|Home | About | Journals | Submit | Contact Us | Français|
SoxR is a transcriptional regulator that controls an oxidative stress response in Escherichia coli. The regulator is primarily activated by superoxide anion-dependent oxidation. Activated SoxR turns on transcription of a single gene, soxS, which encodes a transcriptional regulator that activates a regulon that includes dozens of oxidative stress response genes. SoxR homologues have been identified in many bacterial species, including the opportunistic pathogen Pseudomonas aeruginosa. However, the expected SoxR partner, SoxS, has not been found in P. aeruginosa. Thus, the primary gene target(s) of P. aeruginosa SoxR is unknown and the involvement of this regulator in the oxidative stress response of the bacterium remains unclear. We utilized transcriptome profiling to identify the P. aeruginosa SoxR regulon and constructed and characterized an unmarked P. aeruginosa ΔsoxR mutant. We provide evidence indicating that P. aeruginosa SoxR activates a six-gene regulon in response to O2·−-induced stress. The regulon includes three transcriptional units: (i) the recently identified mexGHI-ompD four-gene operon, which encodes a multidrug efflux pump system involved in quorum-sensing signal homeostasis; (ii) gene PA3718, encoding a probable efflux pump; and (iii) gene PA2274, encoding a probable monooxygenase. We also demonstrate that P. aeruginosa SoxR is not a key regulatory player in the oxidative stress response. Finally, we show that P. aeruginosa SoxR is required for virulence in a mouse model of intrapulmonary infection. These results demonstrate that the E. coli-based SoxRS paradigm does not hold in P. aeruginosa and foster new hypotheses for the possible physiological role of P. aeruginosa SoxR.
Bacteria possess complex molecular biosensors to detect harmful levels of intracellular oxidants originating from endogenous or external sources and transduce this information into activation of an oxidative stress adaptive response (36, 37, 45, 53). This adaptive response includes the expression of genes involved in preventing, counteracting, and repairing oxidative damage. The oxidant biosensors are redox-responsive transcriptional regulators that can exist in reduced or oxidized forms. The oxidized forms of each of these regulators, which arise from oxidant-dependent protein modifications, activate transcription of target genes to trigger a coordinated oxidative stress response. Two such redox-responsive regulators are OxyR and SoxR (36, 37, 45, 53). These regulators are believed to be the primary sensor-transducers controlling the activation of oxidative stress adaptive responses in many bacterial species. OxyR and SoxR have been studied extensively in Escherichia coli. The respective molecular mechanisms by which E. coli OxyR and SoxR sense and transduce redox signals into activation of oxidative stress adaptive responses have been well studied and have become two dominant and valuable archetypal paradigms.
OxyR belongs to the LysR family of transcriptional regulators and is a sensor of H2O2 (7, 48). OxyR is activated by H2O2 via a reversible thiol-disulfide redox switch. H2O2-dependent oxidation of thiols of conserved Cys residues in OxyR produces the formation of an intramolecular disulfide bond to generate the active regulator (54). Although both inactive (reduced) and active (oxidized) regulator forms bind target promoter sequences, only the bound active form activates transcription of the OxyR regulon by interacting with RNA polymerase (47, 49). The E. coli OxyR regulon is believed to include dozens of genes, and its hallmarks are genes such as ahpCF (alkyl hydroperoxidase), dps (a protective DNA-binding protein), fur (ferric uptake regulator), gorA (glutathione reductase), grxA (glutaredoxin I), and katG (hydroperoxidase I) (56, 55). OxyR also activates the expression of oxyS, a small regulatory RNA that exerts its posttranscriptional regulatory function on various genes to achieve homeostatic regulation of the H2O2 produced by metabolic activity (2, 13). The oxyS RNA is believed to activate or repress the expression of as many as 40 genes (2), including flhA (activator of the hyp operon, repressed by oxyS), rpoS (stationary-phase sigma factor, repressed by oxyS), pqqL (endopeptidase, activated by oxyS), and uhpT (hexose phosphate transporter, activated by oxyS) (2, 3, 52).
SoxR belongs to the MerR family of transcriptional regulators and is a sensor for the superoxide anion (O2·−) that is generated, for example, by redox-cycling compounds such as methyl viologen (paraquat [PQ]). SoxR is activated by O2·−-dependent reversible one-electron oxidation of its iron-sulfur center (10, 12, 20). Nitric oxide (NO) also activates SoxR, however, by a different mechanism than that observed for O2·−. The NO-dependent activation takes place via nitrosylation of iron-sulfur centers with displacement of sulfide to form dinitrosyl-iron-dithiol(cysteine) complexes (9). As in the case of OxyR, both the active and the inactive forms of SoxR bind the target promoters, but only the active form stimulates transcription, probably by promoting an allosteric rearrangement of the SoxR-DNA complex that favors transcription initiation (22, 18, 21). Notably, SoxR activates the expression of a single gene, soxS, which encodes the global transcriptional activator SoxS (18). SoxS has homology to the AraC/XylS family of regulators and activates expression of target genes by binding to their promoters and recruiting RNA polymerase (26) and/or by binding to their promoters after forming a complex with the polymerase (15, 28). The SoxS regulon is believed to include at least 60 genes in E. coli (38). Among these are sodA (Mn-dependent superoxide dismutase), fpr (ferredoxin reductase), fumC (redox-resistant fumarase), nfo (DNA repair endonuclease IV), zwf (glucose-6-phosphate dehydrogenase), fldA (flavodoxin A), and acrAB (efflux pump).
OxyR and SoxR homologues have been identified in many bacterial species (53). One such bacterium is the opportunistic human pathogen Pseudomonas aeruginosa (11). The OxyR paradigm noted above appears to be preserved in P. aeruginosa, where OxyR is known to activate the expression of several oxidative stress defense genes, such as katB (catalase B), ahpB (alkyl hydroperoxidase), and ahpCF (alkyl hydroperoxidase) (31). Conversely, the SoxR paradigm does not seem to be applicable to P. aeruginosa SoxR, a homolog with 62% sequence identity and 77% sequence similarity to E. coli SoxR. This notion arises from the observation that SoxR's expected partner, SoxS, is not found in the genome of P. aeruginosa (16, 46). Thus, the primary gene target(s) of P. aeruginosa SoxR is unknown and the involvement of the regulator in the oxidative stress adaptive response of P. aeruginosa remains unclear. Nevertheless, the P. aeruginosa soxR mutant displays a significant delay in systemic dissemination in the mouse burn wound infection model and hypersensitivity to macrophage-mediated killing (16). These observations suggest that SoxR plays an important role in P. aeruginosa virulence and underscore the need to understand the physiological role of P. aeruginosa SoxR. In this study, we describe the P. aeruginosa SoxR regulon, which is activated in response to O2·−- but not H2O2-induced oxidative stress. We also demonstrate that P. aeruginosa SoxR is not a key player in the P. aeruginosa oxidative stress adaptive response and present an alternative hypothesis for the physiological role of P. aeruginosa SoxR. Finally, we provide evidence indicating that P. aeruginosa SoxR is required for virulence in a mouse intrapulmonary infection model.
The unmarked ΔsoxR null mutant of P. aeruginosa PAO1 (46), P. aeruginosa ΔsoxR, was constructed by allelic replacement through site-directed homologous recombination followed by sacB-based counterselection in sucrose-containing medium. Briefly, a 1,692-bp deletion cassette comprised of a P. aeruginosa genomic fragment (positions 2502550 to 2504696) deleted for soxR (positions 2503433 to 2503887) was used to construct the ΔsoxR mutant. The deletion cassette was generated by gene splicing using the overlap extension PCR method (23) with primers X-1 (CACCGCCGCACCAGCACGACCTCAATCC), X-2 (CACGGCTGACCGCTGCCTAGCCGTCATTCTTCATGATTGGCTTGACCTCA), X-1′ (TGAGGTCAAGCCAATCATGAAGAATGACGGCTAGGCAGCGGTCAGCCGTG), and X-2′ (CGAAGGTATCGGTGAGGGTGCGGTTG). The cassette was cloned into pENTR/D-TOPO (Invitrogen) and then transferred to pEXGWGm (kindly provided by S. Lory, Harvard Medical School) using the Gateway Cloning System (Invitrogen). The resulting construct was conjugated into P. aeruginosa PAO1 using E. coli S17-1 as donor (34). Integrants were selected on Luria-Bertani (LB) agar plates containing gentamicin (50 μg ml−1) and Irgasan (10 μg ml−1), a broad-spectrum antibiotic that is inactive against Pseudomonas spp. Counterselection in LB medium containing 8% sucrose was utilized to select for the unmarked ΔsoxR mutant, which was confirmed by PCR and Southern blot analysis. Complementation experiments were performed with pSoxR, a pME6001-based plasmid (5) expressing soxR (genome fragment 2503900 to 2503422 cloned as a KpnI-HindIII insert). The soxR fragment was generated from genomic DNA by standard PCR methods using primers X-rbs (CAAGCAAACTGGTACCAGAGGCAATCATGAAG) and X-stop (CACCACGGCTAAGCTTTGCCTAGCCGTCG). Plasmid pSoxR was electroporated into P. aeruginosa strains, and transformants were selected and cultured in LB medium containing gentamicin (50 μg ml−1).
Microarray experiments were carried out essentially as previously reported (33, 32). For each strain, a 2-ml aliquot of an overnight culture was inoculated into a 1-liter Erlenmeyer flask with 100 ml of tryptic soy broth (17), and the culture was incubated at 37°C with shaking at 220 rpm. When the optical density at 600 nm (OD600) reached 0.5, the culture was split in aliquots of 10 ml into 100-ml Erlenmeyer flasks. The cultures were immediately treated in duplicate by addition of H2O2 (0.5 mM) or PQ (100 μM) to produce sublethal oxidative stress (33, 38) or left untreated. Cells from treated and untreated cultures were collected by centrifugation (3,000 rpm, 10 min, 4°C) after 10 min of incubation (220 rpm, 37°C). RNA was isolated from each cell pellet using the RNeasy Mini kit (QIAGEN). Control spike transcript Mix L10 (kindly provided by S. Lory) was added to the RNA samples as recommended in the GeneChip genome arrays expression analysis protocol technical manual (Affymetrix). cDNA was prepared in triplicate from each RNA sample using pd(N)6 random primers and the SuperScript II kit (Invitrogen). Each cDNA sample was fragmented (50-bp average size) with DNase I in One Phor-All buffer (Invitrogen), and the fragmented cDNA was 3′ labeled using the Enzo BioArray Terminal kit with Bioin-ddUTP (Affymetrix). Each labeled cDNA sample was hybridized to a single GeneChip P. aeruginosa Genome Array (Affymetrix). Array hybridization, scanning, and data analysis were performed as recommended in the GeneChip P. aeruginosa genome arrays expression analysis protocol manual (Affymetrix). Array image acquisition and analysis were performed using Affymetrix Microarray Suite software. Expression data files obtained with MAS software were subsequently loaded into the software GeneSpring 6.0 (Silicon Genetics) for statistical group comparison. The Microarray Suite-derived expression data were normalized with per-chip normalization (using 50th percentile) and per-gene normalization (with a cutoff of 0.01). Statistical analyses to identify genes with statistically significant differences in mRNA levels between compared groups were performed using the Wilcoxon-Mann-Whitney test (P value cutoff, 0.05). Fold change values were calculated as ratios of means of gene signals for genes with statistically significant differences in mRNA levels.
Each cDNA sample prepared as noted above was analyzed in triplicate by real-time PCR using the SYBR Green PCR Core kit (PE Applied Biosystems) and gene-specific primers. Real-time quantitative PCR was performed on an ABI-PRISM 7900HT Sequence Detection System (PE Applied Biosystems) as reported elsewhere (33, 32).
The effects of PQ, H2O2, and the antibiotics mentioned below on the growth of P. aeruginosa strains were determined using a standard disk diffusion method. The cultures were spread on the top of LB agar plates, and disks (Becton Dickinson) impregnated with PQ, H2O2, or antibiotics were then placed onto the agar. Plates contained gentamicin (50 μg ml−1) to maintain the pME600 plasmids. Antibiotic testing was done in plates without PQ and in plates with 50 μM or 100 μM PQ to induce the SoxR regulon. The plates were incubated at 37°C for 24 h, and the zone of growth inhibition around each disk was recorded. The antibiotic disks utilized were amikacin (30 μg), ampicillin/sulbactam (20 μg), aztreonam (30 μg), ceftazidime (30 μg), ceftazidime/clavulanic acid (30/10 μg), cefuroxime (30 μg), cephalothin (30 μg), erythromycin (15 μg), gentamicin (10 μg), imipenem (10 μg), levofloxacin (ciprofloxacin) (5 μg), nalidixic acid (30 μg), novobiocin (5 μg), oxacillin (1 μg), penicillin (10 IU), polymyxin (300 IU), sulfamethoxazole/trimethoprim (23.75/1.25 μg), tetracycline (30 μg), and tobramycin (10 μg). Antibiotic susceptibility was also tested in an automated Vitek 2 susceptibility instrument using a standard antibiotic susceptibility test card for clinical testing of gram-negative bacteria as recommended by the manufacturer (Biomerieux). The card included amikacin, ampicillin, ampicillin/sulbactam, aztreonam, cefazolin, cefepime, cefotetan, ceftazidime, ceftriaxone, cefuroxime, cefuroxime axetil, ciprofloxacin, gentamicin, imipenem, levofloxacin, meropenem, nitrofurantoin, piperacillin, piperacillin/tazobactam, and tobramycin.
Mouse infection experiments were performed as described earlier (51). Briefly, five mice (C57BL/6, 6 to 8 weeks old) per group were challenged by an intratracheal injection of 106 CFU of agarose-encapsulated P. aeruginosa cells diluted in 50 μl of phosphate-buffered saline. Encapsulation of P. aeruginosa and evaluation of the number of encapsulated CFU were verified as previously reported (51). Following intratracheal administration of the P. aeruginosa beads, the survival of the animals was monitored over time. Animal experiments were performed under institutional guidelines.
The pattern search using as a query a 18-bp degenerated sequence that satisfied potential palindromic pairings ([CG]CTCAA[CG]TT[ACTG][ACTG][CG]TTGAG[CG]) deduced for SoxR binding sites was carried out with the program ExhaustivePatterns (L. Shi., unpublished), a program that performs pattern recognition for a given series of patterns and their distance restraints. The pattern search utilizing the iron-sulfur center coordination motif CxxCxCxxxxC as a query was performed with the ScanProsite search tool available via the Expert Protein Analysis System (ExPASy) proteomics server of the Swiss Institute of Bioinformatics (http://us.expasy.org/).
To gain an initial insight into the physiological function of P. aeruginosa SoxR, we sought to define the P. aeruginosa SoxR regulon. To this end, we constructed a ΔsoxR unmarked mutant by creating a chromosomal deletion leaving only the first three and the last three codons, including the stop codon, of soxR. The deletion was confirmed by PCR and Southern analysis (data not shown). The soxR deletion did not appreciably compromise P. aeruginosa's fitness under standard culture conditions (Fig. (Fig.1).1). To reveal possible SoxR-regulated genes, and working under the assumption that SoxR would be activated by O2·−-dependent oxidation, we performed a side-by-side comparison between the transcriptomes of exponentially growing P. aeruginosa ΔsoxR and wild-type (wt) strains treated with the O2·−-generating compound PQ for 10 min. The analysis revealed that the mRNA levels of only six genes (PA2274, PA3718, and PA4205 to PA4208/mexGHI-opmD) showed a statistically significant difference between the two strains (Fig. (Fig.2A2A and Table Table1)1) and that, in all cases, the mRNA levels of these genes were dramatically increased (>3- to 20-fold) in the wt strain relative to the ΔsoxR strain (Table (Table1).1). Conversely, the comparison of the transcriptomes of untreated cells of the ΔsoxR and wt strains showed no genes with statistically significant differences in mRNA levels (Fig. (Fig.2B).2B). By analogy with E. coli SoxR's properties, P. aeruginosa SoxR would not be expected to be activated by H2O2-dependent oxidation. Consequently, P. aeruginosa SoxR-dependent activation of gene expression would not be anticipated in cells exposed to H2O2. To determine whether H2O2 treatment had an effect on the P. aeruginosa SoxR regulon, we compared the transcriptomes of H2O2-treated ΔsoxR and wt strains. As expected, no major differences in the gene expression for the genes of the transcriptomes were observed (Fig. (Fig.2C),2C), and no genes with a statistically significant difference in expression levels between the strains were identified. Particularly, the treatment had no significant effect on the mRNA levels of the P. aeruginosa SoxR regulon genes, which remained essentially unchanged (Table (Table11).
To further validate the microarray data, the changes in the mRNA levels of the P. aeruginosa SoxR regulon genes in PQ- and H2O2-treated ΔsoxR and wt strains relative to their respective untreated controls were determined by real-time RT-PCR analysis. The results of this analysis corresponded well to the results derived from the microarray analysis, taking into account the variation expected due to the different natures of the two methodologies (Table (Table1).1). The regulatory defect of the mutant was unlikely to be produced by a polar effect resulting from the deletion of soxR since the gene located downstream of P. aeruginosa soxR, pbpC, is transcribed in the opposite orientation relative to soxR and encodes a penicillin-binding protein. Nevertheless, we investigated this possibility by conducting a complementation experiment. To this end, we utilized real-time RT-PCR to analyze the expression changes of the P. aeruginosa SoxR regulon genes in the ΔsoxR strain transformed with pSoxR (a plasmid constitutively expressing soxR) in response to PQ treatment. The results of this analysis indicated that the transformant, P. aeruginosa ΔsoxR/pSoxR, had its ability to induce the P. aeruginosa SoxR regulon genes in response to PQ treatment restored by episomal expression of soxR (Table (Table1).1). This observation rules out the possibility that the regulatory defect of the ΔsoxR strain is produced by a polar effect. Overall, these results indicate that, directly or indirectly, P. aeruginosa SoxR activates expression of a six-gene regulon in response to PQ treatment.
Of the six genes in the P. aeruginosa SoxR regulon noted above, PA3718 encodes a probable efflux pump of the multiple facilitator superfamily (40). Notably, the product of PA3718 is a predicted inner membrane protein with 11 transmembrane segments and limited sequence similarity (27% identity, 41% similarity) to SmvA, a Salmonella enterica serovar Typhimurium multiple facilitator superfamily pump that confers increased resistance to PQ (41). PA2274, a second gene of the regulon, is annotated as a hypothetical protein of unknown function (46). A sequence similarity search for PA2274 revealed a marginal match to the monooxygenase ActVA-Orf6 (BLAST E value, 0.003), involved in actinorhodin biosynthesis (25). A protein fold recognition search with 3D-PSSM (24) revealed a potential match to protein data bank structure 1LQ9, the ActVA-Orf6 structure (first match, PSSM E value: 1.09 × 10−5). Sciara et al. described the structure of ActVA-Orf6 and noted the similarity between ActVA-Orf6 and PA2274, particularly in the active sites (43). The remaining four genes correspond to the recently identified mexGHI-ompD operon, which encodes the components of a multidrug efflux pump of the resistance/nodulation/division superfamily (1, 30, 44).
Examination of the promoter regions of PA3718, PA2274, and the mexGHI-ompD operon (Fig. (Fig.3A)3A) revealed the presence of a conserved dyad symmetrical 18-bp sequence in each promoter. This sequence highly resembles the E. coli and S. enterica serovar Typhimurium SoxR binding site in the soxS promoters of the corresponding bacteria (18, 19, 35). Like SoxS promoters in E. coli and S. enterica serovar Typhimurium and other promoters targeted by regulators of the MerR family, the promoters of the P. aeruginosa SoxR regulon have an elongated 19-bp spacer between their predicted −10 and −35 sequences (6, 18, 19, 35) (Fig. (Fig.3B).3B). Interestingly, the gene pairs soxR and PA2274 in P. aeruginosa, soxR and soxS in E. coli (4), and soxR and soxS in S. enterica serovar Typhimurium (35) are organized alike; i.e., the genes in each pair are arranged divergently with their 5′ ends separated by 86 to 77 bp (Fig. (Fig.3B).3B). Furthermore, the SoxR binding site located in each intergenic region of each gene pair is in nearly an identical position in the three species (Fig. (Fig.3B).3B). The finding of SoxR binding site-like sequences in the promoters of PA3718, PA2274, and the mexGHI-ompD operon is in agreement with the P. aeruginosa SoxR-dependent activation of these promoters deduced from the gene expression analysis. This finding also supports the idea that the three transcriptional units of the regulon are activated by direct action of SoxR on their promoters. A genome-wide computational pattern search of the entire P. aeruginosa genome utilizing as a query an 18-bp degenerate sequence ([CG]CTCAA[CG]TT[ACTG][ACTG][CG]TTGAG[CG]) deduced from the SoxR binding sites shown in Fig. Fig.3B3B revealed no matches other than those already identified. This result is in agreement with the inference based on the microarray experiments of only three SoxR-regulated promoters in P. aeruginosa.
In contrast with the scenarios observed in E. coli and S. enterica serovar Typhimurium, the nature of the genes regulated by P. aeruginosa SoxR suggests that the regulator does not have an important role as a controller of the oxidative stress adaptive response. However, the observed PQ-dependent upregulation of efflux pumps by P. aeruginosa SoxR suggests that the regulator may mediate a PQ-dependent increase in antibiotic resistance or even in PQ resistance. To test these hypotheses, we investigated the effects of H2O2, PQ, and various antibiotics on the growth of P. aeruginosa ΔsoxR transformed with plasmid vector pME6001, P. aeruginosa ΔsoxR/pSoxR, and P. aeruginosa wt transformed with pME6001. Disk diffusion assays revealed no difference in susceptibility to PQ and H2O2 between the strains (Fig. (Fig.4).4). Likewise, the antibiotic susceptibility testing in the absence or presence of PQ to induce the P. aeruginosa SoxR regulon did not reveal significant differences between the strains (data not shown).
The results of the oxidant susceptibility experiments are consistent with the microarray analysis-based indication that the activation of the P. aeruginosa oxidative stress adaptive response is independent from SoxR. In a P. aeruginosa soxR-independent manner, both PQ and H2O2 treatments produced an overlapping genetic response consisting of the statistically significant increase in the mRNA levels of 60 genes (data not shown) that, with a few exceptions, have been previously reported to be up-regulated in response to oxidant treatment (33). This group of genes has the hallmarks of an oxidative stress adaptive response. Among these genes were those implicated in oxidative stress adaptation, such as katA, katB, ahpF, ohr, fumC, ankB, sodM, and dps. Each of these genes displayed a more than fivefold mRNA level increase. The overlap in the genetic responses to PQ and H2O2 treatments has been well documented in other bacterial species. PQ is known to induce expression of genes known to be induced by H2O2 via OxyR due to generation of H2O2 by nonenzymatic disproportionation and superoxide dismutase-mediated conversion of O2·− (14, 55). Conversely, genes known to be induced by PQ via SoxRS have been reported to be induced by H2O2 via an as yet unclear mechanism (27, 55).
Overall, the susceptibility testing indicates that deletion of soxR does not significantly reduce P. aeruginosa's ability to withstand peroxide- and superoxide-based oxidative stress or increase antibiotic susceptibility. Thus, under the conditions tested, P. aeruginosa SoxR is neither a significant component of the oxidative stress adaptive response nor a regulator capable of affording increased antibiotic resistance by activation of its regulon. This scenario contrasts with the reduced oxidant and antibiotic sensitivity trigger via the activation of the soxRS regulon in E. coli and S. enterica serovar Typhimurium (36, 37, 45, 53).
Chronic P. aeruginosa lung infection is one of the major causes of morbidity and mortality in cystic fibrosis patients. To investigate the importance of SoxR in pulmonary infection, the virulence of the ΔsoxR strain was compared with the isogenic wt strain and the complemented P. aeruginosa ΔsoxR/pSoxR strain in a time-to-death experiment using a mouse intrapulmonary infection model with a challenge corresponding to a lethal dose of encapsulated P. aeruginosa wt (Fig. (Fig.5).5). The mouse survival data indicated that, as expected, the wt strain caused 60% and 100% mortality within 1 and 3 days postinfection, respectively. Conversely, the ΔsoxR mutant showed dramatic attenuation, producing no deaths in the infected mice during the 7-day test period. Complementation of the ΔsoxR strain with pSoxR partially rescued the wt phenotype. The intermediate phenotype observed with the complemented strain is probably due to the loss of pSoxR by the complemented strain in the absence of antibiotic selective pressure in the mouse. These results demonstrate that soxR plays a critical role in the virulence of P. aeruginosa in mice infected via the pulmonary route and suggest a relevant role for soxR in P. aeruginosa infections of the human host.
The redox-responsive regulator SoxR is believed to be one of the two primary sensor-transducers controlling the activation of the oxidative stress adaptive response in many bacterial species (36, 37, 45, 53). O2·−- and NO-dependent oxidation activates E. coli SoxR, which then activates transcription of the global regulator SoxS. This global regulator activates expression of a regulon that contains genes required for adaptation to oxidative stress (38). The molecular mechanism by which E. coli SoxR senses and transduces redox signals into activation of an oxidative stress adaptive response has become a dominant archetypal paradigm believed to be applicable to the SoxR of other bacterial species.
A SoxR homolog has been identified in the opportunistic human pathogen P. aeruginosa (16). However, P. aeruginosa appears to lack a SoxS homolog (16, 46), suggesting that the SoxRS paradigm does not hold in P. aeruginosa. A long-standing question arising from this observation is: what is the physiological role of P. aeruginosa SoxR? As part of our effort to address this question, we utilized transcriptome profiling to identify the P. aeruginosa SoxR regulon and constructed and characterized an unmarked P. aeruginosa ΔsoxR mutant. Transcriptome comparisons between P. aeruginosa wt and ΔsoxR strains left untreated or treated with the oxidative stress-inducing agent PQ or H2O2 revealed that P. aeruginosa SoxR is required for PQ treatment-dependent upregulation of the expression of six genes. The six-gene P. aeruginosa SoxR regulon is organized into three transcriptional units corresponding to PA2274, PA3718, and the mexGHI-ompD operon. Activation of the P. aeruginosa SoxR regulon is likely to require binding of oxidized P. aeruginosa SoxR to predicted promoters located directly upstream of each of the three transcriptional units. The fact that each of these three promoters contains a sequence that closely resembles the SoxR binding site in the soxS promoters of E. coli and S. enterica serovar Typhimurium (18, 19, 35) supports this view. Taken together, these results strongly suggest that P. aeruginosa SoxR is a functional redox-responsive regulator that binds to a sequence nearly identical to that recognized by E. coli and S. enterica serovar Typhimurium SoxR and that it is activated by O2·−-dependent oxidation, a property also noted for its E. coli counterpart.
Despite the sequence similarity and the apparently shared mechanistic principles for oxidant sensing and signal transduction, the P. aeruginosa and E. coli SoxR homologs have neither equivalent gene targets nor comparable physiological roles. Unlike E. coli SoxR, which activates expression of a single transcriptional unit encoding a global activator of the oxidative stress adaptive response, SoxS, P. aeruginosa SoxR activates three transcriptional units. One of these units contains PA2274, encoding a possible monooxygenase and perhaps with a detoxification function via oxidation of harmful xenobiotic compounds. Notably, the divergent arrangement of PA2274 and P. aeruginosa soxR and the position of the predicted SoxR binding site in the intergenic region are nearly identical to those observed for the soxS and soxR pairs in E. coli and S. enterica serovar Typhimurium. In E. coli, the position of the SoxR binding site permits E. coli SoxR to act as an autorepressor (21). Thus, the presence of the SoxR binding site in the PA2274-P. aeruginosa soxR intergenic region provides the molecular basis for the previously suggested autorepressor function of P. aeruginosa SoxR (16).
The other two transcriptional units of the P. aeruginosa SoxR regulon encode the efflux pumps PA3718 and MexGHI-OmpD. The function of PA3718 has not been investigated. Notably, the pump has weak similarity to S. enterica serovar Typhimurium SmvA, a pump with significant involvement in PQ resistance. The MexGHI-OmpD system has recently been shown to be capable of increasing resistance to some antimicrobial compounds (e.g., erythromycin and tetracycline), but only when the system is overexpressed in a P. aeruginosa mutant lacking the known primary multidrug efflux (Mex) pump systems (30). Given these observations, it is not surprising that P. aeruginosa wt and P. aeruginosa ΔsoxR have comparable antibiotic susceptibilities in the absence or presence of PQ. Another study reported that insertion of a gentamicin resistance cassette upstream of the mexGHI-ompD operon reduced operon expression and rendered a strain with increased resistance to tetracycline, netilmicin, and ticarcillin plus clavulanic acid (1). These unexpected results were attributed to compensatory expression of other multidrug efflux pumps in the mutant (1). Even more surprising, it was reported in the same study that complementation of the mutant with the mexGHI-ompD operon restored tetracycline but not netilmicin sensitivity and further increased resistance to ticarcillin plus clavulanic acid (1). The bases for these results remain unclear.
Although the weak similarity of PA3718 to S. enterica serovar Typhimurium SmvA is suggestive, the overall nature of the P. aeruginosa SoxR regulon and the lack of a SoxS partner to amplify the gene expression response suggest that P. aeruginosa SoxR does not play a primary role in the activation of an oxidative stress adaptive response. This view is supported by the fact that the P. aeruginosa wt and ΔsoxR strains have comparable susceptibilities to both PQ and H2O2 and the observation that both PQ and H2O2 treatments trigger activation of the hallmark genes of the oxidative stress adaptive response in a P. aeruginosa SoxR-independent manner. The latter observation may suggest that P. aeruginosa has at least one as yet unidentified redox-responsive regulator that senses and transduces the O2·− signal into expression of oxidative stress protective genes. Since our analysis of the P. aeruginosa genome shows that SoxR is the only transcriptional regulator with the conserved CxxCxCxxxxC tetracysteine signature (data not shown) that coordinates the iron-sulfur center of SoxR, the unidentified regulator(s) that senses and transduces the O2·− signal is likely to have a redox switch configuration different from that in SoxR.
The MexGHI-OmpD system noted above is one of six characterized resistance/nodulation/division broad-specificity Mex systems in P. aeruginosa (39, 42). Except for the MexGHI-OmpD system, cognate transcriptional regulators have been identified for each of the other Mex systems. None of these regulators is expected to be redox responsive; instead, they are predicted to function as repressors (MexL, MexR, MexZ, and NfB) in the absence of specific small-molecule ligands or as activators (MexT) in their presence. The ligand of each of these regulators is believed to be a substrate(s) of its cognate efflux system (42). To the best of our knowledge, the P. aeruginosa SoxR-dependent regulation of MexGHI-OmpD revealed by our studies represents the first example of a Mex system whose expression is reported to be directly regulated by a redox-responsive transcriptional regulator.
Interestingly, transcription from an as yet uncharacterized promoter upstream of mexG appears to be affected by quorum sensing; it has been reported to be up-regulated in cultures treated simultaneously with N-(3-oxododecanoyl)-l-homoserine lactone and N-butyryl-l-homoserine lactone, but not in cultures treated with either signal alone (50). However, the promoter was proposed not to be directly regulated by the quorum-sensing signals since the upregulation was delayed several hours posttreatment and was not evident until stationary growth phase (50). Expression of the mexAB-oprM operon is also influenced by quorum sensing. In this case, operon expression is increased in cultures treated by exogenous addition of N-butyryl-l-homoserine lactone, whereas N-(3-oxododecanoyl)-l-homoserine lactone has only a moderate effect (29). The molecular mechanisms underlying the possible direct or indirect regulation of these Mex systems by quorum sensing remains unknown. The observation that the MexAB-OprM, MexEF-OprN, and MexGHI-OmpD systems are implicated in quorum-sensing signal homeostasis adds complexity to a still poorly understood regulatory network controlling expression of the Mex systems (42). In particular, the MexGHI-OmpD system is believed to play a role in quorum-sensing signal homeostasis, probably by participating in quorum-sensing signal efflux (1, 8). This view is based on the observation that the extracellular concentration of quorum-sensing signals and multiple quorum-sensing-regulated phenotypes are altered in strains with mutations in the mexGHI-ompD operon (1, 8). It is tempting to speculate that P. aeruginosa SoxR may function as a regulator directly linking O2·−-dependent (and possibly NO-dependent) oxidative stress and the quorum-sensing regulatory network of P. aeruginosa via the control of MexGHI-OmpD expression.
We found that P. aeruginosa soxR is required for virulence in a mouse pulmonary infection model, and an earlier report demonstrated that the soxR mutant displays a significant delay in systemic dissemination in the mouse burn wound infection model (16). While the reasons for the attenuation remain under investigation and the nature of the oxidant(s) that may trigger SoxR activation under physiological conditions is unknown, it is possible that the responses of the mouse to the infection (e.g., phagocyte recruitment and activation) lead to increased levels of O2·− and NO in host environments that result in activation of P. aeruginosa SoxR and, consequently, to an increase in the expression of the MexGHI-OmpD system. Such host environment-induced MexGHI-OmpD expression may be required to produce a programmed alteration in quorum-sensing signal homeostasis that eventually leads to global physiological changes needed for adaptation to the host and virulence. Although the proposed regulation of the P. aeruginosa quorum-sensing network in response to O2·− and, possibly, NO or other oxidative species from the host requires further investigation, the data presented here suggest that the bona fide physiological role of P. aeruginosa SoxR may be to act as a key sensor-transducer in such a regulatory pathway.
This work was supported by Cystic Fibrosis Foundation grant QUADRI00V0. L.Q. is a scholar of the Stavros S. Niarchos Foundation. The Department of Microbiology and Immunology at Weill Medical College of Cornell University acknowledges the support of the William Randolph Hearst Foundation.
We thank J. Xiang for advice with the array analysis.
Editor: A. D. O'Brien