of PA1897 while decreased expression of hmgA and PA2010 as compared with those of the wild-type strain when bacteria were grown in early stationary phase (OD600 = 2.5). Complementation with p-ospR in ΔospR restored the mRNA levels of hmgA, PA2010 and PA1897 to normal levels observed in the wild-type MPAO1 strain (Fig. 6C).
The in vitro results presented above suggest that ospR plays global roles in oxidative stress response, pigment production and antibiotic resistance. We tested whether ospR also impacts the virulence in a mouse model of acute pneumonia. C57BL/6 mice were infected intranasally with approximately 1 × 107 wild-type bacteria (MPAO1/pAK1900), ospR null mutant bacteria (ΔospR/pAK1900) or ospR null mutant bacteria expressing ospR in pAK1900 (ΔospR/p-ospR). Figure 7 shows the ratio of bacteria recovered from the lungs and spleens relative to the initial inoculum at 18 h post infection, with geometric means indicated for each group. In this assay, the wild-type MPAO1/pAK1900 was recovered in numbers approximately at 0.008% and 0.00001% of initial inoculum dose from lungs and spleens respectively. In marked contrast, ΔospR/pAK1900 bacteria were recovered in numbers equal to approximately 0.08% of the inoculum dose from lungs and 0.0005% of the inoculum dose from spleens. These differences achieve statistical significance. Further, bacteria of the complementary strain (ΔospR/p-ospR) were recovered from lungs or spleens with a similar recovery ratio to that of the wild-type MPAO1/pAK1900 strain. These results indicate that ospR has an impact on the capacity for dissemination in this model. Mutation of ospR leads to an increased bacterial virulence.
Reactive oxygen species were originally considered to be exclusively detrimental to cells; however, redox regulation involving ROS is now recognized as a vital component to cellular signalling and regulation (Scandalios, 1997; Imlay, 2008; Poole and Nelson, 2008). In this study, we identified a new gene, ospR, which encodes a MarR family protein in P. aeruginosa. OspR is a functional homologue of the bacterial OhrR/MgrA family oxidative stress sensing and regulatory proteins. OspR binds to the promoter of PA2826, which encodes a glutathione peroxidase, and represses the expression of PA2826 and likely itself. Additional sites (hmgA–PA2010 intergenic region and PA1897 promoter region) in the P. aeruginosa genome may be recognized by OspR. OspR may bind to some of these sites and exert regulatory functions. It is still unclear if OspR can act as a direct transcriptional activator. A cysteine residue, Cys-24, is used by OspR to sense potential oxidative stress and regulates bacterial response. Oxidation of Cys-24 in OspR leads to the dissociation of the protein from promoter DNA. OspR is also involved in pigment production, impacts β-lactam resistance and affects dissemination during infection. Thus, it is a global regulator that controls multiple pathways in P. aeruginosa.
Orthologues of ospR are present in various P. aeruginosa and P. fluorescens strains while absent in some other Pseudomonas species such as P. putida, P. syringae and P. entomophila L48. This gene in P. aeruginosa and P. fluorescens may help to provide an optimal response to the altered redox environment for these bacteria. This response seems to be partially mediated through regulation of PA2826, a glutathione peroxidase. GPx is an enzyme that removes H2O2 with the oxidation of glutathione. Glutathione reductase recycles glutathione for further H2O2 removal (Miller and Britigan, 1997). Thus, it is not surprising that either de-repression of PA2826 in ΔospR or constitutive expression of PA2826 in the wild-type MPAO1 increases the bacterial resistance to H2O2.
Unexpectedly, constitutive expression of PA2826 led to a higher intracellular GSH/GSSG ratio and a reduction in bacterial paraquat resistance. The mechanism underlying these observations is currently unknown. Perhaps, the higher intracellular GSH/GSSG ratio is caused by a compensation pathway induced by overexpression of PA2826, as proposed in Fig. 8. It has been reported that a P. aeruginosa zwf mutant shows an increased sensitivity to paraquat and it is believed that the NADPH level is essential in defending against paraquat toxicity (Ma et al., 1998). The zwf gene encodes glucose-6-phosphate dehydrogenase (G6PDH), an enzyme that catalyses the conversion of glucose-6-phosphate to 6-phosphogluconate while producing NADPH (Ma et al., 1998). Depletion of NADPH via glutathione reductase-catalysed production of GSH may explain the higher sensitivity of the ΔospR strain towards paraquat. This hypothesis needs to be further tested experimentally in the future.
Distinct from OhrR, OspR plays multiple regulatory roles as a transcriptional regulator in addition to protection against oxidative stress, as evidenced by the effects of ospR on pigment production and β-lactam resistance in a PA2826-independent manner. OspR is not required for β-lactam resistance in P. aeruginosa, considering that ospR mutant showed no increased sensitivity to cefotaxime or ceftazidime compared with the wild-type MPAO1 strain. However, enhanced expression or activation of OspR may contribute to the increased β-lactam resistance in P. aeruginosa. How OspR impacts β-lactam resistance is unclear and awaits further study. PA1874, a potential OspR-regulated gene (Table 1), has been reported to associate with antibiotic resistance in P. aeruginosa (Zhang and Mah, 2008).
Pseudomonas species are well known to produce multiple-coloured phenazine pigments. These molecules can undergo redox cycling to produce toxic superoxide and H2O2, and play roles in bacterial virulence and redox balance (Price-Whelan et al., 2006). OspR affects the expression of two phenazine-modifying genes, phzM and phzS, indicating that OspR impacts on phenazine biosynthesis or modification (Fig. 3). Interestingly, the MPAO1/p-ospR strains exhibit the dark red pigmentation on agar plates while it shows a normal blue-green pigmentation in liquid culture (Fig 3 and Fig 4). The pathways underlying this observation are still unknown. P. aeruginosa produces several pigmented chemicals in addition to phenazine and the final colour is the combination of these pigments. We found that OspR binds to the promoter region of hmgA and affected its expression. The hmgA gene has recently been shown to be involved in pyomelanin pigment (dark-brown) production (Rodríguez-Rojas et al., 2009). This pathway could contribute to the observed pigment phenotypes.
The regulations of pigment production and quorum-sensing-regulated genes by OspR are unique observations for the OhrR/MgrA family redox-active regulatory proteins. Aside from regulating the expression of phzM and phzS (Fig. 3), two well-known quorum-sensing-regulated genes (Wagner et al., 2006), OspR also binds to the promoter of PA1897 and exerts a regulatory function (Fig. 6). PA1897 is a quorum-sensing-regulated gene controlled by QscR, which is a modulator of quorum-sensing signal synthesis and virulence (Chugani et al., 2001; Lee et al., 2006). It has been known that P. aeruginosa quorum sensing controls expression of catalase and superoxide dismutase genes (Hassett et al., 1999). Our results reveal additional links between oxidative response and quorum sensing through OspR in P. aeruginosa.
The chronic lung infection in cystic fibrosis (CF) patients is a state of chronic oxidative stress (Wood et al., 2001; Lagrange-Puget et al., 2004). In CF P. aeruginosa infections, bacteria routinely reach very high densities within the respiratory secretions [108 to 1010 colony-forming units (cfu) ml−1] (Hoiby, 1998) and their infections are thought to involve co-ordinated bacterial activities facilitated by quorum sensing systems (Wagner and Iglewski, 2008; Willcox et al., 2008; Winstanley and Fothergill, 2009). The cross-talk between oxidative stress and quorum sensing system revealed in this study may co-ordinate various pathways in P. aeruginosa to cope with changes in the host environment.
There are two subfamilies of OhrR type redox-active regulatory proteins (Panmanee et al., 2006; Soonsanga et al., 2008). The 1-Cys type OhrR proteins sense peroxides by formation of a sulphenic acid intermediate that may further react with intracellular small molecule thiols to give a mixed disulphide. The 2-Cys type proteins sense peroxides by forming intermonomer disulphides. OspR belongs to the 2-Cys class as indicated in Fig. 5C. Cys-24 is a key residue involved in pigment production, drug resistance, and expression of oxidative stress-related genes such as PA2826 and ohr in OspR (Fig. 4). Our data indicate that Cys-24 is likely oxidized first, and the resulting sulphenic intermediate is trapped by Cys-134 to form an intermonomer disulphide. Unexpectedly, although H2O2 induced the transcription of PA2826 in the wild-type MPAO1 strain, it failed to induce the expression of PA2826 in the complementation strain (ΔospR/p-ospR) in our study (Fig. 4D). It is likely that the constitutive expression of OspR in ΔospR/p-ospR yielded a large excess of OspR which compromised its redox-sensitive regulation of PA2826 when bacteria were treated with H2O2.
Lastly, the ΔospR strain shows enhanced dissemination in a murine model of acute pneumonia (Fig. 7). The de-repression of PA2826 (Fig. 1), and the enhanced expression of phzS and phzM (Fig. 3) may contribute to increased bacterial virulence observed for this strain, considering that de-repression of PA2826 leads to increased resistance to H2O2 (Fig. 2) while phzS and phzM are required for full virulence of P. aeruginosa in murine lungs (Lau et al., 2004a,b). In addition, the downregulation of hmgA in ΔospR (Fig. 6C) may help bacteria adapt to immune systems since disruption of hmgA leads to higher resistance to oxidative stress and increased persistence in chronic lung infections (Rodríguez-Rojas et al., 2009). All of these pathways intersect at OspR, showing a connection between oxidative stress response and dissemination in pathogenic P. aeruginosa. These results also suggest that exposure to certain levels of oxidative stress may switch on defensive pathways in P. aeruginosa, thus rendering the bacterium more resistant to killing by immune cells. Similarly, it has been shown that nitric oxide-mediated activation of bacterial defence is important for the in vivo virulence of Bacillus anthracis (Shatalin et al., 2008). An interesting avenue for future research therefore would be to identify genes targeted by OspR in a genome scale, which should help to shed light on the role of oxidative stress response in P. aeruginosa physiology and pathogenesis.
Pseudomonas aeruginosa and Escherichia coli strains were maintained in LB medium. All P. aeruginosa and E. coli strains used in this study are listed in Table 2. For plasmid maintenance in E. coli, the medium was supplemented with 100 µg ml−1 ampicillin, 10 µg ml−1 tetracycline or 15 µg ml−1 gentamicin. For marker selection in P. aeruginosa, 100 µg ml−1 tetracycline or 30 µg ml−1 gentamicin was used, as appropriate.
FD2825downF: GTTTCTAGACAACGAGATGTTCGACCT GC; FD2825downR: TTTCTGCAGCCTGGAGTACAAGCG TCTGG; FD2825upF: TTTGGATCCGGGTTGCGTTACTG CATCA; FD2825upR: TTTTCTAGAAGCAGCAGCGATT CTTCG; Tet-XbaI-F: ATTTCTAGATTTCAGTGCAATTTAT CT; Tet-XbaI-R: TTTTCTAGAGGACGCGATGGATATGTT CT; D26-25UpF: CTCGAATTCCCACCTGCTTCTGCTGGTA; D26-25UpR: TCCTCTAGAAATCGCTCATGGCTTGTCTT; D26-25DownF: TCCTCTAGACTCAACGAGATGTTCGACC TG; D26-25DownR: TCGAAGCTTCCTGGAGTACAAGCGT CTGG; PA2825FCF: TTAAGCTTGCATCAAGTGGAACT TCACC; PA2825FCR: TTGGATCCACTACCTGGCCAAGC CTTTC; FO2826F: TGCAAGCTTTATTTCGAGACCCAC CCTCA; FO2826R: CGGGGATCCGGCGTACAGCTTGAAA CACA; PA2827FNF: GTTGACCGAAGAGCAGTTCC; PA2827FNR: GTGGCTGAAGTCGTCCAGTT; PA2826FNF: ATCAAGGGCGAACAGAAGAC; PA2826FNR: GAAGCT CACCCCGTAGTTCA; PA2824FNF: CTTCGTAGCCGTC CATCACT; PA2824FNR: CTACGGCAGTTTCCTCGAAC; phzM-FNF: CTGCTGCGCGTAATTTGATA; phzM-FNR: CAA CAGGCTGGAAAGGTTGT; phzS-FNF: GGAAAGCAG CAGCGAGATAC; phzS-FNR: CGGGTACTGCAGGATC AACT; C24SF: GCTCGACAACCAGCTGAGTTTCAAGCTG TACGC; C24SR: GCGTACAGCTTGAAACTCAGCTGGT TGTCGAGC; hmgAFNF: GAGGTCAGCACGGTGAAGAT; hmgAFNR: CTACCAGTACCTGGCCAACC; PA2010FNF: CGCCTCCACCAATCATTACT; PA2010FNR: CAACTGAT AGCCCGAGTCGT; PA1897FNF: AGATCGGGAAGTCG CTGTAG; PA1897FNR: CGGGTGATCTTCCTCAACAT; Pa2825C2toSF: TGCGCCAGCAGCTGATCTCCAGCACCG GTTTCGACCT; Pa2825C2toSR: AGGTCGAAACCGGT GCTGGAGATCAGCTGCTGGCGCA; PA2850FNF: CTCGA CGTGAAACTCAGCAC; PA2850FNR: GTTGGAGTAGGG GCAGACCT; PA2825F-NdeI: TTGGACACATATGATGAG CACCCGGGGAAAAGT; PA2825R-XhoI: CACACTCGA GCTAGCCTCCTACCACCAGACGGA.
For gene replacement, a sacB-based strategy (Schweizer and Hoang, 1995) was employed. To construct the ospR null mutant (ΔospR), polymerase chain reactions (PCRs) were performed to amplify sequences upstream (706 bp) and downstream (764 bp) of the intended deletion. The upstream fragment was amplified from MPAO1 genomic DNA using primers FD2825upF (with BamHI site) and FD2825upR (with XbaI site), while the downstream fragment was amplified with primers, FD2825downF (with XbaI site) and FD2825downR (with PstI site, see in Experimental procedures). The two PCR products were digested with BamHI–XbaI or XbaI–PstI, as appropriate, and then cloned into BamHI/PstI-digested gene replacement vector pJQ200mp18 via a three-piece ligation, which yielded pJQ200mp18::2825UD. A tetracycline resistance cassette was amplified from EZ-Tn5 <TET-1> (EPICENTRE® Biotechnologies) with primers, Tet-XbaI-F and Tet-XbaI-R (see in Experimental procedures). The PCR product was digested with XbaI and cloned into XbaI-digested pJQ200mp18::2825UD. The resultant plasmid, pJQ200mp18::2825UTD, was electroporated into MPAO1 with selection for tetracycline resistance. Colonies were screened for gentamicin sensitivity and loss of sucrose (5%) sensitivity, which typically indicates a double-cross-over event and thus of gene replacement occurring. The ΔospR strain was further confirmed by PCR and Southern blot analysis.
The ΔPA2826-ospR strain was constructed by a similar strategy using the suicide vector pEX18Ap::26-25UD. Briefly, the 972 bp fragments of the upstream region of PA2826 gene were amplified using primers D26-25UpF (with EcoRI site) and D26-25UpR (with XbaI site). The primers D26-25DownF (with XbaI site) and D26-25DownR (with HindIII site) were used for amplification of 765 bp of PA2825 downstream region. The two PCR products were digested with EcoRI–XbaI or XbaI– HindIII, as appropriate, and then cloned into EcoRI/HindIII-digested gene replacement vector pEX18Ap via a three-piece ligation, yielding pEX18Ap::26-25UD. A 1.8 kb gentamicin resistance cassette was cut from pPS858 with XbaI and then cloned into pEX18Ap::26-25UD, yielding pEX18Ap::26-25UD.
In order to construct the plasmid for constitutive expression of ospR, the following were amplified using primers PA2825FCF (with HindIII site) and PA2825FCR (with BamHI site): a 705 bp PCR product covering 115 bp of the ospR upstream region, the ospR gene, and 98 bp downstream of ospR gene. The product was digested with HindIII and BamHI and ligated into PAK1900 in the same orientation as plac to generate p-ospR. To construct the plasmid for constitutive expression of PA2826, a 605 bp PCR product covering 36 bp of the PA2826 upstream region, the PA2826 gene, and 84 bp downstream of PA2826 was amplified using primers FO2826F (with HindIII site) and FO2826R (with BamHI site) and then cloned into PAK1900, yielding p-PA2826. The three mutations, p-OspRC24S, p-OspSPRC134S and p-OspRC24SC134S, were obtained by using QuikChange II site-directed mutagenesis kit (Stratagene). Primer pairs C24SF/C24SR and Pa2825C2toSF/Pa2825C2toSR were used for generating C24S and C124S mutation respectively. All the constructs were sequenced to ensure that no unwanted mutations resulted.
Total RNA was isolated using a Qiagen RNeasy kit according to the manufacturer’s recommendations. RNA concentration and purity were determined by absorbency at 260 and 280 nm. Northern blotting was performed following previously reported procedures (Chen et al., 2008a).
OspR was cloned into pET28a with a thrombin-cleavable N-terminal His-tag and expressed in E. coli strain BL21 (DE3) (Novagen). Two sets of primers were used to amplify the OspR gene from P. aeruginosa MPAO1 chromosomal DNA: PA2825-NdeI and PA2825-XhoI. The amplified fragments were ligated into similarly cut pET28a (Novagen) to produce the plasmids pET28a-his-OspR. OspRC24S and ospRC134S was amplified from p-ospRC24S and p-ospRC134S by using primer pair PA2825F-NdeI/PA2825R-XhoI and then cloned into pET28a, respectively, as described above. Clones were verified by DNA sequencing and transformed into BL21 (DE3) for expression.
The expression and purification procedures for OspR, OspRC24S and OspRC134S are described as follows. The strains were grown at 37°C overnight in 10 ml of LB medium containing 50 µg ml−1 kanamycin (LBkan50). The next day, the cultures were transferred into 1 l of LBkan50, incubated at 37°C until the OD600 reached 0.6, after which IPTG (isopropyl-1-thio-β-d-galactopyranoside) was added to a final concentration of 1.0 mM.After 4 h incubation at 30°C, the cells were harvested by centrifugation and stored at −80°C. The cells were lysed at 4°C by sonication in lysis buffer [10 mM Tris (pH 7.4), 300 mM NaCl, 1 mM PSMF and 2 mM DTT]. Clarified cell lysate was loaded onto a HisTrap HP column (Amersham Biosciences), washed with Ni-NTA washing buffer and eluted with Ni-NTA elution buffer. The fractions containing OspR, OspRC24S or OspRC134S were concentrated and loaded onto a Superdex-200 gel filtration column with a running condition of 10 mM Tris (pH 7.4), 300 mM NaCl and 2 mM DTT.
The electrophoretic mobility shift experiments were performed by using [γ-32P]-ATP labelled method. Duplex DNA (5′–3′ annealed to its complementary strand) containing various promoter regions was used for the assay. DNA fragments were labelled using [γ-32P]-ATP (Perkin Elmer) and a T4 polynucleotide kinase. Unincorporated dATP were removed using illustra MicroSpin™ G-50 Columns (GE Healthcare). The electrophoretic mobility shift experiments were performed by adding 0.04 pmol of P32-labelled duplex DNA to 24 µl of reaction buffer (10 mM HEPES at pH 8.0, 1 mM EDTA, 50 mM KCl, 0.05% Triton X-100, 10% glycerol, 10 µg ml−1 salmon sperm DNA). Reactions were placed on ice for 30 min before loading. Designated amounts of H2O2 or CHP were used as appropriate. Gels were run in 0.5× TBE at 85 V at room temperature. The gels were dried and subjected to autoradiography using the storage phosphor screen (Image Screen-K, Kodak) and the Molecular Imager PharosFX Plus System (Bio-Rad).
LB agar plates were made with designated amounts of paraquat, H2O2 and cefotaximine. The overnight culture was diluted 100-fold in fresh LB medium. P. aeruginosa were grown to early stationary phase (OD600 = 2.0) in LB broth, and 10-fold dilutions were made. Aliquots (10 µl) of the diluted cultures for each strain were spotted onto the solid media and grown at 37°C.
The overnight culture was diluted 100-fold in fresh LB media and incubated at 37°C for 6 h until the culture reached OD600 = 2.2–2.3. The pellet from 10 ml of a cell culture was stored at −80°C and re-suspended in 0.25 ml of double-distilled water before used. GSH/GSSG ratio was measured by using GSH/GSSG Cuvette Assay Kit (Catalog # GT35, Oxford Biomedical Research).
Purified His6-tagged OspR proteins were washed with the thiol-assay buffer (100 mM KH2PO4/K2HPO4, 200 mM NaCl and 1 mM EDTA at pH 7.0) four times to remove extra DTT using Microcon YM-10 (Amicon) ultrafiltration device. Protein samples (50 µM, monomer) were treated with four equivalents of CHP (200 µM) at room temperature for 10 min followed by washing with the thiol-assay buffer three times to generate the oxidized OspR. Disulphide bond formation was monitored by non-reducing SDS-PAGE. The OspRC134S-SOH modification was confirmed by the NBD-Cl assay (Chen et al., 2006).
Mouse infections were carried out as described previously (Laskowski et al., 2004), using 8-week-old female C57BL/6 mice obtained from the National Cancer Institute and housed under specified pathogen-free conditions. All studies were approved by the Yale University Institutional Animal Care and Use Committee. Mice were lightly anaesthetized with isoflurane and intranasally infected with c. 2 × 107 cfu of each bacterial isolate; the actual inoculum titre for each group was determined by plating serial dilutions. Animals were sacrificed 18 h post infection. Lungs and spleens were aseptically removed and homogenized in PBS plus 0.1% Triton X-100 to obtain single-cell suspensions. Serial dilutions of each organ were plated on VBM (Vogel–Bonner minimal) agar plates. Bacterial burden per organ was calculated and is expressed as a ratio of the inoculum delivered per animal. Statistical analysis was performed using Prism software (GraphPad).
We would like to thank Professor H.P. Schweizer (Colorado State University) for kindly providing us with plasmids pEX18Ap and pPS858, Ms S. Frank for editing the manuscript, and valuable suggestions from reviewers. This work was supported by National Institutes of Health (NIH/NIAID AI074658 to C.H.), a Burroughs Wellcome Fund Investigator in the Pathogenesis of Infectious Disease Award (C.H.).
Additional supporting information may be found in the online version of this article.
Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.