Pseudomonas aeruginosa causes serious acute and chronic infections in humans. Major differences exist in disease pathogenesis, clinical treatment and outcomes between acute and chronic infections. P. aeruginosa acute infection characteristically involves the type III secretion systems (T3SS) while chronic infection is often associated with the formation of biofilms, a major cause of difficulties to eradicate chronic infections. The choice between acute and chronic infection or the switch between them by P. aeruginosa is controlled by regulatory pathways that control major virulence factors and genes associated with biofilm formation. In this study, we characterized a hybrid sensor kinase PA1611 that controls the expression of genes associated with acute and chronic infections in P. aeruginosa PAO1. Expression of PA1611 completely repressed T3SS and swarming motility while it promoted biofilm formation. The protein PA1611 regulates two small RNAs (sRNAs), rsmY and rsmZ which in turn control RsmA. Independent of phosphate relay, PA1611 interacts directly with RetS in vivo. The positive effect of RetS on factors associated with acute infection could presumably be restrained by PA1611 when chronic infection conditions are present. This RetS–PA1611 interaction, together with the known RetS–GacS interaction, may control disease progression and the lifestyle choice of P. aeruginosa.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Pseudomonas aeruginosa is a major pathogen, causing a range of infections in humans. These include pneumonia and urinary tract infection, bloodstream infection, and infection in wounds and in burn patients. P. aeruginosa is one of the most common nosocomial pathogens creating serious problems in health care associated situations. P. aeruginosa pulmonary infection is also the major cause of morbidity and mortality in cystic fibrosis (CF) patients (Lyczak et al., 2000; 2002). The infections caused by P. aeruginosa can be either acute or chronic. Substantial differences exist between these two types of infections in terms of disease pathogenesis, treatment options and clinical outcome. Specific virulence mechanisms, especially the type III secretion system (T3SS), are required for acute infection and the transition to chronic infection is often accompanied by the formation of bacterial biofilm communities, which is the main cause of the difficulty in eradicating P. aeruginosa chronic infections (Fleiszig et al., 1997; Parsek and Singh, 2003; Hauser, 2009).
In the lungs of CF patients the initial infection is of an acute nature and transition to chronic infection occurs when disease progresses, probably in response to the CF lung environment (Goodman et al., 2004; Wagner and Iglewski, 2008). Evidence indicates that during the early stage of infection P. aeruginosa produces virulence factors that are associated with acute infection while the persistence of long-term chronic infection is dependent on the regulated development of biofilms and the production of a different set of virulence factors (Goodman et al., 2004). Chronic infection is often associated with selection for mutations that confer a chronic phenotype however regulatory switching between lifestyles is not well known in P. aeruginosa. After a chronic infection is established, the CF lung is permanently colonized by P. aeruginosa despite extensive antibiotic treatments (Lyczak et al., 2000). Investigating the transition from acute to chronic infection is fundamental for understanding disease progression and these transition pathways may provide a potential target for preventing the establishment of chronic infection.
Two-component system (TCS) signalling pathways are widespread in bacteria and archaea and are also found in unicellular eukaryotes and higher plants (Wolanin et al., 2002). Pathogens often rely on TCS pathways to monitor environmental stimuli and translate these signals into adaptive responses. It has recently been shown that the development of acute versus chronic infection or the switch between the two is controlled by regulatory pathways composed of TCSs in P. aeruginosa. The first-level key regulatory components involved in such transition are the hybrid sensor kinases LadS and RetS which sense yet unknown environmental signals. RetS is required for the expression of T3SS genes but it represses genes involved in biofilm formation such as genes responsible for biofilm matrix production (Goodman et al., 2004; Laskowski et al., 2004; Zolfaghar et al., 2005; Laskowski and Kazmierczak, 2006). In contrast, LadS has a negative impact on T3SS genes but a positive effect on biofilm formation (Ventre et al., 2006). Thus, RetS and LadS are proposed to control the choice of acute and chronic infection reciprocally (Ventre et al., 2006; Goodman et al., 2009). Both RetS and LadS proteins contain internal domains encoding a sensor histidine kinase and response regulator-like receiver domains (Ventre et al., 2006); but neither LadS nor RetS sensor kinases are genetically linked to a typical response regulator. Both RetS and LadS intersect with the second-level TCS, the GacS/GacA system where GasS is the sensor kinase and GacA is the response regulator. While the mechanism of how LadS interacts with GacS/GacA is still unknown, it has been shown that RetS inhibits GacS activity by forming a RetS/GacS heterodimer, blocking the autophosphorylation of GacS and phosphorelay from GacS to GacA (Goodman et al., 2009). Only phosphorylated GacA activates the expression of two small RNAs (sRNAs), rsmY and rsmZ (Kay et al., 2006; Brencic et al., 2009). These small RNAs are antagonists of the RNA binding regulator RsmA, blocking RsmA function by titration. RsmA is the central post-transcriptional regulator which can regulate pathogenicity determinants such as motility, biofilm formation and T3SS in direct or indirect manner (Pessi et al., 2001; Heurlier et al., 2004; Burrowes et al., 2006; Mulcahy et al., 2006; Brencic and Lory, 2009; Irie et al., 2010). In addition to RetS and LadS, the histidine phosphotransfer protein HptB has also been found to modulate the sRNA levels. This protein together with the response regulator PA3346 and the putative anti-anti-σ factor PA3347 form a regulatory cascade in P. aeruginosa (Lin et al., 2006; Hsu et al., 2008). The HptB signalling pathway appears to exclusively control the expression of rsmY and this control is GacS/GacA-dependent (Bordi et al., 2010).
In this study, we identified a hybrid sensor kinase PA1611 in P. aeruginosa whose expression completely repressed T3SS genes and swarming motility but promoted biofilm formation. Independent of phosphorelay pathways, PA1611 controlled the expression of the two sRNAs, rsmY and rsmZ and consequently the function of RsmA. We report that the hybrid sensor kinase PA1611 plays an important role in controlling acute and chronic infection transition through interacting directly with RetS. The activation of genes and pathways associated with acute infection by RetS could presumably be counteracted by PA1611 when chronic infection is preferred.
Identification of a P. aeruginosa mutant with decreased exoS expression
Type III secretion system plays an important role in P. aeruginosa acute infection and alterations in T3SS expression often determine changes in disease progression. As part of a broader attempt to investigate the regulatory mechanisms that govern P. aeruginosa disease progression, we screened the P. aeruginosa genome for genes that affect T3SS expression. A transposon-insertion-mutant library of P. aeruginosa PAO1 carrying the reporter exoS-lux on its chromosome was constructed and mutants with changed promoter activity of T3SS effector gene exoS were collected. One mutant strain, E7 was isolated which exhibited decreased exoS promoter activity while having a similar growth rate when compared with that of the wild-type PAO1 (Fig. 1A). Sequence analysis showed that the transposon inserted into gene PA1612, at a site 332 bp upstream of the translation start codon (ATG) of the adjacent gene PA1611. PA1612 encodes a hypothetical protein while PA1611 encodes a hybrid sensor kinase.
To verify that the mutation in PA1612 was responsible for the decreased exoS expression, we generated a PA1612 mutant of PAO1 by inserting a Gmr-lacZ cassette into this gene. The promoter activity of exoS was then measured in the mutant. Unexpectedly, the promoter activity of exoS in the PA1612 mutant remained virtually the same as that in the wild-type PAO1 (data not shown). Neither did the deletion of the downstream gene PA1611 result in the phenotype of the transposon insertion mutant E7. In fact, the level of exoS promoter activity was actually increased in the PA1611 mutant when compared with that in the wild type (Fig. 1B).
Expression of PA1611 inhibits type III secretion and swarming motility but promotes biofilm formation
Considering the fact that the transposon used in this study harbours an outward directed tac promoter, we presumed that the decreased exoS expression in the transposon mutant could be due to enhanced expression of the downstream gene PA1611 by the transposon-carried promoter rather than the inactivation of PA1612 or PA1611. Interestingly, the unique regulators QteE and the Roc TCS were also identified respectively using the same transposon when they were overexpressed by the tac promoter on the transposon (Kulasekara et al., 2005; Liang et al., 2011). To test such a possibility, PA1611 was cloned into pAK1900 under the control of a lac promoter. The resulting plasmid, pAK-1611, was transferred into the wild-type PAO1, yielding the strain PAO1(pAK-1611). The promoter activity of exoS was then measured in this PA1611 expression strain. As shown in Fig. 2A, when PA1611 was expressed on the plasmid the promoter activity of exoS was completely abolished.
In addition to exoS, the promoter activity of other T3SS genes, exoY, exoT, exsCEBA and exsD-pscB-L, was also obliterated in the PA1611 expression strain (data not shown) by using the lux-based reporter. The effect of PA1611 on T3SS was confirmed by examining the secretion of T3SS effectors under T3SS inducing conditions (Fig. 2B). The wild-type PAO1 secreted proteins of approximately 53 kDa and 49 kDa, corresponding to ExoT and ExoS respectively. The mutant strain ΔexoT and ΔexoS did not have the respective band corresponding to ExoT and ExoS. However, the PA1611 expression strain did not produce any of these secreted proteins, displaying a phenotype identical to that of T3SS defective strain ΔexsA. The ExoT and ExoS production were both increased slightly in the PA1611 mutant when compared with that in PAO1. These results demonstrate that the expression of PA1611 in P. aeruginosa had extensive and dramatic effects on both the transcription and protein production of T3SS.
To define the scope of PA1611 regulation, its effect on biofilm formation and bacterial motility was examined. Elevated expression of PA1611 resulted in the loss of the swarming motility of P. aeruginosa (Fig. 3A) and a corresponding increase in biofilm formation, measured by a static crystal violet assay (Fig. 4A). The swarming motility was increased in the PA1611 mutant compared with PAO1 (Fig. 3B). Decreased swimming and twitching motilities were also observed in the PA1611 expression strain (data not shown). All these phenotypes are consistent with those previously reported in the retS mutant (Goodman et al., 2004), and both RetS and PA1611 are hybrid sensor kinases. To analyse further the similarity and differences between PA1611 and RetS, a retS mutant was generated in a PAO1 background. The effect of retS mutation and that of PA1611 expression on T3SS, together with bacterial motility and biofilm formation, were compared. The results obtained indicate that the expression of PA1611 in P. aeruginosa had a very similar effect on these phenotypes as the retS deletion (Figs 2, 3 and 4A).
The hyper-biofilm phenotype in retS mutant is linked to the over-production of exopolysaccharides (Bordi et al., 2010). To determine whether this was also the case with the PA1611 expression strain, polysaccharide production was tested by Congo red staining. Both the retS mutant and PA1611 expression strains grown on plates containing Congo red dye displayed similarly stronger staining than the wild type (Fig. 4B), suggesting that both strains produced substantially more polysaccharides.
There are 12 genes in the P. aeruginosa genome which are predicted to encode hybrid sensor kinases (Hsu et al., 2008). To verify that the regulatory effect of PA1611 is specific, instead of being a general influence of any hybrid sensor kinases, two other hybrid sensor kinases, PA1243 and PA2824 together with LadS were tested, and the effect of the expression of these genes on T3SS was measured. As shown in Fig. S1, the expression of PA1243 or PA2824 had no effect on exoS expression. The expression of ladS on the plasmid in PAO1 resulted in repression of exoS, although the effect was weaker than PA1611 expression. This effect of ladS was somewhat expected because LadS is known to have an opposite effect than RetS which is required for the expression of T3SS genes (Ventre et al., 2006). Clearly, like RetS and LadS, PA1611 could play a key role in regulating T3SS, bacterial motility and biofilm formation, hallmark differences between acute and chronic infections.
PA1611 expression diminishes cytotoxicity
As T3SS is a central virulence determinant of P. aeruginosa, enabling the bacteria to penetrate eukaryotic cells (Frank, 1997; Coburn et al., 2007; Hauser, 2009) and as the expression of PA1611 in P. aeruginosa blocked T3SS gene expression, we carried out experiments to determine the impact of PA1611 expression on P. aeruginosa cytotoxicity. The LDH release assay was used to monitor bacterial cytotoxicity to murine mammary carcinoma cells EMT6. The PA1611 expression strain together with the wild-type strain and the other control strains were used to infect EMT6. Four hours post infection, the extent of LDH release was measured. As presented in Fig. 5, the LDH release of EMT6 cells infected by both retS and exsA mutants was akin to the EMT6 cell spontaneous release without bacterial infection. The wild-type strain and ΔPA1611 were able to mediate apparent cytotoxic response. Like the retS and exsA mutants, PA1611 expression strain exhibited markedly less cytotoxicity than the wild-type strain and ΔPA1611. These results demonstrate that PA1611 expression could indeed attenuate the cytotoxicity of P. aeruginosa, further confirming its control over the T3SS.
Expression of PA1611 could bypass the HptB/PA3346/PA3347 signalling pathway
Previous studies have shown that three sensor kinase hybrids including PA2824, PA1976 and PA1611 could transfer phosphate onto the response regulator PA3346 through the histidine phosphotransfer protein HptB in PAO1 (Lin et al., 2006; Hsu et al., 2008). To determine whether the observed regulation of T3SS (Fig. 2) and biofilm formation (Fig. 4) by PA1611 is mediated by the HptB/PA3346/PA3347 signalling pathway, we constructed mutations in PA3346 and hptB and the promoter activity of exoS was measured in these mutants. Importantly, the expression of PA1611 still had the same effect on exoS promoter activity in the PA3346 mutant background (Fig. 6A). Results of swarming motility assay further confirmed the inability of PA3346 mutation to affect PA1611 function (Fig. 6B). There was a decrease in exoS promoter activity in the hptB mutant compared with the wild-type PAO1, but the decreases in exoS promoter activity was much less than in the PA1611 expression strain (Fig. S2). Just like PA3346, mutation in hptB was also unable to affect the function of PA1611 (Fig. S2). These observations indicate that expression of PA1611 could bypass the HptB/PA3346/PA3347 signalling pathway.
The residues in classical phosphorelay are not required for the function of PA1611
PA1611 harbours a conserved histidine residue H275 in the histidine kinase domain and an aspartate residue D565 in the response regulator receiver domain (Fig. 7A). To test whether the classical phosphorelay residues were required for the function of PA1611, we introduced H275L and D565A single amino acid residue substitutions and double residue substitutions into PA1611. These derivatives of PA1611 were expressed in PAO1 and the promoter activity of exoS was measured. As shown in Fig. 7B, these mutated PA1611 derivatives functioned as well as the wild type in terms of its effect on exoS promoter activity. Apparently, the canonical phosphorelay residues, as is the case of RetS (Goodman et al., 2009), were not required for PA1611 function. The results from cytotoxicity assays also indicated that the mutations in these two conserved phosphorelay residues did not affect the cytotoxicity of the strains expressing these derivatives when compared with the strain expressing wild-type PA1611 (data not shown), further indicating that the function of PA1611 was independent of phosphorelay.
PA1611 functions through the GacS/GacA-RsmY/Z signalling pathway
Previous studies showed that the GacS/GacA-RsmY/Z signalling pathway is required for RetS to regulate downstream target genes (Goodman et al., 2004; Brencic et al., 2009; Bordi et al., 2010). To test whether the function of PA1611, like RetS, was dependent on the GacS/GacA system, we investigated the regulation of PA1611 in a gacA mutant background. A gacA mutant of PAO1 was constructed and exoS promoter activity was compared in the gacA mutant and PAO1 with and without the presence of the PA1611 expression plasmid pAK-1611. As shown in Fig. 8A, without pAK-1611, the level of exoS promoter activity, as expected, was much higher in the gacA mutation compared with the wild type. But the expression of PA1611 had no effect on exoS expression in a gacA mutant background in contrast to the observation that it completely abolished exoS expression in the wild-type strain. Similarly, PA1611 expression had no effect on swarming motility in the gacA mutant (Fig. 8B). These results indicate that, like retS, PA1611 exerts its regulation on the target genes through GacA.
To determine whether the function of PA1611 requires the other known regulator LadS, a ladS mutant was constructed, pAK-1611 was introduced into the mutant, and the promoter activity of exoS was measured. No change in PA1611 regulation was observed in the PA1611 expression strain whether ladS was intact or not (data not shown), suggesting that the function of PA1611 does not require LadS.
It has been shown that the GacS/GacA signal transduction system works exclusively through its control over the transcription of the regulatory small RNAs, RsmY and RsmZ (Brencic et al., 2009). Using the lux-based reporter, the impact of PA1611 on rsmY and rsmZ was therefore investigated. As shown in Fig. 9, higher promoter activity of both rsmY and rsmZ was observed in the PA1611 expression strain when compared with the control PAO1(pAK1900). The promoter activity of rsmY and rsmZ was slightly lower in the PA1611 deletion mutant compared with PAO1 (data not shown). Together, these results illustrate that PA1611 functions through the GacS/GacA-RsmY/Z signalling pathway, causing changes in the expression of both rsmY and rsmZ regulatory small RNAs.
PA1611 interacts with RetS directly
The results presented above showed that the conserved phosphorelay residues were not required for the function of PA1611; hence PA1611 probably operates through a pathway unrelated to phosphorelay pathways. Because PA1611 expression strain displayed similar phenotypes to the retS mutant and RetS directly interacts with GacS (Goodman et al., 2009), we speculated that the expressed PA1611 protein may interact with and sequester RetS, thus leaving more free GacS protein to activate the Gac/Rsm pathway. To test this hypothesis, we carried out a bacterial two-hybrid assay to detect the possible interaction between PA1611 and RetS. The kinase core HisKA/H+ ATPase domains of PA1611, RetS and GacS were cloned respectively into bacterial two-hybrid expression vectors to create C-terminal fusions to λ cI or RNAPα. Interaction between these sensor proteins would allow the λ cI to bind to the operator region and recruit RNAPα to initiate transcription of the HIS3 and aadA reporter genes, allowing the reporter strain to grow on dual selective screening plates lacking histidine and supplemented with streptomycin. As shown in Fig. 10, interactions were detected between the hybrid sensor kinases PA1611 and RetS as reflected by the significant bacterial growth. Consistent with previous reports, interaction between RetS and GacS was also detected. No interaction was detected between PA1611 and GacS. In addition, the mutations in the conserved histidine phosphorelay residue in PA1611 did not affect the interaction between PA1611 and RetS (Fig. 10). These results strongly suggest that the hybrid sensor kinase PA1611 directly interacts with RetS in vivo via the kinase core HisKA/H+ ATPase domains and such interaction does not depend on the conserved phosphorelay residue.
Isolation of mutants with elevated expression of PA1611
The effect of PA1611 was most prominent when it was expressed on a plasmid. Using the lux-based reporter, we tested the promoter activity of PA1611 and found that PA1611 was expressed at low levels in the wild-type PAO1 under normal laboratory conditions (Fig. 11A). Like retS and other TCSs, the lingering question is regarding the identity of the signals to which PA1611 responds. In an attempt to answer such a question, we tested the promoter activity of PA1611 in a variety of conditions. Different media conditions (Table S3) were set up and PA1611 expression levels measured. Although variations in the expression levels were observed, PA1611 expression was generally low under all these tested laboratory conditions (data not shown).
As phosphorelay seems to play no significant role in PA1611 function, a non-typical pathway may be involved in signal transduction. In addition to the classical signal transduction through ligand binding, we assumed that there might be gene products enabling PA1611 function simply by controlling its expression. Another round of whole-genome mutagenesis was carried out and mutant that affected the promoter activity of PA1611 was screened. Out of the approximately 30 000 mutants generated, several mutants that dramatically influenced PA1611 promoter activity were obtained. Sequence analysis revealed that the transposon inserted into the gene PA1775 at different sites in these mutant strains. PA1775 encodes a conserved cytoplasmic membrane protein CmpX with unknown function (Winsor et al., 2011). As shown in Fig. 11A, disruption of PA1775 by transposon insertion dramatically activated the expression of PA1611. Motility test in the PA1775 mutant (C36) showed that the inactivation of PA1775 also inhibited swarming motility (Fig. 11B), an effect observed in PA1611 expression strain. To confirm the function of PA1775, a complementation study was carried out. The intact PA1775 gene was cloned into pAK1900 and the constructed plasmid was then transformed into PA1775 transposon mutant C36. As shown in Fig. 11, the promoter activity of PA1611 and the swarming motility were both restored to the wild-type level. Based on these results, we speculated that the environmental signal which PA1611 responds to may come first to PA1775. In other words, PA1775 may be the outpost of the PA1611 regulation over P. aeruginosa infection transition.
The hallmark differences between P. aeruginosa acute and chronic infection are T3SS expression and the activity of genes associated with biofilm formation (Goodman et al., 2004). P. aeruginosa chronic infections associated with biofilms are almost impossible to eradicate. CF pulmonary infection with P. aeruginosa advances from the initial acute phase to the later chronic infection phase. Once a chronic infection is established, P. aeruginosa persists permanently in the lungs of these patients (Lyczak et al., 2000; 2002). Pulmonary exacerbations occur as intermittent episodes during which the symptoms of lung infection increase and the function of the lung decreases. CF pulmonary exacerbations cause irreversible loss of lung function, ultimately leading to pulmonary failure and mortality. Although it is predictable that CF exacerbations arise probably due to changes of patient condition or/and the infecting agents in the lung, the actual causes of exacerbations remain to be completely explained. Clinical prevention of such exacerbations is therefore extremely challenging. Our previous observations prompted us to speculate that the changes of gene expression including T3SS expression in P. aeruginosa influenced by other bacteria may be a cause of CF exacerbation (Duan et al., 2003). Data from experiments using a Drosophila melanogaster killing model and a rat lung infection model seem to support such a possibility (Duan et al., 2003; Sibley et al., 2008). It is possible that transient acute infections signified by the activated expression of virulence factor genes may exist during exacerbations as a result of the interactions among the host, the pathogen and the environmental factors in the lung. Transitions between acute and chronic infection may, therefore, occur in both directions, at least in some situations.
The expression of the hybrid sensor kinase PA1611 on a plasmid abolished T3SS expression and swarming motility while it promoted biofilm formation, indicating that PA1611 potentially plays an important role in regulating transitions between acute and chronic infection. PA1611 seems to counteract the regulatory function of RetS by directly interacting with RetS; such a direct interaction is demonstrated by the bacterial two-hybrid protein interaction assay. It has been shown that RetS regulates the expression of its target genes in an unorthodox way, by the novel mechanism of direct interaction with another sensor kinase GacS (Goodman et al., 2009). The GasS/GacA system works in the classical way by which GacS autophosphorylates and transfers the phosphate group to the response regulator GacA. Unlike the typical TCSs, phosphorelay is not involved in the RetS–GacS interaction. The switch or the choice of acute and chronic infection is largely controlled by the interaction between RetS and GacS on the pathogen side. During the acute virulence phase RetS forms heterodimers with GacS, preventing GacS from forming homodimers and hence blocking GacS phosphorelay; whereas during the transition to chronic infection RetS and GacS each form homodimers in response to unknown signals, allowing phosphorylated GacS to initiate downstream phosphorelay and gene regulation. Similar to the RetS–GacS interaction, the interaction between PA1611 and RetS does not require phosphorelay. Bioinformatics analysis of the predicted protein indicates that PA1611 shares similar domains with GacS, including the HAMP domain, HisKA/H+ ATPase domain and response regulator receiver domain (Fig. 7A), even though GacS has the Hpt domain while PA1611 does not. It is possible that PA1611 and RetS interact in a similar way as RetS and GacS.
Two situations have been proposed in the working model of RetS and GacS interaction. In one situation (acute infection) RetS binds GacS and in the other situation (chronic infection) the binding is affected in response to environmental cues (Goodman et al., 2009). Despite their importance, the environmental cues and the mechanisms that affect the RetS and GacS interaction remain to be identified. The signals which trigger the activation of RetS are unknown. While it is possible that the RetS and GacS interaction is affected directly by environmental signals, the presence of PA1611, another RetS-interacting hybrid sensor kinase, suggests that the RetS and GacS interaction may be altered by this new player. It is tempting to speculate that the induced expression of PA1611 in response to a changed environment would trigger interactions between RetS and PA1611, resulting in the release of GacS from GacS–RetS heterodimers to form GacS homodimers. The GacS homodimers are able to phosphorylate GacA which in turn blocks RsmA function through activation of rsmY/Z (Fig. 12). Because RetS and GacS are constitutively coexpressed under most conditions (Goodman et al., 2009), the expressed RetS may activate the factors associated with acute infection under most conditions. When chronic infection occurs in response to changed conditions, the expression of PA1611 could potentially shift the balance of RetS and GacS interaction and curtail the already active RetS regulon. Clearly, further studies are required to verify such a possibility.
The observed function of PA1611 is similar to that of LadS, another key regulator in regulating the switch between acute and chronic infection. However, LadS seems to function through mechanisms distinct from those of RetS and PA1611. The observations that rsmZ expression is similarly de-repressed in a retS mutant and a retS–ladS double mutant indicate that LadS activity is not required for RetS function, but they instead function independently (Ventre et al., 2006). Consistent with the divergence of the LadS pathway, genes regulated by RetS only partially overlap the genes regulated by LadS (Ventre et al., 2006). Modelling the periplasmic sensor domain DISMED2 (Diverse Intracellular Signalling Module Extracellular 2) of LadS and RetS also reveals potential ligand binding differences which suggest distinct underlying mechanisms (Vincent et al., 2010). Similarly, PA1611 does not function through LadS. Deletion of ladS did not affect the regulatory function of PA1611 (data not shown), demonstrating that the function of PA1611 is independent of LadS. The effect of PA1611 expression is almost identical to that in the retS mutant, which confirms the direct interaction between RetS and PA1611 and the deactivating effect of PA1611 on RetS function.
Like RetS, PA1611 has conserved phosphorelay residues in its histidine kinase and response regulator receiver domains, suggesting that PA1611 is able to participate in phosphotransfer reactions. Indeed, it has been shown that PA1611 and two other hybrid sensor kinases (PA1976 and PA2824) can phosphorylate the histidine phosphotransfer protein HptB in P. aeruginosa PAO1 (Lin et al., 2006; Hsu et al., 2008). Previous research also has shown that the hptB mutant exhibits a hyper-biofilm phenotype and downregulated T3SS in P. aeruginosa PAK strain (Bordi et al., 2010). Although the mutants of hptB and retS display similar phenotypes, the HptB and RetS pathways are distinct. Introduction of a PA3346 or PA3347 mutation in the hptB mutant abolish its hyper-biofilm phenotype. In contrast, introduction of a PA3346 or PA3347 mutation in the retS mutant has no influence on biofilm formation. RetS has much stronger effects on regulating both T3SS and biofilm formation than HptB, and HptB signalling only controls expression of rsmY whereas RetS signalling modulates both rsmY and rsmZ expression (Bordi et al., 2010). Similarly, our data showed that conserved phosphorelay residues in the histidine kinase and receiver domains of PA1611 were not essential for the function of PA1611. Mutation of PA3346 did not affect the phenotypes of the PA1611 expression strain, indicating that PA1611 could bypass the HptB signalling pathway to function when its expression was activated. Consistence with such differences, a recent study revealed that the HptB/PA3346/PA3347 signalling pathway regulates the σ28-dependent motility genes through a partner switching mechanism (Bhuwan et al., 2012).
It has been reported that the hybrid sensor kinases PA1976 (ErcS') and PA2824 (SagS) are respectively involved in controlling aerobic ethanol oxidation and biofilm formation in P. aeruginosa (Mern et al., 2010; Petrova and Sauer, 2011). Although PA2824 also participates in the biofilm formation process, it functions in a manner different from PA1611. PA2824 can directly interact with BfiS, the sensor kinase of the TCS BfiS/BfiR (Petrova and Sauer, 2009), and modulation of the phosphorylation state of BfiS in a growth-mode-dependent manner (Petrova and Sauer, 2011). Inactivation or overexpression of PA2824 does not affect the swarming, swimming or twitching motilities in P. aeruginosa (Petrova and Sauer, 2011). Our data also confirmed that overexpression of PA1243 or PA2824 had no effect on exoS expression (Fig. S1). Despite the fact that PA1611 together with PA1976 and PA2824 could phosphorylate the HptB protein, the results from the literature and our study suggest they may play different regulatory roles through distinct pathways in P. aeruginosa. Although the effect of PA1611 on the transition between acute and chronic infection does not seem to require the HptB pathway, PA1611 still can function through the phosphorelay in the HptB pathway. PA1611 potentially represents a unique sensor kinase hybrid that plays regulatory roles with different strength, through two independent pathways.
PA1611 is expressed at low levels in all the tested laboratory conditions but its effect is most prominent when it is expressed on a plasmid. Identification of the environmental cues or ligands that activate PA1611 will provide further understanding of the mechanism of transition between acute and chronic infection and lifestyle choice of P. aeruginosa. While the search for direct activating cues for PA1611 activation will continue, the data showing that the expression of PA1611 was activated by mutation of PA1775 points to an alternative route of signal transduction. It is possible that the bacterium responds to environmental changes through a regulator of PA1611. Considering that the transition between acute and chronic infection is probably a result of P. aeruginosa responding to many cues in the host environment, the regulatory network governing the transition may be a complex one. Further investigations are required to improve our understanding of P. aeruginosa disease progression and our ability to prevent or control it.
Strains, plasmids and growth conditions
Bacterial strains and plasmids used in this study are described in Table S1. P. aeruginosa and Escherichia coli were routinely grown on Luria–Bertani (LB) agar or in LB broth at 37°C unless otherwise specified. LB was used as a T3SS non-inducing medium and LB supplemented with 5 mM EGTA and 20 mM MgCl2 as a T3SS inducing medium (calcium-deplete). Antibiotics were used at the following concentrations: for P. aeruginosa, gentamicin (Gm) at 50 μg ml−1 in LB or 150 μg ml−1 in Pseudomonas isolation agar (PIA), tetracycline (Tc) at 70 μg ml−1 in LB or 300 μg ml−1 in PIA, carbenicillin (Cb) at 250 μg ml−1 in LB and trimethoprim (Tmp) at 300 μg ml−1 in LB; for E. coli, kanamycin (Kn) at 50 μg ml−1, ampicillin (Ap) at 100 μg ml−1, chloramphenicol at 25 μg ml−1, Tc at 15 μg ml−1 and Gm at 15 μg ml−1 in LB.
Construction of gene expression detecting systems
The plasmid pMS402 carrying a promoterless luxCDABE reporter gene cluster was used to construct promoter-luxCDABE reporter fusions as reported previously (Duan et al., 2003). Promoter regions of rsmY, rsmZ, exsCEBA, exsD-pscB-L and PA1611 were PCR-amplified using high-fidelity Pfu DNA polymerase (Fermentas) and primers designed according to the PAO1 genome data (Winsor et al., 2011). The primers used are listed in Table S2. The primer pair used for rsmY is designated as pKD-rsmY-S (sense) and pKD-rsmY-A (antisense); for rsmZ, they are pKD-rsmZ-S and pKD-rsmZ-A and so on. The promoter regions were cloned into the BamHI–XhoI site upstream of the lux genes on pMS402. The resultant plasmids pKD-rsmY, pKD-rsmZ, pKD-exsC, pKD-exsD and pKD-1611 were transformed into P. aeruginosa respectively by electroporation. Cloned promoter sequences were confirmed by DNA sequencing.
Besides the plasmid-based reporter system, an integration plasmid CTX6.1 originating from plasmid mini-CTX-lux (Becher and Schweizer, 2000) was used to construct chromosomal fusion reporters. This plasmid has all the elements required for integration, the origin of replication, and a tetracycline-resistance marker. The pMS402 fragment containing the kanamycin-resistance marker, the MCS, and the promoter-luxCDABE reporter cassette was then isolated and ligated into CTX6.1. The plasmid generated was first transferred into E. coli SM10-λ pir (Simon et al., 1983) and the P. aeruginosa reporter integration strain was obtained using biparental mating as reported previously (Hoang et al., 2000; Liang et al., 2008).
Using these lux-based reporters, gene expression in liquid cultures was examined as counts per second (cps) of light production using a Synergy 2 Multimode Microplate Reader (BioTek). Expression was measured every 30 min for 24 h. Bacterial growth was monitored at the same time by measuring the OD600 in the Microplate Reader.
The transposon mutagenesis library was constructed as previously described (Kulasekara et al., 2005). Briefly, the donor strain E. coli SM10-λ pir containing pBT20 and the recipient P. aeruginosa PAO1 carrying the reporter exoS-lux on its chromosome were grown separately on solid media overnight and the cells were collected, resuspended in LB and spotted on a fresh LB agar plate at a ratio of 2:1. After incubation for 2 h, the mixed culture was diluted and spread on PIA containing 5 mM EGTA, 20 mM MgCl2 and 150 μg ml−1 gentamicin. After overnight incubation, mutants with altered exoS promoter activity, as indicated by changes in light production under a LAS-3000 imaging system (Fuji Life Science), were collected. Verification of the mutant phenotype was carried out in liquid cultures. The promoter activity of exoS in collected mutants was examined in a Synergy 2 Multimode Microplate Reader (BioTek). The sites of transposon insertion in the selected mutants were determined by arbitrary primed PCR and subsequent sequencing of the PCR products (Friedman and Kolter, 2004; Liang et al., 2008). The same procedures were followed to screen mutants with altered PA1611 promoter activity.
For construction of gene knockout mutants, the previously described sacB-based strategy was employed (Hoang et al., 1998). The DNA regions of the target genes were PCR-amplified using primers listed in Table S2. Restriction sites were incorporated into the primers to facilitate cloning. The PCR products obtained were digested with restriction enzymes; and then cloned into pEX18Tc. To construct the retS, PA1611 and PA3346 unmarked knockout mutant, DNA fragments of 1159 bp, 1209 bp and 309 bp within the individual PCR products were deleted by digesting with PstI, PstI or XhoI respectively. For other mutants, the DNA fragment containing the Gmr-lacZ from pZ1918-lacZGm (Schweizer, 1993) was inserted into the target genes. Gene knockout mutants were obtained using the triparental mating procedure in which the strain carrying the helper plasmid pRK2013 was used together with the donor and recipient (Ditta et al., 1980). The resultant mutants were verified by PCR.
Overexpression of hybrid sensor kinase genes
The multi-copy-number E. coli–P. aeruginosa shuttle vector pAK1900 (Poole et al., 1993) was used to overexpress hybrid sensor kinase genes in P. aeruginosa. The DNA regions covering the entire target genes were PCR-amplified. The primers used are listed in Table S2. The PCR products were digested with corresponding restriction enzymes, and then cloned into pAK1900. The plasmids constructed were then each transformed into P. aeruginosa by electroporation.
Generation of PA1611 point mutations
The construction of the H275L mutation in the sensor kinase domain of PA1611 was accomplished by PCR using oligonucleotide primers containing the point mutation and a XhoI restriction site. The XhoI restriction site was generated from the base alterations at H275 without affecting adjacent codons. The DNA fragment upstream of codon H275 was amplified using primers pAK-1611-S containing a SalI site and 1611H-A containing a XhoI site; the downstream region of H275 was generated using the primers 1611H-S containing a XhoI site and pAK-1611-A containing a HindIII site (Table S2). These two PCR fragments were cloned into the vector pCR2.1-TOPO in sequence and joined together. The resulting construct carries the H275L substitution mutation. To overexpress this mutated PA1611, the above construct was digested with XbaI and HindIII, and the XbaI–HindIII fragment carrying the mutated PA1611 was then cloned into pAK1900 and transferred to P. aeruginosa.
The D565A mutation in the receiver domain of PA1611 was constructed by using the In-Fusion HD Cloning Kit (Clontech). The upstream fragment was amplified by PCR using primers 1611D-S1 and 1611D-A1; the downstream region was generated using primers 1611D-S2 and 1611D-A2 (Table S2). The fragments were fused with pAK1900 by In-Fusion Enzyme and the aspartate D565 of PA1611 was replaced by alanine.
The double substitution mutations on PA1611 were generated by swapping the fragment containing H275L mutation with the wild-type fragment upstream of the D565A mutation using the BamHI restriction site located between the two mutations. All these mutations were verified by sequencing analysis.
Protein secretion measurement
Bacteria were grown in T3SS inducing conditions for 6 h at 37°C. Following the removal of bacterial cells by centrifugation at 14 000 g, proteins in the supernatant were precipitated following addition of an equal volume of 100% TCA, washed with acetone and pelleted. The proteins were resuspended in sample buffer, separated by SDS-PAGE and visualized with Coomassie blue stain (Goodman et al., 2004).
Bacterial motility examination
Bacterial motilities were assessed as described previously with slight modifications (Rashid and Kornberg, 2000). The media used for the swarming motility assay consisted of 0.5% agar, 8 g l−1 nutrient broth mix, and 5 g l−1 glucose. The media used for the swimming motility assay was tryptone agar containing 10 g l−1 tryptone, 5 g l−1 NaCl and 0.3% agar. LB broth solidified with 1% agar was used for the twitching motility assay. For swarming and swimming motilities, bacteria were spot inoculated onto plates as 2 μl aliquots taken directly from overnight LB cultures. For twitching motility, strains were stab inoculated with a sharp toothpick to the bottom of the Petri dish from overnight LB cultures. After inoculation, swarm plates were incubated at 37°C and swim plates at 30°C for 12–14 h, and twitch plates were incubated at 37°C for 24 h. Photographs were taken with the LAS-3000 imaging system.
Exopolysaccharide production assay
Tryptone (10 g l−1) agar (1%) plates supplemented with Congo red (40 μg ml−1) and Coomassie blue dyes (20 μg ml−1) were used. Bacteria were inoculated on the surface of the plates with a toothpick and grown at 30°C. The colony morphology and staining were recorded after 2 days (Bordi et al., 2010).
Biofilm formation quantification
Biofilm formation was measured as described by O'Toole and Kolter (1998) with minor modifications. Cells from overnight cultures were inoculated at 1:100 dilutions into M63 minimal medium supplemented with 0.2% glucose, 1 mM MgSO4 and 0.5% casamino acids in 96-well polystyrene microtitre plates (Costar) and grown at 30°C for 4 h and 6 h. A 100 μl volume of 1% crystal violet was added to each well and staining was allowed for 15 min prior to removal of the liquid by aspiration. Wells were rinsed three times with distilled water, and the remaining crystal violet was dissolved in 200 μl of 95% ethanol. A 125 μl portion of this solution was transferred to a new polystyrene microtitre plate, and the absorbance was measured at 600 nm.
Cytotoxicity was assayed by monitoring the release of the cytoplasmic enzyme lactate dehydrogenase (LDH). Murine mammary carcinoma cells (EMT6) were cultured in RPMI-1640 medium (HyClone) supplemented with 10% heat-inactivated fetal calf serum (YHJM) in a 5% CO2 humidified incubator at 37°C. For LDH assays, EMT6 cells were dispensed into a 96-well plate (Costar) at 104 cells per well. The cells were infected with P. aeruginosa at a multiplicity of infection of 100. After 4 h of infection, the extent of LDH release was assayed with the Cytotox96 Kit according the manufacturer's instructions (Promega).
Detection of protein–protein interaction by a bacterial two-hybrid system
Bacterial two-hybrid experiments were conducted by using the BacterioMatch II Two-Hybrid System Vector Kit (Agilent Technologies). PA1611 and PA1611 H275L subclones encoding from amino acid 202 to 505, and retS from amino acid 393 to 652 were cloned, respectively, at the 3′ end of the gene encoding λ cI repressor protein domain carried on pBT vector. Meanwhile, gacS subclone encoding from amino acid 219 to 520 and retS from amino acid 393 to 652 were cloned at the 3′ end of the gene encoding α-subunit RNA polymerase (RNAPα) domain on pTRG vector respectively. The primers used are listed in Table S2. For PA1611 and PA1611 H275L they were pBT-1611-S and pBT-1611-A; for retS: pBT-retS-S and pBT-retS-A; for gacS: pTRG-gacS-S and pTRG-gacS-A. The resulting gene fusion constructs, pBT-1611, pBT-1611H, pBT-retS, pTRG-gacS and pTRG-retS, were confirmed by DNA sequencing. The positive controls used were pTRG-GAl11P and pBT-LGF2. Derivatives of pBT and pTRG were co-transformed into E. coli XL1-BlueMRF' Kan cells and selected on selective screening plates and the positive colonies were verified by patching on dual selective screening plates according to the manufacturer's instructions.
We thank Dr Colin Dawes for critical reading of the manuscript and Saiprasad Pydi for help with the bioinformatics analysis of PA1611. This work was supported by grants from Natural Science and Engineering Council of Canada, Manitoba Medical Services Foundation and PCSIRT (No. IRT1174).