|Home | About | Journals | Submit | Contact Us | Français|
Colonization of the gastrointestinal tract is the first event in Klebsiella pneumoniae nosocomial infections, followed by colonization of the bladder or respiratory tract or entry into the bloodstream. To survive in the host, bacteria must harbor specific traits and overcome multiple stresses. OxyR is a conserved bacterial transcription factor with a key role both in the upregulation of defense mechanisms against oxidative stress and in pathogenesis by enhancing biofilm formation, fimbrial expression, and mucosal colonization. A homolog of oxyR was detected in silico in the K. pneumoniae sequenced genome and amplified from the LM21 wild-type strain. To determine the role of oxyR in K. pneumoniae host-interaction processes, an oxyR isogenic mutant was constructed, and its behavior was assessed. At concentrations lower than 107 ml−1, oxyR-deficient organisms were easily killed by micromolar concentrations of H2O2 and exhibited typical aerobic phenotypes. The oxyR mutant was impaired in biofilm formation and types 1 and 3 fimbrial gene expression. In addition, the oxyR mutant was unable to colonize the murine gastrointestinal tract, and in vitro assays showed that it was defective in adhesion to Int-407 and HT-29 intestinal epithelial cells. The behavior of the oxyR mutant was also determined under hostile conditions, reproducing stresses encountered in the gastrointestinal environment: deletion of oxyR resulted in higher sensitivity to bile and acid stresses but not to osmotic stress. These results show the pleiotropic role of oxyR in K. pneumoniae gastrointestinal colonization.
Bacterial nosocomial infections are an important cause of morbidity and mortality, and Klebsiella spp. account for up to 8% of them (39). Klebsiella pneumoniae is an enterobacterium mostly responsible for urinary tract infections, pneumonia, soft tissue infections, and bacteremia. K. pneumoniae strain isolates are frequently resistant to multiple antibiotics, which leads to a therapeutic dilemma. A better understanding of the physiological process of this bacterium therefore seems necessary to limit colonization and hence the infection process. The reservoir for K. pneumoniae strains is the gastrointestinal (GI) tract of patients (11, 45), and GI colonization depends on the ability of the bacteria to adhere to mucosal surfaces, to form biofilm within the mucus layer (32), and to resist the specific stresses encountered in the GI tract. Several adhesins have been characterized at the cell surface and on the fimbriae of K. pneumoniae and were shown to play a role in adhesion to intestinal epithelial cells (10, 13). We previously demonstrated that capsular polysaccharides were also involved in the GI colonization step (18), as well as other structures and functions, including membrane transport, the tripartite efflux pump, and metabolic pathways (8, 33). The ability of K. pneumoniae to form biofilm is also likely to play a role in GI colonization since sessile bacteria are more capable of resisting to GI stresses than their luminal counterparts. Several surface organelles and metabolic pathways are involved in biofilm formation by K. pneumoniae (1), some of which may confer the bacteria protection. In Escherichia coli, the presence of Ag43, a self-recognizing surface-displayed protein conferring autoaggregation of bacteria and involved in biofilm formation (9), provides a mechanism for reducing local oxygen concentrations, thereby protecting the cells from damage-causing oxidizing agents (43). The expression of flu, coding Ag43 is phase variable and tightly regulated due to concerted action of the Dam methylase and the transcriptional regulator OxyR (22).
OxyR, a member of the LysR regulator family, is involved in the regulation of virulence factors in several bacteria. In addition to its role in Ag43 expression in E. coli, OxyR contributes to early stages of biofilm formation in Serratia marcescens by influencing fimbrial gene expression; fimbriae possess adhesins that play a role in the initial attachment to epithelial cells (46). In E. coli, OxyR enhances the ability of the bacteria to colonize the urinary tract (27). However, OxyR was first reported to play a role in the oxidative stress response (24, 31, 44, 46, 47, 52). In E. coli, the OxyR transcription factor activates the expression of several antioxidant defensive genes in response to elevated levels of reactive oxygen species (ROS) such as hydrogen peroxide (H2O2) (20, 26). Since H2O2 is a significant component of the antimicrobial armament of human phagocytic cells, OxyR-mediated resistance represents a major defensive strategy during the infection process. The effect of ROS can be exacerbated by the presence of free iron, which can react with H2O2 to enhance oxidative stress via the Fenton reaction (51). In addition, an unusual phenotype has been associated with OxyR: diluted (<10−7 ml−1) oxyR-deficient Pseudomonas aeruginosa suspensions are unable to form single colonies on aerobic rich medium but can form single colonies on anaerobic rich medium (38), probably owing to the high levels of H2O2 in the medium.
In the present study, we identified an oxyR homolog in K. pneumoniae and investigated its role in several phenotypes associated with the process of GI colonization. Using a genetic approach, we discovered that oxyR was involved in biofilm formation, fimbrial synthesis, intestinal colonization, resistance to several GI stresses, and resistance to H2O2 exposure.
The microorganisms used in the present study are listed in Table Table1.1. All bacterial strains were stored at −20°C and −80°C in lysogeny broth (LB) medium containing 20% glycerol. Unless otherwise stated, strains were grown in LB broth or agar at 37°C for 18 to 24 h. The plasmid pDrive (PCR cloning kit; Qiagen, Courtaboeuf, France) was used for cloning experiments, and the plasmid vector pBK-CMV (Stratagene, Amsterdam, The Netherlands) was used for complementation. When needed, media were supplemented with relevant antibiotics: kanamycin (30 μg ml−1), tetracycline (30 μg ml−1), spectinomycin (50 μg ml−1), and streptomycin (50 μg ml−1). LM21CH774 strain (ΔoxyR mutant/pBK-CMV) was cultured under anaerobic conditions when plated onto agar for numeration. Bacterial growth was monitored by measuring the optimal density at 620 nm, and the CFU counts were quantified by plating serial dilutions of the suspension onto LB agar plates. Aerobactin production was monitored by a cross-feeding bioassay with E. coli LG1522 as the indicator strain on M9 minimal medium containing 200 μM 2,2′-dipyridyl (5). The experiment was performed twice.
The primers oxyR-F and oxyR-R were designed from the complete sequence of the oxyR of the K. pneumoniae MGH78578 genomic sequence (GenBank accession number CP000647) and were used to amplify a 918-bp fragment from K. pneumoniae LM21. Chromosomal DNA extraction and amplifications were performed by using standard protocols, and the sequences were obtained commercially from Genome Express (Meylan, France).
The oxyR-defective mutant was created by allelic exchange (2, 6) after replacement of the entire gene by the selectable spectinomycin resistance gene aadA7. The 1,031-bp aadA7 cassette was generated by using the primers oxyR-aad-5′ and oxyR-aad-3′ (Table (Table2)2) and the template MG1655 strain. The resulting aadA7 cassette flanked by 60-bp fragments, which are homologs of the oxyR 3′ and 5′ regions, was electroporated in the K. pneumoniae LM21 strain harboring the lambda-red protein-encoding plasmid pKOBEG199. Mutants were selected onto LB agar containing spectinomycin, and the loss of the pKOBEG199 plasmid was then checked on LB containing tetracycline. The deletion in the ΔoxyR mutant, designated LM21CH773, was further checked by PCR performed with primers oxyR-F and oxyR-R.
For complementation assays, the oxyR gene was amplified by using the primers oxyR-F and oxyR-R, ligated into the pDrive vector, and transformed into JM109. This recombinant vector was then HindIII and BamHI restricted, and the oxyR fragment was subcloned into the corresponding sites of the pBK-CMV vector. The recombinant pBK-CMVoxyR was then electroporated into both of the oxyR-defective mutants, giving rise to a transcomplemented strain designated LM21CH779 (ΔoxyR+pΔoxyR). Control strains (with the plasmid vector alone) were constructed by electroporating the native pBK-CMV vector into both the oxyR-defective mutant and the wild-type K. pneumoniae LM21. The resulting strains were, respectively, designated LM21CH774 (ΔoxyR mutant/pBK-CMV) and LM21CH775 (wild-type strain/pBK-CMV).
To examine bacterial growth in vitro, overnight cultures were diluted 1:100 (or 1:10,000) and subcultured for 7 h.
Disk diffusion assays were performed by spreading ~5 × 108 bacteria and placing a sterile 6-mm paper disk on the plate, to which 10 μl of 3% H2O2 was added, and incubating the preparation. Each experiment was repeated at least three times.
Capsular polysaccharides were extracted from overnight bacterial suspensions adjusted to ~108 cells ml−1 with Zwittergent 3-14 detergent. The amount of uronic acid was then measured according to the method described by Domenico et al. and Favre-Bonté et al. (14, 17). Each experiment was performed in triplicate.
All experiments were performed in triplicate in Dulbecco modified Eagle medium (Cambrex Bioscience, Paris, France) at 37°C. We used a slightly modified version of the microtiter plate assay developed by O'Toole and Kolter (37). Briefly, 4 μl of overnight culture were inoculated into 100 μl of Dulbecco modified Eagle medium in a 96-well culture-treated polystyrene microtiter plate (Nunc). Wells filled with growth medium alone were included as negative controls. After 24 h of incubation at 37°C, surface-adherent biofilm formation was measured by staining bound cells for 15 min with a 0.5% (wt/vol) aqueous solution of crystal violet. After a rinsing with distilled water, the bound dye was released from stained cells using 95% ethanol, and the optical density at 620 nm was determined.
Commercial baker's yeast (Saccharomyces cerevisiae) was suspended in phosphate-buffered saline (5 mg [dry weight] per ml). Equal volumes of yeast cell suspension and bacterial suspension from an overnight culture were mixed on a glass slide. Aggregation was monitored visually.
RNA was extracted from 1 ml of a 5-h culture by using the NucleoBond NucleoSpin NucleoTrap kit from Machery-Nagel according to the manufacturer's instructions. The real-time reverse transcription-PCR (RT-PCR) step was carried out with the LightCycler RNA Master SYBR green I kit (Roche Diagnostics, Mannheim, Germany) on a LightCycler system (Roche Diagnostics) with the primers MrkD2F and MrkD2R (19, 23) or fimA and inv (Table (Table2).2). As an internal control, the 16S rRNA gene was amplified with the universal primers TM1 and TM2. A 20-μl mixture containing 1 μl of RNA and 2 μl of primers was prepared and amplified according to the manufacturer's instructions (the RNA preparations were checked for absence of DNA contamination). To compare the relative gene expression between the wild-type strain and its transconjugants, the following ratio was determined: 2[(CPtransconjugant−CPwild type)control-(CPtransconjugant-CPwild type)gene] was calculated, where CP is crossing point. The experiment was done at least three times.
Female specific-pathogen-free mice (OF1; 8 to 18 weeks old; 20 g; Iffa Credo, L'Arbresle, France) were used. They were individually caged and given sterile water containing 5 g of streptomycin per liter throughout the experiment and had ad libitum access to feed. After 24 h, 100-μl bacterial suspensions containing ~108 CFU were given intragastrically to groups of five mice. On the first day and subsequently every day after inoculation, feces were collected and homogenized in saline, and serial dilutions were plated onto streptomycin agar. The experiment was performed twice.
Intestine-407 (Int-407) cells derived from human embryonic jejunum and ileum, and HT-29-MTX, a mucus-secreting cell line derived from a human colon carcinoma, were cultivated as described elsewhere, and adhesion to the different cell lines was assayed as described previously (17). These experiments were performed three times.
For the growth assay, 5 ml of LB broth with or without the addition of various concentrations of crude ox bile extract (Sigma; 1 and 10% [wt/vol]) or NaCl (0.25 and 1 M) was inoculated with cells from an overnight culture to reach 4 × 107 and 4 × 105 CFU ml−1. The cultures were incubated aerobically with shaking for 5 and 3 h, respectively, and the numbers of viable bacteria were determined. The results are expressed as the ratio of the number of CFU obtained from LB cultures containing different concentrations of bile or NaCl to the number of CFU obtained from control cultures (LB alone). These experiments were performed three times.
The behavior of Klebsiella strains under inorganic acid conditions was assessed according to the method previously described (8, 35). Briefly, overnight cultures were diluted to reach an optical density of 0.2. The bacterial cells from 1 ml of the suspensions were harvested by centrifugation (16,000 × g for 1 min at room temperature) and resuspended in 100 μl of LB broth (pH 7.0). Samples were divided into two aliquots: half of the cells were mixed with 950 μl of LB broth (unadapted cells), and the other half were mixed with 950 μl of LB broth (pH 4.0) (HCl) (adapted cells). After 1 h of incubation at 37°C with aeration, the adapted (pH 4.0) and unadapted (pH 7.0) cultures were pelleted at 16,000 × g for 1 min, and the cells were resuspended in 1 ml of LB broth (pH 3.0) (HCl). Viable cell counts were determined at t = 0 and after 60 min of incubation at 37°C. The number of surviving microorganisms was determined by dividing the total number of viable cells after the hour-long acid shock by the initial number of viable cells at t = 0, and the percentage of adaptation was determined by dividing the results of adapted assays by those of the unadapted assays. To compare relative acid tolerance in the wild-type and the oxyR mutant or the transcomplemented oxyR mutant, we calculated the ratio percentages of adaptation wild type/oxyR mutant or the transcomplemented oxyR mutant. These experiments were performed three times.
The GenBank accession number for the sequence determined in the present study is GQ228030.
The amino acid sequence deduced from the 918-bp DNA fragment obtained from K. pneumoniae LM21 shared 96% identity with the OxyR sequence of E. coli. This oxyR homolog shares 99 and 95% identity with the oxyR sequences of K. pneumoniae MGH78578 and K. pneumoniae 342, respectively. To determine the role of oxyR, an ΔoxyR mutant was created by allelic exchange in the wild-type K. pneumoniae LM21 and complemented in trans. Disk diffusion assay showed that the ΔoxyR mutant had significantly greater sensitivity to H2O2 (inhibition zone = 46 ± 1.0 mm) than the wild-type (inhibition zone = 24 ± 0.0 mm) and the transcomplemented (inhibition zone = 24.7 ± 0.6 mm) strains (wild-type/ΔoxyR mutant, P < 0.001; wild-type/transcomplemented, P = 0.091), thereby demonstrating that the response of K. pneumoniae oxyR is conserved in oxidative stress. In addition, the ΔoxyR mutant had a different colony morphology with smaller and less regular colonies than the wild-type and the transcomplemented strains (Fig. (Fig.11).
When diluted to ≥107 cells per ml in LB broth and incubated aerobically with shaking at 37°C, the growth rate of the ΔoxyR mutant was similar to that of the wild-type and the transcomplemented strains (Fig. (Fig.2A).2A). When diluted to 105 cells per ml, ΔoxyR mutant growth was slower than that of the wild-type strain and partial recovery was observed with the transcomplemented strain (Fig. (Fig.2B).2B). The ΔoxyR mutant was unable to form isolated colonies on LB medium under aerobic conditions (+O2) when the inoculum concentration was lower than 107 CFU/ml, as shown by spotting 5 μl of serially diluted overnight aerobic cultures onto LB agar plates (Fig. (Fig.3A).3A). In contrast, the wild-type and transcomplemented organisms grew well even starting with low inocula such as 103 cells/ml. When the same serial dilutions were plated under anaerobic conditions, the mutant grew as well as the wild-type strain (Fig. (Fig.3B3B).
When serial dilutions of ΔoxyR mutant organisms were spotted onto LB agar plates under aerobic conditions (+O2) in the absence of iron, colonies were observed even with a low inoculum (105 CFU ml−1) (Fig. (Fig.3C),3C), whereas an inoculum of 107 CFU ml−1 was required in the presence of iron (Fig. (Fig.3A).3A). The expression of the aerobactin iron uptake system was not impaired in all of these strains, since under iron-deprived conditions of growth, the wild-type strain and the ΔoxyR mutants were able to cross-feed LG1522 (data not shown).
Our next investigation was to determine whether there were any alterations in biofilm development as a result of the loss of oxyR. Using a microtiter plate experimental model, we showed that the biomass of the biofilm formed by the ΔoxyR mutant was significantly lower (wild type/ΔoxyR mutant, P < 0.001; wild type/transcomplemented, P = 0.292) than that of the wild-type and transcomplemented organisms (Fig. (Fig.4),4), suggesting that oxyR contributes to biofilm formation by promoting the attachment of bacteria to surfaces or between bacterial cells. Since type 3 pili is an adhesive factor facilitating adherence and biofilm formation on abiotic surfaces in K. pneumoniae (12, 29), we looked for pilus synthesis (both type 3 and type 1 pili) in the ΔoxyR mutant. When overnight cultures of the different strains were compared, we observed that the wild-type strain induced an agglutination of yeast cells, while the ΔoxyR mutant and the transcomplemented strains did not, suggesting a decrease in type 1 fimbrial synthesis (Table (Table3).3). Type 1 and type 3 fimbrial expressions were investigated by real-time RT-PCR by targeting a 614-bp fragment of the fim-encoding gene (7) and a 226-bp fragment of the mrkD-encoding gene (19). When considering the reference, the comparative CP values between the wild-type strain and the ΔoxyR mutant and between the wild-type and the transcomplemented strains were 0.2 and 0.4, respectively, indicating that fimbrial expression was downregulated in the ΔoxyR mutant, although not totally restored in the transcomplemented strain.
Since capsule, one of the main virulence factors of K. pneumoniae, plays a role in biofilm formation (1) and is potentially able to prevent cellular receptor-target recognition by bacterial adhesins (42), we investigated the amount of capsular polysaccharides produced by the different strains. The wild-type strain and the oxyR transcomplemented organisms harbored 20.6 ± 1.4 μg and 20.4 ± 0.9 μg of glucuronic acid in 0.5 ml of overnight culture, respectively, and this level was not significantly changed in the ΔoxyR mutant (17.8 ± 3.4 μg, P = 0.2), meaning that oxyR was probably not involved in capsule synthesis.
To investigate the influence of oxyR in K. pneumoniae intestinal tract colonization, the ΔoxyR mutant, the transcomplemented mutant and the parent strain were fed to mice individually, and the bacterial counts of K. pneumoniae in feces were recorded over 4 days. Although the wild-type strain efficiently colonized the murine intestinal tract day after day, the ΔoxyR mutant was detected only a few hours after feeding and not later (Fig. (Fig.5).5). No restoration of the intestinal colonization phenotype was observed with the transcomplemented mutant (data not shown).
The wild-type strain and the ΔoxyR and the transcomplemented mutants were also compared for their ability to adhere to Int-407 and HT-29 cells (Fig. (Fig.6).6). The ΔoxyR mutant was significantly less adherent than the wild-type strain, mounting 5-fold and a 60-fold lesser adhesions to Int-407 and HT-29 cells, respectively, with the transcomplemented mutant showing partial recoveries (wild type/ΔoxyR mutant, P < 0.01; wild type/transcomplemented, P = 0.243 and P < 0.01, respectively).
To determine the role of oxyR in intestinal colonization, bacteria underwent specific GI stress associated with bile, osmotic, and acid stresses. For the bile resistance assay, no significant difference was observed between the strains using high inoculum and either 1 or 10% of bile (data not shown). However, at low inocula and using bile extract concentrations as low as 1%, the ΔoxyR mutant showed a dramatic decrease in the percent survival of cells (Fig. (Fig.7)7) (wild type/ΔoxyR mutant, P < 0.01; wild type/transcomplemented, P = 0.123). No decrease was observed in the growth capacities of the ΔoxyR mutant in the presence of NaCl regardless of the inoculum and the salt concentration (data not shown).
In acid tolerance experiments with inorganic acid, bacteria were resuspended in LB broth, either pH 7.0 (unadapted) or 4.0 (HCl, adapted) and then in acid shock medium (LB [pH 3.0]). The percent survival of cells after 60 min acid shock was 5.91-fold (±0.89) lower in the ΔoxyR mutant than in the wild-type strain, with a level of recovery of the transcomplemented mutant similar to that obtained with the ΔoxyR mutant (6.36-fold [±0.44]).
The oxyR regulon has been detected in a wide variety of bacterial species, and its most commonly observed function is to detect elevated levels of ROS such as hydrogen peroxide and to regulate the expression of antioxidant genes that protect the bacterial cells from killing (52). ROS are produced when oxygen enters cells and oxidizes intracellular redox enzymes and when hydrogen peroxide reacts with iron via the Fenton reaction (51). The ability to survive oxidative stress may thus contribute to virulence because bacteria must be able to cope with the oxidative host defenses and persist in oxygenated tissues during the course of infection (47).
In K. pneumoniae, the oxyR gene was identified from the K. pneumoniae MGH78578 and sequenced from the clinical isolate LM21 DNA. The product of this latter gene seemed well conserved in all strains since it was amplified from 20 other clinical K. pneumoniae strains (data not shown). To investigate the role of oxyR in K. pneumoniae virulence, we constructed an oxyR-deficient mutant and complemented it in trans. As expected, K. pneumoniae was remarkably sensitive to H2O2 without oxyR, which was completely abolished in the transcomplemented strain. Bacteria lacking oxyR exhibited aerobic phenotypes, which means that only bacterial suspensions with initial inocula of ≥107 mutant bacteria per ml survived under aerobic conditions. This inoculum effect was first related in P. aeruginosa and assigned to LB broth composition, which was found to generate ~1.2 μM H2O2 via autoxidation, a level sufficient to kill serially diluted oxyR-deficient bacteria. When diluted, mutant organisms seemed unable to cope with such environmental oxidative stresses, whereas when not diluted they were able to do so. Likewise, when 5-μl portions of overnight K. pneumoniae aerobic cultures were serially diluted and spotted onto LB agar plates containing a ferrous iron chelator, inocula of ≥105 oxyR mutant bacteria per ml survived compared to an inoculum of ≥107 per ml without chelator. Free iron medium participates in the Fenton reactions, stimulating the formation of the destructive and short-lived OH· (4, 48), which increases oxidative stress. A bacterial suspension of 107 may represent the threshold for this H2O2 concentration, which would result in the growth of surviving bacteria whenever the initial concentration was higher. Another explanation would be that a high density of bacteria directly triggers an anti-oxidative stress response without OxyR contribution. One could hypothesize that, in the absence of OxyR, a quorum-sensing process that allows the bacteria to modulate gene expression according to changes in population density is responsible for bacteria O2 stress resistance, whereas OxyR-mediated response would be critical for the survival of planktonic individual organisms.
In comparison to the parental strain, the oxyR mutant formed smaller colonies on both Drigalski and LB agar. Colony morphology variants are frequently observed in P. aeruginosa and correspond to mucoid conversion, enhanced after contact with H2O2 (34). This mucoid conversion is indicative of an overproduction of a capsulelike polysaccharide in P. aeruginosa (15), but in our study no significant difference in capsule synthesis between the K. pneumoniae wild-type strain and its oxyR mutant was detected. However, other factors are known to be involved in colony morphology, such as the production of exopolysaccharides, pilus synthesis (3), lipopolysaccharide composition (30), or the expression of autotransporter proteins (49), which could be responsible for the colony variants observed with the oxyR mutant.
In addition to its role as anti-O2 factor, OxyR regulates several pathogenesis processes. This transcription factor is involved in the expression of pili in S. marcescens (46) and Porphyromonas gingivalis (50), the regulation of Ag43 biosynthesis in E. coli (22), and in biofilm formation in many bacteria (9, 24, 31, 44, 46). In the present study, we found a decrease in the expression of types 1 and 3 pilus genes for the K. pneumoniae oxyR mutant compared to the parental strain, as well as a decreased ability to form biofilm on static polystyrene microtiter plate. A previous study showed that type 3 fimbriae were involved in biofilm formation in K. pneumoniae (25). The restoration of the oxyR gene in trans did not increase pilus gene expression, whereas it restored the morphotype and biofilm formation, suggesting that the influence of oxyR on morphology and biofilm was independent of fimbria synthesis. Further understanding would involve identifying target genes regulated by OxyR in both biofilm formation and fimbrial synthesis.
Since biofilm formation occurred within the GI tract (32), we looked at the role of oxyR in GI colonization in a murine model. When oxyR mutants were fed to mice, bacteria were not found after the first day, suggesting that unlike the wild-type strain, they went through the intestinal tract without colonizing. Both low bacterial concentrations and local O2 pressure may explain the impact of this phenotype. Using a similar animal colonization model with the wild-type K. pneumoniae LM21, Favre-Bonté et al. detected bacterial concentrations of 1 × 104 and 1.35 × 106 CFU/g of contents in the jejunum and ileum, respectively (18), meaning that K. pneumoniae was present in the GI tract at concentrations lower than the critical threshold for ΔoxyR organisms. Although the intestine is commonly thought to be anaerobic, the tissues surrounding the lumen are oxygen rich, and oxygen diffuses into the intestine at appreciable levels. Whole-animal measurements gave oxygen tensions in the 2 to 7% air saturation range for the mouse colon and so the intestine was considered microaerobic (21, 28), which make it difficult for the oxyR mutant to grow in the GI tract.
The ability of microorganisms to initiate colonization depends also on the ability to adhere to mucosal surface and intestinal cells in the GI tract. The oxyR K. pneumoniae mutant was defective in adhesion in vitro to Int-407 cells and to HT-29 cells, two intestinal epithelial cells, indicating that oxyR is likely to play a role in GI colonization by K. pneumoniae by enhancing intestinal adhesion. Such a phenomenon has already been described in E. coli, in which oxyRS mutants were outcompeted by the wild-type strain in a mouse model of ascending urinary tract infection (27). Since oxyR was shown to be involved in types 1 and 3 fimbrial expression, we could speculate that the impaired adhesion observed is related to a defect in fimbrial expression in the oxyR mutant. However, other surface-displayed proteins could also be regulated by OxyR since adhesion deficiency, unlike fimbrial expression, was partially transcomplemented.
During intestinal colonization, bacteria are successively exposed to acid pH in the stomach, to the detergentlike activity of bile, to a decreasing oxygen supply, and to the presence of multiple metabolites produced by the normal gut microflora (41). K. pneumoniae is able to survive in hostile conditions, reproducing stresses encountered in the intestinal tract (33). The behavior of the wild-type strain and the oxyR-deficient mutants was therefore investigated under some of these hostile conditions to further understand how oxyR interacts in the GI tract. In the presence of NaCl, no difference was observed between the wild type and the oxyR mutant in the experimental conditions tested. The oxyR mutant was more sensitive to bile salts than the wild-type strain at low inocula, and its level of adaptation to inorganic acid was severely impaired. It is likely, therefore, that the OxyR factor controls functions required for survival of bacteria in the upper parts of the GI, where they encounter both high acidity and bile salts in a microaerobic environment.
In conclusion, we showed that oxyR was not only necessary for K. pneumoniae to resist O2 stress but also required to colonize the GI tract and was involved in the expression of several phenotypes classically associated with mucosal colonization. Effective transcomplementation of the ΔoxyR mutant with a multicopy plasmid was observed only with some of the phenotypes studied and did not occur in vivo (murine GI tract colonization) nor in pilus synthesis and resistance to acid tolerance. Although we cannot exclude instability of the plasmid without antibiotic selection in the animal model, this partial transcomplementation could also be due to impaired stoichiometry of the OxyR regulator within the cells, which would affect the expression of some phenotypes. It is noteworthy that in our construct the expression of oxyR was not controlled by its own promoter, which therefore rules out any autoregulatory effect of OxyR potentially involved in some aspects of the regulatory process. Several authors have related difficulties in complementing oxyR mutants. Sund et al. reported that multiple copies of oxyR had a deleterious effect on cells (47), and only modest levels of complementation were obtained with a plasmid construct by Mukhopadhyay and Schellhorn (36) in E. coli. The regulatory mechanism of OxyR is still unclear; a fuller understanding will probably shed new light on our findings, but it nevertheless remains a good candidate as a therapeutic target against K. pneumoniae colonization and subsequent infection.
We thank Laure Fauchery for providing technical help.
Editor: J. N. Weiser
Published ahead of print on 28 September 2009.